this paper might be a pre-copy-editing or a post-print...
TRANSCRIPT
This paper might be a pre-copy-editing or a post-print author-produced .pdf of an article accepted for publication. For the
definitive publisher-authenticated version, please refer directly to publishing house’s archive system.
Global Mortality Attributable to Aircraft Cruise Emissions Steven R.H. Barrett,1 Rex E. Britter1 and Ian A. Waitz2
Author Affiliations:
1 Department of Engineering, University of Cambridge, Trumping Street CB2 1PZ, UK.
2 Department of Aeronautics and Astronautics, Massachusetts Institute of Technology, Cambridge
MA 02139, USA.
Date: July 28, 2010
One-sentence Brief:
Aircraft cruise emissions may be more significant than landing and takeoff emissions in terms of their contribution to the risk of premature mortality associated with poor surface air quality.
Abstract Aircraft emissions impact human health though degradation of air quality. The majority of previous
analyses of air quality impacts from aviation have considered only landing and takeoff emissions. We
show that aircraft cruise emissions impact human health over a hemispheric scale and provide the first
estimate of premature mortalities attributable to aircraft emissions globally. We estimate ~8,000
premature mortalities per year are attributable to aircraft cruise emissions. This represents ~80% of the
total impact of aviation (where the total includes the effects of landing and takeoff emissions), and
~1% of air quality-related premature mortalities from all sources. However, we note that the impact of
landing and takeoff emissions is likely to be under-resolved. Secondary H2SO4-HNO3-NH3 aerosols are
found to dominate mortality impacts. Due to the altitude and region of the atmosphere at which aircraft
emissions are deposited, the extent of transboundary air pollution is particularly strong. For example,
we describe how strong zonal westerly winds aloft, the mean meridional circulation around 30-60°N,
interaction of aircraft-attributable aerosol precursors with background ammonia, and high population
densities in combination give rise to an estimated ~3500 premature mortalities per year in China and
India combined, despite their relatively small current share of aircraft emissions. Subsidence of
aviation-attributable aerosol and aerosol precursors occurs predominantly around the dry subtropical
ridge, which results in reduced wet removal of aviation-attributable aerosol. It is also found that aircraft
NOx emissions serve to increase oxidation of non-aviation SO2, thereby further increasing the air
quality impacts of aviation. We recommend that cruise emissions be explicitly considered in the
development of policies, technologies and operational procedures designed to mitigate the air quality
impacts of air transportation.
Introduction Aviation currently accounts for approximately 3% of annual fossil fuel energy usage (1) and has been
forecast to grow at an average of 5% per annum until 2027 (2). While climate perturbations attributable
to aircraft emissions have been extensively studied (3,4), the impact of aviation on air quality and
human health globally has not been reported.
Current regulatory practice is to neglect the effects of aircraft cruise emissions on surface air quality
(5,6). Only landing and takeoff cycle (LTO) emissions – conventionally up to an altitude of 3000 ft or
approximately 1 km – are generally accounted for. This corresponds to a typical planetary boundary
layer height, within which pollutants mix rapidly. However, recent regional modeling work indicates
that cruise emissions may contribute a significant fraction of aircraft-accountable ground-level
pollutant concentrations on a regional scale (7).
We report simulation results indicating that aircraft cruise emissions are implicated in increased
premature mortality on a hemispheric scale. Furthermore, meridional and zonal circulation patterns at
cruise altitudes displace impacts from flight paths by several thousand kilometers. Our approach is to
use a recent aircraft emissions inventory, a global chemistry-transport model, population density and
disease statistics, and concentration-response functions derived from epidemiological studies to assess
the impact of aircraft emissions globally on premature mortality. Parametric uncertainties in aircraft
emissions and concentration-response functions are propagated throughout the analysis, along with
estimates of modeling uncertainty.
Methods Fuel burn and emissions inventories were based on the US Federal Aviation Administration’s
AEDT/SAGE tool (8), which has been evaluated and its uncertainty quantitatively assessed (9).
AEDT/SAGE calculates aircraft fuel burn in 2006 at 188 Tg, which we estimate to be accurate to ±5%
overall and ±25% for LTO emissions. The global average emissions index (emitted pollutant mass per
fuel burn mass) for oxides of nitrogen is EI NOx as NO2 = 13.8 g/kg-fuel ±25%. Current evidence
suggests that the globally-averaged aviation fuel sulfur concentration is in the range 400-800 ppm
(10,11). Significant regional and inter-annual variability exists, but at the present time insufficient fuel
survey data is available to support regionally-differentiated fuel sulfur concentration specifications. We
assumed a fleet average fuel sulfur concentration of 600 ±200 ppm, which corresponds to EI SOx as
SO2 = 1.2 ±0.4 g/kg-fuel. Nominal black carbon (BC) and particulate organic carbon (OC) emissions
indices were estimated at EI BC = 0.04 g/kg-fuel and EI OC = 0.02 g/kg-fuel, respectively.
Uncertainties in BC and OC emissions are discussed in the Supporting Information (SI). Primary
sulfate aerosol emissions were estimated by assuming a nominal conversion efficiency of ! = 2% from
fuel sulfur to SVI, with a range of ! = 0.5-6%. LTO particulate matter emissions indices were based on
the First Order Approximation version 3 (12). Emissions indices with associated uncertainty ranges are
summarized in Table 1.
!"#$%&'(!"#$$%!&'$(&)$*!$+,--,./-!,/*,0$-!1-$*!,/!01(($/%!&--$--+$/%!23!$+,%%$*!-4$0,$-5!!
)*&+,-.,/& 012& 31456"$& 75,8&39!"":&39;& 6758! 695:! 6;59!<9!"":&<9;& 75:! 65<! 65=!=>& 7576! 7578! 75<!<?*&":&7;<9@& 7576! 7578! 756>!9>& 7576! 757<! 75=!
The global three-dimensional, coupled oxidant-aerosol model, GEOS-Chem, was chosen for this study
(13). It is driven by assimilated meteorological observations from the Goddard Earth Observing System
and has been extensively applied and evaluated. GEOS-Chem has been used to study the
intercontinental transport of aerosols and aerosol precursors (14-19) and in a recent air quality
mortality assessment for shipping (20). A model sharing substantially common code for
photochemistry, emissions and deposition was applied for the 1999 IPCC Special Report on Aviation
and the Global Atmosphere (3). A detailed wet deposition scheme is included, which has been
constrained by observations (21). BC, OC and H2SO4-HNO3-NH3 aerosols are simulated (18).
Epidemiological studies indicate that long-term population exposure to fine particulate matter with an
aerodynamic diameter less than 2.5 µm (PM2.5) is associated with increased risk of health impacts
including premature mortality (22-28). Although associations between other pollutants and premature
mortality have been found, a relatively strong evidence base exists for PM2.5 as the exposure metric that
is most consistently and independently associated with premature mortality and other health effects. No
exposure threshold is likely to exist for PM2.5 (28).
Concentration-response functions relate baseline pollutant concentrations and baseline health incidence
rates to expected changes in incidence rates. Concentration-response functions and confidence intervals
were based on the World Health Organization (WHO) guideline method (28). Premature mortalities
caused by lung cancer and cardiopulmonary diseases due to long-term exposure to aviation-attributable
PM2.5 were considered for adults over age 30. Global population density data derived primarily from
national censuses (29) and WHO disease statistics aggregated on the WHO sub-region level were used
in the analysis. An unquantified uncertainty is the potential differential toxicity of the PM modeled,
which is important to the extent that aviation-attributable PM differs in chemical composition and size
relative to the polluted urban air upon which epidemiological concentration-response functions are
derived. We note that the PM precursor pollutants emitted from aviation are not different than those
emitted from other combustion sources. This implies that the impacts of secondary H2SO4-HNO3-NH3
aerosols are comparable to non-aviation sources assuming no differential toxicity among secondary PM
species.
Results from 50 GEOS-Chem integrations of 15 months duration are reported herein. A set of
simulations was used to represent the range of uncertain parametric inputs. Coupling between primary
sulfate aerosol emissions and SO2 emissions was accounted for through the fuel sulfur to SVI
conversion efficiency. Results of simulations were used to construct a multi-dimensional interpolation
(covering fuel burn, EI NOx, EI SOx, !, EI BC, and EI OC), through which input probability density
functions were passed to produce probabilistic outputs of population exposure. Additional simulations
captured the impact of LTO emissions only, European cruise emissions, North American cruise
emissions, and South-East Asian cruise emissions. The role of aircraft NOx emissions in oxidizing non-
aviation SO2 was explored with a set of simulations at different aviation fuel sulfur concentrations (0,
200, 400, 600 and 800 ppm) and with aircraft NOx emissions at their nominal value, perturbed by
±25%, and set to zero. Two different aerosol thermodynamic equilibrium codes were employed.
Sensitivity simulations were conducted to quantify the influence of uncertainty in background
ammonia and selection of meteorological year. Further details are given in the SI.
Results and Discussion We first discuss results from simulations with all uncertain parameters at their nominal values. Figure 1
shows aircraft fuel burn, and separately, ground-level perturbations in BC only and all PM2.5 (BC, OC,
sulfate, nitrate and ammonium) due to aircraft emissions. LTO-only and full flight (LTO+Cruise) fuel
burn and air quality impacts are shown in the right and left columns, respectively.
A5,BC%&'(!?#.2&#!4#.%-!@.(!&/*!ABC!D7E6!F+G!&/*!@1##!@#,)H%!DABC!I!J(1,-$G!.4$(&%,./-K!'$(%,0&##3!-1++$*!&'$(&)$!@1$#!21(/!DF)L-!4$(!6°"6°GM!)(.1/*E#$'$#!NJ!4$(%1(2&%,./!Dµ)L+9GM!&/*!)(.1/*E#$'$#!%.%&#!OP<5>!4$(%1(2&%,./!Dµ)L+
9G!&%%(,21%&2#$!%.!&',&%,./5!
Aircraft-attributable BC is formed during combustion; as such, transport and removal processes are
most relevant when considering ground-level BC perturbations. In the case of LTO emissions and
ground-level BC perturbations, Fig. 1 (right column) show a strong spatial correlation, although
impacts are visibly displaced to the east due to prevailing ground-level westerly winds at midlatitudes.
Considering all aircraft emissions [Fig. 1 (left column)], impacts are dispersed throughout 30-60°N and
to the east of peak emissions by ~10,000 km.
Features of the mean circulation of the atmosphere that contribute to this are depicted in Fig. 2, which
shows mean meridional streamlines and zonal mean winds as a function of altitude. It can be seen that
the peak in zonal westerly winds occurs at about 35°N and 270 hPa (9.5 km), which corresponds
closely to the latitude of peak aircraft emissions and typical cruise altitudes. A large fraction of aircraft
emissions are thus injected at a latitude and altitude where pollutants are transported east at greater than
15 m/s on average, resulting in longitudinally displaced ground-level impacts.
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
Peak aircraft emissions occur in the upper part of the Ferrel cell – a thermally indirect mid-latitude
meridional circulation – which can be characterized in an average sense by air masses at ~40°N and
cruise altitudes moving south to ~30°N before descending towards ground-level. Subsiding air masses
at ~30°N may be entrained into the Hadley circulation and so continue further south past 20°N, or be
entrained into surface southerlies. While upwelling regions transferring polluted air masses from the
lower to free troposphere experience efficient wet scavenging of aerosol, subsiding air masses
transporting pollution from cruise altitudes to the lower troposphere are not associated with rapid
aerosol removal. The typical aircraft pollution transport path depicted in Fig. 2 thus provides for a low-
removal rate path for aerosol. The mean fuel burn-weighted flight latitude is 34°N, while the mean
aviation BC perturbation latitude is 28°N, i.e. a 600 km southerly shift. Moreover, as can be seen in
Fig. 2, there is a 900 km southerly shift between the mean cruise emissions latitude and the center-of-
mass of surface #BC impacts. #BC is approximately symmetric around the subtropical ridge.
Figs. 1 and 2 show that PM2.5 perturbations do not penetrate the intertropical convergence zone (ITCZ),
where strong convective activity occurs. Impacts of LTO emissions (Fig. 1) do not extend as far south
due to prevailing northerly surface winds and faster removal rates in the lower atmosphere.
Aviation-attributable ground-level PM2.5 consists largely of primary particulate matter (BC and OC),
and secondary sulfate-ammonium-nitrate aerosols formed though reaction with aircraft-attributable
HNO3 and H2SO4, from NOx and SOx, respectively, and background NH3. In regions with available
NH3, aqueous then solid phase ammonium concentrations increase until all sulfate is neutralized. If
NH3 is still available, aerosol nitrate begins to form. Aerosol liquid water responds nonlinearly over
these regimes. On a global basis, 99% of population-weighted aircraft attributable PM2.5 is secondary
sulfate-ammonium-nitrate aerosol and 1% is primary particulate matter.
A5,BC%&D(!Y4$0,&%$*!)(.1/*E#$'$#!OP<5>!4$(%1(2&%,./!Dµ)L+9G!*1$!%.!&',&%,./5!DBH$!NJ!4$(%1(2&%,./!,-!
),'$/!,/!",)1($!65G!
Figure 3 shows speciated plots of the ground-level PM2.5 perturbation due to aviation. These constitute
the components of aviation-attributable PM2.5 with the exception of secondary organic aerosols (see the
SI). The influence of the lack of available ammonia over the Sahara can be seen: nitrate formation is
suppressed in this region, with only (acidic) sulfate formation occurring. Note that the OC pattern is on
account of the nominal assumption of low OC cruise emissions relative to LTO, which are spatially
variable depending on aircraft age and type.
Applying the WHO concentration-response functions, disease statistics, population density data, and
GEOS-Chem modeling results, we estimate a global total of 9970 premature mortalities per year due to
aircraft emissions (given 2001-2 meteorology). Premature mortalities per year for selected countries
are: Canada, 67; China, 1890; Ethiopia, 43; France, 380; Germany, 545; India, 1640; Iran, 76; Spain,
189; United Kingdom, 362; and United States, 458. LTO emissions account for 20% of the total
premature mortalities attributable to aviation globally, however this is likely a lower bound due to the
global model resolution. We estimate that LTO impacts are underestimated by a factor of two in the US
by comparing results from a GEOS-Chem simulation with a high (~36 km) resolution CMAQ
simulations of the impact of US LTO operations (30). While noting that the underestimate of LTO
impacts is likely lower in other countries that receive a relatively higher fraction of transboundary
pollution from aircraft, applying this factor globally leads to an estimate of LTO emissions being
responsible for ~30% of aviation-attributable premature mortalities. This is discussed in the SI.
A5,BC%&@(!O$(%1(2&%,./!,/!%H$!)(.1/*E#$'$#!NJ!0./0$/%(&%,./!Dµ)L+9G!&%%(,21%&2#$!%.!&,(0(&@%!$+,--,./-!@(.+! ./#3! @(.+! Z[\! D<>°]R9:°^! E! 97°ZR;7°^GR! ^T! D6=>°]R67°^! E! >7°]R9:°^G! &/*! YZT! D=>°ZR=°Y! E!6>7°ZR>8°^GR!($-4$0%,'$#35
Impacts due to cruise emissions from specific regions were considered. Fig. 4 shows the perturbation in
surface BC concentrations due to emissions only from Europe (EUR), North America (NA) and South-
East Asia (SEA), respectively, which illustrates the impact of strong zonal westerlies at cruise altitudes
and meridional transport for EUR and NA emissions. In the case of SEA emissions, which occur in the
subtropics with prevailing easterly winds and upwelling air, the majority of the impacts are to the west.
EUR, NA and SEA emissions account for 34%, 34% and 29% of global aviation-attributable premature
mortalities, respectively, accounting for 97% of the total mortalities. Results differed by less than 1%
when regional attribution was calculated by subtracting aircraft emissions from EUR, NA and SEA
with all other aircraft emissions present.
Figure 1 shows a local peak PM2.5 concentration in China. The global background ammonia
concentration as calculated by GEOS-Chem is shown in Fig. 5. Peak ammonia concentrations and
emissions occur in China (31). The high availability of ammonia in China allows for efficient
conversion of aerosol precursors transported from the north and west into secondary aerosol. The
country with the second highest available ammonia is India, particularly in the north, where again a
peak in aircraft-attributable PM2.5 is found. These ammonia and PM2.5 concentration peaks are also
correlated with local peaks in population density (29). Vertically integrated aircraft fuel burn over India
and China account for 2% and 8% of global fuel burn, respectively, while their combined share of
global aviation-attributable premature mortalities is 35%.
A5,BC%&E(!BH$!2&0F)(.1/*!)(.1/*E#$'$#!^S9!0./0$/%(&%,./!Dµ)L+9G5!
The role of aircraft NOx emissions in producing ozone in the upper troposphere and lower stratosphere
(UT/LS) has been extensively studied in a climate context (4). We show here that aircraft NOx
emissions are not only responsible for increased surface nitrate concentrations, but also for increased
sulfate concentrations. This is because aircraft NOx emissions increase the oxidizing capacity of the
atmosphere, which increases oxidation of (non-aviation) SO2 to sulfate.
Unger et al. described the link between ozone precursor emissions and sulfate concentrations (32).
Specifically, increased NOx emissions result in increased ozone, which yields increased OH for
(predominantly) gas phase oxidation of SO2 to sulfate. We note that NOx emissions in the UT/LS result
in greater ozone production than the same emissions at the surface level (4).
A5,BC%&F(!JH&/)$!,/!&'$(&)$!-1(@&0$!-1#@&%$!#.&*,/)!&-!&!@1/0%,./!.@!&,(0(&@%!^C!!&/*!YC!!$+,--,./-5!
A5,BC%&G(!_$0($&-$!,/!%H$!)(.1/*E#$'$#!YC<!0./0$/%(&%,./!&/*!%H$!,/0($&-$!,/!%H$!)(.1/*E#$'$#!-1#@&%$!&/*!.U./$!0./0$/%(&%,./-!&%%(,21%&2#$!%.!&,(0(&@%!$+,--,./-5!".(!%H$-$!@,)1($-R!&!7!44+!&',&%,./!@1$#!-1#@1(!0./0$/%(&%,./!H&-!2$$/!&--1+$*5!BH$!2&0F)(.1/*!YC<!0./0$/%(&%,./! ,-!&#-.!-H.Q/5! `%!0&/!2$!-$$/! %H&%! &',&%,./! ,/0($&-$-! -1(@&0$! -1#@&%$! #.&*,/)! +.-%! QH$($! ,%! ,/0($&-$-! .U./$! D$5)5! .'$(! %H$!O&0,@,0!C0$&/G!&/*L.(!,/!($),./-!QH$($!%H$!2&0F)(.1/*!YC<!0./0$/%(&%,./!,-!H,)H!D$5)5!./!%H$!Z&-%$(/!Y$&2.&(*!.@!%H$![YG5!]$!&#-.!/.%$!%H&%!.'$(!.0$&/-!,/0($&-$*!.V,*&%,./!.@!_PY!.001(-5
In order to investigate the relationship between aircraft NOx emissions and surface sulfate loading,
GEOS-Chem integrations were performed perturbing aircraft NOx emissions by ±25% and varying
aviation fuel sulfur concentrations (which are directly proportional to SOx emissions) over 0-800 ppm.
Results are given in Fig. 6, which shows the change in the average surface sulfate concentration
attributable to aviation as a function of the assumed aviation fuel sulfur concentration. We find that at
the nominal assumed level of aircraft NOx emissions and fuel sulfur concentration (600 ppm), more
than half of aviation-attributable sulfate at the surface is associated with aircraft NOx emissions and not
with aircraft SOx emissions. This can be discerned from Fig. 6 by noting that the upper solid line does
not intersect the origin when aircraft NOx emissions are included; the vertical offset between the two
solid lines in the plot represents the sulfate attributable to aircraft NOx emissions, while the increase to
the right is due to increasing fuel sulfur concentration.
The regional distribution of the role of aircraft NOx emissions increasing the oxidation of background
(non-aviation) SO2 and DMS to sulfate is depicted in Fig. 7, where a 0 ppm aviation fuel sulfur
concentration has been assumed. (Surface DMS loading decreased by ~0.01 µg/m3 over the northern
Pacific and Atlantic oceans.) It can be seen that the latitude band of increased ozone due to aviation
gives rise to a corresponding increase in sulfate, except that this increase is magnified in regions of
high background SO2 emissions.
Accounting for parametric uncertainty in emissions, concentration-response functions and estimated
modeling uncertainties and biases (discussed in the SI), our expected number of premature mortalities
per year due to aviation is 12,600, with a 95% confidence interval of 6000-19,900. We find that the
single largest contributor to uncertainty is the concentration-response function, followed by estimated
modeling uncertainty. EI NOx is the third most significant contributor, while uncertainty in fuel sulfur
to SVI conversion efficiency has a negligible (less than 1%) effect on estimates.
We note additional sources of uncertainty that were addressed with sensitivity simulations or
considered qualitatively. Formation of H2SO4-HNO3-NH3 aerosols is sensitive to background NH3
emissions. Sensitivity simulations perturbing ammonia emissions by +30% and -30% (discussed in the
SI) resulted in global mortalities changing by +10% and -15%, respectively, with changes concentrated
in regions of highest ammonia emissions. A sensitivity simulation based on a different year (2006) and
an updated meteorological data product resulted in aviation-attributable mortalities increasing by 21%
relative to the base period (spanning 2001-2002). A further year (2007) with the same updated
meteorological product resulted in a relative change in aviation-attributable mortalities of +5%. We
also note that under-resolution of LTO impacts by a global atmospheric model (discussed in the SI)
may increase the total mortalities attributable to LTO emissions from ~20% to ~30% or more.
We have shown that ~8000 premature mortalities (or on the order of 60,000 life years lost) per year are
attributable to aircraft cruise emissions worldwide. This can be compared with a recent analysis for
shipping by Corbett et al., which estimates ~60,000 mortalities per year attributable to that sector (20).
The quantity of fuel burned by the two sectors is similar, but the fuel sulfur concentration of marine
bunker fuel is an order of magnitude higher (8,20). Crobett et al. found their premature mortality
estimate for California was ~3 times higher than a previous analysis by the California Air Resources
Board that excluded the effects of ships more than 24 nm from the shoreline, although other differences
in modeling may have contributed to this. We note that the methods used in the present study are
comparable to those used in the shipping study – both applied GEOS-Chem and the same WHO
concentration-response functions (although in a more aggregate form in Corbett et al.). The World
Health Organization estimates that globally 0.8 million premature mortalities per year are due to
anthropogenic air pollution (26). Our estimates for aviation-attributable mortalities represent less than
1% of this figure.
Liu et al. (33) estimated globally ~90,000 premature deaths are due to intercontinental transport of
PM2.5 and its precursors. While aviation accounts for ~1% of total air quality-related premature
mortalities, it is equivalent to ~10% of intercontinental impacts. Precise quantitative comparison of
aviation’s “efficiency” of exporting pollution relative to other sources on a like-for-like basis requires
further work. However, our calculations indicate that some countries are net exporters of aircraft
pollution and some are net importers. For example, the United States incurs ~7 times fewer mortalities
than would be expected based on its aviation fuel burn alone, while India incurs ~7 times more
mortalities than would be expected by scaling total global aviation mortalities by its fraction of aviation
fuel burn (see the SI).
The air quality impacts of aviation are uniquely influenced by mean features of the general circulation,
which serve to reduce the correlation between countries responsible for the majority of emissions, and
populations exposed to the largest health risk perturbations. A key implication of our results is that
future policy analyses of the air quality impacts of aviation should account for cruise emissions.
Consideration of the LTO phase in isolation when compiling emissions inventories for air quality
purposes may miss the majority of the health-relevant emissions. Because aircraft operational and
technological factors that control cruise emissions are different from those that set LTO emissions, it is
important that development of new technologies, operational procedures, and policies designed to
mitigate environmental impacts of air transportation explicitly address cruise operations. Furthermore,
as aviation is estimated to contribute 55 mWm-2 (3.5%, with an uncertainty range of 1.3-10%) of
present-day radiative forcing of climate (4), the present analysis provides a case for the co-benefits of
emissions mitigation in the aviation sector.
References and Notes (1) International Energy Agency/OECD. IEA Energy Balance; International Energy Agency: Paris, 2007.
(2) The Boeing Company. Boeing Current Market Outlook 2008-2027; Boeing, U.S., 2008.
(3) Intergovernmental Panel on Climate Change. Aviation and the Global Atmosphere; Cambridge University Press: UK, 1999.
(4) Lee, D. S.; Pitari, G.; Grewe, V.; Gierens, K.; Penner, J. E.; Petzold, A.; Prather, M. J.; Schumann, U.; Bais, A.; Berntsen, T.; Iachetti, D.; Lim, L. L.; Sausen, R. Transport impacts on atmosphere and climate: Aviation. Atmos. Environ. 2010, in press (doi:10.1016/j.atmosenv.2009.06.005).
(5) Ratliff, G.; Sequeira, C.; Waitz, I.; Ohsfeldt, M.; Thrasher, T.; Graham, G.; Thompson, T.; Graham, M.; Thompson, T. Aircraft Impacts on Local and Regional Air Quality in the United States. PARTNER report (Report No. PARTNER-COE-2009-002).
(6) UK Department for Transport. Project for the Sustainable Development of Heathrow: Report of the airport air quality technical panels; UK Department for Transport, 2006.
(7) Tarrasón, L.; Eiof Jonson, J.; Berntsen, T. K.; Rypdal, K. Study on air quality impacts of non-LTO emissions from aviation; Report to the European Commission under contract B4-3040/2002/343093/MAR/C1, 2004. Available from http://www.europa.nl/environment/air/pdf/air_quality_impacts_finalreport.pdf
(8) Kim, B.Y.; Fleming, G.G.; Lee, J.J.; Waitz, I.A.; Clarke, J.-P.; Balasubramanian, S.; Malwitz, A.; Klima, K.; Locke, M.; Holsclaw, C.A.; Maurice, L.Q.; Gupta, M.L. System for assessing Aviation's Global Emissions (SAGE), Part 1: Model description and inventory results. Trans. Res.: Part D: Transport Environ. 2007, 12 (5), 325-346.
(9) Lee, J.J.; Waitz, I.A.; Kim, B.Y.; Fleming, G.G.; Maurice, L.; Holsclaw, C.A. System for assessing Aviation's Global Emissions (SAGE), Part 2: Uncertainty assessment. Trans. Res.: Part D: Transport Environ. 2007, 12 (6), 381-395.
(10) Hileman, J. I.; Ortiz, D. S.; Bartis, J. T.; Wong, H. M.; Donohoo, P. E.; Weiss, M. A.; Waitz, I. A.; Alternative Jet Fuels; PARTNER/RAND report (Report No. PARTNER-COE-2009-001).
(11) Defense Energy Support Center. Petroleum Quality Information System Report; Defense Energy Support Center, DESC-BP, 1999-2006.
(12) International Civil Aviation Organization. Airport Air Quality Guidance Manual; International Civil Aviation Organization, 2007.
(13) Bey, I.; Jacob, D. J.; Yantosca, R. M.; Logan, J. A.; Field, B.; Fiore, A. M.; Li, Q.; Liu, H.; Mickley, L. J.; Schultz, M. Global modeling of tropospheric chemistry with assimilated meteorology: Model description and evaluation. J. Geophys. Res. 2001, 106 (23), 23,073-23,096.
(14) Chin, M.; Diehl, T.; Ginoux, P.; Malm, W. Intercontinental transport of pollution and dust aerosols: Implications for regional air quality. Atmos. Chem. Phys. 2007, 7 (21), 5501-5517.
(15) Heald, C. L.; Jacob, D. J.; Park, R. J.; Alexander, B.; Fairlie, T. D.; Yantosca, R. M.; Chu, D. A. Transpacific transport of Asian anthropogenic aerosols and its impact on surface air quality in the United States. J. Geophys. Res. 2006, 111 (14), D14310.
(16) Park, R. J.; Jacob, D. J.; Kumar, N.; Yantosca, R. M. Regional visibility statistics in the United States: Natural and transboundary pollution influences, and implications for the Regional Haze Rule. Atmos. Environ. 2006, 40 (28), 5405-5423.
(17) Park, R. J.; Jacob, D. J.; Chin, M.; Martin, R. V. Sources of carbonaceous aerosols over the United States and implications for natural visibility. J. Geophys. Res.108 2003, (D12), 4355.
(18) Park, R. J.; Jacob, D. J.; Field, B. D.; Yantosca, R. M.; Chin, M. Natural and transboundary pollution influences on sulfate-nitrate-ammonium aerosols in the United States: Implications for policy. J. Geophys. Res. 2004, 109 (15), D15204.
(19) Duncan Fairlie, T.; Jacob, D. J.; Park, R. J. The impact of transpacific transport of mineral dust in the United States, Atmos. Environ. 2007, 41 (6), 1251-1266.
(20) Corbett, J. J.; Winebrake, J. J.; Green, E. H.; Kasibhatla, P.; Eyring, V.; Lauer, A. Mortality from ship emissions: A global assessment. Environ. Sci. Technol, 2007, 41 (24), 8512-8518.
(21) Liu, H.; Jacob, D. J.; Bey, I.; Yantosca, R. M. Constraints from 210Pb and 7Be on wet deposition and transporting a global threee-dimensional chemical tracer model driven by assimilated meteorological fields. J. Geophys. Res. 2001, 106 (D11), 12109-12128.
(22) Cohen, A. J.; Anderson, H. R.; Ostro, B.; Pandey, K. D.; Krzyzanowski, M.; Künzli, N.; Gutschmidt, K.; Pope, A.; Romieu, I.; Samet, J. M.; Smith, K. The global burden of disease due to outdoor air pollution. J. Toxicol. Environ. Health A. 2005, 68 (13-14), 1301-1307.
(23) Roman, H. A.; Walker, K. D.; Walsh, T. L.; Conner, L.; Richmond, H. M.; Hubbell, B. J.; Kinney, P. L. Expert judgment assessment of the mortality impact of changes in ambient fine particulate matter in the U.S. Environ. Sci. Technol. 2008, 42 (7), 2268-2274.
(24) Cooke, R. M.; Wilson, A. M.; Tuomisto, J. T.; Morales, O.; Tainio, M.; Evans, J.S. A probabilistic characterization of the relationship between fine particulate matter and mortality: Elicitation of european experts. Environ. Sci. Technol. 2007 (41), 6598-6605.
(25) Pope III, C. A. Mortality effects of longer term exposures to fine particulate air pollution: Review of recent epidemiological evidence. Inhalation Toxicol. 2007, 19, 33-38.
(26) Cohen, A. J.; Anderson, H. R.; Ostro, B.; Pandey, K. D.; Krzyzanowski, M.; Kuenzli, N.; Gutschmidt, K.; Pope, C. A.; Romieu, I.; Samet, J. M.; Smith, K. R. Mortality impacts of urban air pollution. In Comparative quantification of health risks: Global and regional burden of disease due to selected major risk factors, eds. Ezzati, M.; Lopez, A. D.; Rodgers, A.; Murray, C. U. J. L. vol. 2. World Health Organization, 2004.
(27) Pope III, C. A.; Burnett, R. T.; Thun, M. J.; Calle, E. E.; Krewski, D.; Ito, K.; Thurston, G.D. Lung cancer, cardiopulmonary mortality, and long-term exposure to fine particulate air pollution. J. Am. Med. Assoc. 2002, 287 (9), 1132–1141.
(28) B. Ostro. Outdoor air pollution: Assessing the environmental burden of disease at national and local levels. Environmental Burden of Disease Series, No. 5 World Health Organization, 2004.
(29) Center for International Earth Science Information Network (CIESIN), Centro Internacional de Agricultura Tropical (CIAT). Gridded Population of the World (GPW), Version 3. Columbia University, 2004.
(30) Brunelle-Yeung, E.; Masek, T.; Rojo, J. J.; Levy, J. I.; Arunachalam, S.; Miller, S. M.; Barrett, S. R. H.; Kuhn, S. R.; Waitz, I. A. Methods for assessing the impact of aviation environmental policies on public health. To appear in Transport Policy. 2010.
(31) Bouwman, A. F.; Lee, D. S.; Asman, W. A. H.; Dentener, F. J.; Van Der Hoek, K. W.; Olivier, J. G. J. A global high-resolution emission inventory for ammonia. Global Biochem. Cycles, 1997, 11 (4), 561-587.
(32) Unger, N.; Shindell, D. T.; Koch, D. M.; Streets, D. G. Cross influence of ozone and sulfate precursor emissions changes on air quality and climate. Proc. Nat. Acad. Sci. 2006, 103 (12), 4377-4380.
(33) Liu, J.; Mauzerall, D. L.; Horowitz, L. W. Evaluating inter-continental transport of fine aerosols:(2) Global health impact. Atmos. Environ. 2009, 43 (28), 4339-4347.
Acknowledgements. UK Research Councils EPSRC and NERC funded this work. The US FAA and US DOT Volpe Center provided aircraft emissions inventories used in this study.
Supporting Information. Further discussion, analyses and results are available online.
! !"
Supporting Information: Global Mortality Attributable to Aircraft Cruise Emissions Steven R.H. Barrett,1 Rex E. Britter1 and Ian A. Waitz2
Author Affiliations:
1 Department of Engineering, University of Cambridge, Trumping Street CB2 1PZ, UK.
2 Department of Aeronautics and Astronautics, Massachusetts Institute of Technology, Cambridge
MA 02139, USA.
Date: July 28, 2010
!Previous Work
Several experimental studies have been published examining the impact of airports on local air quality
(S1-S3,6). The UK Department for Transport Project for the Sustainable Development of Heathrow
(PSDH) used both experimental and dispersion modeling techniques to quantify the impact of
Heathrow on air quality within ~10 km of Heathrow Airport, London, UK (6). Regional impacts of
Atlanta International Airport, US, were studied using a chemistry-transport model (S4). Under federal
mandate, air quality impacts across the contiguous US attributable to aviation within the US were
estimated (5). This involved chemistry-transport modeling on a regional scale and estimation of
premature mortalities using methods similar to the current work. With the exception of (7), modeling
efforts have been concerned only with landing and takeoff cycle (LTO) emissions (i.e. less than 3000 ft
above ground level). The only study previously examining non-LTO emissions (7) concluded that
“aircraft nitrogen oxide emissions above 1000 m and at cruise level (non-LTO emissions) have a small
but significant impact on regional air quality levels in Europe.” Only NOx emissions were considered.
The study assumed that SO2 and primary PM emissions from aviation would be negligible. No
previous studies were found that estimated premature mortalities globally due to aircraft emissions, or
the impact of cruise emissions on human health and air quality globally.
Particulate Matter Emissions
Aircraft primary volatile and non-volatile particulate matter emissions at cruise have been the subject
of a number of studies (S5-S22). We interpret all non-volatile measurements as BC. Volatile PM
emissions are assumed to be either OC or H2SO4.
Of the studies of BC emissions, Pueschel et al. (S5) could be considered an outlier, with EI BC ~
0.0001-0.001 g/kg. This is orders of magnitude below all other estimates, including measurements by
different groups of the same NASA Boeing 757 aircraft with Rolls-Royce RB211 engines. For
example, Hagen et al. (S7) measured EI BC = 0.026 and 0.042 g/kg using fuel with sulfur contents of
70 and 700 ppm, respectively, while volatile emissions changed by an order of magnitude. Petzold et
al. (S9) presented measurements of a Rolls-Royce/Snecma M45H Mk501 finding EI BC = 0.1 g/kg,
and a CFM56-3B1 finding EI BC = 0.011 g/kg. The newer engine technology (CFM56-3B1) EI BC is
an order of magnitude lower than the older technology (Rolls-Royce/Snecma M45H Mk501). Other
measurements from a number of campaigns are outlined in (S12). Reported EI BC values ranged from
! #"
0.01-0.5 g/kg, with higher values corresponding to older engines. For example, the 0.5 g/kg
corresponds to a Boeing 707. Petzold et al. (S9) applied a correlation to estimate the fleet-averaged EI
BC of 0.038 g/kg. Considering the range of values measured and remaining scientific uncertainties, we
applied a nominal estimate for the fleet average EI BC = 0.04 g/kg, with low and high values of 0.01
and 0.2 g/kg, respectively. While this range spans two orders of magnitude, the overall aviation-
attributable ground-level PM2.5 concentration is dominated by secondary nitrates and sulfates. Aircraft
BC emissions are assumed to be hydrophilic (S23).
Several of the earlier cruise PM emissions studies assumed all volatile PM emissions to be H2SO4
(S5,S6,S8). However, it is now thought that a portion of the volatile PM emissions are OC. Modeling
by Lukachko et al. (S22) indicates a fuel sulfur to SVI conversion efficiency of ! < 6.3%. Gaseous
measurements of SVI by chemical ionization mass spectrometry in a simulated gas turbine found ! = 2.3
±1.2% (S19). Modeling and analyses of previous experiments by Kärcher et al. (S16) suggest ! = 0.5-
5%, while sulfuric acid measurements by Curtius et al. indicate ! = 1.3-5.1% (S14). Additionally, !
may be a function of measurement location, increasing with plume age (S19). Overall we assume a
nominal conversion efficiency of ! = 2%, with an uncertainty range of 0.5-6%. However, as will be
shown, this range has <1% impact on premature mortality estimates.
EI OC is relatively more uncertain than EI BC or !. In the case of in-flight measurements for fuel sulfur
concentrations of 600-700 ppm, total measured volatile emissions indices were 0.1-0.6 g/kg (S5,S6,S8).
We expect EI H2SO4 = 0.02-0.15 g/kg (corresponding to ! = 1-7%). This implies that EI OC = 0-0.58
g/kg. Measurements with a fuel sulfur concentration of 70 ppm resulted in total volatile PM emissions
indices of 0.04-0.08 g/kg. In this case we expect EI H2SO4 = 0.002-0.015 g/kg, given our assumed !
range. This implies EI OC = 0.025-0.06 g/kg. EI OC is likely to be a function of altitude, so we do not
apply ground-based measurements here. Overall we assume a nominal EI OC = 0.02 g/kg, with a range
of 0.01-0.6 g/kg. Despite the uncertainty in EI OC spanning two orders of magnitude, its contribution
to overall uncertainty in our estimates of aviation-attributable premature mortalities will be shown to be
small compared to other sources of uncertainty.
Atmospheric Modeling
Table S1 lists the 50 GEOS-Chem simulations from which results of the current study were derived.
GEOS-Chem version 8-01-04 was used, with the following modifications:
1. The aerosol thermodynamic equilibrium code RPMARES (S24) was substituted for
ISORROPIA (S25,S26) due to known over prediction of aerosol nitrate at low relative
humidity in the current GEOS-Chem version.
2. Existing inventories of time-averaged aircraft NOx and SOx emissions were removed. A
module to incorporate AEDT/SAGE 2006 emissions was developed. This included diurnally
varying BC, OC, primary sulfate, NOx, SO2, CO and speciated HC emissions.
Three months were used as a spin-up period for all integrations. The default simulation period (unless
noted otherwise in Table S1) was July 2001 to September 2002. Simulations were performed at a
! $"
horizontal resolution of 4°"5°. This resolution is thought appropriate for estimating ground-level
impacts of a highly spatially disperse emission source at ~10 km altitude, where ground-level impacts
are mixed over a latitude band spanning 30-60°N. However, the impacts due to LTO emissions are
likely to be underestimated due to correlation between the location of peak LTO emissions (airports)
and peak population densities (cities) not being resolved by a global-scale chemistry-transport model.
!"#$%&'()"%&'()*+,-"./-0123/45."6,7847-,9"847"3+,":077,53".309;<"=+,"9,82013":458/>0723/45"/."2"!?"-453+" ./-0123/45" @48" A+/:+" 3+," 8/7.3" 3+7,," -453+." 48" 7,.013." 27," 9/.:279,9B" 0./5>" 2,74.41"3+,7-49;52-/:",C0/1/D7/0-"@E=&B":49,"FGHEF&(I"A/3+"211"2/7:7283",-/../45." /5:109,9"0./5>"54-/521",-/../45." /59/:,." 259" 34321" 80,1" D075<" J,627307,." 874-" 3+/." :458/>0723/45" 27," 543,9<" =+," 54-/521",.3/-23,"847"2K/23/45)2337/D032D1,"67,-2307,"-47321/3/,."/."21.4">/K,5"847",2:+"./-0123/45<"
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
54-/521"2/7:7283",-/../45."@#WB L2:M>74059"A/3+"Z]$",-/../45."/5:7,2.,9"D;"$VR"@#XB L2:M>74059"A/3+"Z]$",-/../45."/5:7,2.,9"D;"$VR"610."
54-/521"2/7:7283",-/../45."
@$VB J,82013",P:,63"54"2/7:7283"]*",-/../45."@$!B J,82013",P:,63"54"2/7:7283"*'",-/../45."@$#B L2:M>74059"@54"2K/23/45B"A/3+"%&'()?"-,3,47414>;"847"
:21,5927";,27"#VVS"@$$B E11"2/7:7283"610."D2:M>74059",-/../45."A/3+"%&'()?"
-,3,47414>;"847":21,5927";,27"#VVS"@$OB L2:M>74059"@54"2K/23/45B"A/3+"%&'()?"-,3,47414>;"847"
:21,5927";,27"#VVT"@$?B E11"2/7:7283"610."D2:M>74059",-/../45."A/3+"%&'()?"
-,3,47414>;"847":21,5927";,27"#VVT"@$SB L2:M>74059"@54"2K/23/45B"0./5>"Y('FF'GYE"@$TB E11"2/7:7283"610."D2:M>74059",-/../45."0./5>"Y('FF'GYE"
@.:21,9"874-"7,.013."0./5>"67,K/40.",-/../45."/5K,5347;B"@$WB J,82013",P:,63"[(*"U"V"66-"@$XB J,82013",P:,63"[(*"U"#VV"66-"@OVB J,82013",P:,63"[(*"U"V"66-"259"14A"&Y"Z'!"@O!B J,82013",P:,63"[(*"U"#VV"66-"259"14A"&Y"Z'!"@O#B J,82013",P:,63"[(*"U"V"66-"259"+/>+"&Y"Z'!"@O$B J,82013",P:,63"[(*"U"#VV"66-"259"+/>+"&Y"Z'!"@OOB J,82013",P:,63"[(*"U"V"66-"259"54"2/7:7283"&Y"Z'!"@O?B J,82013",P:,63"[(*"U"#VV"66-"259"54"2/7:7283"&Y"Z'!"@OSB J,82013",P:,63"[(*"U"OVV"66-"259"54"2/7:7283"&Y"Z'!"@OTB J,82013",P:,63"[(*"U"SVV"66-"259"54"2/7:7283"&Y"Z'!"@OWB J,82013",P:,63"[(*"U"WVV"66-"259"54"2/7:7283"&Y"Z'!"@OXB L2:M>74059"@54"2K/23/45B"A/3+"%&'()?"-,3,47414>;"847"
:21,5927";,27"#VVS"A/3+"('E":49,",52D1,9"@?VB E11"2/7:7283"610."D2:M>74059",-/../45."A/3+"%&'()?"
-,3,47414>;"847":21,5927";,27"#VVS"A/3+"('E":49,",52D1,9"
The overall PM2.5 response to BC and OC emissions was found to be linear within the range
considered, despite coupling through aerosol effects on photolysis rates. A sensitivity study using
ISORROPIA (S25,S26) instead of RPMARES (S24) for both the aviation and no aviation simulations
resulted in a premature mortalities estimate that is 10% lower.
We now note selected sources of atmospheric modeling uncertainty. Background emissions and
concentrations are likely to influence aviation-attributable secondary PM2.5 estimates. For example,
ammonia emissions may be over-represented in North America in the fall, resulting in a 30% high bias
in total ammonium concentrations in comparison with observations in the US (18). An exponential
dependence of livestock ammonia emissions on temperature is assumed, which may not be sufficient to
describe seasonal variation. Uncertainties in global agricultural ammonia emissions were analyzed by
! O"
Beusen et al. (S27), who estimated an uncertainty range of +19% to -15% about their central estimate.
Overall, to assess the influence of uncertainties in ammonia emissions, we applied an uncertainty range
of ±30% (to both simulations without aviation emissions and with aviation emissions), which is likely
to capture both the maximum seasonal bias and the annual average bias. This resulted in global
mortalities changing by +10% and -15%. Additionally, uncertainty in non-aviation SO2 emissions is
likely to lead to additional uncertainty in the impact of aviation on health due to the coupling between
aircraft NOx emissions and oxidation of non-aviation SO2 to sulfate. This was not investigated with
GEOS-Chem simulations, but we note that a new study estimates that the uncertainty in global total
SO2 emissions is ~10% (S45), although regional uncertainties are higher. Our simulations show that
27% of population-weighted aviation-attributable PM2.5 is sulfate (Table S3), of which approximately
half is related to oxidation of background SO2 (Figure 6). Assuming a linear response of sulfate
concentrations to ~10% changes in SO2 emissions, uncertainty in background SO2 emissions is likely
to have a ~1% impact on aviation-attributable premature mortalities. However, regional variability in
uncertainties may mean this is an underestimate. Background emissions included in GEOS-Chem are
relevant to circa 2000 depending on region, while aircraft emissions are for 2006. This remains an
unquantified uncertainty.
On net, stratosphere-troposphere exchange in GEOS (and other) assimilated meteorological datasets
are likely to be biased high (21,S28), perhaps by a factor of 2-3. However, this relates mainly to
transport from upper parts of the stratosphere to the lower stratosphere and upper troposphere as it is
determined by the net flux of tracers, so is not appropriate here (S29). For this application, exchange
from the lower stratosphere to the upper troposphere is of concern as this corresponds to the altitude of
aircraft emissions. This is more accurately modeled and it is known that transport along isentropes
allows for cross-tropopause transport on a time scale of a few days (S30). Indeed, cruise altitude (~10
km) at 40-60°N corresponds to the 330-340 K isentropic surface identified by Chen (S30) as the
boundary below which stratosphere-troposphere exchange occurs vigorously year-round, and above
which a strong annual cycle is found with weak exchange in the winter.
We note that GEOS-Chem correctly reproduces vertical tropospheric and lower stratospheric ozone
and carbon monoxide profiles (13, S31) indicating that, notwithstanding excessive stratosphere-
troposphere exchange, vertical transport is modeled appropriately from the surface to the lower
stratosphere. Comparisons between measured and modeled 7Be – a natural aerosol tracer that is
produced in the lower stratosphere and upper troposphere – for both surface observations and vertical
profiles suggest GEOS-Chem reproduces vertical aerosol transport and scavenging (21), while records
of surface air concentrations of 7Be demonstrate that aerosol from the lower stratosphere and upper
troposphere can reach the surface. We note that uncertainties remain that may be important for
understanding the impact of aircraft cruise emissions on air quality. In particular, the potential
sensitivity of findings to model wet removal and vertical diffusion rates, and feedbacks between air
quality and climate have not been considered. These uncertainties will be considered in future work.
! ?"
In order to understand the possible contribution of atmospheric modeling uncertainties, we estimated
uncertainties in modeling as follows. It was assumed that the population-weighted #PM2.5 due to LTO
emissions – which is expected to be under-resolved – may be biased low by a factor of 1-2.3
[extrapolating from results of a recent US study (5)], with a mode of 1.3. Overall modeling bias of
population-weighted #PM2.5 was assumed to be high by up to 30%, or low by up to 50%. This is
broadly consistent with the modeled variability in aviation contributions to surface BC concentrations
discussed previously (-50% to +10%) and the estimated ±50% uncertainty in modeled sulfate
enhancement due to Asian emissions (for GEOS-Chem) estimate by Heald et al. in a transpacific
aerosol context (15). Park et al. found a correlation between measured and modeled annual average
sulfate concentrations of R2 = 0.91-0.94 in the US (18). The seasonal correlations were lower (R2 =
0.79-0.90), but an annual average is the appropriate comparison here. Annual average biases were -5%
to -9% over the IMPROVE and CASTNET monitoring sites considered. Modeled ammonium was
biased high by 33% with an R2 of 0.90 in the US, while nitrate concentrations were biased high by
87%-143% with R2 = 0.47-0.66. However, a subsequently improved treatment of ammonia dry
deposition has been identified as the probable cause of an improvement in the annual average
ammonium bias to +10%, although the fall over-estimates remain (16). This resulted in a knock-on
improvement to +30% for nitrate. Analysis of European EMEP sites by Park et al. (18) found that
GEOS-Chem modeled annual average sulfate concentrations with R2 = 0.72 and a bias of -2%, nitrate
concentrations with R2 = 0.48 and a bias of -4%, and ammonium concentrations with R2 = 0.63 and a
bias of +25%. In the case of BC, analysis of monitoring sites in the US gives a model R2 = 0.84 and a
bias of -15% (17). Overall comparisons suggest that PM2.5 concentrations attributable to all surface
emission are modeled by GEOS-Chem with a high bias of +10-15%. Recall that we are primarily
concerned with the marginal contribution of aviation emissions to PM2.5 exposure (i.e. population-
weighted PM2.5). While a +30% bound on model over-estimates may be reasonable on the basis of the
above discussion, we selected a -50% bound on model under-estimates of exposure on the basis that
averaging over grid-cells is likely to cause a low bias.
Health Impacts
We applied the WHO health impacts methodology described by Ostro (28). Specifically, relative risk
was given by
!
RR k ="A +1"B +1#
$ %
&
' (
)
,
where !!is the PM2.5 concentration in µg/m3 and subscripts A and B denote the case under study (with
aviation emissions included) and the background case (without aviation emissions), respectively. For
cardiopulmonary mortality " = 0.15515 (95% CI: 0.05624 – 0.2541) and for lung cancer-related
mortality " = 0.232179 (95% CI: 0.08563 – 0.37873). Premature mortalities are then calculated from
!
Mortalities =RR k -1RR kk
" BkPk,
! S"
where k is the population grid cell index from GPW (29) and corresponding PM2.5 concentrations are
used, Bk is the baseline health incidence rate (Table S2), and Pk is the exposed population (from GPW).
We have considered only long-term exposure to PM2.5. This is consistent with, for example, a US
Environmental Protection Agency (EPA) assessment of the benefits of the US Clean Air Act, which
noted that relying on long-term concentrations of PM2.5 as the exposure metric “is likely to result in a
more complete assessment of the effect of air pollution on mortality risk” (S32). It is also consistent
with the World Health Organization (WHO) assessment of the contribution of anthropogenic pollution
to premature mortality in 3211 cities worldwide (26), which made use of a regression approach based
on ambient monitoring stations in 304 cities rather than atmospheric modeling. Additional health
impacts attributable to aviation may occur due to short-term peaks in PM and O3 exposure. However,
the majority (10 out of 12) of experts in a US-based expert elicitation study on the relationship between
PM2.5 exposure and all-cause mortality believed that “short-term mortality impacts not captured in
[long-term] cohort studies represented a very small percentage of the total mortality impact” on the
basis of the strength of current evidence (23). Quantifying the temporal expression of mortality impacts
from short-term peaks in PM is more uncertain than the impact of long-term exposure (24).
There is limited evidence to support the choice of assuming differing levels of toxicity given differing
particulate matter compositions (S33,S34). Consistent with WHO and EPA methods, we do not apply
any composition-differentiated toxicity factors, although note that a European Commission project,
ExternE, did suggest such an approach (S35). The US-based expert elicitation study on the relationship
between PM2.5 exposure and all-cause mortality did not present composition influences explicitly in
their quantitative judgments (23), and an EU-based expert elicitation study noted that secondary
sulfates and nitrates may be less toxic than BC or OC (24). Overall, current evidence is insufficient to
support quantitative differentiation of the toxicity of PM species, so the concentration-functions strictly
apply to the overall mix of ambient fine particulate matter, of which the aviation-attributable fraction
may be different in composition. Nonetheless, a causal relationship between PM2.5 exposure and
premature mortality is thought more than 90% probable my most experts in the EU and US-based
elicitation studies (23,24).
We note that aircraft BC emitted at cruise has a geometric mean diameter of the same order or smaller
than diesel engine emissions (S9,S47), with geometric mean diameters from ~30 nm (aircraft) to ~100
nm (diesel engines), with an overlap between. Ship BC emissions are distinctly larger with a geometric
mean > 100 nm (S46). Differential toxicity related to size remains unquantified, but ultrafine BC may
have particularly high toxicity (S48).
As most cohort studies were performed in the US and Western Europe, there is additional uncertainty
in the application of concentration-response functions derived from these studies to other regions.
Experts in the EU-based expert elicitation study judged that differences in concentration-response
functions between the US and EU were small compared to their estimates of the uncertainty in either
response (24). Little information is available on other regions. We apply the same concentration-
response function parameters in all regions to determine relative risk. Also, we have based our relative
! T"
risk estimates on cardiopulmonary and lung cancer premature mortality separately as is consistent with
the WHO method (26), rather than on all-cause mortality.
Particle bound water corresponding to 30-40% relative humidity at 20-23°C (the laboratory
equilibration conditions) is included in US Federal Reference Method PM2.5 measurements, upon
which concentration-response functions are based. Assuming that (non-acidic) particle bound water has
no health impact, this may bias low concentration-response functions as compared with considering dry
ions only. We expect this bias to be ~10%, as the average ratio of particle bound water mass to
ammonium plus sulfate mass in the US is 0.24 (S36), and approximately half of the aviation
attributable PM2.5 in the US is ammonium sulfate.
Population density information is likely to have uncertainty associated with it, varying by region, and
baseline health incidence rates may suffer underreporting or misclassification, with probable regional
variability.
We calculated risks on a WHO sub-region basis using baseline risks compiled from WHO Global
Burden of Disease statistics (S37). Table S2 shows our estimates of baseline cardiopulmonary and lung
cancer mortality rates for adults over 30. Cardiopulmonary diseases included upper and lower
respiratory infections, hypertensive heart disease, ischaemic heart disease, inflammatory heart diseases,
chronic obstructive pulmonary diseases and asthma.
Overall we find that 88% of aviation-attributable premature mortalities are due to cardiopulmonary
diseases. Table S3 shows nominal premature mortality estimates by country, a #PM2.5 mass breakdown
and the percentage of premature mortalities attributable to LTO emissions. Next, the aviation fuel burnt
above the land area of each country relative to all fuel burnt above land globally is shown as a
percentage. This was based on a gridded dataset of country codes (29) at 2.5’"2.5’ resolution, upon
which the original 1°"1° AEDT emissions inventory was projected. The percentage of fuel burn below
3000 ft in each country is also given. A correlation between the percentage fuel burn in the LTO cycle
and the percentage of mortalities attributable to LTO emissions is observed. Finally, for each country
an ‘Equity Ratio’ (ER) is calculated, which is defined as the percentage of aviation-attributable
mortalities that a country incurs divided by the percentage of global fuel burn above the country (as
previously defined). When ER > 1, a country incurs a higher proportion of mortalities than the fuel
burn in its airspace alone would imply, while when ER < 1, a country incurs fewer mortalities than the
fuel burn and emissions in its airspace would indicate. Note that the global fuel burn metric has been
defined to emphasize the role of net transport of aircraft pollution. For example, the US has ER = 0.1,
with cruise emissions at ~10 km being exported over the Atlantic and towards Europe, while the bulk
of European emissions do not impact the US In the case of India, ER = 7.4, as it is impacted by aircraft
pollution from North America and Europe, combined with relatively high ammonia concentrations and
population densities. Several major European countries, namely Germany, France, the U.K., Italy and
Spain, have ER ~ 1. As we have shown that North American emissions impact Europe, and European
emissions impact South-East Asia, an interpretation is that Europe ‘breaks-even’.
! W"
!"#$%& '*)& L2.,1/5," :279/4601-4527;" 259" 105>" :25:,7" -47321/3;" 723,." D;" ^]'" .0D)7,>/45" /5"9,23+._!VVV<" *40537/,." /5" ,2:+" .0D)7,>/45" 27," 1/.3,9I"A+/:+" 27," >7406,9"D;" >,4>726+/:" 7,>/45" 259"6460123/45"9/.,2.,":+272:3,7/.3/:.<"
+,-&.%/012345#6.%/012&"27&81529.0%4& :".701;5$<12".=& >52/&8"28%.&
?@A3BC"E1>,7/2I"E5>412I"L,5/5I"L07M/52"[2.4I"*2-,7445I"*26,"`,79,I"*+29I"*4-474.I"&C02347/21"%0/5,2I"%2D45I"%2-D/2I"%+252I"%0/5,2I"%0/5,2)L/..20I"N/D,7/2I"H292>2.:27I"H21/I"H207/325/2I"H207/3/0.I"Z/>,7I"Z/>,7/2I"(24"=4-,"259"G7/5:/6,I"(,5,>21I"(,;:+,11,.I"(/,772"N,45,I"=4>4<"
O<S#" V<VW"
?@A3DC"L43.A252I"L07059/I"*,53721"E87/:25"F,60D1/:I"*45>4I"*a3,"9bYK4/7,I"J,-4:723/:"F,60D1/:"48"3+,"*45>4I"&7/37,2I"&3+/46/2I"c,5;2I"N,.43+4I"H212A/I"H4d2-D/C0,I"Z2-/D/2I"FA2592I"(403+"E87/:2I"(A2d/1259I"\>2592I"\5/3,9"F,60D1/:"48"=25d25/2I"e2-D/2I"e/-D2DA,<"
O<?#" V<!V"
?EA3?C"*25292I"*0D2I"\5/3,9"(323,."48"E-,7/:2<"" O<TS" V<X#"?EA3FC"E53/>02"259"L27D092I"E7>,53/52I"L2+2-2.I"L27D294.I"L,1/d,I"L72d/1I"*+/1,I"*414-D/2I"*4.32"F/:2I"J4-/5/:2I"J4-/5/:25"F,60D1/:I"&1"(21K2947I"%7,5292I"%0;252I"]459072.I"f2-2/:2I"H,P/:4I"G252-2I"G272>02;I"(2/53"c/33."259"Z,K/.I"(2/53"N0:/2I"(2/53"`/5:,53"259"3+,"%7,529/5,.I"(07/52-,I"=7/5/929"259"=4D2>4I"\70>02;I"`,5,d0,12<""
$<?W" V<#S"
?EA3BC"L41/K/2I"&:02947I"%023,-212I"]2/3/I"Z/:272>02I"G,70<" $<$T" V<!!"DEA3FC"L2+72/5I"*;670.I"Y725"@Y.12-/:"F,60D1/:"48BI"f47925I"c0A2/3I"N,D2545I"N/D;25"E72D"f2-2+/7/;2I"'-25I"g2327I"(209/"E72D/2I"(;7/25"E72D"F,60D1/:I"=05/./2I"\5/3,9"E72D"&-/723,.<""
O<#T" V<!X"
DEA3BC"E8>+25/.325I"Jh/D403/I"&>;63I"Y72CI"H474::4I"G2M/.325I"(4-21/2I"(0925I"i,-,5<""
?<SO" V<!?"
DGA3?C"E594772I"E0.37/2I"L,1>/0-I"*7423/2I"*d,:+"F,60D1/:I"J,5-27MI"[/51259I"[725:,I"%,7-25;I"%7,,:,I"Y:,1259I"Y7,1259I"Y.72,1I"Y321;I"N0P,-D407>I"H2132I"H452:4I"Z,3+,71259.I"Z47A2;I"G4730>21I"(25"H27/54I""(14K,5/2I"(62/5I"(A,9,5I"(A/3d,71259I"\5/3,9"c/5>94-<""
O<$O" V<TT"
DGA3FC"E1D25/2I"E7-,5/2I"Ed,7D2/h25I"L4.5/2"259"],7d,>4K/52I"L01>27/2I"%,47>/2I"c;7>;d.325I"G41259I"F4-25/2I"(14K2M/2I"=2h/M/.325I"=+,"[47-,7"i0>4.12K"F,60D1/:"48"H2:,945/2I"=07M,;I"=07M-,5/.325I"\dD,M/.325I"i0>4.12K/2<""
S<!$" V<?T"
DGA3:C"L,1270.I"&.345/2I"]05>27;I"c2d2M+.325I"N23K/2I"N/3+025/2I"F,60D1/:"48"H4194K2I"F0../25"[,9,723/45I"\M72/5,<""
!V<$W" V<SX"
'D?A3FC"Y5945,./2I"(7/"N25M2I"=+2/1259<"" O<OT" V<$$"'D?A3BC"L25>129,.+I"L+0325I"J,-4:723/:"G,461,b."F,60D1/:"48"c47,2I"Y59/2I"H219/K,.I"H;25-27I"Z,621I"=/-47"N,.3,<""
S<$O" V<#S"
+HA3?C"E0.3721/2I"L705,/"J270..212-I"f2625I"Z,A"e,21259I"(/5>2647,<" #<XW" V<SS"+HA3FC"*2-D49/2I"*+/52I"*44M"Y.1259.I"[/h/I"c/7/D23/I"N24"G,461,b."J,-4:723/:"F,60D1/:I"H212;./2I"H27.+211"Y.1259.I"H/:745,./2"@[,9,723,9"(323,."48BI"H45>41/2I"Z2070I"Z/0,I"G2120I"G2602"Z,A"%0/5,2I"G+/1/66/5,.I"F,60D1/:"48"c47,2I"(2-42I"(414-45"Y.1259.I"=45>2I"=0K210I"`250230I"`/,3"Z2-<"
$<WV" V<OS"
! X"
!"#$%&'I)&'5"2"6,7):40537;"D2./." 3+," 32D1,"D,14A".+4A.j&,.3/-23,9"54-/521"67,-2307,"-47321/3/,.k"3+,"6,7:,532>,"48"2K/23/45"2337/D032D1,"GH#<?"90,"34"N='",-/../45."6,7":40537;k"3+,"6,7:,532>,"48"L*I"'*I" .01823,I" 5/3723," 259" 2--45/0-"6460123/45)A,/>+3,9" GH#<?"-2..k" 3+," 6,7:,532>,"48" >14D21" 80,1"D075"4::077/5>"4K,7"3+,"1259"27,2"48",2:+":40537;k"3+,"6,7:,532>,"48"80,1"/5",2:+":40537;"D0753"/5"3+,"N='":;:1,k"259"3+,"&C0/3;"F23/4"847",2:+":40537;<"
""
H%.8%29"/%&1J&#HE*)K&<"44&"99.0#59"#$%&91L&
&!1
9"$&E
1.9"$090%4&
>!-&
F:&&
-:&&
'5$J"9%&&
M09."9%&&
?<<12
05<&&
H%.8%2
9"/%&1J&N
$1#"
$&@5%$&F5.2&-O%.&>"27
&
H%.8%2
9"/%&1J&@5%
$&F5.2&02&>!-
&:=8$%&&
DP509=&A"901&&
QR&1J&!
19"$&E
1.9"$090%4&
C&R&1J&!
19"$&@5%
$&F5.2S&
*+/52" !WXO" #!<S" V<!" V<#" !V<V" SS<S" #$<!" T<S" !#<W" #<?"Y59/2" !SOV" !!<W" V<O" V<#" $T<$" $T<!" #O<W" #<#" W<$" T<O"%,7-25;" ?O?" #?<X" V<#" V<T" ?<!" T!<$" ##<T" $<?" !X<#" !<S"\5/3,9"(323,."48"E-,7/:2" O?W" !S<#" !<T" !<O" OS<T" $V<$" !X<X" $O<?" !#<?" V<!"F0../2" O#?" !S<T" V<$" V<$" #$<?" ?#<T" #$<#" O<S" #<O" V<X"[725:," $T?" #W<X" V<$" !<!" S<T" SX<$" ##<T" $<?" !$<$" !<!"\5/3,9"c/5>94-" $S#" #T<X" V<$" V<W" S<X" SX<$" ##<T" #<?" #W<V" !<?"G2M/.325" $#X" !#<#" V<?" V<#" $T<$" $T<O" #O<S" V<T" ?<W" O<T"\M72/5," #T!" ##<#" V<!" V<#" !!<!" S?<O" #$<#" V<T" O<$" O<#"Y321;" #OS" #S<W" V<$" !<!" !#<S" S$<$" ##<T" !<?" ##<T" !<T"f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c47,2" SX" OS<T" V<#" V<#" !#<#" SO<O" #$<V" V<?" #S<V" !<O"*25292" ST" )OV<?" V<?" V<O" #?<!" ?#<$" #!<S" O<X" O<W" V<!"L,1270." SS" ##<T" V<!" V<#" X<S" ST<V" #$<!" V<#" !<T" $<#"Z,621" S?" S<S" V<#" V<!" !S<T" ?X<O" #$<S" V<V" O<$" #!<S"F4-25/2" S!" !W<$" V<!" V<#" !!<V" S?<?" #$<!" V<?" $<#" !<!"\dD,M/.325" ?X" S<O" V<S" V<$" OO<X" $V<#" #$<X" V<!" ?<O" T<#"*d,:+"F,60D1/:" ?W" #$<!" V<#" V<?" ?<X" TV<T" ##<T" V<O" !!<#" !<O"c2d2M+.325" ??" W<O" V<?" V<$" $W<!" $W<$" ##<W" V<O" $<W" !<$"E8>+25/.325" ?O" S<#" V<S" V<$" O!<#" $$<X" #O<V" V<$" V<$" #<!"`/,3"Z2-" ?V" #W<#" V<?" V<O" OS<?" #T<$" #?<$" V<$" !?<!" !<W"(A/3d,71259" O?" #S<X" V<#" !<V" O<?" T!<T" ##<?" V<O" #!<#" !<#"E0.37/2" OO" #$<#" V<#" V<S" ?<X" TV<S" ##<T" V<?" !V<S" V<W"c47,2"@JGFc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
! !V"
H%.8%29"/%&1J&#HE*)K&<"44&"99.0#59"#$%&91L&
&
!19"$&E
1.9"$090%4&
>!-&
F:&&
-:&&
'5$J"9%&&
M09."9%&&
?<<12
05<&&
H%.8%2
9"/%&1J&N
$1#"
$&@5%$&F5.2&-O%.&>"27
&
H%.8%2
9"/%&1J&@5%
$&F5.2&02&>!-
&:=8$%&&
DP509=&A"901&&
QR&1J&!
19"$&E
1.9"$090%4&
C&R&1J&!
19"$&@5%
$&F5.2S&
=+2/1259" #X" #W<S" !<!" $<!" ?W<S" !!<W" #?<$" V<T" #!<W" V<O"i,-,5" #X" S<#" V<W" V<?" TO<S" V<?" #$<?" V<V" #<!" T<X"(209/"E72D/2" #X" T<T" !<V" V<T" WV<S" !<!" !S<S" !<V" !!<$" V<$"J,5-27M" #W" #S<X" V<#" V<?" T<!" SX<$" ##<W" V<#" #O<#" !<O"(A,9,5" #S" #W<W" V<$" V<T" !$<$" S#<S" #$<!" V<?" #!<#" V<?"G4730>21" #S" $V<X" V<S" $<?" $S<T" $W<$" #V<X" V<#" !!<!" !<!"(14K2M/2" #?" !W<W" V<!" V<$" T<V" SX<W" ##<W" V<#" T<W" !<O"L72d/1" #?" $W<X" $<V" #<!" S!<V" W<S" #?<#" #<!" !!<$" V<!"N/3+025/2" #?" #O<$" V<!" V<$" X<T" SS<W" #$<V" V<!" X<?" $<X"H;25-27" #?" !$<O" V<X" V<W" OS<#" #T<V" #?<!" V<#" #<?" !<V"=2h/M/.325" ##" ?<S" V<?" V<$" $O<T" OV<O" #O<!" V<V" #<#" !W<W"F,60D1/:"48"H4194K2" #V" #O<S" V<#" V<$" !$<S" S#<S" #$<$" V<V" $<W" ?<#"c;7>;d"F,60D1/:" !X" W<?" V<?" V<#" #W<O" OT<O" #$<?" V<V" !<?" X<X"%7,,:," !X" !X<T" V<?" V<T" OS<?" #W<O" #$<X" V<O" !!<W" V<?"*7423/2" !X" #V<$" V<#" V<O" X<?" ST<V" ##<X" V<#" O<O" !<V"(;7/25"E72D"F,60D1/:" !X" !?<T" V<S" V<?" ?T<V" !T<O" #O<?" V<!" W<W" !<?"*0D2" !W" O<T" V<W" V<?" TS<S" !<$" #V<X" V<!" ?<S" !<X"Y7,1259" !T" #X<#" V<$" V<?" !#<!" SO<O" ##<W" V<$" !X<!" V<S"L01>27/2" !T" !S<V" V<$" V<O" ##<X" ?#<X" #$<S" V<#" S<$" V<T"Y.72,1" !?" !S<X" V<W" V<W" T#<S" O<!" #!<W" V<V" $T<O" S<W"L4.5/2)],7d,>4K/52" !O" !W<X" V<#" V<O" !$<V" S$<O" #$<!" V<#" #<#" V<X"Ed,7D2/h25" !$" !!<!" V<?" V<O" O#<S" $!<W" #O<T" V<!" $<$" #<$"Z/>,7" !$" #<S" V<W" V<?" T#<?" V<#" #?<X" V<!" V<S" !<V"Z47A2;" !#" $$<?" V<O" V<T" !S<S" ?X<$" #$<V" V<$" $#<V" V<O"(7/"N25M2" !#" !!<S" V<T" V<O" SX<W" #<$" #S<W" V<V" #V<$" ?<X"]45>"c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`,5,d0,12" W" X<#" #<$" !<O" SW<S" #<V" #?<T" V<#" !S<$" V<O"c,5;2" T" $<?" !<W" V<X" ST<X" #<X" #S<O" V<!" #<V" V<?"*2-D49/2" T" !S<X" V<W" V<?" SS<S" ?<?" #S<S" V<!" !O<X" V<X"J4-/5/:25"F,60D1/:" T" !V<O" V<T" V<?" T?<$" !<#" ##<$" V<V" $O<$" !<W"N/D;25"E72D"f2-2+/7/;2" S" !!<W" V<W" !<V" W!<$" #<$" !O<?" V<#" S<$" V<$"*+29" S" $<O" V<W" V<?" T!<X" V<W" #S<V" V<!" !<#" V<T"(,5,>21" S" T<O" V<S" V<O" WO<T" )V<#" !O<?" V<V" !!<T" $<!"(4-21/2" S" #<$" V<X" V<O" T!<!" V<W" #S<S" V<V" !<O" !<$"f47925" S" !?<W" V<W" V<T" T$<S" $<W" #!<!" V<V" !O<T" !<#"&.345/2" S" ##<O" V<#" V<?" !?<S" SV<$" #$<O" V<!" O<V" V<X"[iF'H_"H2:,945/2" ?" !S<V" V<$" V<O" #$<X" ?!<X" #$<?" V<!" O<?" !<V"(/5>2647," ?" ?W<W" V<W" !<!" T?<O" V<T" ##<V" V<V" OT<$" #<X"
! !!"
H%.8%29"/%&1J&#HE*)K&<"44&"99.0#59"#$%&91L"
"
!19"$&E
1.9"$090%4"
>!-&
F:&&
-:&&
'5$J"9%&&
M09."9%&&
?<<12
05<&&
H%.8%2
9"/%&1J&N
$1#"
$&@5%$&F5.2&-O%.&>"27
&
H%.8%2
9"/%&1J&@5%
$&F5.2&
02&>!-
&:=8$%&&
DP509=&A"901&&
QR&1J&!
19"$&E
1.9"$090%4&C&
R&1J&!
19"$&@5%
$&F5.2S&
\>2592" ?" #<T" !<O" V<W" SO<V" T<?" #S<#" V<V" V<!" #<#"*414-D/2" ?" T<V" !<O" !<!" SO<T" T<V" #?<W" V<#" V<W" V<#"]2/3/" ?" X<V" V<T" V<?" T?<X" !<V" #!<X" V<V" X<V" $<!"E0.3721/2" ?" SV<!" #<?" T<O" ??<O" !#<V" ##<T" #<#" X<O" V<V"=25d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c0A2/3" #" W<!" !<V" V<S" TT<V" !<T" !X<T" V<V" O!<X" V<?"Z/:272>02" #" O<W" V<T" V<$" T?<V" V<X" #$<V" V<V" X<$" V<T"*4.32"F/:2" #" S<T" V<O" V<#" W#<$" V<O" !S<T" V<V" !S<!" V<W"f2-2/:2" #" S<X" V<T" V<O" TT<X" V<S" #V<$" V<V" $V<#" !<!"(/,772"N,45," #" O<$" V<S" V<$" TO<W" V<W" #$<?" V<V" W<X" O<T"G,70" #" !X<T" !<S" V<T" T!<!" !<T" #O<X" V<#" O<X" V<!"L,5/5" !" S<$" !<#" V<S" TV<X" V<W" #S<O" V<V" !#<!" V<X"=4>4" !" O<S" !<#" V<S" TV<S" V<X" #S<T" V<V" !O<#" !<#"H4d2-D/C0," !" W<S" !<O" V<T" SW<?" O<V" #?<?" V<!" X<#" V<#"G2602"Z,A"%0/5,2" !" O<X" !<O" V<S" WV<W" V<!" !T<!" V<V" !#<X" V<$"G252-2" !" T<#" V<X" V<?" T?<?" !<$" #!<W" V<V" #V<$" V<O"L41/K/2" !" #!<$" !<S" V<?" T#<S" !<O" #$<W" V<!" ?<S" V<!"Z,A"e,21259" !" ?W<!" !<!" !<W" SV<#" !$<?" #$<$" V<!" #W<S" V<!"e/-D2DA," !" !$<X" !<#" V<T" ST<V" ?<V" #S<V" V<!" V<?" V<#"H292>2.:27" !" !!<O" !<V" V<?" T$<#" V<$" #O<X" V<V" $<V" V<$"H2132" !" !S<$" V<X" !<$" TX<?" O<V" !O<O" V<V" ?#<#" !!<T"%2-D/2" !" T<?" V<S" V<O" W?<O" )V<#" !$<X" V<V" #?<T" $<X"FA2592" !" S<V" !<X" V<X" SO<S" S<?" #S<!" V<V" !<!" #<T"*;670." !" #!<V" V<W" V<W" ST<X" ?<T" #O<W" V<V" ##<T" V<O"H212A/" !" !<W" !<?" !<$" SW<!" #<W" #S<O" V<V" #<?" !<!"%0/5,2)L/..20" !" S<?" V<?" V<$" W!<T" V<#" !T<O" V<V" O<?" #<?"Jh/D403/" !" ?<?" V<X" V<?" T!<V" V<X" #S<T" V<V" W<O" $<V"f,7.,;" !" $V<W" V<?" !<$" ?<!" TV<T" ##<?" V<V" !!<W" ?<O"L2+72/5" !" !#<T" !<!" V<T" W#<T" !<!" !O<O" V<V" $O<#" !<#"E5>412" !" V<O" !<#" V<?" ?V<?" ##<O" #?<O" V<!" #<!" V<!"N/D,7/2" !" $<!" V<T" V<?" T!<T" V<W" #S<#" V<V" !V<S" !<?"
! !#"
H%.8%29"/%&1J&#HE*)K&<"44&"99.0#59"#$%&91L"
"
!19"$&E
1.9"$090%4"
>!-&
F:&&
-:&&
'5$J"9%&&
M09."9%&&
?<<12
05<&&
H%.8%2
9"/%&1J&N
$1#"
$&@5%$&F5.2&-O%.&>"27
&
H%.8%2
9"/%&1J&@5%
$&F5.2&02&>!-
&:=8$%&&
DP509=&A"901&&
QR&1J&!
19"$&E
1.9"$090%4&
C&R&1J&!
19"$&@5%
$&F5.2S&
e2-D/2" !" O<W" !<$" V<?" S?<W" S<T" #?<T" V<!" V<?" V<!"L07059/" !" ?<T" !<X" V<W" SO<S" S<?" #S<!" V<V" !!<T" $<T"&:02947" !" SW<V" !!<S" ?<!" OX<?" !!<T" ##<V" V<!" !#<W" V<!"g2327" !" !#<T" !<!" V<T" W#<T" !<!" !O<O" V<V" $T<!" V<!"E594772" !" $$<T" V<$" #<S" X<V" S?<S" ##<?" V<V" V<V" O<X"!19"$3?O%."/%& TTU*& *V)*& V)W& V)X& *X)X& WT)(& *I)*& & & &
"
E" 7/.M" 48" 8,A,7" 3+25" V<?" 67,-2307,"-47321/3/,." 6,7" ;,27" A,7," ,.3/-23,9" 847j" =7/5/929" 259" =4D2>4I"H2:24I" *,53721" E87/:25"F,60D1/:I" Y.1," 48" H25I" G272>02;I" *26," `,79,I" %0,75.,;I" %029,1406,I" \70>02;I" Y:,1259I" N/,:+3,5.3,/5I" L2+2-2.I" H219/K,.I"H273/5/C0,I" \(" `/7>/5" Y.1259.I" %02-I" N,.43+4I" Z2-/D/2I" %/D721327I" L27D294.I" L705,/" J270..212-I" L43.A252I" %0;252I" *45>4I"J<F<I" Z,3+,71259" E53/11,.I" (A2d/1259I" H452:4I" [/h/I" &2.3" =/-47I" [2,74," Y.1259.I" (414-45" Y.1259.I" H207/3/0.I" (07/52-,I" (2/53"N0:/2I"Z473+,75"H27/252"Y.1259.I"L,1/d,I"(25"H27/54I"[,9<"(323,"48"H/:745,./2I"F,05/45I"E70D2I"*4-474.I"%2D45I"(2/53"`/5:,53I"`250230I"%7,5292I"L,7-092I"[7,5:+"%0/252I"[7,5:+"G41;5,./2I"E53/>02"259"L27D092I"L7/3/.+"`/7>/5"Y.1259.I"*4--<"48"J4-/5/:2I"*2;-25"Y.1259.I"Z,A"*21,945/2I"^,.3,75"(2-42I"&C02347/21"%0/5,2I"H27.+211"Y.1259.I"(2/53"c/33."259"Z,K/.I"c/7/D23/I"=07M."259"*2/:4." Y.1259.I" (,;:+,11,.I" %7,,51259I" (24" =4-," 259" G7/5:/6,I" =45>2I" H2;433,I" E-,7/:25" (2-42I" E5>0/112I" G2120I" Z2070I"H453.,7723I"*44M"Y.1259.I"^211/."259"[03052I"=0K210I"(2/53"G/,77,"259"H/C0,145I"(K21D279I"(2/53"],1,52I"=4M,120I"Z/0,I"Z47841M"Y.1259I"[21M1259"Y.1259."259"G/3:2/75"
Comparison With Previous Studies
Results were compared to the recent US study of the regional-scale air quality impacts of aviation in
the US that used methods consistent with US EPA practice for regulatory impact analyses (5,S38,S39).
In order to compare results, we performed a full simulation for US LTO-only aircraft emissions and
applied the health impacts methodology previously described. Our central estimate for premature
mortalities in the US due to US LTO emissions is 70-100 (for GEOS-3 and GEOS-5 simulations),
which compares with 210 calculated using high (~36 km) resolution CMAQ simulations of the impact
of US emissions on US premature mortality (30). Accounting for the 50% higher central C-R function
gradient applied by Brunelle-Yeung (30), we find that GEOS-Chem underestimates the impact of LTO
emissions by a factor of 1.4-2.0 relative to the higher resolution CMAQ simulation. If we assume that
globally LTO impacts are underestimated by a factor of 1.4-2.0, this yields an estimate of 26-34% of
aircraft-attributable premature mortalities being due to LTO emissions. However, we note that the
underestimate of LTO impacts is likely to be lower in countries outside of Europe and North America,
where the density of airports is lower and a higher fraction of aircraft-attributable PM is transboundary.
We also compare our global modeling of BC concentrations with Hendricks et al. (S40), who present
zonally averaged plots suitable for comparison. At (30°S,EQ,30°N,60°N) we calculate a zonally
averaged ground-level BC concentrations of (0.04,0.10,1.1,0.37) " 10-4 µg/m3, compared to
(0.05,0.10,0.9,0.20) " 10-4 µg/m3 from the ‘BASE’ simulation of Hendricks et al. Total BC emissions
in our study are a factor of 1.7 higher than Hendricks et al. Dividing our estimates by 1.7 for
comparison with Hendricks et al., we find the ratio of our estimates to theirs are (0.5,0.6,0.7,1.1) at the
same latitudes. A second ‘NOICE’ simulation presented by Hendricks et al. results in similar estimates
for ground-level BC concentrations in the northern hemisphere, but approximately a factor of two
higher at 30°S.
! !$"
Finally, we compare with the simulations of Tarrasón et al. (7), who present percentage change in
inorganic aerosol concentrations in Europe due to global non-LTO emissions. Tarrasón et al.
perturbations in the range 0.5-1% over the UK, France, Germany and northern and central Spain. Our
estimates are around 0.4-0.9%; however, our NOx emissions inventory is 20% higher.
Overall our estimates for the ground-level PM2.5 perturbation due to aircraft cruise emissions are
consistent with, or biased low by ~30%, as compared with results from older studies linearly scaled to
the inventories we have applied. The contribution of LTO emissions is likely underestimated in our
analysis, which has been concerned with the impact of cruise emissions on human health globally.
Parametric Uncertainty Analysis
We performed a parametric uncertainty analysis accounting for uncertainties in aircraft emissions, and
estimates for some modeling uncertainties. Results from simulations described in Table S1, also
presented in Tables S4, were used to construct a multi-dimensional interpolation (covering overall fuel
burn, LTO fuel burn, EI NOx, EI SOx, !, EI BC, EI OC). The output of the multi-dimensional
interpolation was aviation-attributable human exposure to PM2.5 normalized for baseline health risk. As
the interpolation was limited due to the number of GEOS-Chem simulations that could practically be
carried out, certain assumptions were made. Linear interpolation was applied between simulated data
points. Background NH3 emissions were not treated as an uncertain variable due to strong interactions
with aircraft NOx and SOx emissions in terms of secondary PM formation, with results instead being
presented to illustrate sensitivity. Changes in BC and OC concentrations were assumed to be linear in
corresponding aircraft emissions. Coupling between BC and OC with sulfate and nitrate concentrations
was neglected. ! was assumed only to effect sulfate concentrations. Uncertainty in fuel burn was
treated as a proportionate change in all aircraft emission indices. Uncertainty in the premature mortality
concentration-response function gradient was applied multiplicatively to modeled changes in
population exposure to PM2.5, as were estimates in atmospheric modeling uncertainty. All uncertain
input parameters were assigned triangular probability density functions, with minimum and maximum
values corresponding to the low and high values of the ranges previously given, and modes
corresponding to nominal values. This was on account of lack of basis for an alternative distribution
shape, and since triangular distributions are bounded. A Monte Carlo integration was used to estimate
the probability density function for the number of premature mortalities attributable to aircraft
emissions per year. To calculate the result for each member, a draw for each of the uncertain input
parameters was made from the respective triangular distribution, the multi-dimensional interpolation
was applied to determine the population exposure to PM2.5 attributable to aviation, and this was
multiplied by draws for multiplicative atmospheric modeling uncertainty and concentration-response
function gradient.
&&&&&
! !O"
!"#$%&'W)&Z4-/521"67,-2307,"-47321/3;",.3/-23,."874-"./-0123/45.".6255/5>"1/M,1;"&Y"Z'!"259"&Y"('!"K210,.<""
" " DY&M-!&
" " >1Z& M1<02"$& ,0/[&
>1Z& TI##$" XI$TO" !!I?T?"
M1<02"$& TISX$" XIXT#" !!I#?V"
DY&'-
!&,0/[& TIX$W" XIW!!" !!IXX#"
Results of uncertainty analyses considering only emissions uncertainty, and also accounting for
estimated modeling uncertainties are given in the main report. The numerically estimated probability
distribution for the latter case is shown graphically in Fig. S1. To illustrate the relative importance of
the different sources of uncertainty, we calculated the variance of the conditional expectation for each
uncertain parameter considered. (As the parameter interpolation in close to linear, first order
consideration of contribution to variance sufficed.) Results are shown in Fig. S2. It is apparent that the
concentration-response (C-R) function dominates uncertainty, with overall modeling uncertainty the
second highest contributor. The likely low bias in resolving LTO impacts contributes <5% to total
variance. Fuel sulfur to SVI conversion efficiencies have a negligible effect (<3%), while the EI NOx is
the most significant source of uncertainty related to the aircraft emissions inventory.
@0/5.%&'()&=+,",.3/-23,9"674D2D/1/3;"9/.37/D03/45" 847" 3+,"50-D,7"48"67,-2307,"-47321/3/,."6,7" ;,27"2337/D032D1,"34"2K/23/45">14D211;<"
! !?"
@0/5.%&'*)&[/7.3"479,7":4537/D03/45"34"K27/25:,"48",2:+"48"3+,"05:,732/5"6272-,3,7.":45./9,7,9<""
Future Work
Future research may reduce modeling uncertainties or refine our understanding of them. We first
discuss future research needs to further understand the importance of aircraft cruise emissions in
degrading air quality, especially relative to LTO emissions. In particular, multi-scale modeling
applying a nested regional model within a global atmospheric model will help to further resolve the
impact of LTO emissions versus cruise emissions. Regional nested models are needed in regions of
high airport density, such as the US, where a global model is unlikely to capture the impacts of LTO
emissions on population exposure. We anticipate that this will have a smaller effect on estimates of
exposure to secondary PM relative to primary PM, as secondary PM formation is regional in nature so
increased resolution will have relatively less effect. Since even a regional atmospheric model is likely
to under-resolve the impact of primary PM emissions on populations – as airport locations are often
correlated with regions of high population density – the authors are developing an approach to capture
sub-grid exposure to primary PM that can be applied rapidly to all major airports (S41, S42). We note
that for LTO impacts to exceed 50% of total aviation impacts, the LTO-related mortalities would have
to have been underestimated by a factor of four. This is plausible for primary PM where analyses have
shown that accounting for sub-grid variability can double exposure to primary pollutants (S43), a factor
of two is more likely for secondary species by comparison with the US study mentioned earlier
(5,S38,S39), and then only in regions of high airport density such as the US. As mortalities appear
dominated by secondary PM – subject to changes in our understanding of the relative toxicity of
different PM species – the conclusion that cruise emissions are more important than LTO emissions in
terms of their impact on human health appears to be a rational working hypothesis in light of the
analysis of this paper.
We note that secondary organic aerosols (SOA) have not been considered in detail in this paper and
will be considered in future work. However, applying the GEOS-Chem SOA code [described in Liao et
al. (S44)], we conduced sensitivity simulations for given background and aviation (cruise + LTO)
emissions using 2006 meteorology. Including the aviation-attributable difference in SOA mass (along
with the aforementioned PM components) in the C-R function resulted in premature mortalities
! !S"
increasing by 3%. In addition to the difficulty of modeling SOA yields, we also note that the relative
toxicity of SOA is unknown.
References and Notes (S1) Yu, K. N.; Cheung, Y. P.; Cheung, T.; Henry, R. C. Identifying the impact of large urban airports on local air quality by nonparametric regression. Atmos. Environ. 2004, 38 (27), 4501-4507.
(S2) Carslaw, D. C.; Beevers, S. D.; Ropkins, K.; Bell, M. C. Detecting and quantifying aircraft and other on-airport contributions to ambient nitrogen oxides in the vicinity of a large international airport. Atmos. Environ. 2006, 40 (28), 5424-5434.
(S3) Schurmann, G.; Schafer, K.; Jahn, C.; Hoffmann, H.; Bauerfeind, M.; Fleuti, E.; Rappengluck, B. The impact of NOx, CO and VOC emissions on the air quality of Zurich airport. Atmos. Environ. 2007, 41 (1), 103-118.
(S4) Unal, A.; Hu, Y.; Chang, M. E.; Odman, M. T.; Russell, A. G. Airport related emissions and impacts on air quality: Application to the Atlanta International Airport., Atmos. Environ. 2005, 39 (32), 5787-5798.
(S5) Pueschel, R. F.; Verma, A.; Ferry, G. V.; Howard, S. D.; Vay, S.; Kinne, S. A.; Goodman, J.; Strawa, A. W. Sulfuric acic and soot particle formation in aircraft exhaust. Geophys. Res. Lett. 1998, 25 (10), 1685-1688.
(S6) Miake-Lye, R. C.; Anderson, B. E.; Cofer, W. R.; Wallio, H. A.; Nowicki, G. D.; Ballenthin, J. O.; Hunton, D. E.; Knighton, W. B.; Miller, T. M.; Seeley, T. M.; Viggiano, A. A. SOx oxidation and volatile aerosol formation in aircraft exhaust plumes depend on fuel sulfur concentration. Geophys. Res. Lett. 1998, 25 (10), 1677-1680.
(S7) Hagen, D. E.; Whitefield, P. D.; Schlager, H. Particulate emissions in the exhaust plume from commercial jet aircraft under cruise conditions. J. Geophys. Res. 1996, 101 (D14), 19551-19557.
(S8) Hagen, D. E.; Whitefield, P. D.; Paladino, J.; Treblood, M. Particle sizing and emission indices for a jet engine exhaust sampled at cruise. Geophys. Res. Lett. 1998, 25 (10), 1681-1684.
(S9) Petzold, A.; Döpelheuer, A.; Brock, C. A.; Schröder, F. In situ observations and model calculations of black carbon emissions by aircraft at cruise altitude. J. Geophys. Res. 1999, 104 (D18), 22171-22181.
(S10) Hendricks, J.; Kärcher, B.; Döpelheuer, A.; Feichter, J.; Lohmann, U.; Baumgardner, D. Simulating the global atmospheric black carbon cycle: a revisit to the contribution of aircraft emissions. Atmos. Chem. Phys. 2004, 4 (11-12), 2521-2541.
(S11) Arnold, F.; Stilp, TH.; Busen, R.; Schumann, U. Jet engine exhaust chemiion measurements: implications for gaseous SO3 and H2SO4. Atmos. Environ. 1998, 32 (18), 3073-3077.
(S12) Schumann, U.; Arnold, F.; Busen, R.; Curtius, J.; Kärcher, B.; Kiendler, A.; Petzold, A.; Schlager, H.; Schröder, F.; Wohlfrom, K.-H. Influence of fuel sulfur on the composition of aircraft exhaust plumes: The experiments SULFUR 1-7. J. Geophys. Res. 2002, 107 (D15), 4247.
(S13) Yu, F.; Turco, R. P.; Kärcher, B. The possible role of organics in the formation and evolution of ultrafine aircraft particles. J. Geophys. Res. 1999, 104 (D4), 4079-4087.
(S14) Curtius, J.; Arnold, F.; Schulte, P. Sulfuric acid measurements in the exhaust plume of a jet aircraft in flight: Implications for the sulfuric acid formation efficiency. Geophys. Res. Lett. 2002, 29 (7), 1113.
(S15) Brock, C. A.; Schröder, F.; Kärcher, B.; Petzold, A.; Busen, R.; Fiebig, M. Ultrafine particle size distributions measured in aircraft exhaust plumes. J. Geophys. Res. 2000, 105 (D21), 26555-26567.
! !T"
(S16) Kärcher, B.; Turco, R. P.; Yu, F.; Danilin, M. Y.; Weisenstein, D. K.; Miake-Lye, R. C.; Busen, R. A unified model for ultrafine aircraft particulate emissions. J. Geophys. Res. 2000, 105 (D24), 29379-29386.
(S17) Schröder, F.; Brock, C. A.; Baumann, R.; Petzold, A.; Busen, R.; Schulte, P.; Fiebig, M. In situ studies on volatile jet exhaust particle emissions: Impact of fuel sulfur concentration and environmental conditions on nucleation modes. J. Geophys. Res. 2000, 105 (D15), 19941-19954.
(S18) Sorokin, A.; Vancassel, X.; Mirabel, P. Emission of ions and charged soot particles by aircraft engines. Atmos. Chem. Phys. 2003, 3, 325-334.
(S19) Katragkou, E.; Wilhelm, S.; Arnold, F.; Wilson, C. First gaseous sulfur (VI) measurement in the simulated internal flow of an aircraft gas turbine engine during project PartEmis. Geophys. Res. Lett. 2004, 31 (2), L02117.
(S20) Lobo, P.; Hagen, D. H.; Whitefield, P. D.; Alofs, D. J. Physical characterization of aerosol emissions from a commercial gas turbine engine. J. Propul. Power. 2007, 23 (5), 919-929.
(S21) Brown, R. C.; Miake-Lye, R. C.; Lukachko, S. P.; Waitz, I. A. Heterogeneous reactions in aircraft gas turbine engines. Geophys. Res. Lett. 2002, 29 (10), 66-1 - 66-4.
(S22) Lukachko, S. P.; Waitz, I. A.; Miake-Lye, R. C.; Brown, R. C.; Anderson, M. R. Production of sulfate aerosol precursors in the turbine and exhaust nozzle of an aircraft engine. J. Geophys. Res. 1998, 103 (D13), 16159-16174.
(S23) Kärcher, B. Aviation-produced aerosols and contrails. Surv. Geophys. 1999, 20 (2), 113-167.
(S24) Binkowski, F. S.; Roselle, S. J. Models-3 Community Multiscale Air Quality (CMAQ) model aerosol component 1. Model description. J. Geophys. Res. 2003, 108 (D6) 4183.
(S25) Nenes, A.; Pandis, S. N.; Pilinis, C. ISORROPIA: A new thermodynamic equilibrium model for multiphase multicomponent inorganic aerosols. Aquat. Geochem. 1998, 4 (1), 123-152.
(S26) Fountoukis, C.; Nenes, A. ISORROPIAII: A computationally efficient thermodynamic equilibrium model for K+-Ca2+-Mg2+-NH4
+-Na+-SO42--NO3
--Cl--H2O aerosols. Atmos. Chem. Phys. 2007, 7 (17), 4639-4659.
(S27) Beusen, A. H. W.; Bouwman, A. F.; Heuberger, P. S. C.; Van Drecht, G.; Van Der Hoek, K. W. Bottom-up uncertainty estimates of global ammonia emissions from global agricultural production systems. Atmos. Environ. 2008, 42 (24), 6067-6077.
(S28) van Noije, T. P. C.; Eskes, H. J.; van Weele, M.; van Velthoven, P. F. J. Implications of the enhanced Brewer-Dobson circulation in European Centre for Medium-Range Weather Forecasts reanalysis ERA-40 for the stratosphere-troposphere exchange of ozone in global chemistry transport models. J. Geophys. Res. 2004, 109 (19), D19308.
(S29) Stohl, A.; Bonasoni, P.; Cristofanelli, P.; Collins, W.; Feichter, J.; Frank, A.; Forster, C.; Gerasopoulos, E.; Gäggeler, H.; James, P.; Kentarchos, T.; Kromp-Kolb, H.; Krüger, B.; Land, C.; Meloen, J.; Papayannis, A.; Priller, A.; Seibert, P.; Sprenger, M.; Roelofs, G.J.; Scheel, H.E.; Schnabel, C.; Siegmund, P.; Tobler, L.; Trickl, T.; Wernli, H.; Wirth, V.; Zanis, P.; Zerefos, C. Stratosphere-troposphere exchange: A review, and what we have learned from STACCATO. J. Geophys. Res. 2003, 108 (D12), 8516.
(S30) Chen, P. Isentropic cross-tropopause mass exchange in the extratropics. J. Geophys. Res. 1995, 100 (D8), 16661-16673.
(S31) Wu, S.; Mickley, L. J.; Jacob, D. J.; Logan, J. A.; Yantosca, R. M.; Rind, S. Why are there large differences between models in global budgets of tropospheric ozone? J. Geophys. Res. 2007, 112 D05302.
! !W"
(S32) US Environmental Protection Agency. The Benefits and Costs of the Clean Air Act, 1990 to 2010, EPA Report to Congress, November 1999, EPA-410-R-99-001. US Environmental Protection Agency, 1999.
(S33) Davidson, C. I.; Phalen, R. F.; Solomon, P. A. Airborne particulate matter and human health: A review. Aerosol Sci. Technol. 2005, 39 (8), 737-749.
(S34) Harrison, R. M.; Yin, J. Particulate matter in the atmosphere: Which particle properties are important for its effects on health? Sci. Total Environ. 2000, 249 (1-3), 85-101.
(S35) European Commissio. ExternE – Externalities of Energy – Methodology 2005 Update, EUR 21951. European Commission, 2005.
(S36) Frank, N. H. Retained nitrate, hydrated sulfates, and carbonaceous mass in federal reference method fine particulate matter for six eastern US cities. J. Air & Waste Manage. Assoc. 2006 (56), 500-511.
(S37) World Health Organization. Global Burden of Disease (GBD) 2000 Version 1, estimates by region: Mortality World Health Organization, 2001.
(S38) Ratliff, G. L. Preliminary Assessment of the Impact of Commercial Aircraft on Local Air Quality in the US, thesis, Massachusetts Institute of Technology, 2007.
(S39) Sequeira, C. J. An Assessment of the Health Implications of Aviation Emissions Regulations, thesis, Massachusetts Institute of Technology, 2008.
(S40) Hendricks, J.; Kärcher, B.; Döpelheuer, A.; Feichter, J.; Lohmann, U.; Baumgardner, D. Simulating the global atmospheric black carbon cycle: A revisit to the contribution of aircraft emissions. Atmos. Chem. Phys. 2004, 4 (11-12), 2521-2541.
(S41) Barrett, S. R. H.; Britter, R. E. Development of algorithms and approximations for rapid operational air quality modelling. Atmos. Environ. 2008, 42 (34), 8105-8111.
(S42) Barrett, S. R. H.; Britter, R. E. Algorithms and analytical solutions for rapidly approximating long-term dispersion from line and area sources. Atmos. Environ. 2009, 43 (20), 3249-3258.
(S43) Stein, A. F.; Isakov, V.; Godowitch, J.; Draxler, R. R. A hybrid modeling approach to resolve pollutant concentrations in an urban area. Atmos. Environ. 2007, 41 (40), 9410-9426.
(S44) Liao, H; Henze, D. K.; Seinfeld, J. H.; Wu, S.; Mickley, L. J. Biogenic secondary organic aerosol over the United States: Comparison of climatological simulations with observations. J. Geophys. Res. 2007, 112, D06201.
(S45) Smith, S. J.; van Aardenne, J. Klimont, Z.; Andres, R.; Volke, A.; Delgado Arias, S. Anthropogenic sulfur dioxide emissions: 1850-2005. Atmos. Chem. Phys. Discuss. 2010, 10, 16111-16151, In review for Atmos. Chem. Phys.
(S46) Fridell, E.; Steen, E.; Peterson, K. Primary particles in ship emissions. Atmos. Environ. 2008, 42 (6), 1160-1168.
(S47) Burtscher, H. Physical characterization of particulate emissions from diesel engines: a review. J. Aerosol Sci. 2005, 36 (7), 896-932.
(S48) Liao, C.-M.; Chio, C.-P. Assessment of atmospheric ultrafine carbon particle-induced human health risk based on surface area dosimetry. Atmos. Environ. 2008, 42 (37), 8575-8584.