Source identification of ultrafine particulate matter and ammonia emissions in the metropolitan area

Jaeseok Heo; Pilho Kim; Jongsung Park; Jae Young Lee

doi:10.4491/eer.2024.074

Environ Eng Res > Volume 30(1); 2025 > Article

Heo, Kim, Park, and Lee: Source identification of ultrafine particulate matter and ammonia emissions in the metropolitan area

Research

Environmental Engineering Research 2025; 30(1): 240074.

Published online: May 22, 2024

DOI: https://doi.org/10.4491/eer.2024.074

Source identification of ultrafine particulate matter and ammonia emissions in the metropolitan area

Jaeseok Heo¹, Pilho Kim², Jongsung Park², Jae Young Lee^3,^†

¹Environment Research Institute, Department of Environmental and Safety Engineering, Ajou University, Suwon-si 16499, South Korea

²Climate & Air Quality Research Department, National Institute of Environmental Research, Incheon 22689, South Korea

³Department of Environmental and Safety Engineering, Ajou University, Suwon-si 16499, South Korea

^†Corresponding author, E-mail: jaeylee@ajou.ac.kr, Tel: +82 312192404, ORCID: 0000-0002-4159-7099

Received February 7, 2024 Revised May 21, 2024 Accepted May 21, 2024

This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/3.0/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

Previous studies have estimated the sources of particulate matter in the atmosphere. Among these, studying the sources of secondary aerosols harmful to human health is important. However, there is a lack of research on ammonia (NH₃), a precursor to secondary aerosol formation. This study uses positive matrix factorization (PMF) model and conditional bivariate probability function (CBPF) model to estimate the sources of particulate matter and ammonia. The results showed that about 40% of the PM_2.5 mass at both sites was attributable to secondary aerosol. To estimate the emission sources of ammonia that contribute to the generation of secondary aerosols, CBPF was utilized to model and compare the emission characteristics of categorized pollution sources and ammonia, and it was found that SMA had similar emission trends to industry, road dust, oil combustion, and biomass combustion, while GRA had similar emission trends to oil combustion and vehicle (diesel). Considering the results from these two regions, ammonia in the metropolitan area is more likely to be emitted from daily activities than from long distances. The study results demonstrate the major role of secondary aerosols on ambient PM_2.5 concentrations and can help develop effective management strategies and policies for air pollution mitigation.

Keywords: Ammonia, Environment monitoring, Emissions, Particular matter, PMF

Graphical Abstract

Keywords: Ammonia, Environment monitoring, Emissions, Particular matter, PMF

1. Introduction

Industrial expansion and increased vehicle activity, indicative of improved quality of life, have significantly escalated air pollution emissions [1]. Once emitted into the atmosphere, pollutants are dispersed over wide areas through airflow and chemical reactions, exposing many individuals to potential health risks [1].

To address this concern, the World Health Organization (WHO) has issued comprehensive air quality management guidelines for both ambient and indoor air. These guidelines form the foundation for setting policies and standards in air quality management [2]. Particularly, particulate matter (PM), recognized as an air pollutant, gained attention when the International Agency for Research on Cancer (IARC), a division of the World Health Organization, classified it as a human carcinogen in 2013 [3].

Extensive research underscores the adverse health effects of prolonged exposure to high concentrations of PM [4–8]. Styer et al. observed a 0.3% increase in mortality with every 10 μg/m³ rise in PM₁₀ concentration, excluding natural causes [9]. Yin et al. reported a 0.62% (0.43–0.81%) increase in deaths from cardiopulmonary disease for every 10 μg/m³ increase in PM [10]. Moreover, Pope et al (2006) linked a 10 μg/m³ increase in PM_2.5 to a 4.5% surge in acute ischemic coronary disease [11]. Studies indicate that PM from automobile engines can constrict blood vessels, contributing to high blood pressure and heart disease [12, 13].

Kim et al. noted that the main chemical components of particulate matter include secondary inorganic components (sulfate (SO₄ ²⁻), nitrate (NO₃ ⁻), and ammonium (NH₄ ⁺)), organic components, and trace metals [14]. These chemical compositions result in different emissions depending on the source. Therefore, identifying the source of the particulate matter (e.g., secondary aerosols, vehicles, coal combustion, soil, etc.) is important [15, 16]. In particular, secondary aerosols are a major contributor to approximately 40% of the particulate matter in the atmosphere [17]. Recently, it was reported that indirect emissions from the secondary sources of particulate matter accounted for approximately 72% of total emissions in Korea [18], and the growing interest in indirect emissions has emphasized the importance of precursors involved in the secondary sources of particulate matter [19]. In order to reduce particulate matter, it is important to manage emission sources of gaseous substances that increase the concentration of pollutants in the atmosphere, but there has been a lack of prior research on gaseous substances such as ammonia, a precursor to particulate matter in the atmosphere, in Korea [20].

In particular, managing ammonia is crucial because it reacts with nitric and sulfuric acids in the atmosphere to form secondary aerosols such as ammonium nitrate [NH₄NO₃] and ammonium sulfate [[NH₄]₂SO₄], the main components of PM_2.5 [21, 22]. Ammoniated aerosols worsen urban air quality and affect human health [23–25].

In the atmosphere, ammonia is caused by both natural and anthropogenic sources [26]. In a study of ammonia emissions, Boyle noted that agriculture and animal husbandry are generally the largest anthropogenic sources of ammonia [27], and Behera et al. noted that industrial processes, vehicle emissions, and volatilization from soil and oceans are the other sources [28]. In particular, Chan et al. reported that the high concentrations of ammonia measured in cities were attributable to vehicle emissions, but the study relied on laboratory studies and tunnel and roadside measurements to estimate vehicle emission factors [ammonia emitted per unit mass of fuel] [22, 29, 30]. As a result, the typicality of laboratory tests across metropolitan areas or fixed measurements at a single location is not well-characterized [22, 31]. In addition, studies related to the sources and atmospheric behavior of ammonia in the atmosphere are scarce [32].

When reviewing this previous research, international studies on estimating ammonia emission sources currently focus on agriculture and livestock, and these results also report large uncertainties in the contribution of sources to atmospheric ammonia at regional scales, given temporal variations in meteorological and environmental conditions and anthropogenic activities [33, 34].

A variety of receptor models have been applied to identify sources of PM_2.5 emissions, with the most commonly used models being Principal Component Analysis (PCA), Positive Matrix Factorization (PMF), Environmental Protection Agency (EPA) Unmix model, and Chemical Mass Balance (CMB) model [16, 35]. Zikova et al. report that the PMF model is the most useful of the three source apportionment methods mentioned above because it provides both a source profile and a time series of source contribution results [36].

In this study, PM_2.5 concentrations and chemical composition were characterized, and the PMF model was used to estimate and identify potential sources of PM_2.5. In addition, CBPF was performed using meteorological data to better understand PM_2.5 source pathways. Finally, to estimate the emission source of ammonia, the CBPF results of the PMF model and ammonia were compared to estimate the emission source of ammonia. This is the first study to estimate the emission sources of ammonia using the PMF and CBPF model. This study can be used as a basis for future source estimation of particulate matter and ammonia emissions, and the results are expected to be used as basic knowledge for air quality improvement strategies and health protection of residents.

2. Materials and Methods

2.1. Study Areas

To estimate the sources of ultrafine particles and ammonia emissions in the metropolitan area, we meticulously curated and analyzed data from two key sources: the SMA and the GRA, as illustrated in Fig. 1. The data utilized for this analysis spans from January to December 2022, focusing on PM_2.5 mass concentrations, its composition, and ammonia concentrations. The SMA located in Eunpyeong-gu, Seoul, is characterized by high emissions from both vehicles and non-industrial combustion in residential districts, owing to its densely populated nature. On the other hand, the GRA, located in Ansan, Gyeonggi-do, is in the middle of industrial facilities that emit various air pollutants. Notable complexes in the vicinity include the Banyeol Special National Industrial Complex and the Sihwa Industrial Complex, both contributing significantly to the region’s air quality dynamics. Based on data from the Clean Air Policy Support System (CAPSS), Gyeonggi-do was observed to have the highest concentration of particulate matter and the second-highest concentration of ammonia in South Korea [37].

2.2. Measurements

To measure the concentration of PM, a Beta Attenuation Mass Monitor (BAM-1020, Met One Inc., USA) was employed at the air environment research center. Air samples were collected from a flow rate of 16.7 L/min through a filter and subjected to measurement through the beta-ray absorption method. The detection of beta rays both before and after sampling utilized a highly sensitive detector. The resultant beta ray attenuation was employed to calculate the mass concentration of particulate matter (μg/m³). For the measurement of ionic components within PM_2.5, an Ambient Ion Monitor (URG-9000D, URG Co., USA) was utilized. Atmospheric aerosols were collected at a flow rate of 3 L/min through the PM_2.5 inlet. A membrane-type liquid diffusion denuder coated with 30% hydrogen peroxide was employed to eliminate gaseous substances. Ensuring high capture efficiency, particles underwent growth in a super saturation chamber. Subsequently, the grown particles were separated in an inertial particle separator and collected in an aerosol sample collector. Ion Chromatography (ICS-2000, Thermo., USA), was then employed to analyze the collected samples for ionic components. The detection limit may vary based on the measured concentration level but typically falls within the range of 0.05 ~ 0.1 μg/m³. Carbon composition analysis was conducted utilizing a Semi-continuous OCEC Analyzer (SOCEC, Sunset lab., USA) employing thermal-optical-transmittance (TOT) and Non-Dispersive Infrared (NDIR) methods, adhering to NIOSH and EPA standards. For elemental composition analysis, an Ambient Continuous Multi-Metals Monitor (Xact 625i, Sailbri Cooper Inc., USA) utilizing non-destructive XRF (X-Ray Fluorescence spectroscopy) was employed. The obtained results encompass PM_2.5 mass concentration, the ionic composition of PM_2.5 (SO₄ ²⁻, NO₃ ⁻, Cl⁻, Na⁺, NH₄ ⁺, K⁺, Mg²⁺, and Ca²⁺), carbon composition (OC and EC), and elemental composition (Si, S, K, Ca, Ti, V, Cr, Mn, Fe, Ni, Cu, Zn, As, Se, Br, Ba, and Pb). To ensure the reliability of the analyzed data, laboratory blanks (LB) were stored in the refrigerator (< 4ºC) during the actual sampling time and analyzed right before and after the actual sample analysis. Field blanks (FB) were stored and analyzed daily under the field conditions.

For a continuous measurement of ammonia, a Gas Concentration Analyzer (G2123, PICARRO Inc., USA) was employed. This instrument, functioning as a precursor for nitrate and sulfate production, upheld consistent operating conditions. The analysis system maintained the operating conditions of the measuring device to inject the sample into the cavity by aspirating the sample from the ambient air at a flow rate of 1.9 to 2.1 L per minute once the laser value stabilized, and the measurement cycle was analyzed in real time at a minimum interval of 1 second to produce the measurement data.

2.3. Source Apportionment using Positive Matrix Factorization (PMF)

The PMF model stands as a prominent receptor model and is extensively utilized for identifying the individual contribution of pollutants, even in scenarios where a pollutant classification table is not available [38–42]. Source apportionment, an integral component of the PMF model, quantifies the impacts and contributions of both particulate and gaseous air pollutants sampled from a receptor [43]. In atmospheric research, factor analysis plays a crucial role in identifying pollution sources and calculating their respective contributions. However, traditional factor analysis relies solely on the covariance matrix, presenting challenges such as imperfect interpretation as a result of the negative factor loadings and ambiguity in factor rotation [44]. To address these limitations, the PMF model was developed by Paatero and Tapper [45–48]. A conventional factor analysis model is represented by Eq. (1) [48].

(1)

X = G F + E

Matrix X exhibits a structured dimensionality of nm, where n denotes the number of samples and m is the analyzed chemical species. In other words, the rows of the X matrix correspond to each collected sample, while the columns encompass concentrations of the chemical species in a given sample. The G matrix is represented by dimensions np and the F matrix adheres to dimensions pm, with p representing the number of factors. In detail, the rows of the F matrix represent the source profile for a particular source, while the columns of the G matrix delineate the strength or source contribution for a particular source. Both matrices G and F consistently possess positive values. Meanwhile, matrix E functions as the residual matrix, as expressed in by Eq. (2) [49].

(2)

E_{i j} = X_{i j} - Σ_{h = 1}^{p} G_{i k} F_{k j}

The measured and calculated PM_2.5 concentrations greater than ± 50%, and samples with anion and cation balances beyond the mean ± 2 σ [σ: standard deviation] were excluded from the model input.

Finally, from a dataset of 8,760 hourly measurements in 2022, 5,401 samples in this study, the EPA’s PMF 5.0 was employed for source apportionment of PM_2.5, emissions in Seoul and Ansan. The utilized data comprised mass concentrations of chemical components in particulate matter, along with their associated uncertainties. The method outlined by Polissar et al. (1998) guided the application of this method [50]. For data falling below the method detection limit (MDL), the mass concentration and its uncertainty were substituted with 1/2 and 5/6 of the MDL, respectively. In instances of missing data, the mass concentration was estimated utilizing the geometric mean of each component, while four times the geometric mean was employed as the uncertainty.

When using the PMF model, PM_2.5 was set as the ‘total variable’ in the settings, and then an additional modeling uncertainty of 20% was included to mitigate unaccounted errors. The PMF model underwent iterative application to ascertain the optimal number of factors required to elucidate the parameter distribution at the two receptor sites. Based on the distribution of the scaled residuals and interpretability of the resulting source profiles, eight factors were selected as the optimal number of contributors for both sites. In this study, samples with analyte interruptions, those exhibiting a discrepancy between data consisting of 24 chemicals were selected for PMF modeling analysis in the Seoul metropolitan area. In addition, 6,245 samples of data consisting of 25 chemicals were utilized for the same analysis in the Gyeonggi region.

2.4. Conditional Bivariate Probability Function (CBPF)

The conditional probability function (CPF) [51] was utilized to estimate the source magnitude for a given direction utilizing a pollutant source. The wind direction was supplemented with wind speed to separate emission intervals with regard to source impacts on receptors. A conditional bivariate probability function (CBPF) was employed, which can be modeled to track source locations on a local scale [52–54]. To trace the path of emission sources to the SMA and GRA, CBPF was performed using R’s Openair package [55]. This calculation was performed using Eq. (3) [53, 56].

(3)

CBPF = \frac{m_{Δ θ, Δ u} ∣ C \geq x}{m_{Δ θ, Δ u}}

where M_Δθ,Δu is the number of data in a wind direction section (Δθ) and wind speed section (Δu) where concentration C is greater than a threshold x, and n_Δθ, Δu is the total number of data for a wind direction and wind speed section. CBPF is a function of wind direction and wind speed; the closer the CBPF value is to 1, the more inclined it is that an emission source exists in that wind direction and wind speed section. The reference concentrations were above the 50th percentile of source contribution and the wind direction was 16 directions (Δθ=22.5), for wind speed, data below 0.5 m/sec were excluded as a result of high uncertainty in the wind direction [57, 58].

3. Results and Discussion

3.1. Seasonal Variations of Chemical Constituents of PM_2.5

PM_2.5 concentrations vary from 3 to 118 μg/m³ at SMA, and 3 to 132 μg/m³ at GRA during the sampling period, with annual averages of 24.7 ± 16.8 μg/m³ and 27.6 ± 19.7 μg/m³, respectively, which exceed the air quality standards in South Korea (15 μg/m³), The seasonal PM_2.5 concentration in SMA is highest in winter (26.9 ± 18.1 μg/m³), followed by spring (26.9 ± 15.7 μg/m³), autumn (23.1 ± 16.1 μg/m³) and summer (17.3 ± 9.9 μg/m³), that in GRA is highest in winter (34.3 ± 21.7 μg/m³) followed by spring (30.3 ± 18.5 μg/m³), autumn (29.5 ± 21.3 μg/m³) and summer (15.8 ± 9.4 μg/m³). To show the time series and seasonal variation of these particulate matter concentrations, Fig. 2 (a) and (b) show the time series data of SMA and GRA, and (c) shows the seasonal variation of SMA and GRA. The results show that the concentration trends of PM in both SMA and GRA are seasonally similar. These seasonal trends show that the average concentration was high in spring mainly due to the yellow dust (up to 330 μg/m³) that occurred in March and April. Further, PM_2.5 concentration was low in summer due to the removal of particles from the atmosphere owing to frequent precipitation, such as the rainy season (average 15 days/month in summer) and the wind direction, generally south or southwest in that period [59–61]. Elevated concentrations of PM_2.5 in winter are thought to be a result of the higher frequency of temperature inversions, relatively stable atmospheric conditions in winter, and mixing reduction effects, which are common in winter in combination with increased emissions such as heating [62–64]. The mean and maximum concentrations of ions, carbon, and metals of PM_2.5 in the SMA and the GRA are summarized in Table S1.

In both regions, the mean concentrations and contributing percentages of ions were NO₃ ⁻ (5.6 ± 6.2 μg/m³, 48.7%) > SO₄ ²⁻ (2.8 ± 2.0 μg/m³, 24.35%) > NH₄ ⁺ (2.6 ± 2.5 μg/m³, 22.61%) > Cl⁻ (0.26 ± 0.31 μg/m³, 2.26%) > Ca²⁺ (0.08 ± 0.1 μg/m³, 0.7%) > K⁺ (0.08 ± 0.1 μg/m³, 0.7%) > Na⁺ (0.07 ± 0.12 μg/m³, 0.61%) > Mg²⁺ (0.008 ± 0.03 μg/m³, 0.07%), while in GRA, NO₃ ⁻ (7.7 ± 9.3 μg/m³, 46.86%) > SO₄ ²⁻ (4.0 ± 3.0 μg/m³, 24.34%) > NH₄ ⁺ (3.8 ± 3.4 μg/m³, 23.12%) > Cl⁻ (0.53 ± 0.7 μg/m³, 3.23%) > K⁺ (0.16 ± 0.2 μg/m³, 0.97%) > Na⁺ (0.14 ± 0.1 μg/m³, 0.85%) > Ca²⁺ (0.08 ± 0.1 μg/m³, 0.49%) > Mg²⁺ (0.023 ± 0.05 μg/m³, 0.14%). The proportion of ionic components (NO₃ ⁻, SO₄ ²⁻, NH₄ ⁺) contributing to secondary aerosol among total ions was 95.7 and 94.3% in SMA and GRA, respectively.

The carbon component is composed of elemental and organic carbon, with elemental carbon being the primary pollutant emitted directly into the atmosphere from combustion sources and utilized as a tracer for primary organic carbon [65]. Organic carbon is divided into primary organic carbon, which is directly emitted from various combustion processes, such as biological combustion and fossil fuel combustion, and secondary organic carbon, which is converted into primary organic carbon through oxidation and aging processes [66]. The annual average concentration of organic carbon in the carbon component during the measurement period was analyzed as 3.8 μg/m³ in SMA and 4.0 μg/m³ in GRA, which are similar. The elemental carbon content remained the same, at 0.8 μg/m³. For organic carbon, the highest concentration levels were found in winter, when fossil fuel use increases, at 4.3 and 5.3 μg/m³, respectively. Similar to organic carbon, elemental carbon had the highest concentrations of 1.1 and 1.2 μg/m³ during the winter season when combustion emissions increased, compared to 0.5 and 0.3 μg/m³ during the lowest summer months, respectively, which is about 2.2 to 3.5 times higher. The proportion of metals in PM was about 8% in both regions. In addition, both areas exhibited high concentrations of sulfur (S), potassium (K), and iron (Fe), with slightly higher concentrations in spring during Asian dust [67].

The seasonal mean concentrations of observed ammonia ranged from 5.2 to 8.8 ppb in the SMA and from 7.9 to 11.2 ppb in the GRA. Both regions exhibited a pattern of high concentrations during summer and low concentrations during winter. In summer, the higher temperatures are hypothesized to lead to increased volatilization of ammonia into the ambient air, resulting in elevated concentrations. During autumn, ammonia concentrations were slightly higher in the GRA at 10.3 ppb compared to 6.6 ppb in the SMA. This disparity may be attributed to the effects of open burning practices, particularly after harvest, in the agricultural areas (Ansan, Hwaseong, and Suwon) located southeast of the GRA, causing higher ammonia concentrations in this region compared to the SMA. Table S2 shows ammonia concentrations by season.

3.2. Assessing PMF Model Reliability

The reliability of the PMF modeling results is important to verify. This can be confirmed by analyzing the correlation between the predicted and actual mass concentrations for particulate matter [44, 68]. Based on the PMF, the slopes of the SMA and GRA were 0.95 and 0.96, respectively. The modeled and measured masses had coefficients of determination (R²) of 0.94 and 0.93, respectively. Fig. 3 shows the correlation between predicted and actual mass concentrations of particulate matter.

3.3. Source Classification and Identification through PMF Modeling

As shown in Fig. 4, in SMA, the sources of PM_2.5 were categorized as secondary aerosols (36.7%), oil combustion (13.6%), industrial (11.0%), biomass combustion (10.7%), vehicles (diesel, 9.8%), coal combustion (8.0%), sea salt (6.8%), and road dust (3.4%). In GRA, the sources of PM_2.5 were categorized as secondary aerosols (40.1%), Na-rich (16.3%), sea salt (15.7%), vehicles (diesel, 8.3%), oil combustion (7.0%), coal combustion (6.7%), industrial (3.9%), and road dust (2.0%). Commonly categorized sources in SMA and GRA comprised of secondary aerosol, oil combustion, industrial, vehicle (diesel), coal combustion, sea salt, and road dust.

In Factor 4 of the SMA and Factor 1 of the GRA, the main sources of secondary aerosols, such as NO₃ ⁻, SO₄ ²⁻, and NH₄ ⁺, were classified as secondary aerosols due to their high concentrations. NH₄ ⁺ and SO₄ ²⁻ are found to be present in the atmosphere as secondary pollutants in the form of sulfate by uniform and nonuniform reactions, and the generation mechanism of NO₃ ⁻ is also found to be as a secondary pollutant in the form of nitrate by the uniform reaction of HNO₃ and NH₄ ⁺ in the gas phase produced by photochemical reactions that bind strongly with NH₄ ⁺ [69, 70]. Also, nitrate is primarily concentrated in winter when low temperatures and high humidity contribute to the formation of secondary nitrate [71–73]. Sulfate, which is formed by the photochemical reaction of the precursor SO₂ at high temperatures, resulting in the formation of secondary particles, is highly characterized in the summer [71, 74, 75]. Seo et al. reported that in the case of nitrate, nitrogen oxide emissions from automobiles and industrial complexes in the metropolitan area are high, resulting in the mixing of particulate matter from China with nitrogen oxide and ammonia emitted domestically, further increasing the nitrate content in the particles, which can lead to increased emissions [76].

In the case of Factor 5 in SMA and GRA, oil combustion was classified. The major contributors to oil combustion are V and Ni [77–80]. Nigam et al. measured the contribution of V and Ni to emissions by direct sampling for various engines and fuels from the exhaust of auxiliary engines used by ships [81]. Mazzei et al. indicated that the contribution of oil combustion was particularly high in summer, which may be related to the noticeable increase in traffic of passenger ships visiting the port during the holiday season. As significant stationary emission sources burning residual oil (e.g. power plants) do not exist in the Genoa urban area, port related activities were assumed to be the main source of heavy oil combustion [78]. The SMA was classified as oil combustion with contributions of V and Ni of approximately 85.1% and 51.4%, and in GRA of 86.7% and 40.2%, respectively [58, 77, 82]. In the case of oil combustion, emissions are known to occur when diesel fuel and B-C oil are burned, such as in engine combustion, power plant combustion, etc. as noted in previous studies [83]. Oil combustion pollutants typically contribute more during the winter months when temperatures are lower due to heating, but the combustion process of many fuels used by nearby large manufacturing plants or small businesses can also contribute [84].

Industries that mainly contributed Cr, Mn, Fe, Ni, Cu, Zn, etc. were categorized as factor 1 and factor 2 [85, 86]. Both locations showed high contributions of Fe, Cu, Pb, and Cd, which are major tracers from industrial facilities such as sintering plants [49]. In particular, the GRA is influenced by the National Industrial Complex to the west, which is estimated to have high concentrations of industrial factor-related species. However, in GRA, the contribution of the industries pollutant source, where metals are the main indicator, was estimated at 3.9%, which is relatively low compared to other pollutant sources. By the characteristics of the PMF receptor model, which separately considers secondary pollutants and directly emitted primary pollutants, the industry factor is assumed to have a low contribution because of its split with primary emissions and secondary sources generation. Other studies have identified it as an industrial process because of the high content of metallic elements such as Ag, Al, Co, Cr, Cu, Mn, Ni, and V. Ni can be sourced from burning oil in industrial oil boilers or non-ferrous metal smelters, and Cr is an important tracer in the metallurgical industry, according to the researchers [80]. LEE et al. noted that a near 1:1 OC to EC ratio is a common feature in diesel vehicles [84]. Gildemeister et al. noted that high EC content along with S, Ba, and Fe is also observed when using diesel [74]. Previous studies have identified significant emissions of Ba, Zn, Cu, Sb, and Pb from tunnel driving using heavy-duty trucks [87], noting that heavy-duty vehicles emit Ba and Sb more intensely than light-duty vehicles [88]. The contribution of Ba is mainly from diesel vehicles; Factor 8 of SMA and Factor 7 of GRA categorized as vehicles (diesel) exhibited Ba contributions of 86.6% and 93.9%, respectively. In addition, the contributions of Cu, EC, and OC were excessive; therefore, it was categorized as a vehicle (diesel). Ba, also known as an indicator of road dust, is utilized in diesel vehicles as a soot suppressant and lubricant wear inhibitor [89]. In addition, OC and EC are major pollutants from the combustion of automobile oil [64]. In the case of the SMA, the impact of high population density and many vehicles, and in the case of GRA, the impact of diesel engines, is estimated to be higher than that in other areas as a result of the frequent movement of large cargo vehicles resulting from the industrial area’s characteristics as a nearby national industrial area.

Huang et al. noted that coal combustion is characterized by particularly high levels of Cl-associated with Pb, As, OC, and EC, with Pb being classified as a major source in winter coal combustion supply, and OC and EC being primary products of coal combustion [80]. In both regions, As and Pb were the highest contributors, with Zn, Mg, EC, and OC also high. Previous studies have found that as and Se are indicators of coal combustion sources [82]. Consequently, Factor 3 and Factor 6 were identified and categorized as coal combustion sources [90, 91].

In the case of the crustal elements Na, Mg, Al, Si, Ca, and Fe in resuspended road dust, a study by Moosmuller et al. stated that resuspension has increased due to the increase in heavy-duty trucks [92]. Moreover, Factor 2 and Factor 3 were consistently classified as road dust in the SMA and GRA, respectively. The contributions of Si, Ca, and Fe, primary components associated with soil sources and indicative of soil and road fugitive pollution, were prominent in both regions [73, 93–95]. It is postulated that particles from the soil are redispersed into the ambient air, originating from activities such as road traffic and construction sites [73].

Gildemeister et al. estimated that in the early July 2002, unusually large contributions of OC, S, and K⁺ were due to the effects of a large wildfire in Quebec, Canada at the time. In addition, the high levels of the element on other days could be the result of a combination of residential wood burning in winter and meat cooking throughout the year [74]. Factor 6 was categorized as biomass burning in the SMA, with a substantial 88.6% contribution from K⁺, a typical chemical component of biomass burning [74, 80, 96–100]. In the case of biomass burning, the contribution from the west and southwest was high, especially in winter. It is assumed that this is due to various burning activities, including the usage of coal for heating at the redevelopment construction site located near the western part of the SMA.

Factor 4 was categorized as Na-rich in the GRA. This source of pollution is known to be caused by sea salt, which is composed of Na⁺, Cl⁻, SO₄ ²⁻, Mg²⁺, K⁺, and Ca²⁺, but in this study, only Na⁺ contributes significantly, as shown in Fig. 5. This is assumed to be due to chemical reactions of sea salt components in the air. Aged sea salt has a high contribution of Na⁺ and a low contribution of Cl⁻ because NaCl is converted to Na₂SO₄ by reaction with gas phase sulfuric acid (H₂SO₄) and Cl⁻ is depleted [68, 84]. In addition, the use of salt as a de-icer to prevent ice on roads in winter is known to contribute to Na-rich pollution sources [101].

Factor 7 and Factor 8 were categorized as sea salt in the SMA and GRA, respectively. In the case of the SMA, though far from the coast, a high contribution of Cl⁻ and Na⁺ were categorized as sea salt, which was further analyzed by using CBPF. The GRA was estimated to be sea salt because of the coast that is within proximity and the high contribution of Na⁺, Cl⁻ as the primary components of sea salt [49].

3.4. Estimation of Particulate Matter and Ammonia Emission Sources using CBPF

Since air pollutants and wind speed are correlated, data such as wind speed and wind direction were used to estimate the sources of particulate matter and ammonia [103]. The CBPF model, which is suitable for estimating pollutants using wind direction and wind speed data, was used. The data applied to the CBPF model excluded wind speeds below 0.5 m/s due to uncertainty in wind direction at low wind speeds. In addition, the study utilized the 50th percentile of contributions to clarify the wind dependence of the pollutants.

The CBPF results for the categorized pollution sources in the SMA and GRA are shown in Fig. 6. In the SMA, secondary aerosols, industry, road dust, oil combustion, and biomass burning were influenced by the southwest wind. Particulate matter from road dust, oil combustion, and biomass burning is estimated to be from redevelopment construction in the southwest. In addition, secondary aerosols and vehicles (diesel) showed high particulate matter concentrations near the measurement points. As mentioned above, the increase in particulate matter concentration is estimated to be due to the high concentration of nitrogen oxide emission sources, such as vehicles and residential heating in the metropolitan area, which mix with ammonia to form secondary aerosols [76]. Coal combustion and sea salt had high particulate matter concentrations from farther away during northwest winds, meaning that external sources had a greater impact than local sources. The overall results showed that the SMA was more affected by localized sources, such as vehicles and residential emissions. In the GRA, coal combustion and vehicles had high particulate matter concentrations during southwest winds, which is likely influenced by industries located in the southwest.

Other categorized pollution sources showed locally high particulate matter concentrations, but the exact source could not be identified. Accordingly, seasonal changes in particulate matter concentrations were analyzed for vehicles (diesel), road dust, Na-rich, and sea salt as shown in (a)~(d) in Fig. 7. In vehicles (diesel), the particulate matter concentration was high in the north-west, and especially in the autumn, the particulate matter concentration was high in all directions. Due to the Pyeongtaek Siheung Expressway located in the northwest, high particulate matter concentrations were estimated to be influenced by vehicles traveling into residential and industrial areas. Road dust showed high particulate matter concentrations throughout the year, except during the summer months, showing that road dust is significantly influenced by vehicle re-fugitive dust, as known from previous studies [101]. Na-rich and sea salt showed high particulate matter in all directions during the winter, with northwest, southwest, and localized effects in the spring. The overall higher concentrations of fine particulate matter in winter and spring could be due to snow removal substances such as calcium chloride, which are often used for snow removal in South Korea [101].

In the case of ammonia, annual concentrations were observed to be high in the southwest and southeast in the SMA and northwest in the GRA. Fig. S1 shows the results. Comparing these results to the categorized pollution sources in Fig. 6, ammonia from the SMA showed a similar trend to the factors such as industrial, road dust, oil combustion, and biomass combustion, but was not clearly distinguishable in the GRA. Therefore, CBPF modeling was performed to compare ammonia and categorized pollution sources for specific seasons, as shown in Fig. S2 and Fig. S3. The specific season was chosen to be spring, which has high concentrations of both particulate matter and ammonia. In spring, ammonia shows a very similar trend to secondary aerosol factors and combustion-related factors such as industrial, oil combustion, and biomass burning among the categorized pollution sources in SMA. In the GRA, oil combustion and vehicles (diesel) were estimated to be the largest contributors to ammonia emissions. Considering the results from these two regions, ammonia in the metropolitan area is more likely to be emitted from daily activities than from long distances.

4. Conclusions

In this study, the PM_2.5 concentration and chemical composition data collected from two cities in the Seoul Metropolitan Area (SMA and GRA) were used to estimate the emission sources of ammonia using PMF and CBPF models for the first time in Korea. Based on results from the PMF model, the sources of PM_2.5 in the SMA were classified into secondary aerosol (36.7%), oil combustion (13.6%), industry (11.0%), biomass burning (10.7%), vehicle (diesel, 9.8%), coal combustion (8.0%), sea salt (6.8%) and road dust (3.5%). In the GRA, sources of PM_2.5 were secondary aerosol (40.1%), Na-rich (16.3%), sea salt (15.7%), vehicle (diesel, 8.3%), oil combustion (7.0%), coal combustion (6.7%), industry (3.9%), and road dust (2.0%).

CBPF results also showed that the SMA had high particulate matter concentrations, primarily in the southwest. In GRA, coal combustion and vehicles (diesel) showed high particulate matter concentrations in the southwest, while other categorized pollution sources showed locally high concentrations.

Regarding ammonia concentrations, the SMA was more affected in the southwest and southeast, while the GRA was more affected in the northwest. When comparing the CBPF results for ammonia and the categorized pollution sources, the SMA found that ammonia had high concentrations at similar points to factors such as industry, oil combustion, biomass burning, and road dust, thus those factors were assumed to be the source. In the spring, which is the season of high concentrations of both particulate matter and ammonia, the SMA estimated that ammonia is emitted from industry, oil combustion, and biomass burning, while the GRA estimated that ammonia is emitted from oil combustion and vehicles (diesel). In conclusion, ammonia in the metropolitan area is more likely to be emitted from daily activities than from long distances.

Although this study has limitations such as a limited number of data, the results of this study can be utilized to determine emission sources of ammonia as a precursor to PM_2.5 and to develop effective emission control strategies for PM_2.5. In addition, studies on sources of particulate matter and ammonia in different regions and analyzing variables of PMF can improve the reliability of PMF and ammonia source apportionments in the future study.

Supplementary Information

eer-2024-074-Supplementary.pdf

Acknowledgments

This research was supported by the National Institute of Environmental Research (NIER) funded by the Ministry of Environment (ME) of the Republic of Korea (grant number NIER-2023-04-02-056); and the Fine Particle Research Initiative in East Asia Considering National Differences (FRIEND) Project through the National Research Foundation of Korea (NRF), funded by the Ministry of Science and ICT (grant number NRF-2023M 3G1A1090660). Also, this work is financially supported by Korea Ministry of Environment (MOE) as ‘Graduate School specialized in Climate Change’.

Notes

Conflicts of Interest

The authors declare no conflicts of interest.

Author Contributions

J.S.H. (Post-doc) conducted conceptualization, methodology, data curation, writing-original draft preparation, writing-review and editing.; J.Y.L. (Assistant Professor) conducted conceptualization, methodology, writing-review and editing funding acquisition.; P.H.K. (Researcher) and J.S.P. (Researcher) conducted data curation.

References

1. Choi J, Heo J, Ban S, Yi S, Zoh K. Source apportionment of PM_2.5 at the coastal area in Korea. Sci. Total Environ. 2013;447:370–380. https://doi.org/10.1016/j.scitotenv.2012.12.047

2. Krzyzanowski M, Cohen A. Update of WHO air quality guidelines. Air Qual. Atmos. Health. 2008;1:7–13. https://doi.org/10.1007/s11869-008-0008-9

3. Loomis D, Grosse Y, Lauby-Secretan B, et al. The carcinogenicity of outdoor air pollution. Lancet Oncol. 2013;14(13)1262–3. https://doi.org/10.1016/S1470-2045(13)70487-X

4. Ngoc L, Kim M, Bui V, Park D, Lee Y. Particulate matter exposure of passengers at bus stations: A review. Int. J. Environ. Res. Public Health. 2018;15(12)2886. https://doi.org/10.3390/ijerph15122886

5. Anderson J, Thundiyil J, Stolbach A. Clearing the air: a review of the effects of particulate matter air pollution on human health. J. Med. Toxicol. 2012;8:166–75. https://doi.org/10.1007/s13181-011-0203-1

6. Martinelli N, Olivieri O, Girelli D. Air particulate matter and cardiovascular disease: a narrative review. Eur. J. Intern. Med. 2013;24(4)295–302. https://doi.org/10.1016/j.ejim.2013.04.001

7. Heo J, Jang Y, Versoza M, Kim G, Park D. A New Method of Removing Fine Particulates Using an Electrostatic Force. Int. J. Environ. Res. Public Health. 2021;18(12)6199. https://doi.org/10.3390/ijerph18126199

8. Asril M, Sugiarto S, Zurfi A. Airborne microbial quality assessment in the educational buildings during the COVID-19 pandemic. Civ. Eng. J. 2023;9(1)114–26. http://dx.doi.org/10.28991/CEJ-2023-09-01-09

9. Styer P, McMillan N, Gao F, Davis J, Sacks J. Effect of outdoor airborne particulate matter on daily death counts. Environ. Health Perspect. 1995;103(5)490–7. https://doi.org/10.1289/ehp.95103490

10. Yin P, He G, Fan M, et al. Particulate air pollution and mortality in 38 of China’s largest cities: time series analysis. BMJ. 2017;356: https://doi.org/10.1136/bmj.j667

11. Pope C, Muhlestein J, May H, et al. Ischemic heart disease events triggered by short-term exposure to fine particulate air pollution. Circulation. 2006;114(23)2443–8. https://doi.org/10.1161/CIRCULATIONAHA.106.636977

12. Brook R, Brook J, Urch B, et al. Inhalation of fine particulate air pollution and ozone causes acute arterial vasoconstriction in healthy adults. Circulation. 2002;105(13)1534–6. https://doi.org/10.1161/01.CIR.0000013838.94747.64

13. Ibald-Mulli A, Timonen K, Peters A, et al. Effects of particulate air pollution on blood pressure and heart rate in subjects with cardiovascular disease: a multicenter approach. Environ. Health Perspect. 2004;112(3)369–77. https://doi.org/10.1289/ehp.6523

14. Kim S, Kim T, Yi S, Heo J. Source apportionment of PM_2.5 using positive matrix factorization (PMF) at a rural site in Korea. J. Environ. Manage. 2018;214:325–34. https://doi.org/10.1016/j.jenvman.2018.03.027

15. Karagulian F, Belis C, Dora C, et al. Contributions to cities’ ambient particulate matter (PM): A systematic review of local source contributions at global level. Atmos. Environ. 2015;120:475–83. https://doi.org/10.1016/j.atmosenv.2015.08.087

16. Park E, Heo J, Kim H, Yi S. Long term trends of chemical constituents and source contributions of PM_2.5 in Seoul. Chemosphere. 2020;251:126371. https://doi.org/10.1016/j.chemosphere.2020.126371

17. Kim S, Kim T, Yi S, Heo J. Source apportionment of PM_2.5 using positive matrix factorization (PMF) at a rural site in Korea. J. Environ. Manage. 2018;214:325–34. https://doi.org/10.1016/j.jenvman.2018.03.027

18. Oh D, Yoo J, Seo S, et al. Study on the Evaluation of Contribution of Secondary Generation of Fine Dust and Ammonia Using a Photochemical Model in the Metropolitan and Chungnam Area. Journal of the Korean Society of Urban Environment. 2020;20(4)237–52. https://www.earticle.net/Article/A394342

19. Irwin J, Williams M. Acid rain: chemistry and transport. Environ. Pollut. 1988;50(1–2)29–59. https://doi.org/10.1016/0269-7491(88)90184-4

20. AlOmar M, Hameed M, Al-Ansari N, Razali S, AlSaadi M. Short-, medium-, and long-term prediction of carbon dioxide emissions using wavelet-enhanced extreme learning machine. Civ. Eng. J. 2023;9(4)815–34. http://dx.doi.org/10.28991/CEJ-2023-09-04-04

21. Hien P, Bac V, Thinh N. PMF receptor modelling of fine and coarse PM₁₀ in air masses governing monsoon conditions in Hanoi, northern Vietnam. Atmos. Environ. 2004;38(2)189–201. https://doi.org/10.1016/j.atmosenv.2003.09.064

22. Sun K, Tao L, Miller D, et al. Vehicle emissions as an important urban ammonia source in the United States and China. Environ. Sci. Technol. 2017;51(4)2472–81. https://doi.org/10.1021/acs.est.6b02805

23. Xu G, Liu J, Wang Y, et al. Exploring the Impact of Palladium Species’ State on Catalytic Ammonia Elimination over Palladium Catalysts. J. Phys. Chem. C. 2024;128:2856–2866. https://doi.org/10.1021/acs.jpcc.3c07192

24. Yan F, Chen W, Wang X, et al. Significant Increase in Ammonia Emissions in China: Considering Nonagricultural Sectors Based on Isotopic Source Apportionment. Environ. Sci. Technol. 2024;58(5)2423–2433. https://doi.org/10.1021/acs.est.3c07222

25. Kuttippurath J, Singh A, Dash S, et al. Record high levels of atmospheric ammonia over India: Spatial and temporal analyses. Sci. Total Environ. 2020;740:139986. https://doi.org/10.1016/j.scitotenv.2020.139986

26. Zeng Y, Tian S, Pan Y. Revealing the sources of atmospheric ammonia: a review. Curr. Pollut. Rep. 2018;4:189–97. https://doi.org/10.1007/s40726-018-0096-6

27. Boyle E. Nitrogen pollution knows no bounds. Science. 2017;356(6339)700–1. https://doi.org/10.1126/science.aan3242

28. Behera S, Sharma M, Aneja V, Balasubramanian R. Ammonia in the atmosphere: a review on emission sources, atmospheric chemistry and deposition on terrestrial bodies. Environ. Sci. Pollut. Res. 2013;20:8092–131. https://doi.org/10.1007/s11356-013-2051-9

29. Chen Y, Zhang Q, Cai X, et al. Rapid increase in China’s industrial ammonia emissions: Evidence from unit-based mapping. Environ. Sci. Technol. 2022;56(6)3375–85. https://doi.org/10.1021/acs.est.1c08369

30. Zhang Y, Benedict K, Tang A, et al. Persistent nonagricultural and periodic agricultural emissions dominate sources of ammonia in urban Beijing: evidence from 15N stable isotope in vertical profiles. Environ. Sci. Technol. 2019;54(1)102–9. https://doi.org/10.1021/acs.est.9b05741

31. Chang Y, Zou Z, Zhang Y, et al. Assessing contributions of agricultural and nonagricultural emissions to atmospheric ammonia in a Chinese megacity. Environ. Sci. Technol. 2019;53(4)1822–33. https://doi.org/10.1021/acs.est.8b05984

32. Sim C, Lee S, Hahn J, et al. A Study on Integrated Management of Particulate Matter Pollution. Korea Environment Institute; 2021. 1–307. Available from: http://synap.kei.re.kr:8080/SynapDocViewServer/viewer/doc.html?key=56932060653a4b07b994158638d22a84&convType=html&convLocale=en_US&contextPath=/SynapDocViewServer/

33. Miller D, Sun K, Tao L, et al. Ammonia and methane dairy emission plumes in the San Joaquin Valley of California from individual feedlot to regional scales. J. Geophys. Res. Atmos. 2015;120(18)9718–38. https://doi.org/10.1002/2015JD023241

34. Sarwar G, Corsi R, Kinney K, et al. Measurements of ammonia emissions from oak and pine forests and development of a non-industrial ammonia emissions inventory in texas. Atmos. Environ. 2005;39(37)7137–53. https://doi.org/10.1016/j.atmosenv.2005.08.016

35. Owoade K, Hopke P, Olise F, et al. Source apportionment analyses for fine (PM_2.5) and coarse (PM_2.5–10) mode particulate matter (PM) measured in an urban area in southwestern Nigeria. Atmos. Pollut. Res. 2016;7(5)843–57. https://doi.org/10.1016/j.apr.2016.04.006

36. Zíková N, Wang Y, Yang F, et al. On the source contribution to Beijing PM_2.5 concentrations. Atmos. Environ. 2016;134:84–95. https://doi.org/10.1016/j.atmosenv.2016.03.047

37. NAEIR. National Air Pollutant Emissions in 2020 and Recalculations of National Air Pollutant Emissions for The Past Year 2016 to 2019. 2023. Available from: https://www.air.go.kr/capss/emission/sido.do?menuId=31

38. Heo J, Wu B, Abdeen Z, et al. Source apportionments of ambient fine particulate matter in Israeli, Jordanian, and Palestinian cities. Environ. Pollut. 2017;225:1–11. https://doi.org/10.1016/j.envpol.2017.01.081

39. Deng J, Zhang Y, Qiu Y, et al. Source apportionment of PM_2.5 at the Lin’an regional background site in China with three receptor models. Atmos. Res. 2018;202:23–32. https://doi.org/10.1016/j.atmosres.2017.11.017

40. Hopke P, Dai Q, Li L, Feng Y. Global review of recent source apportionments for airborne particulate matter. Sci. Total Environ. 2020;740:140091. https://doi.org/10.1016/j.scitotenv.2020.140091

41. Park E, Heo J, Kim H, Yi S. Long term trends of chemical constituents and source contributions of PM_2.5 in Seoul. Chemosphere. 2020;251:126371. https://doi.org/10.1016/j.chemosphere.2020.126371

42. Khan J, Sun L, Tian Y, Shi G, Feng Y. Chemical characterization and source apportionment of PM₁ and PM_2.5 in Tianjin, China: Impacts of biomass burning and primary biogenic sources. J. Environ. Sci. 2021;99:196–209. https://doi.org/10.1016/j.jes.2020.06.027

43. Agarwal A, Satsangi A, Lakhani A, Kumari K. Seasonal and spatial variability of secondary inorganic aerosols in PM_2.5 at Agra: Source apportionment through receptor models. Chemosphere. 2020;242:125132. https://doi.org/10.1016/j.chemosphere.2019.125132

44. Lee G, Kim M, Park D, Yoo C. Fine Particulate Matter (PM_2.5) Sources and Its Individual Contribution Estimation Using a Positive Matrix Factorization Model. Toxics. 2023;11(1)69. https://doi.org/10.3390/toxics11010069

45. Zhang Y, Sheesley R, Schauer J, et al. Source apportionment of primary and secondary organic aerosols using positive matrix factorization (PMF) of molecular markers. Atmos. Environ. 2009;43(34)5567–5574. https://doi.org/10.1016/j.atmosenv.2009.02.047

46. Paatero P. Least squares formulation of robust non-negative factor analysis. Chemometrics Intell. Lab. Syst. 1997;37(1)23–35. https://doi.org/10.1016/S0169-7439(96)00044-5

47. Paatero P, Hopke P. Utilizing wind direction and wind speed as independent variables in multilinear receptor modeling studies. Chemometrics Intell. Lab. Syst. 2000;60(1–2)25–41. https://doi.org/10.1016/S0169-7439(01)00183-6

48. Paatero P, Tapper U. Positive matrix factorization: A non-negative factor model with optimal utilization of error estimates of data values. Environmetrics. 1994;5(2)111–26. https://doi.org/10.1002/env.3170050203

49. Song Y, Zhang Y, Xie S, et al. Source apportionment of PM_2.5 in Beijing by positive matrix factorization. Atmos. Environ. 2006;40(8)1526–37. https://doi.org/10.1016/j.atmosenv.2005.10.039

50. Polissar A, Hopke P, Paatero P, Malm W, Sisler J. Atmospheric aerosol over Alaska: 2. Elemental composition and sources. J. Geophys. Res.-Atmos. 1998;103(D15)19045–57. https://doi.org/10.1029/98JD01212

51. Kim E, Hopke P, Edgerton E. Source identification of Atlanta aerosol by positive matrix factorization. J. Geophys. Res.-Atmos. 2003;53(6)731–9. https://doi.org/10.1080/10473289.2003.10466209

52. Kim E, Hopke P. Source characterization of ambient fine particles at multiple sites in the Seattle area. Atmos. Environ. 2008;42(24)6047–56. https://doi.org/10.1016/j.atmosenv.2008.03.032

53. Uria-Tellaetxe I, Carslaw D. Conditional bivariate probability function for source identification. Environ. Modell. Softw. 2014;59:1–9. https://doi.org/10.1016/j.envsoft.2014.05.002

54. Bae M, Schwab J, Chen W, et al. Identifying pollutant source directions using multiple analysis methods at a rural location in New York. Atmos. Environ. 2011;45(15)2531–40. https://doi.org/10.1016/j.atmosenv.2011.02.020

55. Carslaw D, Ropkins K. Openair—an R package for air quality data analysis. Environ. Modell. Softw. 2012;27:52–61. https://doi.org/10.1016/j.envsoft.2011.09.008

56. Kim E, Hopke P. Comparison between conditional probability function and nonparametric regression for fine particle source directions. Atmos. Environ. 2004;38(28)4667–73. https://doi.org/10.1016/j.atmosenv.2004.05.035

57. Kim E, Hopke P. Source apportionment of fine particles in Washington, DC, utilizing temperature-resolved carbon fractions. J. Air Waste Manage. Assoc. 2004;54(7)773–85. https://doi.org/10.1080/10473289.2004.10470948

58. Watson J, Chow J, Fujita E, et al. Protocol for applying and validating the CMB model for PM_2.5 and VOC. US Environmental Protection Agency; 2004;1–173.

59. Kim J, Choi D, Koo Y, Lee J, Park H. Analysis of domestic and foreign contributions using DDM in CMAQ during particulate matter episode period of February 2014 in Seoul. J. Korean Soc. Atmos. Environ. 2016;32(1)82–99. https://doi.org/10.5572/KOSAE.2016.32.1.082

60. Kim H, Kim E, Bae C, et al. Regional contributions to particulate matter concentration in the Seoul metropolitan area, South Korea: seasonal variation and sensitivity to meteorology and emissions inventory. Atmos. Chem. Phys. 2017;17(17)10315–32. https://doi.org/10.5194/acp-17-10315-2017

61. Zhang F, Wang Z, Cheng H, et al. Seasonal variations and chemical characteristics of PM_2.5 in Wuhan, central China. Sci. Total Environ. 2015;518:97–105. https://doi.org/10.1016/j.scitotenv.2015.02.054

62. Sun Y, Zhuang G, Tang A, Wang Y, An Z. Chemical characteristics of PM_2.5 and PM₁₀ in haze− fog episodes in Beijing. Environ. Sci. Technol. 2006;40(10)3148–55. https://doi.org/10.1021/es051533g

63. Wang Y, Zhuang G, Sun Y, An Z. The variation of characteristics and formation mechanisms of aerosols in dust, haze, and clear days in Beijing. Atmos. Environ. 2006;40(34)6579–91. https://doi.org/10.1016/j.atmosenv.2006.05.066

64. Liu B, Wu J, Zhang J, et al. Characterization and source apportionment of PM_2.5 based on error estimation from EPA PMF 5.0 model at a medium city in China. Environ. Pollut. 2017;222:10–22. https://doi.org/10.1016/j.envpol.2017.01.005

65. Turpin B, Huntzicker J. Secondary formation of organic aerosol in the Los Angeles Basin: a descriptive analysis of organic and elemental carbon concentrations. Atmos. Environ. A, Gen. Top. 1991;25(2)207–15. https://doi.org/10.1016/0960-1686(91)90291-E

66. Gentner D, Isaacman G, Worton D, et al. Elucidating secondary organic aerosol from diesel and gasoline vehicles through detailed characterization of organic carbon emissions. Proc. Natl. Acad. Sci. 2012;109(45)18318–23. https://doi.org/10.1073/pnas.1212272109

67. Yang Y, Lee S, Park B, et al. Estimation of heavy metal contamination by PM₁₀ inflow pathways while Asian dust in Gwangju. J. Environ. Sci. Int. 2020;29(1)55–68. https://doi.org/10.5322/JESI.2020.29.1.55

68. Song X, Polissar A, Hopke P. Sources of fine particle composition in the northeastern US. Atmos. Environ. 2001;35(31)5277–86. https://doi.org/10.1016/S1352-2310(01)00338-7

69. Chow J, Liu C, Cassmassi J, et al. A neighborhood-scale study of PM₁₀ source contributions in Rubidoux, California. Atmos. Environ. A, Gen. Top. 1992;26(4)693–706. https://doi.org/10.1016/0960-1686(92)90181-J

70. Chang Y, Carmichael G, Kurita H, Ueda H. An investigation of the formation of ambient NH₄NO₃ aerosol. Atmos. Environ. 1986;20(10)1969–77. https://doi.org/10.1016/0004-6981(86)90338-0

71. Ogulei D, Hopke P, Zhou L, et al. Receptor modeling for multiple time resolved species: the Baltimore supersite. Atmos. Environ. 2005;39(20)3751–62. https://doi.org/10.1016/j.atmosenv.2005.03.012

72. Ogulei D, Hopke P, Zhou L, et al. Source apportionment of Baltimore aerosol from combined size distribution and chemical composition data. Atmos. Environ. 2006;40:396–410. https://doi.org/10.1016/j.atmosenv.2005.11.075

73. Lee J, Hopke P. Apportioning sources of PM_2.5 in St. Louis, MO using speciation trends network data. Atmos. Environ. 2006;40:360–77. https://doi.org/10.1016/j.atmosenv.2005.11.074

74. Gildemeister A, Hopke P, Kim E. Sources of fine urban particulate matter in Detroit, MI. Chemosphere. 2007;69(7)1064–74. https://doi.org/10.1016/j.chemosphere.2007.04.027

75. Kim E, Larson T, Hopke P, et al. Source identification of PM_2.5 in an arid Northwest US City by positive matrix factorization. Atmos. Res. 2003;66(4)291–305. https://doi.org/10.1016/S0169-8095(03)00025-5

76. Seo J, Lim Y, Youn D, Kim J, Jin H. Synergistic enhancement of urban haze by nitrate uptake into transported hygroscopic particles in the Asian continental outflow. Atmos. Chem. Phys. 2020;20(12)7575–94. https://doi.org/10.5194/acp-20-7575-2020

77. Viana M, Kuhlbusch T, Querol X, et al. Source apportionment of particulate matter in Europe: a review of methods and results. J. Aerosol. Sci. 2008;39(10)827–49. https://doi.org/10.1016/j.jaerosci.2008.05.007

78. Mazzei F, D’alessandro A, Lucarelli F, et al. Characterization of particulate matter sources in an urban environment. Sci. Total Environ. 2008;401(1–3)81–9. https://doi.org/10.1016/j.scitotenv.2008.03.008

79. Pant P, Shukla A, Kohl S, et al. Characterization of ambient PM_2.5 at a pollution hotspot in New Delhi, India and inference of sources. Atmos. Environ. 2015;109:178–89. https://doi.org/10.1016/j.atmosenv.2015.02.074

80. Huang X, Tang G, Zhang J, et al. Characteristics of PM_2.5 pollution in Beijing after the improvement of air quality. J. Environ. Sci. 2021;100:1–10. https://doi.org/10.1016/j.jes.2020.06.004

81. Nigam A, Welch W, Wayne Miller J, Cocher D. Effect of fuel sulphur content and control technology on PM emission from ship’s auxiliary engine. In : Proceeding international aerosol conference; 10–15 September 2006; St Paul, USA. p. 1531–1532.

82. Harrison R, Smith D, Luhana L. Source apportionment of atmospheric polycyclic aromatic hydrocarbons collected from an urban location in Birmingham, UK. Environ. Sci. Technol. 1996;30(3)825–32. https://doi.org/10.1021/es950252d

83. Willis R, Britt P, Lewis C. Workshop on UNMIX and PMF As Applied to PM_2.5 . US Environmental Protection Agency; 2000. p. 1–32.

84. Lee J, Yoshida Y, Turpin B, et al. Identification of sources contributing to mid-Atlantic regional aerosol. J. Air Waste Manage. Assoc. 2002;52(10)1186–205. https://doi.org/10.1080/10473289.2002.10470850

85. Do T, Vuong Q, Lee S, Song C, Choi S. Comparison of characteristics and sources of water-soluble inorganic ions, trace elements, and polycyclic aromatic hydrocarbons in PM_2.5 on polluted and normal days in Ulsan, South Korea. Air Qual. Atmos. Health. 2023;16(6)1239–56. https://doi.org/10.1007/s11869-023-01340-1

86. Rai P, Chakraborty A, Mandariya A, Gupta T. Composition and source apportionment of PM₁ at urban site Kanpur in India using PMF coupled with CBPF. Atmos. Res. 2016;178:506–20. https://doi.org/10.1016/j.atmosres.2016.04.015

87. Gillies J, Gertler A, Sagebiel J, Dippel W. On-road particulate matter (PM_2.5 and PM₁₀) emissions in the Sepulveda Tunnel, Los Angeles, California. Environ. Sci. Technol. 2001;35(6)1054–63. https://doi.org/10.1021/es991320p

88. Sternbeck J, Sjödin Å, Andréasson K. Metal emissions from road traffic and the influence of resuspension—results from two tunnel studies. Atmos. Environ. 2002;36(30)4735–44. https://doi.org/10.1016/S1352-2310(02)00561-7

89. Lough G, Schauer J, Park J, et al. Emissions of metals associated with motor vehicle roadways. Environ. Sci. Technol. 2005;39(3)826–36. https://doi.org/10.1021/es048715f

90. Zhang Q, Jimenez J, Canagaratna M, et al. Understanding atmospheric organic aerosols via factor analysis of aerosol mass spectrometry: a review. Anal. Bioanal. Chem. 2011;401:3045–67. https://doi.org/10.1007/s00216-011-5355-y

91. Dai Q, Liu B, Bi X, et al. Dispersion normalized PMF provides insights into the significant changes in source contributions to PM_2.5 after the COVID-19 outbreak. Environ. Sci. Technol. 2020;54(16)9917–27. https://doi.org/10.1021/acs.est.0c02776

92. Moosmüller H, Gillies J, Rogers C, et al. Particulate emission rates for unpaved shoulders along a paved road. J. Air Waste Manage. Assoc. 1998;48(5)398–407. https://doi.org/10.1080/10473289.1998.10463694

93. Liu W, Hopke P, Han Y, et al. Application of receptor modeling to atmospheric constituents at Potsdam and Stockton, NY. Atmos. Environ. 2003;37(36)4997–5007. https://doi.org/10.1016/j.atmosenv.2003.08.036

94. Simoneit B, Kobayashi M, Mochida M, et al. Composition and major sources of organic compounds of aerosol particulate matter sampled during the ACE-Asia campaign. J. Geophys. Res.-Atmos. 2004. 109D19 https://doi.org/10.1029/2004JD004598

95. Heo J, Hopke P, Yi S. Source apportionment of PM_2.5 in Seoul, Korea. Atmos. Chem. Phys. 2009;9(14)4957–71. https://doi.org/10.5194/acp-9-4957-2009

96. Andreae M, Merlet P. Emission of trace gases and aerosols from biomass burning. Glob. Biogeochem. Cycle. 2001;15(4)955–66. https://doi.org/10.1029/2000GB001382

97. Birch M, Cary R. Elemental carbon-based method for monitoring occupational exposures to particulate diesel exhaust. Aerosol Sci. Technol. 1996;25(3)221–41. https://doi.org/10.1080/02786829608965393

98. Chow J, Watson J, Pritchett L, et al. The DRI thermal/optical reflectance carbon analysis system: description, evaluation and applications in US air quality studies. Atmos. Environ. A, Gen. Top. 1993;27(8)1185–201. https://doi.org/10.1016/0960-1686(93)90245-T

99. Kang C, Lee H, Kang B, Lee S, Young S. Chemical characteristics of acidic gas pollutants and PM_2.5 species during hazy episodes in Seoul, South Korea. Atmos. Environ. 2004;38(28)4749–60. https://doi.org/10.1016/j.atmosenv.2004.05.007

100. Cheng H, Deng Z, Chakraborty P, et al. A comparison study of atmospheric polycyclic aromatic hydrocarbons in three Indian cities using PUF disk passive air samplers. Atmos. Environ. 2013;73:16–21. https://doi.org/10.1016/j.atmosenv.2013.03.001

101. Jeong C, Evans G, Dann T, et al. Influence of biomass burning on wintertime fine particulate matter: Source contribution at a valley site in rural British Columbia. Atmos. Environ. 2008;42(16)3684–99. https://doi.org/10.1016/j.atmosenv.2008.01.006

102. Magd H, Karyamsetty H, Al-Asmi K. Coronavirus Disease Incidence Resonance with Coastline Dynamics: An Evaluation on Global Resurgence of the Pandemic. Emerg. Sci. J. 2023;7(5)1827–42. http://dx.doi.org/10.28991/ESJ-2023-07-05-024

Fig. 1

Sampling Locations in SMA and GRA.

Fig. 2

Time series of daily and seasonal PM_2.5 concentration in SMA and GRA; (a) Daily average PM_2.5 in SMA, (b) Daily average PM_2.5 in GRA, (c) Seasonal average PM_2.5 in SMA and GRA.

Fig. 3

Mass concentrations of PM_2.5 predicted from PMF modeling versus measured mass concentrations in SMA (a) and GRA (b).

Fig. 4

PMF results of PM_2.5 in SMA (a) and GRA (b).

Fig. 5

Source profile resolved from PM_2.5 in the study area, (a) SMA, (b) GRA.

Fig. 6

CBPF plots for the contributor of PM_2.5 in the study area, (a) CBPF plots for the contributor of PM_2.5 in SMA, (b) CBPF plots for the contributor of PM_2.5 in GRA.