Current Issue

Annex VI of the International Convention for the Prevention of Pollution from Ships (MARPOL) contains regulations regarding the emission of air pollutants from ships. One regulated target of Annex VI is to reduce the sulphur emissions from ships. At the 70^th session of the Marine Environment Protection Committee (MEPC 70), held in October 2016, the MEPC adopted the sulphur content in the fuel oil on board ships to be less than 0.50% m/m from January 1, 2020, from the previous limit of 3.50%. Therefore, all ships now use marine fuel oil with a sulphur content of less than 0.50%. Strengthening this regulation in MARPOL, Annex VI (the 2020 global sulphur limit) aims to reduce sulphate (SO₄^2- in particulate matter (PM)) and sulphur dioxide (SO₂) emissions from ships. Thus, the 2020 global sulphur limit could reduce PM_2.5 concentrations in and around the area with a higher density of shipping traffic.

In Japan, the environmental air quality standard for PM_2.5 is set as 35 μg m^-3 and 15 μg m^-3 for the average daily and annual concentrations, respectively. Uno et al. (2017) reported a decrease in the impact of transboundary air pollution in Japan, and the achievement rate for PM_2.5 has improved in the past several years. For example, 765 out of 818 (93.5%) ambient air monitoring stations in Japan achieved the standard for PM_2.5 in 2018 (Ministry of the Environment Japan, 2020a). However, the achievement rate remains low at the Okayama and Kagawa prefectures. In 2018, 7 out of 18 ambient air monitoring stations in Okayama (38.9%) and 8 out of 12 in Kagawa (66.7%) achieved the standard for PM_2.5. These rates were the first and second lowest among all prefectures in Japan. Fig. 1 shows that both prefectures are geologically located in western Japan and across the Seto Inland Sea (SIS).

Fig. 1. 
Location and area of Seto Inland Sea (21,827 km²) and Tokyo Bay (1,380 km²). The prefectures surrounding these ocean areas are also highlighted.

The coastal area surrounding SIS contains several large point sources of air pollutants, and the density of ship traffic is higher than the average in Japan. Moreover, SIS is located in a semi-enclosed area surrounded by the Chugoku Mountains (north side) and Shikoku mountains (south side). Thus, it is critical to evaluate the local weather characteristics and contribution of each emission to the air quality around SIS. If air pollutants emitted from ships affect the air quality around SIS, strengthening the regulation of sulphur content in marine fuel could improve the achievement rate for PM_2.5 at prefectures around SIS. The possibility of improvement can also be expected in other areas such as Tokyo Bay with enormous shipping traffic. To evaluate the mechanism of increasing concentrations and consider the efficacy of control measures for PM_2.5, an appropriate air quality model is critical, and emission inventories for air pollutants are required as input data of the air quality model.

In this context, this study develops a database of emission inventories from ships around Japan. One database was estimated using the activity data in the 2015 Japanese fiscal year. The other database was estimated by considering using marine fuel oil with a sulphur content of less than 0.50%. The reduced amount of SO₂ and sulphate, as expected by the 2020 global sulphur limit, is also discussed considering the comparison between the two estimated databases.

This research was performed by the Environment Research and Technology Development Fund ( JPMEERF20185002) of the Environmental Restoration and Conservation Agency of Japan. An outline of the fund and its recent activities, including this research, are introduced in Ohara and Ono (2018).

The emission database for ships, developed by the Sasagawa Peace Foundation, formerly known as the Ocean Policy Research Foundation (OPRF), were widely used as input data in studies regarding the air quality model in Japan (OPRF, 2011). However, because the database was developed based on the activity data of ships in 2005, the emission database for ships had to be adjusted using the latest activity data. Therefore, in this study, while following the method of OPRF (2011), the emission of air pollutants from ships was estimated for fuel combustion in main engine, auxiliary engines, and boilers. In addition to emissions in the exhaust gas, non-methane volatile organic compounds (NMVOCs) emissions discharged from the cargo tanks of oil/chemical tankers were considered. However, this study summarized the methodology and part of the estimated results for emissions in the exhaust gas.

The ship’ operational types for the emission estimation in the exhaust gas consist of ships in navigation, ships at anchor, and fishing boats. Fig. 2 and Fig. 3 are schematic diagrams of the emission estimation approach. The activity data and sulphur content in the fuel were updated for the 2015 Japanese fiscal year. The target substances were SO₂, SO₄^2- in PM, nitrogen oxides (NO_x), NMVOCs, and carbon monoxide (CO).

Fig. 2. 
Schematic diagram of the approach for the emission estimation regarding navigation ships.

Fig. 3. 
Schematic diagram of the approach for the emission estimation of anchored ships and fishing boats.

In this study, the areas of SIS (21,827 km²) and Tokyo Bay (1,380 km²), and the prefectures surrounding these sea areas are highlighted as a comparison analysis of the emission between land and sea. As shown in Fig. 1, the 11 prefectures of Wakayama, Osaka, Hyogo, Okayama, Hiroshima, Yamaguchi, Fukuoka, Oita, Ehime, Kagawa, and Tokushima (in total 61,260 km²) surround SIS, and the three prefectures of Chiba, Tokyo, and Kanagawa (in total 9,612 km²) surround Tokyo Bay.

2. 2 Activity Data

2. 2. 1 Activity Data for Navigation Ships

Data from the Automatic Identification System (AIS) received from the stationary stations on land in Japan were used in this study to calculate the activity for navigation ships applying the same method as OPRF (2011). AIS was originally introduced in 2002 for the purpose of collision prevention, and it exchanges dynamic information (position, speed) among ships and AIS stations. Although the target year for the estimation was the entire 2015 in this study, because of costs, only the monthly AIS data for September 2015 were used for the estimation. Then, the monthly data were extrapolated by multiplying it by approximately 12 to obtain the annual emission amount. Fig. 4 shows the number of ships entering the 161 major domestic ports per day in Japan for each month of the 2015 Japanese fiscal year. Though there was no large difference for the number of ships among each month, September 2015 has the highest number of ships entering these major domestic ports; therefore, the estimated annual emission amounts in this study were conservative results compared with the actual situation. The raw AIS data were aggregated geographically by approximately 1 km², and hourly basis.

Fig. 4. 
Number of ships entering the 161 major domestic ports per day in Japan for each month of the 2015 Japanese fiscal year.

The satellite AIS data are also available (Johansson et al., 2017; Winther et al., 2014) for the emission estimation from ships but were excluded from this study because the capture rate of ships by the satellite became lower in congested water areas such as SIS and Tokyo Bay (Aratani et al., 2015).

Eqs. 3 and 4 show the calculation method of NO_x and SO₂ emissions, of which the required activity data are the engine output (kWh) and fuel consumption (g-fuel), respectively.

NOx emissiong/h=EFsg/kWh×engine-rated ouputkW×load factor%,

(3)

SO2 emissiong/h=fuel consumptiong-fuel/h×S in fuel%×MSO2/MS-sulphate×MSO2/MSO4,

(4)

where EFs is the emission factor as the NO_x limitation required in MARPOL Annex VI, M_SO2 is the molecular weight of SO₂, MS is the molecular weight of sulphur, M_SO4 is the molecular weight of sulphate, and S in fuel is the sulphur content (%) in marine fuels. Furthermore, the load factor and fuel consumption are calculated using Eqs. 5 and 6, respectively.

Load factor%=85%×Vn3/Vs3,

(5)

fuel consumptiong/h=SFCg/kWh×engine-rated ouputkW×load factor,

(6)

where Vn is the navigation speed of each ship (knots), Vs is the service speed of each ship (knots), and SFC is the standard fuel consumption.

The value of Vn was obtained from AIS data. The same values of SFC shown in OPRF (2011) were adopted according to the ship size and type. The engine-rated output and service speed are the individual information of each ship. Furthermore, the built year information of each ship is necessary to determine the EFs of NO_x, which is also individual information. Because individual information is excluded from the AIS data, domestic and global register databases were used to derive the necessary individual information with the AIS data (The Japan Shipping Exchange, 2017; IHS Markit, 2017).

2. 2. 2 Activity Data for Anchored Ships

The emission of air pollutants from anchored ships was estimated based on fuel combustion by auxiliary engines and boilers. The number of ships entering each port in Japan was estimated and classified by ship size and type based on the annual report of the port survey obtained from the Portal Site of Official Statistics of Japan website (https://www.e-stat.go.jp/). The engine-rated output of auxiliary engines, which is crucial for the emission, was estimated from the regression fitting between the enginerated output and ship size created for various ship types using the same methods as OPRF (2011).

The total estimated activity data of each port were distributed equally to every geographical mesh in each port area. The spatial and temporal basis aggregation equaled those for the activity data for navigation ships stated in paragraph 2. 2. 1. Refer to OPRF (2011) for other parameters such as engine load factor and SFC.

2. 2. 3 Activity Data for Fishing Boats

The emission of air pollutants from fishing boats was estimated based on the fuel combustion in engines onboard. The number of fishing boats registered in each prefecture in Japan was estimated and classified by ship size and type, based on the Fishery Census in 2013, which was the closest survey year to 2015. The data can be obtained from the Portal Site of Official Statistics of Japan website (https://www.e-stat.go.jp/). The enginerated output of the main engine, critical for the emission estimation, was estimated from the regression equation between the engine-rated output and ship size using the same methods as OPRF (2011).

The spatial and temporal basis aggregation was the same as those for the activity data for navigation ships stated in paragraph 2.2.1. Refer to OPRF (2011) for other parameters such as engine load factor and SFC, and the method for apportioning the activity data to the ocean area.

This study developed a database of emission inventories from ships around Japan. The first set of the database was estimated using the activity data in the 2015 Japanese fiscal year. The second set was estimated by considering using marine fuel oil with a sulphur content of less than 0.50%. The reduced amount of SO₂ and sulphate, as expected by the 2020 global sulphur limit, was also discussed by considering the comparison between the two estimated databases. In this study, while applying the same method of OPRF (2011), the emission of air pollutants from ships was estimated for fuel combustion in main engine, auxiliary engines, and boilers onboard. The ships’ operations for the emission estimation in the exhaust gas consist of ships in navigation, ships at anchor, and fishing boats. For the emission estimation regarding navigation ships, data from AIS received at the stationary stations on land in Japan were used in this study to calculate the activity. Because individual information, such as the engine-rated output, service speed, and the launch year information, was excluded in the AIS data, domestic and global register databases were used to derive the necessary individual information with the AIS data.

Since the developed database in this study has a high geographical resolution (approximately 1 km²), shipping routes were clearly captured on the emission map of the horizontal distribution. The fuel consumption of domestic ships obtained from the bottom-up approach was 2,819 Gg year^-1. On the other hand, the statistical result of fuel consumption using the top-down approach was 3,180 Gg year^-1. Consequently, the estimated result is approximately 89% of the statistics, suggesting an underestimation but consistency in the estimation for this study.

The emission amounts were compared between the sea (SIS and Tokyo Bay) and prefectures surrounding those ocean areas. The emission amounts of SO₂, PM, and NO_x were the largest on the land surrounding SIS, covering a wide land area of 11 prefectures. The ship emission ratios in total (land and sea) reached SO₂: 29%, PM: 40%, and NO_x: 22% for the SIS area. In Tokyo Bay, the ship emission ratios in total (land and sea) reached SO₂: 36%, PM: 28%, and NO_x: 13%, and the emission intensities per unit area for SO₂ and PM_2.5 are approximately three times higher than those in the other regions, respectively. Therefore, the shipping traffic density is relatively higher compared to transportation on land, and the emission is condensed locally. Assuming that the 2020 global sulphur limit switched to a sulphur content of 0.50%, SO₂ emissions could be reduced to 24% in SIS and 22% in Tokyo Bay, sulphate emissions could be reduced to 23% in SIS and 21% in Tokyo Bay, and PM emissions could be reduced to 39% in SIS and 36% in Tokyo Bay. Although the modeling approach is essential to evaluate the efficiency of emission regulation for air quality, the 2020 global sulphur limit should reduce the emission from ships around the Japan coastal area and improve the air quality in congested water areas such as SIS and Tokyo Bay. Furthermore, the database of the emission inventories from ships developed in this study would contribute to the modeling study regarding the air quality in Japan.