The conditional bivariate probability function (CBPF), a receptor model, was used to identify local pollution sources of PM10 and SO2. The hourly data of PM10 showed the highest levels in April and May and the lowest in August and December. CBPF results indicated that the petrochemical industry and road traffic were the main local sources of PM10, while SO2 concentration was strongly influenced by the petrochemical industry.
The PSCF and cluster analysis results showed that the potential LRAT sources of PM10 were China in spring. The local impacts of PM10 and SO2 were greatest in summer and decreased in winter. The LRAT of PM10 was observed when high levels of PM from China occurred in spring.
This study can be a basis to identify local and long-distance sources of CAPs in other cities.
INTRODUCTION
- Particulate matter
- Sulfur dioxide
- Ulsan city
- Objectives of this study
In the world, PM10 and PM2.5 are taken into account, so many countries regulate the standard concentrations of particulate matter in the air environment. The regulation level of both PM10 and PM2.5 relates to the daily average concentration and the annual average concentration. Compared with other metropolitan cities, the amounts of PM10 and SO2 emissions in Ulsan are the highest.
Pollution characteristics of PM10 and SO2 are analyzed using hourly concentration data obtained from monitoring stations in Ulsan. Seasonal and monthly variations of PM10 and SO2 levels are analyzed and pollution characteristics are analyzed by region type such as industrial, residential and road area. The emissions of PM10 and SO2 from the stacks are estimated to be dispersed and affected according to CALPUFF, an air dispersion model.
Specific sources of PM10 and SO2 in Ulsan and their impact on residential areas are identified.
MATERIALS AND METHODS
- Monitoring stations
- Meteorological conditions in Ulsan
- Conditional bivariate probability function
- Clean Air Support System
- California puff model
- Backward trajectory analysis
- Conditional Inference Tree
A grid system is created based on the CALMET field for efficient calculation of wind motion and pollutant concentrations (Figure 11) (Scire et al., 2000). Backtracking analysis can be used to verify changes in wind speed along the path and direction of pollutants (Lin et al., 2014; Sait et al., 2013). The number of clusters was determined based on the total spatial variance (TSV) variation, which was calculated as the sum of the SPVAR values as the amount of TSV change increased or decreased (Song et al., 2017).
In this study, PSCF was analyzed to interpret backward trajectory and cluster analysis (Han et al., 2004; Choi et al., 2011). Decision tree learning, a machine learning method, is commonly used in data mining (Torgyn et al., 2017). The algorithms used for statistical analysis are CART, CHAID, MARS and CIT (Fatin et al., 2017).
In this study, CIT analysis was performed using the "party" package in the R program (Hothorn et al., 2017).
RESULTS AND DISCUSSION
Monitoring results
Especially in summer, the SO2 concentration in the industrial areas increased significantly, but that in the roadside areas remained constant. In this season, the concentrations of PM10 and SO2 increased the SO2 ratio during the day, as in the case of the residential areas. This means that the sources of SO2 affected the pollution levels in the residential and roadside areas during the day.
SO2 concentration in the residential and road areas tended to increase during the day, unlike PM10. High concentrations of PM10 in the road areas during spring were considered to be influenced by vehicle emissions on the road. The high level of PM10 at Yeocheon was influenced by the surrounding petrochemical complexes and the non-ferrous industrial complexes in the south, as in the case of spring.
As in the spring, PM10 in the roadside areas had a large impact on the surrounding roads. SO2 levels in roadside areas were influenced by industrial complexes located to the east and southeast. This means that the sources of pollution were located in the south, which was expected given the location of the industrial complexes of non-ferrous metals.
The results of the road areas were not very different from those obtained in the case of summer. The wind in the central region of Ulsan blew from the urban area towards the petrochemical and non-ferrous industrial complexes. As the surface wind during the day blew from the homes towards the industrial areas, the SO2 level in the residential and roadside areas was expected to be influenced by the industrial complexes.
This means that the influence of the point sources in the industrial complexes was strongest during April to August. This indicated that the level of PM10 in the road along the area was affected to a lesser extent than that in the residential and industrial areas (Figure 36 (c)). A few backward trajectories in the third group indicated the influence of PM10 from Shanghai (Figure 39 (c)).
The atmosphere in Ulsan was polluted by SO2 coming from a chimney in an industrial complex.
Local sources identification
- Determining representative sites to identify the sources
- Relationship between PM 10 and SO 2 hourly concentrationrs
- CBPF result of PM 10 and SO 2
- Surface wind field in Ulsan
- CALPUFF result of PM 10 and SO 2 emission
Long-range atmospheric transport of PM 10
- Backward trajectory analysis result
- Cluster analysis result
- Potential source contribution function result
Between March and , the number of return routes from the East Sea and the Pacific and from China increased. Then, most of the return trips after October again came from China, as in the case of January and February (Figure 38). The first group of backtracks was shown to originate from the northwest side of Ulsan.
A total of 496 backward trajectories were among the second largest number of trajectories in 2012, accounting for 34% (Figure 39 (f)). The backward trajectories went from eastern China, even Mongolia, towards North Korea and were influenced by the industrial and residential areas in Shenyang, but also by North Korea. Even the backward trajectories of this cluster were expected to be affected by the sandstorms originating from the Gobi Desert in Mongolia and by the biomass burning in North Korea (Figure 39 (a)).
The third cluster of receding trajectories originates from the west side of Ulsan, which was influenced by the industrial areas of Beijing and Tianjin in China, as well as the western part of South Korea, as the LRAT effect. They showed the impact of the Pacific Ocean and Shanghai, which is located in the southeast of China (Figure 39 (d)). The backward trajectories in this cluster indicated the influence of southern Japan, including the influence of Kitakyushu Industrial Area in the Kyushu region of southern Japan (Figure 39 (e)).
In autumn and winter, the return routes mostly originated from China, but were not potential sources of high PM10 concentrations (Figure 40 (c), (d)). Most of the potential sources in cluster 1 were in eastern China, including Shenyang, and some in Mongolia. The average concentration of return trajectories in cluster 3 was 57.08 μg/m3, which is the highest among all clusters.
This indicated that the highest concentration of PM10 occurred when the reverse trajectories originating from the Beijing and Tianjin Industrial Areas were in effect. The backtracks in the fourth group are mainly contaminated by the South Sea and Shanghai, which is in the southeastern part of China.
Importance factor of PM 10 and SO 2 concentration
CONCLUSIONS
저는 이 감사의 편지를 쓰고 있습니다. 5년 전 처음 EACL 연구실에 아르바이트로 입사한 이후로 많은 졸업생 형제자매들을 봐왔는데, 이런 감사 편지를 받게 될 날이 올 줄은 몰랐습니다. 지금까지 대학원 생활 동안 저에게 도움을 주신 모든 분들께 감사의 말씀을 전하고 싶습니다.
내가 하고 싶은 공부를 할 수 있도록 묵묵히 지켜봐 주시고, 어려울 때마다 가장 먼저 도와주신 부모님께 감사드립니다. 속상할 때도 있었지만 꿈을 위해 더 열심히 공부하고 열심히 살아서 행복한 삶을 살겠습니다. 그리고 석사학위를 성공적으로 마치시고, 앞으로의 연구생활에 영원한 멘토가 되어 주실 최성득 교수님께 감사의 말씀을 전하고 싶습니다.
앞으로도 좋은 강의와 포스팅으로 많은 가르침 부탁드리겠습니다. 그리고 저희 연구실의 첫 번째 의사이신 혜옥 선생님, 제가 처음 연구실에 왔을 때부터 잘 보살펴주시고 연구실에 잘 적응할 수 있도록 도와주셔서 감사하다는 말씀 전하고 싶습니다. 그리고 연구실 생활 중 가장 오랫동안 뵙던 민규형 선생님께도 진심으로 감사드립니다.
그리고 동생이 졸업할 때까지 더욱 감사한 마음을 전하겠습니다. 저희 대기팀을 이끌어주시는 성준형 님에게도 진심으로 감사하고 빚을 지게 될 것입니다. 환경분석센터가 있어서 EACL에 가입할 수 있었어요.
마지막으로 가장 소중한 가족들에게 다시 한 번 감사 인사를 전하고 싶습니다. 열심히 공부해서 멋진 아들, 멋지고 자랑스러운 사람이 되어 행복하게 살겠습니다.
