A Regional Di ﬀ erence Analysis of Microplastic Pollution in Global Freshwater Bodies Based on a Regression Model

: Based on statistical data of the average abundance of microplastics from 37 global freshwater locations up to November 2019, we classiﬁed the freshwater bodies according to developments in their local countries and geographic positions. We highlighted the di ﬀ erences and causes of microplastic pollution in the waters of both developed and developing countries and urban and rural areas. The results showed that microplastic pollution was highest in Asia. The pollution in developed countries was signiﬁcantly lower than in developing countries. The di ﬀ erences in freshwater pollution between urban and rural areas mainly depended on the extent of human activity. The present study found the following phenomena by comprehensively using simple and multiple regression models and a Pearson correlation analysis to solve the impacts of the features, natural factors, and social and economic factors on the distribution of microplastic pollution. The density of microplastics was higher, which promoted the aggregation of microplastics in sediments. Pursuant to that, microplastic pollution was also inﬂuenced by the space-time pollution of movable surface sources, such as the soil and air. A population increase and the average gross domestic product (GDP) could also worsen microplastic pollution. Y.Q.; methodology, H.C.; software, H.C.; validation, Y.Q., H.H., and W.X.; formal analysis, H.C.; investigation, Y.Q.; resources, H.H.; data curation, Y.Q.; writing—original draft preparation, H.C.; writing—review and editing, H.H.; visualization, W.X.; supervision, W.X.; project administration, W.X.; funding acquisition, W.X.


Introduction
With the continuous progress and development of society, plastics have permeated every major facet of human life and economy. Large scale production and plastic waste have had great negative impacts on our environment [1]. According to statistical data, the global production and widespread use of plastics led to an accumulation of 348 million tons in 2017. China produced 102.3 million tons of plastic materials, accounting for 29.4% of the world's plastic production. As a result of this, China became the world's largest producer and consumer of plastic materials [2].
In 2004, Thompson et al. from the University of Plymouth published an article in Science that first posited that microplastics are small pieces and particles of plastic less than 5 mm in diameter [3]. Microplastics are divided according to their source: primary type and secondary type [4]. Scientific research proved that secondary microplastics have a stronger potential adsorption rate on organics and several heavy metals, like Pb and Hg, compared with primary microplastics. Pollutants will be immediately transferred to aquatic organisms when they ingest food laced with microplastic particles, which is a threat to humans and the environment through the food chain [5][6][7].  18 Goiana River River South America Brazil Developed country Urban type 0.190 [22] 19 KwaZulu-Natal River River Africa South Africa Developing country Urban type 0.487 [23] 20 Lake Hovsgol Lake Asia Mongolia Developing country Urban type 0.044 [14] 21 East Lake (Zhejiang) Lake Asia China Developing country Urban type 220 [24] 22 Ling Lake Lake Asia China Developing country Urban type 350 [24] 23 Dongting Lake Lake Asia China Developing country Urban type 633.5 [25] 24

Units of Measurement
As an emerging research field, microplastic pollution has attracted the attention of many international researchers in recent years. Due to the different research methods and the lack of internationally unified standards, there have been many forms of measurement units for microplastic abundance. In this paper, n/m 3 was selected as the standard unit. There were 13 different forms of units used to describe the abundance of microplastics in the investigated waterbodies, and the conversion methods are shown in Table 2.

Measure Original Unit Conversion Formula Note
Unit volume items/L 1 items/L = 10 3 n/m 3 Items, ind, and pieces are equal to abundance units n.

Research Methods
Modeling analysis methods are generally applied to trace regional differences in pollution. For example, Meng Lixia et al. used the Theil Index to analyze the differences in the intensity of water pollution emissions in China [33]. Through the application of the Hybrid Single Particle Lagrandian Integrated Trajectory Model (HYSPLIT), potential source contribution factor method (WPSCF), and concentration weighted trajectory analysis (WCWT), Yan Yu et al. analyzed the pollution characteristics and source differences of fine particulate matter (PM2.5) [34]. Based on the Stochastic Impacts by Regression on Population, Affluence and Technology model (STIRPAT), Wang Naichun et al. analyzed the influencing factors of carbon emissions [35]. In view of the absence of multicollinearity among the selected influencing factors and the scattered research contents related to microplastic pollution (the significances of all variables were higher than 0.05 in the Pearson correlation analysis), we adopted a regression model to analyze the factors that cause differences in the distribution of microplastic pollution.
Through the literature collection and data collation, this paper first describes the distribution map of the average microplastic abundance in freshwater bodies by ArcGIS (a software that can collect, organize, and analyze the geographic information), and compares the differences of microplastic pollution characteristics in developing countries and developed countries, as well as urban and rural freshwater bodies. After collecting all relevant data on the composition and structure of microplastics, water depth, surface area of the water, distance from the city center, region population, population density, per capita GDP, urbanization rate, and secondary industry structure, the primary influencing factors for the distribution of microplastics were identified using a Pearson correlation analysis, univariate or multiple regression models, and other analytical methods.

Difference Analysis of Microplastic Pollution in Global Freshwater Bodies
Through the collection and collation of the sample data, we obtained the pollution distribution of microplastics in global freshwater bodies ( Figure 1). There were relatively more studies on water microplastics in Asia, Europe, and North America. The order of microplastic pollution in freshwater bodies of six continents was: Asia > North America > Africa > Oceania > South America > Europe. China was the country with the most serious microplastic pollution in the world, with the average concentration of microplastics in the water as high as 3793.38 ± 799.68 n/m 3 . Among Chinese locations, the average abundance of microplastics in Wuhan Bei Lake was the highest, reaching 8925 n/m 3 [18]. Globally, Switzerland had the lowest level of pollution, with an average concentration of 0.04 ± 0.01 n/m 3 . In terms of the concentration of microplastics, the highest was Lake Neuchatel, which was 0.061 n/m 3 [17], and the lowest was Lake Zurich with 0.011 n/m 3 [12].
Water 2020, 12, x FOR PEER REVIEW 6 of 16 Through the literature collection and data collation, this paper first describes the distribution map of the average microplastic abundance in freshwater bodies by ArcGIS (a software that can collect, organize, and analyze the geographic information), and compares the differences of microplastic pollution characteristics in developing countries and developed countries, as well as urban and rural freshwater bodies. After collecting all relevant data on the composition and structure of microplastics, water depth, surface area of the water, distance from the city center, region population, population density, per capita GDP, urbanization rate, and secondary industry structure, the primary influencing factors for the distribution of microplastics were identified using a Pearson correlation analysis, univariate or multiple regression models, and other analytical methods.

Difference Analysis of Microplastic Pollution in Global Freshwater Bodies
Through the collection and collation of the sample data, we obtained the pollution distribution of microplastics in global freshwater bodies ( Figure 1). There were relatively more studies on water microplastics in Asia, Europe, and North America. The order of microplastic pollution in freshwater bodies of six continents was: Asia > North America > Africa > Oceania > South America > Europe. China was the country with the most serious microplastic pollution in the world, with the average concentration of microplastics in the water as high as 3793.38 ± 799.68 n/m 3 . Among Chinese locations, the average abundance of microplastics in Wuhan Bei Lake was the highest, reaching 8925 n/m 3 [18]. Globally, Switzerland had the lowest level of pollution, with an average concentration of 0.04 ± 0.01 n/m 3 . In terms of the concentration of microplastics, the highest was Lake Neuchatel, which was 0.061 n/m 3 [17], and the lowest was Lake Zurich with 0.011 n/m 3 [12].

Difference Analysis of Microplastic Pollution in Developed Countries and Developing Countries
Among the selected samples, 17 waterbodies were located in developed countries, including the United States, Switzerland, Canada, Italy, the United Kingdom, and Australia, and 20 waterbodies were located in developing countries, including Mongolia, South Africa, Chile, Brazil, and China ( Figure 2). Significant statistical difference (Sig) (2-tailed) = 0.001 was obtained using an independent sample T-test of the two groups, indicating a significant difference. The total average abundance of microplastics in the waterbodies of developed countries was 189.2 ± 188.7 n/m 3 , while the total average abundance of microplastics in the waterbodies of developing countries was 3034 ± 725 n/m 3 .
Water 2020, 12, x FOR PEER REVIEW 7 of 16 sample T-test of the two groups, indicating a significant difference. The total average abundance of microplastics in the waterbodies of developed countries was 189.2 ± 188.7 n/m 3 , while the total average abundance of microplastics in the waterbodies of developing countries was 3034 ± 725 n/m 3 . As a developing country, China's waterbodies are polluted by microplastics. In addition to the direct pollution caused by industry and human life, it is also necessary to comprehensively consider the spatial and temporal connections between soil pollution and air pollution. For example, agricultural plastic film is a major source of soil microplastic pollution. The annual growth rate of the area covered by agricultural plastic film in the world was found to be 5.7%. The annual growth rate of the agricultural plastic film coverage area in China from 1991 to 2004 reached as high as 30% [37]. Compared with more than 50% of the global plastic film usage, the plastic film recovery rate in China was less than 2/3 [38]. The residual plastic film can enter waterbodies through the ground rain runoff, resulting in large areas of water with microplastic pollution. There were differences in the composition of microplastics in the waters between developed and developing countries. We selected 16 samples to determine the settling ratio, main microplastic composition, and proportion (Table 3) in the waterbodies. The main difference identified was that the microplastics in the waters from developing countries had more polyethylene terephthalate (PET). As shown in Figure 3, the proportion of light microplastics, like polypropylene (PP) and polyethylene (PE), in waterbodies from developed counties was higher than those from developing countries. The waterbodies from developed countries demonstrated 6.75% expanded polystyrene (EPS), 0.9% polyurethane (PU), and 0.9% acrylonitrile butadiene styrene plastic (ABS), while none of these were recorded in waterbodies from developing countries. As a developing country, China's waterbodies are polluted by microplastics. In addition to the direct pollution caused by industry and human life, it is also necessary to comprehensively consider the spatial and temporal connections between soil pollution and air pollution. For example, agricultural plastic film is a major source of soil microplastic pollution. The annual growth rate of the area covered by agricultural plastic film in the world was found to be 5.7%. The annual growth rate of the agricultural plastic film coverage area in China from 1991 to 2004 reached as high as 30% [36]. Compared with more than 50% of the global plastic film usage, the plastic film recovery rate in China was less than 2/3 [37]. The residual plastic film can enter waterbodies through the ground rain runoff, resulting in large areas of water with microplastic pollution.
There were differences in the composition of microplastics in the waters between developed and developing countries. We selected 16 samples to determine the settling ratio, main microplastic composition, and proportion (Table 3) in the waterbodies. The main difference identified was that the microplastics in the waters from developing countries had more polyethylene terephthalate (PET). As shown in Figure 3, the proportion of light microplastics, like polypropylene (PP) and polyethylene (PE), in waterbodies from developed counties was higher than those from developing countries. The waterbodies from developed countries demonstrated 6.75% expanded polystyrene (EPS), 0.9% polyurethane (PU), and 0.9% acrylonitrile butadiene styrene plastic (ABS), while none of these were recorded in waterbodies from developing countries.   The types and densities of microplastics were different, which directly influenced the locations of microplastics in waterbodies. High-density microplastics were difficult to contain in waters and later became sediments. As a result, there were microplastic distribution variations between the waterbodies and the sediments from developed and developing countries. This section used a Pearson correlation analysis to screen out the major effects (Table 4) on the concentration distribution of microplastics. These compositions, including PET, polyamide (PA), polyvinyl chloride (PVC), PU, ABS, and EPS, had correlations with the concentration distribution of microplastics. In addition, the correlation pecking order for these compositions was PET > PA > PVC = PU = ABS = EPS. The PET contents had the most significant statistical difference (Sig = 0.005) for the concentration distribution of microplastics, and the PET density was as high as 1.38 g/m 3 , which could explain the phenomenon that the settling ratio of microplastics in waters from developing countries was higher than that from developed countries. As shown in Figure 4, there were 31 urban waterbodies in the global sample, and the total average value of microplastic abundance was 1861 ± 520 n/m 3 . The total number of rural waterbodies was six, and the total average value of microplastic abundance was 1038 ± 981 n/m 3 . Sig (2-tailed) = The types and densities of microplastics were different, which directly influenced the locations of microplastics in waterbodies. High-density microplastics were difficult to contain in waters and later became sediments. As a result, there were microplastic distribution variations between the waterbodies and the sediments from developed and developing countries. This section used a Pearson correlation analysis to screen out the major effects (Table 4) on the concentration distribution of microplastics. These compositions, including PET, polyamide (PA), polyvinyl chloride (PVC), PU, ABS, and EPS, had correlations with the concentration distribution of microplastics. In addition, the correlation pecking order for these compositions was PET > PA > PVC = PU = ABS = EPS. The PET contents had the most significant statistical difference (Sig = 0.005) for the concentration distribution of microplastics, and the PET density was as high as 1.38 g/m 3 , which could explain the phenomenon that the settling ratio of microplastics in waters from developing countries was higher than that from developed countries.

Difference Analysis of Microplastic Pollution in Urban and Rural Freshwater Bodies
As shown in Figure 4, there were 31 urban waterbodies in the global sample, and the total average value of microplastic abundance was 1861 ± 520 n/m 3 . The total number of rural waterbodies was six, and the total average value of microplastic abundance was 1038 ± 981 n/m 3 . Sig (2-tailed) = 0.519 was obtained by an independent sample T-test of the two samples, with no significant difference. Considering the differences of environmental governance between developed countries and developing countries, it was of little significance to explore the differences in suburban waterbodies on a global scale. Therefore, the scope of discussion was limited to developed countries and developing countries (Table 5). 0.519 was obtained by an independent sample T-test of the two samples, with no significant difference. Considering the differences of environmental governance between developed countries and developing countries, it was of little significance to explore the differences in suburban waterbodies on a global scale. Therefore, the scope of discussion was limited to developed countries and developing countries (Table 5).  The main difference between urban and rural freshwater bodies is the surrounding population and building density. The rural waterbodies are usually far away from cities with low populations and building density. There may be a correlation between the distance between each waterbody and the city center and the distribution of microplastic pollution. Based on the classification in Table 5, this section explores the above correlations. The determination coefficient R 2 of the linear regression model was close to 0 ( Figure 5), indicating that the distance was not the main effect affecting the concentration of microplastics in freshwater bodies. Due to the independence between the waterbodies (even urban waterbodies), developed countries can reduce the microplastics content in their waterbodies through efficient sewage treatment processes, for instance, Lake Zurich in Switzerland, where the average abundance of microplastics in waterbodies was only 0.11 n/m 3 [14].
On the contrary, for developing countries, even if it is a rural waterbody, if the sewage treatment system is not mature enough to treat the sewage discharged by the surrounding factories, it will cause water pollution more serious than some urban lakes, such as Lake Ulansuhai in China, where the concentration of microplastics in seawater was as high as 5940 n/m 3 [30]. In addition, some of the rural lakes in scenic spots, such as the Three Gorges Reservoir, are frequented by tourists all year  The main difference between urban and rural freshwater bodies is the surrounding population and building density. The rural waterbodies are usually far away from cities with low populations and building density. There may be a correlation between the distance between each waterbody and the city center and the distribution of microplastic pollution. Based on the classification in Table 5, this section explores the above correlations. The determination coefficient R 2 of the linear regression model was close to 0 ( Figure 5), indicating that the distance was not the main effect affecting the concentration of microplastics in freshwater bodies. Due to the independence between the waterbodies (even urban waterbodies), developed countries can reduce the microplastics content in their waterbodies through efficient sewage treatment processes, for instance, Lake Zurich in Switzerland, where the average abundance of microplastics in waterbodies was only 0.11 n/m 3 [14].
On the contrary, for developing countries, even if it is a rural waterbody, if the sewage treatment system is not mature enough to treat the sewage discharged by the surrounding factories, it will cause water pollution more serious than some urban lakes, such as Lake Ulansuhai in China, where the concentration of microplastics in seawater was as high as 5940 n/m 3 [29]. In addition, some of the rural lakes in scenic spots, such as the Three Gorges Reservoir, are frequented by tourists all year round and inevitably influenced by human activities, which can also aggravate the degree of pollution.
Water 2020, 12, x FOR PEER REVIEW 10 of 16 round and inevitably influenced by human activities, which can also aggravate the degree of pollution.
(a) (b) Figure 5. A scatter diagram of distance correlation between the concentration of microplastics in the waterbodies and the distance from the city center: (a) in developed countries; (b) in developing countries.

Natural Factors: Water Depth and Water Area
Water depth and water area are two significant limnological and morphological variables, which are closely related to the water storage capacity and heat capacity of waterbodies. There are differences in the hydrodynamic processes of waterbodies at different water depths. The different flow field distribution and wind and wave characteristics lead to the different self-purification capacities of waterbodies [39]. Therefore, the difference in the average water depth and water area may become factors that affect the concentration of microplastics in waterbodies. We selected the basic information of 28 freshwater bodies and extracted the relevant data of the average water depth and water area of each waterbody. The selected freshwater bodies were located in Europe, Asia, and North America ( Table 6). The average depth of the waterbodies was of the order of Europe > North America > Asia, while the order of the degree of microplastic pollution was exactly the opposite. No obvious correlation law was found between the water area and the water microplastic concentration. Based on this, we explored the influence of the average water depth and water area differences on the water microplastic concentration, and then screened out the main effect. As shown in Table 7, Sig = 0.009 was obtained by the regression model of the average water depth and the water microplastic concentration, indicating that there was a weak correlation. In order to control the interaction of country development level with other variables, we fit the linear model of developed and developing countries, respectively. After fitting, the R 2 and P-value for developed countries were 0.024 and -0.155 and for the developing countries were 0.128 and 0.357, which revealed that the x-value parameter did not support this model ( Figure 6). According to the above analysis, when the water depth was large, although the distribution of the water flow field was different, the self-purification capacity did not change significantly. Especially for developed countries, the concentration of microplastics was lower than for developing countries, so the effect was not obvious.

Natural Factors: Water Depth and Water Area
Water depth and water area are two significant limnological and morphological variables, which are closely related to the water storage capacity and heat capacity of waterbodies. There are differences in the hydrodynamic processes of waterbodies at different water depths. The different flow field distribution and wind and wave characteristics lead to the different self-purification capacities of waterbodies [38]. Therefore, the difference in the average water depth and water area may become factors that affect the concentration of microplastics in waterbodies. We selected the basic information of 28 freshwater bodies and extracted the relevant data of the average water depth and water area of each waterbody. The selected freshwater bodies were located in Europe, Asia, and North America ( Table 6). The average depth of the waterbodies was of the order of Europe > North America > Asia, while the order of the degree of microplastic pollution was exactly the opposite. No obvious correlation law was found between the water area and the water microplastic concentration. Based on this, we explored the influence of the average water depth and water area differences on the water microplastic concentration, and then screened out the main effect. As shown in Table 7, Sig = 0.009 was obtained by the regression model of the average water depth and the water microplastic concentration, indicating that there was a weak correlation. In order to control the interaction of country development level with other variables, we fit the linear model of developed and developing countries, respectively. After fitting, the R 2 and p-value for developed countries were 0.024 and −0.155 and for the developing countries were 0.128 and 0.357, which revealed that the x-value parameter did not support this model ( Figure 6). According to the above analysis, when the water depth was large, although the distribution of the water flow field was different, the self-purification capacity did not change significantly. Especially for developed countries, the concentration of microplastics was lower than for developing countries, so the effect was not obvious.    . Developed countries had a small population, a high level of urbanization and GDP per capita, and the contribution rate of secondary industry was slightly lower than that of developing countries (Table 8). Compared with developing countries, developed countries were generally less polluted by microplastics. In this section, the multiple regression analysis method was used to solve the regression equations, with population, population density, GDP per capita, urbanization level, and secondary industry contribution rate as independent variables, to explore the effects of the above five factors on microplastic pollution. The test result of residual independence was Durbin-Watson (DW) = 2.74, indicating that there was no obvious autocorrelation in the residual sequence. Based on the P-P diagram (Figure 7), the distribution of residual disability was relatively concentrated, and the histogram was in line with the characteristics of normal distribution, indicating that the regression equation model was suitable to  [39]. Developed countries had a small population, a high level of urbanization and GDP per capita, and the contribution rate of secondary industry was slightly lower than that of developing countries (Table 8). Compared with developing countries, developed countries were generally less polluted by microplastics. In this section, the multiple regression analysis method was used to solve the regression equations, with population, population density, GDP per capita, urbanization level, and secondary industry contribution rate as independent variables, to explore the effects of the above five factors on microplastic pollution. The test result of residual independence was Durbin-Watson (DW) = 2.74, indicating that there was no obvious autocorrelation in the residual sequence. Based on the P-P diagram (Figure 7), the distribution of residual disability was relatively concentrated, and the histogram was in line with the characteristics of normal distribution, indicating that the regression equation model was suitable to explain the variation law of microplastic pollution with population, population density, GDP per capita, urbanization level, and the contribution rate of secondary industry.  Based on the analysis results of further research, the fitted equation Sig = 0.002 was obtained, from which the regression equation was judged to be highly significant. The regression equation model indicated that the determination coefficient R 2 = 0.993, indicating that the fitting effect was excellent. The Sig values of the two independent variables of the population scale and per capita GDP were 0.001 and 0.033, respectively (Table 9), which passed the test at the significance level of 5%, while the other three factors failed the significance test. According to the coefficients obtained from the non-standardized coefficient B in the table, the regression equation can be obtained as: y = 0.029x1 + 1.403x2 -0.076x3 -10.279x4 -0.303x5 + 310.453 (1) where x1 is the population (ten thousand people), x2 is the GDP per capita (ten thousand yuan), x3 is the contribution rate of the secondary industry (%), x4 is the urbanization rate (%), and x5 is the population density. Based on the analysis results of further research, the fitted equation Sig = 0.002 was obtained, from which the regression equation was judged to be highly significant. The regression equation model indicated that the determination coefficient R 2 = 0.993, indicating that the fitting effect was excellent. The Sig values of the two independent variables of the population scale and per capita GDP were 0.001 and 0.033, respectively (Table 9), which passed the test at the significance level of 5%, while the other three factors failed the significance test. According to the coefficients obtained from the non-standardized coefficient B in the table, the regression equation can be obtained as: where x 1 is the population (ten thousand people), x 2 is the GDP per capita (ten thousand yuan), x 3 is the contribution rate of the secondary industry (%), x 4 is the urbanization rate (%), and x 5 is the population density. From the signs of the indicators, the coefficient of population number and per capita GDP of the independent variables were positive, indicating that population factors and human living standards were positive for microplastic pollution. With the increase of population and the improvement of living standards, the demand for plastic products increases, and economic development accelerates, leading to the increase in the degree of microplastic pollution. There was no correlation between the population density, the contribution rate of the secondary industry, and the urbanization rate to the concentration of microplastics. Therefore, it was of little significance to discuss them separately. Only their combined effect with other indicators should be considered.

Conclusions
After the above comparative analysis, the conclusions are as follows: • The degree of microplastic pollution in the freshwater bodies of the six continents in the world were ranked as follows: Asia > North America > Africa > Oceania > South America > Europe. China was the most seriously polluted and Switzerland was the least polluted. The pollution levels in developed countries were significantly lower than those in developing countries.

•
The average density of microplastics in the water environments of developed countries was lower than that of developing countries. Therefore, microplastics in the water environment of developed countries did not easily sink, and were mostly stored in waterbodies. In developing countries, microplastics were mostly found in sediments. The geographical location and the size of the waterbodies had no significant influence on the distribution of microplastic pollution, so they were not the primary factors affecting the distribution of microplastic pollution. • The regional differences in the distribution of microplastic pollution may depend on factors such as the population, GDP per capita, national economic production level, the receiving waterbody of sewage, and the city pollution treatment technology. Among them, the population and the GDP per capita were directly proportional to the concentration of microplastics. When the waterbody was used as the receiving waterbody of sewage, it depended on the maturity of the urban sewage treatment technology.
The above conclusions are based on the previous research. The research on microplastics is still not fully mature. The majority of the research objects are large urban waterbodies. There is no uniform measurement method for microplastic content internationally; therefore, it is inevitable that there may be errors when comparing the experimental results. In addition, hydrology was the key element missing from this study, which would require another systematic analysis and be difficult to present in the limited space. In this case, taking the regional differences of microplastic pollution distribution as the starting point, we attempted to use the existing research results to further analyze the fundamental factors affecting microplastic pollution, which has significance as a reference for the formulation of governance strategies.