Particulate matter (PM) in the atmosphere, which absorbs or diffuses the sun’s rays, disturbs the global thermal balance (Ramanathan et al., 2001). Moreover, it causes climate change by reacting as a cloud condensation nucleus and generating acidification or eutrophication of forest and surface water, which serves as a means of transportation of air pollutants (Suzumura, 2004). Particulate matter can be hazardous to human health because it contains harmful chemicals, including heavy metals, and results in visibility obstruction. Particulate matter has a harmful effect on the respiratory and cardiovascular systems, increasing the rate of death from related diseases with increasing concentrations of PM (Kang, 2006).

Particulate matter under a diameter of 10 μm is labelled PM10, and it can be categorized as ultrafine (about 0.1 μm), fine (0.1–2.5 μm), and coarse (2.5–10 μm). Furthermore, particulate matter under 2.5 μm is described as PM2.5 or fine particles (Whitby, 1978; Baron and Wileke, 2001).

Particulate matter occurs in the surroundings of human life, as well as in industrial activities. In the surroundings of life, particulate matter is seen in biological dust components, such as bacteria, as well as pollen during summer time, salt particles in seawater, cigarette smoke, automobile exhaust, smoke from burning garbage, and chimney smoke. Particulate matter can be discharged by processes of natural origin, such as forest fires, or by processes of anthropogenic genesis, such as smelting (Fraser et al., 2003).

Harmful PM occurs directly in the form of primary particles, such as in the process of combustion, or in the form of secondary particles, such as via conversion from gas to particles formed after release (Lighty et al., 2000). Primary particles are from such sources as factories, construction sites, and vehicles. Secondary particulate matter comprises the pollutants generated and changed by chemical reactions with light, oxygen, and so on in the atmosphere. These substances include sulfate, nitrate, non-volatile organics, and organic carbon, which are formed by chemical reactions. In addition, 30% to 60% of secondary particles are occupied by ionic substances (Hu et al., 2000, Kang et al., 2004, Lee et al., 2015).

The occurrence of particulate matter was generally conceptualized as yellow sand in the past, and this appeared in large quantities during spring. However, a high extent of particulate matter has recently been observed, regardless of seasons, in the context of long-distance transport from China and stable atmosphere conditions (Lee and Hills, 2003; Jeon, 2012). During spring and summer, continental high pressure is predominantly affected in the area of Asia, and this can be fatal to individuals with respiratory diseases. Heavy fine dust lasts over the long term, with a thin mist in a stable atmosphere (Lee and Hills, 2003).

Park reported (2021) that the levels of PM10 and PM2.5 represent the local distinction because urban structures, such as geological structure, stream environment, and industrial area, varied in Pusan. In cities, pollutant accumulations can be intensified when green areas are insufficient, and the dispersion of pollutants through the city, with its high building densities and impervious surface, is difficult. Thus, in cities, regional characteristics should be considered after the determination of the differentiated causes affecting air quality, as well as the average air quality in the city, and management methods for air environments should be established (Park 2021).

In this study, data on PM10 and PM2.5 from four seasons of the year were analyzed and interpreted to determine variations in their seasonal character in Yangju City, which is located in the north part of Gyeonggi Province, South Korea. The aim of gathering these data was to present a management plan to secure adequate atmosphere quality with consideration of regional circumstances. The emission characteristics of PM10 and PM2.5 on the different sources of potential pollutants and the contribution rate with the neighboring city by air quality modeling were analyzed.

2.1. Data Collection

The measurement data were collected over a 2-year period in 2018 to 2019 via the website of Air Korea operated with the Korea Environment Corporation, which is an air pollution data system in Korea (https://www.airkorea.or.kr). The official measurement sites for PM2.5 and PM10 are the Baekseok-eup (BS) and Go-eup (GO) monitoring stations (Fig. 1), which were installed in 2005 and 2017, respectively. The BS monitoring station was originally installed at Gwangjeok-myeon, but it was then moved and reinstalled at the current location in 2015. Around the GO site, there is a residential area, and a four-lane road is located nearby. Around the north side of the BS monitoring site, there is an industrial complex and places of business massed in is an area of 1 km. In addition, residential and commercial areas are located nearby. Meteorological data, such as information on the direction and velocity of the wind, are also taken from both sites.

Figure 1. Location of the air quality monitoring stations in Yangju.

In this study, two additional monitoring stations; the Gumjun (GJ) industrial complex and the Bongyang-dong (BY) community hall were set up. They were consistently measured to verify that the measurements of PM concentrations from the other locations. The β-ray absorption method was used over the course of a week in December 2019, January 2020, February 2020, and May 2020 to determine the concentrations of PM2.5 and PM10; measurements were not taken on rainy days. The cities of Paju (Geumchon dong), Pocheon (Seondan-dong), Uijeongbu (Uijeongbu 1-dong), and Dongducheon (Bosan dong) were selected to compare the influence of PM levels in Yangju from neighboring cities.

2.2 Release Characteristics and Contribution

Released amounts of PM2.5 and PM10 were taken via the Clean Air Policy Support System (CAPSS) 2016, supplied by the National Institute of Environmental Research (NIER). This system classifies sources of particulate matter into the industry, domestic, energy, road, fugitive dust, agriculture, and non-road categories.

In this study, atmosphere modeling was performed with Weather Research and Forecast (WRF), community multiscale air quality (CMAQ), and SMOKE (Sparse Matrix Kernel Emission). The WRF v. 5.0.2 meteorological model was applied, and boundary line data and initial input data were based on National Centers for Environmental Prediction (NCEP) analysis meteorological data. The chemical mechanism applied was carbon bond-V (CB-05), and Aerosol module version 5 (AERO5) was applied for the estimation of aerosol concentration in CMAQ. Numerical simulation covers all Korean peninsula regions, but the focus area for modeling was latitude 38° and longitude 126°, with a term of 7 months (October 1, 2018 to April 30, 2019).

The contributions of PM2.5 and PM10 among the regions were enumerated using the brute-force method, which is a conventional algorithm of sensitivity analysis. Analysis of the contribution of PM occurrence for businesses of types 4 and 5 and neighboring sites proceeded at the four stations (BS, GO, BY, GJ) via air quality modeling.

The artificial release amount of pollutants by CMAQ was based on the data of Intex-B made public by NASA in 2006 for eastern Asia, excluding North Korea, and CAPSS 2016 on units of a 1 km lattice for South Korea. The naturally released amount was enumerated with the Biogenic Emission Inventory System (BEIS). Input data were collected via the Environmental Geographic Information Service (EGIS) of the Ministry of Environment in Korea and the Forest Geographic Information System (FGIS) by the Korea Forest Service. Foliar density and the discharge coefficient on vegetation were applied according to the indicated values in BEIS v. 3.14.

3.1. Meteorological and Geological Conditions

Yangju City has the following characteristics: Flatlands and hilly areas are situated around Bulkok Mountain and Dorak Mountain. The Chonbo mountain range is located on the eastern side. The Cheongdam stream, Gokneung stream, and Hongjuk stream flow to the north side and enter the Hantan River. The city was characteristically surrounded by mountains. Fine dust did not disperse much and accumulated in the basin area, obstructing by the mountains.

The wind direction in 2019 was often a west wind series in the first quarter (January, February) and fourth quarter (October, November, December). However, in the second and third quarters (April to September), the occurrence frequency of the west wind series decreased. The atmosphere in Yangju City seems to be influenced by the dust introduced from China from October to April. Influence from outside of Korea was analyzed as low in summer to fall when there was a low frequency of west wind series.

The highest velocities of wind in January were 2.6 and 3.4 m/s in 2018 and 2019. The durations of fine dust watches and warnings were 28 and 4 days, respectively, in 2019 (Table 2). Fine dust warnings were only issued in January and March in 2019. In January, during a period with low wind velocity, a significant concentration of fine dust was recorded. PM generated around the area did not diffuse, and it remained in the atmosphere. For Yangju City, based on the days of fine dust watches and warnings, high concentration occurrences of fine dusts were characterized as west wind series or a wind velocity under 1 m/s. Kim et al. (2013) reported that low mean daily wind velocities were specifically observed, and the daily maximum temperature was higher on days showing high concentrations of particulate matter.

Table 1 . Analysis of wind rose in Yangju (Data from seven automatic weather stations in Yangju).

	1st quarter	2nd quarter	3rd quarter	4th quarter
Baekseok-eup	NWW	SWS	SWS	NEN
Eunhyeon-myeon	SWW	SWW	NEN	W
Nam-myeon	W	SWW	NEN	W
Ganap-ri	NWN	SW	N	NWN
Jangheung-myeon	NWN	NWN	N	NWN
Yangju 1-dong	NW	NW	SWS	NW
Hoecheon 1-dong	NW	W	SWS	NW

Table 2 . Meteorological condition with fine dust watches and warnings in 2019 (Data from https://www.airkorea.or.kr).

Contingency action	Date	BS			GO
		Con. ㎍/㎥	Direction of the wind	Wind Velocity (m/s)	㎍/㎥	Direction of the wind	Wind Velocity (m/s)
	2019-Jan.-05	49	SW	0.4	56	NW	1.3
	2019-Jan.-12	91	NW	0.3	96	NW	0.5
	2019-Jan.-13	84	N,W	0.1	96	NW	0.4
	2019-Jan.-14	135	SW, N	0.2	152	NW	0.2
	2019-Jan.-15	91	NW	0.6	106	NW	1.1
	2019-Jan.-19	86	NE	0.2	84	NW	0.3
	2019-Jan.-20	45	NW	0.8	45	NW	1.5
	2019-Jan.-23	70	NW	0.4	59	NW	0.9
	2019-Feb.-07	32	NW	0.8	30	NW	1.9
	2019-Feb.-21	71	NE, W	0.2	86	W	0.4
	2019-Feb.-22	52	NE	0.4	58	NW	0.5
	2019-Feb.-28	66	NW	0.4	75	NW	0.6
	2019-Mar.-01	77	NW	0.4	85	NW, W	0.7
Watching	2019-Mar.-02	79	NE	0.3	94	SW	0.5
	2019-Mar.-03	71	NW	0.3	72	NW	0.4
	2019-Mar.-04	104	NW	0.4	93	W	0.7
	2019-Mar.-05	153	SW	1.1	151	SW	0.7
	2019-Mar.-06	108	NW	0.9	109	NW	0.6
	2019-Mar.-07	40	NW,W	1.6	48	NW	1
	2019-Mar.-12	53	SW, NW	0.8	52	NW	1.3
	2019-Mar.-20	78	SW	0.4	75	S, SW	0.4
	2019-Mar.-21	26	SW, NW	0.9	22	NW	1.5
	2019-Mar.-27	56	SW	1.7	55	SW	1.9
	2019-Mar.-28	49	NW	0.6	49	NW	1.2
	2019-May-04	49	NW, SW	0.5	47	NW	0.6
	2019-May-25	53	SW	0.9	54	SW	0.8
	2019-Dec.-10	70	W, NW	0.8	64	NE	0.5
	2019-Dec.-11	46	SW, NW	2.6	46	W	1.9
	2019-Jan.-14	135	NW, N	0.2	152	NW	0.2
Warning	2019-Jan.-15	91	NW	0.6	106	NW	1.1
	2019-Mar.-05	153	SW	1.1	151	SW	0.7
	2019-Mar.-06	108	NW	0.9	109	NW	0.6

The two existing monitoring stations (BS, GO) are located in the southern part of Yangju. At these stations, it was found that various discharge sources of particulate matter were scattered from such sources as small workplaces, barns, and farmland in the northern parts of the city. Both monitoring stations are significantly distant from the pollutant discharge distribution area. In this study, to verify the representativeness of the public monitoring stations, two additional stations were installed in the upper regions of this city to check the representative values of PM2.5 and PM10 taken from the BS and GO. The values of ρ for PM10 and PM2.5 in BS and GO were over 0.9, which means that data had a strong correlation between monitoring stations (Fig. 2). Therefore, current data from the BS and GO monitoring stations can be used as representative values for particulate matter in the city.

Figure 2. Pearson’s correlation coefficient based on concentrations of PM2.5 (a) and PM10 (b) among four monitoring sites in Yangju.

The annual data for PM2.5 and PM10 from 2019 in Yangju were compared with data from neighboring cities, such as Dondochon, Paju, and Pocheon. The annual mean concentrations of PM2.5 were 25, 28, 26 ㎍/m³ for Dondochon, Paju, and Pocheon, respectively. The mean concentration

of PM2.5 for both BS and GO was 26 ㎍/m³ in Yangju. This is similar to the level of the neighboring city. Pearson’s correlation coefficient (ρ) was over 0.9 for all cities compared in this study. This result indicated that PM2.5 was related to the values of the other cities (Fig. 3). This tendency suggested that the concentration variation of PM2.5 in this city was affected by those of neighboring cities. The distribution of PM2.5 shows a zone type rather than a spot type. The coefficient was over 0.9 with the neighboring cities, and it mostly showed a tendency whereby the smaller the distance between two cities was, the higher the correlation coefficient would be. The value of ρ for PM10 in Yangju indicated that measurement values in BS and GO were highly correlated with those in the other cities.

Figure 3. Pearson’s correlation coefficient based on concentrations of PM2.5 and PM10 among four administrative units in Kyunggi-do.

3.2. Distribution characteristics of PM2.5 and PM10

The mean values of PM2.5 and PM10 at GO and BS from 2010 to 2019 are presented in Fig. 4. The concentrations of PM10 for both years (2018 and 2019) were under 50㎍/m³. However, all annual mean values of PM 10 were shown to be over the legal limits from 2010 and 2017. Moreover, in 2017 to 2019, PM2.5 was over the domestic legal limit of 15 ㎍/m³. The current conditions are problematic, and in the future, management and a consistent focus on this issue should still be required. This is the case even though recent values of PM2.5 and PM10 have been observed to be decreased. These lower values were due to a decreased amount of long distance transport from abroad stemming from the effects of the Covid-19 pandemic. Post-pandemic, they are expected to rise again.

Figure 4. Annual mean values of PM2.5 and PM10 observed in Yangju (2010–2019). (Data from an air pollution data system in Korea :https://www.airkorea.or.kr)

Monthly concentration variations of PM2.5 and PM10 are presented in Fig. 5. The highest values of PM2.5 and PM10 in 2019 at BS were 46 and 81 ㎍/m³, whereas the highest concentrations at GO for the same year were 50

Figure 5. Monthly variation of average PM2.5 and PM10 concentrations at GO and BS in 2019.

and 86 ㎍/m³. The highest values were observed in January at both sites. The values of particulate matter decreased from July to September, likely affected by the high amount of rainfall, and they increased sharply in November.

The variation in PM2.5 was not large at either site. The mean value of PM2.5 in 2019 was 26 ㎍/m³ for both sites, which was the same as that for the province of Kyung-do. Seasonal mean concentrations of PM2.5 were recorded in the order of winter > spring > summer > fall at BS and winter > spring > fall > summer at GO. Seasonal mean concentrations of PM10 at BS and GO were observed in the order of winter > spring > fall > summer. When water-soluble ionic substances and soil substances composed of particulate matter dropped to the surface of the ground with precipitation, the content of PM in the atmosphere rapidly decreased (Mouli et al., 2005).

Low levels of particulate matter in the summer season were caused by rainout and washout. In rainout, PM2.5 entered the upper atmosphere layer; it was then removed via seed making raindrops. In washout, PM2.5 was removed from the air by frequent rainfall. Ouyang et al. (2015) reported that particulate matter showed a negative correlation with accumulated precipitation amounts in China, a country appearing to have a high level of PM. Moreover, Park (2021) reported that PM10 was more affected by precipitation than PM2.5 was. Testing showed that there were higher levels of PM10 in winter than there were in spring. Indeed, the highest levels were shown in winter, followed by temperature inversion in winter or spring and fall, which have high daily temperatures that vary during the night time and in which the surface of the earth is cooled rapidly, a phenomenon often known to occur in the geological basin. Yangju City, which has a mountain range and mountains with high peaks is more likely to have a high PM concentration.

The hourly variation for PM2.5 in January 2019 is presented in Fig. 6. A rapid increase in PM2.5 was shown in the morning. Until 7 a.m., the mean levels of PM2.5 were 46 and 50 ㎍/m³ for BS and GO, respectively. The values of PM2.5 then increased from 7 a.m. The maximum levels were 78 and 93 ㎍/m³ for BS and GO, respectively, and these levels were recorded at 11 a.m. The values then gradually decreased until 6 p.m. Following this, the level of PM2.5 was maintained between 8 p.m. and 7 a.m. the next day. These levels are affected by vehicular traffic, especially during morning rush hours and the operation time of industrial facilities responsible for air emissions. In previous literature, the first peak was shown between 7 a.m. and 9 a.m. in Shanghai and New York, which is different from Yangju. The highest point was delayed to 2 h later, coinciding with the starting time of industrial activities. In Taiwan, two peaks were shown in a day; the first peak was at 10 a.m., and it was explained by emissions from cars during rush hour; the second peak was at 1 p.m., and it was explained by weak wind velocity and photochemical reactions (Lee et al., 2006).

Figure 6. Hourly variation of PM2.5 concentration in Jan. 2019 at GO and BS in 2019.

The variation of seasonal mean levels for PM2.5/PM10 in 2018 and 2019 are presented in Table 3. The highest values at BS were 0.81 and 0.75 in 2018 and 2019, respectively. These appeared in the summer season. The highest values at GO were 0.57 and 0.58 in 2018 and 2019, respectively. These emerged in the winter season. The values of PM2.5/PM10 at BS were higher than those at GO.

Table 3 . Seasonal mean of PM2.5/PM10 at Go-eup (GO) and Backseok-eup (BS) in 2018 and 2019.

	2018				2019
	Spring	Summer	Fall	Winter	Spring	Summer	Fall	Winter
BS	0.61	0.81	0.56	0.52	0.58	0.75	0.61	0.6
GO	0.52	0.51	0.42	0.57	0.52	0.57	0.52	0.58

The mean values throughout all seasons for PM2.5/PM10 for 2018 and 2019 were 0.63 and 0.64, respectively, for BS. Conversely, the levels were 0.51 and 0.55 for GO. Thus, the values of PM2.5/PM10 at BS were higher than those at GO. In the spring, the values observed were shown to be lowest owing to the influence of the significant size of particles, such as yellow dust. Querol et al. (2004) reported that PM2.5/PM10 levels in industrial areas were higher than those in residential areas. The relatively high values of PM2.5/PM10 detected during the summer at BS may have been due to the contribution of secondary aerosols, such as sulfate and nitrate, and the phenomenon of re-scattering by strong wind, although this does not often occur (Tecer et al., 2008).

To illustrate the variations in PM2.5/PM10, the monthly values of PM2.5/PM10 in 2019 are presented in Fig. 7. The lowest values were shown in May, with values of 0.48 and 0.42 for BS and GO, respectively. Moreover, the highest values at BS in July reflected the effect of precipitation during the rainy season. Both sites showed a similar tendency for the monthly variation of PM2.5/PM10. The value of PM2.5/PM10 was high in July, and the values of PM2.5/PM10 at BS were higher than those at GO for most months. The mean values of PM2.5/PM10 in 2018 and 2019 were 0.57 and 0.59, respectively. The mean value of PM2.5/PM10 in Daegu City was reported to be 0.55 (Park and Lim, 2006), the industrial area in Pusan was 0.73 (Jeon and Hwang, 2014), and that in Jeju was 0.68 (Lee et al., 2020).

Figure 7. Monthly variation of the PM2.5/PM10 coefficient at GO and BS in 2019.

Consistency among monitoring stations with the correlation coefficient was checked based on the results for each pollutant item (Fig. 8). PM2.5 was strongly correlated with PM10, and PM10 was positively related with the following pollutants: SO2, NO2, O3, and CO. PM2.5 was positively related with SO2, NO2, and CO. However, ozone had low relations to PM10 and PM2.5 (0.033 and -0.002, respectively).

Figure 8. Pearson’s correlation coefficient among six atmosphere pollutants in Yangju in 2019.

3.3. Emissions Analysis Based on CAPSS

Based on CAPSS 2016 data supplied by the NIER, air pollutant emission sources were analyzed in Yangju City according to the eight following pollutants: NO_X, SO_X, NH3, Total. Suspended Particles (TSP), CO, VOC_S, PM2.5, and PM10 (Table 4). The total emission amount in 2016 was two times higher than that in 2013. The highest emission amount was shown for CO. PM2.5 and PM10 reached 385.7 and 1,019.2 tons, respectively, in 2016. These levels are both approximately 1.2 times higher than those for PM2.5 and PM10 in 2015. The results reflected the increase of emissions by new town development in this city that started in 2014, as well as an increase in the number of vehicles. Major emission sources of PM10 and PM2.5 were analyzed for fugitive dust and biomass burning in Yangju in CAPSS 2016, and those in Yangju both occupied 0.4% compared with national emissions of PM2.5 and PM10.

Table 4 . Air pollutant emissions in Yangju in 2016 (Unit: ton).

Discharge Sources	CO	NOx	SOx	TSP	PM-10	PM-2.5	VOC	NH3
Fuel Combustion -Energy production	703.3	386.3	4.5	16.3	16.3	16.3	95.3	23.1
Fuel Combustion -Non-industry	417	414.5	132.6	9.9	8.2	5	12.7	7.8
Fuel Combustion -Manufacturing industry	64.1	211.1	27.2	1.5	1.4	0.9	9.0	3.3
Industrial Process	0.2	9.8	6.4	0.3	0.2(295.6)+	0.2(354.7)+	11.8	66.2
Energy Transport and Storage	0.0	0.0	0.0	0.0	0.0	0.0	146.3	0.0
Solvent Use	0.0	0.0	0.0	0.0	0.0	0.0	2,263.6	0.0
Road Transport	1,046.9	2,677.9	1.2	66.2	66.2	60.9	241.8	23.2
Non-road Transport	446.3	944.7	0.6	50.9	50.9	46.8	120.3	0.5
Waste Disposal	19.6	126.2	32	3.5	2.5	2.1	531.5	0.1
Agriculture	0.0	0.0	0.0	0.0	0.0	0.0	0.0	1,831.6
Other Surface-pollutant Source	58.7	1.4	0.0	4.5	2.9	2.6	6.0	51.4
Fugitive Dust	0.0	0.0	0.0	2,305.4	714.9	110.6	0.0	0.0
Biomass Burning	1,344.7	36.5	0.7	242.3	155.7	140.3	271.0	0.1
Total	4100.8	4808.4	205.2	2700.8	1019.2(1373.6)+	385.7(681.2)+	3709.3(4556.9)+	2007.3

+: A revised value after the reflection of emissions from 4 and 5 types of businesses in this study.

In 2016, emissions of sources of carbon monoxide were mostly identified as incomplete combustion, especially from biomass combustion or transportation and energy industry combustion from power production facilities. In addition, the occurrence of non-road transport of PM2.5 and PM10 in 2016 was 2.8 times the amount found in 2013.

In summary, what the greatest pollutant sources for NO_X, SO_X, NH₃, TSP, CO, VOC_S, PM2.5, and PM10 were analyzed and found to be road transport, non-industry (fuel combustion), agriculture, fugitive dust, biomass burning, solvent use, biomass burning, and fugitive dust, respectively. In this study, to easily describe countermeasures in each field, the existing 13 categories were reorganized into the four following fields: industry, transport, agriculture and life, and fugitive dust.

The results found in this study were somewhat different from the tendency throughout the nation (Fig. 9). In Korea, the biggest emission source of PM2.5 was fugitive dust, whereas that of PM10 was industry. Kim et al. (2019) recommended creating forest in the city and planting trees along the street, as well as conducting research on the spatial extent and height of dust scattering. Cavanagh et al.(2009) reported that the concentration of PM was shown to have lower values from the outside to the inside of urban forests. The extent of PM2.5 removal has been found to according to vegetation type, in terms of whether the forest comprised needleleaf trees, broadleaf trees, or mixed forest (Nguyen et al. 2015).

Figure 9. Comparison of emissions for PM2.5 and PM10 from the different fields for the entire country and Yangju City.

In Yangju City, the released emission by transport was measured, and the number of cars registered increased from 2010 to the research period in this study. In particular, the number of trucks using diesel increased. Vehicles using diesel are well known to discharge much more emissions than those using gasoline. Metrological factors such as temperature and precipitation were negatively related with PM2.5, which increased with the increase of the real-time traffic congestion index in Seoul (Jin and Jin, 2021). Moreover, the concentration of high-rise buildings in the area created street canyons, where PM stays for long periods (Zwack et al., 2011). Fine dust increased with the conditions of low-velocity driving (40 km/h) and idling (Keuken et al., 2010). Practically, roundabout was shown to have a 2% reduction of fine dust release compared to a general crossroad (Gastaldi et al., 2014). Wind velocity was not constantly related to the level of fine dust (Giri et al., 2008).

The main compounds discharged were NO_X and CO in the transport field (Table 4). These are related to the development of new towns and increased numbers of excavators, bulldozers, and forklifts used in the construction field. Earthy materials enter construction sites via transport vehicles, and the materials can fall out of the vehicles and accumulate on the road. Many cars travel on the road, increasing the content of road fugitive dust. Microplastics from tires when driving cars and brake systems, as well as painting abrasion on road, can be sources of particulate matter.

For the field of agriculture and life, the releases of PM2.5 and PM10 were assessed as 150,072 and 169,256 kg (Table 5). These emissions can be generated, for instance, through by-product incineration during farming and tree fuel use. Biomass burning occupied the biggest share in terms of the release of PM2.5 in the field of agriculture and life; the biggest discharged source was charcoal kiln (67.2%). Other than this factor, the main causes were by-product incineration (25.7%), wood heating (3.2%), and open air-incineration (2.5%). These tendencies were similar to those for PM10.

Table 5 . Emission sources of various air pollutants in Yangju in 2016 in the field of agriculture and life (Unit: kg).

Items		PM-2.5	PM-10	TSP	NOx	SOx	CO	VOC	NH3
Fuel Combustion-Non-industry	commercial· public facilities	645	1,417	1,536	176,342	23,687	44,763	4,639	4,485
	residential facilities	4,350	6,721	8,247	237,069	108,899	371,957	7,999	3,288
	agricultural and livestock facilities	47	74	80	1,135	4	284	14	45
Waste Disposal	waste incineration	2,093	2,518	3,471	126,241	32,033	19,610	531,467	-
	other ways of disposing of waste	-	-	-	-	-	-	-	113
Agriculture	farmlands that use fertilizers	-	-	-	-	-	-	-	38,530
	livestock manure management	-	-	-	-	-	-	-	1,793,071
Other Surface-pollutant Source	natural resources, fires,	2,588	2,875	4,529	1,444	-	58,687	6,000	-
	animals	-	-	-	-	-	-	-	51,378
Biomass Burning	open burning	3,519	3,969	5,798	2,377	-	18,140	193,86	10
	crop residue incineration,	36,024	43,081	114,789	25,398	-	712,789	116,678	21
	grilled meat and fish	1,681	1,830	1,830	27	6	36	436	-
	wood stoves and boilers	4,449	6,774	14,409	6,280	199	195,325	57,396	21
	traditional fireplaces	327	407	616	1,884	33	21,510	6,016	6
	charcoal kilns.	94,349	99,590	104,832	566	440	396,894	71,076	-

Farmland occupied 19% of Yangju City. Agriculture byproducts are often incinerated during farming; these can be materials accumulated from pruning of fruit trees, dead plant matter from weather disasters or insect incursion, or rice straw remaining after harvesting. These were important release sources of particulate matter, and they were intensively released during early spring (February to March) and late fall (October to November). Therefore, such methods as executing crackdowns, providing consistent education, and expanding visiting education should be enforced. Incineration in open spaces should be monitored to eliminate waste in the general home.

Brown coal was often used in the construction field during the winter season, and civil complaints were made because of dust and bad smells. The application of brown coal in concrete curing can cause suffocation accidents for workers and aggravate air quality. Therefore, suitable guidelines and regulations should be followed for the use of this material.

For the industry area, data on air pollutant emissions for 4 and 5 types of business are not currently sufficient. Details of small-scale business are not included in the Stack Emission Management System (SEMS) database because reports on operating information are not mandatory for 4 and 5 types of business. Based on licensee event records and SEMS, there were 313 and 399 places of business where the sums of air pollutants per year were between 2 and 10 tons and under 2 tons, respectively, for 4 and 5 type businesses. The main types of business in this field are plastic and textile goods, rubber, and chemical goods.

The 4 and 5 types of businesses occupied 92.6% of the businesses causing air pollutant emission (Fig 10), and general methods should be prepared for the emission of air pollutants for small-scale businesses, which are not properly managed and supervised. The revised PM2.5 and PM10 values after the reflection of emissions from 4 and 5 types of businesses in this study were 681.2 and 4556.9 tons.

Figure 10. Distribution of 4 and 5 types of business in Yangju.

After reorganizing the data with the new classification groups, higher percentages of emission sources in terms of the level of PM10 were identified in fugitive dust (52.0%), industry (27.10%), agriculture and life (12.30%), and transport (8.50%). With the revised classification groups, higher percentages of emission sources in terms of the level of PM2.5 were identified in industry (45.9%), agriculture and life (22.0%), fugitive dust (16.20%), and transport (15.8%).

3.4. Simulation with WRF-CMAQ

To analyze the contribution of crucial pollutant factors for PM2.5 and PM10 at the four monitoring sites of GO, BS, GJ, and BY in Yangju City, the sites were evaluated for 4 and 5 types of businesses with WRF-CMAQ (Tables 6 and 7). In this study, the contribution was highest in November at the GJ site for PM2.5 and PM10. The highest concentrations of PM10 for each site were as follows: BS, 2.19 in November; GO, 1.60 ㎍/m³ in February; BY, 3.62㎍/m³ in January; and GJ, 8.29 ㎍/m³ in November. The highest contribution of PM10 was 29.85% for GJ, and the lowest contribution of PM10 was 1.12% for GO.

Table 6 . Monthly variation of the PM10 contribution of each site in Yangju for the emission of 4 and 5 types of businesses between Oct. 2018 and Apr. 2019; simulated using WRF-CMAQ.

Sites	Items	Oct.	Nov.	Dec.	Jan.	Feb.	Mar.	Apr.
BS	Con.(㎍/㎥)	1.83	2.19	1.12	0.79	0.91	0.44	0.56
	Contribution(%)	10.31	6.49	4.73	2.25	3.68	1.51	2.25
GO	Con.(㎍/㎥)	0.91	1.27	1.11	1.58	1.60	0.35	0.50
	Contribution(%)	5.14	3.75	4.13	4.03	5.87	1.12	1.94
BY	Con.(㎍/㎥)	1.60	2.03	1.96	3.62	2.34	1.34	1.30
	Contribution(%)	9.07	6.09	7.65	9.22	9.26	4.13	4.95
GJ	Con.(㎍/㎥)	4.46	8.29	5.33	4.91	4.71	2.95	3.74
	Contribution(%)	29.85	20.20	18.17	11.71	16.00	8.72	12.53

Table 7 . Variation of contribution of PM2.5 at each site in Yangju for the emission of 4 and 5 types of businesses between Oct. 2018 and Apr. 2019.

Sites	Items	Oct.	Nov.	Dec.	Jan.	Feb.	Mar.	Apr.
BS	Con.(㎍/㎥)	1.49	1.81	0.88	0.62	0.70	0.34	0.44
	Contribution(%)	11.67	7.27	4.82	2.27	3.68	1.48	2.37
GO	Con.(㎍/㎥)	0.71	1.01	0.87	1.24	1.26	0.27	0.39
	Contribution(%)	5.83	4.15	4.44	4.28	6.29	1.10	2.03
BY	Con.(㎍/㎥)	1.25	1.62	1.51	2.78	1.83	1.00	0.99
	Contribution(%)	10.21	6.71	8.00	9.54	9.70	4.00	5.06
GJ	Con.(㎍/㎥)	5.83	6.55	4.08	3.69	3.61	2.21	2.85
	Contribution(%)	32.02	21.36	18.33	11.45	16.08	8.28	12.61

The highest concentrations of PM2.5 were as follows: BS, 1.49 ㎍/m³ in October; GO, 1.26 ㎍/m³ in February; BY, 2.78 ㎍/m³ in January; and GJ, 6.55 ㎍/m³ in November. The highest contribution observed from PM2.5 was 32.02% for GJ in October, and the lowest one was 1.10%for GO in March. Although the contribution rate was the highest, the concentration was not highest in GJ, BY, or BS. Mostly, the contribution rate of PM2.5 was higher than that of PM10.

Variation of contribution of 4 and 5 types of in terms of times of the day are evaluated in Fig. 11. In GO, the concentrations of PM10 and PM2.5 were the highest between 0 a.m. and 1 a.m. At the other three sites, the highest concentration was shown between 7 a.m. and 8 a.m. The lowest concentration was shown between 2 p.m. and 3 p.m. for all sites. The reason the highest values of PM occurred in the morning was that pollutants were confined by the inversion layer.

Figure 11. Contributions to PM2.5 and PM10 from 4 and 5 types of business in Yangju with time variation as simulated by WRF-CMAQ.

Particulate matter in Yangju with the contribution of the neighboring cities is presented in Table 8. The mean contribution rates in the neighboring cities were over 2.21 (BS), 3.36 (GO) and 1.46 (BY) times higher than those for 4 and 5 types of business for PM10. The contribution rates of the neighboring cities for PM2.5 were more 1.95 (BS), 2.79 (GO), and 1.31 (BY) times higher than those in small-scale businesses. Only GJ showed a higher contribution rate for 4 and 5 types of business compared with other sites. Gong et al. (2021) indicated that PM2.5 in Chungcheong province contributed 17% to that of Gyeonggi province; they also mentioned that cooperative efforts to analyze and mitigate PM among neighboring local governments should be required to reduce particulate matter because PM2.5 was mutually influenced by adjacent local government.

Table 8 . Contribution of PM2.5 and PM10 from neighboring cities to Yangju; simulated using WRF-CMAQ.

Item		BS	GO	BY	GJ
PM-10 (㎍/㎥)	Con.	1.8	2.3	2.0	1.6
	Contribution rate (%)	9.3	12.1	10.4	6.8
	Max. Con.	3.53	4.78	3.89	2.80
	Max Contribution rate (%)	10.47	14.07	12.34	7.53
PM-2.5 (㎍/㎥)	Con.	2.5	3.5	3.0	2.2
	Contribution rate (%)	8.6	10.9	9.7	6.3
	Max. Con.	2.56	3.39	2.83	2.04
	Max Contribution rate (%)	10.28	13.91	1.68	6.94

3.5. Reduction Plan of Particulate Matter in Yangju

To reduce the occurrence of particulate matter in Yangju City, in this study, the main causes of PM were evaluated. Based on the results analyzed, suggestions were determined for a particulate matter reduction plan. These are described below.

Data on the emission amounts of type 4 and 5 workplaces should be gathered to address the missing CAPSS data from NIER in Korea. The database was built with mostly type 1 to 3 workplaces, which have a duty to register. In contrast, the status reports of some type 4 and 5 workplaces are autonomous. Many data are not added to the database in Yangju, and some workplaces have been identified with our investigation. Further surveillance and crackdowns should be followed; moreover, the current situation should be characterized and a detailed database set up for the establishment of air pollutant emission facilities, including for small-scale businesses. In Yangju, 4 and 5 types of businesses accounted for 92.6% of the air pollutant emission facility. Therefore, the contribution of PM from these businesses was significant in Yangju. The policy priority for reducing PM2.5 in this city should be as follows: industry > agriculture and life > fugitive dust > transport.

Yangju is a mixed urban and rural city; thus, illegal incineration in rural areas, such as farming waste incineration and open incineration, should be monitored. In the field of farming, the contribution of ammonia was high, and CAPSS identified emissions that were mostly related to livestock manure management. Mobile measurement vehicles with real-time measurement of air pollutants can be applied regularly in business-congested areas. In addition, relocation, or closure of business was continued with the construction of a new town. So, total inspection, including changes in the condition of the business, should be conducted in the city.

To prevent scattering of fine dust on the road, the operation of water sprinklers after selecting the road should be managed intensively. In the construction field, management services, such as offering sprinkling water facilities and installing soil proof nets, should be developed. Beyond this, control methods to counter fine dust in workplaces, such as construction waste treatment, cement production facilities, should be followed. In particular, it was determined by a spot survey that a considerable amount of dust scattering occurred during the process of loading and unloading material from vehicles.

Conversion to cleaner fuels should be required for businesses using boilers with solid fuels and bunker C fuel oil. A progression plan, such as withdrawing decrepit diesel cars and worn-out construction and agricultural machineries, as well as improving of conditions for public transportation, and bicycle infrastructure, could be implemented in the transport sector.

In the field of agriculture and life, incinerations were performed on the field for crops. Biodegradable bags should be disseminated, and farming waste collection and appropriate disposal systems should be prepared with consistent publicity and crackdowns. Control of wood-burning heaters and charcoal kilns should be performed.

In this study, occurrence sources for the reduction of particulate matter in Yangju were observed to determine suitable countermeasures that will fit the situation in this city. From the results, the following conclusions were drawn:

1.PM2.5 levels were shown in the order of winter > spring > summer > fall in BS and winter > spring > fall > summer in GO. For the monthly variation, the highest concentration of PM2.5 appeared in January and the lowest concentration of PM2.5 was observed in September. Over the day, 11 a.m. was shown to afford the highest level of particulate matter, and daily variation of PM2.5 was like that of PM10 in this study.

2.In terms of the characteristics of Yangju, 4 and 5 types of businesses occupy 92.6% of facilities discharging air pollutants. However, these businesses are currently off the government watch list, and stronger administration management should be followed, as well as preparation of systems for emission quantity counts with category establishments for the 4 and 5 types of businesses, to address this issue.

3.Two monitoring station in Yangju, BS and GO, are located on the southern side of the city. Therefore, the items of PM2.5 and PM10 were also analyzed at GJ and BY, which are located on the northern side of the city, to conduct a data representativeness check using correlation analysis. Pearson’s correlation coefficient (ρ) was over 0.87 for all sites compared in this study.

4.The order of priority of PM2.5 was industry > agriculture and life > fugitive dust > transport. According to this ranking, basic data accumulation and implementation of particulate matter reduction steps in each field should be undertaken by the local government. The regional emission characteristics and amount should be considered in setting up the reduction policy. However, percentages of contribution to the neighboring city for PM2.5 were shown to be between 6.3% and 10.9%in simulation with WRF-CMAQ.