Plastic pollution is ubiquitous in the marine environment. Plastics, due to their durability, low cost, and lightweight, are important in our lives and have been shed to the environment in the last few decades like never. Since the beginning of plastic production in the early 20th century, it has continuously increased and reached 368 million tons globally in 2019 (Plastics Europe, 2020). Microplastics (MPs) are frequently defined as particles with lengths of less than 5 mm (Arthur et al., 2009; Cole et al., 2011).

MPs’ sources and pathways are of utmost importance to control and prevent plastics from entering the ecosystem (He et al., 2019). The exuberance of MPs in the aquatic environment is due to inappropriate human behavior and improper waste management (Jambeck et al., 2015). There is limited information on the amount of plastic waste entering the oceans, but it is widely cited that land sources contribute approximately 80% of the marine plastic debris (Jambeck et al., 2015). Rivers, surface water runoff, sewage treatment, and wind-induced air transport are the major gateways of plastic debris in the aquatic ecosystem (He et al., 2019). In addition, the plastic manufacturing industries release plastics in the form of pellets and resin powders that, via air-blasting, can contaminate the aquatic environment (Eriksen et al., 2013). Plastic pellets are also released from marine accidents during handling and transportation (Veerasingam et al., 2016). Coastal activities, including fisheries, aqua tourism, and marine industries, are also the sources of MP pollution in the marine environment (Eriksen et al., 2013).

MPs can be of two types, primary and secondary MPs. Primary MPs are produced as microscopic particles present before entering the environment and exist as microbeads found in personal care products and plastic pellets (or nurdles). Secondary MPs are formed by physical, biological, and chemical degradation of macroscopic plastic parts and are the main source of microparticles released into the environment (Boucher and Friot, 2017; Zhu et al., 2020). They are formed by the degradation of improperly disposed plastic waste, tire abrasion, and washing of synthetic textiles (Boucher and Friot, 2017; Zhu et al., 2020).

MP particles have different shape classes: fragments, films, filaments, foams, and pellets (GESAMP, 2019). In this study, we divided all MP particles into fibers and fragments. The fragments category includes all non-filament particles like films, foams, and pellets, and the fibers category includes both filaments and fibers. Some polymers, such as PVC, polyester, polyamide, and acrylic, are denser than seawater and, thus, sink to the bottom of the sea (Hidalgo-Ruz et al., 2012). Polymers with a lower density than seawater usually float on the surface, including polyethylene, polypropylene, and expanded polystyrene (Hidalgo-Ruz et al., 2012). Low-density MP accumulates especially within a layer of a few cm below the air–water interface (Andrady, 2011). Thus, most studies target MP abundance and distribution confined to the surface layer (Collignon et al., 2014). Plastic products, traditionally made of monomers, are linked to the form of the polymer structure. During plastic production, several additives are added for promoting specific properties towards its use (Lithner, 2011). As the plastics degrade over time, these chemicals tend to leach out, including coloring agents, and accumulate in animals’ stomachs, resulting in bioaccumulation and biological effects (Mato et al., 2001; Teuten et al., 2009). Worldwide, individuals are accustomed to seafood consumption, which makes it possible for people to be exposed to MPs (Wright and Kelly, 2017). Chemical additives or remaining monomers can pose a potential danger to human health and the ecosystem (Lusher et al., 2017).

Colored plastics impose a high threat to aquatic species (Li et al., 2021). From one side, marine species have difficulty distinguishing between transparent and colored plastics, which increases the risk of MPs ingestion. Moreover, colored MPs may be accidentally ingested by fishes, turtles, and birds (Zhao et al., 2016; Gago et al., 2019). On the other hand, the impact of color affects the selection of food made by aquatic species.

As demonstrated by studies from different Baltic Sea basins, the occurrence of MPs in the Baltic Sea is evident (Setälä et al., 2016; Tamminga et al., 2019; Uurasjärvi et al., 2021). However, the knowledge about the spatial distribution and temporal variability of MPs in the Baltic Sea is limited (Aigars et al., 2021). Also, the methodology used to collect MPs varies by instruments, mesh size, sampling depth, and sampling area.

The main objective of the present paper is to assess the spatiotemporal variability of MPs in the eastern Baltic Sea and analyze environmental drivers affecting it. We report and analyze 5-year measurements of MPs in the surface layer in the eastern Baltic Sea. Specifically, we address the following main questions in this study: What is the mean spatial distribution of MPs in the eastern Baltic Sea? Is there seasonality in the MP concentrations? What is the share of fibers/fragments, different particle sizes, and colors in the MP pool? What processes cause the short-term variability in the MP concentration?

MPs were found at all 16 sampling stations. In total, 9,414 MP particles were extracted from 23,199 m³ water of 122 surface water samples. When total MP particles were divided by the total water volume, the mean was 0.41 counts/m³. However, the arithmetic mean of MP concentrations in samples was 0.49 counts/m³ (0.08 counts/m²), and in the regions of BP, GOFC, GOFW, GOFE, GOR, and VS, the mean concentrations were 0.65, 0.59, 0.65, 0.46, 0.33, and 0.11 counts/m³ (0.11, 0.10, 0.11, 0.08, 0.06, and 0.02 counts/m²), respectively. The results show high variability in concentrations (STD ±0.46 counts/m³) and heterogeneity in distribution patterns of MPs in the eastern Baltic Sea. The relative abundance of MP fibers and MP fragments at different sampling sites is presented in Figure 2 . The average concentration of MP fibers and MP fragments across the dataset was almost the same: 0.25 and 0.24 counts/m³ (0.04 and 0.04 counts/m²), respectively. The annual mean of the share of fibers was higher in 2016, 2019, and 2020 ( Figure 2 ).

We have reported results from the 5-year MP observations in the eastern Baltic Sea. Next, we compare our findings with previous studies that collected data using surface trawling (Manta trawl) like us in the Baltic and global oceans. In the sea surface layer of the Baltic Sea, Aigars et al. (2021) detected an MP concentration of 0.09–4.43 counts/m³, Karlsson et al. (2020) found 0.18–0.92 counts/m³ in the Gullmar fjord at the Swedish west coast, 0.05–0.09 counts/m³ were observed in the South Funen Archipelago (Tamminga et al., 2018), Gewert et al. (2017) measured 0.19–7.73 counts/m³ in the Stockholm Archipelago, and 0–0.8 counts/m³ were found in the GOF (Setälä et al., 2016). In the Arctic waters, Lusher et al. (2015) revealed MP concentrations of 0–1.31 counts/m³, 0.07–9.25 counts/m³ were measured in the Eastern Mediterranean Sea (Adamopoulou et al., 2021), 1.82 counts/m³ were observed in the Mediterranean Sea (Zeri et al., 2018), and 0.06–25.9 counts/m³ were found in the Eastern Indian Ocean (Li et al., 2021).

Although Aigars et al. (2021) collected samples at numerous stations during 1 year, while we conducted measurements at three stations, during 5 years, the mean concentrations in the central GOR were roughly in the same order. Our observations revealed lower concentrations in the northern GOR and the Pärnu Bay; however. Setälä et al. (2016) identified an average MP concentration of 0.3 counts/m³ in the GOFC region, and our investigation revealed 0.59 counts/m³. We can conclude that the concentrations registered in our study—in the range of 0.01 to 2.45 counts/m³ (0.002–0.43 counts/m²) with a mean concentration of 0.49 counts/m³ (0.08 counts/m²)—are in the same order as previous studies in the Baltic Sea.

The mean concentrations in the three subregions in the GOF were in the range of 0.46–0.65 counts/m³, while it was 0.33 counts/m³ in the GOR. The difference between concentrations in the GOR and GOF could be explained by human pressure. Population in the catchment area per surface area of the GOF is ca. 400 inhabitants km⁻² while it is 150 inhabitants km⁻² in the GOR catchment (HELCOM, 2004). Note that in the easternmost part of the GOF, in the area we did not cover, the MP concentrations are probably higher than we observed due to the impact of river Neva (Martyanov et al., 2021). If we combine our results with Aigars et al. (2021), a meridional pattern is revealed in the GOR: higher MP concentrations in the south and lower MP concentrations in the north. This could be related to the population in the catchment areas as well. The water entering the southern part of the GOR is impacted by ca. 2.4 million inhabitants while the total population in the catchment area of the GOR is ca. 2.7 million.

The low human pressure is a likely reason behind the small mean concentration in the VS. The catchment area of the VS has an extremely low population, and there are no larger towns or other considerable point sources at the coast of the VS. Low MP concentrations in a similar area, in the South Funen Archipelago in Denmark, were explained by the sheltered position of the study area, low human pressure on adjacent islands, and the absence of any major potential point sources (Tamminga et al., 2018).

We registered considerable amounts of MPs at stations Pal (2.10 counts/m³) and 2 (1.49 counts/m³) in proximity to the city of Tallinn. Our observations follow previous studies that reported that the MP particle concentrations are frequently greater near densely populated urban areas with pollution sources such as industry and wastewater treatment plants (Yonkos et al., 2014; Gewert et al., 2017; Schönlau et al., 2020).

High mean MP concentration was detected at offshore station 85. This is a somewhat controversial result as earlier studies have reported higher values near coasts and rather low values offshore in the BP (e.g., Aigars et al., 2021). No major rivers enter the area, or remarkable point sources (cities and towns) are nearby. On the one hand, a possible explanation could be that the northern BP is the accumulation zone, where the discharge and buoyant particles from different basins (the GOF, GOR, Bothnian region, and south- and eastern BP) are concentrated. Secondly, the mean cyclonic current structure of the BP (Placke et al., 2018; Liblik et al., 2022) recirculates/traps the surface water in the basin for a longer period. In addition, MPs exhibit different buoyancy characteristics based on their density, shape, size, and biofouling rate (Adamopoulou et al., 2021). Convergence and downwelling act as a sorting mechanism, with relatively larger particles staying in the surface layers and smaller particles getting transported deeper in the water column (van Sebille et al., 2020). The biofilm growth is generally faster for smaller particles due to their high surface-to-volume ratios (Tsiaras et al., 2021). Thus, the buoyancy patterns described above and the fact that nearly 81% of MPs detected were smaller than 1 mm allow us to justify the high mean MP concentration at station 85. Recently, it has been shown that the region is prone to be affected by the submeso- and mesoscale activity (Väli and Zhurbas, 2021; Zhurbas et al., 2022), which can contribute to the convergence and divergence of MPs in the surface layers. We showed that high variability and convergence at the (sub)mesoscale could be a factor leading to high MP concentrations, but further investigations are needed to understand the pathways and reasons behind the phenomenon.

We considered two shapes of the MP particles in the current study: fibers and fragments, which accounted for 96% of the encountered particles in the eastern Baltic according to the recent study (Aigars et al., 2021). MP fragments are more likely to break up into smaller pieces and are caught in the Manta trawl than other MPs (Li et al., 2021). Synthetic fibers derived from textile materials could enter the aquatic ecosystem through sewage systems, surface runoff, or atmospheric transport and deposition (Bai et al., 2018; Liu et al., 2019; Wang et al., 2020).

The share of fibers (51%) and fragments (49%) was approximately equal in the current study. The latter is valid for the whole dataset, as well as the subregions. However, the share of shapes and concentrations might be influenced by the sampling method: Manta trawling with the mesh size of 330 µm. Many studies reported more fibers on the sea surface (Setälä et al., 2016; Bagaev et al., 2017; Gewert et al., 2017; Tamminga et al., 2018; Aigars et al., 2021), compared with fragments. On the other hand, some studies showed a lower proportion of fibers (Zhang et al., 2017; Pan et al., 2019; Karlsson et al., 2020). When samples are collected using the Manta trawl, MP fibers might leak in high numbers through the 330-µm mesh but could be more efficiently trapped when the mesh size is smaller (<100 µm; Setälä et al., 2016). However, even when a smaller mesh size (100 µm) is used, the number of fibers does not appear to rise, as fiber size, particularly from clothes, is less than 20 µm (Setälä et al., 2016).

Despite the equal share of fibers and fragments, the correlation between the concentration of the two shapes in the whole dataset was weak, although significant. However, high correlations were found near the Pärnu river mouth at station K5 (r ² = 0.87) and near the outlet of the Paljassaare wastewater treatment plant, at station Pal (r ² = 0.60). In the rest of the stations, there was a significant weak correlation, except at station 2 in the GOF and offshore station 85 in the BP, where the correlation was not found. Due to disturbance-induced vertical transport, small size MPs get resuspended to the surface layer (Xia et al., 2021). As K5 is a shallow station (depth, 5 m), and nearly three-fourths of MP detected are small, we suggest that the high correlation is related to the resuspension of MP to the sea surface layer. Correlation further off the sources was weaker due to the impact of marine processes, e.g., biofouling and vertical mixing, which could have a different effect on the fibers and fragments.

The MP fragments were mostly white, blue/green, and gray/black, whereas MP fibers were mostly gray/black and blue/green in the current study. This result is consistent with other studies where MP fragments and MP fibers were reported (Zhang et al., 2017; Karlsson et al., 2020; Aigars et al., 2021). The share between the two size classes 330–999 µm and 1,000–4,999 µm was 75% and 25%, respectively. The higher number of particles with decreasing size has been documented earlier in the eastern Baltic Sea (Setälä et al., 2016; Aigars et al., 2021).

The mean concentration was higher in autumn than in summer in all regions and stations. This seasonal signal is likely related to the biofouling and consequent sinking of the MP (Kaiser et al., 2017). Spring and summer are biologically active seasons in the Baltic Sea (Lips et al., 2014; Kahru et al., 2016; Purina et al., 2018). Decay and deepening of the seasonal thermocline and cooling of the upper mixed layer water start in the second half of August in the eastern Baltic Sea (Liblik and Lips, 2011; Skudra and Lips, 2017), which leads to the decrease of organic matter production (e.g., Gasiūnaitė et al., 2005), reduced biofouling, and consequently declined sinking rate of the MP. Moreover, the density of the upper mixed layer increases in autumn, which increases the buoyancy of the MP and reduces the sinking probability as well.

The shorter-term and smaller-scale spatial variability of the MP concentration in the sea surface is shaped by various processes such as advection, divergence, convergence, and vertical mixing (Auta et al., 2017; Lebreton et al., 2018; Zhang et al., 2020). The wind mixing distributes the MP vertically (Kukulka et al., 2012); thus, the MP concentration in the surface layer and wind speed can be negatively correlated (e.g., Schönlau et al., 2020). We found a significant negative, but weak correlation between the wind speed and the MP concentration only at offshore station 14.

The highest concentrations of the MP at the coast of Narva Bay were observed during the downwelling event while the lowest value was detected during the coastal upwelling event. There was a 60-fold difference between the highest and lowest case. The low values during upwelling can be explained by the subsurface origin of the water. The upwelling water originates from the cold intermediate layer in the GOF (Lips et al., 2009). The downwelling causes convergence of the upper layer water and the upper mixed layer in summer could deepen over 40 m (Liblik et al., 2017). Despite the downward movement, the buoyant-enough particles tend to stay at the surface and accumulate (Kooi et al., 2016; Waldschläger and Schüttrumpf, 2019). Thus, in the enclosed sea, where wind from any direction causes downwelling/upwelling along some coastal sections (Myrberg and Andrejev, 2003), the coastal mesoscale processes can potentially cause remarkable variability in the MP concentrations. Moreover, mesoscale eddies could redistribute the MP. The anticyclonic eddies converge the debris, and thus, concentrations there can be much higher compared to cyclonic eddies as shown in other areas (Brach et al., 2018). It is probable that the submesoscale processes, which are evident in the observations (e.g., Lips et al., 2016) and which converge and diverge tracers according to simulations (e.g., Zhurbas et al., 2022) in the smaller spatiotemporal scale, affect the MP concentrations and pathways as well in the Baltic Sea. Our samples were collected along a 1- to 4-km long line; thus, to study the MP in the submesoscale in more detail, other measurement methods, e.g., in situ pumping (Karlsson et al., 2020), should be implemented.

The dataset analyzed in the present study provides the first view on MP pollution and its spatiotemporal variability in the surface water of the eastern Baltic Sea. MPs were found in all 122 samples, and their concentration varied from 0.01 to 2.45 counts/m³ with a mean concentration of 0.49 counts/m³. The obtained concentration ranges, the share of the MP fragments and fibers, and the color composition of MPs generally agree with previous studies in the neighboring areas. The regional differences in the mean MP concentrations in the GOF, GOR, and Väinameri Archipelago Sea are likely related to the human pressure (population) in the catchment areas. The seasonal increase in the concentration from summer to autumn can likely be explained by the decline in the biofouling in autumn and related decrease in the sinking rate of particles.

The high variability in the observations was probably the result of multiple processes, which could not be fully captured by the design of the monitoring program. However, we managed to show that upwellings and downwellings, and wind mixing play a role in the variability of the sea surface MP concentration. It is likely that other (sub)mesoscale processes alter the MP concentrations in the surface layer as well. To improve the knowledge on the pathways of the MPs, the processes in the (sub)mesoscale from the sources to offshore should be addressed by further dedicated observational and modeling studies. Likewise, the measurement and modeling effort to estimate the land–sea and water column–sediment fluxes of MPs should be sought.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation. The whole dataset is available via EMODnet Chemistry and Estonian Environmental Database (KESE).

AM: Conceptualization, Data curation, Formal analysis, Methodology, Validation, Software, Visualization, and Writing—original draft. NB and KL: Methodology, Data curation, Writing—original draft, and Resources. IL: Conceptualization, Resources, Validation, Writing—review and editing. TL: Conceptualization, Data curation, Visualization, Validation, and Writing—review and editing. GV: Conceptualization, Data curation, Validation, Software, Visualization, and Writing—review and editing. UL: Conceptualization, Methodology, Formal analysis, Visualization, Validation, and Writing—review and editing. All authors contributed to the article and approved the submitted version.

Microplastics monitoring was financed by the Estonian Ministry of the Environment, Environmental Agency and Environmental Investment Center. This work was supported by the Estonian Research Council grant PRG602, H2020 project CLAIM (grant agreement no. 774586), and JPI Oceans projects ANDROMEDA (agreement no 4-1/20/160) and RESPONSE (agreement no 4-1/20/161).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.