First assessment of microplastic consumption in sympatric black bass species in a temperate river system

Abstract

Microplastics (i.e. plastic particles ranging in size from 1 μm to 5 mm) are an understudied emerging pollutant in rural rivers that may pose significant health risks to organisms, and their monitoring is resource‐intensive and costly. This study quantifies microplastic consumption in the Monongahela River, West Virginia, USA in three sympatric black bass species: largemouth bass (Micropterus nigricans), smallmouth bass (Micropterus dolomieu) and spotted bass (Micropterus punctulatus). Utilizing an existing sampling program and low‐cost methodology, the spatial variability of microplastic ingestion between navigational pools within the Monongahela River and limited temporal variability within a pool was examined to determine differences between the sympatric species in total and shape‐size fractions. Microplastics were found in all fish surveyed and Bayesian regression models revealed that microplastic consumption is influenced by species, spatial location and temporal factors, with significant interspecies and intraspecies differences in microplastic consumption. The results emphasize the variability that exists even among similar species and reinforce the need to monitor microplastic pollution in rural river systems to evaluate the ecological risks of microplastic pollution.

1. INTRODUCTION

Microplastics (i.e. plastic particles ranging in size from 1 μm to 5 mm, including both primary and secondary plastics; Rochman et al., ²⁰¹⁹) have infiltrated aquatic and terrestrial ecosystems around the globe and pose increasing health risks to humans (Petersen & Hubbart, 2021; Sincihu et al., ²⁰²³; Vince & Hardesty, ²⁰¹⁸). Freshwater riverine ecosystems, especially those in rural areas (Liro et al., 2023; Mihai, ²⁰²³), are vulnerable to microplastic pollution because of land use practices in their drainage areas and microplastic retention (van Emmerik et al., ²⁰¹⁹; van Emmerik & Schwarz, ²⁰²⁰; Wickasono et al., ²⁰²¹). However, the extent of microplastic pollution and its impact on freshwater fisheries, especially in rural systems that are already at risk due to anthropogenic activities (He et al., 2021; Warren Jr & Burr, ¹⁹⁹⁴), is largely unknown.

While previous studies have shown that riverine dynamics determine spatiotemporal distribution and accumulation of microplastic in the ecosystem (Kallenbach et al., 2022; Watkins et al., ²⁰¹⁹), their consumption by aquatic organisms is influenced by an organism's feeding behaviour and life history (Vendel et al., 2017; McNeish et al., ²⁰¹⁸). Fish can directly consume microplastics by mistakenly ingesting plastic particles (Churchill & Bettoli, 2015). They can also be consumed indirectly through trophic transfer or accidental ingestion (Nelms et al., 2018; Parker et al., ²⁰²²). However, this is highly debated, and variability in these findings can be attributed to uncertainties introduced when fish are grouped by trophic level as well as differences in methodology (Au et al., 2017; Bellasi et al., ²⁰²⁰; O'Connor et al., ²⁰²⁰; Bowen et al., ²⁰²¹; McNeish et al., ²⁰¹⁸; Vendel et al., ²⁰¹⁷). Microplastic ingestion impacts vary by life stage, with sublethal effects on fish impacting numerous body systems and growth (Foley & Parnell, 2018; Huang et al., ²⁰²¹; Luo et al., ²⁰²²). Studying sympatric species complexes (i.e. groups of genetically distinct species with overlapping spatial distributions and similar morphology, habitat and life‐history traits) can help determine whether ecologically similar species experience the same levels of microplastic exposure (Aliko et al., 2022; Di Benedetto & Ramos, ²⁰¹⁴; Messinetti et al., ²⁰¹⁸).

The rivers of West Virginia, the second‐most rural state in the United States, are a sensitive biodiverse region, home to 311 fish, 172 crayfish and 136 mussel species (Cincotta & Welsh, 2024; Rice et al., ²⁰¹⁹). Sport fishing for largemouth bass (Micropterus nigricans), smallmouth bass (Micropterus dolomieu) and spotted bass (Micropterus punctulatus) is prevalent in this area. Thus, the West Virginia Division of Natural Resources routinely monitors this sympatric species complex, black bass. Unfortunately, numerous anthropogenic stressors, including fishing, dynamiting and industrial pollution, impact these fish populations. To our knowledge, the potential impact of microplastic pollution has yet to be quantified in the fish or rivers of West Virginia. Previous studies on largemouth and smallmouth bass in other state rivers have shown that black bass species consume microplastics (Hurt et al., 2020; Toner, ²⁰²⁰; Wagner et al., ²⁰¹⁹). Microplastic consumption may vary among individual black bass due to differences in feeding behaviour and diet breadth, as well as among populations because of environmental conditions or prey availability. These factors suggest that the impact of microplastic pollution may vary on a finer scale, with different populations and individuals exhibiting distinct consumption patterns (Brown, 2022; Easton et al., ¹⁹⁹⁶; Schindler & Pert, ¹⁹⁹⁷; Smith & Page, ¹⁹⁶⁹; Stepien et al., ²⁰⁰⁷; Taylor et al., ²⁰¹⁹). Quantifying microplastics at such fine spatial and temporal scales using high‐resolution, microplastic‐specific techniques requires intensive in situ sampling that is both costly and time‐consuming. As such, utilizing a low‐cost approach such as visual identification paired with a hot needle test enables microplastic monitoring while continuing to address current management priorities and thus makes such additional monitoring feasible with limited funds and time.

This study assesses microplastic consumption, shape and size composition variability in three sympatric species (i.e. largemouth, smallmouth and spotted bass) in the Monongahela River. Data were collected by partnering with an existing sampling program to determine potential differences due to navigational pool (i.e. space) and time as well as the potential for variation due to species and diet. We hypothesized that spatially overlapping black bass species would consume similar amounts and compositions of microplastics because they are sympatric species complexes. However, this is expected to fluctuate between pools due to locks and dams that have been shown act as barriers to microplastic flow generating sinks for fibres and fragments, potentially impacting bioavailability (Watkins et al., 2019) as well as temporally, potentially reflecting changes in diet and/or hydrological conditions. Hydrological conditions can be highly variable due to seasonal conditions and discharge has been shown to impact the concentrations of other pollutants in the Monongahela River (Wilson & Van Briesen, 2013).

2. METHODS

The Monongahela River is a 210 km, fourth‐order tributary of the Ohio River, draining a 7340 square mile basin in West Virginia and Pennsylvania (Figure 1). It encompasses six Hydrologic Unit Code (HUC) 8 watersheds, three of which contribute to this study area: the Tygart Valley, West Fork and Monongahela. The Monongahela features a series of locks and dams that enhance navigability but alter its natural hydrological regime and define navigational pools within the river. The three navigational pools that were sampled are the Opekiska, Morgantown and Point Marion. The Opekiska Pool spans from the headwaters to the Opekiska Lock and Dam, the Morgantown Pool from the Hildebrand to the Morgantown Lock and Dam, and the Point Marion Pool from the Morgantown Dam to the state border. The Morgantown Pool and parts of the Point Marion Pool encompass the built sprawl of Morgantown, the second‐largest city in the state. Wastewater treatment plants, stormwater, agricultural runoff, urban debris sources and other effluent sources may influence the amount of microplastic in Monongahela River, a river with a notable trash issue, both directly and via numerous tributaries (Upper Monongahela River Association, 2025).

FIGURE 1.

(a) Land cover in three Hydrologic Unit Code (HUC) 8 wastershed districts (i.e. Monongahela, West Fork and Tygart Valley River) that feed into the Monongahela River, West Virginia, USA. Boundaries of the HUC 8 districts are shown in black. State boundaries are white. (b) Location of dams along the Monongahela River (brown circles). The Opekiska Pool spans from the Opekiska Dam to the Hildebrand Dam. The Morgantown Pool spans from Hildebrand Dam to Morgantown Dam. The Point Marion Pool spans from Morgantown to the state line to Point Marion, PA. Morgantown (black star), a significant population centre as well as other towns (black diamonds) are indicated.

2.1. Ethics Statement

The WV Department of Natural Resources collected all biological samples used in this study, adhering to applicable national and local regulations. No live animals were handled or subjected to experimentation by the authors, therefore ethical approval was not required for this study. The collection and use of the samples complied with all relevant animal welfare and environmental guidelines as enforced by the WV Department of Natural Resources.

2.2. Sample collection

Three sympatric species of black bass (i.e. largemouth, smallmouth and spotted bass) were collected by the West Virginia Division of Natural Resources using night‐time, pulsed direct current (DC)‐mounted boat electrofishing in the Point Marion pool between 12 and 15 September 2019 and in all navigational pools between 19 and 28 May 2020 (Table 1). Each river segment sampling occurred in 10‐ to 15‐minute transects at four to six sites representative of black bass habitat consisting of tributary mouths, embayments and dam tailwaters. After each transect, species were morphometrically identified, recorded, given a unique identifying number and fish were humanely euthanized by West Virginia Division of Natural Resources personnel in accordance with agency protocols and approved animal care guidelines. Fish were frozen until processed.

TABLE 1.

The number of fish collected of each species in each navigational pool during each monitoring trip by the West Virginia Division of Natural Resources.

Pool	Month and year	Smallmouth bass (N)	Largemouth bass (N)	Spotted bass (N)
Opekiska	May 2020	11	17	9
Morgantown	May 2020	17	12	20
Point Marion	September 2019	40	24	28
Point Marion	May 2020	14	13	15

Note: Pools are listed from upstream to downstream within the Monongahela River.

At West Virginia University, fish were thawed and then weighed and measured (total length, TL in mm). The gastrointestinal (GI) tract was removed by an incision anterior to the vent along the ventral side of the fish to the clavicle for each fish. The opercular cavity was opened and the gill arches were carefully removed to expose the oesophagus, stomach and GI tract, which were then removed as a single intact structure. After removal, GI tracts were stored individually in labelled Whirl‐Pak bags and re‐frozen until microplastic processing. Whirl‐Pak bag contamination potential was measured. Results are reported in the Supporting Information.

2.3. Stomach content and microplastic processing

On thawing, a control for airborne microplastic contamination was established (i.e. air control), and viscera was transferred to a clean glass Petri dish. Details of this procedure are described in the Supporting Information. Dietary analysis was conducted broadly with the abundance of crayfish and fish consumed enumerated. The stomach was opened, and its contents were identified and enumerated into crayfish and fish. When sets of crayfish claws were present, each set of claws was assumed to belong to a single individual. Incomplete spinal columns were also assumed to each come from a single individual.

After stomach contents were examined, a 10% KOH weight/volume solution was added to the petri dish to dissolve the GI tract and stomach contents (Hajovsky, 2019; Karami et al., ²⁰¹⁷). This solution was used to dissolve the contents as previous research has shown that it results in minimal degradation of plastic polymers (Karami et al., 2017). The air control and sample were then closed and stored for ~24 h at 40°C in a Fisher Scientific Isotemp Oven Model 655F to ensure a maximum recovery rate and minimal degradation. The sample and air control were opened before filtering, and the sample was rinsed using Ultrapure Water into a metal sieve tower with mesh sizes of 500, 300, and 212 μm to segregate microplastics by size (Estahbanati & Fahrenfeld, 2016; Karami et al., ²⁰¹⁷). Sieves were removed from the tower, washed into the receiving basin of a vacuum filtration device using Ultrapure water, and filtered onto cellulose 47 mm 11 μm Whatman Grade 1 Qualitative Filter Paper. The paper was removed from the device using stainless‐steel forceps and stored in closed glass Petri dishes. Once all sieves were washed and filtered, the air control was closed, and Petri dishes were stored in a Fisher Scientific Isotemp Oven Model 655F to dry and aid in identification.

After drying for ~24 h, samples were visually inspected under a dissecting microscope (Kotar et al., 2022). Microplastic particles were identified using morphological characteristics (Hidalgo‐Ruz et al., 2012; Figure 2). When needed, plastic particles were moved to a compound microscope for further examination. Counts of microplastic particles by sieve size (i.e. 500, 500–300 and 300–212 μm) and shape (i.e. fibre, foam, fragment and sheet) were recorded and will be referred to as their shape‐size fraction (e.g. fibre >500 μm, fibre 500–300 μm, fibre 300–212 μm). All suspected microplastics underwent the hot needle test and were touched with a heated probe (Beckingham et al., 2023; Figure 2). In a single‐blind trial of researchers applying different hot point conditions to a set of synthetic, semi‐synthetic and natural fibres, synthetic and some natural fibres (i.e. the most often difficult to identify microplastic) were correctly identified >70% of the time, making it a reasonable alternative to more expensive methods such as Fourier transform infrared (FTIR) spectroscopy, Raman spectroscopy and pyrolysis gas‐chromatography with mass spectrometry (Py‐GCMS) (Beckingham et al., 2023).

FIGURE 2.

Representative microplastic particles observed at 40× magnification under light microscopy. (a) and (b) Fragments found within the specimens sampled. (c) Foam, (d) fibre and (e) sheets are also shown. (f) Microplastic sheet from (e) after the hot needle test.

2.4. Data analysis

Each species' mean, median, maximum and minimum total plastic counts are reported for each navigational pool during each month. Response counts, total microplastics and shape‐size fractions were cleaned of potentially extreme influencing values by removing outliers greater than the first and third quartiles. Because microplastic counts were highly skewed and zero‐inflated, mean and median values are often uninformative, therefore we report ranges, individual‐level observations and modelled rate ratios to capture variation among species, pools and size‐shape classes.

Bayesian regression models were built using the R package ‘brms’ with a Poisson distribution and independent weakly informative normal priors in R statistical software, Angel Food Cake 4.3.3 (Bürkner, 2017) to allow each coefficient to vary independently of the others and for independent derivation of posterior distributions. To determine intraspecies variation in microplastic consumption between pools during May 2020, total microplastic counts by species and shape‐size fractions were treated as response variables with the pool as a predictor variable:

2.4.1. Base model

$${\mathrm{X}}_{\mathrm{i}}\sim \text{Poisson}\left({\mathrm{\lambda }}_{\mathrm{i}}\right),$$

$$\log \left({\mathrm{\lambda }}_{\mathrm{i}}\right)={\mathrm{\beta }}_0+{\mathrm{\beta }}_{\text{pool}}·{\text{Pool}}_{\mathrm{i}},$$

$${\mathrm{\beta }}_0~\text{Normal}\left(0,10\right),{\mathrm{\beta }}_{\text{pool}}~\text{Normal}\left(0,10\right).$$

where X is the observed count, i is the species (i.e. largemouth, smallmouth or spotted bass) and λ is either the total microplastic expected count or shape‐size fraction expected counts (e.g. count of fibres >500 μm). To evaluate whether there was a significant difference between pools, a hypothesis test was performed on the difference in their regression coefficients. Specifically, we tested the null hypothesis that the effect of the Opekiska pool is equal to the effect of the Point Marion pool on the expected total microplastic count and expected shape‐size fraction using the hypothesis function in the brms package, which provides a Bayesian test for the specified contrast between species effects:

$${\mathrm{H}}_0:{\mathrm{\beta }}_{\text{pool Opekiska}}-{\mathrm{\beta }}_{\text{pool Point Marion}}=0$$

The base model was expanded to include the individual's diet (i.e. both the abundance of crayfish and fish) as explanatory variables due to their hypothesized significance to microplastic ingestion, but the model structure was conserved.

2.4.2. Expanded model

$${\mathrm{X}}_{\mathrm{i}}\sim \text{Poisson}\left({\mathrm{\lambda }}_{\mathrm{i}}\right),$$

$$\log \left({\mathrm{\lambda }}_{\mathrm{i}}\right)=\begin{array}{l}{\mathrm{\beta }}_0+{\mathrm{\beta }}_{\text{pool}}·{\text{Pool}}_{\mathrm{i}}+{\mathrm{\beta }}_{\text{Abundance}\_\text{Crayfish}}·\text{Abundance}\_{\text{Crayfish}}_{\mathrm{i}}\\ +{\mathrm{\beta }}_{\text{Abundance}\_\text{Fish}}·\text{Abundance}\_{\text{Fish}}_{\mathrm{i}},\end{array}$$

$$\begin{array}{l}{\mathrm{\beta }}_0~\text{Normal}\left(0,,,10\right),{\mathrm{\beta }}_{\text{pool}}~\text{Normal}\left(0,,,10\right),{\mathrm{\beta }}_{\text{Abundance}\_\text{Crayfish}}~\text{Normal}\left(0,,,10\right),\\ {\mathrm{\beta }}_{\text{Abundance}\_\text{Fish}}~\text{Normal}\left(0,,,10\right).\end{array}$$

The same model structure was applied to explore interspecies differences in expected total microplastic counts and expected shape‐size counts within pools (i.e. i is the navigational pool). Temporal variation within the Point Marion pool was also explored with the model structure above (i.e. i is the month). These models were also expanded to include diet while conserving model structure. Temporal variation was explored with the model structure above. However, i represent the month in the Point Marion pool. All models were run with 20 chains, each with 100,000 iterations and a warmup period of 10,000 iterations, to allow for model convergence of the Markov Chain Monte Carlo chains. All Bayesian models were run at a high‐performance computing facility using an EPYC node, 96‐core and 480 GB at the University of Maine Advanced Computing Group to handle high computational demands efficiently. All reported rate ratios and predicted values from the Bayesian models represent expected microplastic counts.

Leave‐one‐out cross‐validation (LOO) was used to compare base and expanded models. A model is determined to represent the data better when the difference in the expected log pointwise predictive density exceeds 2, and the standard error on the log pointwise predictive density supports the conclusion by not dropping below the predictive density threshold. Rate ratios, which allow for the interpretation of the relative differences in event rates, calculated as the exponentiated value of the coefficient, were used to assess the effect of each predictor. The 95% credible interval of the rate ratios is reported, which conveys the range of plausible event rates.

3. RESULTS

While the number of individuals collected varied slightly among species, pools and months (Table 1), all species were represented in comparable numbers, allowing for meaningful comparisons of microplastic ingestion across species and sites. Additionally, the size distribution of collected fish remained comparable between species and between months, minimizing the potential impact of ontogenetic shifts in diet and habitat on our findings (Table 1 and Figure S1).

All fish examined contained microplastics, although the size classes and shapes varied among individuals (Figure 2 and Table 2). Fragments and sheets were observed in multiple species (Table 2), but were generally less frequent than foam in some cases. Foam was primarily found in largemouth bass and, to a lesser extent, in smallmouth bass, with counts varying seasonally (e.g. absent from SMB in May). When present, foam occurred across a range of microplastic sizes, but mean and median counts for some species were zero, reflecting its absence in many individuals.

TABLE 2.

Microplastic counts (mean, median [range]) for each species by shape and size in each navigational pool during each monitoring trip by the West Virginia Division of Natural Resources.

Pool	Month and year	Species	Shape	>500 μm	500–300 μm	300–212 μm
Opekiska	May 2020	Largemouth bass	Fibre	2, 1 [0–5]	1, 1 [0–9]	2, 1 [0–13]
			Foam	6, 1 [0–40]	5, 2 [0–39]	6, 2 [0–41]
			Fragment	3, 0 [0–36]	2, 0 [0–24]	3, 0 [0–34]
			Sheet	1, 0 [0–3]	0, 0 [0–2]	1, 0 [0–3]
		Smallmouth bass	Fibre	2, 2 [0–6]	3, 2 [0–9]	3, 1 [0–11]
			Foam	0, 0 [0–0]	0, 0 [0–0]	0, 0 [0–0]
			Fragment	0, 0 [0–2]	1, 1 [0–6]	2, 1 [0–8]
			Sheet	3, 1 [0–21]	3, 2 [0–9]	5, 3 [0–18]
		Spotted bass	Fibre	1, 0 [0–4]	1, 1 [0–5]	1, 1 [0–3]
			Foam	2, 0 [0–10]	0, 0 [0–2]	0, 0 [0–3]
			Fragment	0, 0 [0–0]	1, 0 [0–2]	0, 0 [0–2]
			Sheet	0, 0 [0–1]	1, 0 [0–5]	0, 0 [0–3]
Morgantown	May 2020	Largemouth bass	Fibre	0, 0 [0–3]	0, 0 [0–2]	1, 0 [0–5]
			Foam	3, 0 [0–10]	4, 2 [0–20]	3, 0 [0–13]
			Fragment	0, 0 [0–2]	0, 0 [0–2]	1, 0 [0–5]
			Sheet	2, 1 [0–9]	1, 0 [0–7]	2, 0 [0–9]
		Smallmouth bass	Fibre	5, 2 [0–22]	4, 4 [1–9]	5, 4 [0–13]
			Foam	0, 0 [0–2]	0, 0 [0–1]	0, 0 [0–2]
			Fragment	1, 0 [0–11]	1, 0 [0–4]	2, 1 [0–6]
			Sheet	1, 0 [0–10]	2, 1 [0–7]	3, 1 [0–11]
		Spotted bass	Fibre	1, 0 [0–4]	1, 0 [0–4]	1, 0 [0–5]
			Foam	2, 0 [0–16]	2, 0 [0–16]	2, 0 [0–11]
			Fragment	0, 0 [0–2]	1, 0 [0–5]	0, 0 [0–4]
			Sheet	2, 0 [0–19]	2, 1 [0–10]	2, 1 [0–12]
Point Marion	May 2020	Largemouth bass	Fibre	2, 2 [0–5]	1, 1 [0–3]	1, 1 [0–5]
			Foam	12, 9 [0–52]	12, 9 [0–40]	8, 3 [0–30]
			Fragment	0, 0 [0–2]	1, 0 [0–5]	1, 0 [0–5]
			Sheet	1, 0 [0–6]	0, 0 [0–4]	0, 0 [0–2]
		Smallmouth bass	Fibre	1, 0 [0–6]	3, 2 [0–9]	4, 4 [0–7]
			Foam	0, 0 [0–3]	0, 0 [0–0]	0, 0 [0–0]
			Fragment	3, 2 [0–11]	4, 2 [0–23]	4, 3 [0–19]
			Sheet	1, 0 [0–7]	3, 2 [0–9]	3, 2 [0–16]
		Spotted bass	Fibre	1, 0 [0–10]	2, 0 [0–9]	1, 1 [0–6]
			Foam	3, 0 [0–22]	2, 0 [0–23]	2, 0 [0–22]
			Fragment	1, 0 [0–11]	3, 0 [0–33]	1, 0 [0–6]
			Sheet	0, 0 [0–1]	3, 2 [0–9]	0, 0 [0–1]
Point Marion	September 2019	Largemouth bass	Fibre	2, 0 [0–11]	2, 1 [0–7]	2, 2 [0–10]
			Foam	1, 0 [0–12]	2, 0 [0–44]	2, 0 [0–47]
			Fragment	4, 0 [0–26]	3, 2 [0–13]	3, 2 [0–11]
			Sheet	3, 2 [0–10]	3, 2 [0–11]	3, 2 [0–20]
		Smallmouth bass	Fibre	1, 1 [0–16]	2, 1 [0–8]	2, 1 [0–8]
			Foam	1, 0 [0–10]	0, 0 [0–10]	0, 0 [0–9]
			Fragment	2, 0 [0–24]	3, 0 [0–27]	3, 1 [0–18]
			Sheet	6, 2 [0–54]	4, 1 [0–36]	5, 1 [0–36]
		Spotted bass	Fibre	3, 1 [0–54]	3, 2 [0–13]	3, 2 [0–10]
			Foam	0, 0 [0–5]	0, 0 [0–1]	0, 0 [0–0]
			Fragment	3, 0 [0–37]	3, 0 [0–29]	4, 1 [0–32]
			Sheet	2, 0 [0–14]	2, 0 [0–15]	1, 1 [0–5]

Note: Pools are listed from upstream to downstream within the Monongahela River.

3.1. Intraspecies variation between navigational pools

Significant variation was detected within species between navigational pools in the expected total microplastic count rate ratios. Rate ratios indicate smallmouth and spotted bass ingested fewer microplastics in the Opekiska and Point Marion pools (Table 3 and Figure 1). Conversely, rate ratios demonstrated that largemouth bass had higher ingestion of microplastics in the Opekiska and Point Marion pools (Table 3). LOO demonstrated that expanding the base model pool to include diet, crayfish and fish abundance did not enhance predictive performance for any of the three species.

TABLE 3.

Intraspecies variation in microplastic consumption between navigational pools based on Bayesian regression models listed by species and upstream to downstream.

Species	Comparison	Response variable	Rate ratio	95% CI	Direction of effect
Smallmouth bass	Opekiska vs. Morgantown	Total	19%	8%–26%	Lower
Smallmouth bass	Point Marion vs. Morgantown	Total	17%	8%–26%	Lower
Smallmouth bass	Opekiska vs. Morgantown	>500 μm fibre	46%	41%–886%	Higher
Smallmouth bass	Point Marion vs. Morgantown	>500 μm fibre	242%	42%–887%	Higher
Smallmouth BASS	Point Marion vs. Morgantown	>500 μm foam	148%	63%–285%	Higher
Largemouth bass	Opekiska vs. Morgantown	Total	36%	22%–54%	Higher
Largemouth bass	Point Marion vs. Morgantown	Total	43%	27%–63%	Higher
Largemouth bass	Opekiska vs. Morgantown	>500 μm fibre	232%	41%–793%	Higher
Largemouth bass	Point Marion vs. Morgantown	>500 μm fibre	243%	42%–893%	Higher
Largemouth bass	Point Marion vs. Morgantown	>500 μm foam	148%	63%–283%	Higher
Spotted bass	Opekiska vs. Morgantown	Total	23%	9%–23%	Lower
Spotted bass	Point Marion vs. Morgantown	Total	50%	42%–57%	Lower
Spotted bass	Point Marion Morgantown	300–212 μm sheets	90%	24%–100%	Lower

Note: Only statistically meaningful relationships are presented. Estimates represent the expected percentage of microplastic counts in each pool, derived from Poisson‐distributed models using the R package ‘brms’ with independent weakly informative normal priors. Credible intervals (95% CI) indicate uncertainty in the posterior estimates.

For smallmouth and largemouth bass, rate ratios indicated expected counts of >500 μm fibre were higher in the Opekiska pool and higher in the Point Marion pool than in the Morgantown pool (Table 3). Rate ratios indicated expected counts of foam >500 μm were also higher in the Point Marion pool than in the Morgantown pool for smallmouth and largemouth bass. Spotted bass only demonstrated one significant difference in the accumulation of shape‐size fractions between pools. For sheets of 300–212 μm in diameter, rate ratios indicate expected counts decreased in the Point Marion pool compared to the Morgantown pool (Table 3). The remaining comparisons of shape–size fractions between pools for each species showed no significant differences, as the rate ratio credible intervals overlapped. LOO indicated that expanding these intraspecific models to include diet did not improve the explanation of the variation in counts over the base model.

3.2. Interspecies variation within navigational pools

Rate ratios indicated that the expected counts of total microplastics differed between each sympatric black bass species within each navigational pool. Specifically, the rate ratios indicated that expected total microplastic counts were lower in spotted bass than smallmouth bass and lower in largemouth bass than smallmouth bass across all three pools. Expected total microplastic counts were higher in largemouth bass than in spotted bass.

In general, size‐shape fractions in stomach contents are similar across species within each pool (Table 4). However, fragments display a unique patterning in the size fractions between pools. In the Morgantown pool, rate ratios indicate that expected counts of fragments of 500–300 μm in size were 198% lower in largemouth bass than in smallmouth bass; however, there is a high degree of uncertainty due to wide credible intervals (Table 4). Smallmouth bass exhibits higher microplastic consumption across shapes and size fractions. For all shapes and size fractions, expanding these models to include diet did not improve the explanation of the variation in counts over the base model.

TABLE 4.

Based on Bayesian regression models, there is interspecies variation in microplastic consumption between species within pools.

Pool	Comparison	Response variable	Rate ratio	95% CI	Direction of effect
All	Spotted bass vs. smallmouth bass	Total	73%	62%–86%	Lower
All	Largemouth bass vs. smallmouth bass	Total	23%	5%–21%	Lower
All	Largemouth bass vs. spotted bass	Total	35%	30%–40%	Higher
All	Largemouth bass vs. spotted bass	>500 μm fibre	56%	35%–87%	Higher
All	Spotted bass vs. smallmouth bass	>500 μm fibre	63%	2%–160%	Lower
All	Largemouth bass vs. smallmouth bass	>500 μm foam	12%	5%–17%	Higher
All	Largemouth bass vs. spotted bass	>500 μm foam	50%	29%–84%	Higher
All	Spotted bass vs. smallmouth bass	>500 μm foam	24%	10%–56%	Lower
All	Spotted bass vs. smallmouth bass	500–300 μm fibres	313%	220%–448%	Lower
All	Spotted bass vs. smallmouth bass	300–212 μm fibres	256%	186%–353%	Lower
All	Largemouth bass vs. smallmouth bass	500–300 μm fibres	467%	309%–718%	Lower
All	Largemouth bass vs. smallmouth bass	300–212 μm fibres	223%	163%–303%	Lower
All	Largemouth bass vs. smallmouth bass	500–300 μm foam	99%	94%–100%	Higher
All	Largemouth bass vs. spotted bass	500–300 μm foam	85%	71%–93%	Higher
All	Spotted bass vs. smallmouth bass	500–300 μm foam	94%	58%–99%	Lower
Morgantown	Largemouth bass vs. smallmouth bass	500–300 μm fragments	198%	72%–432%	Lower
All	Largemouth bass vs. spotted bass	300–212 μm fragments	46%	9%–69%	Higher
All	Spotted bass vs. smallmouth bass	300–212 μm fragments	77%	14%–177%	Lower

Note: Only statistically meaningful relationships are presented. Estimates represent the expected percentage of microplastic counts in each pool, derived from Poisson‐distributed models, and credible intervals (95% CI) indicate uncertainty in the posterior estimates.

3.3. Temporal variation in the point Marion Pool

Overall, temporal variation in microplastic counts had species‐specific responses. Rate ratios indicate smallmouth bass had lower expected total microplastic counts in September 2019 than in May 2020. Expected total microplastic counts in this species significantly decreased from September 2019 to May 2020 (Table 5). Conversely, rate ratios indicated that spotted bass did exhibit significant differences in the expected counts. Spotted bass significantly increased in the expected counts of total microplastics in September 2019 compared to May 2020. Largemouth bass showed no significant differences in expected total microplastic counts between months.

TABLE 5.

Bayesian regression models demonstrate the difference between September 2019 and May 2020 microplastic counts within species in the Point Marion pool, West Virginia, USA.

Species	Response variable	Rate ratio	95% CI	Direction of effect
Smallmouth bass	Total	22%	11%–34%	Lower
Smallmouth bass	>500 μm foam	99%	31%–100%	Higher
Smallmouth bass	500–300 μm fibres	60%	11%–129%	Lower
Smallmouth bass	300–212 μm fibres	159%	70%–294%	Lower
Smallmouth bass	500–300 μm fragments	70%	14%–153%	Lower
Smallmouth bass	500–300 μm sheets	80%	9%–195%	Lower
Largemouth bass	>500 μm sheets	53%	17%–75%	Lower
Largemouth bass	500–300 μm fibres	40%	25%–80%	Higher
Largemouth bass	500–300 μm fragments	69%	32%–87%	Higher
Largemouth bass	300–212 μm fragments	78%	55%–90%	Higher
Largemouth bass	500–300 μm sheets	76%	46%–91%	Higher
Largemouth bass	300–212 μm sheets	89%	68%–97%	Higher
Spotted bass	Total	43%	37%–51%	Higher
Spotted bass	>500 μm fragments	68%	21%–89%	Lower
Spotted bass	300–212 μm fragments	55%	8%–80%	Higher
Spotted bass	>500 μm sheets	84%	33%–98%	Higher
Spotted bass	300–212 μm sheets	81%	51%–94%	Higher
Spotted bass	500–300 μm fibres	52%	17%–74%	Higher
Spotted bass	300–212 μm fibres	51%	21%–70%	Higher

Note: Only statistically meaningful relationships are presented. Estimates represent the expected percentage of microplastic counts in September, derived from Poisson‐distributed models, and credible intervals (95% CI) indicate uncertainty in the posterior estimates.

Shape‐size fractions were significantly variable for all three species. In smallmouth bass, the expected counts of all three shapes of 500–300 μm in size and 300–212 μm fibres significantly decreased from September 2019 to May 2020, as indicated by the rate ratios (Table 5). Only foam of >500 μm in size had rate ratios that indicated higher expected counts in September 2019 than in May 2020. Spotted bass experienced a notable increase in fibres and sheets but a decline in large fragments (Table 5). Despite the lack of significant differences in the rate ratios of total microplastics, the shape‐size fractions contribute to the total variation observed in largemouth bass (Table 5). Rate ratios indicated that expected sheets >500 μm counts were lower in September 2019 than in May 2020. In September 2019, rate ratios indicate that expected counts of fibres, fragments and sheets size 500–300 μm and fragments and sheets size 300–212 μm were significantly higher than in May (Table 5). LOO indicated that expanding the temporal models to include diet did not improve their utility.

4. DISCUSSION

This study focuses on three sympatric black bass species found in the Monongahela River: largemouth bass, smallmouth bass and spotted bass. It leverages an existing sampling program and a low‐cost approach to assess microplastic consumption in recreationally important species. To our knowledge, it is the first to report microplastic consumption in the Monongahela River in black bass species. Microplastics were found in all fish sampled; however, the amount, type and size of microplastics varied between species within a navigational pool and within species between navigational pools.

Smallmouth bass exhibited higher microplastic consumption across shapes and size fractions, suggesting it may be more vulnerable to microplastic pollution than spotted and largemouth bass. All three species demonstrated unique independent changes in total microplastic consumption across pools and between months. This demonstrates that ecologically similar sympatric black bass species consume microplastics in different amounts, a finding echoed by other studies of sympatric ecologically similar species (Aliko et al., 2022; Di Benedetto & Ramos, ²⁰¹⁴; Messinetti et al., ²⁰¹⁸). Intraspecies variation in microplastic counts may be due to variations in behaviours due to differences in how each predator uses habitats and food resources, as has been shown in sympatric dolphin species (Vendel et al., 2017; McNeish et al., ²⁰¹⁸). Smallmouth bass are conventionally thought to be found in fast‐running water while largemouth bass are found in embayments and spotted bass are thought to inhabit intermediate zones. Differences in habitat partitioning may impact bioavailability (van Emmerik & Schwarz, ²⁰¹⁹).

Microplastic shape‐size consumption patterns in the Monongahela River pools were not dramatically different within species between navigational pools. Spatially, the patterning between pools within species was similar, exhibiting higher total microplastic counts in the Morgantown Pool, adjacent to an urban centre. A higher number of significant relationships within species was exhibited in the smaller size fractions, suggesting that their uptake may be more dynamic than in larger size fractions. These findings are significant when considering where to target potential monitoring efforts, suggesting that those populations most at risk for high total microplastic loads are not those downstream but those adjacent to the population centre.

While this study was limited in its temporal scope, we found evidence suggesting seasonal differences in microplastic consumption. These temporal differences may be due to seasonal changes in riverine dynamics that influence microplastic availability in the water (Kallenbach et al., 2022; Watkins et al., ²⁰¹⁹). Largemouth bass exhibited significant compositional changes in microplastic but stable counts between months, suggesting potential differences in bioavailability of the type of plastic it was exposed to. Pools on the Monongahela River are bookended by lock and dam systems that restrict flow and potentially impact the system's movement and accumulation of microplastics (Watkins et al., 2019). The movement of black bass in the Monongahela River and interactions with the locks and dams is currently unknown. Work conducted by Frued (2003) in the Ohio River, which the Monongahela River is a tributary of, found differences in habitat use between main river and off‐channel habitats during spring and winter seasons in largemouth bass. Additional research is warranted to assess the bioavailability of microplastic consumption within this heavily modified system and to explore the potential effects of dams on the bioavailability and movement of microplastics and black bass in the Monongahela River. Future research should also examine the potential sources of microplastic into the Monongahela River and its tributaries, including wastewater treatment plants, stormwater, agricultural runoff, urban debris sources and other effluent sources.

Diet has been explored extensively as a contributing factor to microplastic consumption (Estahbanati et al., 2016; McNeish et al. ²⁰¹⁸; Vendel et al., ²⁰¹⁷; O'Connor et al., ²⁰²⁰). Although the black bass in this study would have been classified as piscivores with similar feeding morphology, they exhibited species‐specific microplastic consumption independent of their diet. Including diet as a covariate did not improve model performance or explain additional variation in observed microplastic counts. The findings may elucidate why correlations between microplastic counts and feeding guilds have been inconclusive and contradictory, as it may be more based on the species or individuals involved than the feeding guild (Estahbanati et al., 2016; McNeish et al., ²⁰¹⁸; Vendel et al., ²⁰¹⁷; O'Connor et al., ²⁰²⁰). Like other studies, this research assumes that the dietary snapshot observed in the stomach contents accurately reflects consumption patterns related to the identified microplastics, a standard premise in diet and microplastic research (Estahbanati et al., 2016; McNeish et al., ²⁰¹⁸; Vendel et al., ²⁰¹⁷; O'Connor et al., ²⁰²⁰). Dividing the prey item into an abundance of crayfish and fish may not be enough to substantiate relationships. Although outside the scope of this study, diet items included taxa such as Etheostoma and Percina, darters, benthic dwelling fish, as well as pelagic fishes such as gizzard shad, Dorosoma cepedianum, and white bass, Morone chrysops. Future research should investigate the influence of prey habitat residency and predator dietary breadth variation on microplastic consumption. Additionally, the retention time of microplastics in the digestive tract and variable digestion rates of crayfish compared to fish can influence perception. Future work to understand shape‐size accumulation between species may benefit from incorporating abiotic factors and diet characteristics with habitat suitability modelling to understand accumulation differences between and within riverine environments, improving the understanding of microplastic consumption composition.

Large amounts of uncertainty in estimates may be due to methodological choices, analysis decisions or small sample sizes. Careful consideration of statistical methods is essential when analysing microplastic consumption counts, as different treatments of overdispersion and zero‐generating processes can affect interpretation. Additionally, conflicting results are typical in microplastic studies because different response variables are examined and methodological choices are made. Further research is needed to understand how analysis choices influence results in this evolving field. Future studies could benefit from employing techniques such as FTIR spectroscopy, Raman spectroscopy and Py‐GCMS to help identify microplastic sources in regions of elevated consumption, facilitating mitigation efforts. However these specialized methods are costly, and it should be assessed when their use is most justified or offers a clear advantage over more economic alternatives.

Given the pervasive nature of microplastics in the environment (Roch & Liu, 2020), the presence of microplastics in all collected fish is not surprising. While the presence of microplastics is not novel, detecting these particles in black bass for the first time in the Monongahela River enhances awareness of plastic pollution in this ecologically sensitive area and reinforces the reality of its presence in an economically important species. The findings highlight the need to evaluate microplastics' potential impacts in this ecosystem due to their potential sublethal effects on local biota (Foley & Parnell, 2018; He et al., ²⁰²¹; Huang et al., ²⁰²¹).

Furthermore, microplastics are not presently monitored in the Monongahela River, despite its role as a drinking water source for communities like Morgantown, West Virginia, USA, a significant population centre. Due to the discovery of microplastics further research into their potential presence in drinking water should be conducted due to potential human health impacts. Identifying microplastics in these economically and socially essential species aids in the broader scientific discourse to consider potential ways to expand awareness, assess and manage this issue using legislation such as the Clean Water Act, the Safe Drinking Water Act, and the Toxic Substances Control Act (TSCA) (Sorensen et al., 2023). However, this research is labour and resource intensive, making it difficult to process large sample sizes. Despite this limitation, long‐term data collection and monitoring should be considered as it would strengthen our understanding of this emerging pollutant.

While microplastics have garnered attention globally, the research has disproportionately emphasized marine environments, leaving freshwater ecosystems, especially rural river systems, relatively under‐studied (Estahbanati & Fahrenfeld, 2016; Lusher et al., ²⁰²⁰; Pew Charitable Trust & SYSTEMIQ, ²⁰²⁰; Roch & Liu, ²⁰²⁰). This is the first examination of microplastic consumption in fish in West Virginia and the Ohio River drainage headwaters. Microplastic was found in variable but high amounts in all smallmouth, largemouth and spotted bass examined. The findings suggest that black bass microplastic consumption varies between species and is temporally dependent. Diet did not improve understanding of microplastic count variation, but species‐specific effects were noted. The findings suggest that this issue is pervasive, warranting more research to examine the presence in other rivers in the region to help understand the risk this emerging pollutant poses. Rural rivers have been a blind spot of microplastic research in the United States, potentially due to restricted funding. This paper presents a low‐cost approach for analysing microplastic consumption in existing sampling programs and a robust statistical framework for preliminary analysis to analyse microplastic consumption in these systems for a baseline understanding of the risk and scale of the issue.

AUTHOR CONTRIBUTIONS

Conceptualization: K.Z. and K.J.H. Developing methods: K.Z. and K.J.H. Conducting the research: K.Z. Data analysis: K.Z. Data interpretation: K.Z. Preparation of figures and tables: K.Z. and C.M.L.S.P. Writing: K.Z., K.J.H. and C.M.L.S.P.

CONFLICT OF INTEREST STATEMENT

There is no conflict of interest for the authors involved.

Supporting information

TABLE S1. Contamination results from procedural blanks employed to measure microplastic transfer through the sieve tower measured for each sieve (n = 35).

TABLE S2. Contamination results from controls used throughout sampling processing to account for and ensure minimal contamination.

FIGURE S1. Size distribution (total length, TL, mm) of fish collected from the Monongahela River by species and pool.

DATA AVAILABILITY STATEMENT

Data are available from the authors upon reasonable request.