Anthropogenic pollutants derived from agriculture, industry, and urban sources are severely affecting aquatic ecosystems worldwide and may have adverse impacts on aquatic organisms (Häder et al., 2020). Anthropogenic debris, such as microplastics (1 μm to 5 mm), has been widely detected in aquatic systems throughout the world, thus causing growing public concern (reviewed in Xu et al., 2020). Due to their ubiquity, microplastics are extremely likely to interact with aquatic biota (Scherer et al., 2018) and consequently have been found in a range of marine (reviewed in Lusher et al., 2017, Rezania et al., 2018) and freshwater organisms (reviewed in O'Connor et al., 2019). Primary pathways of ingestion are suspected to be direct, either passively through filter feeding (e.g., Besseling et al., 2015; Su et al., 2018; Van Cauwenberghe et al., 2015), or through confusion with prey (e.g., ingestion by planktivorous fish) (Ory et al., 2017; Roch et al., 2020), and though little is known about the extent to which trophic transfer occurs in the wild, dietary transfer of microplastics has been demonstrated in experimental (Farrell & Nelson, 2013; Setälä et al., 2014; Van Colen et al., 2020) and semi-natural environments (Nelms et al., 2018).
In addition to the possible physical, physiological, and reproductive effects incurred from direct ingestion such as reduced growth, increased mortality, and reduced embryonic development (Horn et al., 2020; Redondo-Hasselerharm et al., 2018; Ziajahromi et al., 2018), exposure to chemical additives contained in microplastics (e.g., phthalate plasticizers) can also induce developmental defects among lower level aquatic organisms (e.g., invertebrates and small fish) (Capolupo et al., 2020; Mu et al., 2018). In order to inform relevant ecological risk assessments of freshwater microplastics as well as potential monitoring strategies, microplastic interactions with top predators need to be assessed.
While microplastic research in freshwater ecosystems is increasing, only a handful of studies have assessed the physical presence of microplastics among top predators, most of which have been fish and bird species (e.g., Northern pike [Esox lucius], white-throated dipper [Cinclus cinclus], common kingfisher [Alcedo atthis] [Campbell et al., 2017; D'Souza et al., 2020; Winkler et al., 2020]). Field studies from the marine environment also indicate that a number of marine top predators ingest microplastics (e.g., Hernandez-Milian et al., 2019; Nelms, Barnett, et al., 2019; Rebolledo et al., 2013), while microplastics were also recovered in one of three Eurasian otters (Lutra lutra) assessed along the Norwegian coast (Haave et al., 2021). Evidence from a semi-natural experiment suggests that mammalian top predators, such as pinnipeds, are likely ingesting microplastics through their prey (i.e., secondary ingestion) (Nelms et al., 2018). While stomachs, intestines, and gastrointestinal tracts (GITs) are generally the main components investigated for microplastics in aquatic biota (reviewed in Lusher et al., 2017, O'Connor et al., 2019), a number of studies have also examined feces as a noninvasive assessment tool for microplastic ingestion (Bessa et al., 2019; D'Souza et al., 2020; Eriksson & Burton, 2003; Le Guen et al., 2020). Moreover, dietary analysis of otters in Italy incidentally revealed the presence of microplastics in fecal samples (Smiroldo et al., 2019).
The Eurasian otter, which is a member of the Mustelidae family, is a top predator of freshwater ecosystems and also occurs in estuarine and coastal waters. As the most widely distributed otter species, its range extends into three continents: Europe, Asia, and Africa (Hung & Law, 2016; Roos et al., 2015). Due to their high trophic status and ecological characteristics, otters are generally considered a sentinel species of organochlorine pollution within freshwater ecosystems (Lemarchand et al., 2011). Population declines experienced throughout northern and western Europe during the latter half of the 20th century are reputed to coincide with elevated levels of persistent organic pollutants, particularly polychlorinated biphenyls (PCBs) (Olsson & Sandegren, 1991; Roos et al., 2001). Ireland, which was not as impacted by such pollutants, is traditionally considered a stronghold for the species, as highlighted in the early 1980s where a high incidence rate was recorded throughout the country (91.7% of survey sites assessed) (Chapman & Chapman, 1982; Reid, Thompson, et al., 2013), which remained relatively high into the 21st century (Reid, Hayden, et al., 2013). Like most mustelids for which scent-marking is an important form of communication (Hutchings & White, 2000), otters deposit feces (“spraints”) as scent marks throughout their home range (Kruuk, 1992), and these can be easily collected over large areas (Mason & Macdonald, 1994). As a result, the “Standard Otter Survey” method developed by Lenton et al. (1980) primarily focuses on the detection of spraints in prominent areas (e.g., boulders and under bridges) and is the main method adopted by most European countries for otter species conservation assessments (Reid et al., 2014). As well as providing information on diet (e.g., Smiroldo et al., 2019), spraints have been used for noninvasive genetic sampling of otter populations (White et al., 2013) and for assessing organochlorine pesticide and PCB concentrations (Lemarchand et al., 2007; Mason & Macdonald, 1994; O'Sullivan et al., 1993; van den Brink & Jansman, 2006). Based on the initial evidence of microplastics in otter spraints (Smiroldo et al., 2019), as well as in the fecal remains of other top predators, it appears spraint samples may provide a noninvasive means of assessing microplastic ingestion for the species.
The research presented here was undertaken as part of a large-scale study assessing the potential pathways of microplastic uptake and transfer within the River Slaney food web located in southeast Ireland. Opportunistic collection of otter spraints presented a basis to investigate the prevalence and abundance of microplastics in otter fecal remains and determine whether microplastics are transferred to this top level predator within the River Slaney catchment. Where possible, and in order to explore the viability of monitoring microplastics in otter spraints and their applicability to different regions, samples were also obtained from a number of additional catchments in the southwest and west of the country (Figure 1). Another objective of this study was to characterize otter diet so as to identify possible trophic links, which would assist in verifying the food web of the River Slaney, as well as the other regions assessed. It was hypothesized that:
- Spatial variations in microplastic concentrations would be observed, specifically higher concentrations in spraints collected from regions with greater microplastic input (i.e., River Slaney) (Mahon et al., 2017). As well as this, spraints collected downstream of potential microplastic sources (i.e., “higher” exposure), which were defined using spatial vector data (ESRI, 2011), would contain higher microplastic concentrations as a result of increased microplastic input (e.g., Kay et al., 2018; McCormick et al., 2014).
- Older spraints categorized as “dry” or “drying” would contain more microplastics than those freshly obtained, due to a greater exposure to particles that may deposit from the atmosphere (Allen et al., 2019; Stanton et al., 2019).
- Microplastic concentrations in spraints would vary by season (i.e., spring, summer, autumn), on account of temporal variations in microplastic inputs (e.g., increases in land runoff) (e.g., Campanale et al., 2020) as well as possible seasonal changes in otter diet (Kyne et al., 1989).
- Microplastics in spraints would be related to dietary composition as suggested by Nelms, Parry, et al. (2019).
Enlarge this image.FIGURE 1. Map showing spraint collection sites (lower exposure: White circles and higher exposure: Black circles) for the western region (a, b), southwestern region (c, d), and the River Slaney catchment (e), along with potential sources of microplastic pollution; urban waste water treatment plants (UWWTPs) (triangles), sites of biosolid application (diamonds), and licensed waste facilities (squares)
Spraint samples were collected from eight catchments located in the southeast, southwest, and west of Ireland. The eight catchments were divided into three distinct study regions that included (1) the River Slaney catchment in the southeast (Figure 1e), (2) the southwestern region comprising three catchments (Figure 1c,d), and (3) the western region, comprising four catchments (Figure 1a,b; Table 1). Of the eight catchments sampled, four comprise Special Areas of Conservation (SACs) designated under the European Union Habitats Directive (EC Directive 92/43/EEC) and include the otter as a qualifying species. These consist of the Slaney River Valley SAC in the southeast, the Blackwater River SAC in the southwestern region, and the Mweelrea/Sheeffry/Erriff and Lough Carra/Lough Mask SACs in the western region. While the overall microplastic exposure in six of the catchments was expected to be low, respective exposure levels of very high and medium were anticipated for the River Slaney and Munster Blackwater catchments, owing to a higher density of urban waste water treatment plants (UWWTPs), UWWTP biosolid application sites, and licensed waste facilities (Mahon et al., 2017).
TABLE 1 Information for microplastics recovered from spraints collected in each region assessed (bold font) as well as the catchments of the southwestern and western regions (normal font), including sample size per exposure level, spraint condition, and season with microplastic prevalence expressed as a percentage (%)
Overall, 105 otter spraint samples were opportunistically collected during the period of March 2017 to July 2019 by citizen scientists (e.g., Irish Wildlife Trust), staff of the Irish National Parks and Wildlife Service (NPWS), colleagues, and members of the project team. Due to limited sampling in certain regions, coupled with a number of samples originating from the marine environment, 53 spraints representing 29 marking sites were deemed viable for analysis. Of these, 51 were from inland sites, while two samples from the western region originated from freshwater streams adjacent to the Atlantic coast (i.e., Achill Island and the Bundorragha river near Killary Fjord) (Figure 1a,b).
In order to determine whether otter spraints could indicate local microplastic exposure levels, spraints were classified as either “higher” (n = 22) or “lower” (n = 31) exposure based on their site of origin. Higher exposure spraints originated from areas downstream of microplastic sources (i.e., UWWTPs, UWWTP biosolid application sites, licensed waste facilities), including the outflow of lakes in receipt of possible microplastic discharges (e.g., upper Lough Corrib, Figure 1a), while lower exposure spraints originated from reference areas, upstream of potential sources where the contribution from local microplastic sources was expected to be lower.
All spraint samples used in this study were collected through the same protocol, which was outlined in a document supplied to all participating personnel prior to the commencement of sampling. Spraint freshness was recorded according to color, moistness, and contents (as per Prigioni et al., 2006), before samples were transferred to 540-ml sterile polyethylene sample bags using wooden disposable tongue depressors that were exchanged between samples. Spraints were stored at −20°C until further processing and analysis. All spraint samples were confirmed as otter spraints by a sole observer prior to their inclusion in microplastic analysis.
Isolation of microplastics was performed in the laboratory following modification of the protocols described in Bessa et al. (2019) and Le Guen et al. (2020) for the extraction of microplastics from the fecal remains of gentoo penguin (Pygoscelis papua) and king penguin (Aptenodytes patagonicus), respectively. Spraints were removed from the freezer and allowed to defrost in their sample bags overnight. Wet weight was obtained, and samples were placed on individual strips of aluminum foil, placed in prerinsed aluminum foil trays that were also covered with aluminum foil and dried in a heating chamber at 40°C for 48 h. Laboratory dried mass (hereinafter referred to as dry weight [dw]) was recorded to the nearest 0.01 g, and each sample transferred to individual acid washed (0.05% nitric acid [HNO3]) glass jars (344 ml) that were sealed with metal screw cap lids. In order to isolate microplastics from otter spraints, a two-stage filtration process was performed. Firstly, each jar received a 3:1 (solution:sample) ratio of 10% potassium hydroxide (KOH) (w/v) and was incubated at 40°C for 48 h in order to digest organic material. Following digestion, the initial supernatant of each sample was vacuum filtered onto glass fiber filter paper (1.2 μm particle retention) using a prerinsed glass pipette (Bessa et al., 2019) and sealed in sterile Petri dishes (55 mm diameter) for microplastic enumeration and characterization. Secondly, a hypersaline solution (NaCl 1.2 g/cm3) was added to the remaining residue (3:1 ratio) for density separation (modified from Le Guen et al., 2020). Samples were stirred for 30 s using a stainless steel spatula and allowed to settle for a minimum of 10 min, and the filtration process repeated to improve microplastic recovery.
Polymer characterization and validationAll suspected microplastic particles found were recorded by morphology type (i.e., fiber, fragment, and film) and color under a stereomicroscope with a polarized attachment (Olympus SZX10, 1.6× magnification), while size (longest axis) and fiber width were manually measured to the nearest 0.1 μm using Image Pro Plus software (QImaging 2000R digital camera). Size classes were assigned to suspected particles following recommendations for reporting microplastics in natura and improving comparability of field studies (Frias & Nash, 2019). However, any particles <100 μm were omitted from analysis due to detection and handling limitations, thus resulting in two size classifications (100 ≤ 350 μm and 350 μm to ≤5 mm).
Particles recovered in spraint samples were manually isolated onto a clean glass fiber filter paper using a stainless steel fine tip forceps and analyzed by micro-Fourier transform infrared (μ-FTIR) spectroscopy. Analysis was conducted using a Bruker Hyperion 2000 microscope (15× objective) (reflectance technique) equipped with a liquid nitrogen cooled single element mercury cadmium telluride detector, which was coupled to a Bruker Tensor 27 spectrometer. Spectra for each particle were collected in absorbance mode using 128 scans (wavenumber range 4000–600 cm−1) at a spectral resolution of 4 cm−1 and recorded using OPUS 7.8 software. Background spectra were collected from blank areas of the filter using the same measurement parameters. To improve comparability with existing and future freshwater microplastic studies, microplastics found in spraints were expressed as both the abundance of microplastics per spraint (MPs/spraint) and concentration of microplastics per gram dw (MPs/g dw).
Quality assurance and quality control proceduresMeasures to account for and reduce background contamination during spraint processing were those described in detail by O'Connor et al. (2020); briefly, a 100% cotton laboratory coat was worn during all stages of processing and microplastic characterization, while nitrile gloves were worn during sample processing. All solutions used in the laboratory (e.g., KOH and NaCl) were prepared using ultrapure water (resistivity: 15.0 MΩ•cm), and all final solutions were vacuum filtered prior to use. Work surfaces were wiped down with 70% ethanol prior to processing of samples. Foil trays, glass pipettes, spatulas, sieves, and all glassware used in the filtering process (e.g., Büchner funnel and flask) were rinsed thoroughly (three times) with ultrapure water so as to reduce cross contamination. Jars used for treating samples were first washed in 0.05% HNO3 and, together with screw cap lids, rinsed thoroughly with ultrapure water. Procedural blanks (negative controls) (n = 13), as well as controls for airborne contamination (n = 10), were evaluated for exogenous particles in conjunction with spraint samples.
Due to observed variation in contamination levels between batches of spraints processed, mean values of contaminating particles from both air controls and blank samples were calculated on a batch-by-batch basis for each microplastic criteria. These were categorized based on their combined morphology type (i.e., fiber, fragment, and film), color, size in the case of fragments, and film (longest axis) (i.e., 100 < 350 μm and 350 μm to ≤5 mm), or width in the case of fibers (i.e., <20, 20–40, 40–60, and >60 μm). Hence, the mean values of contaminating particles in air controls and blank samples were subtracted from the raw number of matching particles in the field samples of each corresponding batch. The mean abundance of exogenous particles was 2.2 ± 0.6 particles/filter (mean ± SE) in air controls (n = 10) and 0.9 ± 0.4 particles/filter in blank samples (n = 13). Fibers were the dominant particle type found in both sets of laboratory controls (air: 68%; blanks: 92%).
Diet analysisFollowing microplastic isolation, spraints were soaked for 24 h in a 10% hydrogen peroxide solution (H2O2) at room temperature so as to degrade remaining organic material (Smiroldo et al., 2019). Samples were then gently washed through sieves of 710-μm mesh and 355-μm mesh, and hard parts (e.g., otoliths, vertebrae, jawbones, and frog scapula) as well as benthic invertebrate remains, fruit seeds, and plant material removed with a featherweight forceps and stored. Remains were identified under a stereo microscope (Micros Austria—Hornet Micro Zoom 1280) using relevant identification keys (e.g., Conroy et al., 2005), though fish were mostly identified from the otoliths (sagittae and asterisci) as these were better preserved. As the microplastic isolation process may have resulted in the destruction of certain hard part remains, prey abundance was not likely to be accurately estimated, and thus, only presence/absence data were recorded for prey items. Salmonid (e.g., brown trout [Salmo trutta] and Atlantic salmon [Salmo salar]) caudal vertebrae that were in good condition were measured for a number of samples (n = 16) (centrum length) using Image-Pro Plus software (QImaging 2000R digital camera) and fork lengths back calculated using a regression described in Hajkova et al. (2003). Dietary results were expressed as percent frequency of occurrence (F%) (number of occurrences of a particular prey item/number of spraints examined containing dietary remains) × 100.
Data analysesMicroplastic data for spraints were visually inspected using Q–Q plots and tested for normality using the Shapiro–Wilk test, which revealed that the data followed a non-normal distribution. Although data transformations were performed in order to normalize the data (e.g., log transformation, square-root transformation), microplastic concentrations (MPs/g dw) remained non-normal and were subjected to nonparametric analyses only. All data analyses were carried out using the “base” package in RStudio (version 3.5.1) unless otherwise stated, and the significance threshold for all tests was set at p ≤ 0.05.
Kruskal–Wallis tests were performed to investigate the variation in median microplastic concentrations between sampling regions (n = 3), while a Mann–Whitney U test was employed to test for differences in median microplastic concentration between spraints expected to be of higher and lower microplastic exposure. Kruskal–Wallis tests were also performed to investigate the variation in median microplastic burden between spraint condition class (n = 3; fresh, drying, and dry) as well as sampling seasons (spring, summer, and autumn). Two Kruskal–Wallis tests were performed in this regard, one for all regions combined and another for the River Slaney catchment independently, which was the only region to represent all three sampling seasons. One sample opportunistically collected from the River Slaney catchment in January was combined with autumn samples as it had the lowest sample size (n = 10). Where significant, a Dunn's test of multiple comparisons was performed using the “dunn.test” package (Dinno, 2017) with pairwise p-values adjusted using the Bonferroni correction method.
Multivariate dispersions (variances) of dietary groups (e.g., regions and seasons) were analyzed using the Betadisper test and the dietary composition of groups with the same multivariate spread (i.e., homogeneity of multivariants) (assumption of permutational multivariate analysis of variance [PERMANOVA]) investigated for differences with Adonis (PERMANOVA) using a Bray–Curtis dissimilarity distance matrix following binary (presence/absence) standardization. In order to visualize the dietary composition of otter spraints by region, a nonmetric multidimensional scaling (NMDS) ordination was produced based on the same Bray–Curtis dissimilarity matrix, which was generated using the step-across shortest path method. Microplastic abundance (MPs/spraint) and microplastic concentration (MPs/g dw) were fitted to the ordination in order to establish whether there was any relationship with diet. All dietary statistical analyses including the Betadisper test, Adonis PERMANOVA, NMDS, and fitting of microplastic abundance and concentration were performed using the “vegan” package (Oksanen et al., 2007).
RESULTS Microplastic prevalence and characteristicsOverall, 53 spraints opportunistically collected from eight river catchments spanning three regions of Ireland were assessed for microplastics. A total of 60 potential microplastic particles (100 μm to 5 mm) were recovered and isolated for μ-FTIR analysis, but two smaller particles (∼100 μm) proved too difficult to verify and were excluded. Of the 58 particles analyzed, approximately 25% were determined to be semisynthetic (e.g., cellulosic) or natural (e.g., chitin) and the identity of approximately 7% could not be determined due to ambiguity in the interpretation of the spectrum result. The remaining 40 particles included acrylic, acrylic copolymers, nylon, polyamide, polyacrylonitrile, polycarbonate, polyisoprene, polypropylene, polypropylene copolymers, vinyl copolymers and terpolymers (e.g., vinyl pyridine and methyl acrylate), and other copolymers (e.g., styrene methyl methacrylate) and synthetics. Having accounted for background contamination, a total of 36 particles were accepted as microplastics and carried forward for data analysis.
Microplastics were recovered in 57% of spraints at a mean abundance of 1.2 ± 0.1 MPs/spraint (mean ± SE) and a mean concentration of 3.8 ± 0.6 MPs/g dw. Fibers were the dominant microplastic type recovered (85%) (width: 5.3–30.9 μm), followed by film (10%) (Figure 2a), while microplastics in the 350 μm to ≤5 mm range were the dominant size class (75%) (Figure 2b). Based on the size categories employed, the minimum microplastic size recorded was 119.3 μm, while the maximum size was 3.2 mm.
Enlarge this image.FIGURE 2. Total abundance of (a) microplastic (MP) types and (b) microplastics size classes for spraint samples analyzed (Slaney: n = 35, western region: n = 8, and southwestern region: n = 10)
Microplastics were present in five spraints from the western region (63%, n = 8), 21 spraints from the River Slaney catchment (60%, n = 35), and four spraints from the southwestern region (40%, n = 10) (Figure 1; Table 1). Although spraints collected from the River Slaney catchment generally contained more microplastics (Table 1), no significant difference in median microplastic concentrations was observed between the three regions (Figure 3a, Kruskal–Wallis, χ2 = 4.75, df = 2, p = 0.093). Overall, microplastics were recovered in 13 spraints collected from higher exposure sites (59%, n = 22), which were located downstream of identified sources and 17 spraints collected from lower exposure sites (54%, n = 31), located upstream. Again, no significant difference in the median concentration of these two exposure levels was detected (Figure 3b; Mann–Whitney U, W = 348.0, and p = 0.903).
Enlarge this image.FIGURE 3. Microplastic (MP) concentration (no. microplastic particles/g dw) (n = 53) for spraint samples per region (a), exposure level (b), condition (c), and season (d). Boxplot midline shows the median, while lower and upper limits show the first quartile (Q1) and third quartile (Q3), respectively, with the box representing the interquartile range (IQR). Upper whisker represents Q3 + IQR × 1.5, while the lower whisker represents Q1 − IQR × 1.5 with open circles indicating the outliers
While there was no significant variation in median microplastic concentration across the three condition classes assigned to otter spraints (i.e., fresh, drying, and dry) (Figure 3c; Kruskal–Wallis, χ2 = 2.45, df = 2, and p = 0.294), it was found that fresh (n = 31) and drying (n = 11) spraints did contain more microplastics, on average, than dry spraints (n = 11) (fresh: 3.9 ± 0.1 MPs/g dw; drying: 4.1 ± 0.2 MPs/g dw; and dry: 2.6 ± 0.4 MPs/g dw) (mean ± SE). A significant seasonal variation was observed however (Figure 3d, Kruskal–Wallis, χ2 = 11.66, df = 2, and p = 0.003), with spraints collected in autumn (n = 10) observed to have a significantly higher median microplastic concentration than spraints collected in spring (Dunn's test; p = 0.002 and n = 28) and summer (Dunn's test, p = 0.020, and n = 15). This was also true for the River Slaney catchment when assessed independently (Kruskal–Wallis, χ2 = 8.04, df = 2, and p = 0.018), where autumn spraints (n = 10) were found to have a significantly higher concentration than spraints collected in spring (n = 14) (Dunn's test and p = 0.014).
DietA total of 20 dietary items were recovered from 53 spraint samples, the majority of which were fish (F% = 85%) (Table 2). Of the fish remains identified, salmonids were the most frequently encountered taxon (F% = 62%) having been found in 33 spraints (26 from the River Slaney catchment), followed by three-spined stickleback (Gasterosteus aculeatus) (F% = 25%), cyprinids (e.g., Phoxinus phoxinus) (F% = 19%), and European eel (Anguilla anguilla) (F% = 8%). A large proportion of spraints (F% = 26%) also contained nondiagnostic fish remains (e.g., head bones) that were recorded as “unidentified.” The estimated mean fork length of predated salmonids ranged from 66 to 117 mm in the River Slaney catchment (n = 11) and 87 to 150 mm in the southwestern region (n = 5).
TABLE 2 Diet of Eurasian otter expressed as percentage frequency of occurrence (%) for the overall study area as well as each region (Slaney:
| Diet item | Frequency (%) | |||
| Overall | Slaney | Southwest | West | |
| Fish | 85 | 89 | 100 | 63 |
| Salmonid | 62 | 74 | 60 | 13 |
| Anguilla | 8 | 9 | 10 | - |
| Cyprinid | 19 | 26 | 10 | - |
| Gastereous | 25 | 23 | 40 | 13 |
| Unidentified | 26 | 20 | 30 | 50 |
| Inverts | 53 | 46 | 50 | 50 |
| Terrestrial | 2 | 3 | - | - |
| Coleoptera | 13 | 14 | 10 | 13 |
| Ephemeroptera | 2 | - | 10 | - |
| Trichoptera | 9 | 9 | 20 | - |
| Gastropoda | 8 | 6 | 10 | 13 |
| Diptera | 4 | 6 | - | - |
| Bivalvia | 9 | 11 | - | 13 |
| Amphipoda | 4 | - | 10 | - |
| Austropotamobius pallipes | 2 | - | - | 13 |
| Hemiptera | 2 | - | 10 | - |
| Adult insects | 6 | 3 | - | - |
| Unidentified | 25 | 26 | 20 | 25 |
| Amphibians | 6 | 3 | - | 25 |
| Vegetation | 13 | 14 | 10 | 13 |
| Fruit | 2 | 3 | - | - |
| Other | 11 | 11 | 20 | - |
Macroinvertebrates were quite prevalent throughout the spraints examined (F% = 53%) and included a mixture of benthic macroinvertebrates (e.g., Coleoptera, Trichoptera, and Bivalvia), terrestrial invertebrates, winged adult insects, as well as white-clawed crayfish (Austropotamobius pallipes), Ireland's only indigenous crayfish species, which was present in one spraint collected from the western region (River Robe, Co. Mayo). Amphibians, most likely common frog (Rana temporaria), were present in three spraints (F% = 6%), while fruit seeds and vegetation were present in a total of eight spraints (F% = 15%). Six spraints also contained hard parts that could not be identified following the microplastic isolation process but may have been of avian or mammalian origin.
Permutational multivariate analysis of variance revealed that the diet of otter did not vary significantly between sampling region (i.e., River Slaney catchment, southwestern region, and western region) (Adonis test F = 1.67, p = 0.096) or season (Adonis test, F = 1.60, and p = 0.096), while the fitting of microplastic variables to the NMDS ordination (stress = 0.078) showed no existing trend between diet and microplastic abundance (p = 0.558), or microplastic concentration (p = 0.810).
DISCUSSIONWe found microplastics in 57% of otter spraints at a mean abundance of 1.2 ± 0.1 MPs/spraint (mean ± SE) and a mean concentration of 3.8 ± 0.6 MPs/g dw. This confirms the species is ingesting microplastics in Ireland and adds to the work of Smiroldo et al. (2019), who incidentally recovered microplastics in otter spraints for the first time but did not report microplastic levels. We found that median microplastic concentrations in spraints did not significantly differ between sampling region (River Slaney catchment, southwestern region, and western region) nor did concentrations vary depending on whether spraint samples were collected upstream (lower exposure) or downstream (higher exposure) of the microplastic sources identified. While no significant spatial variations in median microplastic concentrations were detected, it was observed that the River Slaney catchment, which is deemed to be a high-risk catchment due to a higher density of likely sources (Mahon et al., 2017), had the highest microplastic concentration of all three sampling regions assessed (mean: 4.8 ± 0.8 MPs/g dw and median: 1.0 MPs/g dw).
This study presents evidence to support the assumption that secondary ingestion (i.e., trophic transfer) is the most likely pathway for microplastic interaction among top predators in freshwater ecosystems. D'Souza et al. (2020) found microplastics (500 μm to 5 mm) in the fecal and regurgitate remains of white-throated dipper from four catchments in south Wales (UK) and discovered that particles were largely similar in morphology and size as those previously described in key prey items (i.e., Ephemeroptera and Trichoptera) for the same region (Windsor et al., 2019). We too observed similarities in the characteristics of microplastics recovered in spraints from the River Slaney catchment and those previously found in brown trout from the same catchment (O'Connor et al., 2020), particularly in regard to the dominance of fibers (94%) and larger particles (78%) (350 μm to ≤5 mm), though for brown trout, there was a considerable proportion of fragments recovered (GITs: 25%; stomach contents: 24%). Interestingly, concentrations recorded by D'Souza et al. (2020) in dipper regurgitates (Wales, UK) were double that of spraints in the present study (7.65 ± 1.64 particles/g dw), while fecal samples contained over four times the concentration (15.85 ± 2.85 particles/g dw). However, it should be noted that 26% of the particles recovered in these samples also constituted meso- and macroplastics. In contrast, Winkler et al. (2020) observed much lower levels when looking at common kingfisher regurgitates sampled from the Ticino river valley in northern Italy, with just 12 items (63 μm to 3.1 mm) detected in 7.5% of regurgitate samples (n = 133). Discrepancies in handling procedures as well as particle sizes analyzed renders the comparability of these bird studies challenging however, particularly as digestion procedures were not utilized in either, although Winkler et al. (2020) did perform a density separation using NaCl to improve recovery of smaller particles.
Reasons as to the observed differences in seasonal microplastic concentrations are unclear, but a more robust seasonal assessment of individual sites would better inform us as to likely causes, which, due to the opportunistic nature of the present study, was not feasible. As well as adverse weather events possibly increasing overland flow and subsequent runoff of microplastics from land during the autumn period (Campanale et al., 2020; Xia et al., 2020) (e.g., first-flush phenomenon), rainfall washout and air mass movements (i.e., wind speed and wind direction) have also been found to be important predictors of microfiber deposition from the atmosphere (Roblin et al., 2020). In the present study, it was found that the total rainfall data recorded at a meteorological station in close proximity to the River Slaney catchment (Oakpark, Co. Carlow) were not that different between the two collection periods, in which a significant difference in microplastic concentration was found (spring: 86.6 mm; autumn: 90.5 mm) (Met Éireann, 2020), though this does not account for rainfall duration and intensity. However, as there is greater agricultural activity during summer and early autumn, it is possible that water courses experience a greater influx of microplastics associated with overland flow following rainfall events around this time of year (Crossman et al., 2020).
As the priority of this study was to assess otter spraints for microplastics and not diet, it is likely that some material was damaged, and thus, dietary remains are possibly underestimated. Of the remains identified, fish, most notably salmonids, represented the main prey item in otter spraints (F%, fish: 85%; salmonids: 62%) conforming to previous Irish dietary assessments (Reid, Thompson, et al., 2013), though invertebrates were also well represented (F% = 53%) and likely resultant of secondary ingestion when consuming fish stomachs. Though sample sizes here were small, we found no difference in dietary composition between sampling season and sampling region, while no existing trend was observed with microplastic abundance or microplastic concentration. Interestingly however, the River Slaney catchment, which had the highest microplastic concentration of all regions, also had the highest occurrence of salmonids (Table 2) though estimated fork lengths of 66–117 mm were below the mean fork length of brown trout previously assessed for microplastics in this catchment (149 ± 42 mm; mean ± SD) (O'Connor et al., 2020). Overall, the estimated fork lengths of salmonids in the present study (66–150 mm) (River Slaney and southwestern region combined) are comparable to those documented elsewhere in Europe, where size classes recovered in otter diet were described as mostly “small” (Lyach & Čech, 2017; Marcolin et al., 2020; Miranda et al., 2006), suggesting a possible preference for smaller individuals, which have a higher catchability (Marcolin et al., 2020; Miranda et al., 2006).
The propensity for particles, particularly fibers, to deposit from the atmosphere to remote areas, is a facet of microplastic research that has received much interest in recent times (Allen et al., 2019; Stanton et al., 2019). Depending on the duration and level of exposure, it is possible that spraints may be contaminated with exogenous particles from the surrounding environment (i.e., postdeposition). Through the deployment of four atmospheric samplers (see Appendix S1), we showed that three sites within the River Slaney catchment experienced deposition rates of 157.9–307.9 MPs/m2/day over a 10-day period. While this approach can offer some insight into potential deposition rates within a given site, it is not envisaged to be a feasible option to account for atmospheric deposition in spraints, particularly in large-scale studies that may comprise multiple sampling sites. In this regard, quality control measures employed by D'Souza et al. (2020) and Winkler et al. (2020), such as the testing of surrounding surfaces and deployment of air controls during sampling, may offer more accessible alternatives to account for atmospheric contamination. Contrary to our expectations, we found that older spraints did not contain higher microplastic concentrations (fresh: 3.9 ± 0.1 MPs/g dw and dry: 2.6 ± 0.4 MPs/g dw), indicating that old spraint may be less impacted by postdeposition than initially thought, or else old spraints, due to their brittle nature, may be more susceptible to particle loss due to wind erosion and wash off from rain.
The current rapid population assessments for otter conducted by state bodies, such as the NPWS, coupled with a rise in the popularity of citizen science initiatives (e.g., NGOs) present a good framework to collect otter spraints over a large geographic area. However, it is also important to note that spraints are purportedly employed by otters as a signaling system, to indicate the use of a food patch for instance (Kruuk, 1992), and so this should be taken into consideration before sampling a large number of spraints from a given area or site. For aquatic biomonitoring, an indicator species should provide a long-term set of observations for a particular water quality parameter, facilitating the measurement of a pollutant in a particular area over a given period of time. For comparability, the species need to be widespread with stable populations that exhibit a moderate tolerance to exposure, but more importantly, retain detectable levels of contaminant concentrations in the body or associated residues, which are reflective of background environmental levels (Cleansea Project, 2016; Gerhardt, 2009). Moreover, the selection of an aquatic bioindicator species depends on the specific aims of the monitoring program proposed, the scale of that program (e.g., site specific and regional level) and whether it intends to assess short- or long-term exposure. Here, we were interested in assessing the suitability of otter spraints for monitoring local microplastic exposure levels within a freshwater system. While it is known that otters ingest microplastics, it is not known how long they are retained internally and what time scale they represent. Such uncertainties were emphasized in a feeding trial conducted by Carss and Parkinson (1996) who found that while most prey remains were evacuated from the GITs of captive individuals after 24 h of feeding, some remains were discovered in the spraints of two individuals 10 days after a single meal (60 subsequent spraints). Furthermore, Eurasian otter are considered to have the widest distribution of all 13 otter species (Yoxon & Yoxon, 2019), with large home ranges, broad trophic niche breadth, and the ability to exploit multiple food resources, including those of terrestrial origin. For instance, a 3-month Irish telemetry study observed a mean home range of 7.5 ± 1.5 km (mean ± SD) for females occupying mesotrophic rivers and a home range of 13.2 ± 5.3 km for males occupying rivers of various productivity (mesotrophic and oligotrophic) (Ó Néill et al., 2009). Therefore, it is highly likely that spraints collected in a given area may not contain microplastics or even prey collected and ingested from that area, with some coastal spraints (e.g., Achill Island and Killary fjord) also possibly containing marine species, which could not be determined following the microplastic isolation process. Though not reflected in the present study, otters are also opportunistic predators and can exhibit a high spatio-temporal variation in diet, consuming common frog at particular times of the year, adult insects, and even fruit (Reid, Thompson, et al., 2013). Spraint contents, therefore, may not necessarily even be of aquatic origin. Therefore, while it is acknowledged that the criteria used to determine expected exposure levels in the present study may have been too simplistic, having not accounted for additional sources upstream (e.g., atmospheric deposition), or the proximity and scale of sources identified (e.g., UWWTP size and population equivalent), otter mobility and niche breadth mean that spraints likely provide a poor representation of site-specific microplastic levels. This was highlighted by the fact that no spatial variations were observed in microplastic concentrations, either between regions of contrasting risk (as defined by Mahon et al., 2017) or higher and lower exposure areas, thus rendering it difficult to establish “reference” concentrations (i.e., “low” exposure spraints) for the purposes of monitoring microplastics. As a candidate aquatic bioindicator species of microplastic pollution, it is evident that otter spraints would not be suited as biomonitoring tool, at least on a local level, and efforts should focus toward the development of more reliable bioindicators of microplastics in freshwater ecosystems such as benthic macroinvertebrates or bird species (e.g., white-throated dipper), which are more representative of local environmental conditions, both in terms of site fidelity and resources utilized, but are also feeding at lower trophic levels, which may be at greater risk of microplastic exposure (Walkinshaw et al., 2020).
CONCLUSIONSDespite the relatively small sample size, we present the first account of microplastic levels in otter spraints, confirming the species are encountering and ingesting microplastics in three regions of Ireland. No significant differences in microplastic concentrations were detected between sampling regions, nor were concentrations in spraints explained by whether the point of collection was upstream (lower exposure) or downstream (higher exposure) of identified point and diffuse sources. The characteristics of microplastics recovered in spraints from the River Slaney catchment were largely comparable in their size distribution and type (i.e., fibers) to those previously found in their main prey, brown trout, from the same region, providing some evidence to suggest that trophic transfer is the most likely route of microplastic exposure for this top predator species. Though largely an opportunistic and exploratory study, a number of limitations were identified that would limit the viability of spraints for aquatic biomonitoring purposes, at least on a local scale. The extent of otter home ranges means that spraint contents are representative of the entire range and may not be necessarily associated with specific marking sites, and while otter diet in the present study was dominated by fish, the presence of amphibian remains as well as other terrestrial and unidentified food items (possibly avian or mammalian), though low, suggests that microplastics found in spraints may not be necessarily of freshwater origin. However, for the purposes of assessing the ecological risk of freshwater microplastics to this protected top predator, a more robust sampling design that accounts for seasonal variation would augment the work presented here.
ACKNOWLEDGMENTSThis study was funded under the Irish Environmental Protection Agency (EPA) research program 2014–2020, through the research project “Sources, Pathways and Environmental Fate of Microplastics in Freshwater Systems (2016-W-LS-10).” The EPA research program is a Government of Ireland initiative funded by the Department of Communications, Climate Action and Environment. It is administered by the EPA, which has the statutory function of coordinating and promoting environmental research. Though all spraint samples received were not used in this study, the authors would like to extend their gratitude to the following: Members of the Irish Wildlife Trust (IWT) Galway Branch, in particular Lenny Antonelli, staff of the Irish National Parks and Wildlife Service (NPWS), in particular Clare Heardman, Cyril Saich, Declan O'Donnell, Eoin McGreal, Helen Carty, and Jimi Conroy, to Patrick Smiddy of University College Cork (UCC), Denise O'Meara of Waterford Institute of Technology (WIT), Andrew Power, Elena Pagter, María Pérez-Tadeo, and Paula Silvar-Viladomiu of the Marine and Freshwater Research Centre (GMIT), as well as Sorsha Kennedy, Florence Willis, and Brian Power. Finally, the authors wish to thank Steve J. Ormerod of Cardiff University (Wales, UK), as well as one other anonymous reviewer for their comments on improving the manuscript.
CONFLICT OF INTERESTThe authors declare no conflict of interest.
DATA AVAILABILITY STATEMENTData (O'Connor et al., 2021) are available from Dryad:
© 2022. This work is published under http://creativecommons.org/licenses/by/4.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.
Details
; Lally, Heather T 1 ; Mahon, Anne Marie 1 ; O'Connor, Ian 1 ; Nash, Róisín 1 ; O'Sullivan, John J 2 ; Bruen, Michael 2 ; Heerey, Linda 2 ; Koelmans, Albert A 3 ; Marnell, Ferdia 4 ; Murphy, Sinéad 1 1 Marine and Freshwater Research Centre, Department of Natural Resources & the Environment, Galway‐Mayo Institute of Technology, Galway, Ireland
2 UCD School of Civil Engineering, UCD Earth Institute, UCD Dooge Centre for Water Resources Research, Dublin, Ireland
3 Aquatic Ecology and Water Quality Management Group, Department of Environmental Science, Wageningen University & Research Centre, Wageningen, The Netherlands
4 Science & Biodiversity Unit, National Parks and Wildlife Service, Dublin, Ireland