The Effects of Wastewater Reuse on Smallmouth Bass (Micropterus dolomieu) Relative Abundance in the Shenandoah River Watershed, USA

Municipal and industrial wastewater effluent is an important source of water for lotic systems, especially during periods of low flow. The accumulated wastewater effluent flows—expressed as a percentage of total streamflow (ACCWW%)—contain chemical mixtures that pose a risk to aquatic life; fish may be particularly vulnerable when chronically exposed. Although there has been considerable focus on individual‐level effects of exposure to chemical mixtures found in wastewater effluent, scaling up to population‐level effects remains a challenging component needed to better understand the potential consequences of exposure in wild populations. This may be particularly important under a changing climate in which wastewater reuse could be essential to maintain river flows. We evaluated the effects of chronic exposure to wastewater effluent, as measured by ACCWW%, on the relative abundance of young‐of‐year (YOY), juvenile, and adult smallmouth bass (Micropterus dolomieu) populations in the Shenandoah River Watershed (USA). We found that increases in ACCWW% in the previous year and during the prespawn period were negatively correlated with the relative abundance of YOY, resulting in an average 41% predicted decrease in abundance (range = 0.5%–94% predicted decrease in abundance). This lagged effect suggests that adult fish reproductive performance may be compromised by chemical exposure during periods of high ACCWW%. No relationships between ACCWW% and juvenile or adult relative abundance were found, suggesting that negative effects of ACCWW% on YOY abundance may be offset due to compensatory mechanisms following higher ACCWW% exposure. Understanding the effects of wastewater effluent exposure at multiple levels of biological organization will help in the development of management strategies aimed at protecting aquatic life. Environ Toxicol Chem 2024;43:1138–1148. © 2024 The Authors. Environmental Toxicology and Chemistry published by Wiley Periodicals LLC on behalf of SETAC. This article has been contributed to by U.S. Government employees and their work is in the public domain in the USA.


INTRODUCTION
Chemical pollution of freshwaters is a ubiquitous problem and represents unique challenges involving the understanding of processes governing chemical transport, fate, and the potential risks posed to aquatic life (Schwarzenbach et al., 2006).Inorganic and organic contaminants often occur at low, but potentially biologically relevant concentrations, and as complex mixtures that may increase the risk to aquatic organisms through interactive or cumulative effects (Barber et al., 2022;Bradley et al., 2017;Kortenkamp, 2007;Relyea, 2009;Schäfer et al., 2016).Laboratory and mesocosm experiments and field investigations are rapidly increasing our knowledge of the effects of individual contaminants and biologically relevant chemical mixtures on aquatic animals (see Fleeger et al., 2003;Miller et al., 2020;Roznere et al., 2023;Vijayanand et al., 2023).These studies are critical for understanding organismal response mechanisms, informing risk assessments, and developing biomarkers for monitoring wild populations (Fitzgerald et al., 2021;Jeffries et al., 2021;Miracle & Ankley, 2005).However, scaling up from individual molecular and physiological endpoints to infer population-level consequences of exposure is difficult (Li et al., 2020).Accordingly, complementary studies examining potential contaminant-exposure effects on populations of wild animals are needed to bridge this gap and lend insight into how individual exposure translates to changes in abundance (Hamilton et al., 2016).Although investigations of wild populations are correlative and often lack direct information about the specific mechanisms involved, they provide biologically relevant insights directly relevant to fish and wildlife management decisions and can also be the basis for generating hypotheses that can be further tested under controlled conditions.
Wastewater effluent is an important source of complex contaminant mixtures to many streams and rivers (Ball et al., 2023;Faunce et al., 2023;Qiang et al., 2013;Weisman et al., 2021).For example, Faunce et al. (2023) found that over 8000 stream kilometers in the Potomac River Watershed (USA; spanning Washington, DC and the states of Maryland, Pennsylvania, Virginia, and West Virginia) may be influenced by chemical mixtures derived from municipal and industrial wastewater effluent.The effects of wastewater effluent on aquatic life are well documented and include potential disruptions to metabolic processes, physiology, and endocrine systems (Jobling et al., 2009;Topić Popović et al., 2023;Woodling et al., 2006) as well as impacts on community structure (McCallum et al., 2019) and abundance (Brown et al., 2011;Galib et al., 2018;Smith & Suthers, 1999).Although these effects depend on the specific chemical composition of the effluent and the biological and physiochemical characteristics of the receiving systems (Pinheiro et al., 2021), the volume of wastewater effluent relative to the total river discharge partly determines the magnitude of exposure to aquatic life.During periods of low flow, wastewater effluent can constitute a large proportion of the total river discharge, resulting in organismal exposure to higher contaminant concentrations (Galib et al., 2018;Karakurt et al., 2019;Rice et al., 2013).Furthermore, anticipated decreases in average streamflow due to climate change have the potential to increase wastewater signals, especially in drought-prone areas (Rice et al., 2013).Faunce et al. (2023) found that accumulated wastewater effluent flowsexpressed as a percentage of total streamflow (ACCWW%)-may pose a risk to aquatic life in the Potomac River Watershed and posited that fish may be more susceptible to negative effects resulting from chronic exposure to the chemical mixtures.In fact, wild fish in the Shenandoah River have shown signs indicative of chemical exposure and endocrine disruption, including intersex and oxidative damage (Blazer et al., 2007(Blazer et al., , 2010)).Blazer et al. (2007) found that the prevalence of testicular oocytes in smallmouth bass (Micropterus dolomieu) collected from the Shenandoah River ranged from 80% to 100%, and Blazer et al. (2012) determined that 50% of male smallmouth bass sampled from the North Fork of the Shenandoah River had detectable levels of plasma vitellogenin, suggesting exposure to chemicals interacting with estrogen receptors.In addition, the Shenandoah River Watershed has been the location of robust assessments evaluating the biotic integrity of aquatic communities, with an emphasis on endocrine disruption.For instance, mobile laboratory studies have found negative reproductive and survival outcomes in fathead minnows (Pimephales promelas) exposed to river water from the Shenandoah River (Barber et al., 2019;Bertolatus et al., 2023).Based on a cumulative risk index, Barber et al. (2022) concluded that the chemical mixtures found in the waters of the Shenandoah River Watershed have the potential to cause adverse effects on aquatic organisms, and although multiple fish species have been observed with signs of chemical exposure, the effects on the ecologically and socioeconomically important smallmouth bass have been notable and have co-occurred with mortality events of sexually mature smallmouth bass (Blazer et al., 2010).
As in most aquatic systems, multiple stressors (e.g., thermal, chemical, pathogens) are likely interacting to affect the health of riverine fish populations in the Shenandoah River Watershed (Barber et al., 2022;Blazer et al., 2010).Among the various sources of chemical inputs, accumulated wastewater effluent flows resulting from continuous municipal and industrial wastewater treatment plant discharges represent a chronic, relatively stable exposure source of contaminants to aquatic life, including industrial chemicals, pharmaceuticals, and domestic pesticides (Barber et al., 2022).There is also evidence that endocrine-disrupting chemicals (EDCs) from wastewater effluent represent a risk to fish in this system (Barber et al., 2019;Blazer et al., 2007), with exposure experiments resulting in vitellogenin induction in adult male fathead minnows (Barber et al., 2019).Endocrine-disrupting chemicals are a diverse group that can affect the endocrine system through a variety of mechanisms, with potential effects on both male and female fish, including reproduction (e.g., inhibition of gametogenesis and decreased fertility rates; Carnevali et al., 2018).
Because of the diversity of potential mechanisms through which EDCs may affect fishes, the effects of EDCs are likely life stage (age) dependent, and exposure during important developmental stages is particularly important (Carnevali et al., 2018).For example, exposure during early life stages and recrudescence has been shown to have significant impacts on various physiological and metabolic processes (Ankley & Johnson, 2004;Duffy et al., 2014;Koger et al., 2000).Although population-level changes have been documented (see Blanchfield et al., 2015;Kidd et al., 2007), evidence that shows chronic exposure to sources of EDCs affecting the abundance of wild fish populations is much less common than evidence of EDC exposure in individual fish.Linking chemical exposure, or proxies thereof (e.g., ACCWW%), to changes in population abundance in wild fish has been identified as a challengebecause quantifying abundance is difficult and because fish populations exhibit large natural variation and are influenced by a complex suite of biotic and abiotic factors (Rose, 2000).An additional challenge is separating the potential direct effects of streamflow from flow-dependent chemical exposure on fish abundance.This is because the calculation of metrics such as ACCWW% includes information about streamflow (Barber et al., 2019), and streamflow can also independently affect the survival of riverine smallmouth bass (Smith et al., 2005).For instance, extreme low flows combined with elevated water temperatures (>30 °C) can lead to increased mortality (Hafs et al., 2010), and high flows during spawning are more likely to result in nest failure by causing nest desertion and flushing away of swim-up fry (Li et al., 2020;Lukas & Orth, 1995).However, an understanding of population-level outcomes is needed to better discern whether exposure translates to changes in population structure (Mills & Chichester, 2005).Therefore, the objectives of the present study were to: (1) evaluate the potential effects of chronic exposure to wastewater effluent, as measured by ACCWW%, on the relative abundance of young-of-year (YOY), juvenile, and adult smallmouth bass populations in the Shenandoah River Watershed; and (2) evaluate whether ACCWW% provided a better fit to the smallmouth bass abundance data than streamflow alone.

Study area
The Shenandoah River Watershed is a sub-basin of the Potomac River Watershed that encompasses 7874 km 2 of northern Virginia and eastern West Virginia in the Valley and Ridge province (USA; Figure 1).Three major tributaries (North River, Middle River, and South River) converge south of Staunton, Virginia to form the South Fork Shenandoah River (a Strahler 5th-order stream; Strahler, 1957), which flows 165 km north before joining the North Fork Shenandoah River (168 km in length) northeast of Front Royal, Virginia to form the Shenandoah River (5th-order stream).The Shenandoah River (91 km in length) is a major tributary of the Potomac River (7th-order stream), which flows 282 km from its confluence with the Shenandoah River before emptying into the Chesapeake Bay.The estimated water travel time from the South River to the Potomac River is 15 to 20 days (Taylor et al., 1986).
The Shenandoah River Watershed supported a population of over 430,000 people in 2020 (US Census Bureau, 2020).Towns located in the South Fork Shenandoah River drainage include Harrisonburg (population of ~51,800), Staunton (population 25,750),and Waynesboro (population 22,200).Towns located in the North Fork Shenandoah River drainage include Strasburg (population 6700) and Woodstock (population 5800).Front Royal (population 15,000) is located at the confluence of the North Fork and South Fork Shenandoah Rivers.Approximately 50% of the total Shenandoah River Watershed population lives in incorporated urban areas with public water supply, centralized sewage collection, and wastewater treatment systems; the other 50% lives in unincorporated areas with household water supply wells and on-site wastewater disposal systems (Central Shenandoah Planning District Commission, 2011; Northern Shenandoah Valley Regional Commission, 2011).In addition to effluent flows, other potential contaminant sources within the watershed include agricultural activities such as row cropping (primarily corn and soybean), landfills, and animal feeding operations (Gordon et al., 2017).Agricultural land use comprised approximately 30% and developed land use comprised approximately 10% of the Shenandoah River Watershed in 2019 (Dewitz & US Geological Survey, 2021), with river valley streams typically bordered by these land use types relative to highland tributaries.

Smallmouth bass data
Smallmouth bass data were obtained from fisheryindependent surveys performed by the Virginia Department of Wildlife Resources (Figure 1).Data were collected in 33 study reaches from 1998 to 2018, and analysis was restricted to fall surveys, when capture efficiency was greatest for age-0 fish.Sites were sampled once a year, and the number of years a site was sampled varied from 1 to 13 (Supporting Informaion, Tables S1-S3).Survey catch data consisted of individuals caught, their total length (TL; mm), survey effort (h), and river reach location.Because we hypothesized that exposure to high ACCWW% could be size (age) dependent, we classified individuals as YOY (i.e., age-0 fish; <130 mm), juvenile (i.e., age-1 fish; 130 mm < length ≤ 179 mm), and adults (>179 mm TL).

Accumulated wastewater
As described by Weisman et al. (2021), effluent flow data for 2000 to 2019 were obtained from Discharge Monitoring Report databases compiled by state environmental departments for National Pollutant Discharge Elimination System-permitted facilities discharging into the Shenandoah River (Adams, 2019; US Environmental Protection Agency [USEPA], 2020).The compiled data set included 98 wastewater treatment plants (WWTPs), 94 of which were in Virginia.Twenty-two WWTPs were industrial (i.e., nonpublicly owned treatment works; wastewater derived from manufacturing, mining, and/or refining activities), and 76 were municipal (i.e., publicly owned treatment works; wastewater derived from homes, institutions, and small businesses).
A process-based model (Faunce et al., 2023;Weisman et al., 2021) was used to estimate the volume of accumulated wastewater (from which ACCWW% was calculated) from municipal-plus-industrial effluent flows at each smallmouth bass sampling site after linking the sampling sites and effluent flows to their corresponding National Hydrography Dataset Plus Ver.2.1 (NHDPlus Ver.2.1; USEPA, 2012) stream segments.Model inputs included the reported effluent flows, NHDPlus Ver.2.1 stream networks and hydrologic attributes, and measured streamflow discharge data from US Geological Survey (USGS, 2020) continuous monitoring streamgages.Streamflow at ungaged fish sampling sites was estimated as the product of measured flow for a given period from the nearest upstream or downstream streamgage and an adjustment factor.The adjustment factor represented the proportional difference in the estimated NHDPlus Ver.2.1 mean annual enhanced runoff method streamflow condition (USEPA, 2012) at the fish sampling site and the measured streamflow from the nearest streamgage (Wiesman et al., 2021).Adjustment factors ranged from 0.7 to 1.4.The ACCWW% was summarized to annual and quarterly timescales (Q1: Jan-March, Q2: Apr-Jun, Q3: July-Sept, Q4: Oct-Dec) because the effects of exposure may be time dependent-for example, if exposure occurs during a critical developmental life stage.The ACCWW% calculations were performed in Python (Python Software Foundation, 2017) and assumed no loss of wastewater between upstream receiving reaches and downstream locations, consistent with other studies (see Barber et al., 2019;Rice et al., 2013).

Statistical analysis
To investigate the potential effects of ACCWW% on smallmouth bass relative abundance, we fitted a Bayesian negative binomial generalized linear regression model, with fish counts at each site/year as the response variables.Expected fish counts for each site and year ( [ ]) E Y sy were modeled with a log- link as a function of standardized median ACCWW%.The model also included the sample year and site as random effects (γ γ , s y ), and sampling effort (in hours) was included as an offset term.The linear predictor of the model was as follows: where β 0 is the intercept, β accww is the effect of ACCWW% on relative abundance and the site and year random effects (γ s and γ y ) are assumed to be normally distributed with mean 0 and variances σ s 2 and σ y size class (YOY, juvenile, and adult) separately.In addition, a separate model was fitted for each timescale of ACCWW%, including the previous year's (year prior to sampling, i.e., lagged by 1 year) median annual ACCWW% and quarterly ACCWW%'s, as well as the current year's (i.e., year of sampling) and current year quarters 1-3.The Q4 of the current year was not investigated because this would be ACCWW% that the fish were not yet exposed to (i.e., fish are sampled in Q3 [fall] and Q4 would be after the fish are sampled).All models were fitted using the rstanarm package assuming default priors using the stan_glmer function (Goodrich et al., 2023) in the program R (R Core Team, 2023).Three Markov chains were run for 5000 iterations, half of which the stan_glmer assigns as burn-in and half of which are saved samples used to summarize the posterior distribution.Convergence was assessed through inspection of trace plots and plots of R and N eff statistics (Gelman & Rubin, 1992).The effect of ACCWW% on smallmouth bass abundance was considered significant if the 95% credible interval for β accww did not overlap with 0. For those models that had statistically significant effects of ACCWW% on smallmouth bass relative abundance, we fitted a model of the same form as described for ACCWW% but used streamflow as the predictor variable.We calculated the Pareto smoothed importance-sampling leave-one-out cross-validation information criterion (PSIS-LOO-IC; Vehtari & Gabry, 2019;Vehtari et al., 2017) to compare models with statistically significant ACCWW% effects with the competing model's hypothesis that variation in relative abundance would be better described by streamflow alone.

RESULTS
The ACCWW% ranged from 0.25% to 21.97%, with a mean and standard deviation of 3.68 and 6.47, respectively across all sites and years (Supportin Information, Tables S1-S3).The ACCWW% varied across years (with the largest median value for all sites ranging from 4% to 15%) and across river forks (with the south and north forks shown in Figure 1 displaying the largest and smallest median ACCWW%, respectively).Median stream discharge ranged from 1.7 to >42.5 m 3 /s across sites and years (Supporting Information, Tables S1-S3).There was a total of 20,034 individual smallmouth bass used in the analysis sampled across 44 sites (Figure 1; n = 5386 YOY; n = 4167 juvenile; n = 7361 adult fish).
The effects of ACCWW% were statistically significant and negative for YOY relative abundance at all lagged timescales (the previous year's average and previous year's Q1-Q4) and the current year's Q1.However, the effect of ACCWW% was not significant for YOY in Q2 or Q3 (Figure 2).The effect of ACCWW% was not statistically significant at any timescale for juvenile or adult relative abundance.The YOY model fits for ACCWW% were compared with models that only included streamflow using PSIS-LOO-IC.For all eight models, ACCWW % fit the data better (i.e., smaller PSIS-LOO-IC) compared with the streamflow model (Table 1; see the Supporting Information, Figure S1, for estimated streamflow effects).The magnitude of the effect of ACCWW% can be visualized in Figure 3, where predicted catch-per-unit effort of YOY smallmouth bass is plotted against the largest change (positive or negative) in ACCWW% observed at a site.Observed decreases in ACCWW% led to a predicted 3% to 672% increase in YOY catch-per-unit effort, and increases in ACCWW% led to a predicted decrease in catch-per-unit effort of between 0.5% and 94% (Figure 3).

DISCUSSION
We found that exposure to high ACCWW% in the previous year and during the prespawn (Q1) period was negatively correlated with the relative abundance of YOY smallmouth bass.We did not find evidence suggesting that the effect of ACCWW% was due to a streamflow artifact.Our models incorporating ACCWW% provided a better fit to the data compared with those using streamflow alone.In addition, if flow was driving our observed ACCWW% results, we would predict a positive relationship between YOY abundance and ACCWW% in Q1 and Q2 (indicating a negative effect of streamflow due to the inverse relationship with ACCWW%), and there would not be any lagged effects on YOY abundance.We found no significant lagged effects and negative effects of flow for Q2 and Q3 (Supporting Information, Figure S1).These results support the notion that ACCWW% exposure is capturing more than simply a flow effect on smallmouth bass abundance.
The lagged ACCWW% effect suggests that adult fish reproductive performance may be compromised by exposure to chemical mixtures during periods of higher ACCWW% and that these effects manifest through decreased YOY abundance.The potential for compromised reproductive capacity in wild fish populations in response to exposure to estrogenic compounds originating from wastewater effluent is well documented (Jobling et al., 1998;Tyler & Routledge, 1998;Vajda et al., 2008).In fact, it has been shown that improvements to WWTPs aimed at reducing the estrogenicity of the effluent can lead to improvements in the reproductive condition of wild fish (Nikel et al., 2023)-indicating a strong causal link.This link may help explain our finding of a negative correlation between ACCWW% and YOY relative abundance.
Several, nonmutually exclusive, mechanisms could be involved, whereby adult (male and/or female) exposure to high ACCWW% results in poor reproductive success.Iwanowicz et al. (2009) found that the gonadosomatic index of female bass (Micropterus spp.) downstream of a WWTP was lower compared with fish sampled upstream of the WWTP, and that the exposure to effluent could negatively affect fecundity, egg quality, and survival of fry resulting from lower plasma vitellogenin concentrations.Carnevali et al. (2010) found reduced reproductive capacity in zebra fish (Danio rerio) exposed to di-(2-ethylhexyl)-phthalate-a chemical commonly found in aquatic environments and wastewater effluent (Zolfaghari et al., 2014).In a study examining the effects of wastewater effluent on wild roach (Rutilus rutilus), Jobling, Beresford, et al. (2002) found that female fish living in rivers receiving wastewater had a higher incidence of oocyte atresia compared with reference fish not exposed to wastewater effluent.Evidence from both laboratory and field studies suggests that exposure of female fish to estrogenic compounds in wastewater effluent has the potential to have population-level consequences (Miller et al., 2007).
The effects of EDC exposure on male smallmouth bass could also contribute to the observed negative correlation between ACCWW% and YOY abundance.The EDC exposure of male fishes has negative effects on spermatogenesis, through affecting the function and development of gonads and disrupting the function and synthesis of sex hormones (Delbès et al., 2022).The proportion of male smallmouth bass in the Shenandoah River with testicular oocytes was estimated to be between 80% and 100%, with the highest prevalence and severity of intersex observed during the prespawn and spawning season (Blazer et al., 2007).Jobling, Coey, et al. (2002) demonstrated that intersex in wild roach resulted in reduced gamete production.Intersexed wild roach had reduced ability to successfully fertilize eggs and produce viable offspring FIGURE 2: Effects of effluent flows as a percentage of total streamflow (ACCWW%) on young-of-year (YOY), juvenile, and adult smallmouth bass (Micropterus dolomieu) relative abundance in the Potomac River Watershed, USA.The ACCWW% was summarized at annual and quarterly (Q1: Jan-March, Q2: Apr-Jun, Q3: July-Sept, Q4: Oct-Dec) timescales for the previous year (lagged full year and quarterly) and current year quarters 1-3.Estimates with 95% credible intervals that do not overlap with 0 (blue points and lines) are statistically significant.compared with nonintersex male roach, and gamete quality was negatively correlated with the severity of intersex (Jobling, Coey, et al., 2002).Nonlethal effects of higher ACCWW% on adult smallmouth bass reproductive success is one pathway by which chronic exposure to wastewater effluent could affect YOY abundance.However, mortality events of adult, sexually mature smallmouth bass during the spring (prespawn/spawning period) that have been observed in the Shenandoah River could also contribute to poorer reproductive output in some years.The combination of multiple stressors, including high ACCWW% (exposure to chemical mixtures), thermal stress, and pathogens may result in adult mortality in some years (Blazer et al., 2010) that compounds any nonlethal effects on adult reproductive performance.
Although we observed a negative correlation between ACCWW% and YOY relative abundance, we did not detect relationships between ACCWW% and juvenile or adult relative abundance.This lack of an observed effect of ACCWW% with older age classes may suggest that any increase in mortality of eggs or fry is compensated for by density-dependent processes, that is, any decrease in fecundity/reproductive success is offset due to compensatory mechanisms (e.g., increased growth and survival) because of lower YOY densities in years following higher ACCWW% exposure.For example, increased growth rates during years of lower densities could result in larger YOY in the fall and increased overwinter survival, since overwinter survival in many fish, including Micropterus spp., has been shown to be size dependent in both experimental and field studies (Garvey et al., 1998;Miranda & Hubbard, 1994;Oliver et al., 1979).This is not to suggest that the potential for compensatory mechanisms to offset negative effects of ACCWW% exposure on the reproductive performance of smallmouth bass would be sufficient to mediate effects on older age classes indefinitely.For example, using a simulation model for brook trout (Salvelinus fontinalis), Power (1997) showed that variability in juvenile abundance was most sensitive to stressor intensity and that adult abundance became equally sensitive after the population's compensatory capacity was exceeded.This finding suggests the potential for toxicity thresholds, whereby compensatory mechanisms no longer offset negative effects on population-level outcomes.Furthermore, there are likely interactions between ACCWW% and other environmental stressors, such as increasing water temperatures and low river discharge, that will affect exposure and toxicity dynamics (Noyes et al., 2009;Patra et al., 2015;Schiedek et al., 2007).Accordingly, a better understanding of the density-dependent processes regulating populations for many species could improve our ability to predict the effects of contaminant exposure on population dynamics in a rapidly changing climate.
Previous research has shown that upgrades to WWTP treatment technologies can reduce contaminant concentrations in effluent and downstream receiving waters, followed by improvements in biochemical and tissue biomarkers of exposure in wild fish populations (Nikel et al., 2023;Wilhelm et al., 2017).In our study system, Weisman et al. (2021) reported that nine of the 14 large WWTPs (i.e., WWTPs with calculated annual average flows that were ≥ 0.044 m 3 /s) had advanced nutrient reduction treatment processes installed in or around 2010.For smaller systems (average flows < 0.044 m 3 /s), any previous or current information on upgrades is unavailable.Unfortunately, the data do not currently exist for us to elucidate the potential effects of WWTP upgrades in our study system, but it remains an important area of future research.In addition to changes in WWTP operations (e.g., upgrades and changes in discharge rate), other factors can contribute to observed spatiotemporal variation in ACCWW%, including variation in the duration and magnitude of low flow periods, stream size, and water withdrawals (Hamdhani et al., 2020;Luthy et al., 2015).An increased understanding of the spatiotemporal variation in ACCWW% and the concentration of complex chemical mixtures-which can vary seasonally and with human use patterns (Zhi et al., 2021)-will also be important for developing strategies to manage the health of lotic ecosystems.
Because our study was observational in nature, we cannot rule out other factors that could be affecting smallmouth bass abundance, including both landscape and stream segment habitat characteristics (Brewer et al., 2007).Additionally, we cannot disregard the potential impact of other sources of chemical mixtures entering the river system, such as EDCs derived from agricultural activities and animal feeding operations (Blazer et al., 2012), which may affect the reproductive performance of smallmouth bass.
Wastewater treatment effluent is a well-known source of complex chemical mixtures, including estrogenic compounds, into freshwater streams and rivers.Despite great progress in understanding the biological pathways and mechanisms involved in the response of many organisms to chemical exposure, the challenge of scaling-up to population-level consequences remains.We have shown that the relative abundance of YOY smallmouth bass is negatively correlated with ACCWW% in the Shenandoah River watershed.Although previous research in this river system has shown effects of chemical exposure at the individual level (e.g., high intersex prevalence), our study suggests that reproductive impairment may translate to lower YOY abundance in the year following exposure to high ACCWW%.Changes in climate are predicted to result in decreased flows, at least seasonally, in many rivers (van Vliet et al., 2013), which will effectively increase ACCWW% and the intensity of exposure to wastewater contaminants.Understanding the effects of such exposure at multiple levels of biological organization will help in the development of management strategies aimed at protecting aquatic life.
Supporting Information-The Supporting Information is available on the Wiley Online Library at https://doi.org/10.1002/etc.5849.The accumulated wastewater effluent flows-expressed as a percentage of total streamflow (ACCWW%) on young-of-year (YOY) relative abundance in the Potomac River Watershed (USA).The plot illustrates the difference between predicted YOY catch-per-unit effort given ACCWW% the entire lagged year before and after the largest observed change in ACCWW% across all years at a given site.Arrows connect the predictions at a single site after the change in ACCWW%; decreases in ACCWW% led to an increase in YOY abundance, shown by the blue arrows, and increases in ACCWW% led to decreases in YOY abundance, shown by the red arrows.
Disclaimer-Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the US Government.

FIGURE 1 :
FIGURE 1: Map of study area in the Potomac River Watershed, USA, including sample reaches with smallmouth bass (Micropterus dolomieu), relative abundance data, and location of wastewater treatment plants.Blue lines are river systems.USGS = US Geological Survey.

FIGURE 3 :
FIGURE 3:The accumulated wastewater effluent flows-expressed as a percentage of total streamflow (ACCWW%) on young-of-year (YOY) relative abundance in the Potomac River Watershed (USA).The plot illustrates the difference between predicted YOY catch-per-unit effort given ACCWW% the entire lagged year before and after the largest observed change in ACCWW% across all years at a given site.Arrows connect the predictions at a single site after the change in ACCWW%; decreases in ACCWW% led to an increase in YOY abundance, shown by the blue arrows, and increases in ACCWW% led to decreases in YOY abundance, shown by the red arrows.

TABLE 1 :
Pareto smoothed importance-sampling leave-one-out information criterion (PSIS-LOO-IC) for models comparing accumulated wastewater effluent flows-expressed as a percentage of total streamflow (ACCWW%) with streamflow (m 3 /s) as predictors of youngof-year smallmouth bass (Micropterus dolomieu) relative abundance in the Shenandoah River Watershed (USA)Lower PSIS-LOO-IC values represent a model with a better fit to the data.See Accumulated wastewater and Statistical analysis sections in Methods for details describing the models.Q = quarter (of the year).