ABSTRACT: Aerosol emissions from international shipping are recognized to have a large impact on the Earth’s radiation budget, directly by scattering and absorbing solar radiation and indirectly by altering cloud properties. New regulations have recently been approved by the International Maritime Organi- zation (IMO) aiming at progressive reductions of the maximum sulfur content allowed in marine fuels from current 4.5% by mass down to 0.5% in 2020, with more restrictive limits already applied in some coastal regions. In this context, we use a global bottom-up algorithm to calculate geographically resolved emis- sion inventories of gaseous (NOx, CO, SO2) and aerosol (black carbon, organic matter, sulfate) species for different kinds of low-sulfur fuels in shipping. We apply these inventories to study the resulting changes in radiative forcing, attributed to particles from shipping, with the global aerosol-climate model EMAC-MADE. The emission factors for the different fuels are based on measurements at a test bed of a large diesel engine. We consider both fossil fuel (marine gas oil) and biofuels (palm and soy bean oil) as a substitute for heavy fuel oil in the current (2006) fleet and compare their climate impact to that resulting from heavy fuel oil use. Our simulations suggest that ship-induced surface level concentrations of sulfate aerosol are strongly reduced, up to about 40-60% in the high-traffic regions. This clearly has positive consequences for pollution reduction in the vicinity of major harbors. Additionally, such reductions in the aerosol loading lead to a decrease of a factor of 3-4 in the indirect global aerosol effect induced by emissions from international shipping.

1. Published: March 23, 2011
r 2011 American Chemical Society 3519 dx.doi.org/10.1021/es1036157 |Environ. Sci. Technol. 2011, 45, 3519–3525
ARTICLE
pubs.acs.org/est
Climate Impact of Biofuels in Shipping: Global Model Studies of the
Aerosol Indirect Effect
Mattia Righi,*,†
Carolin Klinger,†
Veronika Eyring,†
Johannes Hendricks,†
Axel Lauer,‡
and Andreas Petzold†
†
Deutsches Zentrum f€ur Luft- und Raumfahrt (DLR), Institut f€ur Physik der Atmosph€are, Oberpfaffenhofen, Germany
‡
International Pacific Research Center, University of Hawaii at Manoa, Honolulu, Hawaii 96822, United States
bS Supporting Information
1. INTRODUCTION
Ocean-going ships contribute significantly to the fuel con-
sumption of all transport related activities. According to ref 1, the
total fuel consumption of the shipping sector in 2000 was higher
than that for aviation (280 versus 207 Mt). Although shipping
contributes only about 16% of the total fuel consumption of all
transport sources, the lack of strict international regulations on
the composition of fuels burned by ships leads to relatively high
emissions of pollutants. In 2000, this sector was responsible for
the bulk of SO2 emissions among all transport modes, a factor of
3 higher than road traffic, and had large contributions to NOx and
CO2 emissions.
Ships also emit various kinds of particulate matter. Aerosols
have an important effect on climate, as they alter the Earth’s
radiation budget in several ways: scattering and absorption of
solar and terrestrial radiation (direct effect) and modification of
cloud properties such as, for instance, potential increase in the
concentration of cloud droplets and reduction of their size
leading to an increased cloud reflectivity and, consequently, to
an increased backscattering of solar radiation (first indirect
effect2
). Under the current regulations, ship traffic can contribute
significantly to the total anthropogenic aerosol indirect radiative
forcing effect resulting from an aerosol-induced increase of the
cloud albedo.3,4
This effect is of particular interest for low marine
clouds being much more effective over the dark oceanic surface
than over land. Furthermore, ship emissions are released in
relatively clean marine air masses with frequent low-level clouds.
Such clouds are more sensitive to the changes in aerosol
concentrations as demonstrated by satellite observations of ship
tracks.5,6
Particulate matter has also an adverse effect on human health:
in the case of shipping, this is especially important for the
population living along the coastlines because about 70% of all
ship emissions are released within 400 km off the coast.4
According to ref 7, shipping-related emissions of particulate
matter are responsible for approximately 60 000 premature
cardiopulmonary and lung cancer deaths annually (year 2002),
most of them occurring near the coastlines of Europe, East Asia,
and South Asia. Under year 2000 regulations, this number would
Received: October 26, 2010
Accepted: February 9, 2011
Revised: January 27, 2011
ABSTRACT: Aerosol emissions from international shipping
are recognized to have a large impact on the Earth’s radiation
budget, directly by scattering and absorbing solar radiation and
indirectly by altering cloud properties. New regulations have
recently been approved by the International Maritime Organi-
zation (IMO) aiming at progressive reductions of the maximum
sulfur content allowed in marine fuels from current 4.5% by
mass down to 0.5% in 2020, with more restrictive limits already
applied in some coastal regions. In this context, we use a global
bottom-up algorithm to calculate geographically resolved emis-
sion inventories of gaseous (NOx, CO, SO2) and aerosol (black
carbon, organic matter, sulfate) species for different kinds of
low-sulfur fuels in shipping. We apply these inventories to study the resulting changes in radiative forcing, attributed to particles from
shipping, with the global aerosol-climate model EMAC-MADE. The emission factors for the different fuels are based on
measurements at a test bed of a large diesel engine. We consider both fossil fuel (marine gas oil) and biofuels (palm and soy bean oil)
as a substitute for heavy fuel oil in the current (2006) fleet and compare their climate impact to that resulting from heavy fuel oil use.
Our simulations suggest that ship-induced surface level concentrations of sulfate aerosol are strongly reduced, up to about 40-60%
in the high-traffic regions. This clearly has positive consequences for pollution reduction in the vicinity of major harbors.
Additionally, such reductions in the aerosol loading lead to a decrease of a factor of 3-4 in the indirect global aerosol effect induced
by emissions from international shipping.

2. 3520 dx.doi.org/10.1021/es1036157 |Environ. Sci. Technol. 2011, 45, 3519–3525
Environmental Science & Technology ARTICLE
be expected to grow by 40% by 2012, but control scenarios would
reduce premature deaths by ∼43500, if the sulfur content is limited
to 0.1% within 200 nautical miles (∼370 km) of coastal areas, and by
∼41200 if the sulfur content is reduced globally to 0.5%.8
International ship traffic is regulated by the International
Maritime Organization (IMO). Emissions from ships are ad-
dressed by ANNEX VI of MARPOL 73/78 (the International
Convention for the Prevention of Pollution from Ships, 9
).
Regulation 14 of Annex VI sets a cap of 4.5% by mass in the
sulfur content of marine fuels and introduces the first two Sulfur
Emission Control Areas (SECA), in the Baltic and in the North
Sea, where this cap is currently set to 1.5%. More restrictive limits
will enter into force in 2012 (3.5%) and 2020 (0.5%). In the
SECAs, the cap will be reduced to 0.1% starting from 2015.
In view of these regulations, low-sulfur fuel will be more widely
used in the future than today. We therefore analyze the climate
impact of different kinds of low-sulfur fuels in shipping. As a
reference, we use a standard fuel inventory (composed of a
mixture of heavy-fuel oil (HFO) and marine-gas oil) and
introduce three new inventories in which the HFO component
is replaced by low-sulfur fuels: marine gas oil, a mixture of marine
gas oil and palm oil, and a mixture of marine gas oil and soy bean
oil. All emissions are calculated for the total fuel consumption of
ships in the year 2006 (321 Mt9
).
We use the ECHAM/MESSy Atmospheric Chemistry model
(EMAC10-12
) with the aerosol module MADE 13
in the same
configuration used by 3
in a previous study. In this particular
setup, the aerosols are coupled with radiation and clouds,
allowing studies of the impact of ships on the Earth’s radiation
budget by comparing the results with a simulation without ship
emissions. The aerosol species considered by the model are
sulfate (SO4), nitrate (NO3), ammonium (NH4), black carbon
(BC), particulate organic matter (POM), aerosol liquid water,
mineral dust, and sea salt.
In recent years, several different methods have been developed
to generate geographically resolved ship emission inventories on
both local and global scales, as well as to analyze the ship traffic
impact on atmospheric composition, air quality, human health,
and climate, with special attention to future scenarios and
mitigation strategies (see ref 4 and references therein for a recent
and comprehensive assessment on this subject). The present
study analyzes, for the first time, the global climate impact of
biofuels in shipping, related to changes in the global aerosol and
in the cloud microphysical properties. Issues connected with the
CO2 emission changes, fuel life cycle, fuel production, or possible
changes in land use are not covered here and will be the subject of
a follow-up study.
2. EMISSION INVENTORIES
In this work, we consider four emission inventories, which are
calculated using the global bottom-up SeaKLIM algorithm
developed by.14
This algorithm provides better spatial resolution
as compared to the previous top-down approaches and for the
first time has a global coverage,14
in contrast to existing regional
bottom-up approaches. The emission factors for the reference
inventory (hereafter REF), which consists of a mixture of HFO
and marine gas oil, are taken from refs 14 and 15. Following ref
14, the emission factors are averaged over nine ship classes
(container, tanker, general cargo, bulk carrier, reefer, roll-on/roll-
off, passenger, fishing, and miscellaneous). To study the climate
effect of substituting HFO by cleaner fuels (marine gas oil or
biofuels), we assume a complete replacement of HFO in all three
emission inventories (Table 1). In the inventory MGO, the HFO
component is completely replaced by marine gas oil: this implies
that the entire fleet is running with marine gas oil. In the other
inventories, the HFO component is replaced either by palm oil
(inventory PALM) or by soy bean oil (inventory SOY). The
marine gas oil component is not changed in these two inventories
and is, therefore, identical to the reference case (REF). The
emission factors used for inventories PALM and SOY are
representative for large container ships and were directly mea-
sured on engine test beds.16
They are then rescaled, to get
representative values for each of the ship types in the fleet (see
the Supporting Information for details).
These emission factors are used in the SeaKLIM algorithm to
calculate global emission inventories for gases (NOx, CO, SO2)
and aerosols (BC, POM, primary SO4). These geographically
resolved emissions inventories are subsequently used as input for
the global model, assuming that the species are released in the
lowest model level. The corresponding global emission totals are
summarized in Figure 1. We shall note that sulfate derives both
from direct emission of primary SO4 and from SO2 oxidation in
the model. The difference in the emission totals for biofuels with
respect to standard fuel is a consequence of their composition,
which is characterized by a low carbon content, a high oxygen
content, and a negligible sulfur content. Emissions of NOx are
almost unchanged because NOx formation is mainly controlled
by the engine combustion temperature.
Table 1. List of the Simulations Performed in This Worka
no. inventory fuel type size distribution
1 NO SHIPS
2 REF heavy fuel oil þ marine gas oil fresh
3 REF heavy fuel oil þ marine gas oil aged 1
4 REF heavy fuel oil þ marine gas oil aged 2
5 MGO marine gas oil aged 2
6 PALM palm oil þ marine gas oil aged 2
7 SOY soy bean oil þ marine gas oil aged 2
a
The parameters of the size distributions are given in Table S2.
Figure 1. Total annual emissions of aerosol (BC, POM, primary SO4)
and gaseous (SO2, NOx, CO) species from global ship traffic in 2006.
Units for NOx are Tg(NO2)/yr.

3. 3521 dx.doi.org/10.1021/es1036157 |Environ. Sci. Technol. 2011, 45, 3519–3525
Environmental Science & Technology ARTICLE
Additional assumptions have been made to split the emissions
of particulate matter into different aerosol modes: three different
size distributions (one for fresh17
and two for aged particles18
)
have been used to test the sensitivity of our results to the number
and size of the emitted particles (details about these distributions
are given in the Supporting Information and in Table S2).
In the analysis reported in ref 16, several types of biofuels were
investigated, including palm oil, soy bean oil, waste edible fat, and
sun flower oil. From this set of biofuels, only palm oil and soy
bean oil are of global availability and therefore of relevance for the
present study. Waste edible fat is used in stationary engines for
local power production only, while sun flower oil is only of local
interest close to the production region. The work of 16
showed
that all biofuels could be used directly in the marine diesel
engines without technical modifications. All fuels showed good
combustion conditions. If biofuels are to be considered for adop-
tion on global scale, other issues like production, distribution
network, and competition for land use will have to be addressed.
This is beyond the scope of the current work. We refer to refs
19-21 for more information on these subjects.
Detailed information on the generation of the emission
inventories and their application to the global model is provided
in the Supporting Information.
3. MODEL SIMULATIONS
The simulations performed in this study were run using
EMAC10-12
with the modal aerosol dynamics module MADE.13
A more detailed description of EMAC-MADE can be found in
the Supporting Information.
We use a T42 spectral horizontal resolution, corresponding to
a horizontal grid size of about 2.8° Â 2.8° (∼300 km at the
equator). The vertical grid has 19 nonequidistant vertical layers
from the surface up to 10 hPa (∼30 km). The simulated period
covers 6 model years, plus one spin-up year to allow aerosol
concentrations and other quantities in the troposphere to reach
an equilibrium state. To obtain significant results with such a
limited amount of simulated years, we minimize the dynamical
differences between the different runs. This is obtained by
constraining the model dynamics (prognostic equations for vorti-
city, divergence, temperature, and surface pressure) by operational
analysis data of the European Centre for Medium-Range Weather
Forecasts (ECMWF) from 1998 to 2004 using the nudging
technique.22
The nudging time-scales are 6 h for vorticity, 24 h
for temperature and surface pressure, and 48 h for divergence.
The dynamical data are given for the period 1998-2004, while
the ship emission inventories are representative for 2006. Aerosol
and SO2 background emissions from other sources follow the
recommendations of AeroCom 23
and are specified for the year
2000. As shown by ref 24, the uncertainty introduced by this
inconsistency between ship and nonship emissions is usually
lower than 5% for both global average aerosol burdens and the
indirect aerosol effect. This is well below the uncertainty of the
ship and nonship emissions and thus not a problem here.
A summary of the simulations and emission inventories
considered in the present study is given in Table 1. We estimate
the respective effect of ship emissions by calculating the differ-
ence between a model run with one of the four ship emission
inventories (REF, MGO, PALM, and SOY) and an additional run
without ship emissions. For the REF inventory, we run simulations
for each of the different size distributions assumed for calculating
particle number emissionsresulting fromshipping (fresh, aged 1, and
aged 2). In this way, we can study the sensitivity of the results on this
assumption. All other inventories use the aged 2 size distribu-
tion only.
4. RESULTS AND DISCUSSION
4.1. Global Aerosol Loading. Here, we analyze the global
effect of ship traffic on annual mean aerosol concentrations at the
surface level. We consider the changes in this quantity deriving
from the substitution of HFO with low-sulfur fuels in the
shipping fleet: we express this change as the relative difference
between each of the low-sulfur fuel experiments (MGO, PALM,
and SOY) and the reference experiment (REF).
The largest contribution of ships to the global aerosol loading
in terms of mass is related to SO4 from oxidation of SO2 and
primary emissions. Important reductions in the global SO4
concentration in fine particles (<1 μm) at the near-surface level
are obtained when the HFO component is replaced by one of the
alternative clean fuels. This is shown in Figure 2 for the SOY
inventory: There is a significant decrease in the surface level
sulfate concentration, up to about 40-60% over frequently
traveled routes over the Indian Ocean, the northern Pacific,
and along the routes over the central Atlantic, where the largest
impact of shipping is simulated (MGO and PALM inventories
give similar reductions and are not shown). This reduction is
Figure 2. Top: Simulated multiyear average (1999-2004) of surface-
level SO4 mass concentration in the REF case. Bottom: Corresponding
relative changes for the SOY inventory. Similar changes are modeled for
the MGO and PALM inventories.

4. 3522 dx.doi.org/10.1021/es1036157 |Environ. Sci. Technol. 2011, 45, 3519–3525
Environmental Science & Technology ARTICLE
consistent with the emission totals (Figure 1) for these inven-
tories, which are an order-of-magnitude smaller for SO2 (and
primary SO4) when compared to REF. The reduction in aerosol
sulfate might have large beneficial effects on human health (see
also ref 8): according to our results, the population situated in the
coastal areas of southern Asia, western U.S., and northern and
Mediterranean Europe could gain a large benefit from the
adoption of such fuels in shipping.7
This conclusion is supported
also by the simulated reduction in total aerosol mass in fine
particles (<1 μm) for low-sulfur fuels, as shown in Figure S3.
However, this decrease in aerosol sulfate is partly offset by a
corresponding increase in aerosol nitrate (NO3), because the
NOx emission does not change much using low-sulfur fuels.
Indeed, if less SO4 is present in the troposphere, less ammonia is
involved in the neutralization reaction forming ammonium
sulfate. This excess ammonia thus becomes available for forma-
tion of ammonium nitrate. The relative increase of aerosol nitrate
is most pronounced along the shipping routes in the equatorial
regions.
BC and POM (not shown) are also sensitive to the choice of
fuel composition with the BC surface level concentration show-
ing a significant decrease: this reduction is largest for the PALM
and MGO inventories, especially along the routes in the northern
Pacific where the concentrations can be reduced by up to about
60% (with respect to average surface level concentrations of the
order of 0.01 μg/m3
). A similar pattern is simulated with the SOY
inventory, but with a slightly lower maximum reduction (about
50%) consistent with the emission totals. For POM, the situation
is very different: in case of MGO, the reduction in POM is quite
small, around 20% over the northern Pacific, and about 10%
along other main routes. In the PALM inventory, POM increases
by 10-20% as compared to REF because the total POM
emissions of this fuel type are the highest. The run using the
SOY inventory shows basically no significant change in the POM
surface level concentration.
We simulated an overall reduction in total aerosol number
concentration of fine particles (<1 μm) in particular over the
northern Indian Ocean and the northern Pacific. The model runs
with the MGO and SOY inventories show a similar pattern with
reductions of about 25% in these regions (with respect to average
surface level concentrations of about 500 particles/cm3
; see top-
left panel of Figure S2). Slightly smaller reductions of around
10% are found when using the PALM inventory, likely due to the
associated increase in POM.
In summary, the substitution of HFO with low-sulfur fuel can
have a very positive impact in terms of reducing surface level
aerosol particle concentrations, in particular along the most
traveled shipping routes. This implies beneficial reductions in
air pollution for the population living along the coastlines and in
the vicinity of major harbors, where the ship traffic density may
be high.
4.2. Cloud Microphysical Properties. Ship-emitted aerosols
are known to potentially modify cloud properties by increasing
the concentration of cloud droplets and consequently reducing
their effective radius, given that the total liquid water content is
practically unchanged (our model results show a ship-induced
change in liquid water path of only about 0.5-2%). Regions
where a large amount of low-level clouds and high ship emissions
coincide are most sensitive to such changes: according to our
model, these regions correspond to the most frequented ship-
ping routes (northern Pacific, northern Indian Ocean, and
northwestern and southwestern Atlantic). The modifications of
cloud properties in these regions are usually limited to boundary
layer clouds (up to about 1-1.5 km). In Figure 3, we show the
global multiyear average changes in cloud droplet number
concentration and cloud droplet effective radius (CDNC and
Reff, respectively, average over cloudy periods only) due to
shipping at model layers 17 (∼0.3-0.6 km) and 16 (∼0.6-
1.1 km). For REF, we show the results for all the three aerosol size
distributions considered in emission calculations (Table S2).
The sensitivity of the results to the aerosol size distribution is
discussed in the Supporting Information. In comparison with
REF (size distribution aged 2), the low-sulfur inventories (MGO,
PALM, and SOY) clearly show a smaller effect on the cloud
microphysical properties: changes in cloud droplet number
concentration are reduced by about a factor of 3 from 6.5 cm-3
for REF down to about 2 cm-3
at 0.3-0.6 km. This is also the case
for the corresponding changes in cloud droplet effective radius,
which are about -0.09 μm for REF and -0.03 μm for the low-
sulfur inventories at 0.3-0.6 km. A similar relative reduction is
obtained also at 0.6-1.1 km, although absolute changes at this
level are very small.
We compare our results for the REF inventory with model
calculations from ref 3, who considered three HFO-based ship
emission inventories. At model layer 17, they give a range [5.9;
13.6] cm-3
and [-0.18; -0.06] μm, for changes in CDNC and
effective radius, respectively. For our REF inventory, we estimate
[6.5; 9.0] cm-3
and [-0.12; -0.09] μm, respectively. However,
Figure 3. Multiyear average (1999-2004) of ship-induced changes in cloud droplet number concentration (CDNC, left) and effective radius (Reff,
middle) at model layers 17 (∼0.3-0.6 km, solid pattern) and 16 (∼0.6-1.1 km, checked pattern). Both CDNC and Reff are averages over cloudy periods
only. The right panel shows the multiyear average of the aerosol indirect effect from shipping. The error bars represent the interannual variability.

5. 3523 dx.doi.org/10.1021/es1036157 |Environ. Sci. Technol. 2011, 45, 3519–3525
Environmental Science & Technology ARTICLE
it is most appropriate to compare their “Inventory A” experiment
with our “REF fresh”, because these two model runs consider
similar total SO2 emissions (12 and 14 Tg(SO2)/yr, respectively)
and assume similar size distributions in the emission calculation,
but different geographical emission patterns. The relative differ-
ence between these two specific cases is about 30% in both the
CDNC and the Reff changes. This agrees with the conclusions of
ref 3 that the aerosol effect on clouds does not only depend on
the total sulfur emissions or on the size distribution but also on
the geographical distribution of the ship emissions. As compared
to the REF inventory, “Inventory A” shows distinctively higher
emissions along the frequently traveled routes, which often
coincide with regions of frequent occurrence of low clouds
(see top-left panel of Figure Figure S1 in the Supporting
Information versus left panel of Figure 1 in ref 3; color coding
is the same). On the other hand, REF shows higher emission in
regions of low traffic. This turns out in a similar total emission of
SO2, but a different impact in terms of cloud properties and
radiative forcing.
To further support these findings, we analyze how the size
distribution of global aerosol number concentration changes
with the different emission inventories, as compared to an
inventory without ships. This is shown in Figure S4, where we
analyze the aerosol size distributions in the four lowermost
model layers (altitude below ∼1.1 km). The left panel shows
the global median distribution in the NO SHIPS case, while the
changes compared to this case for the various inventories are
given in the right panel. It is clear that consideration of ship
emissions in the model leads to an increase of aerosol number
concentration in the larger Aitken and accumulation modes
(∼0.04-1 μm). Particles in this size range are more likely to
act as cloud condensation nuclei. This increase is very pro-
nounced for the REF case, becoming less prominent for the low-
sulfur inventories, thus supporting the above conclusions on
cloud droplet number concentrations.
4.3. Radiative Forcing Effects. Aerosols exert a direct effect
by scattering and absorbing incoming solar radiation. According
to our simulations, the direct effect of the ship-induced particles
is small as compared to the corresponding cloud albedo effect, for
all the ship emission inventories considered, thus confirming the
results of ref 3 (see also ref 4). Our model does not include the
albedo effect of black carbon on snow, which is estimated to be
around 0.1 ( 0.1 W/m2
for total anthropogenic black carbon.25
However, we expect the contribution of shipping to such an
effect to be negligible, given that most of the black carbon
deposited in the Arctic, for example, is of continental origin.26
Furthermore, ship emissions of black carbon are low as compared
to other aerosol compounds (Figure 1). Therefore, in the
following, we will focus on the indirect aerosol effect, given also
the large potential for its reduction (see, e.g., ref 24).
An increase in cloud droplet number concentration and the
corresponding decrease in effective radius leads to an increase in
the cloud reflectivity, which is known as the first indirect aerosol
effect.2
We compute the ship-induced first indirect aerosol effect
ΔRFindirect as the changes in the shortwave cloud forcing defined
as the difference of the all-sky shortwave minus the clear-sky
shortwave radiation at the top of the atmosphere:
ΔRFindirect ¼ ΔðRFallsky - RFclearskyÞ
The change in the cloud forcing also includes the second indirect
effect (aerosol-induced changes in cloud liquid water content,
precipitation, and cloud lifetime). As reported in ref 3, changes in
the precipitation pattern in the model are not statistically
significant, suggesting that the second indirect aerosol effect
from shipping is small. The contribution of the long-wave
(thermal) part of the spectrum is also negligible as cloud lifetime
changes are small and the indirect aerosol effect from shipping is
mostly confined to low-level clouds, which are expected to show
only a weak dependence of their long-wave radiance on cloud
microstructure.
The geographical patterns of the changes in multiyear average
cloud forcing due to shipping are shown in Figure S5 (upper
panels). For the REF experiment (left panel), local changes of up
to about -2 W/m2
are found where the busy shipping routes in
the northern Indian Ocean, northeastern Pacific, and eastern
Atlantic coincide with regions of frequent low-level clouds. On
the other hand, very limited changes in the cloud forcing are
found for the low-sulfur scenarios. This is shown in the right
panel of Figure S5 for the SOY experiment (a similar pattern is
obtained for MGO and PALM). The zonal means (lower panels
in Figure S5) reveal a statistically significant radiative forcing
from shipping in the latitude range of 50°S to 60°N for the REF
scenario, which is strongly reduced in the PALM experiment. If
the global mean values are considered (Figure 3, right panel), the
use of alternative fuels results in a smaller indirect effect with a
reduction by a factor of 3 from -0.28 W/m2
(REF aged 2) to
about -0.10 W/m2
(clean fuels), as expected from the results for
ΔCDNC and ΔReff.
To set these numbers into perspective, it is useful to compare
the results of our study with the radiative forcing from other
sources. According to ref 25, the radiative forcing due to the
cloud albedo effect of all anthropogenic aerosols ranges between
-1.8 and -0.3 W/m2
today (year 2005); hence the contribution
of shipping might be very relevant. According to a recent
assessment in ref 4, the contribution of ship emissions to the
indirect effect is estimated to be in the range -0.74 to -0.05 W/
m2
(in 2005); therefore, it far outweighs the warming effect from
ship-induced greenhouse gases such as CO2 (þ0.03 to þ0.05 W
/m2
) or ozone (þ0.01 to þ0.05 W/m2
), overall causing a
negative radiative forcing of shipping.
Given these numbers, one may argue that such a reduction
toward less negative values of RF might accelerate global warm-
ing by canceling less of the warming effect caused by increasing
CO2. It should be kept in mind, however, that a simple cancella-
tion of global warming and cooling effect might not be reasonable
from a physical point of view: the aerosol-induced cloud forcing
is a strongly localized effect, with very high regional variations
(Figure S5), while CO2 exerts relatively homogeneous warming
across the hemispheres. Moreover, the time scales of these two
components are very different: CO2 remains in the atmosphere
for many decades, and its climate response could be relevant on
time scales of the order of centuries or more. SO4, on the other
hand, has a residence time in the order of a few days and impacts
the climate only for decades (see refs 27,28 for a more thorough
discussion on this issue). Climate effects of fuel production,
which might be very different for biofuels and for HFO, are not
considered here. Hence, a complete evaluation of the climatic
impact is not possible, but should be the subject of future
investigations.
Wecompareagainwith thestudyofref3, whichgaveanestimate
of the indirect aerosol effect for three different HFO-based
ship emission inventories: they found values in the range
[-0.60; -0.19] W/m2
. Our estimate for the REF experiment is

6. 3524 dx.doi.org/10.1021/es1036157 |Environ. Sci. Technol. 2011, 45, 3519–3525
Environmental Science & Technology ARTICLE
[-0.40; -0.28] W/m2
, depending on the assumed size distribu-
tion (Figure 3, right panel). Consistent with the results obtained
for changes in cloud properties, the comparison between the
experiments considering the “Inventory A” 3
and “REF fresh”
inventories gives a 30% lower indirect radiative forcing for the
latter (Figure 3, right panel). This reinforces the conclusion on
the dependence of the cloud albedo effect on the geographical
distribution of the ship emissions and demonstrates the impor-
tance of applying accurate ship emission inventories in global
simulations. In a previous work,29
a global CTM was used to
compute the first indirect effect due to sulfate and organic aerosol
from shipping: they report a lower value of [-0.21; -0.06]
W/m2
, probably related to the lower SO2 emission total of 8.4
Tg(SO2)/yr in their inventory (as compared to 14 Tg(SO2)/yr
for REF in the present study).
The above discussion reveals that the current estimates of the
aerosol indirect effect are still very uncertain. The IPCC assigned
a low level of scientific understanding to this effect.25
Besides the
uncertainties inherent in the emission inventories (which are
considered to be small, at least for the sulfur-containing com-
pounds, given the negligible sulfur content of pure biofuels) and
in the size distribution of the emitted aerosol (which we
addressed by considering three different distributions), large
uncertainties are related to the treatment of cloud microphysics,
which is an essential element in the calculation of the aerosol
effects. Our model uses the parametrization of ref 30 to compute
aerosol activation, and the cloud scheme of refs 31,32 to deal with
cloud microphysical processes like growth of cloud droplets due
to water vapor condensation or the interactions of colliding drop-
lets. The latter process results in a self-conversion of cloudwater
or a conversion of cloudwater to rain. Using other schemes and
parametrizations currently available, however, might turn out
large differences in the resulting effects (see the review in ref 33).
Another source of uncertainty is the description of the updraft
velocity, which is a key parameter in the above cloud schemes.
Peng et al.34
showed that this could become especially important
in the case of low clouds, particularly due to subgrid-scale
variability. Despite these limitations, it is likely that reducing
the sulfur content of fuels leads to a reduction in atmospheric
aerosol burden, which can also reduce the impact of shipping on
cloud droplet number concentration. Although the magnitude of
this effect is highly uncertain, the general conclusions drawn by
our study probably hold.
As shown above, the indirect effects of particulate matter
emissions on climate do not differ significantly between low-
sulfur fossil fuels and low-sulfur biofuels. As long as the produc-
tion conditions of the biofuels, including the fuel life cycle, are
such that the CO2 emissions are reduced as compared to fossil
fuel, biofuels may be more favorable. Our analysis is based on a
full replacement of the heavy-fuel oil component by alternative
fuels in the global fleet. This implicitly assumes that the biofuel
production can completely meet the requirements of the ship-
ping sector in the future. Although a full assessment of biofuels
viability in future shipping scenarios is beyond the scope of this
paper, we can estimate the feasibility of such scenarios on the
basis of the results of other studies. Shipping contributes about
16% to the total fuel consumption of the transportation sector
today.1
Global biofuels production could account for 13-23% of
the total fuel demand in the transportation sector in 2050
(according to refs 19 and 35, respectively). Therefore, biofuels
can theoretically meet the fuel demand of the whole fleet,
provided that the shipping share of total fuel consumption will
not increase. The actual use of biofuels in the future will probably
be limited by many practical factors. The analysis of ref 16,
which constitutes the basis of our modeling study, demonstrates
that some types of biofuels (including palm and soy bean oil) can
be used in present-day engines with no need for technical
modifications and are currently in use for stationary engines
and on short-range routes. Their use on the global scale is
therefore not an issue from a technical point of view. A much
more important issue is the possible competition of biofuel and
food production because large areas of arable land will need to be
converted to cropland used for biofuels production. The Inter-
national Energy Agency (IEA)35
showed that such a competi-
tion could be mitigated by an increasing amount of arable land
and by improving agricultural productivity. The latter, in parti-
cular, has a large potential, especially in the developing countries.
IEA concluded that “the development of biofuels, if accompanied
by effective transfer of technologies and improved agricultural
practices, may not conflict with food production”.35
Biofuels can
be effectively mixed with low-sulfur fossil fuel,16
and the blend-
ing fractions might be changed according to the availability of
each component. According to ref 19, South and Central
America and Africa have the largest potential for biofuel produc-
tion, given the large potential increase of arable land. Ships
moving from these regions might adopt blends with a high
biofuel fraction. On the other hand, the biofuel fraction can be
lowered and replaced with low-sulfur marine gas oil, where the
availability of biofuels is lower (a possibility which we explored in
our MGO inventory). Given these possible future limitations, the
results of this work should therefore be regarded as an upper limit
estimate and may represent a useful reference for future studies
addressing the impact of biofuels in more detailed traffic
scenarios.
Although our model provides a robust estimate of the aerosol
radiative forcing effects, several improvements are desirable. A
ship plume parametrization of the effects of particle aging might
improve the representation of particulate matter emissions from
shipping. Moreover, a complete characterization of the climate
impact of biofuels should take into account the effects related to
fuel production and transportation and the changes in the land-
use following a higher request for boifuels.
’ ASSOCIATED CONTENT
bS Supporting Information. Details on the SeaKLIM algo-
rithm and emission factors, and a brief description of the EMAC-
MADE global model. This material is available free of charge via
the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION
Corresponding Author
*Phone: þ49 8153 28 1813. E-mail: mattia.righi@dlr.de.
’ ACKNOWLEDGMENT
This work was funded by the BIOclean project on alternative
fuels in shipping, which is funded by the Bundesministerium
f€ur Bildung und Forschung (BMBF) under the Klimazwei
program. The work was also supported by the Young Investiga-
tors Group SeaKLIM, which is funded by the German Helmholtz-
Gemeinschaft Deutscher Forschungszentren (HGF) and the
Deutsches Zentrum f€ur Luft- und Raumfahrt (DLR). We kindly

7. 3525 dx.doi.org/10.1021/es1036157 |Environ. Sci. Technol. 2011, 45, 3519–3525
Environmental Science & Technology ARTICLE
acknowledge the provision of MADE by the University of
Cologne, Germany (RIU/EURAD-project). We are grateful to
the whole MESSy team for the development of EMAC and to
Valentina Aquila (now at NASA/GSFC, U.S.) and Irene Cionni
for helpful discussions.
’ REFERENCES
(1) Eyring, V.; K€ohler, H.; van Aardenne, J.; Lauer, A. Emissions
from international shipping: 1. The last 50 years. J. Geophys. Res. 2005,
110, D17305.
(2) Twomey, S. The influence of pollution on the shortwave albedo
of clouds. J. Atmos. Sci. 1977, 34, 1149–1152.
(3) Lauer, A.; Eyring, V.; Hendricks, J.; J€ockel, P.; Lohmann, U. Global
model simulations of the impact of ocean-going ships on aerosols, clouds,
and the radiation budget. Atmos. Chem. Phys. 2007, 7, 5061–5079.
(4) Eyring, V.; Isaksen, I.; Berntsen, T.; Collins, W.; Corbett, J.;
Endresen, O.; Grainger, R.; Moldanova, J.; Schlager, H.; Stevenson, D.
Transport impacts on atmosphere and climate: Shipping. Atmos. En-
viron. 2010, 44, 4735–4771.
(5) Schreier, M.; Mannstein, H.; Eyring, V.; Bovensmann, H. Global
ship track distribution and radiative forcing from 1 year of AATSR data.
Geophys. Res. Lett. 2007, 34, L17814.
(6) Schreier, M.; Joxe, L.; Eyring, V.; Bovensmann, H.; Burrows, J.
Ship track characteristics derived from geostationary satellite observa-
tions on the west coast of southern Africa. Atmos. Res. 2010, 95, 32–39.
(7) Corbett, J.; Winebrake, J.; Green, E.; Kasibhatla, P.; Eyring, V.;
Lauer, A. Mortality from ship emissions: a global assessment. Environ.
Sci. Technol. 2007, 41, 8512–8518.
(8) Winebrake, J. J.; Corbett, J. J.; Green, E. H.; Lauer, A.; Eyring, V.
Mitigating the health impacts of pollution from international shipping:
an assessment of low-sulfur fuel mandates. Environ. Sci. Technol. 2009,
43, 4776–4782.
(9) International Maritime Organization. Second IMO GHG study;
London, UK, 2009.
(10) Roeckner, E.; Brokopf, E.; Esch, M.; Giorgetta, M.; Hagema, S.;
Kornblueh, L. Sensitivity of simulated climate to horizontal and vertical
resolution in the ECHAM5 atmosphere model. J. Climate 2006,
19, 3771–3791.
(11) J€ockel, P.; Sander, R.; Kerkweg, A.; Tost, H.; Lelieveld, J.
Technical Note: The Modular Earth Submodel System (MESSy) - a
new approach towards Earth System Modeling. Atmos. Chem. Phys.
2005, 5, 433–444.
(12) J€ockel, P.; Tost, H.; Pozzer, A.; Br€uhl, C.; Buchholz, J.;
Ganzeveld, L.; Hoor, P.; Kerkweg, A.; Lawrence, M. G.; Sander; et al.
The atmospheric chemistry general circulation model ECHAM5/
MESSy1: consistent simulation of ozone from the surface to the meso-
sphere. Atmos. Chem. Phys. 2006, 6, 5067–5104.
(13) Lauer, A.; Hendricks, J.; Ackermann, I.; Schell, B.; Hass, H.;
Metzger, S. Simulating aerosol microphysics with the ECHAM/MADE
GCM - Part I: Model description and comparison with observations.
Atmos. Chem. Phys. 2005, 5, 3251–3276.
(14) Paxian, A.; Eyring, V.; Beer, V.; Sausen, R.; Wright, C. Present-
day and future global bottom-up ship emission inventory including polar
routes. Environ. Sci. Technol. 2010, 44, 1333–1339.
(15) Lack, D. A.; Corbett, J. J.; Onasch, T.; Lerner, B.; Massoli, P.;
Quinn, P. K.; Bates, T. S.; Covert, D. S.; Coffman, B.; Sierau, B.; et al.
Particulate emissions from commercial shipping: Chemical, physical,
and optical properties. J. Geophys. Res. 2009, 114, D00F04.
(16) Petzold, A.; Eyring, V.; Lauer, P.; Fleischer, F.; Hasselbach, J.;
Lichtenstern, M.; Klinger, C.; Righi, M.; Fritsche, U.; Hendricks, J.; et al.
Reduktion der Emissionen klimarelevanter Gase und Partikeln aus
Diesel-Grossmotoren f€ur die Schifffahrt und f€ur den station€aren Einsatz
durch den Einsatz regenerativer Treibstoffe. BIOclean final report,
Klimazwei, BMBF, 2010.
(17) Petzold, A.; Weingartner, E.; Hasselbach, J.; Lauer, P.; Kurok,
C.; Fleischer, F. Physical properties, chemical composition, and cloud
forming potential of particulate emissions from a marine diesel engine at
various load conditions. Environ. Sci. Technol. 2010, 44, 3800–3805.
(18) Petzold, A.; Hasselbach, J.; Lauer, P.; Baumann, R.; Franke, K.;
Gurk, C.; Schlager, H.; Weingartner, E. Experimental studies on particle
emissions from cruising ship, their characteristic properties, transforma-
tion and atmospheric lifetime in the marine boundary layer. Atmos.
Chem. Phys. 2008, 8, 2387–2403.
(19) Doornbosch, R.; Steenblik, R. Is the cure worse than the
disease? OECD 2007, SG/SD/RT(2007)3/REV1, 57 pp.
(20) Fehrenbach, H.; Fritsche, U. R.; Giegrich, J. Greenhouse gas
balances for biomass: issues for further discussion. Available at http://
www.oeko.de/service/bio/dateien/en/ghg_balance_bioenergy.pdf
(21) von Lampe, M. Biofuel support policies: An economic assessment.
OECD 2008, 119 pp.
(22) Jeuken, A. B. M.; Siegmund, P. C.; Heijboer, L. C.; Feichter, J.;
Bengtsson, L. On the potential of assimilating meteorological analyses in
a global climate model for the purpose of model validation. J. Geophys.
Res. 1996, 101, 16939–16950.
(23) Dentener, F.; Kinne, S.; Bond, T. C.; Boucher, O.; Cofala, J.;
Generoso, S.; Ginoux, P.; Gong, S.; Hoelzemann, J. J.; Ito, A.; et al. Emissions
of primaryaerosol andprecursorgasesinthe years 2000 and1750 prescribed
data-sets for AeroCom. Atmos. Chem. Phys. 2006, 6, 4321–4344.
(24) Lauer, A.; Eyring, V.; Corbett, J. J.; Wang, C.; Winebrake, J. J.
Assessment of near-future policy instruments for oceangoing shipping:
Impact on atmospheric aerosol burdens and the Earth’s radiation
budget. Environ. Sci. Technol. 2009, 43, 5592–5598.
(25) Intergovernmental Panel on Climate Change: Climate change
2007. The physical science basis. In Contribution of Working Group I to
the Fourth Assessment Report of the Intergovernmental Panel on Climate
Change; Solomon, S., Qin, D., Manning, M., Chen, Z., Marquis, M.,
Averyt, K. B., Tignor, M., Miller, H. L., Eds.; Cambridge University
Press: Cambridge, UK and New York, NY, 2007.
(26) Koch, D.; Hansen, J. Distant origins of Arctic black carbon: A
Goddard Institute for Space Studies ModelE experiment. J. Geophys. Res.
2005, 110, D04204.
(27) Fuglestvedt, J.; Shine, K.; Bernsten, T.; Cook, J.; Lee, D.;
Stenke, A.; Skeie, R.; Velders, G.; Waitz, I. Transport impacts on
atmosphere and climate: metrics. Atmos. Environ. 2010, 44, 4648–4677.
(28) Fuglestvedt, J.; Berntsen, T.; Eyring, V.; Isaksen, I.; Lee, D. S.;
Sausen, R. Shipping emissions: From cooling to warming of climate - and
reducing impacts on health. Environ. Sci. Technol. 2009, 43, 9057–9062.
(29) Capaldo, K.; Corbett, J. J.; Kasibhatla, P.; Fischbeck, P. S.;
Pandis, S. N. Effect of ship emissions on sulfur cycling and radiative
climate forcing over the ocean. Nature 1999, 400, 743–746.
(30) Abdul-Razzak, H.; Ghan, S. A parameterization of aerosol
activation2. Multiple aerosoltypes.J.Geophys. Res. 2000, 105, 6837–6844.
(31) Lohmann, U.; Feichter, J.; Chuang, C.; Penner, J. Prediction of
the number of cloud droplets in the ECHAM GCM. J. Geophys. Res.
1999, 104, 9169–9198.
(32) Lohmann, U.; K€archer, B.; Hendricks, J. Sensitivity studies of
cirrus clouds formed by heterogeneous freezing in the ECHAM GCM.
J. Geophys. Res. 2004, 109, D16204.
(33) Lohmann, U.; Feichter, J. Global indirect aerosol effect: a
review. Atmos. Chem. Phys. 2005, 5, 715–737.
(34) Peng, Y.; Lohmann, U.; Leaitch, R. Importance of vertical
velocity variations in the cloud droplet nucleation process of marine
stratus clouds. J. Geophys. Res. 2005, 110, D21213.
(35) International Energy Agency. Energy Technology Perspectives
2006; Paris, 2006.