Introduction

As an emerging contaminant, microplastics are distributed in each sphere (atmosphere, hydrosphere, lithosphere, pedosphere, biosphere) of the earth’s environmental system (Danopoulos et al. 2022). Microplastic contamination has raised concerns among governments and scholars worldwide. Microplastics are harmful to human health. For example, microplastics in the atmosphere can accumulate in human lungs (Amato-Lourenco et al. 2021) and cause pneumonia and cancer (Prata 2017). To prevent and control microplastic contamination, it is essential to reveal their sources, transport, and distribution.

Microplastics are divided into two categories according to their sources: primary and secondary. Primary microplastics are intermediate or final plastic products directly discharging into the environment (Zeng 2020). The leakage of these plastics during production, storage, and transportation induces the release of related primary microplastics to the environment (Sobhani et al. 2020). The use of some products (e.g., personal care and cosmetic products) containing primary microplastics (microbeads or pellets) also causes the release of primary microplastics (Hunt et al. 2020). Generally, primary microplastics have certain shapes (microbead, pellet, and cylinder) and polymer types (polyethylene (PE), polypropylene (PP), polymethyl methacrylate (PMMA), polystyrene (PS), and polyurethane (PU)). Secondary microplastics originate from the fragmentation of plastic products or wastes (Zeng 2020). Secondary microplastics and their original plastic products commonly have similar morphology and the same polymer type. Therefore, the morphology and polymer types can be used to identify the sources of both primary and secondary microplastics. This method of identification has been used in many studies of airborne microplastics. Those results indicated that plastic microbeads originated from coatings of additive materials (Liu et al. 2019c), microplastic fibers originated from clothes and other textiles (Liu et al. 2019b), and fragments, films, and foams of microplastics originated from various plastic products (Cai et al. 2017). Another common source identification method is the tracing simulation model, and the Hybrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) model is the most widely used one. Based on this model, existing studies indicated that microplastic deposition in the Pyrenees Mountains came from places less than 95 km away (Allen et al. 2019), airborne microplastics in the northwestern Pacific Ocean mainly came from land and the adjacent oceanic atmosphere (Ding et al. 2022), and microplastic deposition in the Badain Jaran Desert mainly originated from populated areas (Jinchang and Yinchuan) southeast of the desert (Wang et al. 2021). The first method is based on physical and chemical analysis, and the second method focuses on the transport process. The experimental method or simulation method was used separately in existing studies. The characteristics of the two methods suggest that their combined use may represent an efficient research method.

The atmosphere has been demonstrated to be an important pathway through which microplastics are transferred in the earth’s environment (Long et al. 2022). Atmospheric microplastic deposition has become the hot spot of current research. The translocation of microplastics between environmental spheres can be investigated by analyzing the spatiotemporal distribution of microplastic deposition (Wang et al. 2021). A study on the temporal distribution of microplastic deposition commonly needs to be performed for a long period of time. Therefore, current studies on this topic were limited, and the results were varied. A previous study showed that the deposition flux of atmospheric microplastics in a coastal city was less in autumn than those in other seasons (Zhou et al. 2017), and another study showed that the deposition fluxes of winter were higher than those of other seasons in another coastal city (Li 2021). The deposition fluxes of the dry seasons (winter and spring) were higher than those of the wet seasons (summer and autumn) in a semi-arid city (Abbasi and Turner 2021). Research on the spatial distribution of atmospheric microplastic deposition is still in its infancy. Generally, the deposition fluxes were higher in populous urban areas and lower in underpopulated suburban areas (Dris et al. 2016). However, another study found that the deposition fluxes in suburban areas were higher than those in urban areas (Klein and Fischer 2019). This study indicated that population density was not the only influencing factor and natural factors (e.g., wind) should be considered. An ignored potential microplastic contamination was found in atmospheric dust fall surrounding a glass recycling base (Zhang et al. 2022). Overall, the spatiotemporal distribution of atmospheric microplastic deposition still needs further study. The research of spatial distribution for a long period of time is particularly lacking.

From February to August of 2020, we had investigated the influence of the COVID-19 pandemic on the distribution and possible sources of atmospheric microplastic deposition in the urban area of Lanzhou city (Liu et al. 2022). That study might not show the general spatiotemporal distribution of atmospheric microplastic deposition, and the used source identification method was just the experimental one. In this study, the morphology, polymer types, and abundance of atmospheric microplastic deposition in the main area of Lanzhou city for 4 seasons were clarified. The sampling period was almost unaffected by the pandemic. On the basis of these, the spatiotemporal distribution of atmospheric microplastic deposition was investigated through geostatistical analysis. The simulation method (HYSPLIT modeling) was used to identify the potential sources of microplastic deposition. The findings of this study will show the detailed spatiotemporal distribution and potential sources of atmospheric microplastic deposition in a semiarid urban environment and can improve our understanding in this research field.

Materials and methods

Overview of the study region

The study region is the urban area of Lanzhou city and includes the population centers of Chengguan district, Anning district, Qilihe district, and Xigu district (Fig. S1). The area of the study region is 177.5 km2. In the study region, factories are distributed in the north of Xigu district, mainly producing petroleum products, synthetic resin, and synthetic rubber; universities are mainly distributed in the east and south of Anning district; commercial centers are mainly distributed in the middle and west of Chengguan district and the east of Qilihe district; and the other regions are mainly residential area. The study region is a typical semiarid region. During the sampling period (March 2021–February 2022), the average recorded temperature, precipitation, and relative humidity of the study region were 10.5 °C, 355.2 mm, and 39.6%, respectively. The total number of precipitation days during the sampling period was 41. The Yellow River flows through the study region from west to east. As a valley basin, the study region is characterized by a lower wind speed. The prevailing wind direction was from the east. The numbers of days with wind force levels 1, 2, and 3 were 92, 220, and 53, respectively. From October 26, 2021, to November 9, 2021, closed-off management was conducted in a few residential quarters to prevent the COVID-19 pandemic from spreading. The management had no obvious influence on the human activity in the study region. Except for this period, closed-off management was not conducted in any place of the study region.

Sample collection

According to the recommended method for setting atmospheric monitoring sites (SEPA 2003), 23 sampling sites were set in the study region to collect atmospheric deposition (Fig. 1). The sampling sites covered the main urban function zones. The sampling period was divided into four seasons: spring (March–May 2021; 11 sampling days), summer (June–August 2021; 16 sampling days), autumn (September–November 2021; 12 sampling days), and winter (December 2021–February 2022; 12 sampling days). The samples were collected according to the method in previous research (Abbasi and Turner 2021). In addition, the wet deposition samples were collected referring to a state standard of China (SEPA 1993), and the dry deposition samples were collected referring to another state standard of China (MEE 2021). During sampling, dry and wet deposition were separately collected by using the passive atmospheric deposition sampler (Fig. S2). On days when precipitation was forecasted, the old sampler in each sampling site was covered with aluminum foil and transferred to the laboratory for collecting dry deposition, then a new clean sampler was set on the sampling site to collect wet deposition. At the end of precipitation, the sample bottle was covered with aluminum foil, and the inside wall of the sample collector was rinsed 3 times with 600 mL of deionized and filtered water. The washing water was collected in another clean sample bottle and covered with aluminum foil. The 2 sample bottles were transferred to the laboratory immediately and stored in a refrigerator at 4 °C for the next test. After precipitation, another clean sampler was set on the sampling site for continuously collecting dry deposition. In all wet deposition samples, the liquid in the sample bottle did not spill. If a precipitation event was not observed for 10 days after previous precipitation, dry deposition samples were collected by using the method described above to avoid the degradation of microplastics and environmental pollution in the samples. For collecting dry deposition samples in the laboratory, all particles adhering to the inside wall of the sample collector were rinsed into the sample bottle three times with 600 mL of deionized and filtered water. The sample bottle was covered with aluminum foil and stored in a refrigerator at 4 °C for the next test. In addition to the above sampling date, all samples were collected at the last day of every month. All collected dry or wet deposition samples every month were used to calculate the dry or wet deposition flux of monthly microplastics, respectively. The condition of each sampling site was checked every 5 days. No bird droppings were found in each sampler. We collected 1173 samples (690 dry deposition and 483 wet deposition) during the sampling period.

Fig. 1
figure 1

The study region and locations of sampling sites

Full size image

Sample pretreatment

Due to few organic matters in atmosphere deposition samples, the samples were usually not digested to remove organic matters and were filtered directly (Prata et al. 2020). In this study, microplastics in the dry and wet samples without digestion were extracted in a saturated zinc chloride solution (1.55 g/mL) through density separation at the end of every month. According to the sampling method of dry and wet deposition, some liquid (deionized and filtered water and atmospheric fallout) remained in each collected sample bottle. A dividing rule was made to measure the volume of liquid in each bottle. If the liquid in the sample bottle was little, deionized and filtered water was added to make the liquid into 600 mL. Different weights of zinc chloride were added into the liquid of each sample bottle to make a saturated solution. The liquid in each sample bottle was mixed well by using a clear and dry glass stirring rod. One glass stirring rod was used in one sample bottle. After used, the glass tiring rod was washed with the filtered saturated zinc chloride solution, and the washing solution was added into the related sample bottle. Then, the sample bottle was covered with aluminum foil, waiting at least 12 h for the solution to clear. The clear supernatant solution was filtered through 1.0 μm pore size glass microfiber filters (Whatman GF/B) with a vacuum pump. The filtrate was returned to the sample bottle with sediment, repeating the above-mentioned process twice. After filtering the clear supernatant solution, the filter was washed by 20 mL of deionized and filtered water for three times to remove residual zinc chloride. The filters were dried in a constant temperature drying oven at 40 °C and then stored in clear and dry glass petri dishes sealed with aluminum foils. The petri dishes were kept in a refrigerator at 4 °C.

Identification and characterization of microplastic samples

The potential microplastics in the filter were taken photos through visual inspection referring to the morphological inclusion criteria (Zhang et al. 2020) by using a stereo microscope (Olympus SZX-16) with a digital camera (Olympus DP74). In this study, the shapes of microplastics were divided into 4 kinds: fiber, fragment (particle and foam), film, and pellet. Each shape of potential microplastic was selected by an ultramicro sampling needle (ϕ 100 μm, silver, prong) through visual inspection and then placed on the KBr salt tablet. The KBr salt tablet was placed on the microscope carrier of a Fourier-transform microscopic infrared spectrometer (Thermo Scientific Nicolet iN10) for collecting the infrared spectrum. The spectrum of potential microplastics was compared with the standard FTIR spectra to determine the polymer type based on more than 85% similarity, and the real microplastics were identified. In the image analysis software (Olympus Stream), the photos of identified microplastics were analyzed to record their numbers, shapes, sizes, and colors. The morphological analysis was conducted on the identified microplastics in a scanning electron microscope (Zeiss Sigma300). Due to the limitation of the analysis, microplastics with sizes less than 20 μm were not counted.

Quality control

The 100% cotton laboratory coats were worn throughout the outdoor sampling and indoor experiment. All indoor experiments were conducted in a clean glass booth equipped with a fan filter screen. The deionized water was filtered through 1.0 μm pore size glass microfiber filters (Whatman GF/B) beforehand. The glass microfiber filters were heated to 500 °C for 3 h prior to their use. All experimental apparatuses were rinsed with deionized and filtered water and then dried at 75 °C in an oven. All liquid and experimental apparatuses were always covered with aluminum foil to avoid contact with plastic materials. Two blanks were conducted during the process of each batch of samples. Just synthetic fibers were observed in all blanks, and the maximum quantity was 33 fibers in blanks. We deducted the average quantity of observed fibers in blanks out of the abundance of the corresponding batch of samples.

Potential sources identification of microplastics

Based on the present methods, we improved the potential source identification method conducted in our previous study. The modified method included two parts: transport trajectory analysis and survey of potential sources.

Transport trajectory analysis

The back trajectory analysis in the HYSPLIT model was widely applied to identify potential sources of microplastics in previous studies (Allen et al. 2019; Wang et al. 2021; Ding et al. 2022). In this study, the starting point of the backward trajectory was set at the central point (36.054789°N, 103.829298°E, 1529 m ASL) for the highest microplastic deposition flux distribution area. The model was run for 48 h starting from 0:00 a.m. on the 19th of each month at the same elevation level as the starting point.

Survey of potential sources

Firstly, the plastic products with the same polymer types of microplastic samples were confirmed referring to the literatures (Bruder 2019; Wang et al. 2019; Hunt et al. 2020; Wang et al. 2020; Zeng 2020). Secondly, searching for the related plastic products (or wastes) in the origin areas. If the related plastic products (or wastes) were found and they were numerous and widely distributed, the corresponding microplastic samples were believed to be from the origin areas; if not, the simulation was rerun with changing relevant parameters.

According to the results of transport trajectory analysis, the related plastic products and wastes were collected from 21 survey sites in the origin area (Fig. S3). The origin area was an urban area, and the plastic products were mainly articles for daily use. The plastic products were searched in composite market, commercial center, and supermarket where people frequently bought them. The plastic wastes were searched in the salvage station. The representative composite market, commercial center, supermarket, and salvage station were chosen for the survey in the origin area (Table S1). The disposable tableware was searched in the restaurant near the above survey sites (except the salvage station, because of no restaurant near it). The survey was conducted on the 17th and 18th of each month. According to the distribution of survey sites, the authors were divided into 3 groups for conducting the survey: the first group (2 people) was responsible for the survey sites in Chengguan district, the second group (2 people) was responsible for the survey sites in Qilihe district and the east part of Anning district, and the third group was responsible for the survey sites in Xigu district and the west part of Anning district (Fig. S3). The names, survey sites, and polymer types of collected plastic products and wastes were recorded (the data were not shown). If the polymer type was not marked on the plastic material (mostly plastic wastes), the polymer type was identified through infrared spectroscopic analysis.

Geostatistical analysis

The spatiotemporal distribution of contaminants can be investigated by using geostatistical analysis. The spatial interpolation is a crucial geostatistical analysis method. In the spatial interpolation method, the data of limited sites could be used to estimate the data of other sites (Tang and Yang 2012). Being limited by urban environmental conditions, there are few ideal sampling sites for atmospheric microplastic deposition. So, the method could be applied to investigate the spatiotemporal distribution of atmospheric microplastic deposition. We successfully used the inverse distance weight interpolation to investigate the temporal and spatial distribution of microplastic deposition in previous study (Liu et al. 2022). In this study, the ordinary Kriging interpolation was used to investigate the spatiotemporal distribution of microplastic deposition in ArcGIS 10.4 (ESRI, Redlands, CA, USA). The normal test was conducted through Normal QQPlot in ArcGIS 10.4. The test showed data obeyed normal distribution (Fig. S4). The cross-validation indicated that the predicted values were very close to the measured values (Fig. S5). Microplastic pellets (PE and PMMA) were only found in 5 sampling sites. The spatiotemporal distribution of pellets (PE and PMMA) was not investigated through general Kriging interpolation.

Results

Morphological characteristics and polymer types of microplastic samples

Different shapes of microplastics consisted of different polymer types (Fig. 2). The shapes of the microplastic samples were separated into 4 types: fiber (58.6%), fragment (23.9%), film (16.2%), and pellet (1.3%). Through infrared spectroscopic analysis, 7 polymer types were identified: polyamide (PA, 21.8%), polyethylene terephthalate (PET, 34.9%), PP (16.3%), PE (13.6%), PS (4.8%), polyvinyl chloride (PVC, 7.9%), and PMMA (0.7%). The proportions of shapes, polymers, sizes, and colors at the same types of sampling sites were almost the same in different seasons. The data in Fig. 3 were the average values of 4 seasons. The proportions of these kinds of microplastics in sampling sites were different. Pellets (PE and PMMA) were only found in the deposition of factory. Fibers (PET, PA, and PP) had high proportions in commercial center, residential area, university, and park. Fragments (PET, PS, and PE) and films (PE and PVC) had high proportions in salvage station. The PP was found in sampling sites except for the salvage station. In each sampling site, the proportion of microplastics decreased with increasing size. More than half of the microplastic samples were tiny and small size (≤ 500 μm), including fragments (PET, PP, PE, and PS), fibers (PET, PA, and PP), and pellets (PE and PMMA) (Table S2). Interestingly, the proportion of tiny sizes in the salvage station was far more than those of other sizes. Films (PE and PVC) and most fibers (PET, PA, and PP) were medium and large size (> 500 μm). Most microplastics were colorless (Fig. 3). Only a small proportion of fibers (PET, PA, and PP) and films (PE) were colored, including white, red, and blue (Table S3). The SEM photographs of the microplastic samples (Fig. S6) showed different flaws, including flakes in fibers (PET, PA, and PP); fractures, pits, and grooves in fragments (PET, PP, PE, and PS); and deformation and avulsion in the films (PE and PVC). But the SEM photographs of pellet (PE and PMMA) showed the smooth and regular surface without any flaws.

Fig. 2
figure 2

The infrared spectrogram and shape photograph of microplastic samples. PA polyamide, PE polyethylene, PET polyethylene terephthalate, PMMA polymethyl methacrylate, PP polypropylene, PS polystyrene, PVC polyvinyl chloride

Full size image
Fig. 3
figure 3

The proportions of shapes, polymers, sizes, and colors of microplastic samples. PA polyamide, PE polyethylene, PET polyethylene terephthalate, PMMA polymethyl methacrylate, PP polypropylene, PS polystyrene, PVC polyvinyl chloride

Full size image

Potential sources identification of microplastic samples

The results of the transport trajectory analysis are shown in Fig. 4. The origin area with a frequency more than 90% was located in the study area in most months except March. In March, the origin area with a frequency more than 90% was located in the desolate east part of Tengger desert, but the origin area with a frequency more than 80% was located in the study region. Through the survey, the plastic products or wastes with the same polymer types of microplastic samples were found (Table 1). Most plastic products were widely used in people’s daily life. Some plastic products were finished products of factory. Clothes and textiles had the same polymer types as fibers (PET and PA), and the polymer type of surgical mask was the same as fibers (PP). The packing materials of food, drinks, and other daily necessities had the same polymer types as fragments (PET, PP, PE, and PS) and films (PE and PVC). The pellets (PE and PMMA) were the main products of some factories in Xigu district. In salvage stations, most recyclable plastic wastes had the same polymer types as fragments (PET, PE, and PS) and films (PE and PVC). Through the survey, plastic products were found to be bought from composite market, commercial center, supermarket, and restaurant. The plastic products were widely used in residential area, commercial center, university, and park, and almost all of recyclable plastic wastes (bottles, foams, bags, films) were collected by junkmen and sold to salvage station.

Fig. 4
figure 4

The trajectory frequencies calculated by using the HYSPLIT model

Full size image
Table 1 The survey results of plastic products or wastes with the same polymer types of microplastic samples
Full size table

Spatiotemporal distribution of microplastic deposition

The spatiotemporal distributions of microplastic deposition flux showed significant differences (Fig. 5). The total deposition flux in each season decreased in the following order: summer (535.5 p/(m2·d)), autumn (357.3 p/(m2·d)), spring (318.7 p/(m2·d)), and winter (197.5 p/(m2·d)). The months of the highest and lowest total deposition flux were June 2021 (681.4 p/(m2·d)) and January 2022 (112.2 p/(m2·d)), respectively. The microplastic deposition fluxes in populous areas (commercial centers and residential areas in the east and middle of the study region) and salvage stations were higher than those in other areas. The total deposition flux of each type of sampling sites decreased in the following order: salvage station (521.3 p/(m2·d)), commercial center (468.1 p/(m2·d)), residential area (327.9 p/(m2·d)), university (229.9 p/(m2·d)), park (203.1 p/(m2·d)), and factory (138.1 p/(m2·d)). The highest deposition flux (810.0 p/(m2·d)) was observed at salvage stations in summer, and the lowest highest deposition flux (79.5 p/(m2·d)) was observed at parks in winter.

Fig. 5
figure 5

The spatiotemporal distribution of atmospheric microplastic deposition in the study region. The unit of deposition flux was p/(m2·d). PA polyamide, PET polyethylene terephthalate, PP polypropylene, PS polystyrene, PVC polyvinyl chloride

Full size image

The deposition fluxes of different kinds of microplastics were also highest in summer and lowest in winter (Fig. 5). Abundant fibers (PET, PA, and PP) and fragments (PP) were distributed in areas with high population density. Most fragments (PET, PS, and PE) and films (PE and PVC) were distributed around salvage stations with low population and human activity intensity. Pellets (PE and PMMA) were distributed at the factory and the surrounding 4 sampling sites (Fig. 6). The deposition flux of pellets (PE and PMMA) in the factory was the highest among the five sampling sites.

Fig. 6
figure 6

The spatiotemporal distribution of pellets. PE polyethylene, PMMA polymethyl methacrylate

Full size image

Discussion

Potential source analysis of microplastic deposition

Previous research indicated that microplastics were able to transport for far distance in the atmosphere and then settled on the ground (Abbasi and Turner 2021). We also found that the origin area of microplastic deposition with a frequency more than 90% in March was located in the east part of Tengger desert. The east part of Tengger desert is sparsely populated and has scarcely any plastic products or wastes. So, this area was unlikely to be the source of microplastic deposition in March. The origin area with a frequency more than 80% in March was located in the study region. The plastic products or wastes with the same polymer types of microplastic samples were found in the study region. The study region was more likely to be the source of microplastic deposition in March and other months.

The polymer types of microplastics provide strong evidence used to trace their original plastics. Microplastics and their original plastics must have the same polymer types. This could be used to verify the origin area of microplastic deposition. Through the survey in the study region, we found the plastic products or wastes with the same polymer types of microplastic samples. Microplastics were demonstrated to be released during the use of plastic products and the disposal of plastic wastes. The use and airing of clothes and textiles composed of synthetic fiber could release fibers into the atmosphere (Liu et al. 2019a). Since the year of 2020, surgical masks were widely used for personal protection. This also could induce the release of fibers into the atmosphere (Hu et al. 2022). The use of packing materials of food, drinks, and other daily necessities could release fragments and films into the atmosphere (Sobhani et al. 2020). Due to recycling, most plastic packing materials (bottles, foams, bags, films) were sold to salvage stations. In salvage stations, these packing materials were usually stored in the open air for a long time (Fig. S7). Under the influence of sunlight, the fragmentation of these packing materials was accelerated, releasing fragments and films into the atmosphere (Song et al. 2022). The surface flaws of fibers (PET, PA, and PP), fragments (PET, PP, PS, and PE), and films (PE and PVC) indicated that they came from the fragmentation and wear and tear of related plastic products and wastes. In addition, the leakage of resin particles during production, storage, and transportation also released microplastics to the atmosphere (Sobhani et al. 2020). The smooth and regular surface of pellets (PE and PMMA) indicated they did not come from other plastic products and were the intact plastic products. Thus, these indicated that the study region had potential sources of microplastic samples.

The results of the HYSPLIT model indicated that microplastic deposition came from the study region. This could be attributed to the particular topography and meteorological conditions of the study region. The study region is a river valley surrounded by high mountains. The wind speed in the study region was low. This induced a high atmospheric stability. Atmospheric contaminants in the study region are not easily dispersed, and atmospheric contaminants are difficult to transport from other places to the study region (Shang et al. 2001). Therefore, most airborne microplastics only settled on the study region. Through simulation and survey, the study region was identified as a potential source of microplastic deposition.

Spatiotemporal heterogeneity of microplastic deposition distribution

Many studies had shown that the deposition fluxes of microplastics changed with the season (Zhou et al. 2017; Abbasi and Turner 2021; Li 2021). Our study also indicated significant differences in the deposition fluxes of different seasons. However, the seasons with the highest and lowest deposition fluxes were different from those studies. In this study, the highest deposition flux was observed in summer, and the lowest deposition flux was observed in winter. A previous study found that the deposition flux of microplastics was higher in periods with more precipitation than those in periods with less precipitation (Dris et al. 2015). This is because of the effective scavenging ability of precipitation for atmospheric particulates larger than 4 μm (Dai 2006). The sizes of microplastic samples in this study were more than 20 μm. The precipitation was the highest in summer and the lowest in winter. So, the wet deposition flux was the highest in summer and the lowest in winter (Table 2). The total deposition flux was also augmented by the wet deposition flux. In four seasons, the average proportion of wet deposition flux was 50.1%. However, a study conducted in another semi-arid region indicated that the deposition fluxes in dry seasons were higher than those in wet seasons, and wet deposition only contributed 5% of the annual deposition (Abbasi and Turner 2021). This discrepancy could be attributed to the difference in precipitation between the two studies. In this study, the precipitation of four seasons was 355.2 mm, which was far more than the 226.3 mm precipitation of four seasons in that study. The high precipitation induced high wet deposition flux and increased total deposition flux. In addition, high temperature contributes to the release of microplastics from plastics (Ding et al. 2021). The mean temperatures of the four seasons decreased as follows: summer > spring ≈ autumn > winter (Table 2). The total deposition flux showed the similar trend. Therefore, high temperatures in summer increased the release of microplastics, and high precipitation increased the deposition efficiency. This led to temporal heterogeneity in the deposition of microplastics in the study region.

Table 2 The wet and dry deposition fluxes of microplastics and meteorological conditions
Full size table

It was generally acknowledged that microplastic deposition was more in populous areas than that in underpopulated areas (Dris et al. 2016). We also found high deposition flux of microplastics in populous areas such as commercial centers and residential areas in the east and middle of the study region. However, abundant microplastic deposition including fragments (PET, PS, and PE) and films (PE and PVC) was also found in sparsely populated areas surrounding salvage stations. Through the survey, almost all of the recyclable plastic wastes in salvage stations were bottles, foams, bags, and films. These plastic wastes had the same polymer types as microplastic fragments and films. The releasing sources of fragments (PET, PS, and PE) and films (PE and PVC) might be in salvage stations. Exposed to sunlight for a long time, these recyclable plastic wastes could release abundant microplastics into the atmosphere (Song et al. 2022). In the study region, the atmospheric stability is high (Shang et al. 2001), and the wind speed is low. The horizontal diffusion of airborne microplastics was very slow. This induced that most of the airborne microplastics released from plastic wastes were deposited in areas surrounding salvage stations. Similarly, clothes, textiles, surgical masks, and disposable food packagings with the same polymer types as microplastic fibers (PET, PA, and PP) and microplastic fragments (PP) mainly existed in populous areas such as commercial centers and residential areas. The use and carelessly discarding of these plastic products could release related microplastics into the atmosphere (Liu et al. 2019a; Sobhani et al. 2020; Hu et al. 2022). Under the stable atmospheric conditions, these microplastics were deposited in the populous areas. In addition, the microplastic pellets (PE and PMMA) were abundant in the factory which produced resin particles composed by PE and PMMA. These indicated the important influence of potential sources on the spatial distribution of microplastic deposition. So, in the study region with high atmospheric stability, the spatial distribution of microplastic deposition was influenced not only by population density but also by potential sources.

Contamination characteristics of microplastic deposition in the study region

Based on the comparison with deposition fluxes of previous studies, the total deposition flux of microplastics in this study was higher (Table S4). In addition, the total deposition flux in this study was also higher than that reported in our previous study conducted in the same study region from February to August of 2020. On the one hand, the high atmospheric stability of the study region was not conducive to the horizontal diffusion of airborne microplastics, and most airborne microplastics settled in the study region. On the other hand, the abundant potential sources of airborne microplastics in the study region played an important role. The high total deposition flux of salvage stations indicated more contribution of plastic wastes to microplastic deposition than plastic products. Through the survey, we found that the recycle of plastic wastes was poorly organized in the study region (Fig. S7). The garbage classification was not normatively enforced in the study region. This was mainly attributed to the lack of awareness of garbage classification among residents (Ding et al. 2021). The transportation of recyclable plastic wastes needed protective measures to avoid leakage. In salvage stations, the recyclable plastic waste was stored in the open air for a long time. The poorly organized plastic waste recycling system was conducive to the release of microplastics from plastic wastes. This was not exclusive in the study region, and many cities of China had the similar problem (Ma and Sun 2021). The release of microplastics from salvage stations might be a nationwide problem.

The identified microplastics in the study region included all shapes and most polymer types in previous studies (Table S4). Compared with our research in 2020, PP is a new polymer of microplastics found in this study. Some fibers and fragments consisted of PP. The PP fibers mainly originate from surgical masks. During the COVID-19 pandemic, surgical masks were widely used for personal protection. The use and disposal of surgical masks can release abundant PP fibers (Hu et al. 2022). The PP fragments mainly came from disposable meal boxes and drink cups in the take-out catering industry. In recent years, China’s take-out catering industry has developed rapidly (Xinhuanet 2021). This generates a large amount of disposable plastic waste. The using and carelessly discarding of these plastic products induced abundant PP fibers and PP fragments in populous areas such as commercial centers and residential areas.

Conclusion

In this study, we investigated the spatiotemporal distributions and potential sources of atmospheric microplastic deposition in a semiarid city for 4 seasons. The deposition flux of atmospheric microplastics was higher than that in many other cities. Four shapes of microplastics were identified, including fiber, fragment, film, and pellet. Microplastics consisted of 7 polymer types: PA, PE, PET, PMMA, PP, PS, and PVC. The size of more than half of the microplastics was less than 500 μm, and most microplastics were colorless. Through HYSPLIT model analysis, the study region was suggested as the potential source of microplastic deposition. According to the survey, the microplastic deposition might come from plastic products in populous areas and plastic wastes in salvage stations. The distribution of microplastic deposition had significant spatiotemporal heterogeneity. The deposition flux of microplastics decreased as follows: summer > spring ≈ autumn > winter. Abundant fibers (PET, PA, and PP) and fragments (PP) were distributed in populous areas such as commercial centers and residential areas. Most fragments (PET, PS, and PE) and films (PE and PVC) were distributed around salvage stations. Almost all of the pellets (PE and PMMA) were found in the factory. The temporal distribution of microplastic deposition was influenced by precipitation and mean temperature. The spatial distribution of microplastic deposition was influenced by potential sources and population density. For preventing and controlling atmospheric microplastic contamination, the poorly organized plastic waste recycling system requires normative management.