Microplastics and fibers from three areas under different anthropogenic pressures in Douro river
Joana C. Prata a,⁎,1, Veronica Godoy b,1, João P. da Costa a, Monica Calero b, M.A. Martín-Lara b, Armando C. Duarte a, Teresa Rocha-Santos a
H I G H L I G H T S
• Microplastic concentrations in the sampled areas were 231 MP L−1
• Concentrations of 46 F L−1 for total fibers and 6 F L−1 for synthetic fibers
• Wastewater treatment plant and boat dock and maintenance are potential sources.
• High concentrations are a factor of the detection down to a lower size limit of 2 μm.
Abstract
This study aimed at assessing the concentrations of microplastics (2–5000 μm) and fibers (18–5667 μm) in three areas of distinct influences in the Douro river, Porto, Portugal: (i) a countryside area; (ii) a wastewater treatment effluent release zone; and (iii) an area in proximity to a boat dock and maintenance station. Nile Red staining coupled with microscopy allowed the identification of small microplastics (≥2 μm) with a median concentration of the three areas of 231 MP L−1. Most were fragments (69%). Sizes <40 μm were the most abundant (84%). Highest concentrations of microplastics were found near the boat dock/maintenance and lowest in the countryside area. Fibers were mostly natural (non-synthetic, 63%). Highest concentrations of fibers were found in the area influenced by the wastewater effluent, especially of synthetic fibers, and lowest in the countryside area. Concentration of all fibers and synthetic fibers was 46 F L−1 and 6 F L−1, respectively. High concentrations of microplastics and fiber contamination suggest that the wastewater treatment plant effluent and boat dock/maintenance are the likely sources originating hotspot areas.
Keywords:
Microplastics
Douro river
Nile Red
Sources of microplastics
1. Introduction
The widespread use of synthetic materials has led to environmental contamination with small polymeric particles or fibers, known as microplastics (<5 mm). These microplastics originate from fragmentation of larger pieces during use or under environmental conditions (Rocha-Santos and Duarte, 2015). Alternatively, microplastics are intentionally produced for use in cosmetic products and in industrial activities (Cheung and Fok, 2017). Microfibers may also be released to the environment from garments, thus leading to its contamination (Hernandez et al., 2017). The presence of these materials in the environment is a cause of concern due to the potential impact on organisms, but most importantly, due to their persistence, irreversibility, and possibility to affect essential systems, being considered a planet boundary threat (Villarrubia-Gómez et al., 2018).
Rivers have been demonstrated as major pathways for microplastics, revealing high concentrations. These originate from surface runoff, discharges, and leachates, especially from regions with high population densities and a higher number of sources. One potential entry source of microplastics is the release from wastewater treatment plants, which are important sources of contamination due to the release of high volumes of effluents (Ziajahromi et al., 2017). Indeed, release of effluents from wastewater treatment plants are considered point sources of microplastics, often contaminating the surrounding areas near the effluent release (Estahbanati and Fahrenfeld, 2016; Kay et al., 2018; McCormick et al., 2014). So far, these studies on areas surrounding WWTPs used sampling meshes larger than 150 μm, and thus, underestimate the total concentrations by disregarding the assessment of smaller microplastics. Besides the release from WWTPs, another focus of the present study is set on boating activities that can also contribute to the release of microplastics, either during use, from the release of polymeric paint or synthetic fibers from ropes, or during repair, due to the air-blasting with microplastics (microbeads) methods used to clean the paint from boat hulls (Syberg et al., 2015). Indeed, high boat traffic has been related to high concentrations of microplastics in marine waters (Gewert et al., 2017). Thus, both boating activities and maintenance and wastewater treatment plants are potential sources of microplastics to riverine environments, leading to high contamination pressure. In contrast, countryside areas under low anthropogenic pressures generally present lower microplastic concentrations in sediments compared to areas under the influence of high population densities and receptors to high sewage effluent (Horton et al., 2017).
The objective of this work was to estimate and compare the concentrations of small microplastics (<500 μm) and fibers found in a riverine environment near areas of different anthropogenic pressures, including potential contamination sources, namely, (i) an effluent stream of an urban wastewater treatment plant, (ii) a dock and boat maintenance shop, and (iii) a countryside area.
2. Material and methods
2.1. Description of the sampling areas
Sampling was conducted in Douro river, near Porto, Portugal. Douro river originates in Soria province in Spain, crossing Portugal's territory and flowing into the Atlantic Ocean in the west coast of the Iberian Peninsula, being frequently interrupted by hydroelectric dams leading to an irregular freshwater flow (Azevedo et al., 2014). The city of Porto, the second largest city and metropolitan area of Portugal, is in the northern margin of the Douro estuary contributing to high anthropogenic pressures, namely high population (1.7 million inhabitants) and tourism pressures, and important commercial and industrial sectors (Balsas, 2004; PORDATA, 2018). Moreover, the Douro estuary inflow is controlled by an electrical dam at 21.6 km from the mouth with a variable discharge regime, characterized by an average of 505 m3 s−1, and under tidal influence (Azevedo et al., 2014). The Douro river flows ~930 km and its basin covers an area of 97,682 km2 of predominantly granite and metasedimentary geology, crossing plains, mountain, valleys, and reaching a funnel-shaped estuary with a southward sandbank originating from coastal sedimentary drift (Balsinha et al., 2009; Ribeiro et al., 2018). Sediments are characterized mostly by quartz (40%), followed by mica (26%), feldspars (18%), and aggregates, with the average grain sizes in the sampled area of the upper estuary representing sizes 0.5–4.0 mm, corresponding to very coarse to coarse sand (Balsinha et al., 2009). Douro estuary has an average depth of 8 m, a semidiurnal tidal range of 2–4 m, and a freshwater flow of 0.1 m3 s−1 to 13,000 m3 s−1 in winter (Ribeiro et al., 2018). Porto city, at the estuary of the Douro river, has a mean temperature of 14.4 °C and a mean annual rainfall of 1178 mm (Climate Data, 2015).
To understand concentrations of microplastics in the Douro river estuary, as well as potential sources originating from these anthropogenic pressures, eight sampling points in three areas (A, B, C) were identified ranging from 8 to 14 km of the mouth of the river. The choice of points for sample collection was based on potential sources, distance, and availability of accessible locations. The collection of 3 replicates of 1 L samples of surface water in glass bottles for each of the 7 sampling points, in 3 areas under different pressures, was conducted on November 7th 2019, during the high tide, in a rainy day (10 mm) with average temperatures (9–16 °C).
A wastewater treatment plant (WWTP) was chosen in Gramido, serving Gondomar county in a median-sized European city, the Metropolitan Area of Porto, Portugal. Gramido WWTP performs secondary treatment for residential homes in the suburbs surrounding Porto, serving almost 50,000 habitant equivalents with a wastewater flow of 3000 to 11,000 m−3 day−1, transported by pipes of polyethylene, polyvinyl chloride, and stoneware (Barreto, 2011). The effluent from the WWTP is discharged in the last 300 m of the Archeira Stream, which flows into Douro river. Sampling of Gramido WWTP (A) took place in five points in the North margin of the Douro river: Archeira stream, two points upstream (at 85 and 215 m), and two points downstream (at 130 and 300 m) (Table S-1). The Archeira stream sampling point is at about 11 km from the mouth of Douro river.
The second sampling area was in Freixo (B), 2.9 km downstream of Gramido, and at 8 km from the mouth of the river. The two Freixo sampling points were near a boat dock and near a boat maintenance shop. This area could also be previously affected by the nearby effluent release of Freixo WWTP downstream, which was now changed further downstream. A third sampling area was conducted in the countryside, in Marecos village (C) situated in the countryside lacking easily identifiable sources of microplastics. Marecos is located 3.5 km upstream of Gramido, and at 14 km from the mouth of the river. Thus, sampling took place in multiple sample points which were combined and organized into three sampling areas.
2.2. Sampling and sample treatment
Glass bottles were previously washed with HNO3 (reagent grade, Chem-lab, Belgium) followed by distilled water, covered with aluminum foil, and washed with river water at the sampling site. After sampling, bottles were sealed with aluminum foil, caps, and finally parafilm to avoid losses during transportation. Only aluminum foil and glass were in direct contact with the water sample. Physicochemical parameters were read from 500 mL samples, collected solely for this purpose, using a multiparameter probe (HI98194 Multiparameter, HANNA Instruments) (Table S-2, Supplementary Material). Water samples of 1 L, in three replicates, were filtered in 0.7 μm pore glass microfiber filters (Whatman GF/F, USA) in a glass filtration system previously washed with HNO3 and distilled water. After filtration, 5 mL of 30% H2O2 (Labkem, Spain) and 5 mL of 0.05 M Fe(II) (Labkem, Spain) of previously filtered solutions were added to the filtration system and left to react for 15 min at room temperature (20 °C) to remove biogenic organic matter, followed by filtration and washing with Elix® water (Prata et al., 2020a). Then, a few drops (0.5–1 mL) of 0.01 mg mL−1 of Nile Red in ethanol (Sigma-Aldrich, Germany) were added to cover the filter and left to react for 5 min, followed by washing with Elix® water (Prata et al., 2020b). Filters were kept in closed clean glass Petri dishes, kept inside a cardboard box, and dried over the counter for a week. All sample preparation treatments followed a previously tested and published methods (Prata et al., 2020a; Prata et al., 2020b).
2.3. Identification of microplastics and microfibers
Microplastics stained with Nile Red present a yellow to red fluorescence (Prata et al., 2019). Smaller fluorescent microplastics were identified by optical microscopy. Photographs with a digital camera (Canon 1200D) coupled to an orange filter (Standard ProMaster® Orange Filter) attached to an optical microscope (Olympus BX41) allowed the identification of fluorescent particles down to 2 μm (scale: 1.789 px μm−1). This observation was conducted using the 10× objective which allowed side illumination with the 470 nm lantern (SPEX Forensic, USA). Since it was impossible to correctly observe all the filter area under this amplification, the full length of the filter was observed by only moving the yaxis. The observed width was the image size under 10× objective, of 2897 μm, while the length of filter containing sample was approximately of 36.2 mm (in a total filter length of 47 mm). Thus, an area of 104,9 mm2 was observed in a total filter area of 1029 mm2 containing sample, allowing a rough estimation of the number of smaller particles present in the sample. The use of the 10% area in the estimation, while contributing to some uncertainty, was a compromise in the identification of such small particles (≥2 μm). Due to the presence of particles in different levels in the z-axis under microscopy, these particles were quantified manually in ImageJ, which was used for measuring larger and smaller dimensions (Ferret, MinFerret, respectively) and qualifying as fragments or spheres (including circularity), instead of using automated quantification (e.g., MP-VAT). Due to the low number of fibers identified under this method, resulting from the lack of fluorescence of individualized fibers under Nile Red staining, these fibers were not considered. Since microfibers are not consistently fluorescent with Nile Red, possibly due to the refraction of fluorescent emissions on their surface (Prata et al., 2019), these were quantified manually in the whole filter under visible light under the digital stereomicroscope (Leica DMS300), counting these fibers through the glass without opening the Petri dish, and measuring the obtained images in ImageJ. These fibers were classified as natural or synthetic following a diagram based on surface morphology, namely more regular surface of synthetic fibers (Prata et al., 2020c). When fibers could not be confidently characterized as synthetic or natural, they were labeled as unidentified.
2.4. Control of cross contamination
To avoid contamination, all procedures were conducted in a laminar flow hood in a clean room, with limited access and after thorough cleaning. Only glass or metal materials were used. All glass materials were previously washed with 10% HNO3 (reagent grade, Chem-lab, Belgium) and distilled water and covered with aluminum foil. The laminar flow hood was previously cleaned with 70% ethanol in a paper towel. Glass microfiber filters were burnt out at 470 °C for 3 h to remove potential contaminations on their surface. Glass petri dishes were cleaned with a jet of N2 to remove remaining fibers after washing and drying. The filtration system was dipped in 30% HNO3 followed by distilled water between samples to avoid cross contamination. Cotton lab coats were used throughout the procedure. Procedural blanks, subjected to the same procedures as samples, and a filter in an open petri dish in the laminar flow hood allowed to control for airborne and cross contamination.
2.5. Data analysis
Data was recorded in Microsoft 365 Excel and non-parametric statistical analyses were conducted in IBM SPSS Statistical Package version 26, considering an α = 0.05. Statistical analysis consisted in Likelihood ratio (LR), Spearman correlation, and Kruskal-Wallis (H), followed by Pairwise Comparisons (Supplementary Material). Analyses were performed without subtracting or considering blank contamination, since it was limited. Concentrations were expressed as microplastics per liter (MP L−1) and fibers per liter (F L−1). Reports include concentration for each sample point, for the area (median of its samples), and for Douro river (median of all samples). All results are expressed as medians, unless stated otherwise. Due to the low number of particles found on procedural blanks, these were not subtracted from the reported concentrations.
3. Results
3.1. Cross contamination with microplastics and microfibers in controls
In addition to strict contamination control procedures, including the use of laminar flow hood in a clean room, cross contamination was controlled through an open filter and procedural blanks. For small microplastics, only a total of 3 particles were found in the investigated area of 10% of the whole filter of the 2 procedural blanks (1.5 particles per blank). No particles were found in open air filters. Particles in procedural blanks had median smaller and larger dimensions of 16.4 and 29.7 μm and were mostly comprised of fragments with a circularity of 0.795. In the case of fibers, only a total of 48 fibers were identified in the controls and open filters, with a mean of 16 fiber per filter. These fibers were mostly white (41.7%), with an average size of 348 μm, and of natural origin (50.0%), likely originating from the deposition of airborne fibers released from paper towels and white cotton lab coats. Microplastics contamination corresponded to 0.5% and fiber contamination to 3.7% of the total number of analyzed particles.
3.2. Small microplastics in Douro river
3.2.1. Concentration of small microplastics in three sites sampled at Douro river
Small microplastics were quantified under the microscope (×10) by observing an area equivalent to 10% of the filter, covering edges and central sections, then extrapolated to the overall area of the filter which corresponded to a liter of processed river water (Fig. 1). In total, 665 fluorescent particles ≥2 μm, in the 10% areas of filters, were manually classified regarding the largest and smallest dimensions, as well as shape and circularity. For the three sampled areas in Douro river, the median concentration was 231 MP L−1. The lowest concentration was found for Marecos countryside (<1 MP L−1), followed by Gramido effluent area (265 MP L−1), and Freixo boating area (334 MP L−1). Significant differences were found between sampling areas (H2 = 8.512, p = 0.014), namely Marecos presenting significantly lower concentrations than Freixo (p = 0.012) and Gramido (p = 0.039). Differences were also found between sampling points (H7 = 16.515, p = 0.021), possibly also caused by differences from Marecos samples, but lost in pairwise comparisons due to Bonferroni correction. Both samples in the effluent stream and boat dock presented the highest concentrations of small microplastics, of 350 MP L−1.
3.2.2. Size of small microplastics in Douro river
The overall largest dimension of particles varied from 2.2 to 483.1 μm. For Douro river, the median smallest and largest dimensions of small microplastics were of 12.8 and 18.6 μm, respectively. Between areas, no differences were found for larger (H2 = 0.457, p = 0.796) and smaller (H2 = 1.973, p = 0.373) particle dimensions. Between samples, significant differences were found for both the larger (H7 = 65.365, p < 0.001) and smaller dimensions (H7 = 60.649, p < 0.001). The most upstream area in Gramido presented the smallest dimensions in all samples, of median smallest and largest dimensions of 7.7 and 11.9 μm, being significantly different to all locations (p ≤ 0.007), except the most downstream location in Gramido and Marecos countryside. When comparing larger dimensions, the boat dock in Freixo presented smaller particles than the first upstream location in Gramido (p = 0.017). Moreover, 83.9% of particle's had largest dimensions <40 μm, highlighting the abundance of small microplastics. Only 4 particles (0.6%) had largest dimensions >200 μm.
3.2.3. Morphology of small microplastics in Douro river
Fragments were the most common shape in Douro river (69%), translated in a corresponding circularity of 0.839, with the remaining particles being spheres (exemplified in Fig. 2). Different areas did not present significant differences in shape (LR2 = 2.157, p = 0.340), but different sampling points did (LR7 = 21.640, p = 0.003). This is translated in the abundance of fragments in Marecos (100%) and the first upstream area in Gramido (85%) compared to the remaining areas (<77%). Circularity did not present significant differences between areas (H2 = 0.352, p = 0.839) or point (H7 = 11.841, p = 0.106). Nonetheless, shapes varied from long shaped fragments to perfectly circular spheres (circularity: 0.043–1.000). The largest dimension was significantly different between shapes (H1 = 30.098, p < 0.001), with fragments presenting median largest dimensions of 21.3 μm (2.8–483.1 μm) and spheres of 15.7 μm (2.2–87.8 μm). No differences were found between shapes for smallest dimensions (H1 = 0.561, p = 0.454).
3.3. Fibers in Douro river
3.3.1. Concentration of fibers in Douro river
A total of 1302 fibers were identified and measured in all river samples with a median concentration of 46 F L−1 and of 6 synthetic F L−1 in the Douro river. The total concentrations of fibers found was 53 F L−1 in Gramido, the effluent release area, 45 F L−1 in Freixo, the boat dock and maintenance shop, and finally 26 F L−1 in Marecos, the countryside. Between areas, the number of fibers was statistically different (H2 = 7.297, p = 0.026), namely between Marecos and Gramido, (p = 0.051), a non-significant pairwise comparison after Bonferroni correction. There were no significant differences in fiber concentration between points (H7 = 11.258, p = 0.0129). Despite the different concentrations of synthetic fibers between Marecos (2 fibers L−1), Gramido (7 fibers L−1), and Freixo (9.5 fibers L−1), they were not statistically different (H2 = 5.847, p = 0.054). Conversely, concentrations of synthetic fibers were significant different between sampling points (H7 = 17.118, p = 0.017), likely due to differences between Marecos countryside and Freixo boat dock, as suggested by the post-hoc Pairwise comparison but lost after Bonferroni correction.
3.3.2. Sizes of fibers in Douro river
The overall length of the observed fibers under the stereomicroscope varied from 18 to 5667 μm, with an average and a median of 522 and 315 μm, respectively (Fig. 3). Fibers seemed to increase in size downstream, with medians of 243 μm for Marecos, 324 μm for Gramido and 324 μm for Freixo. However, this relationship was non-significant (H2 = 1.141, p = 0.565). On the contrary, there were significant differences in fiber sizes between sampling point (H7 = 48.388, p < 0.001). Marecos countryside is significantly different from both upstream locations (p < 0.001) and the most downstream location (p = 0.001) in Gramido. In Gramido, sizes found in the first upstream location are also significantly smaller than fibers on the effluent stream (p < 0.001).
3.3.3. Morphology of fibers in Douro river
The nature of fibers was mostly natural (62.6%), followed by unidentified (23.0%), with only 14.4% of fibers being suspected of being synthetic (Fig. 4). For fiber nature, there were significant differences for area (LR6 = 41.694, p < 0.001) and for points (LR21 = 137.049, p < 0.001). The abundance of synthetic fibers seemed to increase downstream, corresponding to 5.6% of fibers in Marecos, 13.1% in Gramido, and 22.3% in Freixo. In Gramido, the proportions for synthetic fibers varied from 3.3 to 28.0%, with the highest proportion found in the effluent stream.
The most frequent color of fibers observed was brown (41.6%), mostly due to the covering of fibers with residues from the hydrogen peroxide reaction, followed by white (27.5%), black (23.5%), and other including blue, red, and green. For color, there were also significant differences for area (LR12 = 63.429, p < 0.001) and points (LR42 = 263.390, p < 0.001). Indeed, the most abundant fiber colors changed between areas, with Marecos presenting mostly black (36.0%), and Gramido brown (44.9%), and Freixo equally white and brown fibers (each 34.3%). The effluent stream in Gramido mostly presented brown (37.6%) and white (34.9%) fibers.
When comparing between the variables there were significant relationships between sizes and color (H6 = 533.539, p < 0.001) and origin (H3 = 95.276, p < 0.001) as well as between color and nature (LR18 = 292.818, p < 0.001). Black fibers (80 μm) were significantly smaller than all other colors (p ≤ 0.002) except green, likely having a natural origin based on their surface characteristics. White fibers were significantly larger than brown fibers (554 vs. 363 μm, p < 0.001). Regarding nature, synthetic fibers (602 μm) were significantly larger than natural (289 μm, p < 0.001) and unidentified fibers (261 μm, p < 0.001), possibly by being less prone to fragmentation. These synthetic fibers were mostly of white (59.9%) and blue colors (20.3%), while natural fibers were mostly brown (53.2%) and black (23.8%) and unknown fibers of black (33.3%) and brown (29.3%) colors.
4. Discussion
4.1. Concentrations and distribution of microplastics and fibers in Douro river estuary
Median concentrations determined in the three sampling locations in the Douro river were of 231 MP L−1 for microplastics and 46 F L−1 for fibers (Table 1). Despite these high concentrations, blanks presented low contamination, which was mostly comprised by airborne deposition of white natural fibers, likely originating from paper towels, which should be avoided in the cleaning of the laminar flow hood, and cotton lab coats, for which rolled up sleeves would be preferred whenever possible. Thus, reported concentrations are unlikely to originate from sample contamination. Concentrations in Marecos countryside were different from both remaining areas regarding microplastic concentrations and likely from Gramido effluent release when considering fiber concentrations. The abundance of microplastics and fibers increased towards the estuary, with lowest concentration in the countryside and highest near potential sources, namely the wastewater effluent release and boat dock and maintenance shop. Similarly, the concentration of synthetic fibers increased downstream, with the highest concentration found in the effluent stream (28.0%). An increase in concentrations of microplastics and fibers towards the estuary was also found in the Magdalena river in Colombia (Silva and Nanny, 2020). Fiber length also increased downstream, likely related to their higher synthetic nature which was correlated to larger sizes. Nonetheless, most fibers were identified as natural (62.8%). In addition to the three sampling areas in the current work, sampling points were also compared individually, allowing for a more detailed evaluation inside each area. Some variation between these points is expected since they may be more or less influence by the potential source of microplastics. Nonetheless, regarding concentrations, differences were only found between sampling points in different areas (e.g., the countryside and others). Other differences were found for particle characteristics, whichlikely dependent on factors involved in the accumulation and distribution of particles.
In terms of microplastics, most were comprised of fragments (69%) while sizes varied from 2.2 to 483.1 μm, varying with sampling point likely influenced by dynamic factors involving the river water and sediment deposition. However, no relationship was found between total dissolved solids and concentrations of microplastics (rS = −0.299, p = 0.471; Table S-2, Supplementary Material), total fibers (rS = −0.190, p = 0.651), or synthetic fibers (rS = 0.012, p = 0.977). Spherical microplastics (15.7 μm) had smaller dimensions than fragments (21.3 μm). While most spherical microplastics were well above the ≥2 μm, perception of very small particles could lead to their misidentification. Nonetheless, it is most likely that these spherical particles originate from the presence of microbeads in Douro river. Microbeads could originate from wastewater, as one of their popular uses is in cosmetics, presenting previously assessed sizes of 5–2188 μm (Godoy et al., 2019). Alternatively, they could originate from air-blasting methods used in the cleaning of boat hulls, having sizes varying from 12 to 80 μm (Hanzel, 2000). Both are plausible sources considering the presence of an effluent stream and a boat maintenance shop in Douro river.
Wastewater treatment plants have been identified as sources of microplastics in river catchments, originating from textile fibers, cosmetics, and consumer products, leading to greater contamination of downstream environments (Kay et al., 2018). Boat docks, marinas, and shipyards are known sources of microplastics, either generated during the cleaning and natural abrasion of boat hulls or the release of paint flakes during use (Magnusson et al., 2016). The irregular shape of paint flakes classifies them as fragments, which were abundant near the boat dock. In the future, chemical characterization of the particles could aid in the identification of sources, such as paints (e.g., polyacrylate; Gaylarde et al., 2021). In Douro river, highest concentrations of microplastics were found in the proximity of a wastewater effluent and a boat dock and shipyard, supporting their role as potential sources of these anthropogenic particles. These areas were found to be microplastic hotspots (i.e., areas of severe contamination) requiring further assessment of microplastics and fibers' adverse effects in organisms and the environment. Ways to reduce the release of microplastics and fibers from sources while conserving economic activities are urgently needed.
Moreover, Douro river has an irregular flow, relatively low discharge (505 m3 s−1) and is influenced by tides (Azevedo et al., 2014), which may influence the accumulation of microplastics in surface waters. Areas of slow water movement, in addition to multiple potential of sources (e.g., WWTP, touristic activity), may also contribute to the accumulation of these synthetic particles in the Douro estuary (Rodrigues et al., 2018). Moreover, the immediate downstream vicinity from Freixo boating area is comprised not by very coarse to coarse sand but by finer sediments of coarse sand to very fine sand as well as higher values of mud (silt, clay) originating from minor creeks and sewage discharges (Balsinha et al., 2009). Differences in estuary sediments could lead to different retention and release of microplastics, facilitating or not the creation of hotspots. Therefore, future works should address what characteristics (e.g., sediment composition, geomorphology) could contribute to the distribution of microplastics in general, and more specifically in the Douro river at the time of sampling.
4.2. Methodological challenges in the identification of small microplastics and fibers
The present study shows one of the highest concentrations of microplastics found so far in riverine environments and suggests a higher degree of underestimation, especially when considering that 83.9% of microplastics were <40 μm, lower that most pore or mesh sizes (e.g., 300 μm). To allow the assessment of such small sizes and controlled conditions to avoid contamination of samples, compromises had to be made, namely by assessing only 10% of the filter for fluorescent particles and using grab samples with replicates of 1 L of surface river water. The use of a lane corresponding to 10% of the filter area allowed observation of fluorescent particles ≥2 μm, under the common optical microscope with external light source and orange filter, without the risk of recounting the same areas. The definition of smaller areas for the detection of small particles under magnification is not uncommon in microplastics analytics. For instance, Schymanski et al. (2018) scanned an area of 4.4 × 4.4 mm and Oßmann et al. (2018) scanned five areas of 1 × 1 mm, with both works covering ~4% of the total filter area when analyzing microplastic from drinking water using microRaman spectroscopy. The lane in the present work covered distinct areas of the filter, from the edges to the center, accounting for the variability in particles distribution. Nonetheless, the resulting estimation could account for some variability in the results, which should be further addressed in future studies.
Grab samples of 1 L were used to assess the presence of small microplastics. A previous study tested laboratory spikes and environmental samples finding that 1 L samples (with replicates) produces good results regarding interquartile range, accuracy (<5%), coefficient of variation (<20%), and recovery rates (<10%) while being feasible in terms of effort and the amount of organic and mineral matter clogging the filter (Prata et al., 2020a). The use of smaller volumes for the accurate quantification of microplastics is not surprising when accounting for the abundance of smaller sizes (i.e., 83.9% under 40 μm). For instance, 10-fold greater concentrations of microplastics in seawater were obtained when using 100 μm meshes than by using 500 μm meshes (Lindeque et al., 2020). Covernton et al. (2019) found that concentrations 8.5 times higher were obtained by filtering a 1 L grab sample in 8 μm pore membrane than by passing 10 L through a 63 μm mesh in situ. Green et al. (2018) found concentration ~3 orders of magnitude higher when filtering 1 L grab samples in 0.45 μm pore membranes than when using bongo nets (>500 μm), manta nets (>300 μm), and plankton nets (>200 and 400 μm). Thus, 1 L grab samples can be useful in the detection of smaller microplastic, which is required for a more comprehensive overview of microplastics contamination. The need for a procedure capable of addressing smaller particles sizes must also be considered in the harmonization of methods.
The present study bases microplastic identification on the use of Nile Red after biogenic organic matter removal instead of visual inspection, which often produces numerous false positives and false negatives (Hidalgo-Ruz et al., 2012). Microplastics are stained by this lipophilic dye and fluorescence can be observed under specific excitation wavelengths (Erni-Cassola et al., 2017). Staining is a cost-effective method of improving the accuracy of microplastics identification, allowing the identification of smaller particles under a common optical microscope by coupling with an external light source and orange filter. However, the presence of particles on the z-axis under microscopy compromises the use of automated quantification scripts, requiring manual measurement of each particle (e.g., in ImageJ). This effort combined with the abundance of smaller particles also contributed to the characterization of 10% of the filter area. Staining following proper sample preparation (e.g., biogenic organic matter removal, in this case) produces accurate identification of plastics. To confirm the accuracy of Nile Red staining, Erni-Cassola et al. (2017) subjected sediment and water samples after biogenic organic matter removal and staining to Raman spectroscopy, revealing that all scanned fluorescent particles (n = 37) were synthetic, and all scanned non-fluorescent particles (n = 23) were non-synthetic. Similarly, Mason et al. (2018) subjected 1000 fluorescent particles >100 μm from bottled water to identification using Fourier Transform Infrared spectroscopy (FTIR), finding that all had a synthetic origin, either as a plastics polymer, additive, or binder. Similarly, Tamming et al. (2019) subjected 277 fluorescent particles, from water samples obtained from a lake after biogenic organic matter removal, to microRaman spectroscopy, from which 273 could be confidently identified as plastics and 4 could not be identified. The use of Nile Red in water samples, when preceded by proper biogenic organic matter removal, can thus be deemed accurate. Despite not being conducted in the present study, the use of spectroscopic methods is recommended in future studies since it can provide information beyond the simple distinction between biogenic and synthetic materials, such as polymer identification.
The identification of fibers using Nile Red is not consistent, either due to the refraction of fluorescent emission on their surface (Prata et al., 2019) or the high crystallinity that compromises staining (Shim et al., 2016). Therefore, fibers were identified under the stereomicroscope following a strict identification key. When no confident identification could be made based on surface characteristics, fibers were classified as unidentified. This often resulted from the small sizes of fibers or the presence of residues from biogenic organic matter removal which cover the fibers. For instance, the presence of brown fibers in Gramido may results from residues of large amounts of biogenic organic matter from the wastewater effluent, while the white fibers observed in Freixo may originate from boats. These residues also hinder efforts of distinguishing natural textile fibers from fibrous biogenic organic matter, all being classified as natural fibers. These methodological difficulties could be overcome in the future by coupling the present visual method with chemical characterization by spectroscopic methods, which would allow precise identification of fiber nature.
4.3. Contextualization of microplastics and fibers in Douro river estuary
Comparison of results between studies must account for the previously described variations in methodology. Following the same procedures but quantifying fluorescent particles >50 μm, concentrations of 18 MP L−1 were found in the Aveiro Lagoon, Portugal (Prata et al., 2020a). Despite differences in particle size (2 vs. 50 μm), the concentrations found can be considered in the same range as the current work. Conversely, the previous work conducted in the Douro river estuary found 1.7 × 10−4 particles L−1 using a 500 μm mesh (Rodrigues et al., 2018). Surface water samples of 2 L from channels in the City of Amsterdam contained concentrations of 48–187 particles L−1, considering both fibers and microplastics >10 μm based on visual identification and confirmation of 6% of particles by Fourier-Transform Infrared spectroscopy, with 0–83% of particles being <300 μm (Leslie et al., 2017). This study also found microplastics in concentrations of 11–105 particles g−1 (dry weight) in biota (e.g., Gammarus spp., Littorina littorea, Mytilus edulis, Crassostrea gigas) and 9–91 MP L−1 in wastewater effluents, for which fractions <300 μm varied from 0 to 100%. In the current study in the Douro river estuary, concentrations of 231 MP L−1 were found, with 83.9% of particles being 2–40 μm, and only 0.6% ≥200 μm. Results from the current study are closer to the concentration found in channels in the City of Amsterdam than previously published results on the Douro estuary due to critical differences in methodology, with a high relevance of particle sizes being identified. Moreover, Leslie et al. (2017) quantified particles by light microscopy, which makes the detection of smaller particles more difficult than when using a fluorescent probe such as Nile Red.
In the Douro estuary, fibers of 315 μm median length were found in concentrations of 46 F L−1. This is in accordance with the concentrations of 22 to 251 F L−1 reported in the Saigon river, Vietnam, were sizes 40–300 μm represented 82% of fibers (Strady et al., 2020) and of 6 to 398 F L−1 in the Marne river (a tributary of the Seine), Paris, France, with mean sizes of 973 μm (Dris et al., 2018). Just as with microplastics, comparisons must be made accounting for methodological differences. These studies conducted sampling of fibers using 80 μm meshes, revealing higher fiber length was found in Paris, while the current study used grab samplings, with 18 μm being the smallest fiber length identified. Despite the current study being able to find smaller fibers, the Saigon river presents fibers of smaller sizes likely due to the large influence of textile and garment industry, while the Douro river is mostly impacted by household effluents and boating activities. Moreover, in Douro river, microplastic concentrations were not correlated to total fiber concentration (rS = −0.065, p = 0.762) but were correlated to synthetic fiber concentrations (rS = −0.602, p = 0.002). This suggests that microplastics and synthetic fibers share the same sources and/or factors affecting their distribution in Douro river, which is an interesting topic to be addressed in future works.
5. Conclusions
Microplastics found in three sampled sites in the Douro river estuaries presented median concentrations of 231 MP L−1, with concentrations increasing towards the estuary. Lowest concentrations were found in the countryside and highest near the wastewater effluent release and boat dock and maintenance shop. Synthetic fibers, which comprised only 14% of total fibers in the river, followed a similar pattern, with higher concentration found in wastewater effluent stream (28%). By coupling Nile Red staining with microscopy, particles ≥2 μm could be counted revealing that 84% of microplastics were <40 μm, highlighting the need for further development of harmonized methods capable of detecting these smaller fractions of particles. These methods could be used to further to address the contributions of effluent release and boating activities to hotspot areas, hopefully helping to find solutions that could couple economic activities with environmental protection.
References
Azevedo, I.C., Bordalo, A.A., Duarte, P., 2014. Influence of freshwater inflow variability on the Douro estuary primary productivity: a modelling study. Ecol. Model. 272, 1–15. https://doi.org/10.1016/j.ecolmodel.2013.09.010.
Balsas, C.J.L., 2004. City centre regeneration in the context of the 2001 european capital of culture in Porto, Portugal. Local Econ. 19 (4), 396–410. https://doi.org/10.1080/ 0269094042000286873.
Balsinha, M.J., Santos, A.I., Alves, A.M.C., Oliveira, A.T.C., 2009. Textural composition of sediments from Minho and Douro estuaries (Portugal) and its relation with hydrodynamics. J. Coast. Res. 56, 1330–1334. https://www.jstor.org/stable/25738005.
Barreto, S.M.C.C., 2011. Impacto da Infliltração e Afluência de Águas Pluviais na Rede de Saneamento. https://recipp.ipp.pt/bitstream/10400.22/7682/1/DM_SusanaBarreto_ 2011_MEC.pdf.
Cheung, P.K., Fok, L., 2017. Characterisation of plastic microbeads in facial scrubs and their estimated emissions in Mainland China. Water Res. 122, 53–61. https://doi.org/ 10.1016/j.watres.2017.05.053.
Climate Data, 2015. https://en.climate-data.org/europa/portugal/porto/porto-161/. (Accessed January 2021).
Covernton, G.A., Pearce, C.M., Gurney-Smith, H.J., Chastain, S.G., Ross, P.S., Dower, J.F., Dudas, S.E., 2019. Size and shape matter: a preliminary analysis of microplastic sampling technique in seawater studies with implications for ecological risk assessment. Sci. Total Environ. 667, 124–132. https://doi.org/10.1016/j.scitotenv.2019.02.346.
Dris, R., Gasperi, J., Tassin, B., 2018. Sources and fate of microplastics in urban areas: a focus on Paris Megacity. In: Wagner, M., Lambert, S. (Eds.), Freshwater Microplastics. The Handbook of Environmental Chemistry. vol 58. Springer, Cham. https://doi.org/ 10.1007/978-3-319-61615-5_4.
Erni-Cassola, G., Gibson, M.I., Thompson, R.C., Christie-Oleza, J.A., 2017. Lost, but found with Nile Red: a novel method for detecting and quantifying small microplastics (1 mm to 20 μm) in environmental samples. Environ. Sci. Technol. 51 (23),13641–13648. https://doi.org/10.1021/acs.est.7b04512.
Estahbanati, S., Fahrenfeld, N.L., 2016. Influence of wastewater treatment plant discharges on microplastic concentrations in surface waters. Chemosphere 162, 277–284. https://doi.org/10.1016/j.chemosphere.2016.07.083.
Gaylarde, C.C., Neto, J.A.B., da Fonseca, E.M., 2021. Paint fragments as polluting microplastics: a brief review. Mar. Pollut. Bull. 162, 111847. https://doi.org/10.1016/ j.marpolbul.2020.111847.
Gewert, B., Ogonowski, M., Barth, A., MacLeod, M., 2017. Abundance and composition of near surface microplastics and plastic debris in the Stockholm Archipelago, Baltic Sea. Mar. Pollut. Bull. 120 (1–2), 292–302. https://doi.org/10.1016/j. marpolbul.2017.04.062.
Godoy, V., Martín-Lara, M.A., Calero, M., Blázquez, G., 2019. Physical-chemical characterization of microplastics present in some exfoliating products from Spain. Mar. Pollut. Bull. 139, 91–99. https://doi.org/10.1016/j.marpolbul.2018.12.026.
Green, D.S., Kregting, L., Boots, B., Blockley, D.J., Brickle, P., da Costa, M., Crowley, Q., 2018. A comparison of sampling methods for seawater microplastics and a first report of the microplastics litter in coastal water of Ascension and Falkland Islands. Mar. Pollut. Bull. 137, 695–701. https://doi.org/10.1016/j.marpolbul.2018.11.004.
Hanzel, D., 2000. Abrasive blasting systems. Metal Finishing https://www.shotpeener. com/library/pdf/1997002.pdf. (Accessed 20 May 2020).
Hernandez, E., Nowack, B., Miltrano, D.M., 2017. Polyester textiles as a source of microplastics from households: a mechanistic study to understand microfiber release during washing. Environ. Sci. Technol. 51 (12), 7036–7046. https://doi.org/10.1021/ acs.est.7b01750.
Hidalgo-Ruz, V., Gutow, L., Thompson, R.C., Thiel, M., 2012. Microplastics in the marine environment: a review of the methods used for identification and quantification. Environ. Sci. Technol. 46 (6), 3060–3075. https://doi.org/10.1021/es2031505.
Horton, A.A., Svendsen, C., Williams, R.J., Spurgeon, D.J., Lahive, E., 2017. Large microplastic particles in sediments of tributaries of the River Thames, UK – abundance, sources and methods for effective quantification. Mar. Pollut. Bull. 14 (1), 218–226. https://doi.org/10.1016/j.marpolbul.2016.09.004.
Kay, P., Hiscoe, R., Moberley, I., Bajic, L., McKenna, N., 2018. Wastewater treatment plants as a source of microplastics in river catchments. Environ. Sci. Pollut. Res. 25 (20), 20264–20267. https://doi.org/10.1007/s11356-018-2070-7.
Leslie, H., Brandsma, S.H., van Velzen, M.J.M., Vethaak, A.D., 2017. Microplastics en route: field measurements in the Dutch river delta and Amsterdam canals, wastewater treatment plants, North Sea sediments and biota. Environ. Int. 101, 133–142. https://doi.org/10.1016/j.envint.2017.01.018.
Lindeque, P.K., Cole, M., Coppock, R.L., Lewis, C.N., Miller, R.Z., Watts, A.J.R., Wilson-McNeal, A., Wright, S.L., Galloway, T.S., 2020. Are we underestimating microplastic abudance in the marine environment? A comparison of microplastic capture with nets of diferente mesh-size. Environ. Pollut. https://doi.org/10.1016/j.envpol.2020.114721.
Magnusson, K., Eliasson, K., Frane, A., Haikonen, K., Hultén, J., Olshammar, M., Stadmark, J., Voisin, A., 2016. Swedish sources and pathways of microplastics to the marine environment. https://www.ivl.se/webdav/files/Rapporter/C183.pdf. (Accessed 20 May 2020).
Mason, S.A., Welch, V.G., Neratko, J., 2018. Synthetic polymer contamination in bottled water. Front. Chem. 6, 407. https://doi.org/10.3389/fchem.2018.00407.
McCormick, A., Hollein, T.J., Mason, S.A., Schluep, J., Kelly, J.J., 2014. Microplastics is an abundant and distinct microbial habitat in an urban river. Environ. Sci. Technol. 48 (20), 11863–11871. https://doi.org/10.1021/es503610r.
Oßmann, B.E., Sarau, G., Hotmannspötter, H., Pischetsrieder, M., Christiansen, S.H., Dicke, W., 2018. Small-sized microplastics and pigmented particles in bottled mineral water. Water Res. 141, 307–316. https://doi.org/10.1016/j.watres.2018.05.027.
PORDATA, 2018. População residente [resident population]. https://www.pordata.pt/ Municipios/Popula%C3%A7%C3%A3o+residente+total+e+por+grandes+grupos+et %C3%A1rios-390. (Accessed 20 May 2020).
Prata, J.C., Reis, V., Matos, J.T.V., da Costa, J.P., Duarte, A.C., Rocha-Santos, T., 2019. A new approach for routine quantification of microplastics using Nile Red and automated software (MP-VAT). Sci. Total Environ. 690, 1277–1283. https://doi.org/10.1016/j. scitotenv.2019.07.060.
Prata, J.C., Manana, M.J., da Costa, J.P., Duarte, A.C., Rocha-Santos, T., 2020a. What is the minimum volume of sample to find small microplastics: laboratory experiments
and sampling of Aveiro Lagoon and Vouga River, Portugal. Water 12 (4), 1219. https://doi.org/10.3390/w12041219.
Prata, J.C., Alves, J.R., da Costa, J.P., Duarte, A.C., Rocha-Santos, T., 2020b. Major factors influencing the quantification of Nile Red stained microplastics and improved automatic quantification (MP-VAT 2.0). Sci. Total Environ. 719, 137498. https://doi.org/ 10.1016/j.scitotenv.2020.137498.
Prata, J.C., Castro, J.L., da Costa, J.P., Duarte, A.C., Cerqueira, M., Rocha-Santos, T., 2020c. An easy method for processing and identification of natural and synthetic microfibers and microplastics in indoor and outdoor air. MethodsX 7, 100762. https://doi.org/ 10.1016/j.mex.2019.11.032.
Ribeiro, C., Couto, C., Ribeiro, A.R., Maia, A.S., Santos, M., Tiritan, M.E., Pinto, E., Almeida, A.A., 2018. Distribution and environmental assessment of trace elements contamination of water, sediments and flora from Douro River estuary, Portugal. Sci. Total Environ. 639, 1381–1393. https://doi.org/10.1016/j.scitotenv.2018.05.234.
Rocha-Santos, R., Duarte, A.C., 2015. A critical overview of the analytical approaches to the occurrence, the fate and the behavior of microplastics in the environment. Trends Anal. Chem. 65, 47–53. https://doi.org/10.1016/j.trac.2014.10.011.
Rodrigues, M.O., Abrantes, N., Gonçalves, F.J.M., Nogueira, H., Marques, J.C., Gonçalves, A.M.M., 2018. Spatial and temporal distribution of microplastics in water and sediments of a freshwater system (Antuã River, Portugal). Sci. Total Environ. 633, 1549–1559. https://doi.org/10.1016/j.scitotenv.2018.03.233.
Schymanski, D., Goldbeck, C., Humpf, H.-U., Fürst, P., 2018. Analysis of microplastics in water by micro-Raman spectroscopy: release of plastic particles from different packaging into mineral water. Water Res. 129, 154–162. https://doi.org/10.1016/j. watres.2017.11.011.
Shim, W.J., Song, Y.K., Hong, S.H., Jang, M., 2016. Identification TBOPP and quantification of microplastics using Nile Red staining. Mar. Pollut. Bull. 113 (1–2), 469–476. https:// doi.org/10.1016/j.marpolbul.2016.10.049.
Silva, P.M., Nanny, M.A., 2020. Impact of microplastic fibers from the degradation of nonwoven synthetic textiles to the Magdalena river water column and river sediments by the city of Neiva, Huila (Colombia). Water 12 (4), 1210. https://doi.org/10.3390/ w12041210.
Strady, E., Dang, T.H., Dao, T.D., Dinh, H.N., Do, T.T.D., Duong, T.N., Duong, T.T., Hoang, D.A., Kieu-Le, T.C., Le, T.P.Q., Mai, H., Trinh, D.M., Nguyen, Q.H., Tran-Nguyen, Q.A., Tran, Q.V., Truong, T.N.S., Chu, V.G., Vo, V.C., 2020. Baseline assessment of microplastic concentrations in marine and freshwater environments of a developing Southeast Asian country, Viet Nam. Mar. Pollut. Bull., 11180 https://doi. org/10.1016/j.marpolbul.2020.111870.
Syberg, K., Khan, F.R., Selck, H., Palmqvist, A., Banta, G.T., Daley, J., Sano, L., Duhaume, M.B., 2015. Microplastics: addressing ecological risk through lessons learned. Environ. Toxicol. Chem. 34 (5), 945–953. https://doi.org/10.1002/etc.2914.
Tamming, M., Stoewer, S.-C., Fischer, E.K., 2019. On the representativeness of pump water samples versus manta sampling in microplastic analysis. Environ. Pollut. 254A, 112970. https://doi.org/10.1016/j.envpol.2019.112970.
Villarrubia-Gómez, P., Cornell, S.E., Fabres, J., 2018. Marine plastic pollution as a planet boundary threat – the drifting piece in the sustainability puzzle. Mar. Policy 96, 213–220. https://doi.org/10.1016/j.marpol.2017.11.035.
Ziajahromi, S., Neale, P.A., Rintoul, L., Leusch, F.D.L., 2017. Wastewater treatment plants as a pathway for microplastics: development of a new approach to sample wastewater-based microplastics. Water Res. 112, 93–99. https://doi.org/ 10.1016/j.watres.2017.01.042.