Analysis of the National Air Pollutant Emissions Inventory (CAPSS 2017) Data and Assessment of Emissions based on Air Quality Modeling in the Republic of Korea
Copyright © 2021 by Asian Association for Atmospheric Environment
This is an open-access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/), which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
Abstract
According to the 2017 National Air Pollutant Emissions Inventory (NEI), air pollutant emissions in the Republic of Korea comprised 817,420 metric tons (hereafter tons) of CO, 1,189,800 tons of NOx, 315,530 tons of SOx, 592,582 tons of TSP, 218,476 tons of PM10, 91,731 tons of PM2.5, 15,555 tons of black carbon (BC), 1,047,585 tons of VOCs, and 308,298 tons of NH3. Emissions of the 13 first-level emission source categories, which constitute the NEI, were estimated and, based on their characteristics, the emission source categories were grouped into five sectors (energy, industry, road, non-road, and everyday activities and others). In addition, the contributions of primary PM2.5 and its four precursors (NOx, SOx, VOCs, and NH3) to the 2017 NEI were assessed in this study. The emission contributions of NOx to the NEI were 36.5% for the road sector, which was the highest of those of all the air pollutants for this sector; NOx emissions for this sector were 4.2% lower than those in the previous year. The emission contributions of SOx and PM2.5 to the NEI were higher than those of the other air pollutants for the industry sector; SOx and PM2.5 emissions for this sector decreased by 9.8% and 19.7%, respectively, compared with those in the previous year. The emission contributions of VOCs and NH3 to the NEI were 65.3% and 83.9% for the everyday activities and others sector, respectively, higher than those of the other air pollutants for this sector; VOCs and NH3 emissions for this sector increased by 0.8% and 2.9%, respectively, compared with those in the previous year. A three-dimensional (3D) chemical transport modeling system was used to validate the emission estimates. These data suggest that simulated SOx emissions were overestimated in areas with dense large-scale industrial complexes, such as Jeollanam-do, Gyeongsangbuk-do, and Ulsan, and that simulated NOx emissions were overestimated in Seoul, Incheon, and Jeollanam-do.
Keywords:
NEI, CAPSS, CMAQ, NEAS, PM2.51. INTRODUCTION
In September 2017, the Government of the Republic of Korea announced the Comprehensive Plan on Fine Dust1 Management, aiming to reduce particulate matter (PM) emissions by 30% by 2022 in response to increasing public demand. Moreover, additional efforts were made to reduce PM emissions by announcing three more plans, namely, the Strengthened Measures to Manage Fine Dust at Normal and Emergency Levels in November 2018, designed to preemptively supplement the previous plan; the Comprehensive Plan on Fine Dust Management in November 2019; and the Special Measures to Respond to High Levels of Fine Dust, which is also called the Seasonal Fine Dust Management System.
Even though the national annual mean concentrations of PM2.5 in 2019 were 3 μg/m3 lower than those in 2015 (2015: 26 μg/m3 → 2019: 23 μg/m3), only 1.9% of the regions in Korea met the national standards for PM2.5 (NIER, 2020). In this regard, basic research is increasingly needed to implement effective plans that help reduce air pollutant emissions (Hwang, 2018; Koo et al., 2018). Primarily, more research is needed to clarify the relationship between PM concentrations and PM precursors. To this end, it is necessary to obtain accurate and reliable emission data and comprehensively analyze current emission trends (Choi et al., 2020).
In Korea, the National Air Emission Inventory and Research Center (NAIR) assesses and declares annual air pollutant emission data. Although efforts have been made to explore trends in, and thereby mitigate, air pollutant emissions, it is difficult to estimate emissions that fluctuate due to various factors, including socioeconomic factors, and to reflect all the variations in emissions. As only qualitative assessment of emission sources and pollutants was conducted when evaluating emission factors and activity data using the data attribute rating system (DARS), the accuracy of the emission data generated in these assessments has often been questioned (Kim and Jang, 2014). Normally, it takes 2 to 3 years to review and improve input data, including raw data (Zhang et al., 2009). For this reason, a bottom-up approach that uses observed data is continuously studied and currently used to rapidly review and improve the reliability of emission data generated (Wang et al., 2012; Martin et al., 2003). In recent research, observed concentrations of air pollutants (i.e. satellite-derived and ground-level observed concentrations) and concentrations simulated using 3D chemical transport models have been used together to evaluate the reliability of emission data (Bae et al., 2020a, b; Kim et al., 2020; Liu et al., 2017; Fioletov et al., 2011).
In line with two previous studies analyzing the 2015 National Air Pollutant Emissions Inventory (NEI) data (Yeo et al., 2019) and 2016 NEI data (Choi et al., 2020), this study analyzed the 2017 NEI data to determine the contributions of various emission source categories to the 2017 NEI and the major factors responsible for changes in emissions from 2016. We here present the current state of emissions. Moreover, we used the 2017 NEI data to simulate the concentrations of the pollutants using a 3D chemical transport model and compared the simulated concentrations with ground-level observed concentrations. Thus, emission uncertainties by region and pollutant were indirectly reviewed.
2. METHODS OF ESTIMATING NATIONAL AIR POLLUTANT EMISSIONS
Emissions of nine air pollutants by emission source category were estimated by using emission data from a telemonitoring system (hereafter TMS)2 operated in places of business or by applying emission factors to activity data. An emission source classification was established based on the CORINAIR classification of the European Environment Agency (EEA) and was adapted in consideration of domestic realities, including industrial structures. Emission sources were grouped into 13 first-level categories, which were further divided into 56 second-level categories, and further classified into 240 third-level categories.
Emission factors are presented as emissions per unit of activity data. Approximately 30,000 emission factors are currently applied to NEI data. Emission factors developed by domestic research institutes such as the National Institute of Environmental Research (NIER) were primarily used for estimating emissions, but for most source categories, emission factors developed by the U.S. Environmental Protection Agency (US EPA) and EEA were adopted, except in the case of a few emission source categories including vehicles, construction machinery, and combustion facilities (NAIR, 2020a).
Activity data were determined using 300 sets of statistical data obtained from approximately 150 institutes related to energy, industry, transport, meteorology, etc. In general, there are three ways to validate these data: comparing the totals of raw data with registered data from the database to detect errors that may have occurred during the data registration process; studying the previous year’s data and analyzing changes; and comparing those data with other similar data (NIER, 2019).
Emissions from power generation and large-scale businesses in the energy and industry sectors were estimated based on activity data (e.g. the amount of fuel·raw material consumed, produced, and incinerated), which were collected using a TMS and the stack emission management system (SEMS). SEMS is a computing system operated by NAIR, and under the law, places of business are legally required to maintain records of air emission control, facility operations, and fuel consumption, among other things.
For the road sector (comprising the passenger cars and freight cars categories), vehicle kilometers traveled (hereafter VKT) data, which reflect the domestic traffic situation, were estimated based on the number of vehicle registrations, daily average VKT, and observed traffic volume, and were used as activity data. To estimate emissions, NIER applied emission factors developed based on tests, including real driving emission-light-duty vehicles (RDE-LDV) tests, to activity data.
For the non-road sector (comprising the railways, ships, aircraft, agricultural machinery, and construction machinery categories), activity data were used to estimate emissions; and the categories related to transport modes were as follows: railways - fuel consumption by section; ships - oil supplies and arrival and departure information obtained from a port management information system (Port-MIS); aircraft - takeoff and landing information by airport; agricultural machinery - the number of registered machinery and working hours by agricultural machinery; and construction machinery - working hours by construction machinery.
For the everyday activities and others sector (comprising the energy transport and storage, fugitive dust, and biomass burning categories), related national statistical data were used as major activity data to estimate emissions. The data used were as follows: energy transport and storage - gasoline production by oil refineries; solvent use - paint production; fugitive dust - VKT and building construction area; and biomass burning - crop cultivation area.
3. 2017 EMISSION ESTIMATES
3. 1 Air Pollutant Emissions
In the 2017 NEI, the nationwide emissions of air pollutants comprised 817,420 tons of CO; 1,189,800 tons of NOx, 315,530 tons of SOx; 592,582 tons of TSP; 218,476 tons of PM10; 91,731 tons of PM2.5; 15,555 tons of BC; 1,047,585 tons of VOCs; and 308,298 tons of NH3, as shown in Table 1 (NAIR, 2020b).
The emission contributions of different emission source categories by pollutant were as follows: road transport (29.0%), biomass burning (27.5%), and non-road transport (21.6%) for CO; road transport (36.5%), non-road transport (26.0%), and manufacturing industry (14.3%) for NOx; industrial process (33.8%), energy production (24.6%), and manufacturing industry (22.9%) for SOx; fugitive dust (71.3%) and manufacturing industry (16.2%) for TSP; fugitive dust (50.3%) and manufacturing industry (25.6%) for PM10; manufacturing industry (31.1%), fugitive dust (19.3%), and non-road transport (16.4%) for PM2.5; non-road transport (43.4%) and road transport (34.2%) for BC; solvent use (53.8%) and manufacturing industry (18.0%) for VOCs; and agriculture (79.3%) and industrial process (13.9%) for NH3 (Fig. 1).
3. 2 Analysis on Changes in Emissions Compared with the Previous Year
This study analyzed the contributions of primary PM2.5 and four of its precursors (NOx, SOx, VOCs, and NH3) to the 2017 NEI based on an emission source classification, which is presented in Table 2. The table shows that thirteen first-level categories of emission sources were grouped into five sectors (energy, industry, road, non-road, everyday activities and others). In addition, the major causes that contribute to changes in emissions from 2016 to 2017 were analyzed and described. Further details on emissions of air pollutants by emission source category can be found in Appendix 1.
The contributions of the energy sector to the 2017 NEI, by pollutant, were as follows: NOx (8.9%), SOx (20.7%), PM2.5 (3.4%), VOCs (0.7%), and NH3 (0.4%). More specifically, NOx, SOx, PM2.5, VOCs, and NH3 emissions decreased by 23.3% (32,099 tons), 17.2% (13,513 tons), 2.9% (93 tons), 2.9% (221 tons), and 14.7% (204 tons), respectively, compared with the previous year (Fig. 2).
The contributions of the public power generation category to the emissions from the energy sector, by pollutant, were as follows: NOx (73.2%), SOx (90.2%), PM2.5 (78.8%), VOCs (57.8%), and NH3 (37.3%). Specifically, NOx, SOx, PM2.5, VOCs, and NH3 emissions from public power generation decreased by 29.6% (32,426 tons), 17.6% (12,597 tons), 4.7% (123 tons), 10.5% (506 tons), and 37.8% (268 tons), respectively, compared with the previous year. Although there were increases in coal consumption (e.g bituminous coal) compared with the previous year, emissions of those pollutants decreased because consumption of oil (e.g. Bunker C oil) and LNG decreased by 66.9% (1.93 million kL) and 20.9% (1.676 billion m3), respectively, and environmental facilities operated by places of business were upgraded. NOx, PM2.5, VOCs, and NH3 emissions from private power generation increased by 0.2% (53 tons), 1.7% (9 tons), 6.3% (143 tons), and 5.8% (30 tons), respectively, compared with the previous year while SOx emissions decreased by 11.3% (662 tons).
Emissions from the industry sector were estimated by totaling the emissions from the manufacturing industry, industrial processes, waste, and oil refinery categories. Overall, the contributions of the industry sector to the 2017 NEI, by pollutant, were as follows: NOx (20.6%), SOx (61.3%), PM2.5 (37.0%), VOCs (23.9%), and NH3 (14.2%). NOx, SOx, and PM2.5 emissions decreased by 3.0% (7,585 tons), 9.8% (20,921 tons), and 19.7% (8,305 tons), respectively, compared with the previous year, while VOCs and NH3 emissions increased by 0.6% (1,468 tons) and 1.1% (479 tons), respectively (Fig. 3).
The contributions of the manufacturing industry category to the emissions from the industry sector, by pollutant, were as follows: NOx (69.3%), SOx (37.4%), PM2.5 (84.0%), VOCs (1.3%), and NH3 (1.6%). NOx, SOx, PM2.5, and VOCs emissions decreased by 3.2% (5,541 tons), 16.5% (14,266 tons), 22.5% (8,284 tons), and 4.3% (143 tons), respectively, compared with the previous year, while NH3 emissions increased by 2.4% (16 tons). Overall emissions decreased mainly because coal consumption (e.g. bituminous coal) decreased by 12.0% (1.707 million tons) compared with the previous year.
The contributions of the industrial processes category to the emissions from the industry sector, by pollutant, were as follows: NOx (21.9%), SOx (55.2%), PM2.5 (15.3%), VOCs (75.3%), and NH3 (98.0%). NOx, SOx, and PM2.5 emissions decreased by 4.1% (2,314 tons), 5.3% (6,004 tons), and 0.1% (6 tons), respectively, compared with the previous year, while VOCs and NH3 emissions increased by 1.2% (2,220 tons) and 1.1% (488 tons), respectively. The overall emissions in this sector declined mainly because emissions from some large oil refineries and steel producers, which are measured by a TMS, decreased despite a 1.9% increase (2.946 million kL) in crude oil consumption by the petroleum industry and a 0.1% increase (86,000 tons) in the output of sintered products in the iron and steel industry.
The contributions of the waste category to the emissions from the industry sector, by pollutant, were as follows: NOx (5.3%), SOx (1.1%), PM2.5 (0.7%), VOCs (23.3%), and NH3 (0.1%). NOx, SOx, PM2.5, and VOCs emissions decreased by 4.2% (575 tons), 1.9% (41 tons), 7.2% (18 tons), and 1.0% (583 tons), respectively, compared with the previous year, and there was a 0.3% increase in NH3 emissions (0.1 tons). The overall decline in the waste category was because the amounts of incinerated municipal solid waste and industrial waste slightly decreased, by 5.5% (296,000 tons) and 0.3% (20,000 tons), respectively, compared with the previous year.
The contributions of the road sector to the 2017 NEI, by pollutant, were as follows: NOx (36.5%), SOx (0.1%), PM2.5 (9.5%), VOCs (4.4%), and NH3 (1.4%). NOx, PM2.5, VOCs, and NH3 emissions decreased by 4.2% (18,956 tons), 10.6% (1,033 tons), 3.5% (1,641 tons), and 12.5% (634 tons), respectively, compared with the previous year, while there was a 19.8% increase in SOx emissions (46 tons). The contributions of the freight cars and RVs categories to the emissions from the road sector were larger than those of the other types of vehicles (Fig. 4).
The overall decline in emissions, despite an increase in the number of car registrations and VKT, was mainly due to the replacement of old cars that produce relatively large amounts of air pollutants with new cars, as shown in Table 3.
The contributions of the non-road sector to the 2017 NEI, by pollutant, were as follows: NOx (26.0%), SOx (11.3%), PM2.5 (16.4%), VOCs (5.7%), and NH3 (0.04%). NOx and SOx emissions decreased by 0.2% (677 tons) and 13.8% (5,732 tons), respectively, compared with the previous year, whereas PM2.5, VOCs, and NH3 emissions increased by 4.5% (648 tons), 45.6% (18,592 tons), and 2.4% (3 tons), respectively (Fig. 5).
The contributions of the ship category to the emissions from the non-road sector, by pollutant, were as follows: NOx (52.5%), SOx (96.9%), PM2.5 (51.5%), VOCs (69.1%), and NH3 (12.3%). NOx, PM2.5, VOCs, and NH3 emissions increased by 0.4% (688 tons), 10.5% (735 tons), 85.1% (18,879 tons), and 2.2% (0.3 tons), respectively, compared with the previous year, while there was a 14.4% decline in SOx emissions (5,819 tons). Since the methodology of obtaining activity data was changed (i.e. the agency in charge of obtaining data on leisure boats was changed), the number of ship registrations increased, which, in turn, contributed to increases in emissions from the non-road sector, while the decrease in SOx emissions resulted from an 11.4% decrease (41,882 kL) from the previous year in the use of oil on which cargo ships operate (e.g. Bunker C oil) (Table 4).
The contributions of the construction machinery category to the emissions from the non-road sector, by pollutant, were as follows: NOx (36.9%), SOx (0.2%), PM2.5 (37.3%), VOCs (24.7%), and NH3 (34.3%). NOx, PM2.5, and VOCs emissions decreased by 2.5% (2,881 tons), 1.4% (80 tons), and 2.3% (339 tons), respectively, compared with the previous year, while SOx and NH3 emissions increased by 28.2% (16 tons) and 5.3% (2.1 tons), respectively. Although the number of registered construction machinery increased by 4.9% (18,000 units), the retrofitting of old construction machinery mainly contributed to these decreases in emissions.
The everyday activities and others sector consisted of the non-industry, energy transport and storage, solvent use, agriculture, other (area sources), fugitive dust, and biomass burning categories. The contributions of this sector to the 2017 NEI, by pollutant, were as follows: NOx (8.1%), SOx (6.6%), PM2.5 (33.7%), VOCs (65.3%), and NH3 (83.9%). NOx, PM2.5, VOCs, and NH3 emissions increased by 0.9% (808 tons), 0.9% (268 tons), 0.8% (5,358 tons), and 2.9% (7,353 tons), respectively, compared with the previous year, while SOx decreased by 13.7% (3,301 tons), as shown in Fig. 6.
The non-industry category included fuel combustion for heating and other purposes in residential, commercial, institutional, agricultural, and livestock facilities. The contributions of the non-industry category to the emissions from the everyday activities and others sector, by pollutant, were as follows: NOx (90.6%), SOx (99.6%), PM2.5 (3.0%), VOCs (0.4%), and NH3 (0.6%). NOx, VOCs, and NH3 emissions increased by 1.1% (979 tons), 3.3% (90 tons), and 1.0% (14 tons), respectively, compared with the previous year, while SOx and PM2.5 emissions decreased by 13.7% (3,300 tons) and 4.4% (43 tons), respectively. Increased NOx emissions were caused by a 5.2% increase (672 million m3) in LNG consumption by residential, commercial, and institutional facilities, while SOx emissions decreased because the consumption of oil (e.g. Bunker C oil) and anthracite coal decreased by 2.2% (236,000 kL) and 14.0% (176,000 tons), respectively, compared with the previous year.
The solvent use category (e.g. painting facilities and other solvent use) was responsible for 82.3% of the VOCs emissions from the everyday activities and others sector. Emissions from this category increased by 1.0% (5,644 tons) compared with the previous year, mainly due to an increase in the paint thinner consumption by coating facilities.
The agriculture category (e.g. fertilizer use and livestock manure management) accounted for 94.4% of NH3 emissions in the everyday activities and others sector. There was a 3.1% increase in NH3 emissions (7,318 tons) from the previous year due to a 0.4% increase (790,000 animals) in the number of livestock (e.g. cattle and pigs) from the previous year.
The fugitive dust category included road dust from vehicles running on the road and dust from construction sites and vacant lands without dust outlets. Fugitive dust accounted for 57.2% of the PM2.5 emissions from the everyday activities and others sector, which was an increase of 2.3% (403 tons) from the previous year. The paved roads category (one of its second-level categories), which accounted for 42.5% of fugitive dust emissions, showed a 6.0% increase in PM2.5 emissions (428 tons) from the previous year. Although there were 12 fewer days with at least 0.254 mm of precipitation, one of the major factors that contribute to changes in the fugitive dust category (US EPA, 2011), than in the previous year, the overall fugitive dust emissions increased because total VKT in Korea increased by 1.4% (4.294 billion km) compared with the previous year.
The biomass burning category included burning in everyday life, such as agricultural residue incineration, and the contributions of this category to emissions from the everyday activities and others sector, by pollutant, were as follows: NOx (9.2%), PM2.5 (38.5%), and VOCs (12.6%). NOx, PM2.5, and VOCs emissions decreased by 2.4% (217 tons), 1.7% (205 tons), and 1.4% (1,188 tons), respectively, compared with the previous year. This was because the cultivation area for fruits (e.g. pears and apples) and specialty crops (e.g. sesame and peanuts) declined by 0.5% (742 ha), compared with the previous year, thereby decreasing the amount of incineration.
3. 3 Air Pollutant Emissions by Region
Data on the 2017 NEI from 17 regions (first-tier administrative divisions3) are shown in Table 5 and Fig. 7. Gyeonggi-do emitted the largest proportions of CO, NOx , and VOCs emissions in Korea, at 15.8% (128,925 tons), 15.5% (184,239 tons), and 18.3% (191,840 tons), respectively. NOx emissions in Gyeonggi-do decreased by 19,416 tons compared with the previous year, mainly due to decreased emissions by the construction machinery category, which, in turn, was attributed to reductions in the VKT of freight cars and in building construction areas. Chungcheongnam-do emitted the largest proportions of SOx and NH3, at 22.2% (69,905 tons) and 17.1% (52,578 tons), respectively. SOx emissions in Chungcheongnam-do declined by 14.6% (11,938 tons), mainly because of a reduction in emissions measured using a TMS as a result of strengthened regulations on air pollution for power generation facilities. Gyeongsangbuk-do generated the largest proportions of TSP, PM10, and PM2.5, at 16.5% (97,910 tons), 18.6% (40,586 tons), and 21.5% (19,738 tons), respectively. PM2.5 emissions in Gyeongsangbuk-do decreased by 12.9% (2,932 tons) compared with the previous year, due to a reduction in the consumption of anthracite coal used for steel production.
3. 4 Air Pollutant Emissions by Country
To quantitatively analyze air pollutant emissions in Korea, a comparison of emissions in Korea and those in 28 European countries and the United States was conducted. The ratios of emissions in South Korea to those of the 29 other countries, by pollutant, ranged between 1.2% and 12.4%. Moreover, emissions per 1 million people (Emission/Population [E/P]) in Korea and the 29 other countries were compared, and the ratio of each pollutant was found to be between 7.7% and 78.3%. It should be noted that simple comparison is not appropriate here because emissions data are partly determined by the economic size and population of a country. Nevertheless, since the comparison revealed that the emissions may have been under- or overestimated, it is necessary to continuously review and improve the accuracy of emission estimates by utilizing various methods, including measurement and observation (i.e. measured or observed data using environmental satellites) (Table 6).
4. ASSESSMENT OF UNCERTAINTY IN EMISSIONS USING AIR QUALITY MODELING
4. 1 Methodology
It is difficult to directly assess the uncertainty in emissions of air pollutants since pollutants are emitted in various forms from different sources. In this regard, a 3D chemical transport model is useful for determining the relationship between emissions and concentrations because the meteorology, emissions, and chemical reactions in the atmosphere can be treated as variables in the model (Ju et al., 2018; Kim et al., 2018; Bae et al., 2017; Kim et al., 2017a; Kim et al., 2017b). Thus, studies are currently being conducted to indirectly review emission data via air quality modeling, which assists in converting emissions data into concentrations and comparing them with different kinds of observed data (Bae et al., 2020a, b; Kim et al., 2020).
In this study, air quality modeling was conducted using the National Emission and Air quality assessment System (NEAS) consisting of the Weather Research and Forecasting (WRF) model, the Sparse Matrix Operator Kernel Emissions (SMOKE) modeling system, and the Community Multiscale Air Quality (CMAQ) model. The model simulation was performed using the Clean Air Policy Support System (CAPSS) 2017, and emission data were reviewed by comparing the air pollutant concentrations obtained during the simulation process with those obtained during the observation of urban air quality monitoring stations. To understand the effect of changes in emissions on air pollutant concentrations in the atmosphere, an air quality simulation was additionally performed using CAPSS 2016. The domains and horizontal resolutions used for the simulation were as follows: Northeast Asia (27 km), the Korean Peninsula (9 km), and South Korea (3 km). The simulated year was 2017, and the simulation configurations are presented in the Supplementary Materials.
4. 2 Assessment of Uncertainty in Criteria Pollutant Emissions
Simulated concentrations based on CAPSS 2017 were compared with observed concentrations from urban air quality monitoring stations (Fig. 8). The target pollutants for the comparison were SO2 and NOx, primary pollutants whose concentrations are directly affected by emission changes. CO, one of primary pollutants, was excluded from this study because its high long-range transport potential makes it unsuitable for assessing local emissions. Considering the spatiotemporal resolution of CAPSS data (by si (city), gun (county), and gu (district) and by month), a comparison was performed using the monthly mean values of each region. The national mean observed SO2 concentrations were 4.4 ppb and the simulated concentrations were 4.3 ppb, showing a 0.1 ppb (2%) error between the observation and simulation data. When classified by region and month, the concentration data of most regions showed an error of ≤2 ppb. As their overall biases were negative, it can be seen that the simulated concentrations were underestimated. In contrast, Jeollanam-do, Geongsangbuk-do, and Ulsan showed the opposite results, with notable errors. In particular, Ulsan had a monthly maximum error of >20 ppb. Since biases and errors of those three regions were positive in a similar range, their simulated concentrations were seen to be consistently overestimated. For simulated SO2 emissions, there was a stark contrast between regions. Since large-scale industrial complexes were responsible for most of the SO2 emissions in the three regions, the reliability of SOx emissions from the industrial sector should be primarily reviewed.
The national mean observed concentrations of NOx were 34.3 ppb, and the simulated concentrations were 28.6 ppb, presenting an underestimation of 5.7 ppb (17%). Errors in yearly NOx concentration data in the different regions ranged from 26% to 55%. Sejong recorded the highest error of 17.1 ppb (55%), followed by Gyeongsangnam-do (16.8 ppb; 50%) and Incheon (15.5 ppb; 40%). The simulated concentrations of NOx in Seoul, Incheon, and Jeollanam-do were deemed to be overestimated, while those in the other regions were deemed to be underestimated. Over- or underestimations of the emissions were found even within the same regions depending on the month, with overestimations generally occurring in the summer months and underestimations occurring in the winter months. These data may have been affected by the uncertainty of the temporal profile of NOx, in addition to the uncertainty in its annual total emissions.
The emissions and concentrations of primary pollutants generally have a linear relationship; thus, the overestimation/underestimation ratios of their emissions were indirectly estimated using the relative ratios of their simulated and observed concentrations (Bae et al., 2020a; Kim et al., 2020). SO2 emissions in Ulsan that showed notable overestimations were overestimated by 60%, while those in Gyeongsangbuk-do and Jeollanam-do were overestimated by 30% each. NOx emissions in Seoul, Incheon, and Jeollanam-do were overestimated by 13%, 23%, and 28%, respectively. However, since the difference between the observed and simulated concentrations may have arisen due to the uncertainty of the air quality models and the meteorological models, the emission estimates presented in this study should be interpreted keeping in mind the possibility of the current emissions being over- or underestimated.
4. 3 Changes in PM2.5 Concentrations Affected by Changes in Emissions
To understand the effect of annual changes in emissions on air pollutant concentrations, an additional experiment using air quality modeling was conducted. The experiment compared simulated concentrations of air pollutants based on CAPSS 2016 and CAPSS 2017, assuming the meteorological conditions to be equal. Changes in concentrations described below mean changes in simulated concentrations for each pollutant caused by changes in emissions. The national mean SO2 concentrations in the CAPSS 2017 simulation showed a 1.0 ppb (19%) decrease, compared with those in the CAPSS 2016 simulation. The emission changes in individual regions also varied, ranging from -11.1 ppb (-51%) to 0.7 ppb (17%). The Seoul Metropolitan Areas (SMA; Seoul, Incheon, and Gyeonggi-do), Chungcheong-do, and the Southeastern areas showed a decline in the simulated SO2 concentrations; as presented in Section 3. 2, SOx emissions in the CAPSS 2017 simulation decreased by approximately 12% across the country, compared with those in the previous year, mainly because both the emissions from public power generation facilities, measured by a TMS, and the use of anthracite coal for steelmaking in the manufacturing industry declined.
On the other hand, some sis (cities) such as Incheon, Pohang, Busan, and Yeosu saw an increase of 1-2 ppb in simulated SO2 concentrations. This may have been because SOx emissions in Incheon and Yeosu from the ship category in the non-road transport sector increased by 48.7% (749 tons) and 25.4% (713 tons), respectively, compared with those in the previous year, and SOx emissions in Busan and Pohang from the iron and steel industry category in the industrial process sector increased by 24.1% (161 tons) and 11.3% (919 tons), respectively.
The national mean NOx concentrations in the CAPSS 2017 simulation decreased by 0.1 ppb (<1%) annually compared with those in the CAPSS 2016 simulation. Further, noticeable changes occurred in the emissions from individual regions, ranging from -3.5 ppb (-11%) to 8.2 ppb (20%). NOx concentrations declined in most of the southern areas of Gyeonggi-do, Chungchung-do, and near highways. That decline occurred mainly due to a decline in the emissions from the road transport sector, which, in turn, happened as a result of effective emission reduction policies, such as the replacement of old vehicles with newer ones being instituted. In contrast, NOx concentrations in the other regions, such as Incheon and Gangwon-do, increased by more than 4 ppb per year. This increase was estimated to be distributed as follows: in Incheon, NOx emissions from the ship and construction machinery categories in the non-road transport sector increased by 74.9% (2,963 tons) and 156.6% (6,281 tons), respectively, from the last year; and in Gangwondo, NOx emissions increased by 114% (709 tons) from the last year as LNG power generation facilities were newly built in this region.
Concentrations of PM2.5, including primary emission and secondary formation, declined by 0.5 μg/m3 (3%) per year, on average. PM2.5 concentrations tended to decrease nationwide; however, these changes varied, ranging from -2.4 μg/m3 (-8%) to 0.7 μg/m3 (4%). Most notably, the PM2.5 concentrations in the CAPSS 2017 simulation in Seoul and Incheon were higher than those in the CAPSS 2016 simulation. This increase was attributed to PM2.5 emissions from the construction machinery category in the non-road transport sector in Seoul and Incheon increasing by 40.6% (241 tons) and 159.4% (310 tons), respectively, compared with the previous year. The decline in simulated PM2.5 concentrations that occurred across the country was brought about by two major factors: 1) reductions in SO2 and NOx emissions which are precursors to PM2.5; 2) reductions in primary PM2.5 emissions. It should be noted that further studies on this need to be conducted since ambient PM2.5 concentrations are affected by various chemical reactions.
5. CONCLUSION
According to the 2017 NEI, air pollutant emissions in the Republic of Korea, estimated using CAPSS, comprised 817,420 tons of CO; 1,189,800 tons of NOx; 315,530 tons of SOx; 592,582 tons of TSP; 218,476 tons of PM10; 91,731 tons of PM2.5; 15,555 tons of BC; 1,047,585 tons of VOCs; and 308,298 tons of NH3. NOx, SOx, TSP, PM10, and BC emissions decreased by 4.7%, 12.1%, 3.1%, 6.3%, 8.5%, and 5.2%, respectively, compared with those in the previous year, while CO, VOCs, and NH3 emissions increased by 2.8%, 2.3%, and 2.3%, respectively.
Emissions of NOx, SOx, VOCs, and NH3, which contribute to the formation of primary PM2.5, by emission source category and their contributions to the NEI were also assessed in this study. Emissions from the energy sector comprised 105,646 tons of NOx, 65,266 tons of SOx, 3,137 tons of PM2.5, 7,484 tons of VOCs, and 1,179 tons of NH3, with their contributions to the 2017 NEI being 8.9%, 20.7%, 3.4%, 0.7%, and 0.4%, respectively. Emissions from the industry sector comprised 244,949 tons of NOx, 193,485 tons of SOx, 33,945 tons of PM2.5, 250,198 tons of VOCs, 43,838 tons of NH3, with their contributions to the 2017 NEI being 20.6%, 61.3%, 37.0%, 23.9%, and 14.2%, respectively. Emissions from the road transport sector comprised 434,038 tons of NOx, 277 tons of SOx, 8,715 tons of PM2.5, 45,920 tons of VOCs, and 4,437 tons of NH3, with their contributions to the 2017 NEI being 36.5%, 0.1%, 9.5%, 4.4%, and 1.4%, respectively. Emissions from the non-road transport sector comprised 309,309 tons of NOx, 35,710 tons of SOx, 15,002 tons of PM2.5, 59,407 tons of VOCs, and 120 tons of NH3, with their contributions to the 2017 NEI being 26.0%, 11.3%, 16.4%, 5.7%, and 0.04%, respectively. Emissions from the everyday activities and others sector comprised 95,858 tons of NOx, 20,791 tons of SOx, 30,932 tons of PM2.5, 684,575 tons of VOCs, and 258,724 tons of NH3, with their contributions to the 2017 NEI being 8.1%, 6.6%, 33.7%, 65.3%, and 83.9%, respectively.
It should be noted that SO2 emissions in the vicinity of industrial complexes and NOx emissions in Seoul, Incheon, and Jeollanam-do may have been overestimated; thus, it is necessary to secure reliable data on monthly activity data. Since the method of estimating air pollutant emissions assumes the errors in simulation to be the uncertainty of input data for emission estimation in each region, it did not consider the transport of ambient air pollutants and was difficult to figure out uncertainty in emissions. In addition, this method may produce different outcomes depending on the model configurations and can be affected by input data that are used for modeling emissions. Going forward, a new methodology that can tackle these limitations needs to be developed to assess uncertainty in a more sophisticated manner. The result produced in simulation included different values of uncertainty, but it was presented in this study because it was seen as reference data that allowed us to promptly assess uncertainty to provide information on regions whose uncertainty is likely to be high. Thus, the result of emission estimation is more useful for understanding characteristics of input data than as quantitative values.
The NEI is being used for establishing, implementing, and assessing national atmospheric environmental policies, as well as for conducting studies designed to understand and interpret atmospheric conditions. To improve the accuracy and reliability of the NEI, it is crucial to continuously conduct various studies on emission estimation, such as identifying missing emission sources, including the defense sector, upgrading and developing antiquated emission factors, and advancing methodologies for estimating emissions.
References
- Bae, C., Yoo, C., Kim, B.U., Kim, H.C., Kim, S. (2017) PM2.5 Simulations for the Seoul Metropolitan Area: (III) Application of the Modeled and Observed PM2.5 Ratio on the Contribution Estimation. Journal of Korean Society for Atmospheric Environment, 33(5), 445-457, (in Korean with English abstract). [https://doi.org/10.5572/KOSAE.2017.33.5.445]
- Bae, C., Kim, H.C., Kim, B.U., Kim, Y., Woo, J.H., Kim, S. (2020a) Updating Chinese SO2 emissions with surface observations for regional air-quality modeling over East Asia. Atmospheric Environment, 228, 117416. [https://doi.org/10.1016/j.atmosenv.2020.117416]
- Bae, C., Kim, H.C., Kim, B.U., Kim, S. (2020b) Surface ozone response to satellite-constrained NOx emission adjustments and its implications. Environmental Pollution, 258, 113469. [https://doi.org/10.1016/j.envpol.2019.113469]
- Choi, S.-W., Kim, T., Lee, H.-K., Kim, H.-C., Han, J., Lee, K.-B., Lim, E.-H., Shin, S.-H., Jin, H.-A., Choi, E., Kim, Y.-M., Yoo, C. (2020) Analysis of the National Air Pollutant Emission Inventory (CAPSS 2016) and the Major Cause of Change in Republic of Korea. Asian Journal of Atmospheric Environment, 14(4), 422-445. [https://doi.org/10.5572/ajae.2020.14.4.422]
- Fioletov, V.E., McLinden, C.A., Krotkov, N., Moran, M.D., Yang, K. (2011) Estimation of SO2 emissions using OMI retrievals. Geophysical Research Letters, 38(21). [https://doi.org/10.1029/2011GL049402]
- Hwang, I.C. (2018) Particulate Matter Management Policy of Seoul: Achievements and Limitations. The Korean Association for Policy Studies, 27(2), 27-50, (in Korean with English abstract). https://www.kci.go.kr/kciportal/ci/sereArticleSearch/ciSereArtiView.kci?sereArticleSearchBean.artiId=ART002365074
- Ju, H., Bae, C., Kim, B.U., Kim, H.C., Yoo, C., Kim, S. (2018) PM2.5 Source Apportionment Analysis to Investigate Contributions of the Major Source Areas in the Southeastern Region of South Korea. Journal of Korean Society for Atmospheric Environment, 34(4), 517-533, (in Korean with English abstract). [https://doi.org/10.5572/KOSAE.2018.34.4.517]
- Kim, B.-U., Bae, C., Kim, H.C., Kim, E., Kim, S. (2017a) Spatially and chemically resolved source apportionment analysis: Case study of high particulate matter event. Atmospheric Environment, 162, 55-70. [https://doi.org/10.1016/j.atmosenv.2017.05.006]
- Kim, E., Bae, C., Yoo, C., Kim, B.U., Kim, H.C., Kim, S. (2018) Evaluation of the Effectiveness of Emission Control Measures to Improve PM2.5 Concentration in South Korea. Journal of Korean Society for Atmospheric Environment, 34(3), 469-485. [https://doi.org/10.5572/KOSAE.2018.34.3.469]
- Kim, J., Jang, Y.-K. (2014) Uncertainty Assessment for CAPSS Emission Inventory by DARS. Journal of Korean Society for Atmospheric Environment, 30(1), 26-36. [https://doi.org/10.5572/KOSAE.2014.30.1.026]
- Kim, O., Bae, M., Kim. S. (2020) Evaluation on Provincial NOx and SO2 Emissions in CAPSS 2016 Based on Photochemical Model Simulation. Journal of Korean Society for Atmospheric Environment, 36(1), 64-83. [https://doi.org/10.5572/KOSAE.2020.36.1.064]
- Kim, S., Bae, C., Kim, B.U., Kim, H.C. (2017b) PM2.5 simulations for the Seoul Metropolitan Area: (I) contributions of precursor emissions in the 2013 CAPSS emissions inventory. Journal of Korean Society for Atmospheric Environment, 33(2), 139-158. [https://doi.org/10.5572/KOSAE.2017.33.2.139]
- Koo, Y.M., Shin, J.W., Lee, M.S. (2018) Cost-Effectiveness Analysis of Policies to Reduce Fine Dust Emissions in Power Generation and Industrial Sectors. Journal of Korean Industrial Economic Association, 31(1), 347-367, (in Korean with English abstract). [https://doi.org/10.22558/jieb.2018.02.31.1.347]
- Liu, F., Beirle, S., Zhang, Q., Zheng, B., Tong, D., He, K. (2017) NOx emission trends over Chinese cities estimated from OMI observations during 2005 to 2015. Atmospheric Chemistry and Physics, 17(15), 9261-9275. [https://doi.org/10.5194/acp-17-9261-2017]
- Martin, R.V., Jacob, D.J., Chance, K., Kurosu, T.P., Palmer, P.I., Evans, M.J. (2003) Global inventory of nitrogen oxide emissions constrained by space‐based observations of NO2 columns. Journal of Geophysical Research: Atmospheres, 108(D17). http://nrs.harvard.edu/urn-3:HUL.InstRepos:14121856 [https://doi.org/10.1029/2003JD003453]
- NAIR (National Air Emission Inventory and Research) (2020a) Handbook of estimation methods for national air pollutant emissions (IV). http://www.air.go.kr/jbmd/sub54_detail.do?tabPage=1&detailKey=61013P06&inputSchTxt=&typeSchOption=sj&menuId=POT049
- NAIR (National Air Emission Inventory and Research) (2020b) 2017 national air pollutant emissions. http://www.air.go.kr/jbmd/sub54_detail.do?tabPage=0&detailKey=60025P06&inputSchTxt=&typeSchOption=sj&menuId=POT048
- NIER (National Institute of Environmental Research) (2019) Standard operations procedure for the construction of supporting data for national air pollutant emissions. https://www.air.go.kr/jbmd/sub54_detail.do?tabPage=1&detailKey=61012P06&inputSchTxt=&typeSchOption=titleNm&menuId=undefined
- NIER (National Institute of Environmental Research) (2020) Annual Report of Air Quality in Korea 2019. http://www.airkorea.or.kr/web/detailViewDown?pMENU_NO=125
- US EPA (U.S. Environmental Protection Agency) (2006) AP-42: Compilation of Air Emissions Factors, Fifth Edition Volume I Chapter 13: Miscellaneous Sources, 13.2.2 Unpaved Roads. https://www3.epa.gov/ttnchie1/ap42/ch13/final/c13s0202.pdf
- Wang, S.W., Zhang, Q., Streets, D.G., He, K.B., Martin, R.V., Lamsal, L.N., Chen, D., Lei, Y., Lu, Z. (2012) Growth in NOx emissions from power plants in China: bottom-up estimates and satellite observations. Atmospheric Chemistry and Physics, 12(10), 4429-4447. [https://doi.org/10.5194/acp-12-4429-2012]
- Yeo, S.-Y., Lee, H.-K., Choi, S.-W., Seol, S.-H., Jin, H.-A., Yoo, C., Lim, J.-H., Kim, J.-S. (2019) Analysis of the National Air Pollutant Emission Inventory (CAPSS 2015) and the Major Cause of Change in Republic of Korea. Asian Journal of Atmospheric Environment, 13(3), 212-231. [https://doi.org/10.5572/ajae.2019.13.3.212]
- Zhang, Q., Streets, D.G., Carmichael, G.R., He, K.B., Huo, H., Kannari, A., Klimont, Z., Park, I.S., Reddy, S., Fu, J.S., Chen, D., Duan, L., Lei, Y., Wang, L.T., Yao, Z.L. (2009) Asian emissions in 2006 for the NASA INTEX-B mission. Atmospheric Chemistry and Physics, 9(14), 5131-5153. [https://doi.org/10.5194/acp-9-5131-2009]