Assessment tools for microplastics and natural fibres ingested by fish in an urbanised estuary

Assessment tools for microplastics and natural fibres ingested by fish in an urbanised estuary

Environmental Pollution 234 (2018) 552e561 Contents lists available at ScienceDirect Environmental Pollution journal homepage: www.elsevier.com/loca...

997KB Sizes 0 Downloads 2 Views

Environmental Pollution 234 (2018) 552e561

Contents lists available at ScienceDirect

Environmental Pollution journal homepage: www.elsevier.com/locate/envpol

Assessment tools for microplastics and natural fibres ingested by fish in an urbanised estuary* Jennifer E. Halstead a, *, James A. Smith a, Elizabeth A. Carter b, Peter A. Lay b, Emma L. Johnston a a b

Evolution & Ecology Research Centre, School of Biological, Earth & Environmental Sciences, University of New South Wales, Sydney, NSW 2052, Australia Vibrational Spectroscopy Core Facility, The School of Chemistry, The University of Sydney, NSW 2006, Australia

a r t i c l e i n f o

a b s t r a c t

Article history: Received 15 May 2017 Received in revised form 26 November 2017 Accepted 26 November 2017

Microplastics and fibres occur in high concentrations along urban coastlines, but the occurrence of microplastic ingestion by fishes in these areas requires further investigation. Herein, the ingestion of debris (i.e., synthetic and natural fibres and synthetic fragments of various polymer types) by three benthic-foraging fish species Acanthopagrus australis (yellowfin bream), Mugil cephalus (sea mullet) and Gerres subfasciatus (silverbiddy) in Sydney Harbour, Australia has been quantified and chemically speciated by vibrational spectroscopy to identify the polymer type. Ingested debris were quantified using gut content analysis, and identified using attenuated total reflectance Fourier transform infrared (ATRFTIR) and Raman microspectroscopies in combination with principal component analysis (PCA). The occurrence of debris ingestion at the time of sampling ranged from 21 to 64% for the three species, and the debris number ranged from 0.2 to 4.6 items per fish for the different species, with ~53% of debris being microplastic. There was a significant difference in the amount of debris ingested among species; however, there was no difference among species when debris counts were standardised to fish weight or gut content weight, indicating that these species ingest a similar concentration of debris relative to their ingestion rate of other material. ATR-FTIR microspectroscopy successfully identified 72% of debris. Raman spectroscopy contributed an additional 1% of successful identification. In addition, PCA was used to nonsubjectively classify the ATR-FTIR spectra resulting in the identification of an additional 9% of the debris. The most common microplastics found were polyester (PET), acrylic-polyester blend, and rayon (semisynthetic) fibres. The potential of using Raman microspectroscopy for debris identification was investigated and provided additional information about the nature of the debris as well as the presence of specific dyes (and hence potential toxicity). © 2017 Elsevier Ltd. All rights reserved.

Keywords: Microplastic Acanthopagrus australis Yellowfin bream Mugil cephalus Sea mullet Gerres subfasciatus Silverbiddy Ingestion Vibrational spectroscopy

1. Introduction Society's consumption of plastic products has increased exponentially in recent years. This consumption combined with inadequate waste management has positioned plastic as one of the most prevalent forms of anthropogenic pollution in the marine environment (Jambeck et al., 2015; Ivar do Sul and Costa, 2014). This form of pollution now has a global distribution (Bergmann et al., 2000; GESAMP, 2016; Helm, 2017) and is consumed by numerous species from a range of habitats with diverse feeding strategies; e.g.

*

This paper has been recommended for acceptance by Eddy Y. Zeng. * Corresponding author. E-mail address: [email protected] (J.E. Halstead).

https://doi.org/10.1016/j.envpol.2017.11.085 0269-7491/© 2017 Elsevier Ltd. All rights reserved.

pelagic and benthic fishes, (Davison and Asch, 2011; Choy and Drazen, 2013; Romeo et al., 2015; Mizraji et al., 2017; McGoran et al., 2017) filter feeders (De Witte et al., 2014) and benthic infauna (Van Cauwenberghe et al., 2015). This study applies the definition of microplastic as artificial polymers and plastics smaller than 5 mm (Arthur et al., 2009; Wright et al., 2013). They can be manufactured in this size class (e.g. beads or pre-production resin pellets), or can result from the fragmentation of larger plastic items. Fibrous synthetic particles (fibres) are a particularly prominent type of microplastic in the marine environment (Claessens et al., 2011). One major source of fibres is clothing, and 70 million tonnes of fibres are used in the global clothing industry yearly. Fibres from clothing, as well as other sources, can enter the marine environment through sewage, or stormwater outputs (Carr et al., 2016; Mintenig et al., 2017). This

J.E. Halstead et al. / Environmental Pollution 234 (2018) 552e561

fibre contamination typically outnumbers other microplastic types with fibres over three times more prevalent in areas that are downstream of wastewater outfalls (Browne et al., 2010) and close to larger urban centres (Dris et al., 2015). Microplastic particles, including fibres, are problematic because they are a form of contamination that is small enough to be ingested by marine animals (GESAMP, 2016; Bergmann et al. 2000; Helm, 2017). This is a concern because ingested polymer fragments of this size can transfer from prey to predators (Farrell and Nelson, 2013), and potentially act as a vector for delivery of toxic chemicals to organisms to cause adverse biological impacts (Chua et al., 2014; Browne et al., 2013, 2015; Rochman et al., 2013, 2015). It is important to note that the potential role of microplastics as vectors for toxic chemical loads is not conclusive, and requires further studies of long-term effects to determine the environmental relevance (Koelmans et al., 2016). Fish are known to ingest microplastic (Foekema et al., 2013; Lusher et al., 2013; Collard et al., 2015; Naidoo et al., 2016; Rummel et al., 2016; Mizraji et al., 2017), making them a potential vector of toxic chemicals through food chains and into human diets (Rochman et al., 2015). Due to typically high concentrations of microplastic (including both natural and synthetic fibres) along urbanised coasts (Mathalon and Hill, 2014), there is a need for more research in these areas to quantify and assess the extent of microplastic ingestion by marine biota, and the ecological signifi€ nnstedt and Eklo €v, 2016). cance of the phenomenon (Lo The three species investigated in this present study were yellowfin bream (Acanthopagrus australis; Owen, 1853), sea mullet (Mugil cephalus; Linnaeus, 1758) and silverbiddy (Gerres subfasciatus; Cuvier, 1830). Each of these species is widely distributed along eastern Australia in coastal areas and estuaries (Kailola et al., 1993), and each is commercially harvested, with sea mullet comprising the largest catch by weight of all fished species in NSW (Stewart et al., 2015). Yellowfin bream is also heavily fished recreationally (Stewart et al., 2015). In NSW, sea mullet and yellowfin bream are classified as ‘fully fished’ and silverbiddy as ‘uncertain’ (Stewart et al., 2015). Each of these three species is consumed by humans, in particular yellowfin bream, which is generally considered a table fish. These species share a general ‘benthic’ feeding strategy, with yellowfin bream eating mostly benthic invertebrates as well as some plant material and fish (Kailola et al., 1993; Hadwen et al., 2007; Truong et al., 2017), sea mullet consuming detritus and benthic microalgae as well as infauna largely indiscriminately (Kailola et al., 1993; Hadwen et al., 2007; Whitfield et al., 2012), and silverbiddy likely consuming benthic invertebrates as well some algae and sediment (based on a closely related species; Platell et al., 1999). This would likely expose them to microplastics and natural fibres on the substrate, and potentially accumulating within the sediment (Barnes et al., 2009; McGoran et al., 2017). Measuring rates of microplastic ingestion enables a better understanding of the threat of microplastics to marine food webs and possibly human health. Sydney is the most populated city in Australia with more than 4.6 million inhabitants (Johnston et al., 2015), and its harbour has been chronically contaminated with metals, organochlorines, hydrocarbons and dioxins (Birch, 2000; Birch and Taylor, 2002). To the authors knowledge there are no published studies to date that have investigated microplastic pollution in this Harbour (Mayer-Pinto, 2015). This lack of research poses a potential issue for properly evaluating the health of the ecosystem. It may also be of concern from a human health perspective, as more research is carried out on the risks of microplastic present in commonly ingested seafoods. Recreational fishing pressure in Sydney Harbour is almost double that in nearby estuaries, despite public health warnings over fish tissue contamination (Hedge et al., 2013).

553

One limitation to our understanding of the extent of microplastic and natural fibre pollution and ingestion has been the difficulty in accurately identifying the constituent polymers and chemical additives of microplastics and fibres. Previous microplastic research has commonly used visual examination for microplastic debris identification, which subsequently lacks information about the specific polymer type or chemical composition of the microplastic (Choy and Drazen, 2013; Romeo et al., 2015; Desforges et al., 2014). Over 200 different chemicals have been detected in marine microplastic sampling (Rani et al., 2015), and without a more detailed understanding of the polymer type, it is difficult to understand how harmful specific types of microplastic can be to biota (Lusher et al., 2013). Additionally, the subjective nature of visual identification means researchers may inaccurately estimate quantities of microplastics when assessing the stomach contents of commonly studied animals, such as fish (Collard et al., 2015), for example by not distinguishing between natural and synthetic fibres (Lenz et al., 2015; Remy et al., 2015), both of which have potential toxicity. In the past five years 50 published studies have used vibrational spectroscopy (Raman and infrared (IR)) to characterise and identify non-destructively the chemical composition of microplastics. Vibrational spectroscopy measures the energy that it takes for the bonds between atoms within molecules to vibrate. Raman and IR spectroscopy provide complementary information about the functional groups (e.g. C-H, C¼C, C¼O) present in a sample. FTIR (Fourier transform infrared) spectroscopy is recognised as a very effective technique for microplastic identification (Mecozzi et al., 2016), although many researchers have used infrared spectrometers and sampling accessories that are not entirely wellsuited to the collection of microscopic samples. Typical infrared sampling accessories used to analyse microplastics include; compression of the sample in a diamond anvil cell (DAC), production of a cast film, or the use of a macroscopic attenuated total reflectance (ATR) accessory (Obbard et al., 2014; Acosta-Coley and Olivero-Verbel, 2015; Nor and Obbard, 2014). More recent studies have employed FTIR microspectroscopy, which collects data using either transmission, reflectance or micro-ATR sampling modes (Song et al., 2014; K€ appler et al., 2016). FTIR microspectroscopy enables investigation of suitable samples as small as 10 to 20 microns, however this can be limited by the sample thickness. An ATR crystal mounted on a microscope objective (micro-ATR) allows samples of any thickness to be measured including dark (highly absorbing) samples as well as providing information about the sample surface (Kazarian and Chan, 2010). Raman microspectroscopy offers non-contact measurements of samples with a spatial resolution in the order of ~1 mm. A range of different laser sources (excitation lines) from the high-energy ultraviolet (UV) through to visible and near-IR (NIR) light can be used to generate a spectrum. It is well known that using visible light for Raman spectroscopic analysis of plastics can generate both a Raman signal as well as a strong fluorescence which is generally attributable to the pigments, additives or impurities within the sample (Siesler, 2011). To bleach or quench the fluorescence, which can mask Raman spectral features, a sample is often exposed to the laser for a longer time period. The use of a 785 nm wavelength excitation line is less likely to excite fluorescence and has a better scattering efficiency c.f. 1064 nm (Siesler, 2011). Microplastics research could benefit from applying multivariate statistical techniques (chemometrics) to the complex spectroscopic data obtained with current technology (Comnea-Stancu et al., 2017). Principal component analysis (PCA) is one such form of chemometrics that analyses the data variance and identifies the independent principal components (PC) within the data set. PCA generates a PC score and PC loadings plots, which together describe

554

J.E. Halstead et al. / Environmental Pollution 234 (2018) 552e561

the distribution of samples within the PCA output. The PC score allows for identification of spectra into similar or dissimilar patterns of variance, based on major spectral differences (Brereton, 2003; Adams, 2004). These analytical techniques have been increasingly applied to microplastic research in order to determine specific polymer types, discriminate between samples of plastic and non-plastic, identify dyes or other additives (Frias et al., 2016; Castillo et al., 2016; Enders et al., 2015; Van Cauwenberghe et al., 2013; Van Cauwenberghe and Janssen, 2014; Imhof et al., 2016). Identification to polymer type offers researchers the ability to more accurately identify the sources, pathways and possible impacts these contaminants pose to ecosystems (Browne et al., 2011, 2015). It is proposed using vibrational spectroscopy in combination with PCA will produce the maximum information possible regarding the chemical identification of microplastic material. The aims of this current study were to investigate the gut contents of three key fish species from Sydney Harbour to: 1) determine if the combined use of vibrational spectroscopy and chemometric analysis would provide a higher level of nonsubjective identification of ingested microplastics and natural fibres, than the use of vibrational spectroscopy alone and 2) determine if these fishes ingest microplastics and natural fibres, and if so how does the type of debris differ between species and diet.

2. Materials and methods 2.1. Study species and fish collection We selected three species for this study: yellowfin bream (Acanthopagrus australis), sea mullet (Mugil cephalus) and silverbiddy (Gerres subfasciatus). These were selected due to their prevalence in Sydney Harbour, their benthic feeding strategy, and importance for fisheries. Each of these species is widely distributed along eastern Australia in coastal areas and estuaries (Kailola et al., 1993), and each is commercially harvested, with sea mullet comprising the largest catch by weight of all fished species in NSW (Stewart et al., 2015). Yellowfin bream is also heavily fished recreationally (Stewart et al., 2015). In NSW, sea mullet and yellowfin bream are classified as ‘fully fished’ and silverbiddy as ‘uncertain’ (Stewart et al., 2015). Each of these three species is consumed by humans, in particular yellowfin bream, which is generally

considered a table fish. Fish were collected from the northern arm of Sydney Harbour, NSW, Australia (Fig. 1). A total of 24 yellowfin bream (average size of 11.1 cm and weight of 40 g) 45 sea mullet (average size of 25.2 cm and weight of 228 g) and 24 silverbiddy (average size of 6.8 cm and weight of 9 g) were caught using nets made from polyamide (nylon) material (beach seine, gill net, cast net) and hook-and-line fishing from three sites within Middle Harbour (Sugarloaf Bay 33.794010 S, 151.227136 E; Sailors Bay 33.803625 S, 151.226895 E; and Long Bay 33.815165 S, 151.231112 E). Caught fish were euthanised using Aqui-S solution and frozen for later stomach content analysis. Samples of net and line from the fishing gear were collected to test for possible fibre contamination during the capture of these fish. Fish were sampled on seven days between March and June 2015. Fish capture was distributed sporadically across sites and times, making an analysis of site or sampling period confounded, so all fish were grouped into a single analysis of Middle Harbour in an Autumn period. 2.2. Stomach content analysis Fish were defrosted, weighed and their digestive tracts were removed (stomach and intestine). The wet mass of the digestive tract was recorded with and without contents to estimate the wet weights of contents (Hyslop, 1980; Hadwen et al., 2007). The diet of each fish was determined using a volumetric analysis of diet components to coarse taxonomic resolution (i.e., sediment, algae, molluscs, crustaceans, worms, fish, insects, unidentified), with diet proportionally estimated using a gridded Petri dish. Potential debris were removed using metal forceps under a dissecting microscope, visually described, recorded and stored in glass vials for later spectroscopic analysis. ‘Debris’ in this study encompassed any suspected non-natural items, and included synthetic fragments and fibres (microplastics) but also anthropogenic particles made from natural materials (wool, cellulose). Debris were dried (50  C, 72 h) to obtain their dry mass. Strict procedural controls were implemented during laboratory work to minimise potential fibre contamination. A cotton laboratory coat and nitrile gloves were worn during dissections (Collard et al., 2015; Van Cauwenberghe and Janssen, 2014; Lusher et al., 2014). All equipment used during the laboratory work, including controls, was washed three times with pre-filtered water before use and between each dissection. Controls were set up for each individual fish dissection process. As each fish gut was removed, weighed and examined, a control Petri dish with a small amount of distilled water was present, to capture any airborne or handling-related contamination that may occur. These controls were examined for debris after each fish was processed to account for any contamination that the samples may be exposed to during the procedure. 2.3. Debris identification

Fig. 1. Sampling locations were confined to three embayments (SLB ¼ Sugarloaf Bay, SB ¼ Sailors Bay and LB ¼ Long Bay) in the northern arm of Sydney Harbour (Middle Harbour), depicted above in the boxed region.

ATR-FTIR spectroscopy was used as the primary tool for identifying debris found in gut contents. Details of the FTIR spectrometer and ATR accessory are in the Supporting Information. Spectra were acquired in triplicate from each sample to ensure reproducibility as the ATR microscope objective has a 100 mm diameter. Initial identification was achieved by comparing the debris sample spectra to those in Bruker FTIR spectral libraries of natural and synthetic fibres, polymers and natural compounds (Browne et al., 2011; Rummel et al., 2016; Mecozzi et al., 2016). Each match was then confirmed by a visual comparison of the spectra comparing the position, height, width and line-shape of the IR bands. A subsample of fibres found in the controls were also analysed using ATR-FTIR spectroscopy and identified using spectral libraries.

J.E. Halstead et al. / Environmental Pollution 234 (2018) 552e561

A subset of samples whose spectra could not be matched using the FTIR spectral libraries were subsequently interrogated using PCA over the spectral region 1797 to 703 cm1. This region provided the best separation between the different spectral classes. Details of the PCA parameters are in the Supporting Information. Raman spectra were recorded using a 785 nm excitation line, in triplicate, with a  50 objective, with laser power (ranging from 0.0001% to 10%), exposure-time (10 or 30 s), number of accumulations (1 or 3) and bleaching time if appropriate (120 s) optimised for similar sample types. Details of the Raman spectrometer can be found in Supporting Information. 2.4. Statistical analyses Generalised linear modelling was used to test whether the number of ingested debris (fibres and fragments or natural materials) differed between fish species (three species: yellowfin bream, sea mullet and silverbiddy). This modelling used the negative binomial family due to considerable over dispersion in debris counts. The debris numbers were also standardised to fish and gut content weights, with the expectation that larger fish (and those with more gut contents) would have a higher probability of ingestion of more debris. This analysis was performed in the generalised linear models (GLMs) by including fish weight or gut content weight as a log-linear offset term. Thus, three metrics were tested: the number of debris per fish; the number of debris per gram of fish; and the number of debris per gram of gut contents. Fish with empty stomachs (n ¼ 3) were excluded from the GLM including gut content weight as an offset. There was likely a correlation between sediment ingestion and exposure to debris as a consequence of the high prevalence of debris in the sediment (Watts et al., 2015). Using negative binomial GLM, we tested if the number of items recovered (per gram gut content; as an offset term) was dependent on the proportion of sediment in the gut content. This analysis necessarily grouped the three fish species due to co-linearity between species and the proportion of sediment in the gut. 2.5. Estimating annual microplastic ingestion The number of ingested microplastic particles reported represent only those in the gut at the time of sampling, but gut contents are digested and replaced frequently, meaning that the number of particles in the gut at any one time may represent only a fraction of the number ingested over longer periods. To place the results of this current study within the broader ecological context of Sydney Harbour, the average gut throughput rate of each species was sourced from the literature (Steele, 1970; Creighton and Twining, 2010) and used to estimate yearly consumption of synthetic microplastic by an average individual. To estimate the number of particles ingested over a year, an average microplastic value was calculated from all individuals in each species and multiplied by the respective gut throughput rates. This method assumes that the throughput rate of microplastic particles is the same as that of natural gut contents, and also assumes that exposure to plastic particles remains relatively constant over time. As such, this calculation is indicative only. It is important to note the variations that could occur from other factors including the changing prevalence of microplastics, particularly fibres, throughout the year. As such, these results would be bolstered by a larger sample, across multiple sampling periods over a year to assess the true exposure in the species sampled. Annual ingestion calculations were limited to microplastic debris only, due to the possible contamination of natural fibres (see Debris Contamination section below).

555

2.6. Loss of sampled debris The size and delicate nature of a number of the particles, given various ages and degradation states, made handing the samples difficult. As a result, some debris samples extracted from the digestive tracts of the fish were lost or damaged, through handling with the metal forceps, or lost once removed from under the microscope. A number of samples (n ¼ 57) were lost as a result of the drying process (either the heat, or handling) and a further 21 samples were damaged when they came in contact with the ATR objective during spectral collection. As a result, 31% of total particles extracted from the fish guts could not be analysed, leaving only 69% of samples to be interrogated with ATR-FTIR microspectroscopy. 3. Results 3.1. Quality control Procedural controls were treated in the same way as standard samples and underwent the drying and spectral analysis processes. A total of 35 fibres were found in the 101 control Petri dishes throughout the entire dissection procedure, however drying led to the loss of a number of these fibres. Ten remaining fibres were analysed using ATR-FTIR microspectroscopy. The samples matched either wool or cellulose reference spectra. This suggests that fibre contamination during sample processing may account for a large proportion of the debris categorised as wool and cellulose, and methods for reducing environmental deposition of these debris are worthy of future investigation (such as sample processing inside cabinets). However, there is no evidence that other polymer types were affected by contamination during sample processing. It is likely that only a minority of the wool and cellulose debris observed in the gut material were ingested during feeding; however, without detailed further forensic examination of both all the contaminant fibres found in the adjacent Petri dish, and all of the fibres of this type found in the gut by a variety of spectroscopic and morphological techniques, we cannot distinguish between contaminants and ingested forms. This is beyond the scope of the current project but may be the subject of future research. Herein, we will confine the remaining of the discussion to those natural fibres (35%), not including wool and cellulose, that are proposed to be the result of ingestion and not contamination. The fishing nets and line used for sampling were polyamide (nylon) (see Fig. S2C showing fishing materials grouping with nylon reference spectra) as nylon fibres were found in both yellowfin bream and sea mullet this debris type was not included as it may potentially be due to contamination. This resulted in the removal of six samples from further analysis. In a future study dyes or other materials in the nylon of the nets and fishing lines could be compared with those obtained from the fish guts. 3.2. Occurrence of debris in fish A total of 93 fish were sampled from Sydney Harbour. A total of 249 particles were extracted from 40 of these fish, which accounted for 43% of the total sampled. Debris was found in the gut of 25% yellowfin bream, 64% of sea mullet and 21% silverbiddy. Of the 249 particles, 83% were fibrous and 17% were granular in shape. We did not find any large fragments (or mesoplastics) present in any of the fish. Due to the loss of some particles, only 171 of the samples underwent further analysis. There was a significant difference in the amount of debris ingested between species (Table 1; likelihood-ratio test of ‘Species’ P < 0.001), with yellowfin bream and sea mullet having

556

J.E. Halstead et al. / Environmental Pollution 234 (2018) 552e561

Table 1 Results of the negative binomial GLMs testing the relationship between total numbers of debris ingested and species or proportion of sediment in the gut.a The intercept represents yellowfin bream, which is used a reference level. Factor

Estimate

Std Error

Z

P

Number of debris per fish Intercept Sea mullet Silverbiddy

0.486 1.053 2.054

0.395 0.481 0.697

1.231 2.190 2.948

0.218 0.029 0.003

Number of debris per gram of fish Intercept 3.283 Sea mullet 0.077 Silverbiddy 0.943

0.415 0.504 0.676

7.915 0.153 1.395

0.000 0.878 0.163

Number of debris per gram of gut contents Intercept 0.111 Sea mullet 0.396 Silverbiddy 0.035

0.367 0.442 0.634

0.303 0.895 0.055

0.762 0.371 0.956

Number of debris per gram of gut contents Intercept 0.336 Proportion sediment 0.152

0.197 2.619

1.702 0.058

0.089 0.954

a Three response metrics were tested (number of debris per fish, debris per gram of fish, debris per gram of gut contents), with the fish weight and gut content weight standardisations achieved using a log-linear offset term.

significantly higher amounts of debris in their gut than silverbiddy (Fig. 2A). However, no difference was found in the amount of debris ingested between species after accounting for fish weight (Table 1, Fig. 2B; likelihood-ratio test of ‘Species’ P ¼ 0.368) or gut content weight (Table 1, Fig. 2C; likelihood-ratio test of ‘Species’ P ¼ 0.585). The diets varied between the three species (see Table S1, Supporting Information), with sediment contributing a lower proportion of the diet in yellowfin bream (mean of 0.09 in yellowfin bream, 0.74 in sea mullet and 0.37 in silverbiddy). However, a negative binomial GLM showed no relationship between proportion of sediment in gut contents and amount of debris ingested (Table 1). There were individuals from each species with ingested debris that had no sediment ingested (Table S1).

3.3. Debris identification Investigation of the debris using ATR-FTIR microspectroscopy highlighted discrepancies between spectroscopic and visual identifications. Samples that were visually identified and categorised as similar were subsequently identified as chemically distinct (see Fig. S1, Supporting Information), which demonstrated the need for a non-subjective protocol for debris identification. ATR-FTIR microspectroscopy successfully identified 72%, i.e., 123

of the 171 debris samples (119 fibres and 4 fragments) by initial comparison of sample spectra with spectral libraries and further confirmation by direct comparison of spectra by experienced spectroscopists. Of this total, 63 of the fibres and three of the fragments were synthetic. This number excludes fibres made of nylon polyamide, (n ¼ 6), as, without forensic matching of the source fishing material and the fibre, we were unable to determine whether it was ingested during feeding, or was a contaminant from the material used to catch the fish. This aspect could be investigated in future studies. Synthetic microplastics made up 55% of identified debris in sea mullet and 36% in yellowfin bream (Table S2). Three of the five debris samples found in silverbiddy were lost during sampling, which limited further analysis. Acrylic polyester blend represented the highest proportion (17%) in sea mullet, followed by polyester (PET) (9%) and rayon and polypropylene (6% and 5%, respectively) (Fig. 3). Yellowfin bream exhibited a similar ingestion pattern, with the acrylic polyester blend accounting for 16% of debris, and 6% each of both polyester (PET) and rayon (Fig. 3). This study classified rayon as a synthetic fibre, but rayon, chemically known as cellulose xanthate or viscose, is an artificial textile fibre that has properties (colour, shape, buoyancy) that can lead to it being identified as synthetic or semi-synthetic (Remy et al., 2015; Peters and Bratton, 2016). While it is comprised of a natural material, it undergoes a chemical conversion to form cellulose xanthate product that is highly absorbent and easily dyed (Remy et al., 2015). For these reasons, it is likely to pose a similar threat to fish as other synthetic fibres, through the transfer of toxic chemicals and dyes during ingestion. Not all fibres were synthetic, with ~43% from natural origin. Natural fibres that were identified through the spectral libraries included cotton, linen, silk, manila, kenaf, sisal rope. Principal component analysis (PCA) was used to nonsubjectively classify the ATR-FTIR spectra collected from all debris samples, as well as a number of reference samples (wool, rayon, polypropylene, cellulose, nylon (polyamide), bone and polyester), in order to assist in the chemical speciation of the remaining unidentified samples (n ¼ 47). Fig. S2 presents the PCA scores and loadings plots obtained from the second-derivatives of the ATRFTIR spectra. Spectra from six unidentified samples were observed to cluster with the reference spectra collected from bone. The loadings plot reveal that spectra cluster based on the common variable at 1033 cm1 attributable to a phosphate stretching mode (n(PO3 4 )) of hydroxyapatite (Morris and Finney, 2004), the mineral component of bone, likely to have come from animal sources via sewage or wastewater debris. Another six samples were identified as cellulose; the variables at 1024 and 1054 cm1 are both attributed to the stretching modes of cellulose C-OH groups (Cho, 2007).

Fig. 2. Numbers of debris per fish (A), per gram of fish (B) and per gram of gut contents (C) in each of the three fish species from Middle Harbour. These data are those tested in the GLMs (Table 1), and there was a significant difference in debris ingestion among species only for debris per fish (A; P < 0.001).

J.E. Halstead et al. / Environmental Pollution 234 (2018) 552e561

557

Fig. 3. Percentage breakdown of debris identification from total interrogated debris in Sea mullet and Yellowfin bream. Silverbiddy not included to lack of data. (n ¼ 171). *Other natural includes cotton, linen, silk, manila, kenaf, sisal rope. **Other synthetic includes polyethylene, polyurethane, modacrylic, IBC filler, lycra spadex, copper phthalocyanine, red ochre pigment.

Three samples grouped with the polypropylene reference material; the clustering of this group is attributed to the variable at 1376 cm1 assigned to the symmetrical bending mode of the CH3 groups of polypropylene (Silverstein et al., 2014). (Details on these clusters can be found in Supplementary Fig. 2). To summarise, ATRFTIR spectra of 15 samples previously unidentified were subsequently identified using PCA, and are included in Table 2. The combined approach of using ATR-FTIR spectroscopy, and spectral libraries, with PCA resulted in an additional 9% sample identification and a total of 82% of the debris identified. To provide complementary information to improve the chemical speciation of debris, Raman microspectroscopy was used to investigate a subsample of the debris set (n ¼ 24) that had not been identified using either the FTIR spectral libraries, or the combined use of ATR-FTIR spectroscopy and PCA analysis. Good quality spectra were obtained from only 10 of the samples, which provided enough information to confirm the general nature of the debris (i.e., natural, synthetic or pigment/dye). Three samples exhibited polyacrylonitrile components, four had spectral features that were attributable to organic material (protein, D and G carbon bands, cellulose) and three spectra were assigned to the presence of pigments/dyes (red ochre, copper pigment, PY83 yellow pigment). Only two samples were identified conclusively and that included a dyed wool fibre and polyacrylonitrile (PAN), resulting in an additional 1% identified. It is proposed that the FTIR spectral library was

unable to identify the wool sample due to the additional bands of the dye, which were not in the database. The Raman spectra collected from a blue sample (Fig. 4 IIA) were very distinctive with a large number of narrow bands very typical of a dye or pigment containing aromatic rings. Comparison of both the ATR-FTIR and Raman spectra of this sample, shown in Fig. 4, with the literature identified the blue debris containing the common blue pigment copper phthalocyanine (the spectrum of copper phthalocyanine, Fig. 4 IIB), is from an online database (Caggiani et al., 2016).

3.4. Annual ingestion of microplastic The microplastics detected in fish gut are representative of those at the time of sampling only. According to our estimates, an average sized sea mullet, which has a gut throughput rate of between 2 and 6 h (Steele, 1970), may ingest up to 11,000 microplastic particles per year (Table 3). Yellowfin bream and silverbiddy have a much slower throughput rate (between 12 and 24 h) (Creighton and Twining, 2010; He and Wurstbaugh, 1993). Thus, we estimate that yearly exposure to microplastics is up to 600 particles for an averaged sized yellowfin bream (Table 3). We have not provided an estimation for silverbiddy, as insufficient samples were available to be tested with vibrational spectroscopy.

Table 2 Summary of each fish species’ debris and microplastic ingestion. Microplastic identification included samples identified using FTIR spectral libraries (123/171; 72%), PCA analysis (15/171; 9%) and Raman (2/171; 1%).

Number of fish with ingested debris (%) Mean number of debris per fisha (±SE) Max. number of debris particles per fish Total debris Total identified debris Percentage of debris as microplasticb Mean number of microplastics per fishc a

Yellowfin bream (n ¼ 24)

Sea mullet (n ¼ 45)

Silverbiddy (n ¼ 24)

6 (25%) 1.6 (±0.8) 14 39 28 36 0.6

29 (64%) 4.6 (±1.2) 43 205 110 55 2.5

5 (21%) 0.2 (±0.1) 1 5 1 45* 0.1

83% of total debris were fibres. Excludes nylon (3.6% of identified debris) due to possible contamination from fishing gear. Calculated using the percentage of identified debris that were microplastic; no debris was identified for silverbiddy, so the mean percentage from the other two species was used (*). b c

558

J.E. Halstead et al. / Environmental Pollution 234 (2018) 552e561

Fig. 4. Raman (IIA) and ATR-FTIR (I) spectra collected from the same blue fibre. ATRFTIR microspectroscopy identified the presence of polypropylene (bands marked with an asterisk) but the other components could not be identified. Raman microspectroscopy was then used to identify the presence the dye phthalocynanine blue, which was confirmed by comparison of spectra from the literature an example of which is provided in IIB.

4. Discussion This research highlights the importance of using vibrational spectroscopy for determination of the chemical composition of debris ingested by marine organisms such as fish. It has been reported that up to 70% of particles that visually resemble microparticles are not plastic (Hidalgo-Ruz et al., 2012). The microscopic size and degree of degradation and weathering of polymers surfaces makes identification via visual categorisation (such as colour, size and shape) unreliable. Visually natural fibres can often appear to be synthetic having a consistent thickness, no tapering and no cellular structure (Noren, 2012) and could be misidentified as microplastic using visual identification alone. In this study, ATR-FTIR spectroscopy was the primary means of microplastic identification and characterisation and any unidentified samples were subsequently analysed with Raman spectroscopy. The use of an ATR microscope objective enabled small

samples, or sample areas, to be analysed. However, the contact of the sample with the ATR crystal can result in sample destruction. Identification using the FTIR spectral libraries was a powerful tool for debris identification, with 72% of tested debris identified. A recent review by Cannon et al. (2016) recommends at a minimum FTIR accompany visual analysis of microplastic. A novel aspect of this research has been the application of principal component analysis (PCA), a multivariate statistical technique, to nonsubjectively interrogate the ATR-FTIR spectra to establish if similarities existed between reference spectra (wool, rayon, polypropylene, cellulose, nylon (polyamide), bone and polyester) and spectra of the debris samples. Only a small number (n ¼ 6) of reference samples were available for PCA, which limited sample identification, however a further 9% of samples were identified. We also found PCA analysis to be useful in differentiating man-made and natural fibres. The role of PCA in providing this additional level of particle differentiation is only recently being recognised (Comnea-Stancu et al., 2017). In future studies using a larger number of reference spectra, PCA analysis has the potential to provide an accurate and unbiased platform for identification of debris for a vast variety of possible synthetic polymers, as well as non-synthetic materials. Raman spectroscopy provided important information about the natural or synthetic origin of the fibres and the pigments and dyes that were present, and allow for the identification of two further samples. Despite using an excitation line of 785 nm to reduce the likelihood of inducing fluorescence a number of samples did fluoresce which obscured any Raman spectral features. This is commonly observed in both polymer (Siesler, 2011) and biological samples (Lieber and Mahadeyan-Jansen, 2003) and may be due to fluorescence in the sample, the presence of biological material or a biofilm on the sample surface, additives, dyes, the polymer or a combination of any or all of these possibilities. More vigorous sample cleaning may help to reduce this potential contamination and increase the number of good quality spectra collected. Debris degradation could influence the spectral quality as has been suggested in previous work using Raman spectroscopy (Murray and Cowie, 2011). Alternatively, future studies could involve the use of UV excitation, or near IR excitation at wavelengths longer than 785 nm, such as 1064 nm used in FT-Raman spectroscopy. Fibres made up the majority (83%) of the debris observed in the fish in this study. Several studies showed higher frequencies of fibres compared with any other forms of microplastic in a variety of marine environments (Thompson et al., 2004; Moore et al., 2001; Zhao et al., 2014; Lusher et al., 2015; Woodall et al., 2014; Hall et al., 2015; Wright et al., 2013). Fibres also have high prevalence in studies of ingestion in fish (Lusher et al., 2013; Naidoo et al., 2016). For example, 80% of the microplastics found in fish from a USA fishery were fibres (Rochman et al., 2015). In this study, the most prevalent materials were acrylic, polyester, and rayon. The prevalence of acrylic and polyester fibres in our study suggests a dominant source of microplastic in this estuary is clothing fibres. Acrylic and polyester are common textiles used in the clothing industry, and a single garment is known to produce in excess of 1900 fibres per wash (Browne et al., 2011). Rayon is also commonly used

Table 3 Estimated ingestion (number of microplastic particles per year) for the three fish species in this study.a Silverbiddy was not included here due to insufficient samples available to be tested with vibrational spectroscopy.

Yellowfin bream Sea mullet a

Mean microplastic concentration (no. per gut)

Throughput rate (hr)

Particles ingested (per year)

0.59 2.50

12e24 2e6

3-6  102 3.7-11  103

Based on the average microplastic concentration in the guts, and range of gut throughput rates for each species.

J.E. Halstead et al. / Environmental Pollution 234 (2018) 552e561

in clothing, female hygiene products and nappies (Lusher et al., 2013; Remy et al., 2015). There are no sewage outfalls into Sydney Harbour (there is a deep ocean outfall approximately 4 km offshore from the entrance to Sydney Harbour) but there are numerous stormwater drains, and these transport runoff directly into the harbour and are a likely dominant vector for fibres (natural and plastics) and other microplastics in this estuary. In an earlier study within the Sydney region, debris found in water from a number of estuarine habitats matched the types found in a storm water drain, and included similar polymers found in this study, including acrylic and polyester (Browne, 2010). This research supports the current understanding that a significant proportion of debris in estuaries is sourced from nearby land-based wastes. A number of non-synthetic ingested fibres were found in this study that may also present a potential threat to marine organisms when ingested. The presence of natural fibres in the guts of the fish analysed is not a natural phenomenon. It is expected that these fibres entered the estuarine environment via similar sources to the plastic fibres also detected, for example through storm water drains. While it is understood that pollutants and chemicals additives that sorb onto the surface of microplastics can transfer from the gut into the tissue, circulatory system and organs (Browne et al. 2008, 2013; Brennecke et al., 2015) of marine organisms, what is more contested is the notion that chemical load on microplastic or natural fibres is more important than the medium carrying it (Bakir et al., 2014; Lee et al., 2015). There is research to suggest that natural clothing fibres are capable of releasing compounds through washing (Harner et al., 2004; Napper and Thompson, 2016). This research highlights the importance of successfully distinguishing between natural and synthetic fibres, as well as determining the potential toxicity of natural fibres as soon as possible based on the research that already exists suggesting natural fibres’ ability to release compounds (Harner et al., 2004). The results of this present study are in line with published estimates, which have shown that 10e70% of fishes from urban coastal areas can ingest microfibres (McGoran et al., 2017; Naidoo et al., 2016; Phillips and Bonner, 2015). In Sydney Harbour, stormwater and urban runoff continue to be released directly into estuaries, with occasional sewer overflows during wet weather events (Birch and McCready, 2009). This could be contributing to the levels of microplastic and natural fibre ingestion found in this study, and aligns with previous studies that have reported increased exposure to microplastics with proximity to urban areas (Peters and Bratton, 2016; Phillips and Bonner, 2015). These results suggest that Australia's coastline should be a major focus for further microplastic research into the future to investigate the potential ramifications of this level of microplastic exposure in fishes that are consumed by humans. There was a significant difference between the amount of debris (including microplastics) ingested by the three study fish species at the time of sampling, but this appears to be predominately due to differences in the natural ingestion rates and sizes of the individual fish. This was evident when the fibre counts data were standardised to gut content weight (or the correlated metric ‘fish weight’), which removed any species effect. There is an expectation of large interspecific differences in ingestion rates of microplastic in estuaries, but the three species in this study were selected due to them having a generally benthic foraging strategy (Kailola et al., 1993; Hadwen et al., 2007; Whitfield et al., 2012). The stomach content analysis demonstrated that the three species of benthic feeding fishes investigated has diverse diets (i.e. the proportion of sediment, worms etc.), however despite these differences, the three species had similar numbers of debris in their gut relative to the amount of total material ingested. Some individuals had ingested debris without ingesting any sediment. This could indicate that there are

559

pathways of ingestion that do not require ingestion of soft sediments, possibly including exposure in the water column and secondary ingestion via invertebrate prey items such as worms and crustaceans that ingest microplastic (Browne et al., 2013; Remy et al., 2015); however, it could also indicate that microplastic egestion takes longer than sediment/normal diet. This requires further investigation. Understanding the factors that influence differences in microplastic ingestion between individuals and species is important for understanding the impact of this ingestion. Given the results in this study, it is clear that a key factor is the simple allometry of the consumer and their feeding, such that larger fish, or individuals with fuller guts, will generally have higher exposure to microplastics due to simply ingesting more material. Although it could not be tested in this study, it is likely that rates of microplastic ingestion differ across space and time, due to spatial and temporal variation in contaminant loads in the environment (Birch et al., 2001). It is thus prudent for future studies on microplastic ingestion to include numerous sampling events across varying spatial and temporal ranges to better quantify average rates of microplastic ingestion. It is important to acknowledge that gut content analysis reveals only a snapshot of ingestion, revealing only what a consumer recently ingested (Cortes, 1997). When the rates of ingestion of the fish species in this study were extrapolated over one year, based on the gut throughput rate of the species (and assuming that debris is not retained in the gut and has the same throughput rate as other gut contents), it is possible that an average sized sea mullet could ingest 11,000 synthetic microplastic particles a year. Given their larger size and rapid throughput rate (Steele, 1970), sea mullet are likely to ingest far more debris per fish, including microplastics, than the other species investigated in this study. This research identified and quantified the presence of ingested microplastics in fish from Sydney Harbour. These findings indicate that catchment managers should be implementing strategies designed to inhibit the proliferation of these microplastic contaminants within coastal marine environments, given that it is clear they are entering food webs via ingestion by recreationally and commercially important fish species in estuaries. Further research is required to fully understand the ecological impact of microplastic within water bodies like Sydney Harbour, and could encompass: 1) more fish species across diverse feeding strategies; 2) an investigation of how ingestion differs spatially (especially relative to freshwater inputs, stormwater drains, etc.), and 3) an investigation of how ingestion changes through time (days, weeks, seasons) with the use of vibrational spectroscopy, as demonstrated here. Acknowledgements This study was conducted under University of New South Wales Animal Care and Ethics Approval #16/73B, and fish collection was authorised under a scientific research permit issued by the NSW Department of Primary Industries. Derrick Cruz and Gregory Miltenyi assisted collection of fish. We are grateful to Mark Browne for his advice on this study. JEH is grateful to Dr. Joonsup Lee at the University of Sydney Vibrational Spectroscopy Core Facility for his assistance. This research was supported by Australian Research Council (ARS) Discovery grants to ELJ and EAC (DP150103272) and PAL (DP140100176) and LIEF Grants for the FTIR (LE0883036) and Raman (LE0560680) instruments to PAL and EAC. Appendix A. Supplementary data Supplementary data related to this article can be found at https://doi.org/10.1016/j.envpol.2017.11.085.

560

J.E. Halstead et al. / Environmental Pollution 234 (2018) 552e561

References Acosta-Coley, I., Olivero-Verbel, J., 2015. Microplastic resin pellets on an urban tropical beach in Colombia. Environ. Monit. Assess. 187, 1e14. Adams, M., 2004. Chemometrics in Analytical Spectroscopy, third ed. John Wiley and Sons, Cambridge. Arthur, C.A., Baker, J., Bamford, H., 2009. In: Proceedings of the International Research Workshop on the Occurrence, Effects, and Fate of Microplastic Marine Debris. NOAA Technical Memorandum. Bakir, A., Rowland, S.J., Thompson, R.C., 2014. Transport of persistent organic pollutants by microplastics in estuarine conditions. Estuar. Coast Shelf S 40, 14e21. Barnes, D.K., Galgani, F., Thompson, R.C., Barlaz, M., 2009. Accumulation and fragmentation of plastic debris in global environments. Philos. T R. Soc. B 364, 1985e1998. Bergmann, M., Klages, M., Gutoq, L., 2000. Marine Anthropogenic Litter. Springer, Cham Heidelberg New York Dordrecht London. Birch, G.F., 2000. Marine pollution in Australia, with special emphasis on central New South Wales estuaries and adjacent continental margin. Int. J. Environ Pollut 13, 573. Birch, G., Taylor, S.E., Matthai, C., 2001. Small-scale spatial and temporal variance in the concentration of heavy metals in aquatic sediments: a review and some new concepts. Environ. Pollut. 113, 357e372. Birch, G.F., Taylor, S.E., 2002. Application of sediment quality guidelines in the assessment and management of contaminated surficial sediments in Port Jackson (Sydney harbour), Australia. Environ. Manage 29, 860e870. Birch, G.F., McCready, S., 2009. Catchment condition as a major control on the quality of recieving basin sediments (Sydney harbour, Australia). Sci. Total Environ. 407, 2820e2835. Brereton, R., 2003. Chemometrics: Data Analysis for the Laboratory and Chemical Plant. John Wiley and Sons, West Sussex. Browne, M.A., Dissanayake, A., Galloway, T.S., Lowe, D.M., Thompson, R.C., 2008. Ingested microscopic plastic translocates to the circulatory system of the mussel. Mytilus Edulis. Environ. Sci. Technol. 42, 5026e5031. Browne, M.A., Galloway, T.S., Thompson, R.C., 2010. Spatial patterns of plastic debris along estuarine shorelines. Environ. Sci. Technol. 44, 3404e3409. Browne, M.A., 2010. Characterisation of sources and sinks of microplastic debris in the Hornsby Shire Catchment. Final Report. Browne, M.A., Crump, P., Niven, S.J., Teuten, E., Tonkin, A., Galloway, T., Thompson, R.C., 2011. Accumulation of microplastic on shorelines worldwide: sources and sinks. Environ. Sci. Technol. 45, 9175e9179. Browne, M.A., Niven, S.J., Galloway, T.S., Rowland, S.J., Thompson, R.C., 2013. Microplastic moves pollutants and additives to worms, reducing functions linked to health and biodiversity. Curr. Biol. 23, 2388e2392. Browne, M.A., Underwood, A.J., Chapman, M.G., Williams, R., Thompson, R.C., van Franeker, J.A., 2015. Linking effects of anthropogenic debris to ecological impacts. Process Biol. Sci. 282. Brennecke, D., Ferreira, E.C., Costa, T.M.M., Appel, D., da Gama, B.A.P., Lenz, M., 2015. Ingested microplastics are translocated to organs of the tropical fiddler crab Uca rapax. Mar. Poll. Bull. 96, 491e495. Caggiani, M., Cosentino, A., Mangone, A., 2016. Pigments checker version 3.0, a handy set for conservation scientists: a free online Raman spectra database. Microchem J. 129, 123e132. Cannon, S.M., Lavers, J.L., Figueiredo, B., 2016. Plastic ingestion by fish in the Southern Hemisphere: a baseline study and review of methods. Mar. Poll. Bull. 107, 286e291. Carr, S.A., Liu, J., Tesoro, A.G., 2016. Transport and fate of microplastic particles in wastewater treatment plants. Water Res. 91, 174e182. Castillo, A.B., Al-Maslamani, I., Obbard, J.P., 2016. Prevalence of microplastics in the marine waters of Qatar. Mar. Poll. Bull. 111, 260e267. Comnea-Stancu, I.R., Wieland, K., Ramer, G., Schwaighofer, A., Lendl, B., 2017. On the identification of rayon/viscose as a major fraction of microplastics in the marine environment: discrimination between natural and manmade cellulosic fibers using Fourier transform infrared spectroscopy. Appl. Spec. 71, 939e950. Cortes, E., 1997. A critical review of methods of studying fish feeding based on analysis of stomach contents: application to elasmobranch fishes. Can. J. Fish. Aquat. Sci. 54, 726e738. Cho, L., 2007. Identification of textile fiber by Raman microspectroscopy. Forensic Sci. 6, 55e62. Choy, C.A., Drazen, J.C., 2013. Plastic for dinner? Observations of frequent debris ingestion by pelagic predatory fishes from the central North Pacific. Mar. Ecol. Prog. Ser. 485, 155e163. Chua, E.M., Shimeta, J., Nugegoda, D., Morrison, P.D., Clarke, B.O., 2014. Assimilation of polybrominated diphenyl ethers from microplastics by the marine amphipod, allorchestes compressa. Environ. Sci. Technol. 48, 8127e8134. Collard, F., Gilbert, B., Eppe, G., Parmentier, E., Das, K., 2015. Detection of anthropogenic particles in fish stomachs: an isolation method adapted to identification by Raman spectroscopy. Arch. Environ. Con Tox 69, 331e339. Claessens, M., De Meester, S., Van Landuyt, L., De Clerck, K., Janssen, C.R., 2011. Occurrence and distribution of microplastics in marine sediments along the Belgian coast. Mar. Poll. Bull. 62, 2199e2204. Creighton, N., Twining, J., 2010. Bioaccumulation from food and water of cadmium, selenium and zinc in an estuarine fish, Ambassis jacksoniensis. Mar. Poll. Bull. 60, 1815e1821. Davison, P., Asch, R.G., 2011. Plastic ingestion by mesopelagic fishes in the North

Pacific subtropical gyre. Mar. Ecol. Prog. Ser. 432, 173e180. De Witte, B., Devriese, L., Bekaert, K., Hoffman, S., Vandermeersch, G., Cooreman, K., Robbens, J., 2014. Quality assessment of the blue mussel (Mytilus edulis): Comparison between commercial and wild types. Mar. Poll. Bull. 85, 146e155. Desforges, J.P.W., Galbraith, M., Dangerfield, N., Ross, P.S., 2014. Widespread distribution of microplastics in subsurface seawater in the NE Pacific Ocean. Mar. Poll. Bull. 79, 94e99. Dris, R., Gasperi, J., Rocher, V., Saad, M., Renault, N., Tassin, B., 2015. Microplastic contamination in an urban area: a case study in greater Paris. Environ. Chem. 12, 592e599. Enders, K., Lenz, R., Stedmon, C.A., Nielsen, T.G., 2015. Abundance, size and polymer composition of marine microplastics 10mm in the Atlantic Ocean and their modelled vertical distribution. Mar. Poll. Bull. 100, 70e81. Farrell, P., Nelson, K., 2013. Trophic level transfer of microplastic: Mytilus edulis (L.) to Carcinus maenas (L.). Environ. Pollut. 177, 1e3. Foekema, E.M., De Gruijter, C., Mergia, M.T., van Franeker, J.A., Murk, A.J., Koelmans, A.A., 2013. Plastic in north sea fish. Environ. Sci. Technol. 47, 8818e8824. Frias, J.P.G.L., Gago, J., Otero, V., Sobral, P., 2016. Microplastics in coastal sediments from Southern Portuguese shelf waters. Mar. Environ. Res. 114, 24e30. GESAMP, 2016. Sources, fate and effects of microplastics in the marine environment: part two of a global assessment. In: Kershaw, P.J., Rochman, C.M. (Eds.), Joint Group of Experts on the Scientific Aspects of Marine Environmental Protection, vol 93, p. 220. Hadwen, W.L., Russell, G.L., Arthington, A.H., 2007. Gut content and stable isotopederived diets of four commercially and recreationally important fish species in two intermittently open estuaries. Mar. Fresh Res. 58, 363. Hall, N.M., Berry, K.L.E., Rintoul, L., Hoogenboom, M.O., 2015. Microplastic ingestion by scleractinian corals. Mar. Biol. 162, 725e732. Harner, T., Shoeib, M., Diamond, M., Stern, G., Rosenberg, B., 2004. Using passive air samplers to assess urban-rural trends for persistent organic pollutants: polychlorinated biphenyls and organochlorine pesticides. Environ. Sci. Technol. 38, 4474e4483. He, E., Wurstbaugh, W., 1993. An empirical model of gastric evacuation rates for fish and an analysis of digestion in piscivorous brown trout. T Am. Fish. Soc. 122, 717e773. Helm, P.A., 2017. Improving microplastic source apportionment: a role for microplastic morphology and taxonomy? Anal. Methods 9, 1328e1331. Hedge, L., Johnston, E., Ayoung, S., Birch, G., Booth, D., Creese, R., Doblin, M., Figueira, W., Gribben, P., Hutchings, P., 2013. Sydney Harbour: a Systematic Review of the Science (Sydney). Hidalgo-Ruz, V., Gutow, L., Thompson, R.C., Thiel, M., 2012. Microplastics in the marine environment: a review of the methods used for identification and quantification. Environ. Sci. Technol. 46. Hyslop, E., 1980. Stomach contents analysis e a review of methods and their application. J. Fish. Bio 17, 411e429. Imhof, H.K., Laforsch, C., Wiesheu, A.C., Schmid, J., Anger, P.M., Niessner, R., Ivleva, N.P., 2016. Pigments and plastic in limnetic ecosystems: a qualitative and quantitative study on microparticles of different size classes. Water Resour. 98, 64e74. Ivar do Sul, J.A., Costa, M.F., 2014. The present and future of microplastic pollution in the marine environment. Environ. Pollut. 185, 352e364. Jambeck, J.R., Geyer, R., Wilcox, C., Siegler, T.R., Perryman, M., Andrady, A., Narayan, R., Law, K.L., 2015. Plastic waste inputs from land into the ocean. Science 347, 768e771. Johnston, E.L., Mayer-Pinto, M., Hutchings, P.A., Marzinelli, E.M., Ahyong, S.T., Birch, G., Gribben, P.E., 2015. Sydney Harbour: what we do and do not know about a highly diverse estuary. Mar. Fresh Res. 66, 1073e1087. Kailola, P., Willams, M., Stewart, P., Reichelt, R., McNee, A., Grieve, C., 1993. Australian Fisheries Resources. Bureau of Resource Sciences, Department of Industries and Energy and the Fisheries Research and Development Corporation, Canberra. €ppler, A., Fischer, D., Oberbeckmann, S., Schernewski, G., Labrenz, M., Ka Eichhorn, K.J., Voit, B., 2016. Analysis of environmental microplastics by vibrational microspectroscopy: FTIR, Raman or both? Anal. Bioanal. Chem. 408, 8377e8391. Kazarian, S.G., Chan, K.L.A., 2010. Micro- and macro-attenuated total reflection fourier transform infrared spectroscopic imaging. Appl. Spec. 64, 135e152. Koelmans, A.A., Bakir, A., Burton, G.A., Janssen, C.R., 2016. Microplastic as a vector for chemicals in the aquatic environment: critical review and model-supported reinterpretation of empirical studies. Environ. Sci. Technol. 50, 3315e3326. Lenz, R., Enders, K., Stedmon, C.A., Mackenzie, D.M.A., Gissel, T.A., 2015. Critical assessment of visual identification of marine microplastic using Raman spectroscopy for analysis improvement. Mar. Poll. Bull. 100, 82e91. Lieber, C., Mahadeyan-Jansen, A., 2003. Automated method for subtraction of fluorescence from biological Raman spectra. Appl. Spec. 57, 1363e1367. Lee, J., Hong, S., Jang, Y.C., Lee, M.J., Kang, D., Shim, W.J., 2015. Finding solutions for the styrofoam buoy debris problem through participatory workshops. Mar. Policy 51, 182e189. € nnstedt, O.M., Eklo €v, P., 2016. Environmentally relevant concentrations of Lo microplastic particles influence larval fish ecology. Science 352, 1213e1216. Lusher, A.L., McHugh, M., Thompson, R.C., 2013. Occurrence of microplastics in the gastrointestinal tract of pelagic and demersal fish from the English Channel. Mar. Poll. Bull. 67, 94e99. Lusher, A.L., Burke, A., O'Connor, I., Officer, R., 2014. Microplastic pollution in the

J.E. Halstead et al. / Environmental Pollution 234 (2018) 552e561 Northeast Atlantic ocean: validated and opportunistic sampling. Mar. Poll. Bull. 88, 325e333. Lusher, A.L., Tirelli, V., O'Connor, I., Officer, R., 2015. Microplastics in arctic polar waters: the first reported values of particles in surface and sub-surface samples. Sci. Rep. 5. Mathalon, A., Hill, P., 2014. Microplastic fibers in the intertidal ecosystem surrounding halifax harbor, nova scotia. Mar. Poll. Bull. 81, 69e79. Mayer-Pinto, M., 2015. Sydney Harbour: a review of anthropogenic impacts on the biodiversity and ecosystem function of one of the world's largest natural harbours. Mar. Fresh Res. 1088e1105. McGoran, A.R., Clark, P.F., Morritt, D., 2017. Presence of microplastic in the digestive tracts of European flounder, Platichthys flesus, and European smelt, Osmerus eperlanus, from the River Thames. Environ. Pollut. 220, 744e751. Mecozzi, M., Pietroletti, M., Monakhova, Y.B., 2016. FTIR spectroscopy supported by statistical techniques for the structural characterization of plastic debris in the marine environment: application to monitoring studies. Mar. Poll. Bull. 106, 155e161. € der, M.G.J., Primpke, S., Gerdts, G., 2017. Identification Mintenig, S.M., Int-Veen, I., Lo of microplastic in effluents of waste water treatment plants using focal plane array-based micro-Fourier-transform infrared imaging. Water Res. 108, 365e372. Mizraji, R., Ahrendt, C., Perez-Venegas, D., Vargas, J., Pulgar, J., Aldana, M., Ojeda, F.P., n-Malago n, C., 2017. Is the feeding type related with the conDuarte, C., Galba tent of microplastics in intertidal fish gut? Mar. Poll. Bull. 116, 498e500. Moore, C.J., Moore, S.L., Leecaster, M.K., Weisberg, S.B., 2001. A comparison of plastic and plankton in the north Pacific central gyre. Mar. Pollut. Bull. 42, 1297e1300. Morris, M.D., Finney, W.F., 2004. Recent developments in Raman and infrared spectroscopy and imaging of bone tissue. Spectroscopy 18, 155e159. Murray, F., Cowie, P.R., 2011. Plastic contamination in the decapod crustacean Nephrops norvegicus (Linnaeus, 1758). Mar. Poll. Bull. 62, 1207e1217. Napper, I.E., Thompson, R.C., 2016. Release of synthetic microplastic plastic fibres from domestic washing machines: effects of fabric type and washing conditions. Mar. Poll. Bull. 112, 39e45. Naidoo, T., Smith, A.J., Glassom, D., 2016. Plastic ingestion by estuarine mullet Mugil cephalus ( Mugilidae ) in an urban harbour, KwaZulu-Natal, South Africa. Afr. J. Mar. Sci. 38, 145e149. Nor, N.H.M., Obbard, J.P., 2014. Microplastics in Singapore's coastal mangrove ecosystems. Mar. Poll. Bull. 79, 278. Noren, F., 2012. Small Plastic Particles in Coastal Swedish Waters (Lysekil). Obbard, R.W., Sadri, S., Wong, Y.Q., Khitun, A.A., Baker, I., Thompson, R.C., 2014. Global warming releases microplastic legacy frozen in Arctic Sea ice. Earth's Future 2, 315e320. Peters, C.A., Bratton, S.P., 2016. Urbanization is a major influence on microplastic ingestion by sun fish in the Brazos River Basin, Central Texas, USA. Environ. Pollut. 210, 380e387. Phillips, M.B., Bonner, T.H., 2015. Occurrence and amount of microplastic ingested by fishes in watersheds of the Gulf of Mexico. Mar. Poll. Bull. 100, 6e11. Platell, M., Sarre, G., Potter, I., 1999. The diets of two co-occurring marine teleosts, Parequula melbournensis and Pseudocaranx wrighti, and their relationships to body size and mouth morphology, and the season and location of capture. Environ. Biol. Fish. 49, 361e376. Rani, M., Shim, W., Han, G., Jang, M., Al-Odaini, N., Song, Y., Hong, S., 2015. Qualitative analysis of additives in plastic marine debris and its new products. Arch. Environ. Contam. Toxicol. 69, 352e366. re, P., Eppe, G., Lepoint, G., 2015. When Remy, F., Collard, F., Gilbert, B., Compe

561

microplastic is not plastic: the ingestion of artificial cellulose fibers by macrofauna living in seagrass Macrophytodetritus. Environ. Sci. Technol. 49, 11158e11166. Rochman, C.M., Hoh, E., Kurobe, T., Teh, S.J., 2013. Ingested plastic transfers hazardous chemicals to fish and induces hepatic stress. Sci. Rep. 3, 3263. Rochman, C.M., Tahir, A., Williams, S.L., Baxa, D.V., Lam, R., Miller, J.T., Teh, F., Werorilangi, S., Teh, S.J., 2015. Anthropogenic debris in seafood: plastic debris and fibers from textiles in fish and bivalves sold for human consumption. Sci. Rep. 5, 1e10. , C., Consoli, P., Andaloro, F., Fossi, M.C., 2015. First eviRomeo, T., Pietro, B., Peda dence of presence of plastic debris in stomach of large pelagic fish in the Mediterranean Sea. Mar. Poll. Bull. 95, 358e361. € der, M.G.J., Fricke, N.F., Lang, T., Griebeler, E., Janke, M., Gerdts, G., Rummel, C.D., Lo 2016. Plastic ingestion by pelagic and demersal fish from the North sea and Baltic sea. Mar. Poll. Bull. 102, 134e141. Siesler, H.W., 2011. Vibrational spectroscopy of polymers. Int. J. Polym. Anal. Charact. 16, 519e541. Silverstein, R., Webster, F., Kiemle, D., Bryce, D., 2014. Spectrometric Identification of Organic Compounds, eighth ed. John Wiley and Sons, Ottawa. Song, Y.K., Hong, S.H., Jang, M., Kang, J.H., Kwon, O.Y., Han, G.M., Shim, W.J., 2014. Large accumulation of micro-sized synthetic polymer particles in the sea surface microlayer. Environ. Sci. Technol. 48, 9014. Steele, J.H., 1970. Marine Food Chains. University of California Press, Berkeley, pp. 222e240. Stewart, J., Hegarty, A., Young, C., Fowler, A.M., Craig, J., 2015. Status of Fisheries Resources in NSW 2013e14. NSW Department of Primary Industries, Mosman, p. 391. Thompson, R.C., Olsen, Y., Mitchell, R.P., Davis, A., Rowland, S.J., John, A.W.G., McGonigle, D., Russell, A.E., 2004. Lost at sea: where is all the plastic? Science 304, 838. Truong, L., Suthers, I.M., Cruz, D.O., Smith, J.A., 2017. Plankton supports the majority of fish biomass on temperate rocky reefs. Mar. Biol. 164. Van Cauwenberghe, L., Vanreusel, A., Mees, J., Janssen, C.R., 2013. Microplastic pollution in deep-sea sediments. Environ. Pollut. 182, 495e499. Van Cauwenberghe, L., Janssen, C.R., 2014. Microplastics in bivalves cultured for human consumption. Environ. Pollut. 193, 65e70. Van Cauwenberghe, L., Claessens, M., Vandegehuchte, M.B., Janssen, C.R., 2015. Microplastics are taken up by mussels (Mytilus edulis) and lugworms (Arenicola marina) living in natural habitats. Environ. Pollut. 199, 10e17. Watts, A.J.R., Urbina, M.A., Corr, S., Lewis, C., Galloway, T.S., 2015. Ingestion of plastic microfibers by the crab Carcinus maenas and its effect on food consumption and energy balance. Environ. Sci. Technol. 49, 14597e14604. Whitfield, A.K., Panfili, J., Durand, J.D., 2012. A global review of the cosmopolitan flathead mullet Mugil cephalus (Teleostei: Mugilidae), with emphasis on the biology, genetics, ecology and fisheries aspects of this apparent species complex. Rev. Fish Biol. Fish. 22, 641e681. Woodall, L.C., Sanchez-Vidal, A., Canals, M., Paterson, G.L., Coppock, R., Sleight, V., Calafat, A., Rogers, A.D., Narayanaswamy, B.E., Thompson, R.C., 2014. The deep sea is a major sink for microplastic debris. R. Soc. Open Sci. 1, 140317. Wright, Stephanie L., Thompson, R.C., Galloway, T.S., 2013. The physical impacts of microplastics on marine organisms: a review. Environ. Pollut. 178, 483e549. Zhao, S., Zhu, L., Wang, T., Li, D., 2014. Suspended microplastics in the surface water of the Yangtze Estuary System, China: first observations on occurrence, distribution. Mar. Poll. Bull. 86, 562e568.