Are northern communities an overlooked source of microplastics and tire wear particles in the Arctic?

Study area and site selection

Iqaluit, Nunavut (63.7467°N, 68.5170°W), is situated in Koojesse inlet in the northwest corner of Frobisher Bay in northern Canada. Iqaluit is the largest community in Nunavut with an area of 52.5 km² and a population of approximately 8,000. The community has few paved roads, no traffic lights, no stormwater drains, and <8,000 registered vehicles (dominated by trucks, all-terrain vehicles, and snowmobiles). Iqaluit is surrounded by barren land with low vegetation and a polar tundra climate that is strongly influenced by the Labrador Current. The average summer temperature during 2020–2022 was 6.1 °C (June to September) with an average wind speed of 3.6 m/s from the south-east (Government of Canada, 2023). In contrast, the average winter temperature during 2020–2022 was –18.9 °C (December to March) with an average wind speed of 4.3 m/s from the north-west (Government of Canada, 2023). The average annual precipitation in Iqaluit was ~217 mm between 2020–2022 (Government of Canada, 2023). Prior to the study period (14–15 July 2022), the highest precipitation event was on June 24^th, 2022 (0.6 mm; Fig. S1) and the last precipitation event was on July 6^th, 2022 (0.05 mm; Fig. S1; Government of Canada, 2023).

The study sites were randomly selected from commercial and industrial areas in Iqaluit (Fig. 1; Table 1) by subdividing each region into equal area sampling units using k-means clustering of 10 m by 10 m land cover classified grids (Spatial Coverage Sampling and Random Sampling from Compact Geographical Strata; Version 0.4-2; Walvoort, Brus & Gruijter, 2023). A total of eight commercial sites were selected from parking lots (n = 4) and roadsides (n = 4), and eight industrial sites were selected from roadsides. Commercial sites were located within the community downtown core, while the industrial sites were clustered to the northwest of the community (due to limited paved surfaces in industrial regions). Commercial areas consisted of storefronts, recreational centers, and a hotel, while industrial areas consisted of construction enterprises, fuel depots, automobile shops, and storage yards. Field surveys were approved by the Nunavut Research Institute (Scientific Research License 01 011 22R-M/01 015 23R-M).

Quality assurance and quality control

Given the ubiquity of microplastics, strict quality assurance and quality control is vital to ensure that concentrations recorded reflect environmental concentrations. In the field, samples were collected against the wind and between sample sites, sampling equipment was wiped with 100% cotton cheesecloth and sprayed with compressed air to clean and to prevent contamination between sites. In the laboratory, to prevent the contamination of microplastics into environmental samples, all solutions, including reverse osmosis water, hydrogen peroxide, and sodium bromide were filtered through glass-fibre filters (Fisherbrand™ Glass Filter Circle, 1.6 µm pore diameter, G6, filter diameter 4.25 cm). A 100% cotton laboratory coat was worn, and hands/gloves were rinsed with filtered reverse osmosis (FRO) water throughout the extraction process to prevent contamination of the samples. Metal and glass equipment were used where possible to prevent potential contamination, and all equipment was triple-rinsed with FRO water before and in between the handling of environmental samples to remove adhered microplastics. Laboratory blanks (n = 2) were performed in sequence with environmental samples and consisted of running solutions through the same laboratory processes as environmental samples (i.e., digestion and density separation) to quantify potential contamination of environmental samples during the extraction process. No microplastics were detected in laboratory blanks.

Field blanks (n = 6) were performed at a random subset of the commercial and industrial study sites and mimicked the field sampling procedures by sweeping microplastic-free road dust (250 g pre-baked at 500 °C for 12 h) off a cork board into the sampling containers. Field blanks were used to capture any potential microplastic or tire wear particle contamination that occurred during the sweeping process (e.g., atmospheric deposition, insufficient cleaning of field equipment). Further, field blanks (1 g) were also spiked with a known concentration of polyethylene (PE) beads (size ranges 75–90 and 212–250 µm) to assess recovery rates and visual limits of detection; blanks were processed in sequence with environmental samples. The recovery for spiked PE microbeads from the field blanks (n = 4) ranged from 70–100% (average = 83%) for sizes 75–90 µm and 90–100% (average = 98%) for sizes 212–250 µm, which is consistent with reported recovery rates (Dimante-Deimantovica et al., 2024). Samples were not corrected based on microbead recovery. The concentrations of microplastics (0.17 n/g) in the field blanks were low and given that all observed concentrations were greater than the limit of detection (LOD; 1.4 n/g), microplastic observations were not blank-corrected. No tire wear particles were detected in field blank samples. The LOD was estimated as the mean microplastic count for field blanks plus three times their standard deviation (Bertrim & Aherne, 2023). Open-air laboratory blanks (n = 2) consisted of an exposed glass-fibre filter in a Petri dish beside the microscope to capture microplastics from the indoor atmosphere. No microplastics were detected in open-air blanks.

Sample collection

On 14–15 July 2022, road dust was collected using a natural-fibre straw brush and a metal dustpan (O’Brien et al., 2021). Briefly, at each site the pavement was swept along a 10−30 m transect (width = 43.2 cm, length of transect depended on quantity of road dust available). For roadside sites, the start of the transect was 43.2 cm from the shoulder of the road. For parking lot sites, samples were collected from parking spaces. Road dust was collected against the direction of the wind. All contents in front of the metal pan (width = 43.2 cm) were collected using a slow sweeping motion to minimize the resuspension of microplastic particles. Road dust contents in the dustpan were sieved at 2 mm to remove large rocks and non-plastic debris, and the <2 mm fraction was retained in an aluminum dish (pre-baked at 400 °C for 4 h), which was subsequently covered and wrapped with aluminum foil and stored in a brown paper bag until analysis.

Microplastic extraction

Road dust samples were first oven-dried at 45 °C for 48 h, homogenized by gently stirring (~1 min), and 1 g of road dust from each site was processed in 0.5 g aliquots to optimize the analytical procedure (i.e., a total of 8 g each for commercial and industrial areas). Then, organic matter was removed using wet oxidation; in a 250 mL Erlenmeyer flask, 40 mL of filtered 30% hydrogen peroxide was added to each road dust sample and digested at 45 °C for 24 h (Hurley et al., 2018). Following digestion, samples were wet sieved (20 µm), and the residue was retained in a 100 mL tall glass beaker by rinsing the sieve with FRO water. The filtrate was retained from a subset of samples (n = 8) to determine if there was particle loss from the sieving process; no microplastics or tire wear particles were detected in the filtrate. Next, microplastics in the residue were subsequently extracted using one water density separation. The water density separation was accomplished by filling each tall beaker with FRO water (~75 mL) and stirring samples on a stir-plate for ~1 min. After the samples were stirred, the walls of the beaker were rinsed with FRO to wash adhered particles off the beaker wall, and the beaker was filled to the 100 mL line with FRO. The beakers were left on the laboratory bench at ambient temperature for 24 h to allow particulates to settle. Following the settling time, samples were filtered by washing the beaker walls with FRO water as the supernatant was decanted onto a glass-fibre filter. This process reduced the surface tension pull of particles and optimized the extraction process. Filters were stored in clear polystyrene petri dishes until visual analysis. The density separation process was repeated twice using filtered sodium bromide (NaBr; density =1.5–1.6 g/cm³) in replacement of FRO water.

Microplastic and tire wear particle identification

Filters were visually analyzed for microplastics and tire wear particles using a stereomicroscope (AmScope version x64). The entire filter surface was analyzed for microplastics, which were identified using the following set of criteria: (1) particles were unnaturally coloured relative to the sample, and appeared homogeneous in material and texture with no visible cellular structures, (2) fibres were a consistent width throughout, (3) particles remained intact when compressed or poked with dissection instruments, (4) particles had a shiny or glossy appearance, and (5) fibres shared no similarities to natural fibres with limited fraying (Kosuth, Mason & Wattenberg, 2018). Given the abundance of tire wear particles, a quarter of the filter surface was analyzed and scaled up. Tire wear particles were identified as being (1) black, (2) elongated/cylindrical in shape, (3) with a rough surface texture, and (4) spongy when poked with a dissection instrument but remained intact (Leads & Weinstein, 2019).

All suspected microplastic or tire wear particles that met at least two visual identification criteria were photographed (AmScope MU100 camera, 3,584 × 2,748 pixels). Suspected microplastics were touched with a heat source (400 °C; Beckingham et al., 2023) while suspected tire wear particles were not touched with a heat source as they do not melt at high temperatures. The length and width of all microplastics that melted were measured using AmScope (Version 4.11.22004.20230115); 95% of the observed tire wear particles were measured. The average length and diameter per particle shape from each site were used for particles with no measurements. Further, a 540 nm blue light (NightSea Stereo Microscope Fluorescence Adapter) was used to capture fluorescent particles difficult to see under a standard brightfield light. In general, visual identification is limited to particles >50 µm, but where possible, particles down to ~20 µm were also identified, albeit a small proportion (~5%). It is important to note that one individual completed all visual analysis and detection favoured microplastics >50 µm due to visual limitations. Fourier-Transform Infrared spectroscopy (FTIR; LUMOS II, Brucker) was performed on a subset of suspected microplastic and tire wear particles using attenuated total reflection (ATR) to determine polymer type. The particles analyzed by ATR-FTIR were not melted and they were selected randomly. A total of 30% microplastics and 22% tire wear particles were picked with clean tweezers and transferred to a slide with double-sided tape (Thornton Hampton et al., 2022). The diameter of the ATR crystal (Germanium; diameter of tip = 100 µm) generally limited polymer analysis to larger particles (>100 µm); fibrous microplastics were particularly challenging due to their narrow diameter. For all particles, a scan time of 16 s was used under low pressure. Subsequently, tire wear particle spectra were corrected for ‘Black Rubber’ in OPUS (Version 8.7). Despite correcting the tire wear particle spectrum, it is important to note there are limitations to using ATR-FTIR for spectral identification of tire wear particles as their dark nature only reflects a small amount of light (Kang et al., 2022). All spectra were uploaded to OpenSpecy to identify polymer type (Cowger et al., 2021). The plastic polymer with the highest Pearson’s Correlation was chosen, with a minimum correlation value of >0.5 due to the challenges of spectroscopic analysis of aged microplastics and tire wear particles.

Data analysis

Field sites were grouped into industrial roadsides (IND-R), and commercial sites (COM-P+R), with the latter comprising commercial parking lots (COM-P) and commercial roadsides (COM-R). The concentration (n/g) of microplastics and tire wear particles per gram dry weight (dw) of road dust was estimated for each site by summing counts from each duplicate and dividing by the sum of road dust mass from each site (~1 g). The concentration of microplastics and tire wear particles were polymer corrected by multiplying the estimated concentration by the proportion of particles identified as plastic by ATR-FTIR. The proportion of polymers confirmed as plastic by ATR-FTIR was calculated by dividing the number of microplastics or tire wear particles confirmed as plastic by the total number particles analyzed by ATR-FTIR and converted to percentage. In this study, the proportion of particles confirmed as plastic was 75% for microplastics and 87% for tire wear particles (see Table S4). Therefore, concentrations of microplastics and tire wear particles were multiplied by 0.75 and 0.87, respectively. It is important to note that the estimated concentrations of microplastics and tire wear particles represent particles that generally ranged from 20–2,000 µm, however, visually identifying microplastics <50 µm was challenging due to their small size.

Mass concentrations (µg/g) were estimated for both microplastics and tire wear particles by using the volume of each particle (see Table S1 for equations used to calculate the volume of microplastics by shape) and the average polymer density (Simon, van Alst & Vollertsen, 2018). Reporting count and mass concentration develops a concrete understanding on the quantity of plastic in diverse environments. This is especially important because microplastics and tire wear particles exist in unique shapes and sizes, and this influences their mass abundance. The volume of microplastics and tire wear particles was calculated for each shape at all sites (see Table S1). Despite the irregular and diverse shapes of microplastic, all microplastics and tire wear particles were simplified to three-dimensional objects to estimate mass concentrations. Mass was calculated by summing the volume by particle shape and multiplying by the average density for shape-specific polymers (see Table S2).

Further, deposition was calculated by dividing the total mass (g) of road dust collected at each site by the sampling area (area (m²) = width of dustpan × transect length; Table 1) and multiplying that value by the concentration of microplastics or tire wear particles (n/g or µg/g). We assumed concentration and deposition represented an accumulation of atmospheric microplastics and terrestrial particles directly emitted from their sources (such as tire wear particles). Hereafter, concentration and deposition (± standard deviation) are reported as per gram dry weight of road dust and per square metre, respectively. The data was not significantly different from normal (Shapiro-Wilk test); therefore, a one-way analysis of variance (ANOVA) was used to test for a statistical difference between the concentration and deposition of microplastics and tire wear particles in commercial and industrial regions. A confidence interval of 95% was used, followed by a Tukey’s test if there was statistical significance. An ANOVA was also used to test for statistical differences between microplastic and tire wear particle characteristics including shape and length across commercial and industrial regions. Further, to estimate the microplastic diversity across groups, a microplastic diversity integrated index (MDII) was used (Li et al., 2021). Briefly, a Simpson’s 1-D index was used to estimate the diversity of microplastic shape, colour, and size bins in road dust samples from commercial roadsides and parking lots, and industrial roadsides. The microplastic community composition was calculated for each group by multiplying the Simpson’s 1-D index results for shape, colour, and size together and taking the cubic root of the value obtained (Li et al., 2021). All figures and statistical analyses were carried out in the software Past (Version 4.12; Hammer, Harper & Ryan, 2001).

Microplastics in Arctic road dust

Concentration and deposition

In total, 102 suspected microplastics (see Table S3 for character information on all suspected particles and Table S4 polymer matches across areas) were identified at 15 of the 16 study sites. The mean concentration of microplastic particles (excluding tire wear particles) was 55.0 ± 97.3 µg/g (5.80 ± 4.44 n/g) across the study area. The concentration was 36.5 ± 68.4 µg/g (5.41 ± 4.69 n/g) in industrial sites and 73.4 ± 121 µg/g (6.21 ± 4.46 n/g) in commercial sites, ranging from 25.0 ± 30.7 µg/g (4.03 ± 4.33 n/g) in commercial roadsides to 122 ± 165 µg/g (8.38 ± 3.89 n/g) in commercial parking lots (See Table S5 for concentrations across sites). There was no statistical difference in count or mass concentrations across groups (p > 0.05; Fig. 2).

Microplastic deposition followed the same pattern; the mean across all sites was 1,688 ± 4,622 µg/m² (130 ± 248 n/m²) with a mean deposition of 3,250 ± 6,338 µg/m² (17.3 ± 20.2 n/m²) in industrial sites and 3,250 ± 6,338 µg/m² (242 ± 319 n/m²) in commercial sites, ranging from 204 ± 176 µg/m² (36.2 ± 31.6 n/m²) in commercial roadsides to 4,152 ± 8,304 µg/m² (448 ± 352 n/m²) in commercial parking lots (See Table S5 for deposition across sites). There was a statistical difference between groups, the count deposition of microplastics were significantly greater in commercial parking lots than in both commercial and industrial roadsides (p < 0.05).

Tire wear particles in Arctic road dust

Concentration and deposition

Tire wear particles were pervasive across the study area, and 1,317 tire wear particles were quantified across all sites (Table S5). The mean concentration of tire wear particles was 282 ± 408 µg/g (75.8 ± 101 n/g) across the study area, with a mean concentration of 83.2 ± 49.1 µg/g (49.3 ± 30.0 n/g) in industrial sites and 481 ± 514 µg/g (102 ± 139 n/g) in commercial sites, from 103 ± 80.5 µg/g (22.2 ± 6.98 n/g) in commercial roadsides, to 859 ± 478 µg/g (183 ± 166 n/g) in commercial parking lots (see Table S5 for concentration across sites). There was a statistical difference in the mass concentration of tire wear particles across groups, commercial parking lots had a greater concentration than both commercial and industrial roadsides (p < 0.05; Fig. 3).

The deposition of tire wear particles followed the same pattern, with a mean deposition of 7,926 ± 14,900 (2,301 ± 5,383 n/g) across the study area; ranging from 227 ± 263 µg/m² (171 ± 181 n/m²) in industrial sites and 11,079 ± 15,644 µg/m² (5,780 ± 9,340 n/m²) in commercial sites, ranging from 1,013 ± 1,091 µg/m² (278 ± 353 n/m²) in commercial roadsides, to 30,235 ± 14,948 µg/m² (8,584 ± 8,629 n/m²) in commercial parking lots (see Table S5 for deposition across sites). There was a statistical difference between groups, deposition of tire wear particles was significantly greater in commercial parking lots than commercial and industrial roadsides (p < 0.05; Fig. 3).