key: cord-0017027-dt5ica03
authors: Zhang, Shichun; Tong, Daniel Q.; Dan, Mo; Pang, Xiaobing; Chen, Weiwei; Zhang, Xuelei; Zhao, Hongmei; Wang, Yiyong; Shang, Bingnan
title: Chemical Composition and Source Apportionment of Wintertime Airborne PM(2.5) in Changchun, Northeastern China
date: 2021-04-20
journal: Int J Environ Res Public Health
DOI: 10.3390/ijerph18084354
sha: 036e6267c9286215626b8d93299ef5027db498b9
doc_id: 17027
cord_uid: dt5ica03

This study presents field observations and laboratory analyses of wintertime airborne particulate matter (PM(2.5)) and its chemical components in the Changchun metropolitan area, the geographical center of northeastern China. Twenty-four hour PM(2.5) filter samples were collected from 23 December 2011 to 31 January 2012 at four sites in the types of traffic, residential, campus, and a near-city rural village, respectively. Daily PM(2.5) concentrations ranged from 49 to 466 µg m(−3), with an arithmetic average of 143 µg m(−3). Laboratory analyses showed that among all measured chemical species, mineral dust contributed the largest proportion (20.7%) to the total PM(2.5) mass, followed by secondary inorganic aerosols (SIA, including SO(4)(2−), NO(3)(−) and NH(4)(+)), which constituted 18.8% of PM(2.5) mass. Another notable feature of PM(2.5) chemical composition was high halogen (Cl(−) and F(−)) loadings at all sites, which was likely due to emissions from coal combustion, plastic manufacturing, and glass melting. Among the four sampling sites, the suburban site exhibited the highest PM(2.5) levels and extremely high Cl(−) and F(−) loadings due to residential wood burning and nearby industrial facilities lacking effective emission controls. Our results report one of the earliest observations of PM(2.5) composition in this region, providing a baseline of aerosol profiles of aerosol before PM(2.5) was routinely measured by environmental protection agencies in China, which could be useful for assessing long-term trends of air quality and effectiveness of mitigation measures.

Exposure to outdoor air pollution causes 4.2 million premature deaths worldwide every year, making air pollution the single largest environmental risk today [1] . Among all pollutants, ozone (O 3 ) and PM 2.5 (particulate matter with diameter <2.5 µm −3 ) are most closely associated with increased occurrence of lung cancer and cardiovascular diseases [2, 3] . A large portion of the global health burden is associated with certain Asian countries, where high population densities correlate with elevated air pollution levels [4] . Elevated levels of PM 2.5 , resulting from rapid economic growth and urbanization in past decades [5] , are common in many regions of China [6] [7] [8] [9] [10] , posing an unprecedented threat to regional air quality and human health [11, 12] . degrading air quality when combined with intense local emissions, occurs frequently in winter [32] .

In China, coal combustion has been identified as one of the largest contributors to air pollution [33] . In Changchun in 2011, a total of 25.8 million tons of raw coal were consumed, among which 13.8 million tons were consumed by thermal power generation and another 4.8 million tons were consumed by wintertime domestic heating [22] . The major industries in Changchun include the automotive industry, coal-fired power plants, metallurgy, iron and steel mills, machine manufacturing, and electric and electronics manufacturing, and most of these industries are located in the western and northern portions of Changchun [34] .

A field campaign was carried out from 23 December 2011 to 31 January 2012. Ambient air was sampled at four sites in Changchun (Figure 1 ). These sites included the following locations: (1) the Chinese Academy of Sciences Institute of Geography and Agroecology (traffic site, marked as S1) located in the southern part of the Changchun industrial zone;

(2) the Liaoyang Residential Area in the city center (residential site, S2); (3) the Jilin Agricultural University (campus site, S3) site at the southeastern corner of Changchun; and (4) the Guangning Village site (suburban village, S4), a suburban community north of Changchun. The locations of these sites are depicted in Figure 1 , and detailed descriptions and the meteorological variables of these sites are provided in Table 1 and Figure 2 , respectively. a mean wind speed of 4.3 m/s. Temperature inversion, a meteorological condition conducive to degrading air quality when combined with intense local emissions, occurs frequently in winter [32] . In China, coal combustion has been identified as one of the largest contributors to air pollution [33] . In Changchun in 2011, a total of 25.8 million tons of raw coal were consumed, among which 13.8 million tons were consumed by thermal power generation and another 4.8 million tons were consumed by wintertime domestic heating [22] . The major industries in Changchun include the automotive industry, coal-fired power plants, metallurgy, iron and steel mills, machine manufacturing, and electric and electronics manufacturing, and most of these industries are located in the western and northern portions of Changchun [34] .

A field campaign was carried out from 23 December 2011 to 31 January 2012. Ambient air was sampled at four sites in Changchun (Figure 1 ). These sites included the following locations: (1) the Chinese Academy of Sciences Institute of Geography and Agroecology (traffic site, marked as S1) located in the southern part of the Changchun industrial zone; (2) the Liaoyang Residential Area in the city center (residential site, S2); (3) the Jilin Agricultural University (campus site, S3) site at the southeastern corner of Changchun; and 4) the Guangning Village site (suburban village, S4), a suburban community north of Changchun. The locations of these sites are depicted in Figure 1 , and detailed descriptions and the meteorological variables of these sites are provided in Table 1 and Figure 2 , respectively. 

Daily 24-h PM 2.5 samples were collected at the four sampling sites on 47 mm Whatman Teflon filters using BGI samplers (Model Omni, BGI Inc., Waltham, MA, USA) at a flow rate of 5 L/min at about 9 AM local time. The filters were equilibrated in a drying apparatus (temperature 20-25 • C; RH 35-45%) for 48 h and then weighed with an electronic balance (Model XS105DU, Mettler Toledo Inc., Greifensee, Switzerland, reading precision 10 µg) before and after sampling. Sampling time, air temperature and air pressure were automatically recorded, and the standard sampling air volumes (V std ) were calculated from these parameters recorded by the air samplers. The PM 2.5 mass concentrations were calculated using the PM 2.5 mass divided by V std . Samples were stored below 4 • C in a refrigerator until analysis to minimize evaporation of volatile components. A total of 129 filter samples were collected and weighed for bulk PM 2.5 mass over the whole campaign. However, 19 of the samples were accidentally lost before the laboratory analysis. Consequently, the remaining 110 samples were analyzed for chemical compositions. For data quality assessment and control (QA/QC), parallel sampling, a blank test, and a standard curve were performed during the outdoor sampling and laboratory analysis. For one half of each Teflon filter sample and field blank filter was extracted with 10 mL ultrapure water in an ultrasonic bath for 30 min. After passing through microporous membranes (pore size 0.45 mm, diameter 25 mm, Auto Science (Tianjin) Instrument Co., LTD, Tianjin, China), the extracts were stored at 4 • C in a precleaned tube for analysis. Four anions (F − , Cl − , NO 3 − and SO 4 2− ) and 5 cations (Na + , NH 4 + , K + , Mg 2+ and Ca 2+ ) were analyzed by Ion chromatography (ICS-1000, Dionex Inc., Sunnyvale, CA, USA), which consists of a separation column (Dionex Ionpac AS14 for anion and CS12A for cation) and a guard column (Dionex Ionpac AG14 for anion and AG12A for cation). A weak base eluent (3.5 mmol Na 2 CO 3 /1.0 mmol NaHCO 3 solution) was used for anion detection, and 20 mmol methane sulfonic acid (MSA) solution was used for cation detection. Blank values were subtracted from samples. The recovery rates of the measured ions were in the range of 80-120%. The limits of detection are in Table S1 , and uncertainties are ±10% for all ions. 

Daily 24-h PM2.5 samples were collected at the four sampling sites on 47 mm Whatman Teflon filters using BGI samplers (Model Omni, BGI Inc., Waltham, MA, USA) at a flow rate of 5 L/min at about 9 AM local time. The filters were equilibrated in a drying apparatus (temperature 20-25 °C; RH 35%-45%) for 48 h and then weighed with an electronic balance (Model XS105DU, Mettler Toledo Inc., Greifensee, Switzerland, reading precision 10 µ g) before and after sampling. Sampling time, air temperature and air pressure were automatically recorded, and the standard sampling air volumes (Vstd) were calculated from these parameters recorded by the air samplers. The PM2.5 mass concentrations were calculated using the PM2.5 mass divided by Vstd. Samples were stored below 4 °C in a refrigerator until analysis to minimize evaporation of volatile components. Another half of each Teflon filter sample was dissolved at 170 • C for 4 h in a highpressure Teflon digestion tank with mixed acid composed of 3 mL of concentrated HNO 3 , 1 mL of concentrated HClO 4 , and 1 mL of concentrated HF. After the solutions were cooled, 0.1 mL of concentrated HNO 3 and deionized water were added to dilute the digestion liquid to 10 mL. A total of 18 elements (Al, Si, Ca, Fe, Mg, K, Mn, Ni, Cu, Zn, As, Se, Sr, Ba, Cd, Cr, Nd, and Pb) were measured using an inductively coupled plasma-atomic emission spectrometer (ICPE9000, Shimadzu Co. Ltd., Kyoto, Japan). The sample solutions were measured in triplicate, and standard and blank samples were inserted and analyzed every 10 samples for quality control.

Meteorological data, including ambient temperature (T a ), pressure (P a ), and relative humidity (RH), were measured using an automatic meteorological station (Model AMS-II-N, Changchun Meteorological Instrument Research Institute, Changchun, China) at the IGA site. Wind speed (W s ) and wind direction (W d ) were downloaded from a weather data archive (http://www.wunderground.com/history; accessed on 16 April 2021).

In this study, spatial variability of PM 2.5 among the sampling sites were tested through coefficient of divergence (COD) analysis:

where x ij and x ik represent the 24-h average PM 2.5 concentration for day i at site j and site k, and p is the number of observations. A value of 0 for the COD implies that two sites are exactly alike, and a value larger than 0 implies that the two sites are dissimilar. Criteria of COD > 0.2 indicating heterogeneity and <0.1 indicating homogeneity were selected for the spatial variability analysis of PM 2.5 [10, 35] .

Here, the enrichment factor (EF) of the chemical elements was calculated to estimate the relative contributions from natural and anthropogenic sources [10, 36, 37] .

where (X/Al) aerosol is the ratio of measured concentration of the element of interest (X) to aluminum (Al) in PM 2.5 ; (X/Al) crust is the ratio of X to Al in crustal material. The information of (X/Al) crust was based on [38] .

If the EF values of elements of interest are below 10, their sources are identified as predominantly natural sources; if the EF values are between 10 and 100, their sources are believed to be both natural and anthropogenic origins; and if the EF values are larger than 100, the dominance of anthropogenic are confirmed [36] .

The daily average PM 2.5 mass concentrations were 122.7 ± 30.9 µg m −3 , 136.7 ± 60 µg m −3 , 137.2 ± 50.1 µg m −3 and 170.6 ± 93.1 µg m −3 at sites S1, S2, S3 and S4, respectively, with an arithmetic average of 143.5 ± 68.1 µg m −3 for all sites ( Figure 2 and Table 2 ). At all sites, the PM 2.5 levels consistently exceeded the World Health Organization (WHO) air quality guidelines for PM 2.5 (25 µg m −3 ) and China's Level I and II air quality standards for PM 2.5 (35 and 75 µg m −3 , respectively). Our results show that although Changchun is not known as one of the most polluted cities in China, PM 2.5 pollution in Changchun is comparable to PM 2.5 pollution levels in severely polluted Chinese cities such as Beijing [10] , Shanghai [39] and Guangzhou [9] , as well as many other polluted Asian cities ( Table 3) .

As shown in Figure 2 and Table 2 , large spatial variabilities in PM 2.5 concentrations existed among the sampling sites. Coefficient of divergence (COD) analysis showed that the COD values of the site-pairs among the S1, S2, and S3 sites were lower than 0.2, indicating the spatial variabilities among these sites were small. However, the COD values between the S4 site and other three sites were all larger than 0.2 (Table S2) , indicating the S4 site was a distinct site among the four sites. 

37 129 a The mineral dust in PM-2.5 was calculated as the sum of oxides of Al, Si, Ca, Fe, Mg, and K (i.e., Mineral dust = 1.89Al +2.14Si +1.40Ca +1.43Fe +1.66Mg +1.21K) [40] , of which the concentration of Si was estimated according to the average ratio of Si/Al (3.6) in earth's crust [40] [41] [42] Among the four sampling sites, the suburban (S4) site exhibited considerably higher PM 2.5 concentrations than other sites ( Figure 2 and Table 2 ). The higher PM 2.5 level at the S4 site was likely due to the lack of effective emission control measures for local emission sources, e.g., coal and biomass burning for household heating and cooking, fugitive dust from unpaved surfaces, and processes at nearby individual workshops.

During the study period, heavy haze events with PM 2.5 levels higher than 150 µg m −3 frequently occurred, as observed on Dec 30 and 31 in 2011 at all the four sites, 26 January 2012 at S2 site, 25 June 2012 at S3 and S4 sites, and 7, 10, 18, 19 January 2012 at S4 site ( Figure 2 ). A few relatively clean days as observed at S2 and S4 sites were associated with the Chinese New Year (also known as "Spring Festival"), with the lowest PM 2.5 concentration observed on 21 January 2012, which was Chinese New Year's Eve.

On 18 and 19 January, a heavy haze occurred simultaneously at S1, S3 and S4 site with PM 2.5 values nearly or higher than 200 µg m −3 . Meteorological analysis ( Figure 2 ) showed a sudden decrease of wind speeds from 2.8 m/s to 0.8 m/s on 17 January. Meanwhile, the 48-h back trajectory analysis using the HYSPLIT model [52, 53] clearly shows that the air mass switched suddenly on 17 January from long range transport from southeast direction to local origins, as could be easily seen by the swift of Above Ground Level (AGL) of the air mass reaching the sampling site ( Figure 3 ). From January 17 to 20, the static weather condition, coupled with relatively high air temperature and relative humidity, may have facilitated the accumulation of air pollutants and generation of secondary aerosols, resulting in the dramatic increasing of PM 2.5 concentrations. Over the whole campaign, correlation analysis ( Figure S1 ) indicated that PM 2.5 showed strong negative relationship with visibility, suggesting important influences of PM 2.5 on haze pollution. Besides, PM 2.5 showed an obvious negative correlation with wind speed owing to the dilution effects of wind on PM 2.5 in the atmosphere. However, no obvious correlations were observed between PM 2.5 and air temperature, relative humidity, dew point temperature and air pressure. Further analysis showed that the heavy hazy days were usually associated with winds from the directions of potential main emission sources as depicted in Table 1 (also see Figure S2 ). There is a distinctive temporal changes in PM2.5 at the residential (S2) site compared to other sites ( Figure 2 ). The PM2.5 concentrations at the S1 and S4 sites decreased remarkably after 21 January, which was possibly related to the reduction of secondary inorganic aerosols (Figure 4d ). However, the PM2.5 levels at site S2 increased noticeably during the same period, largely due to the increase of mineral dust ( Figure 4b ). As shown in Figure  5 , the elements related to crustal (e.g., Al, K, and Sr), motor vehicles (e.g., Pb and Zn), coal combustion (e.g., As) and fireworks (e.g., Sr) also increased considerably after 21 January, indicating that human activities in the residential area during the Chinese Spring Festival celebration may have appreciably contributed to local PM2.5 pollution at the S2 site, regardless of the intrusion of cleaner background air. There is a distinctive temporal changes in PM 2.5 at the residential (S2) site compared to other sites (Figure 2 ). The PM 2.5 concentrations at the S1 and S4 sites decreased remarkably after 21 January, which was possibly related to the reduction of secondary inorganic aerosols (Figure 4d ). However, the PM 2.5 levels at site S2 increased noticeably during the same period, largely due to the increase of mineral dust ( Figure 4b ). As shown in Figure 5 , the elements related to crustal (e.g., Al, K, and Sr), motor vehicles (e.g., Pb and Zn), coal combustion (e.g., As) and fireworks (e.g., Sr) also increased considerably after 21 January, indicating that human activities in the residential area during the Chinese Spring Festival celebration may have appreciably contributed to local PM 2.5 pollution at the S2 site, regardless of the intrusion of cleaner background air.

At all sites, mineral dust was the most abundant chemical component, contributing 13.6 µg/m 3 to 75.9 µg/m 3 to PM 2.5 mass concentration, with an arithmetic average of 30.4 ± 11.4 µg/m 3 ( Table 2 ). Here the mineral dust in PM 2.5 was calculated as the sum of Al, Si, Ca, Fe, Mg and K oxides (i.e., mineral dust = 1.89 Al +2.14 Si +1.40 Ca +1.43 Fe +1.66 Mg +1.21 K) [40] , assuming that the elements existed as oxides [54] . In this study, the mineral levels were much comparable to those reported in other Chinese cities (e.g., 26 .5 µg/m 3 in Beijing [10] ) and constituted the largest portion (20.7%) of total PM 2.5 mass, indicating prevailing emission of soil dust particles in Changchun. The secondary inorganic aerosols (SIA, including SO 4 2− , NO 3 − and NH 4 + ), with a mean of 27.6 ± 5.2 µg/m 3 , constituted the second largest portion (18.8%) of the total PM 2.5 mass. This percentage was lower than some reported values in Beijing (28%) [8] , Tianjin (34.1%) [36] , and Guangzhou (48.5%) [55] , but similar to or even higher than those in some other studies conducted in wintertime Beijing (12.0% in [10] and 20.5% in [55] ), Harbin (13.7%) [56] , and Changchun (18.7%) [31] . We suggest that the relatively low SIA/PM 2.5 mass ratios in this study might indicate that the low temperatures in such cold regions have suppressed secondary inorganic formation in winter.

Another notable feature of the chemical characteristics of PM 2.5 in the study period is high Cl − (7.1 ± 3.5 µg/m 3 ) and F − (4.1 ± 4.0 µg/m 3 ) loadings at all sites. To our knowledge, the observed F − content (7.3 ± 4.2 µg/m 3 ) at the suburban site (S4) was the highest value ever reported at an urban environment based on literatures. F − concentration typically ranges from 0 to 2 µg/m 3 [57, 58] . Cl − and F − are usually emitted from coal combustion [57] , while Cl − can also be emitted from biomass burning in NE China [59] . Although coal burning is important in Changchun due to domestic heating, the unusually high levels of Cl − and F − concentrations could not be accounted for by coal burning alone because other co-emitted components [60] were not elevated proportionally. After examining the local emission sources, we speculate that the high levels of F − observed in this study can be attributed to emissions from plastic-making and specialized glassmaking facilities near the S4 site, where glass products were treated with an F solution for embroidery designs. Additionally, NH 4 + at S4 was the highest among the four sites, likely due to its proximity to NH 3 emission sources such as poultry farms. 

At all sites, mineral dust was the most abundant chemical component, contributing 13.6 µ g/m 3 to 75.9 µ g/m 3 to PM2.5 mass concentration, with an arithmetic average of 30.4 ± 11.4 µ g/m 3 ( Table 2 ). Here the mineral dust in PM2.5 was calculated as the sum of Al, Si, Ca, Fe, Mg and K oxides (i.e., mineral dust = 1.89 Al +2.14 Si +1.40 Ca +1.43 Fe +1.66 Mg +1.21 K) [40] , assuming that the elements existed as oxides [54] . In this study, the mineral levels were much comparable to those reported in other Chinese cities (e.g., 26.5 µ g/m 3 in Beijing [10] ) and constituted the largest portion (20.7%) of total PM2.5 mass, indicating prevailing emission of soil dust particles in Changchun.

The secondary inorganic aerosols (SIA, including SO4 2− , NO3 − and NH4 + ), with a mean of 27.6 ± 5.2 µ g/m 3 , constituted the second largest portion (18.8%) of the total PM2.5 mass. This percentage was lower than some reported values in Beijing (28%) [8] , Tianjin (34.1%) [36] , and Guangzhou (48.5%) [55] , but similar to or even higher than those in some other studies conducted in wintertime Beijing (12.0% in [10] and 20.5% in [55] ), Harbin (13.7%) [56] , and Changchun (18.7%) [31] . We suggest that the relatively low SIA/PM2.5 mass ratios in this study might indicate that the low temperatures in such cold regions have suppressed secondary inorganic formation in winter. Among the trace elements, As (0.7 ± 0.2 µg/m 3 ), Zn (0.65 ± 0.4 µg/m 3 ) and Ba (0.3 ± 0.2 µg/m 3 ) were the most abundant in PM 2.5 . As is a marker for coal combustion and was present at much higher levels in this study than in other studies [61] , suggesting that wintertime air pollution in Changchun may pose serious health threats to local residents given the well-known health effects associated with such trace elements. Zn and Ba are associated with non-tailpipe emissions, such as brake and tire wear [62] . Ba is also a good indicator of firework emissions [63, 64] . The Zn concentrations observed in this study are comparable to those in polluted cities such as Shanghai [65] and the Beijing-Tianjin-Shijiazhuang city clusters [66] in China. Whereas the observed Ba concentration is much higher than those reported in the Beijing-Tianjin-Shijiazhuang city clusters [36, 66] , indicating contribution from fireworks emissions.

For the S4 site, high SIA loadings on 18 and 19 January, and a sudden falling of SIA on 21 January were observed (Figure 4) . We attributed such changes to both regional-scale transport and secondary formation processes from local precursors. From 17 to 19 January, the dominating air masses were from local origins (Figure 3 ) which provided a meteorological condition (low wind speed and high ambient temperature and humidity) that was favorable to growth and accumulation of SIA [67] . After three days of accumulation, SIA reached the maximum value on 19 January. Meanwhile, on 21 January, the dominating air mass origins switched to remote regions as far as Russia. Such air masses provided clean background air and a meteorological condition in favor of diluting the pollutants, resulting in a sudden decrease of SIA and total PM 2.5 mass.

After 21 January, the mineral components increased notably at the residential (S2) site but decreased at other sites ( Figure 5 ). The opposing temporal trends of mineral contents between S2 and other sites resulted in differences in PM 2.5 trends between these sites, as mentioned above (Figure 2 ). The increase in crustal components at site S2 was unlikely to have been caused by windblown dust emissions from the strong wind during this period or by long-range transport of dust plumes from northern arid lands. During the campaign period, the surface in northern China was largely covered by snow and ice, and soil moisture was too high to allow dust mobilization. Long-range transport was also excluded because only the S2 site observed increased crustal element concentrations, whereas other sites that were likely influenced by dust plumes did not experience increases in crustal elements. Emissions from local sources, including fireworks and bridge and subway construction, resumed after the holidays in urban areas. In addition, fugitive dust was re-suspended by human activities under windy conditions. These factors were the major reasons for the increase in crustal PM 2.5 concentrations at the S2 site. Figure 5 illustrates the temporal variations of the measured ions other than SIA and crustal elements. In general, these elements showed larger variations at the S2 and S4 sites than at the other locations, which is consistent with the variable trends in PM 2.5 concentrations. Site S2 exhibited high loadings of mineral elements (e.g., Al, K), coal ash indicators (e.g., As) and firework indicators (e.g., Sr) after 21 January, which is consistent with the aforementioned temporal trends of mineral PM 2.5 ( Figure 4 ) and overall PM 2.5 ( Figure 2 ). The S4 site exhibited extremely high values for almost all elements on 30 December, the day with a very high PM 2.5 value (321 µg m −3 ). However, high ion and element loadings were not detected on the next day (31 December) when the PM 2.5 concentration peaked (reaching 463 µg m −3 ), suggesting that the peak was caused by different emission sources than those on the previous day. The abrupt changes in PM 2.5 concentrations and chemical compositions highlighted the complexity associated with understanding the sources and behaviors of aerosol pollution in this area.

As shown in the above sections (Table 2; Figure 2 ), in conjunction with COD analysis, large spatial variabilities in PM 2.5 concentrations were observed among different locations, suggesting different source origins of PM 2.5 at different locations. Among the sampling sites, the S2 site and especially the S4 site showed obvious distinction. The large spatial variations in PM 2.5 might come from a couple of causes: (1) The samples over the campaign were sparse, resulting in limited number of pairs ranging from 24 to 33 and thus larger COD values; (2) The sampling sites were far from each other (9-14 km) and characterized with significantly different surroundings, which increased spatial variabilities of PM 2.5 ; (3) Contributions from primary emission sources, i.e., mineral dusts (20.7%), were relatively more important than those from secondary inorganic aerosols, i.e., SIA (18.8%), causing sensitive responses of PM 2.5 to primary sources and thus relatively large spatial variabilities. Indeed, four sampling sites were far less than needed to fully represent the spatial variabilities of PM pollution in such a metropolitan region with a fully urbanized area larger than 1200 km 2 (https://en.wikipedia.org/wiki/Changchun; accessed on 16 April 2021). Actually, nine sampling stations in urban area and one station in outer suburb have been in running by Bureau of Ecology and Environment of Changchun since 2013.

From the results of chemical analysis (Table 2 ; Figures 4 and 5) , and the constructed mass closure of PM 2.5 (Figure 6 ), the PM 2.5 sources of the sampling sites also showed obvious spatial variabilities. For example, the S1 site showed highest mineral components and NO 3 − /SO 4 2− mass ratio, consistent with its source features as influenced significantly by traffic; the S2 site showed high SO 4 2− , As and mineral, indicating the prominent contributions from coal burning and traffic; the S4 site was characterized by high halogen species, K + and NH 4 + , indicating the contributions from multi-sources such as coal burning, glass sculpture, biomass burning, and agricultural sectors. The enrichment factor (EF) analysis for the elements was also used to estimate the relative contributions of natural and anthropogenic origins to wintertime PM2.5 in Changchun, following the methods developed by [38] . The results (Figure 7) show that the calculated EF values of Al, Si, Fe, Mn, K, Mg, Ca, and Na were all below 10, suggesting that these elements came predominantly from natural sources. The EF values of Ba, Cu, Cr, Ni, and Pb were between 10 and 100, suggesting contributions from both natural and anthropogenic sources. The EF values of Zn and As were larger than 100, indicating the dominance of anthropogenic sources such as traffic, coal burning, and industrial processes. The enrichment factor (EF) analysis for the elements was also used to estimate the relative contributions of natural and anthropogenic origins to wintertime PM 2.5 in Changchun, following the methods developed by [38] . The results (Figure 7) show that the calculated EF values of Al, Si, Fe, Mn, K, Mg, Ca, and Na were all below 10, suggesting that these elements came predominantly from natural sources. The EF values of Ba, Cu, Cr, Ni, and Pb were between 10 and 100, suggesting contributions from both natural and anthropogenic sources. The EF values of Zn and As were larger than 100, indicating the dominance of anthropogenic sources such as traffic, coal burning, and industrial processes. The enrichment factor (EF) analysis for the elements was also used to estimate the relative contributions of natural and anthropogenic origins to wintertime PM2.5 in Changchun, following the methods developed by [38] . The results (Figure 7) show that the calculated EF values of Al, Si, Fe, Mn, K, Mg, Ca, and Na were all below 10, suggesting that these elements came predominantly from natural sources. The EF values of Ba, Cu, Cr, Ni, and Pb were between 10 and 100, suggesting contributions from both natural and anthropogenic sources. The EF values of Zn and As were larger than 100, indicating the dominance of anthropogenic sources such as traffic, coal burning, and industrial processes. In summary, the PM 2.5 sources in Changchun were characterized with high spatial variabilities among different sites and noteworthy contributions from mineral and halogen species ( Figure 6 ). In addition, the EF analysis showed high EF values of the As element. The chemical and source characteristics of PM 2.5 were consistent with Changchun as being a typical coal burning and industry city.

To estimate the effectiveness of haze mitigation measures dating back to 2012, historic air quality data in Changchun from Chinese Environmental Protect Agency were collected (https://www.aqistudy.cn/historydata/; accessed on 16 April 2021). Results in Figure 8 show that wintertime PM 2. In summary, the PM2.5 sources in Changchun were characterized with high spatial variabilities among different sites and noteworthy contributions from mineral and halogen species (Figure 6 ). In addition, the EF analysis showed high EF values of the As element. The chemical and source characteristics of PM2.5 were consistent with Changchun as being a typical coal burning and industry city.

To estimate the effectiveness of haze mitigation measures dating back to 2012, historic air quality data in Changchun from Chinese Environmental Protect Agency were collected (https://www.aqistudy.cn/historydata/; accessed on 16 April 2021). Results in Figure 8 show that wintertime PM2. 5 However, the PM2.5 switched to a slightly increasing trend from 2019. Even more, the wintertime PM2.5 in the period between 23 December 2019 and 31 January 2020 reached the highest level since 2014. The PM2.5 level appears to reach a stationary plateau stage, with PM2.5 loadings ranging between 51 and 56, and greater efforts must be made to further lower the pollution level in the future. However, the PM 2.5 switched to a slightly increasing trend from 2019. Even more, the wintertime PM 2.5 in the period between 23 December 2019 and 31 January 2020 reached the highest level since 2014. The PM 2.5 level appears to reach a stationary plateau stage, with PM 2.5 loadings ranging between 51 and 56, and greater efforts must be made to further lower the pollution level in the future.

This study presents detailed field observations of the chemical components and spatiotemporal variations of wintertime PM 2.5 in the Changchun metropolitan area located in the center of the China's breadbasket region. Such information could be used as a baseline for related studies based on the fact that such information for this region was very limited before 2013. We found that severe wintertime haze pollution existed in Changchun due to unique environmental settings, including strong temperature inversion, high emission loadings, scattered sources with mixed fuel types, and frequent use of deicing salts. Observations from the four sites across the city revealed that PM 2.5 levels (143 µg m −3 ) constantly exceeded the WHO air quality guidelines for PM 2.5 (25 µg m −3 ) and China's Level I (35 µg m −3 ) and II (75 µg m −3 ) air quality standards for PM 2.5 , with levels comparable to those observed in Asia's most polluted cities. Mineral dust and secondary inorganic aerosols together contributed about 40% to the total PM 2.5 mass.

While strong dependence of PM 2.5 levels on atmospheric circulation were found, the observed heterogeneous distribution of PM 2.5 mass and chemical compositions among the monitoring sites suggested that localized emission sources significantly affected PM 2.5 pollution in Changchun. In particular, the suburban village site exhibited the highest PM 2.5 levels and extremely high Cl − and F − loadings.

Results from this study shed light on both the severity of air pollution in China and the potential measures required to understand and mitigate air pollution in this region. The fact that the levels of both total PM 2.5 and its inorganic components were comparable to those in China's most polluted cities was surprising because Changchun has generally been considered less polluted than other large Chinese cities. In fact, the sky was blue and sunlight was dominant during the field campaign, even on the most polluted days, unlike the visible haze in other cities. However, pollution was visible from a horizontal viewpoint. Our results suggest that there is likely a mechanism behind the high pollution episodes in this region that is distinct from the mechanisms in the Beijing-Tianjing-Hebei region or the Pearl River Delta region. The distinct characteristics of air pollution in this region also make pollution difficult to detect using satellite remote sensing or to simulate using numerical chemical transport models. The shallow boundary layer confines most of the aerosols to the lowest atmospheric layer, and the land in this region is largely covered by snow and ice during the cold season, making it challenging to remotely sense atmospheric aerosols. Furthermore, emission inventories in China have focused on typical industrial sources derived from province-level [68] or county-level [69] energy consumption. Chemical transport models based on these inventories are likely to underpredict some major components of PM 2.5 unless the missing sources can be accurately accounted for in the emission inventories used to drive these models.

Based on the results of the chemical component analysis, one notable characteristic of air pollution in wintertime Changchun was high loadings of mineral dust, other than the chemical species related to the burning of fossil fuels. Thus, effective measures, i.e., more thorough and frequent sweeping or watering of streets, covering of bare soil on sides of busy roads, and stricter dust controls for the widespread construction activities, should be taken to control the road and off-road dust emissions. In addition, the comparisons among the monitoring sites showed that the upwind suburban sites were more polluted than the sites in the urban area. In the future, more attention should be paid to the emissions from small factories of the township enterprises in the suburban and even rural areas.

Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/ijerph18084354/s1, Figure S1 : Correlation relationships between filter based PM 2.5 and related meteorological factors, Figure S2 : Contouring plot showing the relationship between measured PM2.5 and wind speed and direction for the four sampling sites., Table S1: Method detection limit (MDL) for measuring the concentrations of 9 ions and 18 elements, Table S2 : Coefficients of Divergence (COD) and pearson correlation analysis for the bulk PM2.5 concentrations between site pairs. 

(outdoor)-air-quality-and-health

Lung cancer, cardiopulmonary mortality, and long-term exposure to fine particulate air pollution

Long-term ozone exposure and mortality

An estimate of the global burden of anthropogenic ozone and fine particulate matter on premature human mortality using atmospheric modeling

Cleaning China's air

Chemical characteristics of pm 2.5 and pm 10 in haze−fog episodes in Beijing

Characteristics of particulate matter pollution in the pearl river delta region, China: An observational-based analysis of two monitoring sites

Winter and summer pm 2.5 chemical compositions in fourteen chinese cities

The characteristics of pm 2.5 and its chemical compositions between different prevailing wind patterns in Guangzhou

Chemical characterization and source apportionment of pm 2.5 in beijing: Seasonal perspective

Air pollution and health studies in china-policy implications

Associations between fine particle, coarse particle, black carbon and hospital visits in a chinese city

Simultaneous measurement of particulate and gaseous pollutants in an urban city in north china plain during the heating period: Implication of source contribution

The secondary formation of inorganic aerosols in the droplet mode through heterogeneous aqueous reactions under haze conditions

Formation and evolution mechanism of regional haze: A case study in the megacity Beijing, China

Hygroscopic growth of urban aerosol particles during the 2009 mirage-shanghai campaign

Seasonal trends in pm 2.5 source contributions in beijing

Combined source apportionment, using positive matrix factorization-chemical mass balance and principal component analysis/multiple linear regression-chemical mass balance models

Characteristics and health impacts of particulate matter pollution in China

Water spray geoengineering to clean air pollution for mitigating haze in China's cities

Chemical characteristics and source apportionment of atmospheric particles during heating period in harbin

Characteristics of elemental composition of pm 2.5 in the spring period at tongyu in the semi-arid region of northeast china

Characterization of elemental species in pm 2.5 samples collected in four cities of northeast china

A seasonal study of polycyclic aromatic hydrocarbons in pm 2.5 and pm 2.5-10 in five typical cities of liaoning province

Chemical characteristics and source apportionment of pm 10 during a brown haze episode in harbin

Characteristics of visibility and particulate matter (pm) in an urban area of northeast china

Size distributions and health risks of particulate trace elements in rural areas in northeastern china

Investigation of yearly indoor/outdoor pm 2.5 levels in the perspectives of health impacts and air pollution control: Case study in changchun, in the northeast of china

Typical atmospheric haze during crop harvest season in northeastern china: A case in the changchun region

Pollution characteristics of pm2. 5 aerosol during haze periods in changchun

Model elucidating the sources and formation mechanisms of severe haze pollution over northeast mega-city cluster in china

China's environment in a globalizing world

Assessment of heavy metals contamination in urban topsoil from changchun city

Representativeness of an air quality monitoring station for pm2.5 and source apportionment over a small urban domain

Chemical composition of pm 2.5 during winter in tianjin

Wintertime chemical composition of aerosols at a rural location in the indo-gangetic plains

The geochemical evolution of the continental crust

Contamination characteristics and possible sources of pm 10 and pm 2.5 in different functional areas of shanghai, china

Chemical characterisation of pm 2.5 , pm 10 and coarse particles at urban, near-city and rural sites in Switzerland

Characterization of soil dust aerosol in china and its transport and distribution during 2001 ace-asia: 1. Network observations

Variation of sources and mixing mechanism of mineral dust with pollution aerosol-revealed by the two peaks of a super dust storm in Beijing

Spatial and seasonal distributions of carbonaceous aerosols over china

Study on the seasonal variation and source apportionment of pm 10 in harbin

Atmospheric aerosol compositions in China: Spatial/temporal variability, chemical signature, regional haze distribution and comparisons with global aerosols

Seasonal and diurnal variations of ambient pm 2.5 concentration in urban and rural environments in Beijing

Water-soluble ions in atmospheric aerosols measured in xi'an, china: Seasonal variations and sources

Chemical characteristics of fine particulate matters measured during severe winter haze events in ulaanbaatar

Diurnal and seasonal variations of black carbon and pm 2.5 over new delhi, india: Influence of meteorology

Characterization of gaseous pollutants and pm 2.5 at fixed roadsides and along vehicle traveling routes in bangkok metropolitan region

Wintertime particulate pollution episodes in an urban valley of the western us: A case study

NOAA's HYSPLIT atmospheric transport and dispersion modeling system

Temporal and spatial variations of pm 2.5 and pm 10 aerosol in the southern california air quality study

Effect of chemical composition of pm 2.5 on visibility in guangzhou, China

Exploring the cause of pm2.5 pollution episodes in a cold metropolis in China

The ion chemistry and the source of pm 2.5 aerosol in Beijing

Long-range transport of fluoride in east asia monitored at noto peninsula

Inorganic markers, carbonaceous components and stable carbon isotope from biomass burning aerosols in northeast China

Emissions inventory of pm 2.5 trace elements across the united states

Impacts of aerosol compositions on visibility impairment in xi'an, China

Characterization and source apportionment of pm 2.5 in an urban environment in Beijing

The chemical composition and sources of pm 2.5 during the 2009 chinese new year's holiday in Shanghai

Influences of fireworks on chemical characteristics of atmospheric fine and coarse particles during taiwan's lantern festival

The relationship between physicochemical characterization and the potential toxicity of fine particulates (pm 2.5 ) in Shanghai atmosphere

Characteristics of concentrations and chemical compositions for pm 2.5 in the region of beijing, tianjin, and hebei, China

Characterization of major chemical components of fine particulate matter in north carolina

Asian emissions in 2006 for the nasa intex-b mission

A comprehensive inventory of agricultural atmospheric particulate matters (pm 10 and pm 2.5 ) and gaseous pollutants (vocs, so 2 , nh 3 , co, no x and hc) emissions in china

Acknowledgments: Authors gratefully thank anonymous reviewers for their extensive knowledge and helpful advice. We thank Qi Zhang and the Beijing Municipal Institute of Labor Protection for measuring the ions and elements in the samples in this study.

The authors declare no conflict of interest.