Impact of IMO Sulfur Regulations on Air Quality in Busan, Republic of Korea

Abstract

In this study, we investigate the air quality improvement effect in Busan, the largest port city in South Korea, caused by the implementation of International Maritime Organization (IMO) sulfur regulations. Currently, the Korean government is struggling with problems related to PM_2.5, and ships are one of the major sources of PM_2.5 generation in South Korea. Therefore, we tried to estimate how much the PM_2.5 levels in South Korea could be improved via low-sulfur regulation. According to the Clean Air Quality Policy Support System (CAPSS; National Emission Inventory) in 2016, ship emissions in Busan accounted for 39%, 71%, and 39% of PM_2.5, SO₂, and NO₂ emissions, respectively. To simulate the effect of the IMO’s 0.5 percent sulfur regulation, SOx and PM_2.5 emissions from oil-fueled cargo ships were reduced. Via ship fuel regulation, the PM_2.5 concentration was improved by up to 19% at a site near the port in 2020. In addition, in the case of sulfate, the reduction rate was higher on the downwind side of the Busan port and not near the port, which can be considered as the cause of advection and secondary formation. The PM_2.5 contributions from ships to each of the sub-regions in Busan also decreased by an average of 47% because of IMO sulfur regulation. Although there were limitations in terms of emission estimations because of the application of low-sulfur regulation, we expect that the results of this paper can be used for additional PM_2.5 improvement plans developed by the Korean government and by the local government as well.

1. Introduction

Maritime trade has continued to expand [1]. Air pollutants emitted from ships are known to have negative effects on air quality, human health, and global climate change [2,3,4,5]. Due to these environmental factors, the International Maritime Organization (IMO) has enforced regulations for the emission of nitrogen oxides (NOx), sulfur oxides (SOx), and particulate matter (PM) from ships through the Convention on the Prevention of Marine Pollution [6].

The United States and Canada limited the use of sulfur-based fuel in certain large commercial marine vessels to below 1% in 2012 and to below 0.1% in 2015 when sailing within 200 nautical miles of these countries’ coasts as part of a North American Emissions Control Area (NA-ECA) [7,8]. In this regard, Kotchenruther [7] showed that reductions in PM_2.5 from residual fuel-oil combustion aligned with regulations using PM_2.5 source apportionment. A data analysis on source pollutants also demonstrated that the regulations improved air quality in terms of SO₂ and PM_2.5 levels in Canadian port cities [8]. China set up a domestic emission control area (DECA), located 12 nautical miles away from shore, in 2016. An air quality simulation modeling study reported that the average reductions in the Pearl River Delta were 9.54% and 2.7%, respectively, because of the DECA policy [9]. In addition, a measurement analysis study showed that there was a significant decrease in SO₂ levels by 27−55% after the implementation of the DECA policy in Shanghai [10].

South Korea also started to apply a sulfur content standard for fuel oil for international sailing ships, setting a limit of 0.5% in January 2020 [11]. In addition, emission control areas (ECAs) are designated around five major ports: the Incheon port, the Pyeongtaek/Dangjin port, the Yeosu port, the Gwangyang port, and the Busan port, as well as around the Ulsan port [12]. The permitted sulfur content for ship fuel in emission control areas (ECAs) was 0.1% until 2022 [12]. Song et al. [13] revealed that the implementation of reinforced regulations significantly reduced the SO₂ and NH₄⁺ concentrations at the Busan port through a regression discontinuity design for air-quality data. However, no studies on the results of quantitative effect analyses on the application of IMO regulations using modeling analysis have been conducted in South Korea.

Reducing PM_2.5 concentration is a prioritized subject in South Korea, and the Korean government is struggling to determine what the sources of pollution are and how much this pollution can be reduced. Ship emissions are a major source of PM_2.5 in Korea, and this fact is very important from a policy perspective to examine whether the PM_2.5 concentration can be improved using low-sulfur fuel. In particular, in the case of an area where various emission sources exist within a narrow area, it is necessary to examine the effect of port emission reduction in sub-regions near the port through high-resolution modeling.

This study aims to examine the effect of air pollutants emitted from ships in the Busan port, which is the largest port in South Korea and the seventh-largest port worldwide [1], on air quality in Busan when the IMO strengthens the standards for sulfur content in ship fuels. First, the PM_2.5 contributions in the Busan area from ships were examined based on the national air pollutant emission inventory. Then, based on new ship emissions measured after the application of the low-sulfur content standard, the PM_2.5 contributions were recalculated. To estimate the PM_2.5 and SO₂ improvement effect caused by the implementation of low-sulfur fuel regulations, only the emission reduction effect of low-sulfur fuel was considered under the same conditions as the current national emission inventory (CAPSS). Although there were limitations introduced by not considering all of the processes in detail, such as the changes in vessel activity during the emission calculation process, it is meaningful to estimate the extent of the PM_2.5 and SO₂ improvement effect achieved by applying fuel regulations and to use this result as basic data for future policy planning.

2. Materials and Methods

2.1. Research Area and Model Configuration

Busan is the largest port city in South Korea and was ranked as the seventh-largest port worldwide in 2020, with 21.8 million TEUs for container port throughput [1]. The location of Busan in South Korea is shown in Figure 1. Busan is also an urbanized metropolitan city and has been named South Korea’s second-largest metropolis after the capital city, Seoul, with a population of approximately 3.4 million people. There are various sources of air pollution, such as industrial and non-industrial combustion, as well as on-road and non-road mobile combustion, waste, fugitive dust, etc.

To estimate the effect of IMO sulfur regulation on air quality in the harbor city of Busan, we used the weather research and forecasting model (WRF)–sparse matrix operator kernel emission (SMOKE)–community multiscale air quality (CMAQ) modeling system [14,15]. WRF version 3.9.1 was used for meteorological modeling with the National Center for Environmental Protection (NCEP) FNL 0.25 × 0.25 global product as the initial conditions and boundary conditions.

Table 1. Summary of WRF configuration.

Physics	Selected Option	Reference
Microphysics	WSM 3-class simple ice scheme	[16]
Longwave Radiation	RRTM scheme	[17]
Shortwave Radiation	Dudhia scheme	[18]
Surface Layer	Revised MM5 Monin–Obukhov scheme	[19]
Land Surface	Unified Noah land–surface model	[20]
Planetary Boundary layer	YSU scheme	[21]
Cumulus Parameterization	Kain_Fritsch scheme	[22]

summarizes the configuration of the WRF model. The meteorology–chemistry interface processor (MCIP) was used as a preprocessor to prepare meteorological data for CMAQ.

The following emission inventories were used: the clean air quality policy support system (CAPSS 2016), which is the national emission inventory for domestic anthropogenic emissions; the comprehensive regional emissions inventory for atmospheric transport experiment (CREATE) 2015 for foreign anthropogenic emissions; and the model of emissions of gases and aerosols from nature (MEGAN) for biogenic emissions [23]. To generate CMAQ-ready data and hourly gridded emissions input data, SMOKE was used. CMAQ model v4.7.1 with the AERO5 aerosol module and Statewide Air Pollution Research Center version 99 (SAPRC99) was used, and

Table 2. Summary of CMAQ configuration.

Category	Selected Option	Reference
Chemical Mechanism	SAPRC99	[24]
Advection Scheme	PPM	[25]
Horizontal Diffusion	Multiscale	[26]
Vertical Diffusion	Eddy	[26]
Cloud Scheme	ACM	[27]

summarizes the configuration of the CMAQ model [24].

To set up the modeling domain, we used a nested domain with a 27 km resolution mother domain over East Asia and two daughter domains with 9 km and 3 km resolutions to cover the Korean Peninsula and Busan port area, each as shown in Figure 2.

2.2. Observation Data and Model Validation

In Busan, there are 20 monitoring sites that measure the concentration of air pollutants such as PM_2.5 and SO₂ in real time. The locations of the monitoring sites are shown in Figure 1. The monitoring data are managed and were provided by National Institute of Environmental Research (NIER) [28]. We selected the data from the days measuring more than 90% per day and used these data to identify the status and verify the modeling results.

The assessment of the CMAQ model’s performance was accomplished using the PM_2.5 monitoring data from Busan (Figure 3). Several statistical metrics were used to assess the performance of the model-estimated values. The correlation coefficient (R) value of PM_2.5 between simulated and observed values was 0.74, satisfying the target standard of the R value of 0.4 or more recommended by Emery et al. [29]. The index of agreement (IOA) and root mean square error (RMSE) were 0.81 and 8.9 μg/m³, respectively, and it can be seen that overall model results simulate the observation results well.

2.3. Estimation of Contribution Concentration

To estimate the SO₂, PM_2.5, and NO₂ contribution concentrations, we utilized the brute-force method (BFM) as the contribution estimation method [30,31]. The “brute-force” method consists of estimating the concentration change by performing and subtracting two simulations, one with and one without a specific emission source, to be analyzed [32]. Although this method has some limitations, it is efficient and practical for analyzing the contribution of regional emissions.

We calculated the contribution concentrations and the contribution rates (%) of the SO₂, PM_2.5, and NO₂ concentrations introduced by ship emissions via the following equations:

$$Contribution concentration=Concentratio{n}_{with}-Concentratio{n}_{no}$$

(1)

$$Contribution rate (\%)=\frac{Concentratio{n}_{with}-Concentratio{n}_{no}}{Concentratio{n}_{with}}\times 100$$

(2)

where Concentration_with and Concentration_no represent the averaged surface concentrations of pollutants with and without shipping emissions, respectively [33]. In addition, we calculated the reduction rate of change in the PM_2.5 and SO₂ concentrations caused by IMO sulfur regulation via the following equations:

$$Reduction rate (\%)=\frac{Concentratio{n}_{CAPSS2016}-Concentratio{n}_{IMO}}{Concentratio{n}_{CAPSS2016}}\times 100$$

(3)

where Concentration_CAPSS2016 and Concentration_IMO are the averaged PM_2.5 concentrations with the CAPSS2016 emission regulations and newly applied emission regulations, respectively.

3. Results

3.1. Contributions of PM_2.5 from Ship Emissions in Busan

Figure 4 shows air pollutant emissions from ships in local regions in South Korea based on CAPSS 2016. The Busan area shows the highest level of air pollutant emissions from ships compared to other areas. The SOx emissions from the ships in the Busan port are 7638 ton/year, representing 26.6% of the SOx emissions of all ship ports in South Korea. In the case of PM_2.5 and NOx, the ship port emissions in Busan account for 25% and 26% of the emissions for all of the ship ports, respectively. This reveals that a significant amount of cargo is concentrated in the Busan area.

Ship emissions represent the largest source of pollutants in Busan, representing a non-road mobile source (Figure 5). In the case of SOx emissions, ship emissions accounted for 71% of the total emission sources, representing the highest source. In the case of PM_2.5 emissions, the proportion of ship emissions was 39%, followed by fugitive dust. The amount of NOx emitted from on-/non-road mobile sources, including from ships, accounted for about 70% of the total emissions. This implies that the air quality in Busan is largely affected by the emissions from ships.

To quantitatively calculate the impact of ship emissions in the Busan port before the implementation of IMO2020, the contribution rates (%) of the SO₂, PM_2.5, and NO₂ concentrations were derived using the above-mentioned CAPSS2016 emissions (

Table 3. Annual contribution rates (%) of SO₂, PM_2.5, and NO₂ for each local region from ship emissions from Busan.

Local Region	Contribution Rate (%)
	SO2	PM2.5	NO2
Seoul	0.04	0.04	−0.01
Incheon	0.02	0.02	0.00
Gyeonggi-do	0.04	0.04	0.00
Chungcheongnam-do	0.03	0.05	0.02
Sejong	0.12	0.08	0.04
Daejeon	0.13	0.08	0.01
Chungcheongbuk-do	0.11	0.08	0.04
Gwangju	0.30	0.15	0.07
Jeollabuk-do	0.11	0.10	0.05
Jellanam-do	0.18	0.21	0.14
Busan	50.92	6.79	19.27
Ulsan	0.19	0.35	0.43
Daegu	0.27	0.23	0.14
Gyeongsangnam-do	4.01	0.54	2.83
Gyeongsangbuk-do	0.21	0.20	0.20
Gangwon-do	0.07	0.03	0.03
Jeju-do	0.13	0.13	0.04
Nationwide	4.16	0.43	1.62

). As expected, the ship emissions in the Busan area had the highest effect on each pollutant concentrations in the Busan area compared to in other areas. It was simulated that the ship emissions in the Busan port affect more than 50% of the SO₂ concentration in the Busan area. The effect was calculated at a contribution level of 6.79% for the annual mean PM_2.5 concentration, and this is the result of including both the effects of the PM_2.5 emitted directly from ships and the secondary formation determined by precursors. In addition, the effects of the air pollutants emitted from ports in Busan also affected the air quality in Gyeongsangnam-do, which is located near Busan. The reduction in ship emissions in the Busan port is not only an important means to improve the air quality in the Busan area but is also expected to improve the air quality in neighboring areas.

The monthly variability in the contribution concentration and contribution rate of ships in the Busan port on air quality was analyzed (Figure 6). In the case of SO₂, the contribution concentration was high in June, and the contribution rate was high in July. In the case of PM_2.5, the variability in the contribution rate was larger than the change in the contribution concentration, and the contribution rate was particularly high in August. One of the reasons related to this is that the concentration of PM_2.5 in summer is relatively lower than it is in other seasons; however, the influence of ships is rather large. In the case of NO₂, the contribution concentration and the contribution rate were bimodal, and peaks appeared in May and August. Although the monthly variability was different for each air pollutant, the contribution concentration and contribution rate in summer are higher than in winter. This indicates that the air pollutants that move from the port via the southeasterly wind in summer have a severe effect on the inland portion of the Busan region because of the influence of the monsoons when the monthly variability in ship emissions is not taken into account in the model.

The effects of ship emissions on air quality in the Busan area were examined by dividing them into sub-regions within the Busan area (Figure 7). First, the distribution of the annual average concentration of PM_2.5 in the Busan area showed that the PM_2.5 concentration was found to be higher at the sites separated to the northeast or west, which are industrial and urban areas, than near the port, and exceeded the national ambient air quality standard (15 µg/m³). The PM_2.5 contribution concentration by ships also varies significantly by region in Busan. The PM_2.5 contribution by ship emissions was the highest near the port and was about 5.78 µg/m³ in high-concentration places and about 0.09 µg/m³ in low-concentration places, a difference of about 60 times. In addition, even at the same distance from the port, the concentration is relatively higher in the north-northwest direction, which has been established to be the effect of the wind direction in the coastal area reflecting the topography Figure 8). Figure 8 shows the wind-field modeling with a resolution of 1 km resolution and shows the wind conversion caused by the inland wind and the effects of topography where the mountain is located. These areas are not only close to the port but also close to the main station in Busan, which has a number of mobile emission sources. The high PM_2.5 concentration contribution in those areas is estimated to be due to the combination of the wind effect and emission sources.

Figure 9 shows the annual mean contribution of the PM_2.5 concentration from ship emissions by dividing the emissions between daytime and nighttime. As shown in Figure 7, the contribution of the PM_2.5 concentrations near the port is larger than it is inland because of the influence of the wind direction. The high contribution of the nighttime PM_2.5 concentrations is estimated to be due to there being low emissions from other sources during the nighttime and because the low planetary boundary layer (PBL) in the ocean at nighttime creates conditions that are insufficient for the diffusion of pollutants.

3.2. Characteristics of Changes in Air Quality in Busan According to IMO2020

Among the ship emission sources in Busan, cargo ships account for more than 97% of the total air pollutant emissions from all ships (

Table 4. Air pollutant emissions and proportion by type of ship in Busan in 2016 (unit: ton/year).

Emission Source	SOx	PM2.5	NOx
Cargo ship	7632.78 (99.93%)	958.50 (97.38%)	18,666.89 (97.26%)
Ferry	1.33 (0.02%)	0.12 (0.01%)	1.93 (0.01%)
Fishing boat	4.23 (0.06%)	10.88 (1.11%)	520.63 (2.71%)
Leisure ship	0.00 * (0.00%)	14.75 (1.50%)	3.83 (0.02%)

* 0.01 kg/year.

). The low SO_x emissions from leisure ships are due to their fuel type and fuel usage. Leisure ships have lower fuel usage than other marine emission sources, and leisure ships rely on gasoline, which has a profoundly low sulfur content. Bunker fuel, notably used in cargo ships, can be categorized into Bunker-A, Bunker-B, and Bunker-C. These are classified according to the fuel’s flash point, specific gravity, viscosity, sulfur content, etc. Bunker-C oil has a higher viscosity and sulfur content than Bunker-B oil and Bunker-A oil.

When the sulfur content standard applied was 0.5% under the IMO regulations, a scenario in which the emissions of SOx and PM_2.5 decrease depending on the oil type was constructed. Since SOx emissions are proportional to the sulfur content of the fuel, the reduction rate was calculated using the ratio of the reinforced sulfur content standard and the sulfur content of each oil type [34]. The PM_2.5 reduction rate was calculated based on the equation between the sulfur content and the PM emission factor presented in previous studies [35].

Table 5. Sulfur content for each oil type and reduction ratios of SOx and PM_2.5 emissions in CAPSS2016 when applying IMO2020.

Oil Type	Emission Reduction Percentage
	SOx	PM2.5
Bunker-A oil	66.7%	48.8%
Bunker-B oil	57.9%	22.0%
Bunker-C oil	85.7%	55.1%

summarizes the reduction rate of air pollutant emissions from ships according to the application of the IMO regulations by oil type.

The difference between the contribution of air pollutants calculated based on the CAPSS2016 that is due to ships in Busan and the contribution by the emission-change scenario according to IMO regulations was examined (Figure 10). The SO₂ concentration was significantly reduced because of the introduction of IMO fuel regulations near the Busan port. In addition, some decreasing trends can be confirmed in the northwest direction from the Busan port. PM_2.5 also has a similar concentration change pattern to SO₂, but the region with the second-highest maximum is more pronounced in the region to the northwest of the Busan port. This result is related to sulfate formation; sulfate’s secondary formation showed a higher contribution-rate change in the area to the northwest of the Busan port. This means that the effect of the direct discharge of air pollutants from ships appears near the port, but the effect of secondary generation appears in areas away from the port, so it is necessary to consider the secondary formation of PM_2.5 in air quality management.

As a result of applying the IMO regulations, the annual average PM_2.5 concentration improvement rate in Busan was about 10%, and the local PM_2.5 concentration improved from a minimum of 6.7% to a maximum of 18.8% (Figure 11). In addition, the contribution of the PM_2.5 concentration from Busan port ships in Busan’s sub-regions was reduced by an average of 47.2%, from a minimum of 37.6% to a maximum of 53% because of the implementation of the IMO regulations. The highest improvement effect for PM_2.5 and SO₂ concentration was found at the branch near the Busan port (221241 Sujeong-dong). As an example, the Deokpo-dong (221282) site, which is located to the north-northwest of the Busan port and exceeds the national ambient air quality standard, also showed a significant air quality improvement effect.

4. Conclusions and Discussion

This study investigated the air quality improvement effect in the Busan area in terms of assessment of the policy effect after the application of regulations for the sulfur content in ship fuel in the Busan port, which is the largest port in South Korea and the seventh-largest port in the world, through high-resolution modeling simulation. Among the sources of PM_2.5, SOx, and NOx emissions in Busan, the largest proportion was found to be from ship emissions in the port. Accordingly, the air quality contribution concentrations were determined to mainly be from air pollutant emissions from ships in the Busan Port and were shown to have a significant effect on the air quality in the Busan area, as well as in the surrounding local regions. These concentrations were found to have the greatest impact on the air quality in Busan, as well as in some adjacent areas.

The effect of ship emissions on the Busan area was examined temporally and regionally. Although the monthly contribution concentrations and contribution concentration patterns of ship emissions on air quality were different for each pollutant, they showed a common characteristic, showing a higher contribution rate in summer than in winter. In addition, the relative contribution of the sub-regions in Busan was high not only in the areas near the Busan port but also in the sites located to the north-northwest of the Busan port. It showed a higher contribution rate during the nighttime than during the daytime. Overall, the characteristics of the port being located near the ocean, the influence of monsoon and sea and land breeze, and the topography were comprehensively considered, and it was determined that the influence of the air pollutants from ships in the port might be affecting not only the area near the port but also the major downwind areas, exceeding the national ambient air standard. To apply a scenario in which sulfur content regulations are implemented for ship fuel, a model simulation was conducted by reducing the PM and SO₂ emissions from ships. The concentrations of PM and SO₂ decreased significantly near the port, and it is noteworthy that the sulfate concentration showed a rather high concentration at the point separated from the port to the northwest after secondary formation. This shows that ship-fuel regulations in the Busan port area can effectively improve the air quality of Busan and can improve the air quality in neighboring areas. In addition, when advection and secondary formation are considered, regulations can help to significantly improve the inland air quality in Busan.

However, there are some limitations for the estimation of emissions in this study. The national emission inventory for 2020 when the IMO sulfur regulation was implemented has not been prepared. As an alternative, the sulfur content and PM emissions in ship fuel were directly estimated from the 2016 national ship emissions. In this process, it is difficult to collect the data at the national level to consider the vessel activity, which is a major factor in estimation of the emission, so we assumed the vessel activity at the same level as in 2016.

Although the results have some limitations as mentioned above, it is meaningful to examine the effects of implementation of low-sulfur-fuel regulation and to support future PM_2.5 management planning. In addition, this is the first study to show quantitatively how much IMO regulation can improve air quality in the port city and how the contribution of ships’ emissions to air pollutant concentration changes. Especially, it was possible to examine the effect of port emission on the city center separated from the port using the high-resolution photochemical reaction modeling that considers secondary formation. In addition, through this study, it is possible to know whether additional policies are needed to improve air quality in port cities.

The Korean government is trying to reduce the port emissions in many ways; however, it is not easy to find priority policies. Due to the IMO regulations, the total SOx emissions have decreased, and, as a result, PM_2.5 has also decreased; however, it is still evident that we need a counterplan. The Korean government is planning to increase financial incentives to encourage low vessel speeds in the public sector and to make the conversion to eco-friendly fuels in the private sector. Additionally, the government is supporting the installation of alternative maritime power (AMP) and the reduction in electricity fees to reduce ship emissions.

Although low-level sulfur fuels are used, it is evident that the NOx emissions generated from combustion of ship fuel are still present; therefore, NOx management plans should be prepared. NOx emission control areas (NECAs) have been designated by the IMO and are only in effect in the North Sea and in the Baltic Sea. Korean ECAs only regulate SOx. It is also necessary to manage the VOC emissions from crude oil storage facilities and paints during ship operation. Therefore, to improve air quality, the management of various target materials should be considered.

Furthermore, in the case of ship emissions, more efficient management can be achieved when similar regulations are applied to ships arriving at adjacent ports in China and Japan, as well as to ships on the coast and arriving in port areas of Korea. Therefore, it is necessary to collaborate with neighboring nations to maximize the effect of the regulations and management.

To set up a better air-quality improvement policy, we need to understand how the target emission influences air quality. It is an important role to provide quantitative and scientific data that can be used to set up policies, even if some uncertainties are included at the current level. From this point of view, this study is meaningful and further work is continuing to drive policy priority.