Levels and composition of microplastics and microfibers in the South Saskatchewan River and stormwater retention ponds in the City of Saskatoon, Canada

In recent decades, contamination of the environment with microplastics and microfibers has been recognized as a pervasive and ubiquitous issue of global concern. While much research in this field has been undertaken in marine environments, more recent studies have identified rivers as major conveyors of plastic pollution from terrestrial into marine systems. However, reports on the levels and composition of microplastic and microfiber contamination in rivers of the Canadian prairie region, specifically the South Saskatchewan River (SSR), are scarce, which leaves this vital source of water for societies and ecosystems in a vulnerable state. To fill this gap, we obtained samples from seven sites along the Saskatchewan portion of the SSR, as well as three stormwater retention ponds (SRP) in the city of Saskatoon during the spring, summer, and fall of 2020. We used optical and Raman microscopy to enumerate and characterize particles in these samples. Total levels of particles and fibers in all samples ranged from 32 to 116 particles m−3. Most particles (approximately 77%) were natural fibers, while polymers accounted for the remaining 33%. Average microplastic levels were lower (3.18 ± 3 particles m−3) downstream of Lake Diefenbaker, a large reservoir on the SSR, compared to upstream (12.0 ± 9 particles m−3). Retention of microplastics in the reservoir could explain the lower mean microplastic concentration of 4.43 ± 3 particles m−3 recorded in the SSR compared to mean concentrations of 26.2 ± 18 particles m−3 reported in the North Saskatchewan River, which is not dammed. This study is among the first to describe microplastic and microfiber levels in the SSR and thereby helps improve our understanding of this pervasive environmental contaminant on the Canadian prairies.


Introduction
Global plastic production has increased exponentially since the early 1950s, reaching more than 370 million tons annually in 2020. A cumulative 8.3 billion metric tons of plastics have been estimated to have been produced by 2015, of which more than 75% became waste rather than being recycled or still in use today. This number has grown to 9.54 billion metric tons in 2019 (OECD 2023). While governments and industries around the globe have increased their efforts to reduce plastic waste in recent decades, it is estimated that less than 10% of global plastic production has been recycled (US EPA 2017). Additionally, plastic recycling does not often follow a circular economy scheme, but recycled plastics are spun into fibers or extruded into films that themselves cannot be further recycled. Last, improper waste management at landfills, windblown litter, and deliberate littering have resulted in a considerable fraction of plastic waste being released into the environment. This trend has recently been fueled by the COVID-19 pandemic, which has led to the skyrocketing generation of plastic waste in the form of single-use medical supplies such as face masks, gloves, and gowns (Liang et al. 2022).
Ultimately, many of the improperly disposed of plastics end up in the environment, particularly in aquatic systems, where biota inhabiting these ecosystems can be adversely affected (Wang et al. 2021). Plastic debris occurs in a variety of shapes and sizes, but plastic particles in a size range from 1 to 5 mm, also named microplastics, have received particular attention in the scientific literature and the public discourse (Arthur et al. 2009). Primary microplastics have in the past been produced directly in the form of microbeads used for cosmetic products and industrial scrubbing agents; these have made their way into aquatic ecosystems through wastewater treatment plants (McCormick et al. 2014;Talvitie et al. 2017). Due to significant concerns related to ecosystem health, microbeads have been recently banned in many legislations around the globe. Consequently, secondary microplastics that form during the breakdown of larger plastic items through physical, chemical, and biological degradation processes, including microfibers that are emitted when garments from synthetic textiles are laundered, have become the dominant form of microplastics detected in aquatic systems (Singh and Sharma 2008;Henry et al. 2018).
Many studies to date have focused on microplastic issues in the marine environment, but it is important to examine how these plastics get there. Rivers are today recognized as the most important conveyors of plastic debris and microplastics into our oceans, rather than direct disposal at sea or abandoned fishing gear, etc. (Lebreton and Andrady 2019;Mai et al. 2020). However, accurate estimates of river fluxes have been scarce compared to measurements of densities in marine systems, in particular since sampling methods are more complex and rivers are highly dynamic systems (Lebreton and Andrady 2019;Weiss et al. 2021). This has led to considerable knowledge gaps in terms of developing a global understanding of river fluxes. In particular, inland rivers have received very limited attention.
The objective of the present study was to provide critically needed information on the levels and composition of microplastic and microfiber contamination of an inland river, the South Saskatchewan River (SSR) in Canada. It is a major lifeline of the Canadian prairie region and a water resource of critical importance to the prairie provinces of Alberta, Saskatchewan, and Manitoba. Roughly 70% of Saskatchewan's population relies on the SSR for their water supply, including irrigation, drinking water, and industrial applications.
To this end, we obtained samples from seven sites along the Saskatchewan portion of the SSR, as well as three stormwater retention ponds in the city of Saskatoon during the spring, summer, and fall of 2020. We applied a previously published methodology (Prajapati et al. 2021) to extract microplastics and microfibers from these samples using a plankton net and subjected them to Fenton oxidation and density separation for cleanup in preparation for enumeration and identification of particles using Raman microscopy. Through selecting sampling sites upstream and downstream of Lake Diefenbaker, a major reservoir on the SSR, upstream and downstream of a wastewater lagoon and a treatment plant, as well as in stormwater retention ponds, we also attempt to describe the major sources of particles and fibers in the SSR.

Sampling sites
Water samples were collected at seven sites along the South Saskatchewan River and at three selected storm ponds in Saskatoon (Saskatchewan, Canada) during three individual sampling events between May and September 2020 (Fig. 1). Specific dates were May 20, 23, and 28 in May 2020, June 30 and July 3 and July 7 in June/ July 2020, and August 10, 13, and 14 in August 2020. Many sampling timepoints were preceded by spring rainfall events, while all other sampling dates were preceded by dry weather conditions. The South Saskatchewan River flows primarily in a North-Easterly direction. It passes through one major city (Saskatoon) and one reservoir (Lake Diefenbaker  (52.175632, −106.574932) are all located within Saskatoon urban areas. Evergreen and Aspen Ridge storm ponds collect surface runoff water mostly from residential settlements, while the Northeast Swale is located in a nature and wildlife conservation area of the city. A few of the sample sites have additional distinguishing characteristics. For example, the town of Outlook has a wastewater lagoon that was discharged in late spring between May 18 and 23, 2020. Samples were collected upstream and downstream of the lagoon prior to flushing in May, and after flushing in June/July and August/September 2020.

Sample collection, treatment, and extraction
Water samples were collected by applying the method used by Prajapati et al (2021). Water was collected by pumping 500 L of water just below the surface near the shore through a 10-µm plankton net using a metered pump with the net terminal connected to a 125-mL pre-screened collection bottle (Prajapati et al. 2021). The pump inlet contained a 5-mm screen, effectively limiting the size range of detectable plastics between 10 µm and 5 mm. This ensured microplastics were concentrated about 400fold prior to further treatments. Concentrated samples were stored in a refrigerator to maintain their original identities prior to digestion. The samples were further processed prior to extraction by Fenton oxidation overnight at room temperature to remove any organic matter present. This was done by first acidifying each sample with concentrated sulfuric acid (pH < 3.0), followed by a stepwise addition of Fenton's reagent, i.e., 0.25 g of ferrous sulfate, FeSO 4(aq) (Sigma, > 99%) and 12.5 mL of hydrogen peroxide, H 2 O 2(aq) (Fisher scientific, 15% w/w) (Prajapati et al. 2021).
Microplastics were extracted by density difference using a small glass separator unit as described by Nakajima et al (2019). A schematic of the separator unit is provided in Fig. S1. A known volume of the digested water sample was transferred into the separator device with the upper and lower plates joined and clipped together to prevent any loss of sample. The glass jar was rinsed several times with deionized water (18.2 MΩ cm) to ensure all particles were transferred into the glass separator unit. Approximately 160 g NaI (99 + %, Alfa-Aesar, Thermo Fisher) was added to the water sample in the unit, and the volume was increased to 150 mL with ultrapure water to achieve a solution density of 1.6 g cm −3 . The unit was then covered with aluminum foil to prevent exposure to light and other external contaminants before gently stirring the solution for ~ 30 min using a glass covered stir bar. The solution was allowed to settle for ~ 4 h, and then the upper half of the device -which is expected to contain the microplastics and other less dense particles -was separated from the lower half containing settled denser materials or sediments by sliding the upper plate over the lower plate such that the glass sheet extending from the top of the lower plate acted as a base for the liquid in the upper plate. This enabled separation of the liquids in the upper and lower compartments without loss of materials or disturbances to the system.
The solution in the upper half of the glass unit was carefully transferred to a glass funnel containing a 0.7-μm pore size Whatman glass microfiber filter paper (GFF) and then filtered. The internal wall of the upper half of the glass separator device was rinsed into the funnel several times with deionized water to capture any particles adhered to the glass wall. Following extraction, the GFF was rinsed several times with deionized water to remove all NaI. The GFF was then allowed to air-dry while covered with aluminum foil in a laminar flow hood to ensure all particles collected remained in their original physical and chemical states and to prevent any loss or contamination from airborne plastics. Control experiments using virgin polymer samples (500-1000 μm), including nylon 6, polyethylene (PE), polypropylene (PP), polyethylene terephthalate (PET), polysulfone (PSF), and polyvinyl chloride (PVC) indicated that microplastic recovery was 98.3 ± 0.6% to 100 ± 0.0% for microplastics of densities ranging from 0.90 to 1.68 g cm −3 (Fig. S2). A full description of these experiments is provided in the SI.

Microplastic quantification and characterization
The extracted microplastics on the GFF were quantified visually under white light using a Renishaw confocal microscope. The samples were brought to a sharp focus with a 10 × /0.25 numerical aperture Leica objective. A montage of dimensions ranging from 350 × 350 µm to 500 × 500 µm covering the entire surface of the filter was acquired, and particles were then counted manually. Utmost care was taken such that any recalcitrant plant fibers or animal larvae -characterized by inconsistent color and texture or showing segmented surfaces -were not counted. Individual images of each identified particle were acquired using a 40 × /0.60 numerical aperture Olympus objective. These magnified images were used to better determine particle physical structures and colors.
Chemical characterization of each identified particle was performed by obtaining Raman spectra using a Renishaw confocal Raman microscope equipped with a CCD detector, an 1800 l/mm (vis) grating, and a 50 mW 532 nm continuous wave diode-pumped solid state excitation laser. Instrument parameters such as acquisition time, laser power, and number of accumulations were adjusted for each particle to optimize signal intensity and signal-tonoise ratio while preventing saturation of the CCD detector and photochemical damage of the sample. Acquisition times of 5-10 s and laser powers of 5-10% were generally used for pellets and fragments (described below), while acquisition times of 1-5 s and laser powers of 1-5% were generally used for fiber and film particles, which are prone to photo-damage. Static scans centered around 1400 cm −1 (400-1500 cm −1 ) and extended scans between 100 and 4000 cm −1 were collected for each identified particle using the 40 × objective. The collected spectra were further processed by normalizing the peak intensities and removing any fluorescence contributions to the Raman spectra using the baseline subtraction tool in the Wire5.2 software of the instrument. The spectra were then matched with an in-house Raman polymer library to identify the specific polymer make-up of the microplastics using the component analysis tool of the software. The in-house Raman polymer library was developed by collecting Raman spectra of virgin plastic polymers of low-density polyethylene (LDPE), medium density polyethylene (MDPE), high density polyethylene (HDPE), nylon 6, PET, poly (methyl methacrylate), PP, PSF, and PVC as shown in Fig. S3 of the SI. The suspected microplastics were further characterized by comparison with spectra in open-source polymer databases such as Openspecy (Cowger et al. 2021) and PublicSpectra.

Particle concentrations and classification
Total particle concentrations, including microplastics and natural textile fibers, are shown in Fig. 2. Particle concentrations ranged from 32 to 116 particles m −3 . The mean particle concentration, averaged across all sampling sites and sampling times, was 63.2 particles m −3 . The majority of the particles (~ 77%) were from natural fibers, while microplastics accounted for ~ 33%. Among the river samples, the total microplastic concentrations were higher at Miry Creek (12.0 ± 9.1 particles m −3 ) than at all other sampling locations (mean of 3.18 ± 3.0 particles m −3 ). However, this difference was not statistically significant at the 95% confidence interval due to the considerable variability in concentrations. The higher microplastics loadings were likely because Miry Creek is located upstream of Lake Diefenbaker, while all other sites were downstream. It is known that the Gardiner dam located on Lake Diefenbaker acts as a sediment trap; it might act similarly for microplastics before the water discharges into the South Saskatchewan River just upstream of the Outlook municipality. Previous studies have suggested that sediments collected behind dams could be possible sinks for microplastics (Watkins et al. 2019;Gao et al. 2023). This would explain the lower total mean microplastics concentration of 4.43 ± 2.9 particles m −3 recorded in the South Saskatchewan River compared to the mean concentration of 26.2 ± 18.4 particles m −3 reported in the North Saskatchewan River, where no dam is directly constructed on the river (Bujaczek et al. 2021). Similarly, low levels (~ 5.3 particles m −3 ) have been Particle concentrations in the Saskatoon storm ponds were similar to those in the river samples (48.7 particles m −3 ). The Northeast Swale had the largest relative contribution from microplastics (60% compared to about 19% and 24% for Evergreen and Aspen Ridge). The contribution of microplastics to total particle numbers in the Norteast Swale was also much larger than the contribution for the river sites, where microplastics contributed a mean of (18 ± 9)% with a maximum at Miry Creek of 37%. This finding is particularly interesting because the Northeast Swale is an urban conservation area. However, due to its close vicinity to a new development site, much of this contamination might have originated from construction activities and be of temporary nature.
We classified all microplastics (not including natural fibers) under four morphological categories: (1) polymer fibers, which are flexible with a uniform thickness and color; (2) fragments, with rigid structure and irregular shapes; (3) spheres, which are spherical in shape; and (4) polymer films, which are flat and of varying colours (Rochman et al. 2019;Rosal 2021). The images of some microplastics collected are shown in Fig. S4. Figure 3 shows the distribution of microplastics morphologies in all samples. Synthetic fibers were the dominant microplastics, accounting for about 51% of all microplastics detected. Microplastics in the form of fragments contributed about 42%, while films and spheres contributed only about 4% and 3%, respectively.
The total microplastic concentrations of different samples were examined in relation to their sampling seasons as shown in Fig. 4. While the large variability and small number of replicates precluded more detailed statistical analysis, a few observations are worth discussing. "Outlook upstream" and "Outlook downstream" sites are upstream and downstream of the town of Outlook, SK, respectively. Following the spring sampling period (May 18-23, 2020), the wastewater lagoon was flushed into the South Saskatchewan River (personal communication, Kerry Lowndes). While no microplastics were detected in either April/May or August/September, a relatively high concentration of 10 particles m −3 was observed in the June/July sampling season at Outlook downstream. It is likely that sediments, including potentially associated microplastics, were remobilized and discharged into the river during the flushing of the wastewater lagoon. This would explain the high microplastics concentration detected, as well as the corresponding high concentration of natural fibers detected in the sample as illustrated and discussed in Fig. 2   This suggests that the WWTP in Saskatoon is likely not acting as a significant point sources of microplastics pollution to the river, but that the wastewater lagoon in Outlook can be considered a source during flushing and maintenance. The importance of WWTPs as point sources depends on various factors including population density, different treatment processes employed at various WWTPs, water flow rates, and the percent (%) recovery efficiencies of these WWTPs (Warrack et al. 2017;Vermaire et al. 2017;Talvitie et al. 2017;Bujaczek et al. 2021). For example, WWTPs in Edmonton have been reported to not have a direct influence on the microplastics concentrations on the North Saskatchewan River (Bujaczek et al. 2021), while WWTPs have been reported to be direct point sources of microplastics to Lake Winnipeg and the Ottawa River (Baldwin et al. 2016;Talvitie et al. 2017;Watkins et al. 2019;Bujaczek et al. 2021).

and the accompanying
The mean microplastic concentration in Saskatoon's storm ponds was 6.44 ± 3.6 particles m −3 . Concentrations in Evergreen and Aspen Ridge were similar to those in the river. However, concentrations in the Northeast Swale, a conservation site within Saskatoon, were much higher, with a mean of 12 ± 7.7 particles m −3 , similar to the concentration in Miry Creek. The composition of the microplastics in the Northeast Swale was also distinct from that of other locations (both river and storm ponds), with less than 30% of microplastics consisting of synthetic fibers. For comparison, synthetic fibers accounted for 53% of particles from river samples, and 88% in the Evergreen and Aspen Ridge storm ponds. Polymer fragments accounted for 50% of microplastics in the Northeast Swale, with spheres and films each accounting for 11%. Spheres were not detected at any other sampling site, and only one other sampling site (Outlook upstream, during the August/September sampling period) contained films. The reason for the anomalous microplastic concentration and type is not known.
There are few reports of microplastics concentrations in storm ponds. A mean microplastic concentration of 15,400 particles m −3 was reported in stormwater runoff released into Lake Ontario, suggesting that urban stormwaters contribute significantly to the microplastic pollution of Lake Ontario (Grbić et al. 2020;Shruti et al. 2021). It is likely that microplastic concentrations in storm ponds depend on numerous factors, likely including local population density and land use. It should also be noted that the samples here were taken near the surface of the storm ponds under baseflow conditions and that levels in stormwater runoff and overlaying water in the receiving storm ponds can be expected to be greater under stormflow conditions.

Chemical characterization
Raman spectra were acquired for each particle, and composition was determined by comparison with polymer spectra from the in-house spectral library or the literature (Fig. S5-S12; Table S1-S3). Figure 5 shows the fractional composition of all samples excluding natural fibers. Polyethylene terephthalate (PET) was the dominant polymer, accounting for nearly half of all polymers. Polypropylene accounted for 36%, and polystyrene accounted for 3%. We were unable to definitively characterize the remaining 12% of particles due to scattering or fluorescence from chemical additives masking the Raman signal. Likely sources of PET and PP are textiles, packaging, and various consumer products . Polystyrene (often referred to as Styrofoam in its expanded form) is commonly used to manufacture numerous consumer products ranging from disposable dishes to medical appliances.
We note that the fractional composition shown in Fig. 5 is not necessarily representative of microplastics within the South Saskatchewan River as a whole. For example, PET is denser than water (1.38 g cm −3 ) and is expected to reside primarily near or within the sediment layer (Lenaker et al. 2019). Since sampling in this campaign was performed near the surface, we do not expect to collect most of the PET, even though it is a significant component of microplastics frequently recovered in freshwater samples due to its application in manufacturing of most synthetic textile fibers, fishing nets, and lines (Cheung et al. 2016). The large contribution of PET in our samples is likely due to resuspension of dense microplastics to the subsurface caused by high flow velocity and strong hydrodynamics of the river, especially from rainfall (Talbot and Chang 2022). The actual burden of PET in the South Saskatchewan River is likely much higher than that reported in our study. Similarly, while polystyrene particles were present in our samples, they accounted for only 3% of all microplastics. Polystyrene is very buoyant, and partitions readily to the air-water interface. Sampling this interface (e.g., via skimming techniques) might yield a far greater contribution from polystyrene.
Natural textile fibers accounted for ~ 77% of all particles analyzed. Most (77%) of fibers were dyed; signal from the dyes dwarfed the spectra of the underlying fibers. Figure 6 shows the distribution of dyed and undyed natural fibers. Approximately half of all natural fibers were dyed with Reactive Black 5 dye (RB5). An additional 36% were dyed with Vat Indigo Dye (VIG), and undyed brown cotton and polycotton accounted for 20% of natural fibers. Diarylide Yellow (DY 83) and Copper phthalocyanine (Pigment Blue 15:0) contributed about 1% each to the total natural textile fibers, while an additional 1% of natural fibers were uncharacterized. The Raman spectra of these microplastics and natural fibers, including their peak assignments, are summarized in Figs. S5-S12 in the SI. Figure 7, which summarizes microplastic composition and concentration at each sampling site, highlights the variability in composition across sites. While natural textile fibers accounted for the majority of particles at most locations, they accounted for only 63% and 40% for Miry Creek and Northeast Swale, the two locations with the highest synthetic particle loadings. The relative concentrations of individual polymers were also different at these two sites. At most sites, PET was the dominant polymer, whereas it accounted for less than 11% and 18% of polymers in Miry Creek and Northeast Swale. Beaver Creek was also an outlier in this regard. Although this site had low polymer loadings and microplastics were only detected in the August/September sampling season, all microplastics detected at this location consisted of polypropylene polymer fragments.

Conclusions
While microplastic concentrations and composition have been studied in numerous Canadian waterways, this is, to our knowledge, the first study focused on the South Saskatchewan River. Our results suggest that microparticle concentrations in this river are very low, especially downstream of the Diefenbaker Dam. Purging of the Outlook wastewater lagoon in the spring increased downstream particle loadings, resulting in concentrations five times greater than the mean of all samples. Conversely, the WWTP in Saskatoon did not noticeably affect downstream particle levels.
Most particles in the river were natural textiles, likely from weathering of fibers and microfibers detached from the textiles during laundering (clothing) and potentially from geotextiles. Most of the microplastics identified as fragments and films are suggested to be from degraded plastics commonly used to package and manufacture consumer products. Discharges of surface runoff water into the waterways and land use may have contributed to the varying microplastics concentrations recorded at different sampling sites.
This study also adds to the sparse literature describing microplastic concentrations in storm ponds. Our results suggest that microplastics concentrations and compositions in Saskatoon storm ponds were generally similar to those in the river, although the relatively high levels of polymers in the Northeast Swale -a conservation area -suggests a local, currently unknown source.

Fig. 7
Compositional distribution of particles identified at each sampling site. NTF, natural textile fibers; PET, polyethylene terephthalate; PP, polypropylene; PS, polystyrene; UnC., uncharacterized microplastics Overall, both total particle levels and microplastics levels were low in the South Saskatchewan River and in Saskatoon storm ponds. This work has provided further information to understand microplastic contamination in freshwater ecosystems in Canada and will help further assesses the risk associated with plastic contamination in the environment.