Comparison of Ground-Based PM2.5 and PM10 Concentrations in China, India, and the U.S.

Urbanization and industrialization have spurred air pollution, making it a global problem. An understanding of the spatiotemporal characteristics of PM2.5 and PM10 concentrations (particulate matter with an aerodynamic diameter of less than 2.5 μm and 10 μm, respectively) is necessary to mitigate air pollution. We compared the characteristics of PM2.5 and PM10 concentrations and their trends of China, India, and the U.S. from 2014 to 2017. Particulate matter levels were lowest in the U.S., while China showed higher concentrations, and India showed the highest. Interestingly, significant declines in PM2.5 and PM10 concentrations were found in some of the most polluted regions in China as well as the U.S. No comparable decline was observed in India. A strong seasonal trend was observed in China and India, with the highest values occurring in winter and the lowest in summer. The opposite trend was noted for the U.S. PM2.5 was highly correlated with PM10 for both China and India, but the correlation was poor for the U.S. With regard to reducing particulate matter pollutant concentrations, developing countries can learn from the experiences of developed nations and benefit by establishing and implementing joint regional air pollution control programs.


Introduction
Rapid economic development and increasing urbanization are partly responsible for rising particulate matter (PM) emissions. Exposure to particulate matter pollution has been associated with increased risks of respiratory, cardiovascular, cerebrovascular, and lung diseases [1][2][3]. According to the World Health Organization (WHO), air pollution has become a global environmental burden, with 92% of the world's population currently living in areas where the air quality level exceeds the WHO guideline level of 10 µg m −3 ; about 3 million annual deaths worldwide are related to outdoor air pollution [4]. Assessing PM concentrations can improve our understanding of the effectiveness of pollution control and quantify its impacts on public health [5], regional visibility [6], and global climate change [7].
In order to prevent further worsening of air pollution, protect public health, and reduce economic losses, many countries have taken significant measures to improve air quality. Developed countries such as the U.S. have a long history of legislation against air pollution. In fact, the country is recognized as a global leader in mitigating air pollution. The U.S. Congress has passed and revised many laws in this regard, including the Air Pollution Control Act, the Clean Air Act, and the Motor Vehicle Air Pollution Control Act, which have laid the foundation for reducing emissions [8]. Although the air quality of most U.S. cities has now reached the recommended WHO level, PM pollution is still a significant concern in some western U.S. cities [9,10]. Moreover, developing countries such as China and India, which currently suffer from serious air pollution, are now beginning to establish and implement environmental management regulations and systems to control and mitigate air pollution. For example, the Chinese State Council and Ministry of Environmental Protection (MEP) promulgated "Ten Actions" in June 2013 and the Air Pollution Prevention Action Plan in September 2013. The urban air quality in the country has since improved [11]. The Indian government has also undertaken many steps, including announcing a National Clean Air Plan (NCAP), launching the National Air Quality Index, and banning the burning of biomass [12]. Although these measures have had some positive effects, more than 75% of Indian cities continue to report PM concentrations above the National Ambient Air Quality Standard. The matter is grave as air pollution reportedly causes about 1.1 million premature deaths each year in India alone [13]. Notably, the existing energy, industrial, and consumption structure of the Chinese and Indian economies cannot be completely reformed within a short period of time, and PM pollution management is a long-term task, particularly in economically developed metropolitan areas, which are likely to bear the brunt of air pollution. Therefore, PM pollution remains the focus of future air pollution prevention efforts in these countries.
Many recent studies have investigated the characteristics of PM pollution in different countries and regions. Previous studies on spatiotemporal PM variations have been based on either aerosol optical depth (AOD) data from ground monitoring [14], satellite observations [15], or visibility observations [16]. Indirect observation methods are often influenced by weather conditions and inversion models and contain large uncertainties. Thus, they cannot directly reflect air pollution status on the ground, leading to lower accuracy than ground monitoring. Ground monitoring provides direct observations of PM 2.5 and PM 10 concentrations (particulate matter with an aerodynamic diameter of less than 2.5 µm and 10 µm, respectively), which better reflect the true atmospheric environment. In China, for example, Wang et al. [17] systematically analyzed the spatiotemporal characteristics of six pollutants in 31 provincial capital cities from March 2013 to February 2014. They found that the PM 2.5 and PM 10 concentrations were higher in cities located in the northern region than those in the western and southeastern regions. Zheng et al. [11] studied the changes of PM 2.5 emissions from 2013 to 2015 and demonstrated that PM 2.5 concentrations in China showed a decreasing trend over the study period. Xie et al. [18] investigated the correlation between six pollutants in 31 provincial capital cities and found a strong correlation between PM 2.5 and PM 10 . In India, the Indo-Gangetic Basin has been identified as having the most severe air pollution [19,20]. PM concentrations have been reported to be very high in Delhi [21], Raipur [22], Kolkata [23], and Kanpur [24]. In comparison, lower concentrations have been reported in Chennai [25] and Mumbai [26]. In recent years, there has been relatively little particulate contamination in the U.S. [27]. Thus, research in this area has been relatively scarce and the current studies mainly focus on areas where PM pollution continues to be relatively high, such as California [28] and Arizona [29]. However, as these studies were mainly concentrated in megacities, the results are fragmented and it is difficult to understand these trends from a macroscopic perspective. Further, although such studies systematically analyzed the temporal and spatial trends of PM concentrations, more in-depth research is needed on the relationships between different particle sizes.
Few studies have presented inter-country comparisons-especially between developed and developing countries-of ground-based PM concentrations. Such comparisons are important as they can help developing countries draw lessons from the experiences of developed countries, and suitable targeted control measures can be planned and implemented accordingly. Thus, in this study, we examined the PM pollution characteristics for selected megacity regions in China, India, and the U.S. from 2014 to 2017 by analyzing the temporal and spatial variations in PM 2.5 and PM 10 concentrations and their trends.

Data
We acquired PM data collected at monitoring stations in China, India, and the U.S. (Figure 1)   In order to ensure the integrity and representativeness of the evaluation, all the data were processed to reject spatial and temporal outliers. As each country maintains a different number of monitoring sites, the 24-h PM concentrations of all sites within that country were averaged to represent the overall PM pollution level. The total days of usable PM 2.5 and PM 10 data were 1461 for both concentrations for China, 1366 and 1434 respectively for India, and 1400 for both concentrations in the U.S.

Methods
To evaluate the air quality status with regard to PM in China, India, and the U.S., we analyzed four years (2014 to 2017) of ambient monitoring data for PM 2.5 and PM 10 in all three countries. To better understand each country's spatial variations, 11 megacity regions were chosen for further analysis: Beijing-Tianjin-Hebei (BTH), the Yangtze River Delta (YRD), the Pearl River Delta (PRD), CY (Cheng-Yu), Delhi, Mumbai, Kolkata, Chennai, Hyderabad, San Diego-Los Angeles-San Francisco (SanSan), and Boston-NewYork-Washington (BosWash). More detailed information regarding megacity regions is given in supplementary Figure S1 and Table S2. These cities are the most economically developed and populated in their respective countries and represent different geographic and meteorological conditions. Pearson correlation analysis (r) was used as an indicator of the relationship between PM 2.5 and PM 10 . The coefficient of variation (CV, namely, the standard deviation divided by the average) was used to describe the degree of spatial variation of PM concentrations in a given area, and it can be expressed as Rate of change (ROC) was used to compare variance in PM 2.5 and PM 10 concentrations [30]: where Y represents the average PM 2.5 or PM 10 concentration in 2014, and X represents the average PM 2.5 or PM 10 concentration in a subsequent year.

Temporal Variations
Tables 1 and 2 summarize the characteristics of PM 2.5 and PM 10 concentrations, respectively, for the three countries from 2014 to 2017. These characteristics reflect the different attitudes of the country's respective governments towards air quality (the annual data are presented in Supplementary Tables S3 and S4). PM levels were found to be lowest in the U.S., followed by China, and India. Although the mean PM 2.5 and PM 10 concentrations exceeded the median and showed a right-skewed distribution for all three countries, the U.S. showed lower PM concentrations and smaller standard deviations.
In the U.S., average daily concentrations of PM 2.5 and PM 10  for all four studied years were still much higher than the IT-1 standard and 5-6 (4-5) times that of the U.S., the fraction of days meeting the standard increased from 30% in 2014 to 46% in 2017, reflecting improved air quality resulting from government measures to mitigate PM pollution.
In India, average daily concentrations of PM 2.5 and PM 10 [19,31]. In contrast, the U.S. followed a different pattern, in which both PM2.5 and PM10 concentrations were relatively high in the summer.   [19,31]. In contrast, the U.S. followed a different pattern, in which both PM 2.5 and PM 10 concentrations were relatively high in the summer.  Figure 4 and Figure S2 show the spatial distributions of annual PM2.5 and PM10 concentrations, respectively, in China, India, and the U.S. The spatial patterns of PM2.5 and PM10 in China were similar, exhibiting a structure that is dispersed at large scales but aggregated at small scales. This result agrees with that of Cao et al. [32]. During the study period, higher PM2.5 and PM10 concentrations were generally observed in the North China Plain (e.g., BTH), the East China Plain (e.g., YRD), the Sichuan Basin (e.g., CY), and the Taklimakan Desert. Low PM2.5 and PM10 concentrations were found in the southwest (e.g., Tibet and Yunnan) and southeast (e.g., PRD, Hainan, and Taiwan). However, PM10 pollution was heavy in parts of the northwestern regions whereas the PM2. In India, the spatial distributions of both PM2.5 and PM10 were highest in the north and lowest in the south, in agreement with the findings of Van et al. [33]. The regions with high PM2.5 and PM10 concentrations were mainly located in the Indo-Gangetic Plain; Delhi, Haryana, Uttar Pradesh, and Bihar are the most polluted states in India. PM concentrations were relatively low in most southern Indian states such as Kerala, Karnataka, and Tamil Nadu, which benefitted from the effects of landsea breezes. From 2014 to 2017, the PM2.5 and PM10 concentrations in the most polluted cities continuously worsened due to rapid urbanization and industrialization and weak environmental enforcement [34]. However, most of southern India seemed to comply with the WHO's IT-1 standard.  Figure 4 and Figure S2 show the spatial distributions of annual PM 2.5 and PM 10 concentrations, respectively, in China, India, and the U.S. The spatial patterns of PM 2.5 and PM 10 in China were similar, exhibiting a structure that is dispersed at large scales but aggregated at small scales. This result agrees with that of Cao et al. [32]. During the study period, higher PM 2.5 and PM 10 concentrations were generally observed in the North China Plain (e.g., BTH), the East China Plain (e.g., YRD), the Sichuan Basin (e.g., CY), and the Taklimakan Desert. Low PM 2.5 and PM 10 concentrations were found in the southwest (e.g., Tibet and Yunnan) and southeast (e.g., PRD, Hainan, and Taiwan). However, PM 10  In India, the spatial distributions of both PM 2.5 and PM 10 were highest in the north and lowest in the south, in agreement with the findings of Van et al. [33]. The regions with high PM 2.5 and PM 10 concentrations were mainly located in the Indo-Gangetic Plain; Delhi, Haryana, Uttar Pradesh, and Bihar are the most polluted states in India. PM concentrations were relatively low in most southern Indian states such as Kerala, Karnataka, and Tamil Nadu, which benefitted from the effects of land-sea breezes. From 2014 to 2017, the PM 2.5 and PM 10 concentrations in the most polluted cities continuously worsened due to rapid urbanization and industrialization and weak environmental enforcement [34]. However, most of southern India seemed to comply with the WHO's IT-1 standard. During the study period, all of the stations in the U.S. had the annual PM2.5 and PM10 concentrations less than WHO's IT-1 guideline, and most studied locations were below WHO's IT-3 level. Higher concentrations were found along the west coast (e.g., SanSan and Alaska) and the central Midwest (e.g., Indiana and Illinois), while low concentrations were generally observed in most other parts of the U.S. This finding is in agreement with that of Van et al. [15]. From 2014 to 2017, the number of regions with relatively high PM2.5 and PM10 concentrations decreased substantially, although California continued to record higher levels of PM pollution.

Correlation between PM2.5 and PM10 Concentrations
Correlations between changes in different particulate sizes in the same region are an important indicator of regional particulate pollution [35]. Therefore, we calculated the Pearson correlation coefficients (r) between PM2.5 and PM10 concentrations for all three countries over the study period ( Figure 5). Daily PM2.5 and PM10 concentrations and ratios are presented in Figure 6. In China (Figure  5a), PM2.5 was highly correlated (r = 0.92) with PM10, similar to the correlation (0.94) found by Zhou During the study period, all of the stations in the U.S. had the annual PM 2.5 and PM 10 concentrations less than WHO's IT-1 guideline, and most studied locations were below WHO's IT-3 level. Higher concentrations were found along the west coast (e.g., SanSan and Alaska) and the central Midwest (e.g., Indiana and Illinois), while low concentrations were generally observed in most other parts of the U.S. This finding is in agreement with that of Van et al. [15]. From 2014 to 2017, the number of regions with relatively high PM 2.5 and PM 10 concentrations decreased substantially, although California continued to record higher levels of PM pollution.

Correlation between PM 2.5 and PM 10 Concentrations
Correlations between changes in different particulate sizes in the same region are an important indicator of regional particulate pollution [35]. Therefore, we calculated the Pearson correlation coefficients (r) between PM 2.5 and PM 10 concentrations for all three countries over the study period ( Figure 5). Daily PM 2.5 and PM 10 concentrations and ratios are presented in Figure 6. In China (Figure 5a), PM 2.5 was highly correlated (r = 0.92) with PM 10 , similar to the correlation (0.94) found by Zhou et al. [36]. Approximately 62% of PM 2.5 values were above the IT-1 standard as compared to 55.24% for PM 10 , suggesting that PM 2.5 pollution was more serious. Furthermore, if the concentration of one pollutant reached a problematic level, the concentrations of the pollutant usually increased as well. For example, 53.32% of PM 10 values exceeded this standard when PM 2.5 concentrations exceeded the standard, while 54.21% of PM 2.5 concentrations exceeded the standard when PM 10 concentrations exceeded the standard (Figure 6a). As per Figure 5b, the correlation coefficient between PM 2.5 and PM 10 was 0.64 for India. It was noted that 84.40% of PM 2.5 values were above the IT-1 standard as compared to 91.70% for PM 10 , suggesting that PM 10 pollution was more serious. This result agrees with that of a Global Burden of Disease (GBD) report, which found that India's PM 10 pollution is more widespread compared to PM 2.5 [13]. Furthermore, 78.83% of PM 10 concentrations exceeded this standard when PM 2.5 concentrations did not meet the standard, while 75.09% of PM 2.5 concentrations exceeded the standard when PM 10 concentrations did not comply with it (Figure 6b). The simultaneous variation in differently sized PM provides important evidence pertaining to the regional characteristics of PM pollution [35,37]. These results suggest that PM pollution in China and India have similar regional characteristics. Finally, extremely weak correlations (0.20) were detected between PM 2.5 and PM 10 for the U.S. (Figure 5c), indicating no obvious synchronous variation between the two concentrations.  (Figure 6a). As per Figure 5b, the correlation coefficient between PM2.5 and PM10 was 0.64 for India. It was noted that 84.40% of PM2.5 values were above the IT-1 standard as compared to 91.70% for PM10, suggesting that PM10 pollution was more serious. This result agrees with that of a Global Burden of Disease (GBD) report, which found that India's PM10 pollution is more widespread compared to PM2.5 [13]. Furthermore, 78.83% of PM10 concentrations exceeded this standard when PM2.5 concentrations did not meet the standard, while 75.09% of PM2.5 concentrations exceeded the standard when PM10 concentrations did not comply with it ( Figure 6b). The simultaneous variation in differently sized PM provides important evidence pertaining to the regional characteristics of PM pollution [35,37]. These results suggest that PM pollution in China and India have similar regional characteristics. Finally, extremely weak correlations (0.20) were detected between PM2.5 and PM10 for the U.S. (Figure 5c), indicating no obvious synchronous variation between the two concentrations.

PM2.5 and PM10 Concentrations in Megacity Regions
We compared PM concentrations in 11 megacity regions of China, India, and the U.S. ( Figure  S3). There were significant differences in PM concentrations, with the levels being generally much lower in the U.S. as it is more developed compared to cities in China and India. Delhi and Kolkata were the most polluted cities, followed by BTH, the YRD, and Mumbai, while the two megacity regions of SanSan and BosWash in the U.S. were the cleanest. Spatial variations in the annual average PM concentrations (Table 3)

PM 2.5 and PM 10 Concentrations in Megacity Regions
We compared PM concentrations in 11 megacity regions of China, India, and the U.S. (Figure S3). There were significant differences in PM concentrations, with the levels being generally much lower in the U.S. as it is more developed compared to cities in China and India. Delhi and Kolkata were the most polluted cities, followed by BTH, the YRD, and Mumbai, while the two megacity regions of SanSan and BosWash in the U.S. were the cleanest. Spatial variations in the annual average PM concentrations (Table 3)

Discussion
This study investigated the spatial and temporal features of PM2.5 and PM10 concentrations in China, India, and the U.S. Developing countries suffer from serve PM pollution as they undergo rapid industrialization and urbanization. Most developed countries, such as the U.S., have already experienced this phase. The PM pollution can be attributed to increasing fossil fuel use by powerheavy industries, construction, vehicle exhaust, and biomass combustion [19,40,41]. China and India show significant seasonal variations in PM concentrations, the highest being in winter and the lowest in summer, while the opposite situation is noted for the U.S. This inconsistency is most likely due to the different seasonal variations of PM emissions and meteorological conditions in these countries

Discussion
This study investigated the spatial and temporal features of PM 2.5 and PM 10 concentrations in China, India, and the U.S. Developing countries suffer from serve PM pollution as they undergo rapid industrialization and urbanization. Most developed countries, such as the U.S., have already experienced this phase. The PM pollution can be attributed to increasing fossil fuel use by power-heavy industries, construction, vehicle exhaust, and biomass combustion [19,40,41]. China and India show significant seasonal variations in PM concentrations, the highest being in winter and the lowest in summer, while the opposite situation is noted for the U.S. This inconsistency is most likely due to the different seasonal variations of PM emissions and meteorological conditions in these countries [42,43]. For example, in China and India, the highest concentrations in winter coincide with massive emissions from fossil fuel and biomass burning for heating in winter [44][45][46]. Furthermore, more frequent occurrences of adverse meteorological conditions (e.g., slow winds, high humidity, and lower boundary layer height) may promote the accumulation of PM pollutants in these countries [47]. For the U.S., both PM 2.5 and PM 10 concentrations were relatively high in the summer, which can be attributed to the more frequent wildfire activity in this period [48]. In terms of spatial distribution, PM concentrations are generally higher in the north than in the south, and they are lower in the coastal regions compared to the inland regions in China and India. The most polluted regions were observed in the BTH and the Indo-Gangetic Plain. Similar conclusions have been reported in previous studies [33,44]. In the U.S., most cities, except for parts of the western regions [15], met the IT-1 standard for PM concentrations. This result indicates that different control strategies should be targeted for different regions, depending on regional emissions and meteorological characteristics [49]. Strong correlations between PM 2.5 and PM 10 concentrations were noted for China (0.92) and India (0.64), which could be explained by the similar sources of PM 2.5 and PM 10 pollutants, mainly emissions from combustion sources (e.g., coal, vehicle exhaust, and biomass) and dust [50,51]. However, weak correlations were detected in this regard for the U.S. (0.20). The main likely reason for this difference is that these pollutants arise from different sources. The latest version of the EPA's National Emissions Inventory (NEI) concerning PM sourcing and elemental characteristics demonstrated that fires (i.e., wildfires, prescribed fires, and agricultural burning) are larger contributors to primary PM 2.5 concentrations while dust (i.e., unpaved road dust, construction dust, and paved road dust) contributed largely to PM 10 concentrations [52]. These results suggest that PM pollution in China and India has similar regional characteristics, and thus, regional collaboration in PM pollution control and mitigation could be an important approach to enhance urban and regional air quality. China has successful experiences in air pollution reduction due to its policy of regional collaboration. For example, with the establishment and optimization of the regional coordination mechanism for the BTH region since 2013, the air quality in this region has improved dramatically. Annual mean PM 2.5 (PM 10 ) concentrations of the BTH region decreased by 33% (34%) in 2016 compared to that of 2013 [53].
From 2014 to 2017, the PM 2.5 and PM 10 concentrations showed a decreasing trend in China and the U.S., and an increasing trend (albeit with some fluctuations) in India. Legislation plays a crucial role in air pollution management [54]. Table S5 lists the main legislations on air quality protection adopted by China, India, and the U.S. Many cities in the U.S. (e.g., Donora in Pennsylvania) experienced severe air pollution in the 1940s-1950s, causing serious health crises [55]. In response, the U.S. Congress passed the Clean Air Act and its subsequent amendments, and made remarkable progress in improving the nation's air quality. This can be evidenced by the fact that the combined emissions of the six common pollutants (PM 2.5 and PM 10 , SO 2 , NO x , volatile organic compounds, carbon monoxide, and lead) dropped by 73% from 1970 to 2016 [39]. China enacted the Air Pollution Prevention and Control Law in 1987, and amended it in 1995 and 2000. However, these laws were quite basic and enforcement was difficult due to inadequate financial and human resources. Thus, efforts to improve air quality in developing countries are typically ineffective [56]. Given the worsening air pollution and rising public concern, however, the Chinese government has successfully enacted more stringent and enforceable laws such as the Action Act in 2013 and the Air Pollution Prevention and Control Law in 2015. Thus, PM pollution across China, especially in its most polluted areas, has been significantly reduced [11,36]. However, in India, the laws regarding air pollution control (e.g., the Air (Prevention and Control of Pollution) Act) are outdated and weak. Moreover, while these laws cover the theoretical requirements for pollution control and mitigation, they have failed in terms of implementation [57].
Thus, PM pollution in India continues to be severe despite the enactment of various air pollution control regulations.

Conclusions
In this paper, we analyzed the spatiotemporal characteristics, change rates, and correlations of PM 2.5 and PM 10 values in China, India, and the U.S. Further, we discussed the main reason for the differences in PM pollution characteristics among the three countries and the impacts of their air pollution control laws on PM concentrations. Our main conclusions are as follows. PM levels in China and India were much worse than those in the U.S. from 2014 to 2017. A strong seasonal trend was observed in China and India, with the highest values occurring in winter and the lowest in summer. In contrast, PM pollutant distribution in the U.S. showed relatively high concentrations in summer and the relatively low values in winter. The most heavily polluted areas included the North China Plain and the Indo-Gangetic Plain, while less polluted areas included southwestern and southeastern China and southern India. PM concentrations in most of the studied regions in the U.S. met WHO standards; higher concentrations of PM 2.5 and PM 10 were found in California and the central Midwest. Although both PM 2.5 and PM 10 concentrations declined gradually in the U.S. and China-particularly in the severely polluted regions of China-during the study period, no comparable decrease occurred in India. In addition, PM 2.5 and PM 10 concentrations were highly correlated in both China and India but poorly correlated in the U.S. China and India still record high levels of PM pollution; however, both countries can learn from the experiences of the U.S. pertaining to air pollution laws and management. These results also suggest that regional joint prevention and control over air pollution is an important approach to improve the ambient air quality in cities and regions in China and India.
Our research contributes to the field as it increases understanding of the status and variations of PM 2.5 and PM 10 concentrations in China, India, and the U.S. However, our study also has several limitations. For example, the data used in this study were sourced from unevenly distributed ground stations in these countries. This poses a potential threat to the validity of the results. Moreover, long-term inter-country comparisons among countries are possible using satellite-derived PM data despite the uncertainties associated with them. Thus, in the future, more in-depth inter-nation comparisons should be made using satellite-derived PM concentrations in conjunction with ground-based measurements.
Supplementary Materials: The following are available online at http://www.mdpi.com/1660-4601/15/7/1382/ s1, Table S1: Summary of datasets applied in this study, Table S2: Basic information of megacity regions in China, India, and the U.S., Table S3: Summary of PM 2.5 concentrations in China, India, and the U.S., Table S4: Summary of PM 10 concentrations in China, India, and the U.S., Table S5: Main legislations on air protection adopted in China, India, and the U.S., Figure S1: Locations of megacity regions in China, India, and the U.S., Figure S2: Spatial distribution of PM 10