Paint: a ubiquitous yet disregarded piece of the microplastics puzzle

Abstract

Microplastics are widespread pollutants. Microplastics generated from the wear and tear of paints and coatings have recently been modeled to be a large source of microplastics to the environment. Yet, studies focused on microplastics broadly frequently overlook paint microplastics. In this article, we systematically reviewed the primary literature (turning up 53 relevant articles) on paint microplastic sources, identification methods, environmental concentrations, and toxicity to model organisms. Examples of sources of paint microplastics include paints from buildings and murals, crafts and hobbies, cars and roads, marine boats and structures, and industrial systems like pipes, sewers, and other infrastructure. Paint microplastics have been quantified in several marine samples from Europe and, to a lesser extent, East Asia. Reported concentrations of paint microplastics are up to 290,000 particles per kilogram of sediments, with the greatest concentration reported near a graffiti wall. Out of the toxicity studies testing paint microplastics, there have been 68 tested effects in total across all endpoints and organisms and 17 quantified lethal concentration 50% doses (ranging from 0.001 to 20 g/L). Of the tested effects, 45 observed endpoint values in the paint treatment were significantly different from the control (66%)—most of which were tests using antifouling paints. Overall, the number of studies on paint microplastics is small, limiting a holistic understanding of microplastics. Based on our synthesis of the state of the science on paint microplastics, we suggest a research agenda moving forward informed by research gaps.

Introduction

Microplastics are ubiquitous pollutants that have been documented in remote corners of the Earth: from snow and stream water on Mt. Everest (Napper et al., 2020) to deep-sea sediments in the Pacific Ocean, Atlantic Ocean, Mediterranean Sea, and Indian Ocean (Fischer et al., 2015; Woodall et al., 2014). Microplastics are plastic particles less than 5 mm in size (in any dimension; Arthur et al., 2009; Hartmann et al., 2019). Although simplified to one word, microplastics are a diverse class of contaminants that vary in polymer backbones, chemical additives, size, morphology, and color (Andrady, 2017; Hale et al., 2020; Rochman et al., 2019; Wiesinger et al., 2021). Primary microplastics come from products engineered in this size range (e.g., microbeads, pre-production pellets; Arthur et al., 2009; Hartmann et al., 2019); secondary microplastics are broken-down fragments of larger plastic products degraded by physical, chemical, and biological processes (Jahnke et al., 2017; Rochman et al., 2019; Sobhani et al., 2020a).

Some of the first reports of marine microplastic pollution emerged in the scientific literature in the 1970s, with the identification of microplastics in surface waters and fish collected off the coast of southern New England, United States (Carpenter et al., 1972). Since then, the scientific community has defined what is considered a microplastic, and the array of particles considered microplastics has been expanded on. For example, microfibers and tire wear particles were found in the environment but were not considered microplastics until decades later—well after many articles on microplastics were published.

Microfibers were once an overlooked component of the global microplastic problem. Microfibers may be comprised of natural materials (e.g., cotton, silk, hemp), synthetic materials (e.g., polyester, nylon, polypropylene), or a semi-synthetic combination (e.g., paper, viscose, regenerated keratin; Athey and Erdle, 2022). Microfibers enter the environment during the production, use, and disposal of clothing, agricultural, industrial, and other textiles (Liu et al., 2019). Many studies, especially older studies, rely on methods that exclude microfibers, such as the Manta and Neuston nets (reviewed in Pasquier et al., 2022), which frequently have mesh sizes between 333 and 335 µm (e.g., Eriksen et al., 2014, 2023; Law et al., 2014). Many microfibers are thought to be missed in these studies because the diameter of microfibers frequently falls below 50 µm (Barrows et al., 2018; Hung et al., 2021; Liu et al., 2019). Grab samples of water capture more microfibers than trawl net samples (Covernton et al., 2019; Hung et al., 2021) and are more common today.

It was not until 2011 that researchers began to chemically identify microfibers as microplastics in their samples due, in part, to technical challenges (Browne et al., 2011), reviewed in Athey and Erdle (2022). Since then, anthropogenic microfibers have been reported in every continent and ocean (Athey and Erdle, 2022; Barrows et al., 2018). Biota, such as mussels, shrimp, fish, sea anemones, and freshwater clams, consume microfibers (Mishra et al., 2019; Rochman et al., 2015; Romanó de Orte et al., 2019; Su et al., 2018). Overlooking microfibers results in overlooking a biologically- and environmentally-relevant microplastic.

Similar to microfibers, tire wear particles were noted in the scientific literature in the 1970s, yet about three decades passed before these particles were considered microplastics (Cadle and Williams, 1978; Knight et al., 2020). Tire wear particles are formed from the abrasion between tires and roads and are comprised of elastomers, such as rubber (Brahney et al., 2021; Evangeliou et al., 2020). When less than 5 mm in size, these particles fit within the definition of microplastics. The effects of overlooking tire wear particles as microplastics go beyond semantics; tire wear particles have been modeled to be one of the greatest sources of global microplastic emissions to the environment, but empirical data and studies quantifying tire wear particles are lacking (Knight et al., 2020; Lau et al., 2020).

A similar story appears to be unfolding for microplastics generated from paints in the built environment (e.g., roads, buildings, boats). Like tire wear and microfibers, paints are an overlooked component of the microplastic problem. Recent modeling suggests that paints may be a large source, if not the greatest source, of microplastics to aquatic and terrestrial environments (Eunomia, 2016; Galafassi et al., 2019; Paruta et al., 2022; Xu et al., 2020; Zhu et al., 2024). However, researchers have infrequently characterized paint microplastics in the environment. Several studies have reported the presence of pigments/dyes or suspected paints in environmental samples (e.g., Imhof et al., 2016; Rosso et al., 2022) or as a part of microplastic fragments (e.g., Fischer et al., 2015; Hamilton et al., 2021; Horton et al., 2017). However, robust and certain environmental concentrations of paint microplastics are limited, likely due to methodological limitations and a lack of recognition of paints as plastic.

Paints are multi-component materials designed to meet product specifications as liquids and solids (see Table 1 for a summary of paint components, functions, and example chemical compounds). Paints are often divided into two broad categories: (1) solvent-based paints, which use volatile organic compounds as solvents containing dissolved polymers, and (2) water-based paints, which use water as solvents and contain polymer emulsions (80–1,000 nm; Faber et al., 2021). In solvent-based paints, the solvent quickly evaporates, leaving behind the dissolved components to form a coating (Faber et al., 2021; Stoye and Freitag, 1998). Solvent-based paints were more common before the regulation of volatile organic compounds in 2004, and since then, water-based paints have grown in popularity (Faber et al., 2021; Stoye and Freitag, 1998). Paints are frequently applied in multiple layers, such as the primer, undercoat or basecoat, and gloss or topcoats (Caddy, 2001; Stoye and Freitag, 1998). Paints coat indoor and outdoor substrates, such as wood, metal, concrete, plastic, and stone, and can be used in consumer/household or commercial settings (Faber et al., 2021).

Are paints plastics? Paints often contain synthetic organic polymers/plastics (Hale et al., 2020), primarily in the binder of paint, such as alkyd, acrylic, epoxy, and polyurethane, which creates a polymer matrix to embed the other components (Faber et al., 2021; Stoye and Freitag, 1998). Polymers can also be found in paint emulsions (Faber et al., 2021), solvents, and additives (Faber et al., 2021; Stoye and Freitag, 1998). Estimates suggest that paints are 37% plastic on average (Paruta et al., 2022) and that paint microplastics contain greater additive content (by mass) than most non-paint microplastics, despite sharing some chemical constituents like the polymer, flame retardants, fillers, and plasticizers (Turner, 2021). According to the European Council of the Paint, Printing Ink and Artists’ Colours Industry, some water-based paints also contain primary microspheres (between 5 and 80 µm) and microfibers (0.5–50 mm) to enhance paint elasticity, resistance, thicken paint layers, and reduce the weight of paint; however, these paints only comprise approximately 1% of water-based decorative paints (Faber et al., 2021).

How are paints sources of microplastics? When paint on a surface fragments into particles less than 5 mm in size, microplastics are formed. Paint fragmentation may occur during maintenance/repairs (e.g., scraping, sanding, stripping, sand-blasting), application, use, disposal, and environmental weathering (Turner, 2021). Given the vast array of paints in the indoor and outdoor built environment, researchers have begun to model projected emissions to the environment: global estimates include 156 kilotons (kt) of paint microplastics/year (from marine and road markings; Boucher and Friot, 2017) to 640 kt/year globally (architectural, marine, and road markings; United Nations Environment Programme Technical University of Denmark [DTU] 2018). A recent estimate that accounts for more sources of paints from the architectural, marine, road markings, industrial, automotive, and industrial wood sectors projects that 1,857 kt of paint microplastics are emitted into the global environment annually, which would make paint a large source, if not the largest source, of microplastics to the environment (Paruta et al., 2022). Other estimates of paint’s contribution to microplastic emissions are limited to specific countries/regions, such as the Netherlands, Sweden, Norway, and Europe (Galafassi et al., 2019).

Here, we synthesize primary research and provide a baseline research agenda to advance our understanding of the sources, identification methods, environmental concentrations, and toxicity of paint microplastics. Baseline research can later inform holistic microplastic management strategies that do not overlook a significant source of microplastics. The unintentional and deliberate exclusion of “different” types of microplastics, like tire wear particles, microfibers, and paints, limits our scientific understanding of microplastics. Similarly, applied contributions stemming from microplastic research (Coffin, 2023), such as laws and regulations targeting microplastics (Diana et al., 2022; Karasik et al., 2020; Schnurr et al., 2018; Thornton Hampton et al., 2022; Xanthos and Walker, 2017) as well as remediation and cleanup (Dijkstra et al., 2021; McIlwraith et al., 2019; Schmaltz et al., 2020), will only reflect those microplastics that are studied and may miss an ecologically important part of the problem. This study fills a gap by synthesizing the state of the science on paint microplastics, building on previous reviews discussing paint microplastics (Galafassi et al., 2019; Gaylarde et al., 2021; Hale et al., 2020; Torres and De-la-Torre, 2021; Turner, 2021; Xu et al., 2020) by providing an updated and systematic approach, contributing to a more holistic understanding of microplastics.

Primary research on paint microplastics

We conducted a systematic literature review on paint microplastic sources, identification methods, environmental concentrations, and toxicity to model organisms (Figure 1). This exercise was completed as part of a contract for the Canadian federal government to assess the landscape of the research on paint microplastics to inform the state of the science and priority research gaps.

Methods summary

We systematically reviewed the peer-reviewed literature for articles relevant to paint microplastics by searching “microplastic, paint” in Web of Science: Core Collection (all fields) published between January 1, 1900 and February 27, 2023. Two researchers noted the authors, journal, year, and whether the article was relevant to paint microplastics. Articles were first screened based on the title and abstract. If it was unclear whether the article was relevant, we downloaded and read it to come to a conclusion. Only primary research articles that explicitly mentioned paint microplastics in the full text of the article were retained for data extraction (n = 53 articles). We did not extract data from review articles to avoid double-counting. Review articles were retained to contextualize findings in the Introduction and Discussion sections (n = 9 review articles).

Titles and abstracts were reviewed to categorize articles according to primary foci: sources, identification methods, environmental concentrations, and/or toxicity. Articles that did not fit into these categories were also retained for use in the Discussion section (n = 3 articles). If the article’s focus was unclear from reviewing the title and abstract, the whole article was reviewed. Articles in the “sources” category focused on the products, built environment, or activities that produce paint microplastics and transport pathways in the environment. Articles in the “identification methods” category described methods to characterize paint microplastics from environmental samples. Articles in the “environmental concentrations” category determined the extent of paint microplastic pollution in the environment. The “toxicity” category described articles examining the toxicological impacts of paint microplastics on organisms. Articles that had multiple foci were placed into multiple categories. Data extraction points were tailored to the article’s focus (source, identification methods, environmental concentrations, toxicity; see online supplementary material for a detailed description of the data extraction process).

Synthesis from the scientific literature

We found 53 primary research articles and 9 review articles that discussed paint microplastics published between January 1900 and February 2023, with the majority having at least one focal point on method development (n = 40) and environmental concentrations (n = 38), followed by sources (n = 13) and toxicity (n = 11). The first articles on paint microplastics were published in 2014 (Ivar Do Sul et al., 2014; Lima et al., 2014; Song et al., 2014), with the number of articles published significantly increasing between the years 2014 and 2022 (linear regression, F-test, p = .02, online supplementary material Figure S1). We identified only one primary research article on paint microplastics published in 2023, but the literature search ended in February 2023.

Sources of paint microplastics

Sources of paint microplastic pollution include various products, activities, and the built environment (Figure 2). Products described in the literature include household products, such as water-based dispersion paints, which contain primary nanoplastics (less than 1,000 nm in size; Müller et al., 2022), pigments in manufactured coatings, adhesives, plasters, finger paints (Ragusa et al., 2021), and wall paints (Van Wezel et al., 2016).

Marine sources of paint microplastics include coatings on commercial and recreational ships and other structures. Examples include antifouling coatings, anti-corrosion coatings, cosmetic finishes/paints, and tie coat primers (coating between primer and top coating) that can enter the environment while in use and when undergoing maintenance/repair, such as through high-pressure washing, scraping, and sanding (Kumar and Varghese, 2021; Muller-Karanassos et al., 2019, 2021; Piehl et al., 2021; Turner, 2021). Artisanal and commercial fisheries are reported to be sources of paint microplastics due to the weathering and breakdown of gear, such as tubs, flasks, bottles, and nets, as well as the aforementioned vessel paints (Díez-Minguito et al., 2020; Kumar and Varghese, 2021; Lima et al., 2014). Shipping and port activities were also mentioned as sources of paint microplastics (Díez-Minguito et al., 2020; Piehl et al., 2021).

In the built environment, road marking acrylic and thermoplastic paints (Burghardt et al., 2022; Rosso et al., 2022) and surface paints (Horton et al., 2017) were noted as sources of paint microplastics. Microplastics are produced by abrasion between tires and road paints (Seleznev et al., 2021). Other sources of paint microplastics include construction materials (Seleznev et al., 2021), building paints (Rosso et al., 2022), and sewage paints (Horton et al., 2017).

Quantitative estimates of paint microplastics entering the environment were reported for road markings, antifouling paint, and household paint. Road markings were estimated to be between 0.1 and 4.3 g per person annually in Europe, based on samples collected in Croatia, Austria, Poland, and Sweden (Burghardt et al., 2022). Fifteen leisure boat facilities in the Warnow estuary, Germany were estimated to emit 370 million antifouling paint particles annually during in-service use, high-pressure washing, and maintenance/repair (Piehl et al., 2021). Paint microplastics from consumer household products that end up in wastewater resulted in an average of 2.7 µg/L of microplastics in Dutch sewage-treatment plant effluents, with model estimates ranging from 0.2 to 66 µg/L (Van Wezel et al., 2016).

Paint microplastic identification methods

Paint microplastic identification methods were described in 40 peer-reviewed articles. The methods used to identify paint microplastics from environmental matrices varied across studies but often involved a combination of visual identification using microscopy paired with chemical identification using spectroscopy. After sample preparation, such as filtration, density separation, and digestion, paint particles were frequently identified using visual optical microscopy (Bagaev et al., 2018; Díez-Minguito et al., 2020; Fischer et al., 2015; Ivar Do Sul et al., 2014; Jaini and Namboothri, 2023; Järlskog et al., 2020, 2021; Kang et al., 2015a; Lima et al., 2014, 2016; Mengatto and Nagai, 2022; Oztekin and Bat, 2017).

Paint microplastic density ranges from about 1 to 3 g/cm³ (Turner, 2021), while density separation techniques utilized solutions ranging in density from 0.3 to 1.85 g/cm³ to float paint microplastics and other particles. Density separation techniques used included sodium chloride (NaCl) with densities of 0.3 g/cm³ (Kwon et al., 2020), 1.2 g/cm³ (Caron et al., 2018; Järlskog et al., 2020; Mengatto and Nagai, 2022; Pradit et al., 2022), and 1.8 g/cm³ (Järlskog et al., 2020), sodium iodide with densities of 1.6 g/cm³ (Xu et al., 2022a) and 1.85 g/cm³ (Järlskog et al., 2021), sodium bromide with densities of 1.4 g/cm³ and 1.8 g/cm³ (Xu et al., 2022b), and zinc chloride with a density of 1.6 g/cm³ (Horton et al., 2017), 1.7 g/cm³ (Horton et al., 2017; Imhof et al., 2016; Mani et al., 2019), and 1.8 g/cm³ (Imhof et al., 2016; Xu et al., 2022b). After NaCl density separation, Truchet et al. (2022) visually observed and manually sorted through the precipitate for denser antifouling paint particles. In one study, a novel oleo-extraction process was used to density separate particles as dense as 2.2 g/cm³ (described in detail in Rosso et al., 2022).

Fourier transform infrared spectroscopy (FTIR) was frequently used to identify paint microplastics (Caron et al., 2018; Ehlers et al., 2022; Hall et al., 2015; Herrera et al., 2019; Kang et al., 2015b; Kwon et al., 2020; Leistenschneider et al., 2021; Lorenzi et al., 2021; Mani et al., 2019; Pradit et al., 2021, 2022; Rosso et al., 2022; Song et al., 2014, 2015; Truchet et al., 2022; Xu et al., 2022a,b). Other spectroscopy approaches include X-ray fluorescence spectrometry (Leistenschneider et al., 2021; Muller-Karanassos et al., 2019), Raman spectroscopy (Białowąs et al., 2022; Horton et al., 2017; Imhof et al., 2016; Piehl et al., 2021; Ragusa et al., 2021), and X-ray analyzer (EDX; Truchet et al., 2022). Scanning electron microscopy was used in combination with spectroscopy methods, sometimes to identify metals or persistent organic pollutants sorbed to paint microplastics (Białowąs et al., 2022; Pradit et al., 2021; Seleznev et al., 2021; Truchet et al., 2022; Turner et al., 2022). Two studies (Fang et al., 2020; Sobhani et al., 2020b) focused on method development using Raman but did not test environmental samples. Fang et al. (2020) proposed four methodologies for avoiding false positives and negatives when using Raman to identify paint microplastics and nanoplastics. Sobhani et al. (2020b) utilized Raman to identify paint nanoplastics as small as 100 nm in size.

Environmental concentrations of paint microplastics

Paint microplastic concentrations in the environment have been documented in 38 peer-reviewed articles from sampling at 450 locations (see online supplementary material Table S1 for sites, GPS points, ecosystems, and matrices sampled). Paint microplastics were reported in all matrices (water, sediment, and organisms) in most ecosystems (marine, estuarine, freshwater, terrestrial), with the exception that no paint microplastics were reported in water samples from freshwater ecosystems (though freshwater sediments and organisms were sampled). Paint microplastics were reported in water collected from marine (n = 211 sampling sites) and estuarine (n = 21 sampling sites) ecosystems. Paint microplastics were reported in sediments collected from marine (n = 33 sampling sites), estuarine (n = 100 sampling sites), freshwater (n = 45 sampling sites), and terrestrial (n = 6 sampling sites) ecosystems. Paint microplastics were reported in organisms collected from marine (n = 28 sampling sites), estuarine (n = 5 sampling sites), and freshwater (n = 1 sampling site) ecosystems.

Paint microplastic sampling locations have thus far been concentrated in Europe and, to a lesser extent, East Asia, with fewer sampling sites in the Pacific and off the coast of Africa, the Caribbean, Central America, the Middle East, North America, South America, and South Asia (Figure 3). Paint microplastic concentrations ranged across ecosystems, matrices, geographies, and time (Table 2). Overall, the greatest concentration of paint microplastics observed by count was 290,000 particles per kilogram (dry wt) found in terrestrial sediments near a graffiti wall in Berlin (Xu et al., 2022b). The lowest concentration observed across matrices was 0 paint microplastics, which was noted in the marine water column in Brazil (Lorenzi et al., 2021), washwater (stormwater and road runoff) in Sweden (Järlskog et al., 2020), and estuarine sediments in South Brazil (Mengatto and Nagai, 2022) and Plymouth, United Kingdom (Muller-Karanassos et al., 2019).

Paint microplastic concentrations varied across matrices. In estuarine waters, the greatest concentration observed was 0.0758 particles/m³ in the Goiana Estuary (Brazil; Lima et al., 2016). In estuarine sediments, a maximum of 434 paint microplastic particles/L of wet sediment were observed in Plymouth, United Kingdom (Muller-Karanassos et al., 2019). No studies quantifying paint microplastics in water from freshwater ecosystems were identified. Freshwater sediments had a maximum of 13,600 particles per kg of sediment (dry wt) in Sweden near a construction site (Järlskog et al., 2021). Marine waters had a maximum of 4,661 paint microplastic particles/m³ in surface waters off of Japan’s coast (Xu et al., 2022a). Marine sediments had suspected paint microplastic particles (Fischer et al., 2015; Pradit et al., 2022), but no quantified values were reported. The greatest amount observed in terrestrial sediments includes 290,000 particles per kg of sediment near a graffiti wall in Germany (Xu et al., 2022b).

Most of the organisms in which paint microplastics were identified were from marine (n = 7), followed by estuarine (n = 2), and freshwater (n = 1) ecosystems (see Table 3 for the organism, sampling location, and paint microplastic concentration observed). In the cases where no concentration is recorded (Białowąs et al., 2022; Hall et al., 2015; Muller-Karanassos et al., 2019; Pradit et al., 2021), the authors noted the presence of paint but did not report a quantitative amount. The marine organisms that had documented paint microplastic presence were herring (Clupea harengus; Białowąs et al., 2022), cod (Gadus morhua; Białowąs et al., 2022), green sea turtles (Chelonia mydas; Caron et al., 2018), harbour porpoises (Phocoena phocoena; Philipp et al., 2021), Atlantic chub mackerel (Scomber colias; Herrera et al., 2019), arrow worms (Jaini and Namboothri, 2023), and sea snails (Steromphala cineraria, Phorcus turbinatus, and Phorcus sauciatus; Ehlers et al., 2022). The only freshwater organisms that had documented paint microplastic presence were catfish (Arius maculatus) from a local market near the brackish Songkhla Lake, Thailand (Pradit et al., 2021). Shrimp (Parapenaeopsis hardwickii and Metapenaeus brevicornis) were also sampled from this location, but no paint microplastics were reported, only fibers (Pradit et al., 2021). Estuarine organisms included crabs (Neohelice granulata and Cyrtograpsus angulatus; Truchet et al., 2022) and ragworms (Hediste diversicolor; Muller-Karanassos et al., 2019). In addition to the animals identified, one study found nine paint/coating/dye microplastic particles in the human placenta (Ragusa et al., 2021).

One study examined the ratio of paint microplastics (number per m³) in comparison to the number of organisms living in the surface waters (epineuston). Approximately 0.94 ± 0.96 paint microplastic particles/m³ (mean ± standard deviation) were found before the rainy season, and 1.05 ± 0.94 paint microplastic particles/m³ were found after the rainy season in 2012 (Kang et al., 2015a). In 2012, the total number of epineustonic individuals was 23.92 ± 31.84 individuals/m³ before the rainy season and 256.61 ± 184.24 individuals/m³ after the rainy season (Kang et al., 2015a). A similar trend was seen in 2013: approximately 0.25 ± 0.24 paint particles/m³ were observed before the rainy season and 0.28 ± 0.18 paint particles/m³ after the rainy season (Kang et al., 2015a). The number of epineuston observed over the same time period was 103.68 ± 110.81 individuals/m³ before the rainy season and 282.07 ± 296.34 individuals/m³ after the rainy season in 2013 (Kang et al., 2015a).

Toxicity of paint microplastics

Paint microplastic toxicity tests were conducted on a range of model organisms. The model organisms tested include a mouse cell line (L929 cells; Müller et al., 2022), bacteria (Desulfatitalea tepidiphilia; Tagg et al., 2019), yeast (Papiliotrema laurentii; Barlow et al., 2020), freshwater planktonic crustacean (Daphnia magna; Müller et al., 2022), and freshwater green algae (Raphidocelis subcapitata; Simon et al., 2021). Marine animals were also tested, including zooplankton (Calanus spp. and Tempora spp.) and ichthyoplankton (Lima et al., 2014; Molino et al., 2019), the harbour ragworm (H. diversicolor; Muller-Karanassos et al., 2021), the common cockle (Cerastoderma edule; Muller-Karanassos et al., 2021), and crabs (N. granulata and C. angulatus; Truchet et al., 2022). The ichthyoplankton tested were fish larvae and eggs, while the zooplankton included the nauplii of Cirripedia, Hydromedusa larvae, zoea of Brachyura, the copepod Calanoida, Appendicularia, mollusc larvae, Panaeidae larvae, Amphipoda spp., zoea of Euphasidae, Chaetognatha, Mysis Lucifer faxoni, and Isopoda spp. (Lima et al., 2014).

Across the 11 studies that examined ecotoxicological effects broadly, 5 studies assessed whether or not paint microplastics resulted in toxicological effects to 6 model organisms at different concentrations and lengths of exposure (Barlow et al., 2020; Molino et al., 2019; Müller et al., 2022; Muller-Karanassos et al., 2021; Simon et al., 2021). Statistically significant effects were observed 66% of the time, while the remaining 34% had no significant effect observed (Figure 4; n = 68 effects total). The effects studied included mortality (for marine copepod communities; Molino et al., 2019), burrowing behavior (for ragworms and cockles; Muller-Karanassos et al., 2021), lower feeding rate (for ragworms), altered clearance rate (for cockles; Muller-Karanassos et al., 2021), algal growth inhibition (Simon et al., 2021), weight loss (in ragworms and cockles; Muller-Karanassos et al., 2021), the presence of a cellular starvation response (for yeast; Barlow et al., 2020), cell and cell cluster deterioration (for yeast; Barlow et al., 2020), decreased metabolic activity (in a mouse cell line [Müller et al., 2022] and yeast [Barlow et al., 2020]), and metallothionein-like protein concentrations (in ragworms and cockles; Muller-Karanassos et al., 2021). The effects observed did not frequently follow a dose-dependent pattern within studies, except for altered metallothionein-like protein concentrations.

Three studies quantified 17 lethal concentration 50% (LC50) values for five model organisms using four different paints (see Table 4 for experiment details and LC50 values). In two studies (Molino et al., 2019; Müller et al., 2022), immobility was used as a proxy for mortality. The LC50 values observed varied across the model organism and paint types studied, ranging from 0.0011 g/L to 19.9 g/L. Muller-Karanassos et al. (2021) also found that modern biocidal antifouling paint particles resulted in 100% mortality of the common cockle after 10 days at 3.0 g/L.

Although multiple paint types were tested in toxicity studies, antifouling paints were the most frequently studied. The specific antifouling paints tested include historical and modern biocidal paints (Muller-Karanassos et al., 2021), biocide-free silicone paints (Muller-Karanassos et al., 2021), antifouling paints collected from a shipyard (Molino et al., 2019), and commercial antifouling paints labeled as containing copper (Cu), zinc (Zn), and zinc oxide (Simon et al., 2021). Other paints tested include wall paints (Müller et al., 2022) and two coatings comprised of polyether polyurethane and polyester–polyether polyurethane (Barlow et al., 2020). In addition, paint microplastics from environmental samples were isolated from sediments from the North Atlantic Ocean (Turner et al., 2022), Lake Garda, Italy (Imhof et al., 2016), and the Baltic Sea (Tagg et al., 2019).

One study examined the effects of paint microplastics on bacterial communities. The microbiome of paint particles appeared distinct from the microbiomes of water, sediment, non-synthetic cellulose, and common microplastics (Tagg et al., 2019). The paints identified in this study were isolated from sediment grabs from the Baltic Sea and consisted of polyester, alkyd-based, and epoxy resins (Tagg et al., 2019). Multiple paint microplastic biofilms were dominated by the sulfate-reducing bacteria D. tepidiphilia (Tagg et al., 2019).

Discussion: research needs

Despite the rapid growth in publications focused on microplastics in recent years—with nearly 800 studies published in 2019 alone (Cowger et al., 2020), there are comparatively few on paint microplastics (Paruta et al., 2022). For example, we identified just 62 articles published on paint microplastics over the last two decades. The literature on microplastics primarily began in 2004 (Rochman, 2018; Thompson et al., 2004), while the first articles on paint microplastics were published about a decade later in 2014 (Ivar Do Sul et al., 2014; Lima et al., 2014; Song et al., 2014). Although there are exceptions (e.g., Bargagli and Rota, 2022; De Miranda et al., 2021; Galafassi et al., 2019; Hale et al., 2020; Monira et al., 2021; Torres and De-la-Torre, 2021; Turner, 2021; Xu et al., 2020), we suggest that the scientific community studying microplastics has yet to incorporate paint microplastics into the broad definition of microplastics. Synthetic microfibers and tire wear particles were only relatively recently recognized as a component of microplastics. Now, it is time for scientists to recognize that many paint particles less than 5 mm in size are microplastics. We hypothesize that this lag may be due to the difficulty in identifying paint microplastics and the confusion regarding whether paint contains plastics, though paints are approximately 37% polymer content on average (Paruta et al., 2022). Microplastics do not generally have a minimum percentage plastic threshold that must be met to be considered microplastics; some plastic additives, such as plasticizers, comprise up to 70% w/w of plastics (Hahladakis et al., 2018). Recently, for a bill relevant to microplastics in drinking water, the California State Water Resources Control Board defined microplastics as having a “polymer content of greater than or equal to 1% by mass” (Coffin, 2020). Thus, paint microplastics fit within the currently understood definition of microplastics (Hale et al., 2020; Turner, 2021).

Baseline research is needed to better understand the sources of paint microplastics, identification methods, environmental concentrations, and toxicological effects on organisms and populations. The suggested baseline research builds upon the studies identified in our literature review and will provide information that can be used to strengthen our scientific understanding of microplastics and, in the future, be used to inform policy and management.

Sources of paint microplastics

A comprehensive inventory of the sources of paint microplastics to the environment does not yet exist (to the best of our knowledge). The sources noted in the literature are disparate and do not provide a comprehensive overview of the range of paints and coatings that form microplastics. Paruta et al. (2022) modeled global paint microplastic emissions by sector: architectural, marine, road markings, general industrial, automotive, and industrial wood paints. An overview of paints by product category is provided by Stoye and Freitag (1998). Paints are used on automobiles, commercial transport vehicles, marine surfaces/structures, coils, domestic appliances, packaging/cans, furniture, and buildings (Stoye and Freitag, 1998). A comprehensive inventory of paints paired with production, use, and application frequency data could inform the paints that should be prioritized in research on the environmental concentrations, fate, and toxicity of paint microplastics.

A sectoral (Paruta et al., 2022) and a product-based approach (Stoye and Freitag, 1998) could be combined to build a comprehensive inventory of paint microplastic sources. These sources could be used to identify local sources of paint microplastic pollution and inform detailed inventories of the microplastic pollution contributions in a geographic region. Such inventories have been widely used in climate change discussions to determine contributions in reducing greenhouse gas emissions to meet the Paris Agreement’s goals and have been broadly suggested for plastic pollution (Zhu et al., 2024; Zhu and Rochman, 2022). Given the vast array of paint sources and uses, detailed paint inventories would be helpful to feed into broader plastic pollution inventories.

Paint components include binder and resin, plasticizers, pigments, extenders, additives, and solvents (Stoye and Freitag, 1998). Researchers have cataloged the chemical additives (including known toxicological concerns) in plastics broadly, though few paints were included in these studies (Groh et al., 2019; Hahladakis et al., 2018; Hermabessiere et al., 2017; Lithner et al., 2011; Vincoff et al., 2024; Wiesinger et al., 2021). A comprehensive inventory of the chemical compositions of paint components tied to use/application data would aid scientists in prioritizing paints to use in studies examining fate and toxicity.

The complex chemistries of paints and associated chemical transformations remain understudied. As painted surfaces undergo weathering and form microplastics, the fate of the chemical compounds within paint microplastics and the particle itself remains unclear. In plastics broadly, most additives are loosely bound to the polymer; thus, additives and associated chemical compounds (e.g., contaminants, non-intentionally-added substances) can leach into the surrounding environment (Do et al., 2022; Hermabessiere et al., 2017; Li et al., 2016). A few studies quantified Cadmium, Lead (Pb), Cu, and Zn in paint leachate, but did not find detectable levels of Chromium (Cr; Brennecke et al., 2016; Simon et al., 2021). Others found Pb, Cu, Cr, Molybdenum, Titanium, Iron, and Zn in paint microplastics directly (Seleznev et al., 2021; Truchet et al., 2022; Turner et al., 2022). We did not identify any studies characterizing or quantifying organic leachates from paints. If similar to non-paint plastics, over 10,000 chemical compounds have been associated with plastics, over 2,400 of which have known toxicological concerns (Wiesinger et al., 2021). Paints are thought to contain a greater proportion of additives than non-paint plastics (Turner, 2021). Thus, there is a great need to understand the role these additives may play in paint microplastic toxicity. The additives in paint microplastics and the fate of these additives in the environment should be a future research priority.

Plastics can adsorb environmental pollutants (Andrady, 2017; Wiesinger et al., 2021), including both organic (Rochman et al., 2013) and inorganic contaminants (Rochman et al., 2014; Wan et al., 2021). Due to their small size and greater surface area (Turner, 2021), paint microplastics may be able to adsorb pollution in surrounding environments. Some studies identified metals/elements adsorbed to paint microplastics (Imhof et al., 2016; Muller-Karanassos et al., 2019, 2021). Further research determining if paint microplastics can accumulate other pollutants, such as organic contaminants, and the saturation rates will help researchers understand this contaminant’s fate in the environment and potential toxicological concerns.

The rate at which paints undergo weathering or wear and tear from a surface and shed particles into the environment (i.e., the shedding rate) should be examined for paints under the environmental conditions encountered (e.g., vessel paints would encounter water at various salinities, pH levels, wave energies, and temperatures). Paruta et al. (2022) highlighted shedding rates as a study limitation, and we did not find a study empirically examining shedding rates, though few studies examined particle transport to some degree by examining spatial and seasonal variations in paint microplastic pollution (Lorenzi et al., 2020, 2021). In modeling exercises, road markings were estimated to be between 0.1 and 4.3 g per person annually (Burghardt et al., 2022) and 320 tons per capita annually in European countries (Monira et al., 2021); however, these studies could be built upon with empirical paint shedding rate evaluations. Paint microplastic shedding rates in environmentally relevant scenarios would provide insights into the microplastics that scientists should expect to see in the environment.

Paint microplastic identification methods

Although we identified 40 studies quantifying paint microplastics in environmental samples at 450 discrete sampling locations, researchers lack essential analytical tools to characterize paint microplastics in the environment. Researchers can identify microplastics visually using general light microscopy. However, this becomes difficult when differentiating between organic particles (e.g., sand, sediment) and microplastics of similar colors like black, white, brown, and clear (Primpke et al., 2020). Similarly, researchers’ ability to visually distinguish between non-paint and paint microplastics has not yet been determined. We identified at least 12 studies that relied solely on visual observation to identify paint microplastics (Bagaev et al., 2018; Díez-Minguito et al., 2020; Fischer et al., 2015; Ivar Do Sul et al., 2014; Jaini and Namboothri, 2023; Järlskog et al., 2020, 2021; Kang et al., 2015a; Lima et al., 2014, 2016; Mengatto and Nagai, 2022; Oztekin and Bat, 2017), which suggests that paint microplastics may have a distinct and recognizable morphology. Given that the accuracy of visual paint microplastic observations remains unknown, whether these estimates are likely under- or over-counting is unclear. Visual keys for identifying and classifying paint microplastics would help advance this approach.

Researchers commonly use FTIR and Raman to quantify microplastics in environmental samples and chemically identify the plastic type (De Frond et al., 2021; Munno et al., 2020; Primpke et al., 2018). We identified multiple studies that used FTIR or Raman to identify paint microplastics, with those using FTIR outnumbering Raman. Art preservation and forensic sciences provide insight into the relative advantages and disadvantages of these two analytical methods for studying paint chemical compositions. In art preservation studies, FTIR is noted to identify the chemical composition of the paint binder more readily, which is frequently a polymer, and Raman more readily identified the inorganic (e.g., hematite, goethite, cobalt phosphate) and organic pigments (e.g., those in the azo, quinacridone, and perinone families; Bouchard et al., 2009). Consistent with art preservationists, forensic scientists note that FTIR has advantages for studying the resins (e.g., acrylic, epoxy, vinyl perchloride, polyurethane) in automotive coatings, while Raman has advantages for pigments (e.g., lead chromate, rutile, calcium carbonate) due to Raman spectra extending below 600 cm^–1 where many inorganic pigment bands are detected (Lv et al., 2016). Building on lessons learned from art preservation and forensic sciences, we suggest that paint microplastic quantification in the environmental and toxicological sciences should utilize FTIR for polymer identification and Raman if researchers are interested in the pigments. Despite the frequent use of FTIR to quantify paint microplastics in studies, an open-access FTIR spectral library of broad sources of paints does not yet exist.

Researchers have created spectral libraries for microplastics broadly (De Frond et al., 2021; Munno et al., 2020; Primpke et al., 2018), though, to our knowledge, these spectral libraries include few, if any, paints. Commercial spectral libraries exist but can be costly. Law enforcement agencies, such as the Royal Canadian Mounted Police, have created automobile paint spectral libraries like Paint Data Query and the European Paint Group database. These libraries are often used in solving hit-and-run cases and are limited to automobile paints (Bishea et al., 1999; Chang et al., 2003; Duarte et al., 2022; Lavine et al., 2016). Thus, these libraries have limited applicability to a wide array of paint sources.

Fourier transform infrared spectroscopy spectral libraries designed for microplastics broadly frequently classify paint samples as dyes, pigments, or other synthetics, have a low hit quality index, or appear to match the polymer backbone in the paint. Thus, researchers are limited in their ability to chemically identify paint microplastics in the environment using the available spectral libraries and determine the type of paint identified. An open-access FTIR spectral library of paints is an essential analytical tool to further our ability to quantify paint microplastics. Such a library should be created with pristine paints from a paint can or standard reference material paints and validated with environmental samples. This spectral library would ideally improve on scientists’ ability to quantify paint microplastics in environmental samples and potentially trace the paint back to the emitting source by sector or product.

Environmental concentrations of paint microplastics

Paint microplastics have been modeled to be the greatest source of microplastics in the environment, yet we only identified 38 scientific studies quantifying paint microplastics in environmental matrices. This may, in part, be due to the methodological limitations in separating paint microplastics from environmental samples and chemically identifying paint microplastics. Paint microplastics are smaller (<500 µm) and denser (1–2×) than non-paint microplastics—ranging from approximately 1 to 3 g/cm³ (Turner, 2021), whereas non-paint plastics range in density from 0.85 g/ml in polypropylene up to 1.41 g/ml in polyvinyl chloride and polyethylene terephthalate (Law, 2017; Schütze et al., 2022). Thus, density separation, digestion, and size-based filtration techniques for non-paint microplastics likely exclude most paint microplastics. In articles reviewed here, we identified separation techniques ranging in density from 0.3 to 1.85 g/cm³, suggesting that paint microplastics, especially those ranging in density from 1.85 to 3 g/cm³, are being missed in these studies.

Spike recovery tests that spike a known amount of paint microplastics into different, ideally standardized environmental matrices (e.g., Organisation for Economic Co-operation and Development artificial soil) should be conducted using different density separation solutions (e.g., water, sodium chloride, sodium bromide), digestion solutions (e.g., KOH, NaOH, H₂O₂), and methods (e.g., duration, temperature) to determine the amount of paint microplastics recovered from environmental matrices and any potential contamination or damage to paint microplastics across approaches (Lusher et al., 2017, 2020; Schütze et al., 2022). Standardized guidelines and harmonized approaches (e.g., Cowger et al., 2020; Lusher et al., 2020) for improving the reproducibility and comparability of studies quantifying microplastics broadly can guide the quantification of paint microplastics. Ensuring standardized techniques and reporting for extracting paint microplastics from environmental samples can aid in accurate quantification, improving reproducibility and comparability across geographies and over time.

The studies reviewed here counted the number of paint particles per volume (e.g., number of paint microplastics per m³ of water, number of particles per kg of sediment); however, given that paints can be brittle and fragment readily, researchers may consider reporting paint microplastics by mass in addition to count. This would also help to determine if the ecotoxicological studies were environmentally relevant, as the toxicity studies reviewed in this article measured paints by mass (g/L). Given that paint will likely range in density, it is unclear if the toxicological studies are environmentally relevant. After the aforementioned methods have been developed, more field studies that quantify paint microplastics are needed. Specifically, we noted sampling gaps in the Pacific and off the coast of Africa, the Caribbean, Central America, the Middle East, North America, South America, and South Asia.

We did not identify any studies reporting paint microplastics in waters from freshwater ecosystems, though some studies examined freshwater sediments (e.g., Horton et al., 2017; Imhof et al., 2016; Mani et al., 2019; Seleznev et al., 2021) and biota (Pradit et al., 2021). This bias toward marine ecosystems is not unique to paint microplastics and has been documented for anthropogenic microfibers (Athey and Erdle, 2022) and microplastics broadly (Blettler et al., 2018). Rivers are a major pathway for plastics to enter the ocean (Lebreton et al., 2017; Meijer et al., 2021), and most plastics are generated on land (Jambeck et al., 2015). Thus, further research is needed closer to the sources of microplastics (Rochman, 2018).

Studies identifying paint microplastics in organisms were sparse and also focused on marine ecosystems. Only one study identified paint microplastics in freshwater organisms (Pradit et al., 2021), while another two studies did so in estuarine animals (Muller-Karanassos et al., 2019; Truchet et al., 2022). Interestingly, one study identified paint microplastics in human placentas (Ragusa et al., 2021). Limited studies were available examining paint microplastic concentrations in animals broadly. Overall, furthering our understanding of paint microplastic sources and environmental concentrations will help to determine if most scientific studies on microplastic conducted in the last decade failed to examine the modeled top source of microplastics in the environment—paint.

Toxicity of paint microplastics

Scientists’ understanding of the toxicological effects of paint microplastics is in its nascency and focuses on testing the effects of antifouling paints on marine invertebrates. The exceptions include toxicity tests on a mouse cell line (Müller et al., 2022), bacteria and fungi (Barlow et al., 2020; Tagg et al., 2019), and the commonly used freshwater model crustacean D. magna (Müller et al., 2022). Similar to studies documenting the extent of paint microplastics, further ecotoxicological research should be directed at freshwater and terrestrial environments, which are closer to the source of many paint microplastics.

The type of paints most frequently used in toxicology studies were modern and historic (biocidal) antifouling paint microplastics from a paint can (Muller-Karanassos et al., 2021; Simon et al., 2021), whereas others collected environmental samples of paint microplastics from a shipyard (Molino et al., 2019). The types of paints evaluated should include antifouling paints and other paint sources. In the absence of an understanding of the biochemical properties of underwater adhesion, historical antifouling paints are engineered with long-lived biocides such as copper and tin to deter adhesion, and modern replacements have raised many toxicological concerns (Feng et al., 2012; Kraska and Rittschof, 2015; Rittschof, 2019). Thus, the toxicological effects observed from studies evaluating antifouling paints do not provide insights that are particularly applicable to the vast array of paints. Few studies have been conducted on other sources of paint microplastics—road paints, architectural exterior paints (e.g., concrete, wood), varnishes and coatings, paints on other marine structures (e.g., buoys, moors), mural/graffiti paints, industrial paints (e.g., wastewater treatment, oil/gas), and automotive paints (Paruta et al., 2022). An array of paint sources likely results in diverse microplastics—ranging in size, morphology, polymer type, and additives—which for microplastics may broadly influence ecotoxicological effects (Bucci and Rochman, 2022; Thornton Hampton et al., 2022). Sales, use, and application data paired with data regarding paint microplastic concentrations in the environment will aid scientists in prioritizing paint types to test.

Ecotoxicity studies must also take into account environmentally relevant concentrations and exposure duration. Scientists are limited in their ability to comment on the environmental relevance of the concentrations used in ecotoxicity studies, given that relatively few studies document paint microplastic concentrations in the environment. Further research should also prioritize long-term exposures, as the most prolonged exposure identified in our sample was for 33 days (Barlow et al., 2020). This limitation is not exclusive to studies on paint microplastics, as the same limitation has been noted for studies evaluating the effects of microplastics more broadly (de Ruijter et al., 2020).

In comparison to microplastics broadly, 59% (of the 577 total effects tested) had effects observed, while the remaining 41% did not (Bucci et al., 2020), though the number of studies and toxicological endpoints greatly outnumbered those focused on paint microplastics. Given our focus on paint microplastics, many primary research studies that examined antifouling coatings broadly, rather than as paint microplastics, fell outside our scope. However, antifouling paint microplastics were frequently the focus of the toxicity studies identified, although other paints were occasionally tested. For example, wall and ceiling paint microplastics were tested on D. magna and a mouse cell line, resulting in acute toxicity in both models and decreased metabolic activity in the mouse cell line (Müller et al., 2022). One study examined polyether polyurethane and polyester–polyether polyurethane coatings that have applicability across the marine, aerospace, and automotive sectors (Barlow et al., 2020). Acute toxicity tests that quantified LC50 values tested antifouling and ceiling paint microplastics (Molino et al., 2019; Müller et al., 2022; Muller-Karanassos et al., 2021). For D. magna, paint microplastic LC50 values ranging between 0.078 and 0.48 g/L are comparable, albeit on the lower end, to those observed for polyethylene and polystyrene non-paint microplastics, ranging from 0.0085 and 1.84E + 24 g/L (Frydkjær et al., 2017; Gerdes et al., 2019; Jaikumar et al., 2018; Kim et al., 2017; Rehse et al., 2016; Thornton Hampton et al., 2022). However, comparisons are difficult because there are many more general microplastic toxicity studies than paint microplastic toxicity studies. Moreover, there are many differences in study design, such as the polymer shape and length of exposure. Further studies directly quantifying the LC50 values of paint microplastics compared to non-paint microplastics would be helpful because effect thresholds are often used to inform regulations (de Ruijter et al., 2020).

As is the case for microplastics broadly, evaluating the effects of microplastics at the ecological level (populations, assemblages, and ecosystems) is less frequently conducted (Bucci et al., 2020) but would provide important insights that can inform broad environmental targets such as biodiversity goals or water quality monitoring. At the sub-organismal level, further studies may consider evaluating mutagenicity, teratogenicity, and carcinogenicity in animal models, given that the International Agency for Research on Cancer identified paint exposure as a Class I carcinogen to humans but identified no animal studies (IARC, 2012). Further studies are needed to evaluate the toxicological effects of a range of paint microplastics at all levels of biological organization.

Further work informing paint microplastic concentrations in the environment (exposure) and the concentrations that cause effects (hazard concentrations) can be combined to inform risk assessments that more holistically reflect the diverse contaminants that are microplastics. The studies used to inform risk assessments should undergo quality assurance and quality control review (de Ruijter et al., 2020) to ensure that the most reliable studies are informing policy. A risk-based management framework was developed to address the State of California’s legislative mandate to manage microplastics (Mehinto et al., 2022) and used to examine management options in the Laurentian Great Lakes (Hataley et al., 2023). This framework recommends tiered management strategies ranging from monitoring to implementing pollution control measures and is based on species sensitivity distributions to microplastic toxicity (using food dilution and tissue translocation endpoints) and environmental concentrations (Mehinto et al., 2022). Potential pollution control measures suggested in the literature for paint microplastics include stormwater treatment traps (Lange et al., 2021) or paint innovations to utilize proteins and bio-minerals (Jerri et al., 2022). As studies examining the ecotoxicological effects of paint microplastics are conducted and methods are refined for quantifying paint microplastics in the environment, it will be essential to incorporate paint microplastics into risk-based management frameworks to ensure that microplastics management does not ignore what may be an important type of microplastic.

Conclusions

Paint microplastics are an understudied component of microplastic research. Research is beginning to shed light on the sources of paint microplastics, yet more needs to be understood and well-documented regarding their chemical composition. Researchers have started to detect and quantify paint microplastics in environmental samples ahead of methodological developments that would help to determine if these studies are missing or overcounting paint microplastics. Given paint microplastics’ small size, heavier density, and lack of paint-specific FTIR and Raman spectral libraries (other than automobile paint libraries), we expect that researchers are undercounting paint microplastics. Similarly, toxicological studies primarily focus on a sliver of paint microplastics that are known to be toxic—antifouling coatings—and do not have the baseline data available to determine the most environmentally relevant concentrations to test. Further research in this area is greatly needed across paint microplastic sources, identification methods, environmental concentrations, and toxicity. This work sets the stage for scientifically informed environmental policy and management, if warranted, on potentially the greatest source of microplastics. The discourse around microplastics must shift to incorporate paints. A shift in the narrative will set the stage for holistic conversations about microplastics and ensure that future studies do not overlook a widespread and potentially harmful piece of the microplastics puzzle.

Supplementary material

Supplementary material is available online at Environmental Toxicology and Chemistry.

Data availability

All data are available in the article and supplementary material.

Author contributions

Zoie T. Diana (Data curation, Funding acquisition, Investigation, Project administration, Visualization, Writing—original draft, Writing—review & editing), Yuying Chen (Data curation, Investigation, Methodology, Writing—review & editing), Chelsea M. Rochman (Conceptualization, Funding acquisition, Investigation, Methodology, Project administration, Supervision, Writing—review & editing)

Funding

We thank Environment and Climate Change Canada (ECCC; Y.C., C.M.R.) and the Liber Ero Postdoctoral Fellowship (Z.T.D.) for supporting this research.

Conflicts of interest

The authors declare no competing interests.

Acknowledgment

We thank M. Massey for the photograph of the MV Ithaka in Hudson Bay (Churchill, Manitoba).

References