Metal contamination of intertidal sediment and macroalgae in an area impacted by paint from abandoned boats

Metals commonly employed in boat paint (Ba, Cr, Cu, Pb, Sn and Zn) have been determined in 63- μ m-frac- tionated intertidal sediments and in Ulva lactuca and Fucus vesiculosus sampled in the vicinity of abandoned vessels. Metal concentrations in sediment were elevated but highly variable, both between sites and amongst replicates from the same site (e.g., mean Cu ~ 100 to 1200 mg kg (cid:0) 1 ; mean Pb ~ 130 to 6900 mg kg (cid:0) 1 ) due to heterogeneous contamination by metal-rich boat paint particles. Concentrations of all metals except Zn were higher in U. lactuca than F. vesiculosus but in both species metal levels were also elevated and variable. These observations were attributed to contamination by sediment particles and physical and chemical interactions between fine, suspended or deposited paint particles and the algal surface. The latter interactions act as a means by which boat paint metals may enter the foodchain.


Introduction
Contamination of the coastal zone by metals has been recognised and studied over many decades. Amongst the most important metal sources in this environment are mining waste, treated and untreated municipal and industrial effluents, urban and agricultural runoff, and discharges from boating and shipping activities (Tanner et al., 2000;Pan and Wang, 2012;Naser, 2013). Regarding the latter, much of the focus has been on metals in antifouling applications, like Cu, Sn and Zn, and their direct release from vessel hulls and other structures into water (Valkirs et al., 2003;Michaud and Pelletier, 2006). However, more recent studies have highlighted that antifouling paint particles derived from unregulated boating and shipping maintenance and repair activities represent a significant, local source of metals (Singh and Turner, 2009;Soroldoni et al., 2018).
In addition to antifouling particles, Rees et al. (2014) identified abandoned boats as an important source of a wider array of paint particles, including old lead-based formulations, derived from multiple parts of the vessel and whose immediate impact is contamination of local coastal sediments. The extent and evolution of boat abandonment along the coast of southern England was recently investigated remotely by Payne and Turner (2022) from Google Earth imagery. Significantly, it was found that the majority of the several hundred boats identified had been abandoned on the mudflats of protected areas, and often in concealed and sheltered locations, and that the scale of the problem had been steadily increasing over time.
In a field study conducted by Muller-Karanassos et al. (2019) in south west England, it was shown that Cu can accumulate in worms that are exposed to antifouling paint particles in their sandy or silty habitats or through inadvertent ingestion, thereby affording a means by which biocidal metals may enter the foodchain. While most of the study sites identified by these authors were associated with contemporary boating activities, a sheltered embayment (Hooe Lake, Plymouth) was distinctly different in that its shoreline housed several abandoned yachts and trawlers. In the present study, therefore, this site was selected to investigate a broader array of metals associated with paints more generally in both sediments and macroalgae. The latter, primary producers are effective accumulators of metals in coastal waters and serve as useful time-integrated biomonitors of metal bioavailability (Rainbow, 1995).

Study site
along its southern shoreline that range in length from about 5 to 25 m. The precise study area, noted in Fig. 1c and illustrated in Fig. 2a, contains four timber boats (three trawlers and a yacht) that were abandoned between 1996 and 2007.

Sampling and sample processing
Sampling took place during the spring of 2016. Samples of paint (n = 22) ranging from about 1 to 3 cm in size and <1 mm thick were taken from accessible parts of the four boats where layers were visibly peeling, including the hull, cabin, winch, railings and porthole frames, using a pair of stainless steel tweezers. Samples were transported to the laboratory in individual polyethylene specimen bags and stored in the dark until required for analysis.
Samples of surface (<5 mm) intertidal sediment of about 50 g were taken from fourteen sites within the area shown in Fig. 2a using a plastic shovel and transported in individual polyethylene specimen bags. In the laboratory, large stones and fragments of shell and algae were removed using a pair of polyethylene tweezers before the remaining contents were wet-sieved through a 63 μm nylon mesh with the aid of a few mL of Milli-Q water (MQW). The sieved fractions were transferred to a series of polypropylene centrifuge tubes before being centrifuged for 5 min at 3000 rpm using a Centaur 2 centrifuge. Supernatants were discarded and sediments were frozen at − 18 • C and then freeze-dried in an Edwards Modulyo D freeze-dryer for 48 h. Sediment fractions remaining on the mesh were carefully washed into individual specimen bags with the aid of MQW before being frozen and freeze-dried likewise. All sediment samples were stored under desiccation until required for digestion.
The two most common species of macroalga evident in the study area, the chlorophyte Ulva lactuca (sea lettuce) and phaeophyte Fucus vesiculosus (bladderwrack), were collected by hand into individual polyethylene specimen bags. In triplicate, growing apical tips of F. vesiculosus were sampled at the 14 locations where sediment was taken while triplicate whole individuals of the less abundant U. lactuca were sampled from nine of these locations. In the laboratory, algal samples were washed under MQW water to remove epibionts, surficial salts and associated sediments and blotted dry on tissue paper before being frozen, freeze-dried and stored as above.

Paint particle analysis
Paint particles were analysed for various metals (but with the focus on metals commonly employed in pigmented or antifouling compounds: Ba, Cr, Cu, Pb, Sn and Zn) by energy dispersive X-ray fluorescence (XRF) spectrometry using a Thermo Scientific Niton hand-held XRF analyser (model XL3t 950 He GOLDD+) housed in a laboratory stand (Turner et al., 2016). Samples were placed centrally over the detector window with the outer surface facedown and were counted for 60 s each in a plastics mode with thickness correction (with the precise thickness having been measured through the central region using digital callipers). Instrument performance was checked by regular analyses of polyethylene reference discs impregnated with Ba, Cr and Pb and detection limits were derived from three counting errors for each sample and element and ranged from about 10 mg kg − 1 for Pb and Zn to 100 mg kg − 1 for Ba.

Sediment and macroalgae digestion and analysis
All glassware used for sample digestions was soaked for >24 h in 10 % HCl and rinsed thoroughly in distilled water before being used. Acids for digestion and standard preparation were purchased from Fisher Scientific and were TraceMetal grade. Sediment and macroalgae digestions were performed in aqua regia (a 3:1 mix of concentred HCl to concentrated HNO 3 ) and concentrated nitric acid, respectively (Varma et al., 2011).
In triplicate, 200 mg of both fine (<63 μm) and coarse (>63 μm) sediment fractions were weighed in to a series of 50 mL Pyrex beakers. Aliquots of 5 mL of aqua regia were added to each beaker and the contents covered by watch glasses. After 15 min of cold digestion, beakers were heated to 90 • C on a hot plate for 45 min. Cooled digests and MQW rinsings were vacuum-filtered through Whatman 451 filter papers in a Buchner filtration system before being transferred to 25 mL volumetric flasks and diluted to mark with MQW. For quality assurance purposes, an in-house reference sediment (designed for aqua regia digestion) and controls (in the absence of sediment) were processed likewise in triplicate.
Triplicate digestion of 200 mg portions of macroalgae involved the same protocol but 7 mL of concentrated HNO 3 in place of aqua regia and certified reference sea lettuce (BCR 279) and bladderwrack (ERM CD200) rather than the reference sediment.
Algal digests were analysed for Ba, Cr, Cu, Pb, Sn and Zn by collision cell-inductively coupled plasma-mass spectrometry (ICP-MS) using a Thermo X-series II (Thermoelemental, Winsford UK) with a concentric glass nebuliser and conical spray chamber under operating conditions reported by Rees et al. (2014). The instrument was calibrated externally Fig. 1. Location of Hooe Lake in relation to Plymouth and the Plym Estuary, with the precise study area annotated.
using three blanks and five mixed standards in 2 % HNO 3 (within the range 10 to 1000 μg L − 1 and depending on the metal) in 2 % HNO 3 , and internally by the addition of 50 μg L − 1 of 115 In and 193 Ir to all standards, samples and blanks. Sediment digests were analysed for the same metals and both sediment and algal digests were analysed for Al (an indicator of grain size and potential sediment contamination of macroalgae) by inductively coupled plasma-optical emission spectrometry (ICP-OES) using a Varian 725-ES (Mulgrave, Australia) with a Sturman-Masters spray chamber and V-groove nebuliser. The instrument was operated under conditions given in Turner et al. (2016) and was calibrated externally using three blanks and five mixed standards (within the range 1 to 100 mg L − 1 and depending on the metal) in 2 % HNO 3 .
Measured, dry weight concentrations of Cu, Cr, Pb and Zn in the reference materials are compared with certified or indicative values in Table 1. Agreement between measured and certified values is within 15 % except for Cu and Pb in ERM CD200. Here, low certified concentrations (and considerably lower than corresponding concentrations in the samples) were overestimated and returned with relatively high variation. The Cr concentration measured in BCR 279 is about 50 % lower than its indicative value meaning that some Cr data for macroalgae may be underestimated.

Metallic composition of paint samples
Concentrations of the principal metals in the paints sampled from the four abandoned boats are summarised in Table 2. For each metal, concentrations spanned at least two orders of magnitude, with maximum concentrations of Ba, Cu and Zn exceeding 10 % by weight. Amongst the samples, the six different metals were heterogeneously distributed and contributed between <0.1 to about 50 % of the total paint mass. Most samples were layered and this heterogeneity presumably reflected the different histories and functions of these layers, including antifouling coatings and primers of different ages. Significantly, many metal-based compounds in older formulations (e.g. organotins and lead-based pigments) have since been restricted or banned.

Metal concentrations in sediments
The concentrations of Al in the fractionated Hooe Lake sediments are shown in Table 3. In the fine fraction, concentrations are similar both between sites, and averaging about 13,000 mg kg − 1 , and within each  Table 1 Certified or indicative concentrations in the reference materials compared with values determined in the present study. Errors represent 95 % confidence intervals (certified) or one standard deviation (measured).  site, with relative standard deviations amongst replicates averaging 4.6 %. In the coarse fraction, concentrations are more variable between sites, ranging from <6000 mg kg − 1 to about 14,500 mg kg − 1 , but less variable amongst replicates (average relative standard deviation of 3.1 %). These observations suggest that while fine sediments have a consistent granulometry and coarse sediments have a more diverse granulometry, there is relatively little granulometric variation amongst replicated samples. The concentrations of metals commonly employed in compounds added to paints and identified in Table 2 are shown for Hooe Lake sediments in Table 4. Concentrations of all metals in the fine and coarse fractions are more variable than Al, both between sites and within replicates. For example, average relative standard deviations for fine sediments range from about 13 % for Zn to >30 % for Ba, Cu and Sn, and average deviations for coarse sediments range from about 35 % for Cr and Zn to >60 % for Pb and Sn. Within the coarse fraction of replicate samples, the relative standard deviation sometimes exceeded 100 % for Ba, Cu, Pb and Sn.
According to a series of Wilcoxon sign tests performed in Minitab v19, the only significant differences (p < 0.05) in median concentrations between the fine and coarse fractions are for Cr and Pb (and both greater in the coarse fractions). Results of Spearman's rank correlation analysis revealed the strongest associations (r s > 0.7) between Cu in the fine and coarse fractions, Ba -Cr in the fine fraction and Pb -Zn in the coarse fraction.

Metal concentrations in macroalgae
The concentrations of metals, including Al, in U. lactuca and F. vesiculosus are shown in Tables 5 and 6, respectively. Regarding U. lactuca, metal concentrations are variable both between locations and amongst individuals sampled from the same sites, with a range of mean concentrations exceeding an order of magnitude for Ba and Pb. In F. vesiculosus, metal concentrations are less variable between sites and amongst apical tips from the same individuals. A series of Mann-Whitney U tests performed in Minitab v19 revealed that median concentrations of Cr, Cu, Pb, Sn (and Al) are significantly higher (p < 0.05) in U. lactuca than F. vesiculosus whereas the median concentrations of Zn are significantly greater in the latter.
Spearman's rank correlation analysis revealed a number of significant (p < 0.05) relationships between the concentrations of the same metal or different metals in U. lactuca or F. vesiculosus and in the fine or coarse fraction of sediment, between different metals in either U. lactuca or F. vesiculosus, and between the same metal or different metals in U. lactuca and F. vesiculosus. Amongst the strongest positive associations (r s > 0.7) were: Ba-Al, Cu-Cu and Cr-Sn in coarse sediment-U. lactuca; Ba -Cu, Ba -Pb, Cr -Cu, Cr -Pb, Cr -Zn, Cu -Pb, Cu -Zn and Pb-Zn in U. lactuca; and Pb-Ba, Zn-Pb, Zn-Zn and Zn-Al in U. lactuca-F. vesiculosus.

Discussion
Metal concentrations in sediments of Hooe Lake in the vicinity of abandoned boats are heterogeneous, in both the fine and coarse fractions, with no relationships evident with granulometry. Such heterogeneity may be attributed to the dispersion of metal-rich boat paint particles of diverse compositions and origins, and as exemplified in Table 1, in the intertidal zone. This effect has been observed in other marine systems impacted by paint from abandoned vessels or boats undergoing maintenance (Singh and Turner, 2009;Takahashi et al., 2012;Rees et al., 2014;Wu et al., 2016;Soroldoni et al., 2018) and was directly visible here from colourful paint fragments retained by the 63 μm mesh after sediment fractionation. Metal concentrations in sediment subject to point source particulate contaminants are also likely to vary temporally. Thus, deposited paint particles are eroded into smaller fragments that become buried or washed away, but are replenished by particles derived from both existing and newly exposed paint layers on boat hulls and other painted surfaces.
Overall, metal concentrations are highly elevated compared with a  Table 4 Mean (±one standard deviation; n = 3) of metal concentrations (in mg kg − 1 ) in fine (<63 μm) and coarse (>63 μm) Hooe Lake sediments.  As a measure of the health status of Hooe Lake sediments, mean concentrations of metals in both fine and coarse sediments at each site (Table 4) can also be compared with surficial (<5 cm) sediment quality guidelines for Cr, Cu, Pb and Zn in marine and estuarine sediment and available to aqua regia as published by the Canadian Council of Ministers of the Environment (1999). Thus, for Cr, interim sediment quality guideline (ISQG) and probable effect level (PEL) concentrations are 52.3 mg kg − 1 and 160 mg kg − 1 , respectively, with no exceedances in Hooe Lake sediment. Regarding Zn, the ISQG concentration of 124 mg kg − 1 is exceeded in all lake sediment samples but the PEL of 271 mg kg − 1 is only exceeded in six fine and four coarse sediments. For Cu, both ISQG and PEL concentrations of 18.7 mg kg − 1 and 108 mg kg − 1 , respectively, are exceeded in the present study for all fine sediments and all but one coarse sediment sample, while for Pb, both ISQG and PEL concentrations of 30.2 mg kg − 1 and 112 mg kg − 1 , respectively, are exceeded in all cases.
While these comparisons are useful in providing a general health assessment of Hooe Lake sediment, it should be borne in mind that the forms of Cr, Cu, Pb and Zn upon which quality guidelines are based may be different to those encountered in boat paint particles. For instance, Cr guidelines are centred around toxicological data for the less harmful trivalent form of the metal, whereas many paints contain chromate pigments in which Cr exists in the more toxic, hexavalent oxidation state (La Puma et al., 2001). Aquatic toxicological studies involving Pb and Zn usually refer to the bivalent ions, but in antifouling formulations more hydrophobic organometallic Pb and Zn pyrithione complexes have been employed (Dick and Nowacki, 1970;Turley et al., 2000). With respect to Cu, toxicological data are based on Cu 2+ but in antifouling paints compounds employ the metal in its lower (I) oxidation state (Blossom et al., 2018) and whose ecotoxicology is unknown. Regarding Sn, Canadian quality guidelines are unavailable for sediment, and although a UK action level of 0.5 mg kg − 1 for organotins is reported for the management of dredged sediment (Mason et al., 2021), the methods employed in the present study do not discriminate between inorganic and organic forms of the metal.
In most cases, concentrations of metals in U. lactuca and F. vesiculosus exceed the corresponding ranges of "typical" concentrations in macroalgae for the region, where reported by Rainbow (2020). As biomonitors, macroalgae are net accumulators of metals over a period of time in which concentrations reflect average bioavailable concentrations (Rainbow, 1995;Villares et al., 2007). In theory, therefore, concentrations are less prone to spatial and temporal variations of metal concentrations than sediments that are directly contaminated by heterogeneous point sources, and correlations between metals in algae and sediments would not necessarily be expected. However, there is still considerable variation in metal concentrations in U. lactuca and F. vesiculosus sampled a few tens of metres apart in Hooe Lake, and in particular in U. lactuca where relative standard deviations (n = 14) exceeded 50 % for Ba, Cr and Pb.
Part of this variability could reflect contamination of algae by sediment particles that evade washing under MQW. An estimate of the magnitude of this effect may be gained by assuming that all Al in the macroalgae, [Al alg ], arises from adherent sediment grains as follows (Luoma et al., 1982): where [Me alg ] is the measured metal content of the macroalga, [Me] f and [Al] f represent, respectively, the metal and Al concentrations of fine sediment sampled at the same site, and [Me alg ]* is a "corrected" metal concentration in the macroalga. Results of applying this correction to U. lactuca and F. vesiculosus are shown in Fig. 3. For the latter, corrected values are not dissimilar to measured concentrations but for the former, correction reduced concentrations considerably in many cases (and for Sn, to [Me alg ]* < 0). Although it is suspected that the correction applied to U. lactuca may have been over-estimated because of the ability of this macroalga to naturally accumulate Al (Olsson et al., 2020), correction  failed to reduce the variability in metal concentrations observed in either species. An alternative possibility is that macroalgal surfaces are able to capture and retain microscopic paint particles directly from the water column or from bed sediment at low tide that are not readily removed on cleaning in MQW. Turner et al. (2012) suggested that silver nanoparticles adhered to the polysaccharide-rich mucus layer on the surface of U. lactuca could explain the apparent accumulation of Ag by the alga in the absence of any nanoparticle dissolution in controlled laboratory exposures. More recently, Gutow et al. (2016) observed the attachment of microplastic particles to the surface of F. vesiculosus in laboratory experiments and suggested that the effect was enabled by both adherence to the mucus layer and electrostatic binding between hydroxyl residues on the of polymer surfaces and cellulose. The authors also surmised that such interactions could facilitate the release of additives from the microplastic matrix and their subsequent accumulation by the alga. The stochastic nature of these processes and the greater mobility of metal additives in paint particles than in microplastics or nanoparticles (Turner, 2022) could account for both the enrichment and variation of metals observed in U. lactuca and F. vesiculosus in the present study.
These assertions, more generally, imply that macroalgae are not necessarily good indicators of aqueous metal concentrations where heterogeneous, point sources of particulate metals exist. Nevertheless, and regardless of the precise nature of paint-algal interactions, macroalgae may represent an important vector for paint-bound metals to enter the marine food chain.

Declaration of competing interest
The author declares that he has no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability
Data will be made available on request.