Shipping contributes significantly to air pollutant emissions and
atmospheric particulate matter (PM) concentrations. At the same time,
worldwide maritime transport volumes are expected to continue to rise in the
future. The Mediterranean Sea is a major short-sea shipping route within
Europe and is the main shipping route between Europe and East Asia. As
a result, it is a heavily trafficked shipping area, and air quality
monitoring stations in numerous cities along the Mediterranean coast have
detected high levels of air pollutants originating from shipping emissions.
The current study is a part of the EU Horizon 2020 project SCIPPER (Shipping
Contributions to Inland Pollution – Push for the Enforcement of Regulations),
which intends to investigate how existing restrictions on shipping-related
emissions to the atmosphere ensure compliance with legislation. To
demonstrate the impact of ships on relatively large scales, the potential
shipping impacts on various air pollutants can be simulated with chemical
transport models.
To determine the formation, transport, chemical transformation, and fate of
particulate matter < 2.5 µm (PM2.5) in the Mediterranean Sea in 2015, five different regional
chemical transport models (CAMx – Comprehensive Air Quality Model with
Extensions, CHIMERE, CMAQ – Community Multiscale Air Quality model, EMEP –
European Monitoring and Evaluation Programme model, and LOTOS-EUROS) were
applied. Furthermore, PM2.5 precursors (ammonia (NH3), sulfur dioxide (SO2), nitric acid (HNO3))
and inorganic particle species (sulfate (SO42-), ammonia (NH4+), nitrate (NO3-)) were studied, as they are important for explaining
differences among the models.
STEAM (see “List of abbreviations” in Appendix A) version 3.3.0 was used to compute
shipping emissions, and the CAMS-REG version 2.2.1 dataset was used to calculate
land-based emissions for an area encompassing the Mediterranean Sea at a
resolution of 12 × 12 km2 (or 0.1∘× 0.1∘). For additional input, like meteorological fields and
boundary conditions, all models utilized their regular configuration. The
zero-out approach was used to quantify the potential impact of ship
emissions on PM2.5 concentrations. The model results were compared with
observed background data from monitoring sites.
Four of the five models underestimated the actual measured PM2.5
concentrations. These underestimations are linked to model-specific
mechanisms or underpredictions of particle precursors. The potential impact
of ships on the PM2.5 concentration is between 15 % and 25 % at
the main shipping routes. Regarding particle species, SO42- is
the main contributor to the absolute ship-related PM2.5 and to total
PM2.5 concentrations. In the ship-related PM2.5, a higher share of
inorganic particle species can be found when compared with the total
PM2.5. The seasonal variabilities in particle species show that
NO3- is higher in winter and spring, while the NH4+ concentrations displayed no clear seasonal pattern in any models. In most
cases with high concentrations of both NH4+ and
NO3-, lower SO42- concentrations are simulated.
Differences among the simulated particle species distributions might be
traced back to the aerosol size distribution and how models distribute
emissions between the coarse and fine modes (PM2.5 and PM10). The
seasonality of wet deposition follows the seasonality of the precipitation,
showing that precipitation predominates wet deposition.
Horizon 2020814893Introduction
Exhaust particles emitted from shipping have a large share in total
emissions from the transport sector (Corbett and Fischbeck, 1997; Eyring et
al., 2005), thereby affecting the chemical composition of the atmosphere as
well as the regional air quality. Particularly in coastal areas, maritime
transport contributes a considerable fraction to air pollution (Viana et
al., 2014).
High particulate matter < 2.5 µm (PM2.5) concentrations can be caused by transported particles,
desert dust, or the production of secondary particulate matter (Tomasi and Lupi, 2017). Previous studies have revealed that in Europe the PM2.5 concentration increase caused
by shipping emissions is small (Viana et al., 2009; Aksoyoglu et al., 2016).
Nevertheless, in the Mediterranean region the relative ship impact on the
PM2.5 concentration is large, with a share of 5 % to 20 % of
the total PM2.5 concentration (e.g., Aksoyoglu et al., 2016; Nunes et
al., 2020). The formation of secondary particulate matter from
ship emissions is of particular importance. According to Viana et al. (2009), the
secondary contribution of ship emissions is equivalent to double their
primary contribution. Secondary particles in the atmosphere form from
gaseous precursors, whereas primary particles are directly emitted and
evolve within a short time to form secondary particles. To improve the air
quality in coastal regions, it is important to identify the pollutant
sources and make reliable estimations of their impacts on surrounding PM
levels. It has been shown that the majority of secondary particles
contributing to local PM in ports come from shipping (Song and Shon, 2014).
Furthermore, according to Klimont et al. (2017), the proportion of
international shipping's particulate matter primary emissions to global
anthropogenic emissions is between 3 % and 4 %, which is comparable to
road traffic. Additionally, shipping contributions to total PM2.5 concentrations far from coastlines were found to be responsible for
exceedances of the WHO air quality guideline values (Nunes et al., 2020). The
annual mean PM2.5 limit value in the EU is 25 µg m-3 (EU DIRECTIVE 2008/50/EC, 2008), whereas the annual
mean PM2.5 goal established by the WHO is 5.0 µg m-3 (WHO, 2021). Strong evidence has been found for the relationship between
exposure to PM2.5 and the occurrences of certain diseases affecting the
lungs, cancer, or type 2 diabetes (Heusinkveld et al., 2016; Chen et al.,
2016; Gao and Sang, 2020). According to the WHO, there is no safe level of
PM2.5; thus, the gap between the WHO and EU's PM2.5 values is of
real concern (Karamfilova, 2022).
The MEPC decided in December 2022 to establish a sulfur emission control
area in the Mediterranean Sea by 1 January 2025. In this area, the
limit for sulfur in fuel oils used on board ships is 0.10 % (IMO, 2022).
The global sulfur cap for marine vessels came into effect in January 2020,
which declares that the sulfur content of any fuel oil used from ships must
not exceed 0.50 % m m-1, except for ships using “equivalent” compliance
mechanisms, such as scrubbers. Calculations show that this policy has led to
PM2.5 reductions ranging from 0.5 µg m-3 to
more than 2.0 µg m-3 along the major shipping routes
in the Mediterranean Sea (Jonson et al., 2020). These relatively strict 2020
regulations are expected to lower the number of PM2.5-related premature
deaths by on average 15 % (Viana et al., 2020).
Although the Mediterranean Sea contains one of the busiest shipping routes
worldwide, only a few regional-scale chemical transport modeling studies
have considered this region. Viana et al. (2014) reviewed studies concerning
the impacts of shipping emissions on air quality in European coastal areas,
noting that the highest PM2.5 contributions were found in the
Mediterranean Sea and North Sea. Aksoyoglu et al. (2016) studied PM2.5 concentrations in the Mediterranean Sea followed by a comparison of two
models. Marmer and Langmann (2005) investigated the Mediterranean Sea on a
broader scale and without comparing different CTM systems. Nevertheless,
other studies have concentrated on smaller domains, such as the Iberian
Peninsula (Baldasano et al., 2011; Nunes et al., 2020), the eastern
Mediterranean Sea with the Arabian Peninsula (Večeřa et al., 2008;
Tadic et al., 2020; Celik et al., 2020; Friedrich et al., 2021), or the urban scale and harbor cities (Schembari et al., 2012; Donateo et al., 2014; Prati
et al., 2015). None of these studies, however, analyzed the potential
shipping impacts on PM2.5 concentrations together with individual
aerosol species on a regional basis while additionally comparing the results of
five CTMs.
A wide range of gaseous pollutants, such as sulfur dioxide (SO2) and
nitrogen oxide (NOx= NO + NO2), coming from shipping
emissions can be precursors for particle formation (Jägerbrand et al.,
2019; Karl et al., 2019; Matthias et al., 2010). Sulfur dioxide is released
mainly by human activities such as fossil fuel burning, petroleum refining,
and metal smelting (Zhong et al., 2020). SO2 is oxidized by dissolved
oxidants such as ozone (O3) and hydrogen peroxide (H2O2) in the
aqueous phase and by hydroxyl (OH) in the gas phase to generate sulfuric
acid (H2SO4)
(Seinfeld and Pandis, 2006). H2SO4 and nitric acid (HNO3) react with ammonia
(NH3) to form ammonium sulfate ((NH4)2SO4) and
NH4NO3 aerosols, with H2SO4 neutralization having
preference due to its lower vapor pressure (Hauglustaine et al., 2014).
Nitrogen oxides are primarily removed during the day via the OH
radical oxidation reaction to produce (HNO3) (Seinfeld and
Pandis, 1998). At night, the main NOx removal method involves
interacting with O3 to produce the nitrate (NO3) radical,
which may then combine with nitrogen dioxide (NO2) to form dinitrogen
pentoxide (N2O5) and may subsequently undergo a heterogeneous
reaction with water to produce HNO3. As it is highly soluble, HNO3
disperses quickly in water droplets or is neutralized by reaction with
NH3 to produce NH4NO3 aerosols. Increased emissions
of NH3 or HNO3 formation and their deposition negatively
affect the environment through eutrophication and acidification, thereby
contributing to the loss of ecosystem biodiversity (Remke et al., 2009;
Kleijn et al., 2009; Krupa, 2003).
Furthermore, the air pollution status should be assessed to investigate the
consequences of new legislation.
The current work investigates and analyzes the predictions of five different
CTMs for air pollutant dispersion and transformation. The intercomparison
was carried out in two parts: part one included the photochemistry and
differences among the models regarding NO2 and O3 (Fink et al., 2023). The
present study is part two of the model intercomparison and evaluates the
same CTM simulations but different air pollutants, namely aerosols. This
paper is structured as follows: Sect. 3.1 and 3.2 consider the simulated
overall PM2.5 model performance and spatial distribution. In Sect. 3.3,
precursors (NH3, HNO3, SO2, and NO2) are investigated as
the basis for inorganic particle species. Inorganic aerosol concentration and
wet deposition are regarded in Sect. 3.4.
To date, the present study is the first multimodel study designed to compare
the potential impacts of shipping on PM2.5 and particle species
simulated by five regional-scale CTMs for the Mediterranean Sea.
Materials and methods
In this section the models participating in the intercomparison study are
briefly described. More detailed information about the standard setup of
the models and model internal mechanisms used in the present study can be found
in part one of this intercomparison study (Fink et al., 2023), which focuses
on nitrogen oxides and ozone.
Models
In this study, five different regional-scale CTM systems run by four
institutions participated: CAMx and CHIMERE run by AtmoSud, CMAQ run by
Helmholtz-Zentrum Hereon, EMEP run by the IVL Swedish Environmental Research
Institute, and LOTOS-EUROS run by the TNO Netherlands Organization for Applied
Scientific Research. In order to produce comparable results with respect to the impact of
shipping emissions on PM2.5 concentrations, the models were set up in
a similar way. The same shipping emissions data from STEAM (version 3.3.0.;
Jalkanen et al., 2009, 2012; Johansson et al., 2013, 2017) were used for all CTMs. Land-based emissions
(CAMS-REG, v2.0), grid projection (WGS84_longlat), domain
(Mediterranean Sea), grid resolution (0.1∘× 0.1∘, 12 × 12 km), and the modeled year (2015) were also
consistent (Table 1). The CTM systems were applied in their standard setup
for other input data; i.e., the meteorological input data and the boundary
and initial conditions differed.
The model domains covered the largest part of the Mediterranean Sea, with a
spatial extent ranging in latitude from 33.8 to 44.95∘ and in longitude from -0.95 to
29.95∘
(Appendix A). The appointed grid cell size was 12 × 12 km2 interpolated on a 0.1∘× 0.1∘ grid nested in a 36 × 36 km2 grid
(except EMEP).
A reference run for present air quality conditions was performed using all
models, including all emissions (base case). Furthermore, all models ran
once without shipping emissions (no-ship case). The difference between the
estimates with all emissions and the calculations without shipping emissions
was then used to calculate the potential impact of ships on pollutant
concentrations (zero-out method).
From the results of all models, the annual averaged ensemble mean was
calculated based on the daily files. The model run outputs all contained
PM2.5 in micrograms per cubic meter (µg m-3) at a daily resolution on a 2-D
grid from the lowest layer and provides this as a netcdf file following CF
conventions. Concentrations in the lowest layer close to ground were used for
the intercomparison. The CTM systems calculated PM2.5 concentrations
in different ways depending on the major physical and chemical mechanisms
implemented. Table 1 summarizes the model setups.
The models used in the intercomparison are listed as follows:
CAMx v6.50 (Ramboll Environment and Health, 2020),
CHIMERE 2017r4 (Menut et al., 2013),
CMAQ v5.2 (Byun and Schere, 2006; Appel et al., 2017),
EMEP MSC-W (Simpson et al., 2012, 2020), and
LOTOS-EUROS v2.0 (Manders et al., 2017).
Detailed descriptions of the models used can be found in the first part of
the intercomparison study (Fink et al., 2023).
Main model parameters and input data for the five chemical
transport models.
Model parameterCAMxCHIMERECMAQEMEPLOTOS-EUROSGrid resolution inner domain12 × 12 km212 × 12 km212 × 12 km20.1∘× 0.1∘0.1∘× 0.1∘Grid resolution outer domain36 × 36 km236 × 36 km236 × 36 km2None0.5∘× 0.25∘Meteorological driverWPS and WRFWPS and WRFCOSMO-CLM v5.0ECMWF (IFS)ECWMF (IFS)Boundary conditionsMOZART-4 output is used and downscaled for time- and space-variable boundary conditionsGaseous species: LMDz-INCA model (Folberth et al., 2006), with climatology as average monthly fields Aerosols: Global Ozone Chemistry Aerosol Radiation and Transport (GOCART) model (Ginoux et al., 2001)IFS_CAMS cycle45r1Provided with the open-source model distribution for the year 2015; simple functions for prescribing concentrations in terms of latitude and the time of the year or time of the day (Simpson et al., 2012). Boundary conditions of ozone are developed from climatological ozone-sonde datasets as in the EMEP Status Report 1/2022 (2022).CAMS C-IFS global forecast (lateral and top)Land-based emissionsCAMS-REG v2.2.1CAMS-REG v2.2.1CAMS-REG v2.2.1CAMS-REG v2.2.1CAMS-REG v2.2.1Shipping emissionsSTEAM v3.3.0STEAM v3.3.0STEAM v3.3.0STEAM v3.3.0STEAM v3.3.0Biogenic emissionsMEGAN model v2.03 output for the year 2015MEGAN model v2.04 output for the year 2015MEGAN model v3 output for the year 2015Calculated online: emissions of isoprene and monoterpenes are based on Guenther et al. (1993, 1995). Soil NO emissions from soils of seminatural ecosystems are specified as a function of the N deposition and temperature.Calculated online: emissions of isoprene and monoterpenes are based on Guenther et al. (1993), using actual meteorological data. Emission of NO from soil is based on Manders-Groot et al. (2016).Sea salt emissionsCalculation based on Ovadnevaite et al. (2014)Calculation based on Monahan et al. (1986)Calculation based on Kelly et al. (2010)Calculation based on Monahan et al. (1986) and Mårtensson et al. (2003)Calculation based on Monahan et al. (1986) and Mårtensson et al. (2003)Dust emissionsBased on the approach used in global EMAC (ECHAM/MESSy; Klingmüller et al., 2018; Astitha et al., 2012)Calculated online: after parametrization of Marticorena and Bergametti (1995) and Alfaro and Gomes (2001)Not consideredThe key parameter is wind friction velocity. The parameterization is done as in Marticorena and Bergametti (1995), Marticorena et al. (1997), Alfaro and Gomes (2001), Gomes et al. (2003), and Zender et al. (2003). Daily emissions are from forest and vegetation fires from the Fire INventory from NCAR version 1.0 (Wiedinmyer et al., 2011).Calculated online: emissions used are based on Marticorena and Bergametti (1995) with soil moisture as described by Fécan et al. (1999). Dust from re-suspension by traffic and agriculture follows Schaap et al. (2009).Chemical mechanismCB05MELCHIOR2CB05EmChem 19aCBM-IVAerosol size distributionPM2.5; PM10Eight bins: 40 nm to 10 µmTrimodal size distribution (0.03, 0.3, and 6 µm; Binkowski and Roselle, 2003)PM2.5; PM2.5–10PM2.5; PM2.5–10Inorganic aerosol moduleISORROPIA (Nenes et al., 1998)ISORROPIA (Nenes et al., 1998)ISORROPIA II (Fountoukis and Nenes, 2007)MARS (Binkowski and Shankar, 1995)ISORROPIA II (Fountoukis and Nenes, 2007)Organic aerosol moduleSOAP semivolatile scheme (Strader et al., 1999)Described in Pun et al. (2006)Updates on SOA as described in Pye et al. (2017)For SOA the volatility basis set (VBS) approach is used (Robinson et al., 2007; Donahue et al., 2009; Bergström et al., 2012).There are no organic aerosols in the simulations.Wet-deposition schemeScavenging model for gases and aerosols (Seinfeld and Pandis, 1998)Wet deposition in CHIMERE follows the scheme proposed by Loosmore and Cederwall (2004).Wet deposition is calculated within CMAQ's cloud module as described by Roselle and Binkowsk (1999).Calculation is done as described in Emberson et al. (2000); parametrization for different surfaces is done as in Simpson et al. (2012).Wet deposition is divided between in-cloud and below-cloud scavenging. The in-cloud scavenging module is based on the approach described in Seinfeld and Pandis (2006) and Banzhaf et al. (2012).Dry-deposition schemeThe resistance model of Zhang et al. (2003) is used.Dry deposition is used following Wesely (1989).The dry-deposition scheme M3Dry (Pleim, 2001) is used.As described in Simpson et al. (2012)The resistance approach follows Erisman et al. (1994).Aerosol modules
CAMx includes algorithms for inorganic aqueous chemistry (RADM–AQ),
inorganic gas–aerosol partitioning (ISORROPIA), and two organic gas–aerosol
partitioning and oxidation approaches (VBS or SOAP). Using gas-phase
processes, these approaches produce sulfate, nitrate, and condensable
organic gases. The hybrid 1.5-D VBS is applied to provide a unified
framework for gas–aerosol partitioning and the chemical aging of both
primary and secondary atmospheric organic aerosols (Ramboll Environment and
Health, 2020). One crucial assumption in PSAT is that PM is allocated to the
primary precursor for each type of particulate matter (i.e., PSO4 is
apportioned to SOx emissions, PNO3 is apportioned to NOx
emissions, and PNH4 is apportioned to NH3 emissions).
A detailed description of CHIMERE's inorganic and organic modules can be found
in Menut et al. (2013). CHIMERE's sectional aerosol module includes emitted
TPPM and secondary species such as nitrate, sulfate, ammonium, and SOAs.
Natural dust and sea salt aerosols can also be produced as passive tracers
or interactive species in equilibrium with other ions. Organic matter and
elemental carbon can be speciated if an inventory of their emissions is
supplied. The utilized models include the aqueous, gaseous, and particulate
phases of ammonia, ammonium, nitrate, and sulfate. For instance, in
accordance with the ISORROPIA thermodynamic equilibrium model, the model
species pNH3 represents an equivalent ammonium in the particulate phase
as the sum of the NH4+ ion, NH3 liquid, NH4NO3
solid, and other salts (Nenes et al., 1998).
CMAQ represents aerosol formation and growth using three log-normal-distributed modes: the Aitken and accumulation modes are generally less than
2.5 µm in diameter, while the coarse mode contains significant amounts
of mass above 2.5 µm. PM2.5 and PM10 can be obtained from the
model-predicted mass concentration and size distribution information.
The CMAQ aerosol scheme AERO6 was employed; this scheme expands the chemical
speciation of PM by the species Al, Ca, Fe, Si, Ti, Mg, K, and Mn. Sulfuric
acid (H2SO4), nitric acid (HNO3), hydrochloric acid (HCl), and
ammonia (NH3) gas-phase–aerosol partition equilibria are solved by
the ISORROPIA II mechanism (Fountoukis and Nenes, 2007; Nenes et al., 1998).
Contained within this scheme is the formation of SOA from isoprene,
terpenes, benzene, toluene, xylene, and alkanes (Carlton et al., 2010; Pye
and Pouliot, 2012). CMAQ allows for dynamic mass transfer of semivolatile
inorganic gases to coarse-mode particles, which facilitates the replacement
of chloride by NO3- in sea salt aerosols (Foley et al., 2010).
The EMEP MSC-W model version used was rv4.34 with chemical mechanism EmChem
19a (Simpson et al., 2012; Simpson et al., 2020). The mechanism builds on
surrogate VOC species (as in Simpson et al., 2012, but extended with benzene
and toluene) and has 171 gas-phase and heterogeneous reactions. The model
always assumes equilibrium between the gas and aerosol phases using the MARS
equilibrium module of Binkowski and Shankar (1995). For SOAs a VBS approach
is used (Robinson et al., 2007; Donahue et al., 2009; Bergström et al.,
2012). The semivolatile ASOA and BSOA species are considered to oxidize
(age) in the atmosphere via OH reactions, whereas all POA emissions are
treated as nonvolatile to maintain the emission totals of both the PM and
the VOC components from the official emission inventories (Simpson et al.,
2012). The aerosol module of the EMEP model distinguishes five classes of
fine and coarse particles (fine-mode nitrate and ammonium, other fine-mode
particles, coarse nitrate, coarse sea salt, and coarse dust); for
dry-deposition purposes, these particles are assigned mass median diameters
(Dp), geometric standard deviations (σg), and densities
(ρp). The aerosol components that are taken into account include
sea salt, SO42-, NO3-, NH4+, and anthropogenic
main PM. Aerosol water is also considered.
LOTOS-EUROS uses the TNO CBM-IV scheme, which is a modified version of the
original CBM-IV scheme (Whitten et al., 1980). N2O5 hydrolysis is
described explicitly based on the available (wet) aerosol surface area
(Schaap et al., 2004). The aqueous phase and heterogeneous formation of
sulfate are described by a simple first-order reaction constant (Schaap et
al., 2004; Barbu et al., 2009). Aerosol chemistry is represented using
ISORROPIA II (Fountoukis and Nenes, 2007).
Wet-deposition mechanisms
Wet deposition is the predominant removal process for fine particles. The
CAMx wet-deposition model uses a scavenging method in which the local
concentration change rate inside or under a precipitating cloud is
determined by a scavenging coefficient. From the top of the precipitation
profile to the surface, wet scavenging is estimated for each layer inside a
precipitating grid column. The scavenging coefficients of gases and PM are
calculated differently depending on the correlations given by Seinfeld and
Pandis (2006) (Ramboll Environment and Health, 2020). The wet-deposition
process in CHIMERE follows the scheme proposed by Loosmore and Cederwall (2004). In CMAQ, wet deposition is calculated in cloud chemistry treatments.
The resolved cloud model calculates the contribution of each model layer to
the precipitation. Based on a normalized profile of precipitating
hydrometeors, CMAQ operates a simple algorithm to assign precipitation
amounts to individual layers (Foley et al., 2010). The EMEP model's
parameterization of wet-deposition processes covers both the in-cloud and
the sub-cloud scavenging of gases and particles. The parameterization of wet
deposition is described in Berge and Jakobsen (1998). There are two types of
wet deposition in LOTOS-EUROS: below-cloud scavenging and in-cloud
scavenging. The technique is described in Seinfeld and Pandis (2006), and
Banzhaf et al. (2012) served as the foundation for the in-cloud scavenging
module.
EmissionsLand-based emissions
All five models used anthropogenic land-based gridded emissions from the
CAMS-REG v2.2 emission inventory for 2015, which is described in Granier et
al. (2019) and is essentially a further development of the earlier
TNO_MACC inventories (Kuenen et al., 2014). A more recent
version, CAMS-REG-v4.2, is described in detail in Kuenen et al. (2022).
For each country, the gridded emission files included GNFR emission sectors
for the air pollutants NOx, SO2, NMVOC, NH3, CO, PM10,
PM2.5, and CH4. The spatial resolution of the emissions data was
1/10∘×1/20∘ in longitude and latitude (i.e.,
∼6× 6 km over central Europe). The CAMS-REG
inventory also provides default information in order to apply the emissions in the
CTMs. The height distribution of emissions per GNFR sector was prepared
according to Bieser et al. (2011). Based on the assignment of PM and NMVOC
components at a detailed subsector level, PM and NMVOC speciation profiles
are provided for each country, year, and GNFR sector. The temporal
distribution of emissions is based on the default temporal variation
provided along with the CAMS-REG inventory. The NOx splitting was performed
according to Manders-Groot et al. (2016).
Shipping emissions
The shipping emission dataset produced with the STEAM model has a spatial
resolution of 12 × 12 km2 and a temporal resolution
of 1 h. The STEAM v3.3.0 emissions are divided into two vertical layers
(0 to 36 m; 36 to 1000 m) and are provided for mineral ash, carbon
monoxide (CO), carbon dioxide (CO2), elemental carbon (EC), NOx,
organic carbon (OC), PM2.5, particle number count (PNC), sulfate
(SO4), SOx (containing SO2 and SO3), and VOCs. To reduce
the number of generated emission maps and the computational resources needed
to run the STEAM model, VOC emissions were divided into four categories
according to their properties as a function of the engine load. Emission
factors for VOCs are based on the average values taken from various
publications (Agrawal et al., 2008, 2010; Sippula et al.,
2014; Reichle et al., 2015).
All shipping emissions are included in the lowest layer of CAMx. In CAMx,
all gridded emissions are at the ground level except punctual and linear
emissions. For CHIMERE, 88 % of the emissions below 36 m and all shipping
emissions above 36 m were added to the second layer. Only 12 % of the
emissions below 36 m were allocated to the model's lowest layer. The STEAM
emission dataset, which included stack heights, was used for this procedure.
In CMAQ, shipping emissions were split between the two lowest levels; those
below 36 m were ascribed to the lowest layer, while those above 36 m were
positioned in the second layer. The heights of the lowest and the second
layer in CMAQ are 42 m for each. The STEAM emissions were summed from hourly
to daily emissions and were attributed to the lowest layer (up to 90 m) in the
EMEP simulations. In LOTOS-EUROS, emissions below 36 m were divided into two
layers: the first layer was 25 m thick (∼ 70 % of
emissions), and the second layer was 30 m thick (∼ 30 % of
emissions). Over 36 m, emissions were separated into various height groups:
30 % were between 36 and 90 m, 30 % were between 170 and 90 m, 30 % were between 170 and 310 m, and 10 % were between 310 and 470 m.
These emissions were placed in the second or third model layers because of
the dynamic second model layer, which follows the meteorological boundary
layer. All emissions were placed in this second layer when the
meteorological boundary layer was well mixed and vertically extended (higher
than 470 m), while some emissions were placed in the third layer when the
boundary layer was shallow.
Observational data, statistical analysis, and analysis of model results
The model findings regarding the total surface PM2.5 concentrations
from the five CTM systems were compared with data from the air quality
monitoring network obtained from the EEA's download service (https://discomap.eea.europa.eu/map/fme/AirQualityExport.htm, last access: 6 September 2023). The
locations of the measurement stations are shown in Fig. A1, and detailed
information on the stations can be found in Appendix B.
The stations were chosen based on the following criteria: (i) the station
type was “background”; (ii) the station elevation was less than 1000 m;
and (iii) the station recorded data for more than one of the following
pollutants – NO2, O3, or PM2.5. In the first part of this
intercomparison study (Fink et al., 2023), NO2 and O3 were
discussed. Since simulating the potential impact of ships was the main focus
of this study, stations near the sea were the preferable choice.
The model findings regarding the total surface PM2.5 concentrations
from the five CTM systems were compared with existing observations. The RMSE,
NMB, and correlation coefficient R were determined for each monitoring
station to quantify the model performance, as described in the previous
study (Fink et al., 2023).
A categorization scheme for the correlations was established as described in
Schober et al. (2018), with weak (0.00–0.39), moderate (0.40–0.69), and
strong (0.70–1.00) correlations.
To compare the predicted daily mean concentrations with the measurements
recorded at representative sites, time series were employed. In addition,
based on hourly data, the yearly mean potential ship impact was determined.
Boxplots based on yearly values obtained from hourly data at each station
were used to graphically compare the model performances using the R, NMB,
and RMSE metrics. Annual mean values based on hourly data were utilized for
the intercomparison maps. Based on hourly data, the correlations between
models were determined for each grid cell.
ResultsPM<inline-formula><mml:math id="M210" display="inline"><mml:msub><mml:mi/><mml:mn mathvariant="normal">2.5</mml:mn></mml:msub></mml:math></inline-formula> model performance
Regarding the model performance, time series can give an overview of the
performance throughout the whole year. Figure 1 displays the average values
at all 28 measurement stations. CAMx, CMAQ, EMEP, and LOTOS-EUROS
underestimate the actual measured data. The largest underestimations are
found for CMAQ (NMB =-0.42) and LOTOS-EUROS (NMB =-0.54). Contrary to
the other CTM systems, CMAQ does not consider the contribution of dust, which can
cause underestimations of PM2.5. However, the correlations between the
modeled and measured data are strongest for these models (CMAQ: R= 0.50,
LOTOS-EUROS: R= 0.54; Table 2). No correlation can be found between the
measured and modeled data for CHIMERE (R= 0.02); on the other hand
CHIMERE displays only a slight overestimation of the actual data (NMB = 0.06). The simulated potential impacts of ships at all measurement stations
are between 5.7 % (CMAQ) and 13.8 % (CAMx; Table 2) as annual averages.
The simulated ship impacts on PM2.5 concentrations are within the ranges
stated in other studies. In a review of studies regarding the impact of
shipping emissions on coastal regions, Viana et al. (2014) reported
PM2.5 impacts of shipping to be between 5 % and 14 %. Aksoyoglu et
al. (2016) found PM2.5 concentrations between 10 % and 15 % along
coastal areas due to ship traffic. Ship impacts of approximately 20 % in
the southern coastal region of the Iberian Peninsula were found by Nunes et
al. (2020). Although the models underestimated the
actual measured total PM2.5 concentrations in this study, they slightly overestimated
the relative potential impact of ships on PM2.5 compared with previous
measurement studies. Donateo et al. (2014) measured a proportion of 7.4 %
of ships to the total PM2.5; Pandolfi et al. (2011) measured a
proportion of shipping to PM2.5 concentrations in the Bahía de Algeciras of between 5 % and 10 %. Agrawal et al. (2009) monitored
PM2.5 at the harbor of Los Angeles and found PM2.5 contributions from ships of up to 8.8 %. Predominating secondary particles
in PM2.5 for potential ship impact in the present study can explain the
deviations to the measurement studies.
Time series with daily mean PM2.5 concentrations in 2015,
averaged for all stations, and the respective grid cells of the models for (a) CAMx, (b) CHIMERE, (c) CMAQ, (d) EMEP, and (e) LOTOS-EUROS.
The dashed gray lines indicate measured data, the colored lines indicate modeled data, and the solid gray
lines indicate the modeled potential ship impacts.
The RMSE is very similar for all models with a value between 10.7
and 12.2 µg m-3. However, the
RMSE is strongly determined by high concentrations and can be biased by
outliers. This might explain the similar RMSE derived from CHIMERE despite
the lack of correlation. The mean RMSE from different models for PM2.5
in Europe found in the AQMEII intercomparison study by Im et al. (2015) was
6.19 for rural stations and 10.26 for urban stations and is similar to the
RMSE calculated in the present study.
The underestimation of PM2.5 concentrations by four out of five
models is consistent with results by Im et al. (2015), who reported an
underestimation of particulate matter for all participating models, with
the largest underestimations observed in the Mediterranean region. They stated
that the representation of dust and sea salt emissions had a large impact on
the simulated PM concentrations and that uncertainties remain when trying to
identify the reasons for the model bias (Im et al., 2015). Additionally, in
a study by Gašparac et al. (2020), underestimations were also found when
using EMEP and WRF-Chem to model PM2.5 at rural stations in Europe.
Solazzo et al. (2012) performed an operational model evaluation for 10
models and found that the models underestimated the monthly mean PM2.5 surface concentrations in Europe in most cases.
Correlation (R), normalized mean bias (NMB), root mean square error
(RMSE), and observational (obs) and modeled (mod) mean PM2.5 values for
2015 over all 28 stations. The observed mean value for all stations is 14.6 µg m-3.
CorrelationNMBRMSEModAbsolute potential ship impactRelative potential ship impact(µg m-3)(µg m-3)(annual mean average at all(annual mean average at allstations) in µg m-3stations) in %CAMx0.19-0.3311.58.91.213.8CHIMERE0.020.0611.114.31.813.2CMAQ0.50-0.4210.78.30.55.7EMEP0.17-0.3312.28.90.99.1LOTOS-EUROS0.54-0.5310.96.80.69.5PM<inline-formula><mml:math id="M251" display="inline"><mml:msub><mml:mi/><mml:mn mathvariant="normal">2.5</mml:mn></mml:msub></mml:math></inline-formula> spatial distribution
The highest PM2.5 values are simulated by all five models in northern
Italy, the Balkan Peninsula, and northern Africa (Fig. 2). The PM2.5
annual mean concentration results show that CHIMERE has the highest annual
mean values of 13 to 15 µg m-3 for the eastern part of the domain and over water,
whereas LOTOS-EUROS displays the lowest values with 2.0 to 4.0 µg m-3 in most regions
(Fig. 2). CMAQ, CAMx, and EMEP show similar model PM2.5 outputs with
diverse values distributed between 2.0 and 11 µg m-3 over the domain. The ensemble mean value over
the whole domain is 8.6 µg m-3 (Fig. 8a). All five
models display high PM2.5 concentrations of > 15 µg m-3 in the Po Valley. In this area, Kiesewetter et al. (2015) and Clappier et al. (2021) also simulated high values between 20 and 45 µg m-3 for 2015.
As demonstrated in Table 3, the correlation between the base-run model
results with all emissions is the strongest between EMEP and CMAQ (R= 0.59)
and CAMx and CMAQ (R= 0.42). In Fink et al. (2023), a high correlation was
found between CAMx- and CHIMERE-simulated NO2 and O3 concentrations
because both models used the same meteorology. Nevertheless, the present
study reveals that particle chemistry causes results that differ more due to a
higher complexity in the calculations.
Annual mean PM2.5 total concentrations for (a) CAMx, (b) CHIMERE, (c) CMAQ, (d) EMEP, and (e) LOTOS-EUROS, as well as for the (f) ensemble
model mean. Below the domain figures are the respective frequency distributions
displayed for the annual mean PM2.5 concentrations, referring to the
whole model domain.
Annual mean PM2.5 relative potential ship impacts for (a) CAMx, (b) CHIMERE, (c) CMAQ, (d) EMEP, and (e) LOTOS-EUROS, as well as for the (f) ensemble model mean. Below the domain figures are the respective frequency
distributions displayed for the annual mean PM2.5 potential ship impacts,
referring to the whole model domain.
The potential impacts of PM2.5 from ships simulated by CAMx,
LOTOS-EUROS, and EMEP have the largest areas with values of up to 25 % at the
main shipping routes (Fig. 3). CMAQ and CHIMERE have a potential shipping
impact of 15 % along the main shipping lines close to the African coast.
This impact is lower than that shown in other studies. Aksoyoglu et al. (2016) found the highest impacts of 25 % to 50 % of total PM2.5 concentrations when using CAMx along the main shipping routes. Sotiropoulou
and Tagaris (2017) used CMAQ for simulations and stated that emissions from
shipping are likely to increase PM2.5 concentrations during winter by
up to 40 % over the Mediterranean Sea, while during summer, they
simulated an increase of more than 50 %. In both studies, the modeled
year is 2006, which might explain the deviation from the present study as using a
different year. Regarding coastal areas in the present study, potential
shipping impacts reaching 12 % to 15 % are simulated.
Annual mean PM2.5 absolute potential ship impacts for (a) CAMx, (b) CHIMERE, (c) CMAQ, (d) EMEP, and (e) LOTOS-EUROS, as well as for the (f) ensemble model mean. Below the domain figures are the respective frequency
distributions displayed for the annual mean PM2.5 potential ship impacts,
referring to the whole model domain.
Regarding the absolute potential impacts of ships at the main shipping
routes, CAMx, CHIMERE, and EMEP show values of 2.0 µg m-3, and the values simulated by CMAQ and LOTOS-EUROS are
between 0.5 and 1.0 µg m-3 (Fig. 4). The median of the ensemble mean is 0.85 µg m-3 (Figs. 4 and 8). Aksoyoglu et al. (2016)
simulated similar shipping impacts with CAMx, with values mainly between 0.5
and 1.0 µg m-3.
The sea salt concentrations might partly give an explanation for the
differing PM2.5 concentration distribution among the models. The annual
mean sea salt (NaCl) concentration in fine and coarse PM showed the highest
values for CHIMERE, which might be an explanation for the high PM2.5
absolute concentration (Supplement Fig. S1). The LOTOS-EUROS sea salt displayed
the lowest concentrations; the overall PM2.5 concentration is also the lowest
compared with the other CTMs. The sea salt concentration was the highest (up to
7.0 µg m-3) over the sea in areas with high surface wind
speeds for CHIMERE, CMAQ, EMEP, and LOTOS-EUROS (Fig. S2). This can be
confirmed by the correlation between wind speed and sea salt at several points
over water for CMAQ, EMEP, and LOTOS-EUROS (Fig. S3; Table S1 in the Supplement). CAMx is
excluded from the analysis since sea salt is only present in fine PM.
Solazzo et al. (2012) demonstrated that the chemical components
SO42-, NO3-, and NH4+ were better
reproduced by nine CTMs than the total PM2.5. They concluded from this
result that other components (e.g., organic aerosols) could be simulated
with less accuracy than inorganic components.
Correlations between models for the PM2.5 base runs of the
whole domain (all grid cells), based on daily PM2.5 total
concentration data.
High amounts of NH3, HNO3, SO2, and NO2 are expected to
lead to higher values of the aerosol particles composed of NH4+,
NO3-, and SO42-. The modeled spatial distributions of
these precursors can be found in the Supplement (HNO3: Figs. S4–S6;
NH3: Figs. S8–S10; SO2: Figs. S11–S13; and NOx: Figs. S14–S16).
The highest annual mean HNO3 concentration among the base runs is found
in the CAMx and CHIMERE simulations over water (2.0 to 5.0 µg m-3); over land, the
values are between 0.0 and 1.5 µg m-3, and those in coastal areas reached 2.0 µg m-3 (Fig. S4). The absolute potential ship impact is
also the highest in CAMx and CHIMERE at the main shipping routes and over water
areas (1.0 to 3.0 µg m-3). The relative potential ship impact on the total
HNO3 ranges from 60 % to 85 % along the main shipping routes
simulated by CAMx, CMAQ, and EMEP (Fig. S4). These impacts are slightly
lower for CHIMERE and LOTOS-EUROS (60 % to 75 %).
The high HNO3 concentrations simulated by CAMx and CHIMERE might be
traced back to the NO2 concentrations; these two models also show
higher NO2 concentrations than the other CTMs (Fig. S14; Fink et al.,
2023). This can be explained by the fact that HNO3 is a major NO2
sink, especially during daytime. NO2 is primarily emitted from
anthropogenic fossil fuel burning but also from natural sources (i.e.,
soil emissions, biomass burning, lightning). During daytime, the main
NO2 removal mechanism is oxidation by hydroxyl (OH) radicals to form
HNO3 (Seinfeld and Pandis, 1998).
It can be concluded that in areas with shipping, more NO2 enters the
atmosphere; the total NO2 concentration increases; and as a result of
the subsequent reactions, the HNO3 concentration also increases. The
HNO3 : NO2 ratio can be used to normalize the data (Fig. S7). The
ratio displays low values over land and along the main shipping routes,
indicating that in these areas, both the HNO3 and the NO2
concentrations are high. A low HNO3 : NO2 ratio could also mean that
only a small amount of OH is present, especially in areas with a low O3
concentration.
After its formation, HNO3 can react with NH3 to be neutralized and
to form particles when NH3 is in excess. The annual mean NH3 for
the base case shows very similar patterns and values among all models (Fig. S8). The highest concentrations of NH3 with all emission sources are
located over land areas with values up to 2.5 µg m-3, which can be traced back to agriculture, the main source of NH3
emissions (Behera et al., 2013). Over water areas, the NH3
concentration is very small, typically between 0.0 and 0.3 µg m-3, except for the
slightly higher results modeled by LOTOS-EUROS, with values between 0.2
and 0.8 µg m-3.
Negative potential ship impacts (-0.01 to -1.0 µg m-3 and -2.5 % to -150 %; Figs. S9 and S10)
are found for the whole domain in all five models. The relative ship impacts
are the lowest at the main shipping routes for CAMx and EMEP. The spatial
distribution of the NH3 relative ship impact is opposite to the
simulated HNO3 values at the main shipping routes, with low NH3
and high HNO3 values. These results indicate that available NH3
reacts directly with HNO3 to form particles (i.e., NH4NO3).
Thus, NOx emissions from shipping lead to HNO3 formations and
subsequent NH3 consumption; e.g., shipping impacts on NH3
concentrations are usually negative.
The CAMx simulations show the highest SO2 concentrations with more than 10 µg m-3 in some areas in western Turkey, in urban
areas, and along major shipping lanes (Fig. S11). The results from the
other four CTMs display high values around the Bosporus and in some areas
over the Balkan Peninsula with values of 11 µg m-3
and much lower concentrations along the main shipping routes. The potential
ship impacts are similarly high in CAMx and CHIMERE (1.0 µg m-3; 85 % of the total concentration; Figs. S12 and S13),
with the highest values along the major shipping route north of the African
coast. The CMAQ, EMEP, and LOTOS-EUROS results display similarly high values
but only in small areas. The modeled year is 2015, so the global 0.5 %
sulfur cap of marine fuels was not yet effective. Heavy fuel oils with
sulfur contents reaching 3.50 % were used until 2020 to power ships;
thus, the SO2 emitted from ships in the present study is still high, and
it can be expected that it has a large impact on secondary particle
formation.
Inorganic aerosol speciesConcentrations
In the Northern Hemisphere, secondary inorganic ammonium, sulfate, and
nitrate aerosols represent a large fraction of the PM2.5 composition
(Jimenez et al., 2009). Ammonium preferentially binds to
SO42- in atmospheric aerosols in the form of
(NH4)2SO4. NH4NO3, on the other hand, is formed in
areas characterized by high-NH3 and high-HNO3 conditions and low-H2SO4 conditions. The results of the CTMs with regard to these
three particle species and their potential ship impacts are considered in
the following section. The spatial distributions of the total concentrations
and absolute potential ship impacts of the individual species can be found
in the Supplement (NH4+: Figs. S17 and S18; SO42-:
Figs. S19 and
S20; and NO3-: Figs. S21 and S22). The spatial
distribution of the relative potential ship impact is shown in Figs. 5 to 7.
Annual mean NH4+ relative potential ship impacts for (a) CAMx, (b) CHIMERE, (c) CMAQ, (d) EMEP, and (e) LOTOS-EUROS.
Below the domain figures are the respective frequency distributions displayed
for the annual mean NH4+ potential ship impacts, referring to the
whole model domain.
Annual mean SO42- relative potential ship impacts for (a) CAMx, (b) CHIMERE, (c) CMAQ, (d) EMEP, and (e) LOTOS-EUROS.
Below the domain figures are the respective frequency distributions displayed
for the annual mean SO42- potential ship impacts, referring to the
whole model domain.
Annual mean NO3 relative potential ship impacts for (a) CAMx,
(b) CHIMERE, (c) CMAQ, (d) EMEP, and (e) LOTOS-EUROS. Below the
domain figures are the respective frequency distributions displayed for the
annual mean NO3 potential ship impacts, referring to the whole model
domain.
The spatial distribution of NH4+ shows that the lowest total
annual mean can be found mainly in the southwestern part of the domain
(approximately 0.0 µg m-3) and the highest in the Po
Valley and Bosporus (1.5 µg m-3, Fig. S17). The
relative ship impacts are very similar for all models (0.25 % to 5.0 %
over land, 10 % to 25 % over water; Fig. 5) as well as for the
absolute ship impact (Fig. S15). Aksoyoglu et al. (2016) simulated
NH4+ values between 0.0 and 0.2 µg m-3 in the Mediterranean region, with higher
concentrations (0.4 µg m-3) in the Po Valley. This
is within the same range of concentrations in the present study. Ge et al. (2021) used the EMEP model to simulate global particle species
concentrations and compared them with measured concentrations. They showed in
their study that the NH4+ concentrations simulated in Europe in
2015 were overestimated by a factor of 2 compared with the actual measured
NH4+ concentrations. The measurements displayed a mean of 0.45 µg m-3. The ensemble mean for NH4+ in
the present study (0.6 µg m-3, Fig. 8a) is in
good agreement with these measurements. However, a previous study on
measured compared with simulated aerosol distribution with the CMAQ model
displayed a slight underestimation of NH4+ (Matthias, 2008).
(a) Boxplots for concentrations of PM2.5 and the PM2.5
components SO42-, NO3, NH4+, and “others” as
simulated by the five CTMs. The ensemble mean is “all_mean”.
Others are calculated as PM2.5 minus the sum of SO42-,
NO3, and NH4+. Data are based on the whole domain (all grid
cells) and hourly data for all emission sources (“emisbase”). (b) The same as
(a) but for ships only.
The NH4+ proportion to the total PM2.5 is similar among all
models (5.6 % to 7.8 %; Fig. 8a, Table 4), and only LOTOS-EUROS
displayed a relatively high share (12.2 %). This pattern is similar for
the ship impacts, where all models show proportions between 9.1 % and
12.6 %, but higher values are simulated by LOTOS-EUROS (23.5 %; Fig. 8b, Table 5).
SO42- is the oxidation product of SO2, which is primarily
emitted by anthropogenic processes such as fossil fuel combustion, petroleum
refining, and metal smelting (Zhong et al., 2020). In the present study,
SO42- is the main contributor to the total PM2.5 mass (Fig. 8,
Table 4). Especially in the model ensemble mean for the absolute
ship-related concentrations, SO42- makes up 44.6 % of
PM2.5 (Fig. 8b, Table 5). The annual mean SO42- total
concentration is the highest for CHIMERE in the eastern part of the domain,
reaching 6.0 µg m-2. EMEP displays a SO42-
concentration within the ranges of the other models, CAMx, CMAQ, and
LOTOS-EUROS, in the western part of the domain. These models show very
similar spatial distributions with concentrations of up to 2.0 µg m-3. The median ensemble mean for the run with all
emission sources is 2.0 µg m-3. This ensemble mean
is low in comparison with the results of Solazzo et al. (2012); they found a
mean value of 6.0 µg m-3 but considered a larger
European area that included the areas with the highest SO42-
concentrations in Europe. For this larger area, Solazzo et al. (2012) found
that the models used underestimated SO42- by 7 % to 17 %.
In the present study, the relative potential ship impact on the total
SO42- is the lowest over land, with 0 % to 3.0 %, and higher in
coastal areas, with values from 6 % to 20 % (Fig. 6). Along the main
shipping routes it is the highest, reaching 50 % for CAMx, EMEP, and LOTOS;
for CHIMERE and CMAQ, it is lower, with values reaching 30 %. Aksoyoglu et
al. (2016) showed similar relative potential ship impacts of 50 % to 60 % in the western Mediterranean. In their study, values were between 0.0 to 1.0 µg m-3 over land
areas, but over water along the main shipping routes they were the highest at
2.2 µg m-3.
Mallet et al. (2019) traced back higher SO42- in the eastern part
of the domain due to westerly winds. In the present study, we found this
higher concentration for SO42- in the eastern part of the
Mediterranean as well. In Lampedusa, they found ammonium sulfate contributed
63 % to PM1 mass, followed by organics (Mallet et al., 2019). In our
study, the organics/others had the highest share of total PM2.5 when
considering all emission sources, followed by sulfate and ammonium.
In the present study, CTM systems simulated lower values for ship impacts;
over land, they are 0.0 to 0.03 µg m-3, and along the main shipping routes, they reached 0.9 µg m-3. Regarding the absolute ship impacts on
SO42-, the model simulations display similar
concentrations and are slightly lower for CMAQ and LOTOS-EUROS (Fig. S20)
compared with the other models. Especially over water areas, large areas with
considerable SO2 and SO42- concentrations can be seen.
Because NH4+ is preferentially bound to SO42- in atmospheric
aerosols to form (NH4)2SO4, in areas over water, less
NH4NO3 forms.
Im et al. (2014) suggested in their intercomparison study that over Europe,
SO42- levels were underestimated by most models; only a few models
overestimated SO42- concentrations in Europe. The underestimating
models were WRF-Chem models, and the SO42- underestimations were
attributed to the absence of SO2 oxidation in cloud water in the
heterogeneous phase.
The highest annual mean NO3- total concentrations are simulated
over land areas, especially over Italy and in the Balkan states (> 2 µg m-3; Fig. S18), and the lowest concentrations are over the
sea. CAMx, CMAQ, and LOTOS-EUROS show higher concentrations compared with
results derived from CHIMERE. The concentrations over water are lower than
those over land. The ensemble median of all CTMs over the whole domain is
0.63 µg m-3 (median value; Fig. 8a). The absolute
potential impacts of ships on the total NO3- concentrations are
similar among all models, displaying values mainly between -0.005 and 0.15 µg m-3; only CMAQ
demonstrates relatively low values along the main shipping routes (-0.5 µg m-3), and CAMx has higher values (1.0 µg m-3) in some coastal areas (Fig. S19). This can be
explained by higher SO42- concentrations derived from SO2
emissions. Sulfate replaces nitrate as long as the ammonia concentration is low.
In model simulations with ships, NO3- can decrease because ammonia
is already taken from sulfur emissions from ships. Aksoyoglu et al. (2016)
found similar results for the Mediterranean Sea considering the
NO3- concentrations, with values between 0.0 µg m-3 and 0.2 µg m-3. Im et al. (2014) showed that simulated NO3- levels were overestimated by
most of the CTMs by more than 75 %. Higher concentrations over water than
over land due to NH4NO3 formation are found in areas characterized
by high-NH3 and high-HNO3 conditions and low-H2SO4
conditions. In the present study, the relative potential ship impacts on
NO3- display contradicting tendencies among the models (Fig. 7). The CAMx, EMEP, and LOTOS model results are similar, with relative
potential ship impacts over land from 0.0 % to 5.0 % (in the Balkan
states), those in coastal areas and Italy from 10 % to 25 %, and those
along main shipping routes from 50 % to 65 % or even up to 85 %.
CHIMERE and CMAQ display lower relative potential ship impacts. For CMAQ,
the impact is negative along the main shipping routes, at -25 %.
Sulfur dioxide or ammonia might lead to a negative NO3- impact
because the NO2 emissions from ships would make a positive contribution
to nitrate formation. Therefore, without ships, an (NH4)2SO4
should be formed, which is more stable than NH4NO3. These low
values in the aerosol species for CMAQ but higher values for EMEP, CAMx, and
LOTOS represented the PM2.5 ship impacts and might partly explain the
deviations in PM2.5. Furthermore, in CMAQ the coarse mode in nitrate
and ammonium has a larger share compared with the other CTMs. A more detailed
discussion is given in Sect. 4.
Regarding the PM2.5 composition, the share of other particles, which
contain mainly organics but also, e.g., sea salt, is the highest compared with the
inorganic species (Fig. 8). Nevertheless, the particle composition
revealed varying distributions in the ship-related PM2.5 concentration.
Here, inorganic particle species have relatively high percentages compared
with organic aerosols. In some cases, sulfate has an even higher share of the
total PM2.5 than other particles.
The seasonal variability in particle species shows that NO3- is
more temperature dependent than SO42- and NH4+.
NO3- is higher in winter and spring but lower in summer and
autumn. This pattern can be found in all CTM simulations. For
PM2.5, on the other hand, no discernible pattern is found
regarding seasonal variability. In particular, the ensemble mean PM2.5 concentration remained within the same range in all seasons.
Relative particle species of total PM2.5 emissions.
Concentration of particle species and precipitation divided by
seasons and CTMs. “all_mean” displays the model ensemble.
Spring: March, April, May; summer: June, July, August; autumn: September, October, November; winter: December, January, February.
Concentration is based on the annual median over the whole domain.
Precipitation displays the seasonal sum (in mm).
Wet deposition
Wet deposition can provide indications of the fate of particles. EMEP does not
deliver separate deposition files for individual particle species but for
reduced and oxidized nitrogen. Thus, EMEP is not considered when analyzing
wet deposition in this study.
Regarding the spatial distribution of NH4+ wet deposition, the highest
annual sums are displayed by CMAQ and LOTOS-EUROS (up to 250 mg m-2 yr-1 over land; up to 50 mg m-2 yr-1
over water; Fig. S23). CAMx and CHIMERE show a similar spatial
distribution with values mainly between 10 and 25 mg m-2 yr-1. CAMx and CHIMERE used the same meteorology data,
but despite this the seasonal distribution of wet deposition differs
(Fig. 10). An explanation for this differing behavior might be provided by
the scavenging mechanisms. In CHIMERE the in-cloud mechanism for deposition
of particles is assumed to be proportional to the amount of water lost by
precipitation. In CAMx, the in-cloud scavenging coefficient for aqueous
aerosols is the same as for the scavenging of cloud droplets. Below the
cloud, CHIMERE uses a polydisperse distribution following Henzig et al. (2006), whereas in CAMx for rain or graupel the collection efficiency is
calculated as in Seinfeld and Pandis (1998). The other possible explanation
is that all the emissions in CAMx are emitted in the first layer and in
CHIMERE it depends on the emissions distribution.
Regarding the wet deposition of sulfate, the annual totals for all
emission sources are the highest over the Balkan Peninsula in the CMAQ and
LOTOS-EUROS model outputs (300 to 800 mg m-2 yr-1; Fig. S24). For CAMx over land areas, the
values reach 300 mg m-2 yr-1, and the lowest totals over land
can be seen in the CHIMERE results (0.0 to 50 mg m-2 yr-1). Over water, these values are low in all model
outputs (50 to 150 mg m-2 yr-1),
except in CHIMERE, where, in contrast to the other models, the highest wet deposition was
found over water.
The wet deposition of NO3- is the highest for CMAQ (> 400 mg m-2 yr-1) over the whole domain (Fig. S25). For CAMx and
LOTOS-EUROS, it is generally lower, with most areas displaying 25
to 50 mg m-2 yr-1. The lowest wet deposition of
nitrate is shown in CHIMERE outputs with values not exceeding 50 mg m-2. Regarding the sum for the whole year, the highest
values are found for CMAQ (northern Italy and the Balkan Peninsula, where
the urban-area values reached 400 mg m-2 yr-1). Over water,
deposition is lower than over land in the results of all CTMs. Lower
wintertime precipitation in CMAQ compared with the other models might lead to
high particle concentrations as well as high deposition due to low dilution
(Fig. 10).
Wet-deposition sum (mg per season) of particle species and
precipitation divided by seasons and CTMs. “all_mean”
displays the model ensemble. Spring: March, April, May; summer: June,
July, August; autumn: September, October, November; winter: December,
January, February. Wet deposition is based on the annual sum over the whole
domain. Precipitation displays the seasonal sum (in mm).
Wet deposition depends mainly on the ability of the models to predict the
amount, duration, and type of precipitation. The precipitation data show
that the lowest values are found for CMAQ input data. CAMx and CHIMERE use
the same meteorological input data and thus display the same precipitation
results, with the highest values in winter. CMAQ and LOTOS-EUROS have
precipitation values within a similar range, with the highest values
occurring in autumn and winter.
Although the precipitation results in CAMx and CHIMERE are the same, wet
deposition differed between these two models, indicating that the
concentration as well as model internal mechanisms caused differences rather
than the input data. Additionally, in CMAQ, a lower wet-deposition rate is
expected for nitrate. There are usually two mechanisms that are important for
scavenging in CMAQ: in-cloud and below-cloud scavenging. High wet deposition
for nitrate in CMAQ outputs might be traced back to efficient below-cloud
scavenging of coarse-mode particles containing nitrate, through which the
wet deposition can be high despite precipitation in similar ranges to other
models. Furthermore, the deposition of particulate nitrate crucially depends
on the reactive uptake of HNO3 to larger particles (Karl et al., 2019)
because coarse-mode particles are removed much faster than fine-mode
particles.
Discussion
Various reasons for deviations in PM2.5 concentrations among regional
CTM systems might be traced back to model-specific calculations.
Regarding PM (coarse and fine for sea salt), an uncertainty among models
might be caused by the differences in the calculation of sea salt and dust
emissions. Here, both are considered in all CTMs, except for dust in CMAQ. If
sodium chloride and dust components are not considered, underestimations of
PM and uncertainties in areas near coasts (sea salt) or where dust is
important, e.g., Saharan dust in the Mediterranean region, occur, as
described in Sect. 3.1. Furthermore, if sea salt and dust are omitted from
the pH calculations, it might also cause deviations in sulfur chemistry, as
this factor is very sensitive to pH.
In the CMAQ runs dust was considered at the model boundaries but dust
emissions were not included. The Mediterranean region is frequently affected
by Saharan desert dust (Palacios-Peña et al., 2019), but the main source region
for this dust emission is not included in the model domain; thus the dust
coming from the boundary can be seen as sufficient for the CMAQ model run.
Generally, the boundary conditions for dust and sea salt in CAMx and CHIMERE
were produced by offline models that run on meteorological fields
from GEOS-5, GEOS DAS, and MERRA. For CMAQ and LOTOS-EUROS these boundary
conditions were produced within the boundary condition calculations.
Boundary conditions of EMEP are developed from climatological ozone-sonde
datasets.
All models used offline meteorology in which the ABL heights were
calculated. Annual medians of the atmospheric boundary layer heights at 16:00
and 04:00 UTC were compared among the models. The comparison of spatial
distribution of ABL heights at 16:00 and 04:00 shows that over water, the ABL
heights do not have much variability in all models (Figs. S26 and S27). The
lowest ABL height over water was used for CHIMERE. This corresponds to the
high PM2.5 concentrations simulated by this model over water. Over
land, the comparison of spatial distribution from 16:00 to 04:00 displays more
variable ABL heights: during nighttime the ABL heights are up to 200 m,
whereas during daytime the heights increase to 1000 m or higher (Figs. S26
and S27). Over land the input in CAMx, CHIMERE, CMAQ, and LOTOS-EUROS has a
higher median ABL at 16:00, whereas in EMEP it is the opposite, with the
highest median at 16:00 mainly being over water areas. However, there was not a large
deviation in the PM2.5 concentration simulated by EMEP from concentrations
received from other models. Generally, due to ABL dynamics, deviations between
measured and simulated data can be expected because measurement stations were
chosen close to the coast, which leads to uncertainties. In these areas, the
measurements are influenced by air masses either coming from water or coming
from land. In addition, measured data were received from one measurement
point, which is hardly representative of a whole grid cell of 12×12 km2.
The treatment of dust and sea salt, as well as the boundary conditions used, has an
effect on the analysis and comparison of PM results because these
parameters are part of the PM2.5 formation but differ among the models.
Regarding the CTM performances, reasons for underestimations of PM2.5
have already been discussed in previous studies: for CAMx, Pepe et al. (2019)
linked these underestimations to meteorological parameters and to the
overestimation of the vertical mixing in the lower atmosphere. Tuccella et
al. (2019) found underestimations of PM2.5 in the CHIMERE model and
explained these by an excess of wet scavenging in the model. An excess of
wet scavenging in CHIMERE compared with the other CTM systems is not found in
the present study; thus it cannot be used as an explanation for deviations
here. In EMEP, which differs from the other CTM systems, the MARS module was
used to calculate the equilibrium between the gas and aerosol phases; this
model does not treat sea salt or dust, leading to underestimations of
PM2.5. Kranenburg et al. (2013) linked the underestimation of
particulate matter in LOTOS-EUROS to the missing descriptions of SOA
processes in the model. Thus, various reasons and combinations of reasons
can lead to underestimations of PM2.5 in the CTM systems used herein.
For a better understanding, the inorganic particle species are considered in
the present study. Consideration of inorganic and organic particles
could lead to more uncertainties. Moreover, in shipping emissions the
inorganic aerosols display a higher share.
A large part of PM2.5 is secondary; therefore underestimations can be
linked to underestimations of precursors, e.g., NO2. This has already been
shown in the first part of this intercomparison study, where all five CTM
systems underestimated measured NO2 (Fink et al., 2023). But
SO2 is also usually underestimated by CTMs, as shown in previous studies
(e.g., Eyring et al., 2007). Four out of five CTM systems underestimate the
actual measured PM2.5 concentration in the present study. Gaseous precursors like SO2 and NO2 need to be oxidized before
they can form particles in reactions with ammonia. The hydroxyl radical (OH)
is the main oxidant. The amount of available OH can be analyzed when the
NO2 concentration is set in relation to HNO3 and NO3-
(Fig. S28). This gives an indication of the OH availability. In ship
plumes OH is consumed fast; therefore values are low along the shipping
lanes. In regions with lower NO2 concentrations, more OH is available, and
HNO3 is efficiently formed. In the present study, the HNO3 was
similar within all five CTM systems (Fig. S5).
One reason for the differences in HNO3 might be traced back to the
amount of cloud droplets, since HNO3 is resolved in them. The dissolution
of gases in droplets is usually assumed to be irreversible for HNO3 and
NH3 in CTMs; thus, the amount of formed ammonium nitrate mass depends
on the amount of HNO3 or the cloud droplets. This could in the end
lead to the deviation among the CTM-simulated HNO3.
The preference of NH4+ to bind to SO42- in atmospheric
aerosols to form (NH4)2SO4 explains why in some models
NO3- displays relatively low values when the SO42-
concentration is high. CHIMERE, for instance, has an NO3- share
of 3.1 % to the total PM2.5 and an SO42- share of 27.0 %,
whereas in the CAMx results, NO3- has a share of 11.1 % to
the total PM2.5 and an SO42- share of 14.6 %. This is
confirmed by the low SO2 concentration and high SO42-
concentration in CHIMERE (Figs. S11 and S19), indicating that sulfate is
formed more efficiently compared with CAMx. Furthermore, this leads to lower
NO3- concentrations in the CHIMERE output (Fig. S21). For
SO2 and SO42- the concentration of cloud water and the amount of cloud
droplets also play an important role.
Regarding the thermodynamic equilibrium within the models, ISORROPIA and
ISORROPIA II mechanisms are used in all CTM systems except EMEP, meaning
similar results can be assumed to be obtained from this mechanism. Despite
this similarity, differences in concentrations may be a result of
differences in available cloud water, vertical mixing, the spatiotemporal
distribution of emissions, or aerosol size distributions. EMEP uses the MARS
module to calculate the equilibrium between the gas and aerosol phases.
Although four of the five models use the ISORROPIA or ISORROPIA II mechanisms
for inorganic secondary aerosol formation, many factors within these models
still cause significant differences among the model outputs.
The aerosol size distribution also has an impact on the particle species
distribution. As displayed in Table 1 (Sect. 2.1), there are two concepts
for how the aerosol size distribution is represented within the models:
the distribution is either in bins or in log-normal modes. As already discussed in
Solazzo et al. (2012), the PM chemical composition differs greatly with the
particle size. Consequently, differences in modeling the aerosol size
distribution also affect the chemical composition. In CMAQ, for example,
large fractions of nitrate and ammonium can be found in the coarse mode
where they undergo different removal processes than in the fine mode.
Although there is harmonization in terms of the input emission data in the
present study, the internal model mechanisms used to calculate particulate
matter lead to differences in the particle species distribution, as
discussed in Sect. 3.1.
In addition, the calculations of how to determine PM2.5 vary among CTM
systems or even within one CTM. As an example, there are two possibilities
for calculating PM2.5 within CMAQ: either online during the model run
with the PM2.5 module or subsequently by calculating the value as the
sum of two modes. These different options lead to different results (as
shown by Jiang et al., 2006) and will also affect the particle composition.
In the present study, the sum of two modes is used in CMAQ.
Model simulations with relatively high PM2.5 concentrations display
higher absolute shipping impacts on PM2.5, as presented in Sect. 3.2.
Consequently, relatively low variability in the relative potential ship
impacts among the models compared with that of the absolute values could be
expected. For a more quantitative evaluation, the relative potential ship impact
is plotted against the absolute potential impact. A larger incline of the
regression line can be explained by a higher background PM2.5
concentration; thus the relative ship impact is lower for the same concentration
increase (e.g., EMEP and CHIMERE) (Fig. S29).
From the ISORROPIA and ISORROPIA II mechanisms, it can be expected that the
molar ratios between the acids on the one side (NO3- and
SO42-) and the base on the other side (NH4+) are in
balance. However, the ratio between SO42-, NH4+, and
NO3- shows that the balance in all models except in LOTOS-EUROS is
not given for PM2.5; sulfate plus nitrate is much higher compared with
ammonium (Fig. S30). This balance is almost perfectly given in
LOTOS-EUROS, although both CMAQ and LOTOS-EUROS used the ISORROPIA II
mechanism. An imbalance among the
inorganic particle species is present, especially at the shipping lanes. Differences in the particle species ratio
among the models can be traced back to the differences in the particle size
distribution. In contrast to the other models, CAMx only has three species in
the coarse mode: coarse others primary, coarse crustal, and reactive gaseous
mercury. For NO3- and SO42-, the ratio between the fine
and coarse mode is calculated for the CTMs (Figs. S31 and S32).
NH4+ was not considered here, since it is only in present in
the coarse mode in CMAQ.
These ratios show that CHIMERE and LOTOS-EUROS only have a small proportion
of particles in the coarse mode. For SO42- in LOTOS-EUROS the coarse
particle concentration is zero and for EMEP no SO42- is present in
the coarse mode. In CMAQ a higher concentration of particles is assigned to the
coarse mode and also for NH4+.
The present study has shown that different reasons can cause deviations
among the simulated PM2.5 CTM outputs. The major reasons are the
differences in the size distribution and how models distribute chemical species
between the coarse and fine modes (PM2.5 and PM10). Differences
among the modeled PM2.5 concentrations can also be a result of the
differences in the height of the lowest model layer and the way in which
ship emissions are distributed among the layers. As shown in Fink et al. (2023), the vertical distribution of PM2.5 precursor emissions varies
among the models; e.g., in CAMx all shipping emissions are assigned to the
lowest layer. This leads to differences in chemical transformations because
of different concentration levels close to the source and consequently to
deviations among the particle distributions. Furthermore, precipitation
differences lead to variations among the model outputs for wet deposition.
The limitations of the present study are that only the chemistry of the lowest
layer is evaluated. The model input was standardized as far as possible, but
meteorological input data varied and are not compared in detail here.
Interactions between fine and coarse particles are only studied to a limited
extent, and the same holds for aqueous chemistry, which has an impact on
the oxidation mechanisms of sulfur species.
Summary and conclusion
The current work investigates and analyzes the predictions of five different
CTM systems for PM2.5 and inorganic particle species (NH4+,
SO42-, NO3-) dispersion and transformation in the
Mediterranean region. Additionally, the total concentration focus is on the
potential ship impact. The results show that four of the five models
underestimated the actual measured PM2.5 concentrations at stations
close to the European coastline. The relative ship impacts on PM2.5
simulated by the CTMs at the measurement stations are between 5.7 %
(CMAQ) and 13.8 % (CAMx). The potential impacts of PM2.5 from ships
simulated by CAMx, LOTOS-EUROS, and EMEP have the largest areas with values of
up to 25 % along the main shipping routes in the Mediterranean Sea. CMAQ and
CHIMERE simulated potential ship impacts of 15 % along the main shipping
lines close to the African coast. These impacts are within the range of the
ship impacts obtained in other studies.
The spatial distribution of ammonium displays a low total annual mean mainly
in the southwestern part of the domain (approximately 0.0 µg m-3) and is the highest in the Po Valley and Bosporus (1.5 µg m-3 ). The ensemble mean of NH4+ (0.6 µg m-3 ) is in good agreement with the measurements
provided in previous studies. The relative and absolute ship impacts are
very similar for all models (0.0 to 0.06 µg m-3 over land, up to 0.15 µg m-3 over water; 0.25 % to 5.0 % over land, and 10 % to 25 % over water). This indicates that differences among the
simulated PM2.5 from ships result from differences in sulfate and
nitrate.
The NH4+ proportion to total PM2.5 is similar in all models
(5.6 % to 7.8 %), and only LOTOS-EUROS shows a relatively high share
(12.2 %). The ship impact pattern is similar; all models display
proportions between 9.1 % and 12.6 %, but higher values are simulated
by LOTOS-EUROS (23.5 %).
SO42- is the main contributor to the total PM2.5 concentration
regarding shipping emissions only. In the model ensemble mean for the
absolute ship concentration, SO42- accounts for 44.6 % of
PM2.5. The annual mean sulfate total concentration is the highest for
CHIMERE in the eastern part of the domain, reaching 6.0 µg m-3.
CAMx, CMAQ, EMEP, and LOTOS-EUROS simulate a total SO42-
concentration within one range between 0.4 µg m-3 and
2.0 µg m-3 in the western part of the domain. The
relative potential ship impacts on the total SO42- are the lowest over
land, with values of up to 3.0 %, and are higher in coastal areas, with values
ranging from 6 % to 20 %. Along the main shipping routes, the impacts
are the highest, reaching 50 % for CAMx, EMEP, and LOTOS-EUROS; for CHIMERE
and CMAQ, they are lower, with values reaching 30 %. Regarding the
absolute ship impacts on SO42-,, the model simulations
display similar concentrations and are slightly lower for CMAQ and
LOTOS-EUROS. Concentrations are in particular identified over water areas with relatively high SO2 and
SO42-. Because NH4+
preferentially binds to SO42- in atmospheric aerosols to form
(NH4)2SO4, in areas over water less NH4NO3 forms.
The highest annual mean NO3- total concentrations appear over
land areas in the simulations by CAMx, CMAQ, and LOTOS-EUROS, especially over
Italy and in the Balkan states (> 2 µg m-3). The lowest concentrations are simulated by CHIMERE. The concentrations
over water are lower than those over land areas. The ensemble mean of all
CTMs over the whole domain shows a median value of 0.63 µg m-3. Higher concentrations over land than over water are
expected due to NH4NO3 formation in areas characterized by high-NH3 and high-HNO3 conditions and low-SO42- conditions.
The relative potential ship impact on NO3- differs among the
models. The CAMx, EMEP, and LOTOS-EUROS results are similar; the relative
potential ship impacts over land range from 0.0 % to 5.0 % (in the
Balkan states), those in coastal areas and Italy range from 10 % to 25 %, and those along main shipping routes range from 50 % to 65 % or even reach
85 %. CHIMERE and CMAQ show lower relative potential ship impacts
for NO3-. For CMAQ, the impacts are the lowest along the main shipping
routes; nitrate is even reduced by 25 %. Low values in nitrate can be
explained by the preference to form (NH4)2SO4; thus nitrate
stays in the gas phase or is transferred to the coarse mode. These low
values for SO42- and NO3- in CMAQ but relatively high
values for EMEP, CAMx, and LOTOS are reflected in the PM2.5 ship
impacts and partly explain the deviations in PM2.5 among the models. As
expected, the seasonal variabilities in particle species show that
SO42- and NH4+ are less temperature dependent than
NO3-. Nitrate is higher in winter and spring but lower in summer
and autumn. This pattern is found in all CTM simulations.
The spatial distribution of NH4+ wet deposition shows the highest
annual sums by CMAQ and LOTOS-EUROS (up to 250 mg m-2 yr-1
over land; up to 50 mg m-2 yr-1 over water). CAMx and CHIMERE
show a similar spatial distribution with values mainly between 10
and 25 mg m-2 yr-1. For wet
deposition of SO42-, the annual totals for all emission
sources are the highest over the Balkan Peninsula in the CMAQ and LOTOS-EUROS
model outputs (300 to 800 mg m-2 yr-1). For CAMx over land areas, the values reach 300 mg m-2 yr-1, and the lowest totals over land can be seen in the CHIMERE results
(0.0 to 50 mg m-2 yr-1). Over
water, these values are low in all model outputs (50 to 150 mg m-2 yr-1), except in CHIMERE. Wet deposition
of NO3- is the highest for CMAQ (> 400 mg m-2 yr-1) over the whole domain. For CAMx and LOTOS-EUROS,
it is generally lower, with most areas displaying 25 to 50 mg m-2 yr-1. The lowest wet deposition of nitrate is shown in
CHIMERE outputs, with values not exceeding 50 mg m-2. Over
water, deposition is lower than over land in the results of all CTMs.
The complexity of particle treatments within the models, as well as the
large number of causes for these changes, makes it difficult to find a single
cause for the variable outputs. One point causing uncertainties is that the
aerosol formation mechanisms differ among CTMs. The detailed investigation
of PM2.5 and its chemical composition has demonstrated that differences
among the particle species might be traced back to the aerosol size
distribution. This was shown in particular for CMAQ regarding the balance of
the inorganic particle species nitrate and sulfate on the one side and
ammonium on the other side. CMAQ and EMEP tend to assign a higher particle
mass to the coarse mode compared with the other three CTMs. This has
implications for particle deposition because both wet and dry depositions
are more efficient for larger particles.
An ensemble mean with standard deviations based on several model results can
provide a more reliable assessment of possible ship impacts on air
concentration and deposition. Previous research has demonstrated that using
only one chemical transport model resulted in underestimated model
uncertainty and overconfidence in the conclusions (e.g., Solazzo et al.,
2013; Riccio et al., 2012; Solazzo et al., 2018), indicating that a model
ensemble should be used. Particularly in terms of the study's policy
point of view, the ensemble mean is important: if model simulations are used
to support decision-making regarding shipping regulations, the
uncertainty in individual models must be considered.
The goal of this study was not to make model outputs as similar as possible but
to show the discrepancies that occur among CTM systems despite using similar
input data. Different CTM systems were asked the same question to find the
impact of shipping, for which they got the same emissions as input data.
Nevertheless, to achieve less-varying results in future studies, the
vertical emission distribution as well as the boundary conditions could be
the same in all CTM inputs. This can help to make the modeled output more
similar. Adjustments in using the same meteorology could also be helpful, yet
difficult to realize, since the meteorology and meteorological driver within
each CTM system are closely connected. To gain more insight into certain
mechanisms, one model could be used with, e.g., changing vertical profiles,
emissions, or meteorology. Furthermore, the present study does not use the
same boundary conditions, and the models also do not use the same sea salt or dust
emissions. For more consistent investigations of model results, future
intercomparison studies should be carried out using the same boundary
conditions as well as sea salt and dust emissions as input data.
Regional-scale models with relatively coarse grid resolutions do not account
for chemical transformation mechanisms within a ship's exhaust gas plume.
They typically assume direct dilution and neglect the in-plume chemistry at
high-pollutant concentration levels. To obtain more precise information
regarding the effects of shipping on particle concentrations, the particle size
distribution and the interaction mechanisms from plume to background
concentrations, as well as chemical transformations within ship plumes,
should be considered in future studies.
List of abbreviationsAbbreviationDescription1.5-D1.5-dimensionalASOAAnthropogenic secondary organic aerosolBSOABiogenic secondary organic aerosolCAMxComprehensive Air Quality Model with ExtensionsCB05Carbon bond mechanism 05CBM-IVCarbon bond mechanism (version IV)CMAQCommunity Multiscale Air Quality modelCTMChemical transport modelECMWFEuropean Centre for Medium-Range Weather ForecastsEEAEuropean Environment AgencyEMEPEuropean Monitoring and Evaluation Programme modelEUEuropean UnionGEOS-5Goddard Earth Observing System Model, Version 5GEOS DASGoddard Earth Observing System Data Assimilation SystemGNFRGridded Nomenclature for ReportingIFS-CAMSIntegrated Forecasting System – Copernicus Atmosphere Monitoring ServiceLMDz-INCALaboratoire de Météorologie Dynamique General Circulation Model – INteraction with Chemistry and AerosolsMARSModel for an Aerosol Reacting SystemMEGANModel of Emissions of Gases and Aerosols from NatureMEPCMarine Environment Protection CommitteeMERRAModern-Era Retrospective Analysis for Research and ApplicationsMSC-WMeteorological Synthesizing Centre – WestNMBNormalized mean biasNMVOCNon-methane volatile organic compoundPMParticulate matterPOAPrimary organic aerosolPSATParticulate source apportionment technologyRSpearman's correlation coefficientRADM–AQRegional Acid Deposition Model-Aqueous ChemistryRMSERoot mean square errorSCIPPERShipping Contributions to Inland Pollution – Push for the Enforcement of RegulationsSOASecondary organic aerosolSOAPSecondary Organic Aerosol ProcessorSTEAMShip Traffic Emission Assessment ModelTNONederlandse Organisatie voor Toegepast Natuurwetenschappelijk OnderzoekTPPMTotal primary particulate matterVBSVolatility basis setVOCVolatile organic compoundWHOWorld Health Organization
Domains and measurement stations. The red trapezoid displays the 12×12 km2 domain, the blue icons are the locations of the measurement
stations. On the bottom left the larger 36×36 km2 domain is
displayed. Map source: ArcGIS Pro 2.7.1, Esri (2020).
CAMx source code and documentation can be downloaded from https://camx-wp.azurewebsites.net/download/source/ (Ramboll Environment and Health, 2018) and the Chimere website (https://www.lmd.polytechnique.fr/chimere/2020_getcode.php, Cholakian et al., 2023). CMAQ version 5.2, which was used here, is
available at 10.5281/zenodo.1167892 (US EPA
Office of Research and Development, 2017). EMEP is available at 10.5281/zenodo.3647990 (EMEP MSC-W, 2020), LOTOS-EUROS is
available at https://lotos-euros.tno.nl/open-source-version/
(Schaap et al., 2005), and WPS/WRF is available from the WPS (2023; https://github.com/wrf-model/WPS) and UCAR/NCAR Earth System Laboratory (2023, 10.5065/D6MK6B4K). The COSMO software is available at
https://www.cosmo-model.org/content/support/software/default.htm#models
(COSMO, 2023) and ecmwf-ifs/ifs-scripts at
https://github.com/ecmwf-ifs (ECMWF IFS, 2023).
Data on measurement stations from EEA can be downloaded at https://discomap.eea.europa.eu/map/fme/AirQualityExport.htm (European Environment Agency, 2023). CTM model results are available upon request.
The supplement related to this article is available online at: https://doi.org/10.5194/acp-23-10163-2023-supplement.
Author contributions
LF: CMAQ model runs, evaluation and analysis of model results, preparation
and writing of the paper. MK: analysis of the results, revision of the text.
VM: supervision, analysis of the results, revision of the text. SO: CAMx and
CHIMERE model runs, discussion of the results. RK and JK: LOTOS-EUROS model
runs, land-based emissions data provision, discussion of the results. JM and
SJ: EMEP model runs, discussion of the results. JPJ and EM: STEAM model
runs, shipping emissions data provision, discussion of the results.
Competing interests
The contact author has declared that none of the authors has any competing interests.
Disclaimer
Publisher’s note: Copernicus Publications remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Acknowledgements
AtmoSud acknowledges the continuous support of CAMx by RAMBOLL and CHIMERE
by LMD.
The Community Multiscale Air Quality (CMAQ) modeling system is developed
and maintained by the US EPA. Its use is gratefully acknowledged. Ronny Petrik from Helmholtz-Zentrum Hereon (now at Marinekommando Deutsche Marine,
Rostock) is acknowledged for providing meteorology and boundary conditions
for the CMAQ runs.
The support from the Meteorological Synthesizing Center – West of EMEP at the Norwegian
Meteorological Institute, especially from Peter Wind and David Simpson, in the
implementation of emissions and meteorological fields used in this paper
into the open-source EMEP model is gratefully acknowledged.
Financial support
This research has been supported by the Horizon 2020 (grant no. 814893) program. The computations for the regional modeling using the EMEP model was enabled
by resources provided by the Swedish National Infrastructure for Computing
(SNIC), partially funded by the Swedish Research Council through grant
agreement no. 2018-05973.The article processing charges for this open-access publication were covered by the Helmholtz-Zentrum Hereon.
Review statement
This paper was edited by Fangqun Yu and reviewed by two anonymous referees.
ReferencesAgrawal, H., Welch, W. A., Miller, J. W., and Cockert, D. R.: Emission
measurements from a crude oil tanker at sea, Environ. Sci. Technol.,
42, 7098–7103, 10.1021/es703102y, 2008.Agrawal, H., Eden, R., Zhang, X., Fine, P. M., Katzenstein, A., Miller, J.
W., Ospital, J., Teffera, S., and Cocker, D. R.: Primary particulate matter
from ocean-going engines in the Southern California Air Basin, Environ. Sci.
Technol., 43, 5398–5402, 10.1021/es8035016, 2009.Agrawal, H., Welch, W. A., Henningsen, S., Miller, J. W., and Cocker, D. R.:
Emissions from main propulsion engine on container ship at sea, J. Geophys.
Res, 115, D23205, 10.1029/2009JD013346, 2010.Aksoyoglu, S., Baltensperger, U., and Prévôt, A. S. H.: Contribution of ship emissions to the concentration and deposition of air pollutants in Europe, Atmos. Chem. Phys., 16, 1895–1906, 10.5194/acp-16-1895-2016, 2016.Alfaro, S. D. and Gomes, L.: Modeling mineral aerosol production by wind
erosion: Emission intensities and aerosol size distributions in source
areas., J. Geophys. Res., 106, 18075–18084, 10.1029/2000jd900339, 2001.Appel, K. W., Napelenok, S. L., Foley, K. M., Pye, H. O. T., Hogrefe, C., Luecken, D. J., Bash, J. O., Roselle, S. J., Pleim, J. E., Foroutan, H., Hutzell, W. T., Pouliot, G. A., Sarwar, G., Fahey, K. M., Gantt, B., Gilliam, R. C., Heath, N. K., Kang, D., Mathur, R., Schwede, D. B., Spero, T. L., Wong, D. C., and Young, J. O.: Description and evaluation of the Community Multiscale Air Quality (CMAQ) modeling system version 5.1, Geosci. Model Dev., 10, 1703–1732, 10.5194/gmd-10-1703-2017, 2017.Astitha, M., Lelieveld, J., Abdel Kader, M., Pozzer, A., and de Meij, A.: Parameterization of dust emissions in the global atmospheric chemistry-climate model EMAC: impact of nudging and soil properties, Atmos. Chem. Phys., 12, 11057–11083, 10.5194/acp-12-11057-2012, 2012.Baldasano, J. M., Pay, M. T., Jorba, O., Gassó, S., and
Jiménez-Guerrero, P.: An annual assessment of air quality with the
CALIOPE modeling system over Spain, Sci. Total Environ., 409, 2163–2178,
10.1016/j.scitotenv.2011.01.041, 2011.Banzhaf, S., Schaap, M., Kerschbaumer, A., Reimer, E., Stern, R., van der
Swaluw, E., and Builtjes, P.: Implementation and evaluation of pH-dependent
cloud chemistry and wet deposition in the chemical transport model
REM-Calgrid, Atmos. Environ., 49, 378–390,
10.1016/j.atmosenv.2011.10.069, 2012.Barbu, A. L., Segers, A. J., Schaap, M., Heemink, A. W., and Builtjes,
P. J. H.: A multi-component data assimilation experiment directed to sulphur
dioxide and sulphate over Europe, Atmos. Environ., 43, 1622–1631,
10.1016/j.atmosenv.2008.12.005, 2009.Behera, S. N., Sharma, M., Aneja, V. P., and Balasubramanian, R.: Ammonia in the atmosphere: a
review on emission sources, atmospheric chemistry and deposition on
terrestrial bodies, Environ Sci. Pollut. R., 20, 8092–8131,
10.1007/s11356-013-2051-9, 2013.Berge, E. and Jakobsen, H. A.: A regional scale multilayer model for the calculation of long-term transport and deposition of air pollution in Europe, Tellus B, 50, 205–223, 10.3402/tellusb.v50i3.16097, 1998.Bergström, R., Denier van der Gon, H. A. C., Prévôt, A. S. H., Yttri, K. E., and Simpson, D.: Modelling of organic aerosols over Europe (2002–2007) using a volatility basis set (VBS) framework: application of different assumptions regarding the formation of secondary organic aerosol, Atmos. Chem. Phys., 12, 8499–8527, 10.5194/acp-12-8499-2012, 2012.Bieser, J., Aulinger, A., Matthias, V., Quante, M., and van der Denier Gon,
H. A. C.: Vertical emission profiles for Europe based on plume rise
calculations, Environ. Pollut., 159, 2935–2946,
10.1016/j.envpol.2011.04.030, 2011.Binkowski, F. S. and Shankar, U.: The Regional Particulate Matter Model: 1.
Model description and preliminary results, J. Geophys. Res., 100,
26191–26209, 10.1029/95JD02093, 1995.Binkowski, F. S. and Roselle, S. J.: Models-3 Community Multiscale Air
Quality (CMAQ) model aerosol component 1. Model description, J. Geophys.
Res., 108, 4183, 10.1029/2001JD001409, 2003.Byun, D. and Schere, K. L.: Review of the Governing Equations, Computational
Algorithms, and Other Components of the Models-3 Community Multiscale Air
Quality (CMAQ) Modeling System, Appl. Mech. Rev., 2, 51–77,
10.1115/1.2128636, 2006.Carlton, A. G., Bhave, P. V., Napelenok, S. L., Edney, E. O., Sarwar, G.,
Pinder, R. W., Pouliot, G. A., and Houyoux, M.: Model representation of
secondary organic aerosol in CMAQv4.7, Appl. Mech. Rev., 59, 51–77,
10.1021/es100636q, 2010.Celik, S., Drewnick, F., Fachinger, F., Brooks, J., Darbyshire, E., Coe, H., Paris, J.-D., Eger, P. G., Schuladen, J., Tadic, I., Friedrich, N., Dienhart, D., Hottmann, B., Fischer, H., Crowley, J. N., Harder, H., and Borrmann, S.: Influence of vessel characteristics and atmospheric processes on the gas and particle phase of ship emission plumes: in situ measurements in the Mediterranean Sea and around the Arabian Peninsula, Atmos. Chem. Phys., 20, 4713–4734, 10.5194/acp-20-4713-2020, 2020.Chen, R., Hu, B., Liu, Y., Xu, J., Yang, G., Xu, D., and Chen, C.: Beyond
PM2.5: The role of ultrafine particles on adverse health effects of air
pollution, Biochim. Biophys. Acta, 1860, 2844–2855,
10.1016/j.bbagen.2016.03.019, 2016.Cholakian, A., Mailler, S., Pennel, R., Valari, M., Couvidat, F., Menut, L., and Siour, G.: CHIMERE v2020r3 version (+ WRF 3.7.1 + OASIS-MCT3), model code, https://www.lmd.polytechnique.fr/chimere/2020_getcode.php, last access: 19 January 2023.Clappier, A., Thunis, P., Beekmann, M., Putaud, J. P., and Meij, A. de:
Impact of SOx, NOx and NH3 emission reductions on PM2.5 concentrations
across Europe: Hints for future measure development, Environ. Int., 156,
106699, 10.1016/j.envint.2021.106699, 2021.
Corbett, J. J. and Fischbeck, P.: Emissions from ships, Science, 278, 5339,
823–824, 1997.COSMO Consortium: Models Supported by COSMO, COSMO [model], https://www.cosmo-model.org/content/support/software/default.htm#models, last access: 7 September 2023.Donahue, N. M., Robinson, A. L., and Pandis, S. N.: Atmospheric organic
particulate matter: From smoke to secondary organic aerosol, Atmos. Environ.,
43, 94–106, 10.1016/j.atmosenv.2008.09.055, 2009.Donateo, A., Gregoris, E., Gambaro, A., Merico, E., Giua, R., Nocioni, A.,
and Contini, D.: Contribution of harbour activities and ship traffic to
PM2.5, particle number concentrations and PAHs in a port city of the
Mediterranean Sea (Italy), Environ. Sci. Pollut. R., 21, 9415–9429,
10.1007/s11356-014-2849-0, 2014.ECMWF IFS: ERA-Interim Reanalysis Data, GitHub [data set, code], https://github.com/ecmwf-ifs/era-interim (last access: 7 September 2023), 2023.
Emberson, L. D., Simpson, D., Tuovinen, J.-P., Ashmore, M. R. and Cambridge,
H. M.: Towards a Model of Ozone Deposition and Stomatal Uptake over
Europe, EMEP/MSC-W Note 6/00, Norwegian Meteorological Institute, Oslo, 57,
2000.EMEP MSC-W.: metno/emep-ctm: OpenSource rv4.34 (202001) (rv4_34), Zenodo [model code], 10.5281/zenodo.3647990, 2020.EMEP Status Report 1/2022: Transboundary particulate matter,
photo-oxidants, acidifying and eutrophying components, Joint MSC-W & CCC
& CEIP & CIAM Report, https://emep.int/publ/reports/2022/EMEP_Status_Report_1_2022.pdf (last access: 7 September 2023),
2022.Erisman, J. W., van Pul, A., and Wyers, P.: Parametrization of surface
resistance for the quantification of atmospheric deposition of acidifying
pollutants and ozone, Atmos. Environ., 28, 2595–2607,
10.1016/1352-2310(94)90433-2, 1994.Esri: ArcGIS Pro 2.7.1, Esri, https://www.esri.com/en-us/arcgis/products/arcgis-pro/overview (last access: 7 September 2023), 2020.EU DIRECTIVE 2008/50/EC: European parliament: Directive 2008/50/EC of the European Parliament and of the Council of 21 May 2008 on ambient air quality and cleaner air for Europe, http://data.europa.eu/eli/dir/2008/50/oj (last access: 6 September 2023), 2008.European Environment Agency: Air Quality Data Download Service, Discomap, European Environment Agency [data set], https://discomap.eea.europa.eu/map/fme/AirQualityExport.htm (last access: 7 September 2023), 2023.Eyring, V.: Emissions from international shipping: 1. The last 50 years, J.
Geophys. Res., 110, D17305, 10.1029/2004JD005619, 2005.Eyring, V., Stevenson, D. S., Lauer, A., Dentener, F. J., Butler, T., Collins, W. J., Ellingsen, K., Gauss, M., Hauglustaine, D. A., Isaksen, I. S. A., Lawrence, M. G., Richter, A., Rodriguez, J. M., Sanderson, M., Strahan, S. E., Sudo, K., Szopa, S., van Noije, T. P. C., and Wild, O.: Multi-model simulations of the impact of international shipping on Atmospheric Chemistry and Climate in 2000 and 2030, Atmos. Chem. Phys., 7, 757–780, 10.5194/acp-7-757-2007, 2007.Fécan, F., Marticorena, B., and Bergametti, G.: Parametrization of the increase of the aeolian erosion threshold wind friction velocity due to soil moisture for arid and semi-arid areas, Ann. Geophys., 17, 149–157, 10.1007/s00585-999-0149-7, 1999.Fink, L., Karl, M., Matthias, V., Oppo, S., Kranenburg, R., Kuenen, J., Moldanova, J., Jutterström, S., Jalkanen, J.-P., and Majamäki, E.: Potential impact of shipping on air pollution in the Mediterranean region – a multimodel evaluation: comparison of photooxidants NO2 and O3, Atmos. Chem. Phys., 23, 1825–1862, 10.5194/acp-23-1825-2023, 2023.Folberth, G. A., Hauglustaine, D. A., Lathière, J., and Brocheton, F.: Interactive chemistry in the Laboratoire de Météorologie Dynamique general circulation model: model description and impact analysis of biogenic hydrocarbons on tropospheric chemistry, Atmos. Chem. Phys., 6, 2273–2319, 10.5194/acp-6-2273-2006, 2006.Foley, K. M., Roselle, S. J., Appel, K. W., Bhave, P. V., Pleim, J. E., Otte, T. L., Mathur, R., Sarwar, G., Young, J. O., Gilliam, R. C., Nolte, C. G., Kelly, J. T., Gilliland, A. B., and Bash, J. O.: Incremental testing of the Community Multiscale Air Quality (CMAQ) modeling system version 4.7, Geosci. Model Dev., 3, 205–226, 10.5194/gmd-3-205-2010, 2010.Fountoukis, C. and Nenes, A.: ISORROPIA II: a computationally efficient thermodynamic equilibrium model for K+–Ca2+–Mg2+–NH4+–Na+–SO42−–NO3−–Cl−–H2O aerosols, Atmos. Chem. Phys., 7, 4639–4659, 10.5194/acp-7-4639-2007, 2007.Friedrich, N., Eger, P., Shenolikar, J., Sobanski, N., Schuladen, J., Dienhart, D., Hottmann, B., Tadic, I., Fischer, H., Martinez, M., Rohloff, R., Tauer, S., Harder, H., Pfannerstill, E. Y., Wang, N., Williams, J., Brooks, J., Drewnick, F., Su, H., Li, G., Cheng, Y., Lelieveld, J., and Crowley, J. N.: Reactive nitrogen around the Arabian Peninsula and in the Mediterranean Sea during the 2017 AQABA ship campaign, Atmos. Chem. Phys., 21, 7473–7498, 10.5194/acp-21-7473-2021, 2021.Gao, R. and Sang, N.: Quasi-ultrafine particles promote cell metastasis via
HMGB1-mediated cancer cell adhesion, Environ. Pollut., 256, 113390,
10.1016/j.envpol.2019.113390, 2020.Gašparac, G., Jeričević, A., Kumar, P., and Grisogono, B.: Regional-scale modelling for the assessment of atmospheric particulate matter concentrations at rural background locations in Europe, Atmos. Chem. Phys., 20, 6395–6415, 10.5194/acp-20-6395-2020, 2020.Ge, Y., Heal, M. R., Stevenson, D. S., Wind, P., and Vieno, M.: Evaluation of global EMEP MSC-W (rv4.34) WRF (v3.9.1.1) model surface concentrations and wet deposition of reactive N and S with measurements, Geosci. Model Dev., 14, 7021–7046, 10.5194/gmd-14-7021-2021, 2021.Ginoux, P., Chin, M., Tegen, I., Prospero, J. M., Holben, B., Dubovik, O.,
and Lin, S.-J.: Sources and distributions of dust aerosols simulated
with the GOCART model, J. Geophys. Res., 106, 20255–20273,
10.1029/2000JD000053, 2001.
Gomes, L., Rajot, J. L., Alfaro, S. C., and Gaudichet, A.: Validation of a
dust production model from measurements performed in semi-arid agricultural
areas of Spain and Niger, Catena, 52, 257–271, 2003.Granier, C., Darras, S., van der Denier Gon, H., Doubalova, J., Elguindi,
N., Galle, B., Gauss, M., Guevara, M., Jalkanen, J.-P., Kuenen, J., Liousse,
C., Quack, B., Simpson, D., and Sindelarova, K.: The Copernicus Atmosphere
Monitoring Service global and regional emissions: (April 2019 version),
Copernicus Atmosphere Monitoring Service (CAMS) report,
10.24380/d0bn-kx16, 2019.Guenther, A., Zimmerman, P., Harley, P., Monson, R., and Fall, R.: Isoprene and monoterpene rate variability: model evaluations and sensitivity analyses, J. Geophys. Res., 98, 12609–12617, 10.1029/93JD00527, 1993.Guenther, A., Hewitt, C., Erickson, D., Fall, R., Geron, C., Graedel, T., Harley, P., Klinger, L., Lerdau, M., McKay, W., Pierce, T., Scholes, R., Steinbrecher, R., Tallamraju, R., Taylor, J., and Zimmerman, P.: A global model of natural volatile organic compound emissions, J. Geophys. Res., 100, 8873–8892, 10.1029/94JD02950, 1995.Hauglustaine, D. A., Balkanski, Y., and Schulz, M.: A global model simulation of present and future nitrate aerosols and their direct radiative forcing of climate, Atmos. Chem. Phys., 14, 11031–11063, 10.5194/acp-14-11031-2014, 2014.Henzing, J. S., Olivié, D. J. L., and van Velthoven, P. F. J.: A parameterization of size resolved below cloud scavenging of aerosols by rain, Atmos. Chem. Phys., 6, 3363–3375, 10.5194/acp-6-3363-2006, 2006.Heusinkveld, H. J., Wahle, T., Campbell, A., Westerink, R. H. S., Tran, L.,
Johnston, H., Stone, V., Cassee, F. R., and Schins, R. P. F.:
Neurodegenerative and neurological disorders by small inhaled particles,
NeuroToxicology, 56, 94–106, 10.1016/j.neuro.2016.07.007,
2016.Im, U., Bianconi, R., Solazzo, E., Kioutsioukis, I., Badia, A., Balzarini, A., Baró, R., Bellasio, R., Brunner, D., Chemel, C., Curci, G., van der Denier Gon, H., Flemming, J., Forkel, R., Giordano, L., Jiménez-Guerrero, P., Hirtl, M., Hodzic, A., Honzak, L., Jorba, O., Knote, C., Makar, P. A., Manders-Groot, A., Neal, L., Pérez, J. L., Pirovano, G., Pouliot, G., San Jose, R., Savage, N., Schroder, W., Sokhi, R. S., Syrakov, D., Torian, A., Tuccella, P., Wang, K., Werhahn, J., Wolke, R., Zabkar, R., Zhang, Y., Zhang, J., Hogrefe, C., and Galmarini, S.: Evaluation of operational online-coupled regional air quality models over Europe and North America in the context of AQMEII phase 2. Part II: Particulate matter, Atmos. Environ. 17, 421–441, 10.1016/j.atmosenv.2014.08.072, 2014.Im, U., Bianconi, R., Solazzo, E., Kioutsioukis, I., Badia, A., Balzarini,
A., Baró, R., Bellasio, R., Brunner, D., Chemel, C., Curci, G., van der
Denier Gon, H., Flemming, J., Forkel, R., Giordano, L.,
Jiménez-Guerrero, P., Hirtl, M., Hodzic, A., Honzak, L., Jorba, O.,
Knote, C., Makar, P. A., Manders-Groot, A., Neal, L., Pérez, J. L.,
Pirovano, G., Pouliot, G., San Jose, R., Savage, N., Schroder, W., Sokhi, R.
S., Syrakov, D., Torian, A., Tuccella, P., Wang, K., Werhahn, J., Wolke, R.,
Zabkar, R., Zhang, Y., Zhang, J., Hogrefe, C., and Galmarini, S.: Evaluation
of operational online-coupled regional air quality models over Europe and
North America in the context of AQMEII phase 2. Part II: Particulate matter,
Atmos. Environ., 115, 421–441,
10.1016/j.atmosenv.2014.08.072, 2015.IMO (International Maritime Organization): Marine Environment Protection
Committee (MEPC) – 79th session, 12–16 December 2022,
https://www.imo.org/en/MediaCentre/MeetingSummaries/Pages/MEPC-79th-session.aspx
(last access: 16 January 2023), 2022.Jägerbrand, A. K., Brutemark, A., Barthel Svedén, J., and Gren,
I.-M.: A review on the environmental impacts of shipping on aquatic and
nearshore ecosystems, Sci. Total Environ., 695, 133637,
10.1016/j.scitotenv.2019.133637, 2019.Jalkanen, J.-P., Brink, A., Kalli, J., Pettersson, H., Kukkonen, J., and Stipa, T.: A modelling system for the exhaust emissions of marine traffic and its application in the Baltic Sea area, Atmos. Chem. Phys., 9, 9209–9223, 10.5194/acp-9-9209-2009, 2009.Jalkanen, J.-P., Johansson, L., Kukkonen, J., Brink, A., Kalli, J., and Stipa, T.: Extension of an assessment model of ship traffic exhaust emissions for particulate matter and carbon monoxide, Atmos. Chem. Phys., 12, 2641–2659, 10.5194/acp-12-2641-2012, 2012.Johansson, L., Jalkanen, J.-P., Kalli, J., and Kukkonen, J.: The evolution of shipping emissions and the costs of regulation changes in the northern EU area, Atmos. Chem. Phys., 13, 11375–11389, 10.5194/acp-13-11375-2013, 2013.Johansson, L., Jalkanen, J.-P., and Kukkonen, J.: Global assessment of
shipping emissions in 2015 on a high spatial and temporal resolution, Atmos.
Environ., 167, 403–415, 10.1016/j.atmosenv.2017.08.042,
2017.Jiang, W., Smyth, S., Giroux, É., Roth, H., and Yin, D.: Differences
between CMAQ fine mode particle and PM2.5 concentrations and their impact on
model performance evaluation in the lower Fraser valley, Atmos.
Environ., 40, 4973–4985,
10.1016/j.atmosenv.2005.10.069, 2006.Jimenez, J. L., Canagaratna, M. R., Donahue, N. M., Prevot, A. S. H., Zhang,
Q., Kroll, J. H., DeCarlo, P. F., Allan, J. D., Coe, H., Ng, N. L., Aiken,
A. C., Docherty, K. S., Ulbrich, I. M., Grieshop, A. P., Robinson, A. L.,
Duplissy, J., Smith, J. D., Wilson, K. R., Lanz, V. A., Hueglin, C., Sun, Y.
L., Tian, J., Laaksonen, A., Raatikainen, T., Rautiainen, J., Vaattovaara,
P., Ehn, M., Kulmala, M., Tomlinson, J. M., Collins, D. R., Cubison, M. J.,
Dunlea, E. J., Huffman, J. A., Onasch, T. B., Alfarra, M. R., Williams, P.
I., Bower, K., Kondo, Y., Schneider, J., Drewnick, F., Borrmann, S., Weimer,
S., Demerjian, K., Salcedo, D., Cottrell, L., Griffin, R., Takami, A.,
Miyoshi, T., Hatakeyama, S., Shimono, A., Sun, J. Y., Zhang, Y. M., Dzepina,
K., Kimmel, J. R., Sueper, D., Jayne, J. T., Herndon, S. C., Trimborn, A.
M., Williams, L. R., Wood, E. C., Middlebrook, A. M., Kolb, C. E.,
Baltensperger, U., and Worsnop, D. R.: Evolution of organic aerosols in the
atmosphere, Science, 326, 1525–1529,
10.1126/science.1180353, 2009.Jonson, J. E., Gauss, M., Schulz, M., Jalkanen, J.-P., and Fagerli, H.: Effects of global ship emissions on European air pollution levels, Atmos. Chem. Phys., 20, 11399–11422, 10.5194/acp-20-11399-2020, 2020.Karamfilova, E.: BRIEFING Implementation Appraisal, Revision of the EU
Ambient Air Quality Directives, EPRS European Parliamentary Research
Service,
https://www.europarl.europa.eu/RegData/etudes/BRIE/2022/734679/EPRS_BRI(2022)734679_EN.pdf (last access: 7 September 2023), 2022.Karl, M., Bieser, J., Geyer, B., Matthias, V., Jalkanen, J.-P., Johansson, L., and Fridell, E.: Impact of a nitrogen emission control area (NECA) on the future air quality and nitrogen deposition to seawater in the Baltic Sea region, Atmos. Chem. Phys., 19, 1721–1752, 10.5194/acp-19-1721-2019, 2019.Kelly, J. T., Bhave, P. V., Nolte, C. G., Shankar, U., and Foley, K. M.: Simulating emission and chemical evolution of coarse sea-salt particles in the Community Multiscale Air Quality (CMAQ) model, Geosci. Model Dev., 3, 257–273, 10.5194/gmd-3-257-2010, 2010.Kiesewetter, G., Schoepp, W., Heyes, C., and Amann, M.: Modelling PM2.5
impact indicators in Europe: Health effects and legal compliance, Environ.
Modell Softw., 74, 201–211, 10.1016/j.envsoft.2015.02.022,
2015.Klingmüller, K., Metzger, S., Abdelkader, M., Karydis, V. A., Stenchikov, G. L., Pozzer, A., and Lelieveld, J.: Revised mineral dust emissions in the atmospheric chemistry–climate model EMAC (MESSy 2.52 DU_Astitha1 KKDU2017 patch), Geosci. Model Dev., 11, 989–1008, 10.5194/gmd-11-989-2018, 2018.Kleijn, D., Kohler, F., Báldi, A., Batáry, P., Concepción, E.
D., Clough, Y., Díaz, M., Gabriel, D., Holzschuh, A., Knop, E.,
Kovács, A., Marshall, E. J. P., Tscharntke, T., and Verhulst, J.: On the
relationship between farmland biodiversity and land-use intensity in Europe,
P. R. Soc., 276, 1658, 10.1098/rspb.2008.1509, 2009.Klimont, Z., Kupiainen, K., Heyes, C., Purohit, P., Cofala, J., Rafaj, P., Borken-Kleefeld, J., and Schöpp, W.: Global anthropogenic emissions of particulate matter including black carbon, Atmos. Chem. Phys., 17, 8681–8723, 10.5194/acp-17-8681-2017, 2017.Kranenburg, R., Segers, A. J., Hendriks, C., and Schaap, M.: Source apportionment using LOTOS-EUROS: module description and evaluation, Geosci. Model Dev., 6, 721–733, 10.5194/gmd-6-721-2013, 2013.Krupa, S. V.: Effects of atmospheric ammonia (NH3) on terrestrial vegetation:
a review, Environ. Pollut., 124, 179–221,
10.1016/s0269-7491(02)00434-7, 2003.Kuenen, J., Dellaert, S., Visschedijk, A., Jalkanen, J.-P., Super, I., and Denier van der Gon, H.: CAMS-REG-v4: a state-of-the-art high-resolution European emission inventory for air quality modelling, Earth Syst. Sci. Data, 14, 491–515, 10.5194/essd-14-491-2022, 2022.Kuenen, J. J. P., Visschedijk, A. J. H., Jozwicka, M., and Denier van der Gon, H. A. C.: TNO-MACC_II emission inventory; a multi-year (2003–2009) consistent high-resolution European emission inventory for air quality modelling, Atmos. Chem. Phys., 14, 10963–10976, 10.5194/acp-14-10963-2014, 2014.Loosmore, G. A. and Cederwall, R. T.: Precipitation scavenging of
atmospheric aerosols for emergency response applications: testing an updated
model with new real-time data, Atmos. Environ., 38, 993–1003,
10.1016/j.atmosenv.2003.10.055, 2004.Mallet, M. D., D'Anna, B., Même, A., Bove, M. C., Cassola, F., Pace, G., Desboeufs, K., Di Biagio, C., Doussin, J.-F., Maille, M., Massabò, D., Sciare, J., Zapf, P., di Sarra, A. G., and Formenti, P.: Summertime surface PM1 aerosol composition and size by source region at the Lampedusa island in the central Mediterranean Sea, Atmos. Chem. Phys., 19, 11123–11142, 10.5194/acp-19-11123-2019, 2019.Manders, A. M. M., Builtjes, P. J. H., Curier, L., Denier van der Gon, H. A. C., Hendriks, C., Jonkers, S., Kranenburg, R., Kuenen, J. J. P., Segers, A. J., Timmermans, R. M. A., Visschedijk, A. J. H., Wichink Kruit, R. J., van Pul, W. A. J., Sauter, F. J., van der Swaluw, E., Swart, D. P. J., Douros, J., Eskes, H., van Meijgaard, E., van Ulft, B., van Velthoven, P., Banzhaf, S., Mues, A. C., Stern, R., Fu, G., Lu, S., Heemink, A., van Velzen, N., and Schaap, M.: Curriculum vitae of the LOTOS–EUROS (v2.0) chemistry transport model, Geosci. Model Dev., 10, 4145–4173, 10.5194/gmd-10-4145-2017, 2017.Manders-Groot, A., Segers, A., and Jonkers, S.: LOTOS-EUROS v2.0 Reference
Guide, Report,
https://airqualitymodeling.tno.nl/publish/pages/3175/lotos-euros-reference-guide.pdf (last access: 7 September 2023), 2016.Marmer, E. and Langmann, B.: Impact of ship emissions on the Mediterranean
summertime pollution and climate: A regional model study, Atmos. Environ.,
39, 4659–4669, 10.1016/j.atmosenv.2005.04.014, 2005.Mårtensson, E. M., Nilsson, E. D., Leeuw, G. de, Cohen, L. H., and
Hansson, H.-C.: Laboratory simulations and parameterization of the primary
marine aerosol production, J. Geophys. Res., 108, 4297,
10.1029/2002JD002263, 2003.Marticorena, B. and Bergametti, G.: Modelling the atmospheric dust cycle: 1.
Design of a soil drived dust emission scheme., J. Geophys. Res., 100,
16415–16430, 10.1029/95jd00690, 1995.
Marticorena, B., Bergametti, G., Aumont, B., Callot, Y.,
N'Doumé, C., and Legrand, M.: Modelling the atmospheric dust
cycle: 2. Simulation of Saharan dust sources, J. Geophys. Res., 102, 4387–4404, 1997.Matthias, V.: The aerosol distribution in Europe derived with the Community Multiscale Air Quality (CMAQ) model: comparison to near surface in situ and sunphotometer measurements, Atmos. Chem. Phys., 8, 5077–5097, 10.5194/acp-8-5077-2008, 2008.Matthias, V., Bewersdorff, I., Aulinger, A., and Quante, M.: The
contribution of ship emissions to air pollution in the North Sea regions,
Environ. Pollut., 158, 2241–2250,
10.1016/j.envpol.2010.02.013, 2010.Menut, L., Bessagnet, B., Khvorostyanov, D., Beekmann, M., Blond, N., Colette, A., Coll, I., Curci, G., Foret, G., Hodzic, A., Mailler, S., Meleux, F., Monge, J.-L., Pison, I., Siour, G., Turquety, S., Valari, M., Vautard, R., and Vivanco, M. G.: CHIMERE 2013: a model for regional atmospheric composition modelling, Geosci. Model Dev., 6, 981–1028, 10.5194/gmd-6-981-2013, 2013.Monahan, E. C., Spiel, D. E., and Davidson, K. L.: A Model of Marine Aerosol
Generation Via Whitecaps and Wave Disruption, OCSL,
2, 10.1007/978-94-009-4668-2_16, 1986.Nenes, A., Pandis, S. N., and Pilinis, C.: ISORROPIA: A New Thermodynamic
Equilibrium Model for Multiphase Multicomponent Inorganic Aerosols, Aquat.
Geochem., 4, 123–152, 10.1023/A:1009604003981, 1998.Nunes, R. A. O., Alvim-Ferraz, M. C. M., Martins, F. G., Calderay-Cayetano, F., Durán-Grados, V., Moreno-Gutiérrez, J., Jalkanen, J.-P., Hannuniemi, H., and Sousa, S. I. V.: Shipping emissions in the Iberian Peninsula and the impacts on air quality, Atmos. Chem. Phys., 20, 9473–9489, 10.5194/acp-20-9473-2020, 2020.Ovadnevaite, J., Manders, A., de Leeuw, G., Ceburnis, D., Monahan, C., Partanen, A.-I., Korhonen, H., and O'Dowd, C. D.: A sea spray aerosol flux parameterization encapsulating wave state, Atmos. Chem. Phys., 14, 1837–1852, 10.5194/acp-14-1837-2014, 2014.Palacios-Peña, L., Lorente-Plazas, R., Montávez, J. P., and
Jiménez-Guerrero, P.: Saharan Dust Modeling Over the Mediterranean Basin
and Central Europe: Does the Resolution Matter?, Front. Earth Sci., 7, 290,
10.3389/feart.2019.00290, 2019.Pandolfi, M., Gonzalez-Castanedo, Y., Alastuey, A., La Rosa, J. D. de,
Mantilla, E., La Campa, A. S. de, Querol, X., Pey, J., Amato, F., and
Moreno, T.: Source apportionment of PM10 and PM2.5 at multiple sites in
the strait of Gibraltar by PMF: impact of shipping emissions, Environ. Sci.
Pollut. R., 18, 260–269, 10.1007/s11356-010-0373-4, 2011.Pepe, N., Pirovano, G., Balzarini, A., Toppetti, A., Riva, G. M., Amato, F.,
and Lonati, G.: Enhanced CAMx source apportionment analysis at an urban
receptor in Milan based on source categories and emission regions, Atmos.
Environ. X, 2, 100020, 10.1016/j.aeaoa.2019.100020, 2019.Pleim, J. E., Xiu, A., Finkelstein, P. L., and Otte, T. L.: A Coupled
Land-Surface and Dry Deposition Model and Comparison to Field Measurements
of Surface Heat, Moisture, and Ozone Fluxes, Water Air Soil Poll.-Focus,
1, 243–252, 10.1023/A:1013123725860, 2001.Prati, M. V., Costagliola, M. A., Quaranta, F., and Murena, F.: Assessment
of ambient air quality in the port of Naples, J. Air Waste Manage.,
65, 970–979, 10.1080/10962247.2015.1050129, 2015.Pun, B. K., Seigneur, C., and Lohman, K.: Modeling secondary organic aerosol formation via multiphase partitioning with molecular data, Environ. Sci. Technol., 40, 4722–4731, 10.1021/es0522736, 2006.Pye, H. O. T. and Pouliot, G. A.: Modeling the role of alkanes, polycyclic
aromatic hydrocarbons, and their oligomers in secondary organic aerosol
formation, Environ. Sci. Technol., 46, 6041–6047,
10.1021/es300409w, 2012.Pye, H. O. T., Murphy, B. N., Xu, L., Ng, N. L., Carlton, A. G., Guo, H., Weber, R., Vasilakos, P., Appel, K. W., Budisulistiorini, S. H., Surratt, J. D., Nenes, A., Hu, W., Jimenez, J. L., Isaacman-VanWertz, G., Misztal, P. K., and Goldstein, A. H.: On the implications of aerosol liquid water and phase separation for organic aerosol mass, Atmos. Chem. Phys., 17, 343–369, 10.5194/acp-17-343-2017, 2017.Ramboll Environment and Health: CAMx Source Code and Documentation, CAMx v6.50 (April 30, 2018), https://camx-wp.azurewebsites.net/download/source/ (last access: 7 September 2023), 2018.Ramboll Environment and Health: User's Guide COMPREHENSIVE AIR QUALITY MODEL
WITH EXTENSIONS: Version 7.10, User's Guide,
https://camx-wp.azurewebsites.net/Files/CAMxUsersGuide_v7.10.pdf (last access: 7 September 2023), 2020.Reichle, L. J., Cook, R., Yanca, C. A., and Sonntag, D. B.: Development of
organic gas exhaust speciation profiles for nonroad spark-ignition and
compression-ignition engines and equipment, J. Air Waste Manage.,
65, 1185–1193, 10.1080/10962247.2015.1020118, 2015.Remke, E., Brouwer, E., Kooijman, A., Blindow, I., Esselink, H., and
Roelofs, J. G. M.: Even low to medium nitrogen deposition impacts vegetation
of dry, coastal dunes around the Baltic Sea, Environ. Pollut., 157,
792–800, 10.1016/j.envpol.2008.11.020, 2009.Riccio, A., Ciaramella, A., Giunta, G., Galmarini, S., Solazzo, E., and
Potempski, S.: On the systematic reductionof data complexity in multimodel
atmospheric dispersion ensemble modeling, J. Geophys. Res., 117, D05314,
10.1029/2011JD016503, 2012.Robinson, A. L., Donahue, N. M., Shrivastava, M. K., Weitkamp, E. A., Sage,
A. M., Grieshop, A. P., Lane, T. E., Pierce, J. R., and Pandis, S. N.:
Rethinking organic aerosols: semivolatile emissions and photochemical aging,
Science, 315, 1259—1262, 10.1126/science.1133061, 2007.Roselle, S. J. and Binkowski, F. S.: Cloud Dynamics and Chemistry, Chap. 11, https://www.cmascenter.org/cmaq/science_documentation/pdf/ch11.pdf (last access: 7 September 2023), 1999.Schaap, M., van Loon, M., ten Brink, H. M., Dentener, F. J., and Builtjes, P. J. H.: Secondary inorganic aerosol simulations for Europe with special attention to nitrate, Atmos. Chem. Phys., 4, 857–874, 10.5194/acp-4-857-2004, 2004.Schaap, M., Sauter, F., Boersen, G., and Builtjes, P.: The integration of LOTOS and EUROS: Activities during 2004. Tech. rep. Apeldoorn, The Netherlands (in Dutch): TNOreport B&O-A R2005/209, 2005 (data available at: https://lotos-euros.tno.nl/open-source-version/,
last access: 19 January 2023).Schaap, M., Manders, A. M. M., Hendriks, E. C. J., Cnossen, J. M., Segers,
A. J. S., Denier van der Gon, H. A. C., Jozwicka, M., Sauter, F., Velders, G.,
Matthijsen, J., and Builtjes, P. J. H.: Regional modelling of particulate matter
for the Netherlands, Tech. rep., Netherlands Environmental Assessment Agency
(PBL), https://www.pbl.nl/sites/default/files/downloads/500099008_0.pdf (last access: 7 September 2023), 2009.Schembari, C., Cavalli, F., Cuccia, E., Hjorth, J., Calzolai, G., Pérez,
N., Pey, J., Prati, P., and Raes, F.: Impact of a European directive on ship
emissions on air quality in Mediterranean harbours, Atmos. Environ.,
61, 661–669, 10.1016/j.atmosenv.2012.06.047, 2012.Schober, P., Boer, C., and Schwarte, L. A.: Correlation Coefficients:
Appropriate Use and Interpretation, Anesth. Analg., 126, 1763–1768,
10.1213/ANE.0000000000002864, 2018.
Seinfeld, J. H. and Pandis, S. N.: Atmospheric Chemistry and Physics, John Wiley and Sons, New York, ISBN 9781119221166, 1998.
Seinfeld, J. H. and Pandis, S. N.: Atmospheric Chemistry and Physics:
From Air Pollution to Climate Change. 2nd Edition, John Wiley & Sons, New
York, ISBN 978-1118947401, 2006.Simpson, D., Benedictow, A., Berge, H., Bergström, R., Emberson, L. D., Fagerli, H., Flechard, C. R., Hayman, G. D., Gauss, M., Jonson, J. E., Jenkin, M. E., Nyíri, A., Richter, C., Semeena, V. S., Tsyro, S., Tuovinen, J.-P., Valdebenito, Á., and Wind, P.: The EMEP MSC-W chemical transport model – technical description, Atmos. Chem. Phys., 12, 7825–7865, 10.5194/acp-12-7825-2012, 2012.Simpson, D., Bergström, R., Briolat, A., Imhof, H., Johansson, J., Priestley, M., and Valdebenito, A.: GenChem v1.0 – a chemical pre-processing and testing system for atmospheric modelling, Geosci. Model Dev., 13, 6447–6465, 10.5194/gmd-13-6447-2020, 2020.Sippula, O., Stengel, B., Sklorz, M., Streibel, T., Rabe, R., Orasche, J.,
Lintelmann, J., Michalke, B., Abbaszade, G., Radischat, C., Gröger, T.,
Schnelle-Kreis, J., Harndorf, H., and Zimmermann, R.: Particle emissions
from a marine engine: chemical composition and aromatic emission profiles
under various operating conditions, Environ. Sci. Technol., 48,
11721–11729, 10.1021/es502484z, 2014.Solazzo, E., Bianconi, R., Pirovano, G., Matthias, V., Vautard, R., Moran,
M. D., Wyat Appel, K., Bessagnet, B., Brandt, J., Christensen, J. H.,
Chemel, C., Coll, I., Ferreira, J., Forkel, R., Francis, X. V., Grell, G.,
Grossi, P., Hansen, A. B., Miranda, A. I., Nopmongcol, U., Prank, M.,
Sartelet, K. N., Schaap, M., Silver, J. D., Sokhi, R. S., Vira, J., Werhahn,
J., Wolke, R., Yarwood, G., Zhang, J., Rao, S. T., and Galmarini, S.:
Operational model evaluation for particulate matter in Europe and North
America in the context of AQMEII, Atmos. Environ., 53, 75–92,
10.1016/j.atmosenv.2012.02.045, 2012.Solazzo, E., Riccio, A., Kioutsioukis, I., and Galmarini, S.: Pauci ex tanto numero: reduce redundancy in multi-model ensembles, Atmos. Chem. Phys., 13, 8315–8333, 10.5194/acp-13-8315-2013, 2013.Solazzo, E., Riccio, A., van Dingenen, R., Valentini, L., and Galmarini, S.:
Evaluation and uncertainty estimation of the impact of air quality modelling
on crop yields and premature deaths using a multi-model ensemble, Sci. Total
Environ., 633, 1437–1452, 10.1016/j.scitotenv.2018.03.317,
2018.Song, S.-K. and Shon, Z.-H.: Current and future emission estimates of
exhaust gases and particles from shipping at the largest port in Korea,
Environ. Sci. Pollut. R., 21, 6612–6622,
10.1007/s11356-014-2569-5, 2014.Sotiropoulou, R. E. P. and Tagaris, E.: Impact of Shipping Emissions on
European Air Quality, in:
Perspectives on Atmospheric Sciences, edited by: Karacostas, T., Bais, A., and Nastos, P., Springer Atmospheric Sciences,
10.1007/978-3-319-35095-0_149, 2017.Strader, R., Lurmann, F., and Pandis, S. N.: Evaluation of secondary organic aerosol formation in winter, Atmos. Environ., 33, 4849–4863, 10.1016/S1352-2310(99)00310-6, 1999.Tadic, I., Crowley, J. N., Dienhart, D., Eger, P., Harder, H., Hottmann, B., Martinez, M., Parchatka, U., Paris, J.-D., Pozzer, A., Rohloff, R., Schuladen, J., Shenolikar, J., Tauer, S., Lelieveld, J., and Fischer, H.: Net ozone production and its relationship to nitrogen oxides and volatile organic compounds in the marine boundary layer around the Arabian Peninsula, Atmos. Chem. Phys., 20, 6769–6787, 10.5194/acp-20-6769-2020, 2020.Tomasi, C. and Lupi, A.: Primary and secondary sources of atmospheric aerosol, Atmospheric Aerosols: Life Cycles and Effects on Air Quality and Climate, Wiley, 1–86, 10.1002/9783527336449.ch1, 2017.Tuccella, P., Menut, L., Briant, R., Deroubaix, A., Khvorostyanov, D.,
Mailler, S., Siour, G., and Turquety, S.: Implementation of Aerosol-Cloud
Interaction within WRF-CHIMERE Online Coupled Model: Evaluation and
Investigation of the Indirect Radiative Effect from Anthropogenic Emission
Reduction on the Benelux Union, Atmosphere, 10, 20,
10.3390/atmos10010020, 2019.UCAR/NCAR Earth System Laboratory: WRF – Weather Research and Forecasting Model Users' Page, MMM/WRF [model code], 10.5065/D6MK6B4K, 2023.US EPA Office of Research and Development: CMAQ (5.2), Zenodo [model code], 10.5281/zenodo.1167892, 2017.Večeřa, Z., Mikuška, P., Smolík, J., Eleftheriadis, K.,
Bryant, C., Colbeck, I., and Lazaridis, M.: Shipboard Measurements of
Nitrogen Dioxide, Nitrous Acid, Nitric Acid and Ozone in the Eastern
Mediterranean Sea, Water Air Soil Poll.-Focus, 8, 117–125,
10.1007/s11267-007-9133-y, 2008.Viana, M., Amato, F., Alastuey, A., Querol, X., Moreno, T., Dos Santos, S.
G., Herce, M. D., and Fernández-Patier, R.: Chemical tracers of
particulate emissions from commercial shipping, Environ. Sci. Technol.,
43, 7472–7477, 10.1021/es901558t, 2009.Viana, M., Hammingh, P., Colette, A., Querol, X., Degraeuwe, B., Vlieger, I.
d., and van Aardenne, J.: Impact of maritime transport emissions on coastal
air quality in Europe, Atmos. Environ., 90, 96–105,
10.1016/j.atmosenv.2014.03.046, 2014.Viana, M., Rizza, V., Tobías, A., Carr, E., Corbett, J., Sofiev, M.,
Karanasiou, A., Buonanno, G., and Fann, N.: Estimated health impacts from
maritime transport in the Mediterranean region and benefits from the use of
cleaner fuels, Environ. Internat., 138, 105670,
10.1016/j.envint.2020.105670, 2020.
Wesely, M. L.: Parameterization of surface resistances to gaseous dry
deposition in regional-scale numerical models, Atmos. Environ., 23,
1293–1304, 10.1016/0004-6981(89)90153-4, 1989.
Whitten, G. Z., Hogo, H., and Killus, J. P.: The carbon-bond mechanism: A
condensed kinetic mechanism for photochemical smog, Environ. Sci. Technol.,
18, 280–287, 1980.Wiedinmyer, C., Akagi, S. K., Yokelson, R. J., Emmons, L. K., Al-Saadi, J. A., Orlando, J. J., and Soja, A. J.: The Fire INventory from NCAR (FINN): a high resolution global model to estimate the emissions from open burning, Geosci. Model Dev., 4, 625–641, 10.5194/gmd-4-625-2011, 2011.WPS: The Weather Research and Forecasting (WRF) Preprocessing System, GitHub [data set, code], https://github.com/wrf-model/WPS, last access: 7 September 2023.Zender, C., Bian, H., and Newman, D.: Mineral Dust Entrainmentand Deposition
(DEAD) model: Description and 1990s dust climatology, J. Geophys. Res.,
108, 4416–4437, 10.1029/2002jd002775, 2003.Zhang, L., Brook, J. R., and Vet, R.: A revised parameterization for gaseous dry deposition in air-quality models, Atmos. Chem. Phys., 3, 2067–2082, 10.5194/acp-3-2067-2003, 2003.Zhong, Q., Shen, H., Yun, X., Chen, Y., Ren, Y.'a., Xu, H., Shen, G., Du,
W., Meng, J., Li, W., Ma, J., and Tao, S.: Global Sulfur Dioxide Emissions
and the Driving Forces, Environ. Sci. Technol., 54, 6508–6517,
10.1021/acs.est.9b07696, 2020.