Factors affecting spatial and temporal patterns in perfluoroalkyl acid (PFAA) concentrations in migratory aquatic species: a case study of an exploited crustacean

Matthew D. Taylor *ab
aPort Stephens Fisheries Institute, New South Wales Department of Primary Industries, Locked Bag 1, Nelson Bay, NSW 2315, Australia. E-mail: matt.taylor@dpi.nsw.gov.au
bThe University of Queensland, Queensland Alliance for Environmental Health Sciences, 20 Cornwall Street, Woolloongabba, Queensland 4102, Australia

Received 29th April 2019 , Accepted 23rd June 2019

First published on 25th June 2019


Per- and poly-fluorinated alkyl substances (PFASs, including perfluoroakyl acids [PFAAs]) have been used in a range of applications, and are widely distributed throughout the environment including environmental media in aquatic systems. Recent literature provides multiple reports of these compounds in a range of aquatic species, but temporal and spatial variability in tissue concentrations is rarely assessed in a rigorous way. Using an important fishery species of representative biology as a case study (Eastern School Prawn, Metapenaeus macleayi), temporal (month-to-month, and year-to-year) and spatial (intra-estuarine and oceanic) variability in PFAAs concentrations was assessed alongside potential contributing factors. Perfluorooctane sulfonate (PFOS) was the dominant PFAA detected, and there was significant spatial variation in concentration driven primarily by distance to major point sources. There was also substantial variation in PFOS among months, likely driven by behavioural physiological or ecological factors. Importantly, muscle tissue concentrations were unrelated to surface water inputs of PFAAs into the estuary. A numerical model linking prawn migration data with concentrations in the estuarine nursery accurately predicted PFOS concentrations in adjacent oceanic trawling grounds. The results demonstrate the magnitude of temporal and spatial variation in PFAA concentrations, which has implications for assessing PFAA exposure risk through seafood consumption for free-ranging aquatic animals.



Environmental significance

Per- and poly-fluorinated alkyl substances (PFASs) have been used extensively, and are widely distributed throughout the environment. The relative solubility and mobility of these contaminants mean they are commonly present in aquatic environmental media, including fishery species. This has created some concern regarding exposure risk through seafood and potential impacts on the health of fish, crustacean and mollusc populations, but factors driving variability in tissue concentrations are poorly understood for aquatic species. Using an important fishery species of representative biology as a case study, the results of this work highlight the complex biological, ecological and physical factors affecting tissue concentrations in aquatic species. A novel model is also presented, to highlight how knowledge of animal migration patterns can be used to predict contaminant load in migratory species away from known point sources.

Introduction

Per- and poly-fluorinated alkyl substances (PFASs) have been used in a range of applications, including industrial and manufacturing processes, and industrial and consumer products.1 As a consequence, these chemicals are widely distributed throughout the environment, and are resistant to a range of degradation processes which makes them particularly persistent.2 PFASs are a key ingredient of aqueous film-forming foams (AFFF) which are used to combat hydrocarbon fires. While aviation disasters are relatively rare, deployment of AFFF has occurred regularly through training exercises, and as a consequence PFASs are present in environmental media adjacent to military bases, airports, and firefighting training centres.3,4 The relative solubility and mobility of PFASs in water mean they are frequently detected in adjacent aquatic environments,5,6 including sources of drinking water.7

The environmental fate of PFASs in the aquatic environment includes bioaccumulation in aquatic organisms,1 and the recent literature provides multiple reports of these compounds in a range of aquatic species and environments e.g.ref. 8–15. While numerous PFAS compounds have been detected, longer chain perfluoroalkyl acids (PFAAs) appear to be more prevalent in estuarine and marine species, especially perfluorooctane sulfonate (PFOS) and perfluorohexane sulfonate (PFHxS) (e.g.ref. 16–18). In estuaries, previous studies have also indicated cases where higher concentrations of PFAAs were observed in crustaceans,19,20 relative to fish.

There are considerable data gaps and uncertainty about potential toxic impacts of PFAAs on aquatic biota at concentrations normally observed in estuaries. However, consumption of exploited aquatic species (i.e. seafood) is often considered to be an important exposure pathway for humans, and as a consequence the measurement of PFAA concentrations in exploited species is important in defining exposure risk.21–23 Of the work undertaken on this issue, many studies are of limited spatial and temporal breadth, but quantifying spatial and temporal variability is essential for understanding exposure pathways for aquatic species, and exposure risk through seafood consumption, over time. Also, many aquatic species are known to migrate extensively, particularly estuarine and marine species, meaning they can be exposed in one place and harvested in another. Despite this, assessments of PFAA contamination rarely take animal movement patterns into account, and as such assessment of PFAA exposure risk from seafood harvested away from identified point sources is often overlooked.

Eastern School Prawn (Metapenaeus macleayi, hereafter referred to as School Prawn) is an important exploited species in south-eastern Australia.24 The species is a short-lived (1–2 years), migratory penaeid prawn with a Type-2 life cycle (the most common life cycle for penaeid prawns25) including a juvenile estuarine phase and a oceanic adult phase. School Prawn is harvested across both these components of their life history,26 and generally transition between their estuarine nursery and adjacent inshore waters abruptly and predictably around the new moon primarily between January and April lunar months.27 School Prawn has been shown to accumulate PFAAs in contaminated estuaries,17,20 but can also depurate these contaminants.28 Using School Prawn as a case study, this study provides novel analyses in support of several objectives that are relevant to the assessment of PFAA contamination in aquatic systems. Specifically, this included (1) the relationship between surface water and biota PFAA concentration in relation to flood conditions; (2) seasonal variability in PFAA concentrations; (3) interannual variability in PFAA concentrations; and (4) spatial variability in PFAA concentrations.

In addition, noting the lack of PFAA (and other contaminant) assessments that incorporate animal movement patterns, this study explores how knowledge of animal migration patterns can be employed to assess PFAA exposure risk in areas away from identified point sources. Specifically, objective (5) presents a migratory model for School Prawn to demonstrate how animal migration data (in this case, derived using stable isotope tracers) can be drawn together with PFAA concentrations measured near known estuarine point sources to quantify PFAA concentrations in distant parts of the fishery.

Methods

Study location and Hunter River School Prawn

The Hunter River estuary is a wave dominated estuarine system on the mid-north coast of New South Wales, Australia. The catchment is primarily agricultural and forested, however the lower estuary is highly industrialised and urbanised around the City of Newcastle and the middle and lower estuary supports an important fishery (Fig. 129). In particular, Fullerton Cove represents an extensive estuarine embayment on the north arm of the lower estuary and provides extensive fishing grounds for School Prawn, and other species. The Royal Australian Airforce (RAAF) base at Williamtown to the north of the estuary17 is contaminated with PFASs following historic firefighting training with aqueous film-forming-foam (AFFF3), and PFASs migrating southward in surface water generally discharge into the north-east quadrant of Fullerton Cove (which is the primary point-source of PFASs into the estuary). PFAAs were first detected in biota in the Hunter River estuary in 2015.17
image file: c9em00202b-f1.tif
Fig. 1 The Hunter River estuary, and key features of the estuary and surrounding area relevant to the current study (places names are included in italics). Location numbers are indicated on the map and correspond to those listed in Table 1 and referred to in the text. For each location, the relative contribution of School Prawn originating at that location to the inshore area (i.e. Stockton Bight Trawl Grounds) determined using stable isotope tracers is indicated as a circle (see figure legend). Mangrove and saltmarsh habitats are also indicated as green and light brown polygons (respectively). The inset panel indicates the location of the study estuary on the mid north coast of NSW.

As noted in the Introduction, School Prawn display a Type-2 life cycle30 with an estuarine nursery phase and an oceanic phase. In the Hunter River estuary and adjacent inshore oceanic areas, School Prawn are targeted by commercial fishers using otter trawl in both phases, with estuarine fishing concentrated along the north arm of the estuary and Fullerton Cove, and oceanic fishing occurring in the Stockton Bight Trawl Grounds (Fig. 1). Estuarine School Prawn abruptly run to sea in the last quarter of the moon,27 and appear in the Stockton Bight Trawl Grounds shortly thereafter.31 The Stockton Bight Trawl Grounds essentially represent a mixed group of prawns that have recently emigrated from different areas of the estuary. This abrupt migration of School Prawn has underpinned the development of a novel approach to assign emigrating prawns to originating locations in the estuary using carbon and nitrogen stable isotope tracers,32 as described below.

Sample collection for PFAA analysis

This study presents a synthesis of samples and chemical analysis data from multiple sampling events across 2015–2017. For all sampling events, School Prawn were sampled either using an otter trawl or beam trawl. Sampling targeted 6 general locations within the Hunter River estuary, and the Stockton Bight Trawl Grounds. For all sampling events, samples were stored on ice following landing, before being sorted and processed (dissected for muscle tissue and composited) at the laboratory and analysed for PFAAs (described below). Four composites of 20 School Prawn were nominally analysed for PFAAs from each sampling time point and each location, to ensure concentration estimates were representative (i.e. to ensure information was obtained from enough individuals to provide a reasonable estimate of summary statistics given inter-individual variability in PFAS concentrations14). Samples for analysis of PFAAs were collected from Fullerton Cove (location 1) and the broader Hunter River (locations 2–6) as outlined in Table 1, and each component of the data was used to address respective objectives as outlined below.
Table 1 Summary information outlining PFAA sampling details and stable isotope data. Sampling dates are in the format YYYY/MM, and the objectives that the data from each sampling event was used to assess is indicated in brackets (with numbers corresponding to numbered objectives in the Introduction). Location refers to locations outlined in Fig. 1. Where a range of dates is presented, it indicates monthly samples were collected within that range
Region Location PFAA sampling dates δ 15N δ 13C
Estuary 1 (Fullerton Cove) 2015/09 (3) 10.9 ± 0.1 −17.4 ± 0.2
2015/11 (2, 3, 4, 5)
2016/04 (3)
2016/12–2017/04 (1, 2, 3, 4)
2 2015/11 (4, 5) 11.3 ± 0.1 −17.6 ± 0.1
3 2015/11 (4, 5) 11.7 ± 0.2 −21.5 ± 0.2
2016/12–2017/04 (1, 3, 4)
4 2015/11 (4, 5) 12.3 ± 0.2 −22.0 ± 0.3
5 2015/11 (4, 5) 14.6 ± 0.1 −23.9 ± 0.1
6 2015/11 (3, 4, 5) 11.7 ± 0.1 −16.2 ± 0.2
2016/12 (3, 4)
Ocean Stockton Bight Trawl Grounds 2015/12 (5)


Sampling to assess the relationship between surface water and biota PFAA concentrations during flood conditions (Objective 1) was designed to coincide with a high-rainfall event that mobilised surface water across the flood plain at Williamtown and into the Hunter River estuary. Samples of surface water were collected at 3 time points at both the main point-source of PFAS pollution into the Hunter River estuary (the exit of the Ring Drain into Fullerton Cove, Fig. 1), as well as the main river channel (at location 3, Fig. 1). Surface water samples were collected at these locations in HDPE bottles, with one sampling time-point just prior to the rainfall event, and two sampling time points following the rainfall event. Logger stations (Odyssey®, Dataflow Systems Ltd., Christchurch, New Zealand) were deployed at these locations to measure conductivity and temperature (the Ring Drain logger was deployed on the landward side of the floodgate), and recorded a time-series of conditions near the water surface to track changes in conditions as freshwater run-off moved into the estuary.

To assess seasonal variability in PFAA concentrations (Objective 2), School Prawn samples were collected in Fullerton Cove (location 1), over multiple months between September and April, covering the main period that School Prawn use the estuarine nursery.26 A continuous random predictor variable was created to reflect the number of days from the start of September that had elapsed when each sample was collected. This variable was used to model potential non-linear changes in PFOS (as the only PFAA consistently measured and detected in prawn muscle tissue) tissue concentrations through time over this period, through fitting of a generalised additive model (GAM) using the gam function in the mgcv package in R v. 3.5.2.33

To assess interannual and spatial variability in PFAA concentrations, School Prawn samples were collected from adjacent to the main point source to the Hunter River estuary (Fullerton Cove, location 1), and two locations adjacent to significant industrial areas along the lower estuary (location 3 and location 6, Fig. 1), in November 2015 and December 2016. A two-factor analysis-of-variance was used to compare concentrations among sampling locations (locations 1, 3 and 6), and sampling years (2015 and 2016). Objective 4 was further supported by comparison of concentrations among all six locations in the Hunter River estuary collected in November 2015 (see Table 1).

PFAA analyses

Analysis for PFAAs was based on reference method USEPA 537,34 using the isotope dilution approach, and readers are referred to recent publications reporting the full methodology for analysis of aquatic biota and water samples (e.g.ref. 17 and 20). Samples up to March 2015 were analysed for perfluorooctanoic acid (PFOA), perfluorooctane sulfonate (PFOS), C2H4-perfluorooctane sulfonate (6[thin space (1/6-em)]:[thin space (1/6-em)]2 FTSA), and C2H4-perfluorodecane sulfonate (8[thin space (1/6-em)]:[thin space (1/6-em)]2 FTSA) only, whereas samples collected after this point were analysed for an expanded suite of additional PFASs including perfluorohexanoic acid, (PFHxA), perfluoroheptanoic acid (PFHpA), perfluorononanoic acid (PFNA), perfluorobutane sulfonate (PFBS) and perfluorohexane sulfonate (PFHxS). Qualitative/quantitative analysis for analytes was undertaken using an Agilent 1100 HPLC and a triple quadrupole mass spectrometer (ABSciex 4000 Qtrap). Multiple reaction monitoring (MRM) of two characteristic transitions was employed to confirm identification through detection of target ions in each MRM within known retention time windows. The limits-of-reporting (LORs) for each analyte in each sample was determined from noise and laboratory blank levels. Concentrations were reported on a wet weight basis and corrected for recovery of 13C isotopically labelled surrogates (to overcome matrix suppression/enhancement).

Using animal migration data to predict PFAA concentrations

To demonstrate the potential utility of animal migration data in assessments of PFAAs in seafood, data which quantifies the School Prawn migration patterns described above was synthesised with PFOS concentrations (as the dominant PFAA) measured in School Prawn in the contaminated estuarine nursery habitat, to estimate expected PFOS concentrations in School Prawn outside the estuary following migration to the Stockton Bight Trawl Grounds. The estimation approach involves a simple model that recognises the concentration of PFAAs (in this case PFOS) in prawns in the Stockton Bight Trawl Grounds following migration, is a function of the contribution of individuals exposed to PFAAs in the nursery habitats from which they emigrated. We validated model estimates with actual PFAA concentrations measured from School Prawn captured in the Stockton Bight Trawl Grounds following migration. Migration data was obtained using a stable isotope tracer approach which has previously been applied in studies of prawn migration,35,36 which involves assigning the origin of prawns emigrating from the estuarine nursery by matching the isotopic composition of these prawns to the isotopic composition of prawns collected at known locations-of-origin throughout the estuarine nursery.32

Samples for stable isotope analysis and assignment of origin for emigrating prawns were collected from January to March 2015, and included the collection of School Prawn samples from potential locations-of-origin throughout the estuarine nursery (Fig. 1 and 2) collecting emigrating prawns as they moved through the estuary mouth. Samples were collected from each location within the estuary (∼3 sites within each location) during the last quarter of the January 2015 lunar month, and muscle tissues were pooled into composites of 6 individuals for stable isotope analysis. Pooling of individuals into samples for each location-of-origin ensures that a sufficient number of individuals are analysed for the data to be representative of that location, within a reasonable number of analyses. Collection of emigrating prawns from the estuary mouth occurred over 3–6 nights during the last quarter of each of the January–March 2015 lunar months. Emigrating prawns were analysed as individual samples for stable isotope composition (n = 185), samples were not pooled (as they were for the locations-of-origin) as assignment occurs at the individual level. Tissue samples (composites from each location-of-origin, and individual emigrating prawns) were prepared for stable isotope analysis using conventional methods, and isotopic composition (15N and 13C) was measured on a Sercon 20–20 isotope ratio mass spectrometer (IRMS, Cheshire, United Kingdom). International standards were used to derive delta values for carbon and nitrogen isotopes.37 Isotope data were subsequently analysed using the Stable Isotope Analysis in R (SIAR38) package to calculate proportional contribution (and associated error metrics) of each area of origin in the estuary to emigrating prawns as per the approach of Taylor et al.32 This generated a mean and variance estimate of the proportional contribution (Pl) of prawns from each location-of-origin to the adjacent Stockton Bight Trawl Grounds. These summary statistics were used to derive a normal distribution for each Pl which was sampled in the simulations as outlined below.


image file: c9em00202b-f2.tif
Fig. 2 The relationship between conductivity (solid black line), biota PFOS (red circles) and PFHxS (green circles) tissue concentrations (mean ± SE), and monthly (mean ± SE for values >LOR) surface water ΣPFAA concentration (blue squares; note only 1 sample was collected in April for location 1, and values did not exceed LOR for location 3) for the main PFAS point source into the Hunter River at Fullerton Cove (upper panel, location 1 [see Fig. 1]) and the river channel (lower panel, location 3 [see Fig. 1]). Temperature is also indicated as a dashed black line.

The contribution data derived from isotope tracers were linked with PFOS concentrations measured in prawns at each of these locations-of-origin within a Monte-Carlo Analysis of Uncertainty (MCAoU) framework. The model derived the estimated PFOS concentration (PFOSest) of School Prawn within the Stockton Bight Trawl Grounds using the equation image file: c9em00202b-t1.tif where PFOSl is the PFOS concentration at location l, and n is the number of locations-of-origin in the estuary. Both Pl and PFOSl were specified as normal distributions for each location l, and independently sampled in each of 10[thin space (1/6-em)]000 simulations. Simulation data (PFOSest) was expressed as a probability distribution, and compared with actual PFOS concentrations of School Prawn from the Stockton Bight Trawl Grounds.

Results

General comments

Perfluorooctane sulfonate (PFOS; LOR = 0.3 μg kg−1, recovery 64–148%) was detected in 100% of School Prawn samples from the Hunter River that were analysed for PFAAs. Perfluorooctanoic acid (PFOA; LOR = 0.5 μg kg−1, recovery 61–183%) was detected in ∼16% of samples, and perfluorohexane sulfonate (PFHxS; LOR = 0.5 μg kg−1, recovery 39–202%) was detected in ∼51% of samples that were tested for the expanded suite of PFAAs (n = 75). For samples in which PFHxS was tested and detected, there was a significant positive relationship between PFOS and PFHxS concentrations (β = 0.24, F1.36 = 57.3, P ≪ 0.001). This indicates that where PFHxS was detected, on average this analyte was present at ∼24% of the concentration of PFOS. No other PFAAs were detected in School Prawn samples tested for the expanded suite of PFAAs. Surface water samples collected at the Ring Drain (Fig. 1) were dominated by PFOS (100% detection frequency) and PFHxS (80% detection frequency), but perfluorobutane sulfonate (PFBS; LOR = 0.02 μg kg−1), perfluoropentane sulfonate (PFPeS; LOR = 0.02 μg kg−1), perfluorohexanoic acid (PFHxA; LOR = 0.02 μg kg−1) and perfluoroheptane sulfonate (PFHpS; LOR = 0.02 μg kg−1) were also present at levels just above LOR in some samples.

Surface water and School Prawn PFAA concentrations

A heavy rainfall event occurred during late February 2017. Prior to this, conductivity within the Ring Drain was high, but decreased substantially from 26 February 2017 to 11 μs cm−1 by 8 March 2017 (Fig. 2). This indicates that water was moving off the floodplain at Williamtown into Fullerton Cove (location 1) by this time. Conductivity in the main river channel at location 3 also began to decrease from 26 February 2017, but did not drop sharply until 17 March 2017. Surface water PFAA concentrations in the Ring Drain increased from February to March, and again in April, however no PFAAs were detected at concentrations exceeding LOR at location 3 (Fig. 2). In contrast, School Prawn PFOS and PFHxS concentrations in Fullerton Cove (location 1) decreased substantially over the same period that surface water PFAA concentrations were increasing (Fig. 2). At location 3, School Prawn PFOS concentrations displayed marginal changes over this time frame, and PFHxS concentrations remained <LOR.

Seasonal, inter-annual, and spatial variability in School Prawn PFAA concentrations

There was a relationship between the muscle tissue PFOS concentration in School Prawn collected from Fullerton Cove (location 1) and elapsed time through the recruitment season (Fig. 3). Analysis of this relationship using a generalised additive model identified that this was a statistically significant non-linear relationship (F2.189 = 33.44, P ≪ 0.001, Fig. 3), and time elapsed through the season explained ∼57% of the variation in PFOS concentration. When multiple locations and years were examined, there was a significant difference in School Prawn PFOS concentrations among locations (F2.42 = 183.76, P ≪ 0.001) and between years (F1.42 = 34.64, P ≪ 0.001), but with a significant interaction between these factors (F2.42 = 31.11, P ≪ 0.001). Tukeys test indicated this interaction was driven by concentrations detected at location 6 (Fig. 4). Fullerton Cove (location 1) and location 6 had similar PFOS concentrations in the 2015/16 season, whereas concentrations at location 6 in the 2016/17 season were closer to concentrations at location 3. There was no appreciable difference in concentrations between years at Fullerton Cove and location 3 (Fig. 4). When other locations throughout the estuary were considered (2015/16 data only, Fig. 5), concentrations along the broader estuary decreased rapidly with increasing distance from Fullerton Cove (location 1). Location 6 had the most variable concentrations of all locations sampled (Fig. 5).
image file: c9em00202b-f3.tif
Fig. 3 Seasonal variation in PFOS concentrations for School Prawn collected from Fullerton Cove (location 1 [see Fig. 1]) across 2015–2017. The influence of the smoothed predictor variable (time since September, see text) on PFOS concentration is overlaid (solid black line), and 95% confidence intervals are indicated as dotted red lines.

image file: c9em00202b-f4.tif
Fig. 4 Spatial and inter-annual variation in PFOS concentrations for School Prawn collected at three locations in the Hunter River estuary (locations indicated in Fig. 1).

image file: c9em00202b-f5.tif
Fig. 5 Box-plot showing PFOS concentrations for School Prawn collected at 6 locations across the Hunter River estuary (locations correspond to those indicated in Fig. 1) in November 2015. Orange circles indicate mean values for each location.

Prediction of PFOS concentrations based on migration data

PFOS concentrations detected in School Prawn from the Stockton Bight Trawl Grounds in 2015/16 were 5.2 ± 1.7 μg kg−1 (mean ± SE; 100% detection frequency), but PFHxS was also present at lower concentrations (1.3 ± 0.5 μg kg−1, mean ± SE; 50% detection frequency). Assignment methodology indicated asymmetric contribution of locations-of-origin to the emigrating School Prawn population (Fig. 1). Locations 3 and 4 had the highest contributions, whereas other locations had proportionally similar contributions to School Prawn emigrating to the Stockton Bight Trawl Grounds. This data was successfully synthesised with location-specific PFOS concentrations, to predict expected PFOS concentrations (PFOSest) in the Stockton Bight Trawl Grounds (Fig. 6). Consideration of these data alongside actual PFOS concentrations measured in the Stockton Bight indicated that the predicted and observed data were very similar, and that the PFOSest mean was within the QA/QC threshold employed for the PFAA analyses (±30% of the actual mean34).
image file: c9em00202b-f6.tif
Fig. 6 Outcomes of model simulations predicting PFOS concentrations in the mixed population of School Prawn in the Stockton Bight Trawl Grounds (solid blue line, based on data collected in November 2015) that originated from the adjacent Hunter River estuary. Actual PFOS concentration data for School Prawn collected from the Stockton Bight Trawl Grounds in November 2015 is also indicated as an overlaid box-plot. Vertical dashed red lines indicate the ±30% QA/QC threshold for PFAA analyses.34

Discussion

The composite data set and associated analyses presented above highlight multiple factors that affect PFAA concentrations in School Prawn. There was substantial spatial variation in muscle tissue concentrations which was primarily dependent on distance to the major point source, but variability at location 6 indicates that other activities in the system may also influence PFOS concentrations (this is explored in more detail below). Temporal variation in environmental, behavioural and physiological factors similarly drives substantial variation in accumulation of PFAAs, which is evident across multiple temporal scales (month-to-month, and year-to-year). While School Prawn was chosen as model species to evaluate the influence of these factors, the patterns outlined in this study have much broader implications for studying bioaccumulation and toxicokinetics of PFAAs, and estimating exposure risk in free-ranging aquatic species.

Temporal patterns in concentrations

Previous research examining bioaccumulation of PFAAs in aquatic animals has sought to link surface water or pore-water PFAA concentrations through bioaccumulation factors, but these studies inherently assume that water is the main route of uptake for PFAAs. Khairy et al.39 presented bioaccumulation factors (BAF) for surface water and PFAAs in a range of species (including a crustacean), and showed reasonable agreement between lipid standardised PFAA concentration predicted from river water concentrations, and measured concentrations in fish and crab tissue. These relationships contrast with the patterns in the current study which indicated that tissue concentration in School Prawn decreased over the period when surface water concentrations at the point source were increasing. This suggests that either surface water is not a major PFAA exposure pathway for School Prawn (in lieu of food, or sediment exposure), or that other factors may have overshadowed any increase in uptake from elevated surface water concentrations.

The changes in tissue concentrations observed over seasonal time scales may provide one such explanation. Studies examining PFAA tissue concentrations in free-ranging animals over this scale are rare, and there is little supporting literature to suggest why this may occur in aquatic species. School Prawn may be considered an annual species, with the bulk of recruitment to the estuarine nursery occurring early in spring and early summer.24 Prawns grow throughout the season, periodically emigrating as described above27 As prawns get larger over the growing season, several things occur that could affect the uptake of PFAAs; (1) relative metabolic rate decreases as prawns get larger; (2) relative growth rate decreases as prawns get larger;40 (3) temperature decreases (late in the season), which can affect both (1) (e.g. Kulkarni and Joshi41) and (2) (e.g.ref. 42); (4) other biochemical changes may occur in the prawn throughout the season; and (5) there may be changes in PFAA concentrations in the prawns food. While there is no specific evidence for (1) above in School Prawn, metabolic rate scales with size in penaeid prawns (e.g. Kulkarni and Joshi41), and evidence for (2) is found in general sigmoidal relationship between mass and age evident in most aquatic animals including prawns.30 These changes through the warm season contribute to decreases in both relative food consumption and mass-standardised metabolic rate.43 There is also some evidence which suggests that PFAA tissue concentration in estuarine fish and crustaceans scales with relative metabolic rate (M.D. Taylor; pers. obs.), and this provides a potential mechanistic explanation for the changes observed in School Prawn. Other biochemical changes may be occurring as prawns grow and mature which could affect PFAA tissue concentrations, such as changes in tissue composition (e.g. protein or lipid content). Finally, changes in PFAA concentrations in the food web may affect the patterns observed. In estuaries, School Prawn generally feed on bivalve molluscs, crustaceans and annelid worms,44 however there is little information available to further interpret relationships between PFAA concentrations in food and its influence on changes in School Prawn PFAA concentration. In addition, a potential change in the composition of the food web could occur alongside the shift in conductivity during and after a flood, with prey (or the prawns themselves) emigrating to Fullerton Cove from less contaminated areas elsewhere in the system. This could also explain the post-flood patterns described above.

It is important to highlight other factors which could also contribute to the contrasting findings of Khairy et al.39 and the current study. For example, our study occurred in a much more saline environment; PFAAs are known to partition out of the aqueous phase as salinity increases,45 and this may occur close to the point source as the outward flowing surface water meets incoming saline tidal water in Fullerton Cove. Also, surface water PFAA concentrations were much higher in Khairy et al.39 than the current study. While bioaccumulation through surface water exposure may be occurring, it appears that this exposure pathway may not be sufficient to offset other concurrent physiological or ecological factors affecting tissue concentrations.

In general, few studies present seasonal and inter-annual comparisons of PFAA concentrations in estuarine and marine species.9 Despite the within-year variation in PFOS concentrations observed, at two-thirds of the locations there was very little between-year variation in the period sampled (which notably only covered two years). The exception was location 6, where School Prawn had 6× the PFOS concentration in 2015/16 than that measured in 2016/17. At location 6, the Hunter River is surrounded by a large industrial area, and the City of Newcastle lies directly to the south. Consequently, this part of the estuary could possibly be impacted by other contaminant sources, which could be intermittently mobilised by industrial activities or through disturbance occurring during development and construction activity. In addition, the lower south arm and mouth of the Hunter River is also heavily dredged on an ongoing basis. The removal of sediment may influence year-to-year variability in this area, and could explain the high variability in PFOS concentrations at location 6.

Spatial patterns in concentrations and incorporating migration data in assessments

Published examples that provide comprehensive spatial analyses of PFAA concentrations are rare for marine animals,20 although such analyses are useful in the characterisation of contaminant sources. The extensive spatial sampling presented here identified a high degree of variation in School Prawn PFOS concentrations across the Hunter River. The highest concentrations were measured in Fullerton Cove (which is the main point source into the estuary) and location 6 (which is subject to additional urban and industrial sources, and ongoing dredging). There was a relatively consistent decline in concentration along the north arm of the estuary with increasing distance to source (locations 1, 2 and 3), which is consistent with this area being influenced by the main point source in Fullerton Cove. This is further supported by the water PFAA concentrations, for which there was no increase detected during flood conditions suggesting minimal inputs to this part of the estuary. In contrast, School Prawn from location 6 had highly variable PFOS concentrations; potential reasons for this are provided above.

This is the first study that has incorporated animal migration data into assessments of PFAAs in a free-ranging aquatic species. While the modelled estimates of PFOS concentration presented here is largely species-specific and relevant to the time and locations sampled, the case study provides a clear example of how the synthesis of these data could be employed to improve assessments, particularly in areas away from known point sources. Animal movement and migration studies are a fundamental field of investigation in fisheries ecology, and underpins many aspects of fisheries management and resource assessment.46–48 Consequently, there is considerable data already available describing movement and migration patterns of exploited aquatic species. Recent work has provided quantitative evidence highlighting that such data are useful for predicting contaminant concentrations in exposed aquatic animals.49 Even with the relatively simple model presented here, employing only the concentrations for animals in contaminated source areas, and the origin of animals prior to emigration, PFOS concentrations in a separate component of the fishery were predicted with remarkable accuracy. Other data streams could potentially enhance this predictive capacity – such as incorporation of depuration rates (e.g.ref. 28) – which would allow concentration in non-contaminated areas to be modelled through time.

In terms of managing exposure risk through consumption of School Prawn, the outcomes of the model has few implications for management, since there were only low levels of PFAAs in School Prawn in the Stockton Bight Trawl Grounds, and this product is usually mixed with product from multiple other sources at the point of sale. For prawns in general, a majority of species exhibit Type-2 life-histories encompassing an estuarine juvenile stage and an oceanic adult stage,25 which means there is the potential for prawns to be exposed in contaminated estuaries, and harvested elsewhere. The approach is also applicable for other aquatic taxa that tend to accumulate higher concentrations of PFAAs (e.g. portunid crabs and Sea Mullet Mugil cephalus20), and display similar migratory patterns and aggregations away from contaminated source areas to those observed for School Prawn. Assessments for other highly migratory species whose life histories span rivers, estuaries and oceanic habitats, such as salmonids, could equally benefit from this approach.50,51 In conclusion, such analysis presents a powerful tool for assessing exposure risk through consumption of seafood from areas away from locations where animals are directly exposed to known sources of PFAAs.

Synthesis and implications

This study demonstrates the magnitude of temporal and spatial variation in PFAA concentrations in a model estuarine crustacean species, and highlights important considerations for (1) assessing PFAA exposure risk through seafood, and (2) understanding factors affecting PFAA concentrations in free-ranging animals. Studies to quantify bioaccumulation are often performed within a laboratory or aquarium setting, which is useful for evaluating variables under controlled conditions. The results of these experiments should also be supported by studies examining patterns in natural systems which are subject to multiple extrinsic factors that can affect PFAA tissue concentrations. Clearly, physiological and behavioural traits have an important role in driving PFAA concentrations in aquatic organisms, and this help explain the variability observed in free ranging animals. Importantly, this study also suggests a relatively poor correlation between surface water concentration and PFAA concentrations in free-ranging School Prawn, due to the influence of other factors. This means that PFAA concentrations estimated from surface water concentrations in aquatic biota could misrepresent exposure risk. Furthermore, incorporating animal migration data in assessments of PFAAs in aquatic animals presents a useful tool to improve evaluation of exposure risk, particularly where species can migrate from contaminated areas and be harvested elsewhere. Ultimately, resolution of the interacting factors affecting PFAA tissue concentrations in aquatic biota in general would benefit from further work on the relationships between uptake, depuration, growth, movement and metabolic rate.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

The author wishes to acknowledge the New South Wales PFAS Expert Panel, the New South Wales Environment Protection Authority, as well as the Australian Department of Defence, and their consultants for Williamtown RAAF Base, AECOM. The authors also wish to acknowledge the contribution of E. Mitchell, D. Cruz, J. McLeod, J. Hewitt, and G. Hyde for technical and logistical support throughout the program. Work was carried out under NSW DPI Animal Care and Ethics permit 14-11, and collection of animals was permitted under the New South Wales Fisheries Management Act 1994 through s37 permit number P01/0059.

References

  1. L. Ahrens, Polyfluoroalkyl compounds in the aquatic environment: a review of their occurrence and fate, J. Environ. Monit., 2011, 13, 20–31 RSC.
  2. J. P. Giesy, J. E. Naile, J. S. Khim, P. D. Jones and J. L. Newsted, in Reviews of Environmental Contamination and Toxicology, Springer New York, New York, NY, 2010, pp. 1–52 Search PubMed.
  3. M. D. Taylor and I. Cesenza, A toxic legacy from firefighting foams, Aust. Sci., 2016, 37, 39 Search PubMed.
  4. M. Filipovic, A. Woldegiorgis, K. Norström, M. Bibi, M. Lindberg and A.-H. Österås, Historical usage of aqueous film forming foam: A case study of the widespread distribution of perfluoroalkyl acids from a military airport to groundwater, lakes, soils and fish, Chemosphere, 2015, 129, 39–45 CrossRef CAS PubMed.
  5. C. Baduel, J. F. Mueller, A. Rotander, J. Corfield and M.-J. Gomez-Ramos, Discovery of novel per- and polyfluoroalkyl substances (PFASs) at a fire fighting training ground and preliminary investigation of their fate and mobility, Chemosphere, 2017, 185, 1030–1038 CrossRef CAS PubMed.
  6. C. Gallen, C. Baduel, F. Y. Lai, K. Thompson, J. Thompson, M. Warne and J. F. Mueller, Spatio-temporal assessment of perfluorinated compounds in the Brisbane River system, Australia: Impact of a major flood event, Mar. Pollut. Bull., 2014, 85, 597–605 CrossRef CAS PubMed.
  7. J. Thompson, G. Eaglesham and J. Mueller, Concentrations of PFOS, PFOA and other perfluorinated alkyl acids in Australian drinking water, Chemosphere, 2011, 83, 1320–1325 CrossRef CAS PubMed.
  8. M. D. Taylor, First reports of poly- and perfluoroalkyl substances (PFASs) in Australian native and introduced freshwater fish and crustaceans, Mar. Freshwater Res., 2018, 69, 628–634 CrossRef CAS.
  9. S. Hong, J. S. Khim, T. Wang, J. E. Naile, J. Park, B.-O. Kwon, S. J. Song, J. Ryu, G. Codling, P. D. Jones, Y. Lu and J. P. Giesy, Bioaccumulation characteristics of perfluoroalkyl acids (PFAAs) in coastal organisms from the west coast of South Korea, Chemosphere, 2015, 129, 157–163 CrossRef CAS PubMed.
  10. G. L. Lescord, K. A. Kidd, A. O. De Silva, M. Williamson, C. Spencer, X. Wang and D. C. G. Muir, Perfluorinated and polyfluorinated compounds in lake food webs from the Canadian High Arctic, Environ. Sci. Technol., 2015, 49, 2694–2702 CrossRef CAS PubMed.
  11. J. P. Giesy and K. Kannan, Global distribution of perfluorooctane sulfonate in wildlife, Environ. Sci. Technol., 2001, 35, 1339–1342 CrossRef CAS PubMed.
  12. H. Yoo, N. Yamashita, S. Taniyasu, K. T. Lee, P. D. Jones, J. L. Newsted, J. S. Khim and J. P. Giesy, Perfluoroalkyl acids in marine organisms from Lake Shihwa, Korea, Arch. Environ. Contam. Toxicol., 2009, 57, 552–560 CrossRef CAS PubMed.
  13. S. Taniyasu, K. Kannan, Y. Horii, N. Hanari and N. Yamashita, A survey of perfluorooctane sulfonate and related perfluorinated organic compounds in water, fish, birds, and humans from Japan, Environ. Sci. Technol., 2003, 37, 2634–2639 CrossRef CAS PubMed.
  14. M. D. Taylor, Survey design for quantifying perfluoroalkyl acid concentrations in fish, prawns and crabs to assess human health risks, Sci. Total Environ., 2019, 652, 59–65 CrossRef PubMed.
  15. W. A. O'Connor, A. Zammit, M. C. Dove, G. Stevenson and M. D. Taylor, First observations of perfluorooctane sulfonate occurrence and depuration from Sydney Rock Oyster, Saccostrea glomerata, in Port Stephens NSW Australia, Mar. Pollut. Bull., 2018, 127, 207–210 CrossRef PubMed.
  16. J. E. Naile, J. S. Khim, S. Hong, J. Park, B.-O. Kwon, J. S. Ryu, J. H. Hwang, P. D. Jones and J. P. Giesy, Distributions and bioconcentration characteristics of perfluorinated compounds in environmental samples collected from the west coast of Korea, Chemosphere, 2013, 90, 387–394 CrossRef CAS PubMed.
  17. M. D. Taylor and D. D. Johnson, Preliminary investigation of perfluoroalkyl substances in exploited fishes of two contaminated estuaries, Mar. Pollut. Bull., 2016, 111, 509–513 CrossRef CAS PubMed.
  18. C. Baduel, F. Y. Lai, K. Townsend and J. F. Mueller, Size and age–concentration relationships for perfluoroalkyl substances in stingray livers from eastern Australia, Sci. Total Environ., 2014, 496, 523–530 CrossRef CAS PubMed.
  19. M. Habibullah-Al-Mamun, M. K. Ahmed, M. Raknuzzaman, M. S. Islam, M. M. Ali, M. Tokumura and S. Masunaga, Occurrence and assessment of perfluoroalkyl acids (PFAAs) in commonly consumed seafood from the coastal area of Bangladesh, Mar. Pollut. Bull., 2017, 124, 775–785 CrossRef CAS PubMed.
  20. M. D. Taylor, J. Beyer-Robson, D. D. Johnson, N. A. Knott and K. C. Bowles, Bioaccumulation of perfluoroalkyl substances in exploited fish and crustaceans: Spatial trends across two estuarine systems, Mar. Pollut. Bull., 2018, 131, 303–313 CrossRef CAS PubMed.
  21. U. Berger, A. Glynn, K. E. Holmström, M. Berglund, E. H. Ankarberg and A. Törnkvist, Fish consumption as a source of human exposure to perfluorinated alkyl substances in Sweden-Analysis of edible fish from Lake Vättern and the Baltic Sea, Chemosphere, 2009, 76, 799–804 CrossRef CAS PubMed.
  22. K. Y. Christensen, M. Raymond, M. Blackowicz, Y. Liu, B. A. Thompson, H. A. Anderson and M. Turyk, Perfluoroalkyl substances and fish consumption, Environ. Res., 2017, 154, 145–151 CrossRef CAS PubMed.
  23. W. D'Hollander, P. de Voogt, W. De Coen and L. Bervoets, in Reviews of Environmental Contamination and Toxicology Volume 208: Perfluorinated Alkylated Substances, ed. P. De Voogt, Springer, New York, NY, 2010, pp. 179–215 Search PubMed.
  24. M. D. Taylor and N. R. Loneragan, Catchment-derived stressors, recruitment, and fisheries productivity in an exploited penaeid shrimp, Reg. Stud. Mar. Sci., 2019, 29, 100628 CrossRef.
  25. W. Dall, B. Hill, P. Rothlisberg and D. Sharples, in The Biology of the Penaeidae, Academic Press, London, 1990, vol. 27, pp. 283–313 Search PubMed.
  26. C. Hart, T. F. Gaston and M. D. Taylor, Utilisation of a recovering wetland by a commercially important species of penaeid shrimp, Wetlands Ecol. Manage., 2018, 26(4), 665–675 CrossRef.
  27. A. A. Racek, Prawn investigations in eastern Australia, Res. Bull. State Fish, 1959, 6, 1–57 Search PubMed.
  28. M. D. Taylor, K. C. Bowles, D. D. Johnson and N. A. Moltschaniwskyj, Depuration of perfluoroalkyl substances from the edible tissues of wild-caught invertebrate species, Sci. Total Environ., 2017, 581, 258–267 CrossRef PubMed.
  29. M. D. Taylor, B. Fry, A. Becker and N. A. Moltschaniwskyj, The role of connectivity and physicochemical conditions in effective habitat of two exploited penaeid species, Ecol. Indic., 2017, 80, 1–11 CrossRef CAS.
  30. W. Dall, B. Hill, P. Rothlisberg and D. Sharples, The Biology of the Penaeidae, Academic Press, London, 1990 Search PubMed.
  31. N. V. Ruello, Migration and stock studies on Australian school prawn Metapenaeus macleayi, Mar. Biol., 1977, 41, 185–190 CrossRef.
  32. M. D. Taylor, J. A. Smith, C. A. Boys and H. Whitney, A rapid approach to evaluate putative nursery sites for penaeid prawns, J. Sea Res., 2016, 114, 26–31 CrossRef.
  33. S. Wood, Fast stable restricted maximum likelihood and marhinal likelihood estimation of semiparametric generalised liner models, J. R. Stat. Soc. Ser. B, 2011, 73, 3–36 CrossRef.
  34. J. A. Shoemaker, P. Grimmett and B. Boutin, Determination of selected perfluorinated alkyl acids in drinking water by Solid Phase Extraction and Liquid Chromatography/Tandem Mass Spectrometry (LC/MS/MS), U.S. Environmental Protection Agency, Washington, DC, 2008, https://cfpub.epa.gov/si/si_public_file_download.cfm?p_download_id=525468 Search PubMed.
  35. M. D. Taylor, A. Becker, N. A. Moltschaniwskyj and T. Gaston, Direct and indirect interactions between lower estuarine mangrove and saltmarsh habitats and a commercially important penaeid shrimp, Estuaries Coasts, 2018, 41(3), 815–826 CrossRef CAS.
  36. M. D. Taylor, B. Fry, A. Becker and N. A. Moltschaniwskyj, Recruitment and connectivity influence the role of seagrass as a penaeid nursery habitat in a wave dominated estuary, Sci. Total Environ., 2017, 584–585, 622–630 CrossRef CAS PubMed.
  37. B. Fry, Stable Isotope Ecology, Springer, New York, 2006 Search PubMed.
  38. A. C. Parnell, R. Inger, S. Bearhop and A. L. Jackson, Source partitioning using stable isotopes: coping with too much variation, PLoS One, 2010, 5, e9672 CrossRef PubMed.
  39. M. A. Khairy, G. O. Noonan and R. Lohmann, Uptake of hydrophobic organic compounds, including OCPs and PBDEs and perfluoroalkyl acids (PFAAs) in fish and blue crabs of the lower Passaic River (NJ, USA), Environ. Toxicol. Chem., 2019, 38(4), 872–882 CrossRef CAS PubMed.
  40. S. S. Montgomery, C. T. Walsh, M. Haddon, C. L. Kesby and D. D. Johnson, Using length data in the Schnute Model to describe growth in a metapenaeid from waters off Australia, Mar. Freshwater Res., 2010, 61, 1435–1445 CrossRef CAS.
  41. G. K. Kulkarni and P. K. Joshi, Some aspects of respiratory metabolism of a penaeid prawn, Penaeus japonicus (Bate) (Crustacea, Decapoda, Penaeidae), Hydrobiologia, 1980, 75, 27–32 CrossRef.
  42. L. R. Lloyd-Jones, Y.-G. Wang, A. J. Courtney, A. J. Prosser, S. S. Montgomery and Y. Chen, Latitudinal and seasonal effects on growth of the Australian eastern king prawn (Melicertus plebejus), Can. J. Fish. Aquat. Sci., 2012, 69, 1525–1538 CrossRef.
  43. J. A. Smith, B. J. Eyre, J. Rosentreter and M. D. Taylor, Modelling estuarine stocking density for crustaceans using net ecosystem metabolism, Bull. Mar. Sci., 2019, 95(2), 217–238 CrossRef.
  44. N. V. Ruello, Burrowing, feeding, and spatial distribution of the school prawn Metapenaeus macleayi (Haswell) in the Hunter River region, Australia, J. Exp. Mar. Biol. Ecol., 1973, 13, 189–206 CrossRef.
  45. S. Hong, J. S. Khim, J. Park, M. Kim, W.-K. Kim, J. Jung, S. Hyun, J.-G. Kim, H. Lee, H. J. Choi, G. Codling and J. P. Giesy, In situ fate and partitioning of waterborne perfluoroalkyl acids (PFAAs) in the Youngsan and Nakdong River Estuaries of South Korea, Sci. Total Environ., 2013, 445–446, 136–145 CrossRef CAS PubMed.
  46. M. D. Taylor, R. C. Babcock, C. A. Simpfendorfer and D. A. Crook, Where technology meets ecology: Acoustic telemetry in contemporary Australian aquatic research and management, Mar. Freshwater Res., 2017, 68, 1397–1402 CrossRef.
  47. M. D. Taylor, Acoustic telemetry technology in contemporary fisheries management: Perspectives of a government scientist, Bull. Ecol. Soc. Aust., 2018, 48, 6–8 Search PubMed.
  48. G. T. Crossin, M. R. Heupel, C. M. Holbrook, N. E. Hussey, S. K. Lowerre-Barbieri, V. M. Nguyen, G. D. Raby and S. J. Cooke, Acoustic telemetry and fisheries management, Ecol. Appl., 2017, 27, 1031–1049 CrossRef PubMed.
  49. M. D. Taylor, D. E. van der Meulen, S. Brodie, G. Cadiou and N. A. Knott, Applying acoustic telemetry to understand contaminant exposure and bioaccumulation patterns in mobile fishes, Sci. Total Environ., 2018, 625, 344–354 CrossRef CAS PubMed.
  50. A. M. Debruyn, M. G. Ikonomou and F. A. Gobas, Magnification and toxicity of PCBs, PCDDs, and PCDFs in upriver-migrating Pacific salmon, Environ. Sci. Technol., 2004, 38, 6217–6224 CrossRef CAS PubMed.
  51. G. Ewald, P. Larsson, H. Linge, L. Okla and N. Szarzi, Biotransport of organic pollutants to an Inland Alaska lake by migrating Sockeye Salmon (Oncorhynchus nerka), Arctic, 1998, 51, 40–47 CrossRef.

This journal is © The Royal Society of Chemistry 2019