A numerical approach to assess air pollution by ship engines in manoeuvring mode and fuel switch conditions

Abstract

Due to increasing worldwide trade, transportations of goods through seaports have been gradually growing in the past years; harbours are main centres of economic activities, with unavoidable consequences on the air quality degradation and human health of many urbanized ports. This study provides a methodology to assess the impact of pollutant emission from marine engines on ambient air quality of the coastal areas. The environmental pollution from sea traffic was evaluated by assessing the production and atmospheric dispersion of exhaust emissions produced by ship engines in manoeuvring mode and in fuel switch conditions from heavy-sulphur residual fuel oil to low-sulphur distillate fuel oil. Specifically, this study analyses the emissive behaviour of merchant ships, equipped with large size diesel engines and transiting in the port of Naples, which is located very close to the densely populated urban centre. The spatial distribution of the air pollutant concentrations was calculated with a specific dispersion modelling approach, delivering significant data about the environmental impact of merchant ship emissions on the coastal urban area. The NO_X, PM and SO_X pollutant concentrations estimated with this methodology were also compared with pollution levels measured in an experimental monitoring campaign performed in the port of Naples. Estimated concentration levels were lower than measured values because the latter were affected by other anthropogenic emissive sources adjacent to port area, above all exhaust emissions from the road transport sector. Both the calculated and measured concentration levels over coastal area were below the European Limit Values.

Keywords

Ship emissions marine diesel engines air pollutant dispersion fuel change-over low-load/manoeuvring mode international policy strategies

Introduction

Nowadays, many efforts have been made all over the world to decrease air pollutant emissions from anthropogenic sources (power generation, industrial, road transport sector), but the relative weight of the marine transport sector to the total emission is still growing. In European coastal areas, pollutant emissions from marine engines, in fact, contribute meaningfully to air pollution worldwide. Since significant contributions to the total anthropogenic NO_X and SO_X emissions are ascribable to seagoing ships, and approximately 70% of ship emissions are assessed to befall inside 400 km of land, intense pollutant emission from ship engines has the potential to worsen meaningfully the air quality state of coastal and port areas.^1,2

Port emissions do not contribute significantly to the global exhaust emissions from the shipping sector; however, ship emissions have a direct consequence on the population of many urbanized coastal areas and ports, which inexorably constitute significant point sources of ship exhaust pollution, such as sulphur oxides (SO_X), nitrogen oxides (NO_X) and particulate matter (PM).³ Ports, in fact, constitute surely intense areas of marine transport, considerably affecting air quality state and human health in the adjacent areas.^4,5 Furthermore, the marine transport sector will grow in the next future and with it ship exhaust pollution and its consequence on coastal areas;^6,7 therefore, it is very important to appropriately assess the impact of marine traffic on the atmospheric composition of coastal regions.

Despite this, even though marine transport sector contributes considerably to the global transportation sector, ship exhaust pollution is one of the least regulated anthropogenic source, and its estimation is not always quantified. For this reason, it is essential to examine the effect of the current ship emissions on the atmospheric composition, and how the foreseen growth of the marine transport and the geographic extension of the ports, conceivably combined with the implementation of international rules, are going to influence the air quality state of the port areas.⁸ It is commonly a hard assignment to evaluate a reliable value to the ship pollution rates in the port areas. Although scientific literature has been improved recently by some investigations on ship exhaust pollution,^9–15 the scientific publications assessing the effect of shipping emissions on air quality state of coastal areas are yet limited, especially in relation to low-load manoeuvring and fuel switch conditions.^16–19

Emission inventories from marine transport in port areas, in fact, are typically less defined than those of vessel cruising (at sea), since port activities are not commonly well refined, and the pertinent emission factors are based on incomplete and obsolete information. Emission inventories in port areas to be dependable must be based upon in-port vessel activity data, but this information is not easy to get to a detail that defines engine type, ship engine mode and operating times, ship movements and fuel type.^20,21

Besides, since ship emissions in port areas are considered as complex sources owing to the aforementioned reasons, atmospheric emissions from these sources do not conform to conventional dispersion modelling source configuration. Therefore, emission quantification for dispersion modelling and air quality impact assessment aims, as well as the mitigation measures for these sources, proves difficult owing to insufficient emission data and methods for assessing emission metrics.²²

Moreover, some results have been obtained using dissimilar methodologies in the different regions, as no common EU-wide modelling approach now exists. Consequently, a more standardized application of these methodologies seems to be a real possibility to obtain comparable information on the effect of shipping emissions on the air quality state in urban areas adjacent to the ports. Also, in order to project and implement actual rules to reduce environmental impacts of ship exhaust pollution in coastal areas, comprehensive data are necessary for the environmental management to evaluate the effect of these emissions on air quality degradation.

This paper responds exactly to all these requirements by suggesting a modelling approach for evaluating the rate of emissions and the environmental impact of marine traffic on air quality state of coastal zones. In this study, a particular dispersion modelling approach is applied to ship engines in manoeuvring and fuel switch conditions, in order to estimate the spatial distribution of air pollutant concentrations in coastal areas. The methodology suggested in this paper can be considered as a valuable tool to characterize (in terms of pollutant concentrations) the impact of the exhaust emissions from ship engines during the approach phase on air quality degradation in the port areas. Manoeuvring mode, in fact, implicates slow speed movement of vessels into and out of berths, which can involve higher exhaust emissions than those relating to optimum operating conditions.

Specifically, this investigation deals with the emissive behaviour of merchant ships equipped with large size two-stroke diesel engines and transiting in the coastal area of Naples (moving towards and away from the port), taking into consideration two particular effects: first, the fuel switch conditions from heavy-sulphur residual fuel oil (HSRFO) to low-sulphur distillate fuel oil (LSDFO) and subsequently, the engine operating conditions at low load during manoeuvring mode.

The port of Naples is one of the key ports of departure and destination for cruises in the Mediterranean, and it is the chief harbour for merchant ships of southern Italy, making a significant contribution to air pollution in the whole coastal area. Both traffic of passengers and goods take advantage of more than 11 km of banks (and 75 mooring points), around 3 km road infrastructures and 2 km of rails connected with the national road network.²³ The polluting concentrations from sailing and arriving ships and the resultant air quality state in the coastal area close to the port of Naples were achieved using a dispersion model, estimating the spatial distribution of air pollutant concentrations in the overall area under examination for two different simulation scenarios. The NO_X, PM and SO_X pollutant concentrations calculated with this dispersion modelling approach were also compared with both a monitoring campaign performed close to the port of Naples and limit values established by European legislation, thus obtaining interesting results.

The calculation procedures presented in this study can be used by the local environmental authorities to evaluate how the forecast future developments of ship traffic and geographic growth of the ports can influence the air quality state of the coastal areas. This methodology, in fact, is rather general and can be used as an effective appraisal tool to evaluate both the agreement of air quality levels with the legislative limit values and the effect of new emissive scenarios from sailing and arriving vessels on the air quality levels. Lastly, it is important to underline that many variables used in this study (engine load factor, engine operating conditions, emission factors, meteorological factors and in-port ship activities) are based on some approximations and assumptions, so influencing the reliability of the results.

Methods

The influence of ship exhaust pollution on air quality state of coastal areas was assessed within some scientific publications worldwide.^24–27 These research advise that a significant percentage of air pollutants in the port areas may come from marine transport sector, in comparison with emissions from other anthropogenic sources. A realistic correlation between the emissions from ship engines (in manoeuvring and fuel switch conditions) and the air quality state in the areas close to the ports is very hard to build. This is due to different aspects characterizing the manoeuvring mode of the ships in port areas. First, the determination of the actual engine operating conditions and the real power rated by engines in manoeuvring mode are possible only with some approximations and assumptions. Besides, to attain a good relationship between ship emissions and air quality, the effect of the weather factors, the dispersion method, the precise position of ships and other factors having some influence on the dispersion process of the pollutants in atmosphere must be analysed with great accuracy.

The aim of the present research is to build a consistent calculation procedure able to correlate exhaust emissions from marine engines in manoeuvring and fuel switch conditions with the air quality state in terms of emissions production and dispersion in the ports areas. In this section, the main criteria and methods applied for the purpose of the paper will be explained in detail.

The fuel changeover

Limits for pollutant emissions from marine engines have been gradually restricted in recent years. In particular, SO_X, NO_X, ozone depleting substances and volatile organic compounds (VOCs) are regulated by Annex VI of MARPOL 73/78 (Marine Pollution), and its further amendments by IMO.

Among them, SO_X are strictly dependent on the sulphur content in fuel. For this reason, regulation has been requiring lower and lower values of it. Currently, the sulphur limit in marine fuel is 0.1% in Sulphur Emission Control Areas and port areas and 3.5% in all other areas. Two strategies can be adopted to comply with these restrictions: the first (more expensive at the moment) is the installation of seawater scrubbers on board, reducing the SO_X content in exhaust emissions;²⁸ the second is the fuel switch from residual (HSRFO) to distillate (LSDFO) fuel oil when crossing emission control areas. The first strategy is being widely adopted in new ships, while generally for existing ships the second is more convenient, at least until (presumably 2025) the limit will drop from 3.5% to 0.5%. The ships approaching the port areas are particularly relevant about emissions for two reasons: first of all, the ship is in manoeuvring mode, and the engine works at part load; moreover, the vessel is performing the fuel switch. Both these conditions alter the normal levels of engine emissions.

This analysis focused on merchant ships transiting in the port of Naples, equipped with large size two-stroke diesel engines. We referred to typical emission levels of engines of about 10,000 KW power, as reported in Tables 1 and 2.

Table 1.

Emission factors from HSRFO combustion in a 9800-kW diesel engine.

	Engine load 100%	Engine load 75%	Engine load 50%	Engine load 25%
Engine power (kW)	9800	7350	4900	2450
NO_x (g/kWh)	14.60	15.70	17.50	17.20
SO_x (g/kWh)	2.96	2.91	3.33	3.l4
PM (g/kWh)	1.7	1.9	2.2	2.4

HSRFO: heavy sulphur residual fuel oil; NO_x: nitrogen oxides; PM: particulate matter; SO_x: sulphur oxides.

Table 2.

Emission factors from LSDFO combustion in a 9800-kW diesel engine.

	Engine load 100%	Engine load 75%	Engine load 50%	Engine load 25%
Engine power (kW)	9800	7350	4900	2450
NO_X (g/kWh)	12.98	14.71	15.43	13.53
SO_X (g/kWh)	0.178	0.175	0.177	0.189
PM (g/kWh)	0.3	0.5	0.7	0.9

LSDFO: low sulphur distillate fuel oil; NO_x: nitrogen oxides; PM: particulate matter; SO_x: sulphur oxides.

NO_X and SO_X emission factors were friendly delivered from MAN DIESEL, while PM ones are those reported in significant scientific publications.^29,30 NO_X emissions are not universal for any type of engine, since they depend on maximum combustion temperatures and combustion process time. By observing the two tables, the specific emissions of NO_X do not show any remarkable differences between the fuels, whereas a considerable difference exists between emission factors of SO_X and PM.

Values for all the intermediate power rates can be evaluated with different approaches.³¹ Many technical and operational issues are related to the switch form residual to distillate fuel, and specific procedure must be adopted in order to avoid loss of power, engine failure and ‘noncompliance’. First of all, the fuel service system is required to be fully flushed through before entry into the ECA-SO_X, in order to comply with regulations. From the technical point of view, due to consistent viscosity differences, the fuel switch procedure involves crucial fuel temperature differences.

The components of the engine, on the other hand, cannot undergo thermal shock and then the changeover procedure must last several hours, and it has to begin well before approaching the ECA. Last but not the least, a different lubricating behaviour and base number of fuels must be taken into account when defining a correct switch procedure.^32,33

The low-load operating mode

In this research, the low-load/manoeuvring mode includes some ship activity in delimited speed areas related to anchorages and ports, in addition to the activity of manoeuvring out and into of berths. The definition of the times during which ships are manoeuvring into and out of anchorage and berths can be based on location and speed, which in theory allow low-load/manoeuvring mode to be dealt as a distinct operating mode.

The engine load factor, lf, is the ratio of the power necessary to the maximum installed power, P/P_max, and it is calculated for individual ships of a trip using data of the ship’s location at stated periods.¹ The movement of each ship is always characterized by the distance d_a–b between two fixed waypoints a and b and the transit time t_a–b of arrival or departure of the ship. In equation (1), these values are used to estimate the average ship speed s_a–b. Fractional load factor, lf, is calculated in equation (2); in this equation, the power necessary, P, is considered proportional to the cube of the vessel speed. This equation is characterized by two characteristic speeds for each ship (specified in commercial databases): the service speed, s_ss, namely, the nominal cruising speed at which the vessel is designed to operate, and the hypothetical maximum ship speed at maximum engine power, s_max. In the same equation, the coefficient f is the fraction of maximum engine power required for the ship’s service speed, s_ss. The work W_a–b delivered by the main engine and required for the propulsion of the vessel between the given waypoints a and b is calculated in equation (3)

s_{a - b} = d_{a - b} / t_{a - b}

(1)

l f = p / p_{max} = {(s_{a - b} / s_{max})}^{3} = f \cdot {(s_{a - b} / s_{s s})}^{3}

(2)

W_{a - b} = P_{m a x} \cdot l f \cdot t_{a - b}

(3)

The share of maximum engine power required for the ship’s service speed is not exactly definite; however, typical values of the coefficient f range between 0.80 and 0.85, as well explained in recent scientific publications.^34,35 The cubic relationship shown in equation (2) is broadly applied to calculate engine power required at several ship speeds, while the engine power necessary at design speed, s_ss, is used as a choice of the size of the main engine with a margin depending on other factors that influence the accuracy of power prediction. These factors are the result of currents, waves, wind, ship trim and fuel quality. However, the deviation from the cubic relationship (equation (2)) on power estimate is substantial only at actual speeds meaningfully higher than the service speed.

The appraisal of vessel exhaust emissions involves the application of emission factors to specific ship activities, whereas the emission factors are characteristic values attempting to correlate the emitted pollutants with the operating mode of the vessel engines. Vessel speed and engine-operating mode are used to calculate engine power rate as explained in the previous equations; subsequently, ship exhaust emissions are calculated adopting suitable emissions factors for the specific ship type, fuel type (residual fuel oil HSRFO, distillate fuel oil LSDFO, gasoline), engine-operating mode (anchored, transit, low-load/manoeuvring) and engine type (slow-, medium- and high-speed diesel, gas turbine, steam turbine for large ships, gasoline 2S and gasoline 4S for small vessels).

In the ‘bottom-up’ method, emission estimates are based on activities of individual ships; exhaust emissions of a generic ship are assessed by using equation (4), in which the emission factors are given as mass of exhaust emissions per unit of engine work, and where

E_i,j,z,m (kg): Emission of pollutant i, from engine type j, using fuel type z, during operating mode m

P_max,j (kW): Maximum installed power for engine type j

lf_,j,m (–): Engine load factor for engine type j, during operating mode m

t_j,z,m (h): Time spent by engine type j, using fuel type z, during operating mode m

ef_i,j,z (g/kWh): Emission factor for pollutant i, from engine type j, using fuel type z

LLA_i (–): Low-load adjustment factor for pollutant i

E_{i, j, z, m} = (P_{max, j} \cdot l f_{j, m} \cdot t_{j, z, m} \cdot e f_{i, j, z} \cdot L L A_{i}) / 1000

(4)

Since this study covers emissions from merchant ships transiting in the coastal area of Naples, emission factors ef_i,j,z are deduced on the basis of these assumptions: PM, NO_X and SO_X as pollutants, large size two-stroke diesel engines as engine type, residual (HSRFO) and distillate (LSDFO) fuel oil in switch conditions as fuel type and low-load/manoeuvring as the operating mode. The pertinent values of PM, NO_X and SO_X emission factors are deduced in Table 2.

It is hard to precisely model the emissive behaviour during the low-load/manoeuvring mode since this specific operating mode can involve meaning swings in engine load; the low-load adjustment factors LLA_i take in consideration this particular engine-operating condition. In scientific literature, the number of studies and experiments performed during low-load operating mode is smaller than for higher engine loads; therefore, the reliability of these adjustment factors is not high.³⁶ As explained in equation (4) and stated by wide review works, the approach assumed in this study to model the engine low loads during the manoeuvring mode is to multiply emission factors ef_i,j,z (resulting by steady-state load conditions) by low-load adjustment factors LLA_i for all diesel-cycle engines.^1,37 However, these adjustment factors will only be accounted if the engine load factors drop below 20%, and consequently, the emissions factors are adjusted in accordance with the following considerations.

Generally, diesel ship engines are not as efficient when they operate at low engine loads, then during harbour maneuvering or when vessels move slowly at sea in reduced speed zones. In fact, while exhaust emissions, expressed in terms of mass per unit time, tend to go down as ship speed and engine load decrease; on the other hand, the emission factors, expressed in terms of grams per kilowatt-hour, tend to grow. This circumstance is based on the remark that compression-cycle combustion engines are less efficient at low engine loads. In other words, when main diesel-cycle engines operate below 20% load, lessened engine efficiency involves increased emissions factors, which are considered for by applying correction factors as a function of engine loads, as shown in equation (5)

L L A_{i} = \frac{a \cdot {(l f_{j, m})}^{- c} + b}{e f_{i, j, z}}

(5)

For NO_X and PM emission factors, the low-load effect for the range of load factors between 2% and 20% is described in equation (5); in this way, the low-load adjustment factors LLA_i are generated as a function of the engine load factors lf_,j,m, whereas Table 3 presents the coefficients used in the same equation. For engine loads below 20%, therefore, the low-load adjustment factors LLA_i increase so as to reflect increased emissions due to engine inefficiency. Obviously, the low-load adjustment factors LLA_i are not applied at engine loads higher than 20%. Besides, these adjustment factors are not applied to steamships or ships having gas turbines.

Table 3.

Low-load emission factor regression equation coefficients for PM and NO_X.

	Coefficient
Pollutant	a	b	c
PM	0.0059	0.26	1.5
NO_X	0.1256	10.45	1.5

NO_x: nitrogen oxides; PM: particulate matter.

The emission dispersion model

Predicting the air quality state in urban environment as result of numerous anthropogenic pollution sources (industry, energy, road transport sectors) involves an assessment of the atmospheric dispersion process (also accounting meteorological parameters, chemical reactions, natural removal processes, etc.) to calculate the concentrations of the air pollutants in the areas around the emissive sources after the release.³⁸

The particular dispersion model adopted in this paper is a Gaussian model, suitably adapted to the aims and main characteristics of this investigation. Gaussian models analyse atmospheric gas dispersion as being carried downwind in a fixed plume, under horizontal and vertical mixing in the area adjacent the emissive source. The plumes of air pollutants diffuse horizontally and vertically, with a resulting decrease in concentration as they move downwind, also considering aerodynamic downwash due to adjacent constructions and obstacles. Mixing with the adjacent atmosphere is greatest at the edge of the plume, so causing lower pollutant concentrations outward (horizontally and vertically) from the centre of the plume.

The pollutant concentration calculated for each source at each receptor is summed to obtain the total concentration produced by the combined source emissions. For a steady-state Gaussian plume, the pollutant concentration is calculated as in equation (6)

C_{i} (x, y) = \frac{E_{i}}{π u σ_{y} σ_{z}} e^{(\frac{- H^{2}}{2 σ_{z}^{2}})} e^{(\frac{- y^{2}}{2 σ_{y}^{2}})}

(6)

C_i(x,y) (g/m³): concentration of pollutant i to the receptor level at downwind distance x and crosswind distance y

E_i (g/s): flow rate of pollutant i

H (m): effective source elevation

u (m/s): mean wind speed at release height

σ_y (m): lateral dispersion coefficient

σ_z (m): vertical dispersion coefficient

Although the basic hypothesis of the Gaussian dispersion approach, that is the assumption of ‘normal’ distribution of the air pollutant concentrations, befalls above all in wholly ideal conditions of atmosphere homogenous turbulence, generally the numerical results obtained with these dispersion models show a good degree of approximation with experimental campaigns, especially for assessments in the long-term version. Instead, the Gaussian dispersion models are not much suitable to study the atmospheric gas dispersion of unstable pollutants, in non-homogenous conditions of atmosphere and under complex orography situations (in presence of obstacles and buildings). For the precise requirements of the present investigation, the model was utilized in the long-term version; the hourly meteorological data, by a statistical processing, were grouped into joint frequencies of event for the particular wind direction, atmospheric stability and wind speed classes. Consequently, this model calculates atmospheric stability, wind speed and wind direction as average values on the whole year.

Results and discussion

The present work has been focused on exhaust emissions from marine engines emitted by merchant ships transiting in the port of Naples, since there is growing awareness of the environmental impact of maritime transport in the coastal area. The port of Naples extends over a great part of the coastal area, with a whole length of 11 km of banks, 75 mooring points berths and about 3 km road infrastructures.²³ Besides, this port is situated close to the main square of the city, very near to the urban and commercial area of the city where many people live and work. The Italian Institute of Environmental Protection and Research estimated that the effect on air quality state caused by the environmental pollution from sea traffic had very increased in the last two decades. More specifically, the fraction of SO₂ pollutant emission due to the port activities of Naples is substantial (near 40%). For these reasons, it is very important to estimate the effect of the exhaust emissions from merchant ships on the human life in such a concentrated and densely populated area.

As already explained in the previous paragraphs, for port emissions, a vessel activity profile is a breakdown of a vessel’s movements into several operating modes, with a characteristic engine type and displacement, engine load factor, fuel type and time spent in each operating mode. In the bottom-up approach adopted in this study, global environmental pollution from sea traffic in the port of Naples is based on activities of individual ships linked with locations so that spatially resolved emissions are developed straight. Therefore, these estimations are based on individual vessel movements and the characteristics of each vessel. The positions of pollutant emission sources are determined by the positions of the predefined navigation routes, therefore, destinations of merchant ships were important to locate the points of berth in relation to the entry/exit of the port.

Analysis domains of the numerical simulations are 10 km × 10 km areas comprising the piers for large ships, the coastal area adjacent and the sea area involved by ship manoeuvring out and into of berths. The whole analysis domain is discretized with 32 × 32 elements square grid, while three main routes are considered for ships leaving and entering the port of Naples. As beforehand assumed and explained, a Gaussian dispersion model was used in the long-term version³⁹ to estimate the atmospheric dispersion of NO_X, SO_X and PM exhaust emissions and the spatial distribution of the pertinent pollutant concentrations along the whole coastal area. This model is valid for steady-state conditions, that is, for one or more fixed point sources that emit constantly, whereas the chimneys of the merchant ships are moving sources. Furthermore, the emission rate is rather inconstant during manoeuvring mode depending on power rate conditions. For these reasons, in order to apply the steady-state model to this specific contest, moving stacks are simulated by a specified number n_vps of virtual point sources, accounted as static emission sources along the three routes and producing the same total quantities of NO_X, SO_X and PM emitted by the merchant ships for year in the port of Naples. In the following simulations, the number n_vps of virtual point sources is equal to 60 (namely 20 for each of the three routes). In Figure 1, approaching routes and virtual point sources are shown on the analysis domain with green arrows and red circular spots, respectively.

Figure 1.

Analysis domain, main approaching routes, virtual point sources and sampling point on the coastal area of Naples.

Fixing large size two-stroke diesel engines as engine type j, residual (HSRFO) and distillate (LSDFO) fuel oil in switch conditions as fuel type z and manoeuvring as operating mode m, the mass flows of NO_X, SO_X and PM produced during the merchant ship’s inbound and outbound manoeuvring can be estimated starting from equation (4), thus obtaining equation (7), in which

Ė_i (g/h): flow rate of pollutant i of each virtual point source

n_s: total number of merchant ships crossing the port for year

n_vps: number of virtual point sources

{\dot{E}}_{i} = \frac{(P_{max} \cdot l f \cdot t \cdot e f_{i} \cdot L L A_{i}) \cdot n_{s}}{n_{v s p} \cdot 365 \cdot 24}

(7)

Considering traffic data acquired by the Maritime Authority of the port of Naples, the total number n_s of merchant ships crossing the port of Naples for year is around 10,500. The other parameters in the previous equation have been already explained in the description of equation (4). Manoeuvring the merchant ships, both on entry and exit from the port of Naples is characterized by rather variable power rates of the engine; characteristic distances between port borders and berth were used to calculate ship’s average in-port speed and consequently the engine load factor lf during the manoeuvring mode as calculated in equation (2). It is very important to estimate the engine load factor because the emissions factors are evidently connected with this factor, as explained and shown previously in Table 2. In equation (7), t is time spent during manoeuvring mode in the port, and it is accounted as the distance travelled between port entry/exit and berth point divided by ship’s average in-port speed. Emission factors EF_i for NO_X, PM and SO_X are taken from Table 2, whereas the low-load adjustment factors LLAi for PM and NO_X are calculated according to equation (5).

Therefore, spreading the total emissions of each pollutant along the routes, all the virtual point sources emit constantly with the emission rates calculated in equation (7). Once known, the mass flows of NO_X, SO_X and PM of each virtual point source, the stacks height and diameter, the temperature of the exhaust emissions and the application of the model provide the spatial distribution of the air pollutant concentrations above the whole analysis domain due to the arriving and sailing merchant ships in the port of Naples.

The concentration and atmospheric gas dispersion of such pollutants in the area under investigation were examined for two different scenarios on maps with receptors positioned at 2 m altitude:

The first scenario refers to gradual fuel switch from residual (HSRFO) to distillate (LSDFO) fuel oil during the whole approaching manoeuvre in the port area.

In the second scenario, the fuel changeover is achieved before the approaching manoeuvre, with the merchant ships burning only LSDFO fuel in the port area (best scenario).

The main results so obtained are shown in the maps of Figures 2 to 7; in these simulations, it is possible to assess, for the examined scenarios, the spatial distribution of SO_X, NO_X and PM in the form of iso-concentration levels and average concentrations of the same air pollutants, calculated by using equation (6) and expressed in µg/m³. However, the pollutant concentrations exposed in the following maps have to be accounted as average values in the year.

Figure 2.

Calculated spatial distribution of NO_x concentration at 2 m altitude – first scenario. NO_x: nitrogen oxides.

Figure 3.

Calculated spatial distribution of NO_x concentration at 2 m altitude – second scenario. NO_x: nitrogen oxides.

Figure 4.

Calculated spatial distribution of SO_x concentration at 2 m altitude – first scenario. SO_x: sulphur oxides.

Figure 5.

Calculated spatial distribution of SO_x concentration at 2 m altitude – second scenario. SO_x: sulphur oxides.

Figure 6.

Calculated spatial distribution of PM concentration at 2 m altitude – first scenario. PM: particulate matter.

Figure 7.

Calculated spatial distribution of PM concentration at 2 m altitude – second scenario. PM: particulate matter.

The NO_X, PM and SO_X pollutant concentrations calculated with this methodology are also compared in Table 4 with an experimental monitoring campaign performed in the port of Naples⁴⁰ to assess the air quality state in the adjacent area. The pollutant concentrations calculated in the previous numerical simulations are reported in Table 4 for each pollutant as the highest values in the port area resulting from merchant ships transiting in the port of Naples. On the other hand, the experimental continuous pollutant concentrations, reported in Table 4 as 1-h average concentrations, were obtained by a mobile laboratory situated in a fixed point in the port of Naples²³ and equipped with continuous gaseous analysers of NO₂, PM₁₀ and SO₂. NO₂ was measured using chemiluminescence detector, SO₂ was measured with fluorescence detector and PM₁₀ was detected using continuous analyzer based on orthogonal nephelometry. The exact localization of this mobile laboratory (sampling point) in the port area is shown in Figure 1. All these numerical and experimental results are also compared in Table 4 with limit values established by European legislation (EC, 2008), that have to considered as1-h or 24-h average concentrations.

Table 4.

Comparison between calculated, measured and limit values of NO₂, PM₁₀ and SO₂.

Pollutant	Highest calculated concentrations (first scenario)	Highest calculated concentrations (second scenario)	Experimental campaign²³	European Limit Values
NO₂ (µg/m³)	10.1	8.5	40.0	200 (1 h)
SO₂ (µg/m³)	1.8	1.2	6.8	350 (1 h)–125 (24 h)
PM₁₀ (µg/m³)	1.4	1.0	29.8	50 (24 h)

First of all, it can be observed that the measured levels of NO₂, PM₁₀ and SO₂, obtained as 1-h average concentration levels, were always below the respective European Limit Values. NO₂ and SO₂ concentration levels were well below the hourly and daily average European Limit Values.

Besides, preliminary comparisons of these results show that, in the port of Naples, only NO₂ and SO₂ concentration levels may be affected by merchant ship emissions, albeit without a crucial percentage contribution. This outcome indicates that merchant ship emissions cannot constitute the sole source of air pollution in the port of Naples; in fact, since estimated concentration levels are very lower than measured values, other anthropogenic activities adjacent to port area produce high concentration levels of SO₂, NO₂ and PM₁₀, so affecting the measured values. In fact, in order to correlate the effect of merchant ships on the air pollution in the port of Naples, other factors must be considered, in particular the manoeuvring of ships other than merchant ships; besides, the port of Naples is very close to the centre of the city, which is characterized by a dense network of road and motorways. Therefore, other sources of atmospheric pollution are located in this area, and between these, a considerable contribution on the air composition comes from the urban traffic roads and the heavy transportation activity, which can deliver a significant impact on the measured values. However, shipping emissions in the port of Naples could increase in the next future due to the increase of global-scale trade.

For all these reasons, in order to attain a consistent monitoring of the air quality in the area under investigation, other research should be involved. However, in order both to validate the computational results obtained in this study and to calculate the real dependence on the several pollution sources existing in the port of Naples, the next step of this research will consist of experimental measurements of pollutant concentrations off the coast, where the impact of other anthropogenic emissive sources can be considered almost insignificantly.

Finally, it must be stressed that the numerical results obtained in this study are characterized by a certain degree of approximation, due to both the exactness of the request input data (such as the ship emission levels and meteorological variables) and the accuracy of the dispersion model adopted in this research. The Gaussian model used in this investigation, in fact, is normally characterized by some approximations concerning the assessment of atmospheric gas dispersion under complex orography situations and for non-homogenous conditions of the atmosphere. Therefore, in order to correlate ship exhaust emissions to the actual pollutant concentration levels in atmosphere, depth knowledge of the atmospheric gas dispersion is necessary also in presence of obstacles and buildings and without flat topography. Anyway, since policy strategies and actions to reduce air pollutants from sea traffic are expected at the local and international level, they must be based on such robust analyses and estimations.

Conclusions

Marine transport sector contributes significantly to pollutant emissions in the coastal areas adjacent to ports. In order to project and implement actual rules to reduce environmental impacts of ship exhaust pollution in coastal areas, comprehensive data are necessary to evaluate the effect of these emissions on air quality degradation. In this paper, the case study of the port of Naples was investigated by estimating the production and atmospheric gas dispersion in the coastal area of SO_X, NO_X and PM associated to the traffic of merchant ships into and out of the harbour, considering both the fuel switch conditions from residual fuel oil (HSRFO) to distillate fuel oil (LSDFO) and the engine operating conditions at low load during the manoeuvring mode. Maritime transport in the port of Naples plays a central role in the assessment of the air quality state, but pollutant concentrations over coastal area are also affected by other emissive sources, above all emissions from the road transport sector. The methodology and all the results obtained in this research can be considered as an effective calculation tool for the national environmental management, since they can be used (together with the measured concentrations) to evaluate the agreement of air quality state with the legislative guidelines and limit values and the impact of new maritime activities in coastal areas on the pollutant concentrations.

Footnotes

Acknowledgements

We are grateful to the Port Authority of Naples for the data related to the maritime traffic and to MAN Diesel for information related to exhaust emissions of diesel engines.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Paolo Iodice received the M.S. degree cum laude in “Mechanical Engineering” and the Ph.D. degree in “Mechanical Systems Engineering” from the University of Naples Federico II, Italy, in 2002 and 2009, respectiely. He is currently a postdoctoral researcher in mechanical engineering with the Department of Industrial Engineering, University of Napoli Federico II. His research interests include: renewable fuels use, co/tri-generation power systems, green mobility, atmospheric dispersion models, emission inventory, motorcycle emissions, vehicle exhaust pollutants, experimental-analytical investigation on SI engines fuelled with gasoline-ethanol blended fuels. Paolo Iodice is author of over 50 scientific papers published in distinguished international journals and in conference proceedings.

Giuseppe Langella graduated cum laude in “Mechanical Engineering” in 1995 at University of Naples Federico II and he postgraduated in “Thermo-mechanical Systems Engineering” in 1999, at the same University. Since 2002, he is researcher at Department of Industrial Engineering of University of Naples Federico II. Since 2004, he is Assistant Professor of “Steam Generators and Thermal Plants” at the University of Naples Federico II. Since 2008, he is Assistant Professor of “Fluid Machines Management and Energy Systems” at Sannio University. Giuseppe Langella is author of several scientific papers about renewable fuels use, combustion processes with specific application of computer fluid dynamics and experimental investigations, and ORC plants.

Amedeo Amoresano took his degree in “Mechanical Engineering” at the University of Naples Federico II in 1991. In 1994 he took his PhD in “Thermo-mechanical and Energetic Systems”. In 1997 he became Researcher of the University of Naples Federico II and from 2015 he is Associate Professor of Fluid Machinery. He is also Associate Researcher at the Combustion Research Institute (IRC) of the National Council Research (CNR). He is advisor for the Italian Government on Innovative Energy Systems. During his career, Amedeo Amoresano tutored several graduated and PhD students and he is author of about sixty scientific papers published in distinguished international journals and in conference proceedings.

References

Goldsworthy

and Goldsworthy

Modelling of ship engine exhaust emissions in ports and extensive coastal waters based on terrestrial AIS data: an Australian case study. Environ Model Software 2015; 63: 45–60.

Viana

Hammingh

Colette

et al . Impact of maritime transport emissions on coastal air quality in Europe. Atmos Environ 2014; 90: 96–105.

Vinken

GCM

Boersma

van Donkelaar

et al . Constraints on ship NOx emissions in Europe using GEOS-Chem and OMI satellite NO₂ observations. Atmos Chem Phys 2014; 14: 1353–1369.

Fuglestvedt

Shine

Berntsen

et al . Transport impacts on atmosphere and climate: metrics. Atmos Environ 2010; 44: 4648–4677.

Bailey

and Solomon

Pollution prevention at ports: clearing the air. Environ Impact Assess Rev 2004; 24: 749–774.

Jalkanen

Johansson

Kukkonen

A comprehensive inventory of the ship traffic exhaust emissions in the Baltic Sea from 2006 to 2009. AMBIO 2014; 43: 311–324.

Eyring

Kohler

Lauer

et al . Emissions from international shipping: 2. impact of future technologies on scenarios until 2050. J Geophys Res 2005; 110: D17306.

Vutukuru

and Dabdub

Modeling the effects of ship emissions on coastal air quality: a case study of southern California. Atmos Environ 2008; 42: 3751–3764.

Dore

Vieno

Tang

et al . Modelling the atmospheric transport and deposition of sulphur and nitrogen over the United Kingdom and assessment of the influence of SO₂ emissions from international shipping. Atmos Environ 2007; 41: 2355–2367.

10.

Derwent

Stevenson

Doherty

et al . The contribution from shipping emissions to air quality and acid deposition in Europe. AMBIO 2005; 34: 54–59.

11.

Endresen

Sorgard

Sundet

et al . Emission from international sea transportation and environmental impact. J Geophys Res 2003; 108: 4560.

12.

Walsh

and Bows

Size matters: exploring the importance of vessel characteristics to inform estimates of shipping emissions. Appl Energ 2012; 98: 128–137.

13.

Agrawal

Malloy

QGJ

Welch

et al . In-use gaseous and particulate matter emissions from a modern ocean going container vessel. Atmos Environ 2008; 42: 5504–5510.

14.

Agrawal

Welch

Miller

et al . Emission measurements from a crude oil tanker at sea. Environ Sci Technol 2008; 42: 7098–7103.

15.

Corbett

and Koehler

HW.

Updated emissions from ocean shipping. J Geophys Res 2003; 108: 4650.

16.

Brook

Alfarra

et al . Identification and characterization of inland ship plumes over Vancouver, BC. Atmos Environ 2006; 40: 2767–2782.

17.

Streets

Guttikunda

and Carmichael

GR.

The growing contribution of sulfur emissions from ships in Asian Waters, 1998–1995. Atmos Environ 2000; 34: 4425–4439.

18.

Giakoumis

and Alafouzos

AI.

Study of diesel engine performance and emissions during a transient cycle applying an engine mapping-based methodology. Appl Energ 2010; 87: 1358–1365.

19.

Raptotasios

Sakellaridis

Papagiannakis

et al . Application of a multi-zone combustion model to investigate the NOx reduction potential of two-stroke marine diesel engines using EGR. Appl Energ 2015; 157: 814–823.

20.

Petzold

Hasselbach

Lauer

et al . Experimental studies on particle emissions from cruising ship, their characteristic properties, transformation and atmospheric lifetime in the marine boundary layer. Atmos Chem Phys 2008; 8: 2387–2403.

21.

Moldanová

Fridell

Popovicheva

et al . Characterisation of particulate matter and gaseous emissions from a large ship diesel engine. Atmos Environ 2009; 43: 2632–2641.

22.

Lai

Lee

et al . Persistent organic pollutants in tropical coastal and offshore environment: part A—atmospheric polycyclic aromatic hydrocarbons. Int J Environ Sci Technol 2015; 12: 1075–1086.

23.

Prati

Costagliola

Quaranta

et al . Assessment of ambient air quality in the port of Naples. J Air Waste Manag Assoc 2015; 65: 970–979.

24.

Tzannatos

Ship emissions and their externalities for the port of Piraeus – Greece. Atmos Environ 2010; 44: 400–407.

25.

Miola

and Ciuffo

Estimating air emissions from ships: meta-analysis of modelling approaches and available data sources. Atmos Environ 2011; 45: 2242–2251.

26.

Petzold

Weinzierl

Fiebig

et al . Particle emissions from ship engines: emission properties and transformation in the marine boundary layer. In: Proceedings of the TAC-Conference, Oxford, UK, 26–29 June 2006.

27.

Schembaria

Cavallia

Cucciab

et al . Impact of a European directive on ship emissions on air quality in Mediterranean harbours. Atmos Environ 2012; 61: 661–669.

28.

Caiazzo

Di Nardo

Langella

et al . Seawater scrubbing desulfurization: a model for SO₂ absorption in fall-down droplets. Environ Progr Sustain Energ 2012; 31: 277–287.

29.

Kasper

Aufdenblatten

Forss

et al . Particulate emissions from a low speed marine diesel engine. Aerosol Sci Tech 2007; 41: 24–32.

30.

Winnes

Fridell

Particle emissions from ships: dependence on fuel type. J Air Waste Manag Assoc 2009; 59: 1391–1398.

31.

Wang

Zhang

et al . Predication emission of a marine two stroke diesel engine based on modeling of radial basis function neural networks. In: Proceedings of the second WRI global congress on intelligent systems, Wuhan, China, 2010, IEEE Computer Society Washington: 184–188.

32.

MAN Diesel & Turbo. Operation on low sulphur fuels. Publication no. 5510-0075-01 paper, 2014, Denmark.

33.

DNV GL Maritime. Preparing for low sulphur operation. Technical update, https://www.dnvgl.com (2014).

34.

Jalkanen

Brink

Kalli

et al . A modelling system for the exhaust emissions of marine traffic and its application in the Baltic Sea area. Atmos Chem Phys 2009; 9: 9209–9223.

35.

Olesen

Winther

Christensen

et al . Ship contribution to air pollution in Denmark – an assessment utilising AIS data. In: Proceedings of the 13th conference on harmonisation within atmospheric dispersion modelling for regulatory purposes, Paris, France, 2010.

36.

Whall

Cooper

Archer

et al . Quantification of emissions from ships associated with ship movements between ports in the European community, prepared for the European Commission. Final Report, Entec UK Limited, Chesire, UK, July 2002.

37.

Cooper

Exhaust emissions from high speed passenger ferries. Atmos Environ 2001; 35: 4189–4200.

38.

Abdul-Wahab

SA.

An application and evaluation of the CAL3QHC model for predicting carbon monoxide concentrations from motor vehicles near a roadway intersection in Muscat, Oman. Environ Manag 2004; 34: 372–382.

39.

United States Environmental Protection Agency. User’s guide for the ISC3 DISPERSION MODELS. Epa-454/B-95-003b, US EPA OAQPS, September 1995. Research Triangle Park, NC: United States Environmental Protection Agency.

40.

Murena

Prati

Quaranta

Monitoring the quality of air in the port of Naples. In: Developments in maritime transportation and exploitation of sea resources. 2014; Taylor & Francis Group, London: 835–839.