Atmospheric Environment 45 (2011) 5415e5424

Contents lists available at ScienceDirect

Atmospheric Environment
journal homepage: www.elsevier.com/locate/atmosenv

Air quality and public health impacts of UK airports. Part I: Emissions

M.E.J. Stettler a, b, S. Eastham a, S.R.H. Barrett b, *
a b

Department of Engineering, University of Cambridge, Trumpington Street, Cambridge CB2 1PZ, United Kingdom Department of Aeronautics and Astronautics, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, MA 02139, USA

a r t i c l e i n f o
Article history: Received 3 March 2011 Received in revised form 4 July 2011 Accepted 5 July 2011 Keywords: Aviation Airport Emissions Air quality Particulate matter

a b s t r a c t
The potential adverse human health and climate impacts of emissions from UK airports have become a signicant political issue, yet the emissions, air quality impacts and health impacts attributable to UK airports remain largely unstudied. We produce an inventory of UK airport emissions e including aircraft landing and takeoff (LTO) operations and airside support equipment e with uncertainties quantied. The airports studied account for more than 95% of UK air passengers in 2005. We estimate that in 2005, UK airports emitted 10.2 Gg [23 to 29%] of NOx, 0.73 Gg [29 to 32%] of SO2, 11.7 Gg [42 to 77%] of CO, 1.8 Gg [59 to 155%] of HC, 2.4 Tg [13 to 12%] of CO2, and 0.31 Gg [36 to 45%] of PM2.5. This translates to 2.5 Tg [12 to 12%] CO2-eq using Global Warming Potentials for a 100-year time horizon. Uncertainty estimates were based on analysis of data from aircraft emissions measurement campaigns and analyses of aircraft operations. The First-Order Approximation (FOA3) e currently the standard approach used to estimate particulate matter emissions from aircraft e is compared to measurements and it is shown that there are discrepancies greater than an order of magnitude for 40% of cases for both organic carbon and black carbon emissions indices. Modied methods to approximate organic carbon emissions, arising from incomplete combustion and lubrication oil, and black carbon are proposed. These alterations lead to factor 8 and a 44% increase in the annual emissions estimates of black and organic carbon particulate matter, respectively, leading to a factor 3.4 increase in total PM2.5 emissions compared to the current FOA3 methodology. Our estimates of emissions are used in Part II to quantify the air quality and health impacts of UK airports, to assess mitigation options, and to estimate the impacts of a potential London airport expansion. 2011 Elsevier Ltd. All rights reserved.

1. Introduction 1.1. Context Aviation affects the environment via the emission of pollutants from aircraft and supporting airport infrastructure, impacting on human health and well-being, and on the climate (Lee et al., 2010). Between 1960 and 2005 worldwide scheduled passenger air travel grew from 109 billion to 3.7 trillion passenger-km travelled. This represents an average growth rate of over 8% per year (IPCC, 1999; ICAO, 2006), while over the next two decades global air travel is forecast to grow by 4.5e6% per year (Lee et al., 2009). In the UK, a signicant political issue has been the proposed expansion of London Heathrow Airport, and potentially other London airports. Heathrow expansion was the policy of the previous administration, although the London Assembly (2010) criticised the plans on air quality grounds and the current administration does
* Corresponding author. E-mail address: sbarrett@mit.edu (S.R.H. Barrett). 1352-2310/$ e see front matter 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.atmosenv.2011.07.012

not plan to increase capacity at Heathrow or Stansted airports (HM Government, 2010). However, the debate on air quality impacts of potential expansion has occurred without quantication of those impacts on human health on a regional scale. Emitted pollutants resulting from aviation include greenhouse gases (GHGs) and particulate matter that contribute to forcing of the climate (Lee et al., 2010) and gases and particulate matter that are harmful to human health (Barrett et al., 2010). Aircraft engine emissions include CO, CO2, H2O, SO2, NOx (NO NO2), a range of hydrocarbons (HC), and volatile (sulphate and organic carbon) and non-volatile (mostly soot) particulate matter (PM). Emitted PM has an aerodynamic diameter much less than 2.5 mm (PM2.5), with modal diameter less than 100 nm (Onasch et al., 2009; Petzold et al., 2005). Non-volatile PM exists at the engine exit plane while volatile PM nucleates as new particles or condenses on existing particles in the cooling exhaust plume (Wayson et al., 2009; Onasch et al., 2009; Petzold et al., 2005). PM2.5 is thought to have adverse health impacts at concentrations down to pre-industrial levels and there is epidemiological evidence to show that adverse effects are associated with both short and long term exposure (WHO, 2006).

5416

M.E.J. Stettler et al. / Atmospheric Environment 45 (2011) 5415e5424

European Union air quality regulations state that there is no lower threshold concentration at which PM2.5 is not harmful to human health (EU, 2008). Barrett et al. (2010) estimated that w10,000 premature mortalities per year can be attributable to current aircraft operations at cruise (w80%) and the landing and takeoff cycle (LTO) (w20%), which when compared to an estimated 0.8 million premature mortalities globally due to anthropogenic air pollution (Krzyzanowski and Cohen, 2008), represents less than 1% of this gure. Ratliff et al. (2009) estimated that there are w160 annual cases of premature mortality in the United States as a result of airport operations (aircraft LTO and auxiliary power units (APUs)) attributable to PM2.5 exposure. Previous UK airport emissions inventories, for example the Project for Sustainable Development of Heathrow (DfT, 2006), pre-date the current method used to estimate aircraft engine particulate matter emissions called FOA3.0 (Wayson et al., 2009). Underwood et al. (2001) assessed the risk of exceedance of NO2 and PM10 limit values in the near vicinity of regional UK airports (not including London airports), however the scope of this study did not include a quantication of the level of exceedance or the subsequent impact on public health. Airport emissions inventory methodologies (Kim et al., 2009; EMEP/EEA, 2009; ICAO, 2007; IPCC, 2006) rely on reference emission indices (EIs) e i.e. g of specied pollutant emitted per kg of fuel burnt e and assumptions on the operational landing and takeoff (LTO) cycle. However, empirical studies of operational aircraft jet engine emissions (Schfer et al., 2003; Schrmann et al., 2007; Herndon et al., 2008, 2009; Wood et al., 2008; Carslaw et al., 2008; Mazaheri et al., 2009; Timko et al., 2010a) have revealed variability and discrepancies between observed and certication EIs in the Engine Emissions Databank (CAA, 2009). For example, Carslaw et al. (2008) found that the NOx concentration in the engine exhaust plume at takeoff could vary by up to 41% for aircraft with the same engines, attributing the variation in emission indices to variation in engine thrust setting. Timko et al. (2010a) observed that EI(CO) showed variability of w25%. 1.2. Purpose of paper This paper details the rst step towards quantifying the environmental impacts of UK airport operations. The objectives of this paper are to: (i) develop an inventory of emissions from aircraft LTO operations APUs, and airside support vehicles; (ii) quantify the scientic and operational uncertainty in the emissions of different species; and (iii) determine the most important sources of uncertainty. This analysis is limited to these three emissions sources, where the LTO cycle is dened up to an elevation of 3000 ft AFE. Although it is now thought non-LTO emissions may impact air quality on a regional-to-global scale (Barrett et al., 2010), these are outside of the scope of this study. Yim et al. (forthcoming) e hereafter referred to as Part II e will apply the inventories and uncertainty estimates developed in this paper to quantify the impacts of UK airports on air quality and human health. Thus, the greatest weighting in this paper is given to pollutants that impact upon air quality, including primary PM. However, we also present the total climate impact of airport operations through the CO2-eq metric because: (i) this is an area of increasing concern amongst airport operators (ACI, 2010); and (ii) while uncertainties in cruise emissions have previously been assessed, the likely larger uncertainties in LTO emissions have not been studied (Barrett et al., 2010). Additionally, Part II will examine potential mitigation strategies and assess the impacts of a potential future expansion of London Heathrow. While airport expansion in London has been ruled out under the current UK administration, the previous administration had supported a third runway at Heathrow and due to capacity

constraints, the issue is likely to be raised again in future (Greater London Authority, 2011). The purpose of this additional analysis is that future debates or decisions on air quality and other impacts are based on a more complete and quantitative understanding of environmental impacts and possible mitigation options. Details omitted from the main text of this paper are presented in the Supporting Information (SI). 2. Methods This section describes the methods used to estimate emissions at airports arising from the aircraft LTO operations (2.3), use of APUs (2.4), and airside support vehicles (2.5). The treatment of uncertainty is discussed in the following subsection. While the methods are similar to those used in the National Atmospheric Emissions Inventory (NAEI) (Watterson et al., 2004), signicant distinctions are the revised PM2.5 methodology (and results), extended consideration of operational assumptions, and assessment of uncertainty in estimates including attribution of sources of uncertainty. 2.1. Uncertainty Uncertainties in model inputs are based on reviews of existing literature and new empirical analysis. In most cases, probability distributions have been assumed to be triangular with corresponding minimum, modal (nominal) and maximum values, unless otherwise specied. Using Monte Carlo 2000-member ensembles, the emissions model outputs (emissions totals for a number of species) take the form of probability distributions. From these, summary statistics are derived and reported as results in Section 3. Variance-based global sensitivity analysis methods are implemented to estimate the contribution of each input parameter to the total output variance. The sensitivity analysis approach is described in the SI. 2.2. Schedules The 20 airports in the study accounted for 96% of UK air passengers in 2005 and 87% of scheduled air trafc movements (ATMs). We develop a 2005 emission inventory that is based on schedule information from the Ofcial Airline Guide (OAG) (OAG aviation, 2005). The OAG covers 87% of all (scheduled and chartered) ights at the 20 airports as a whole when compared to statistics from the Civil Aviation Authority (CAA, 2005) (see SI); no correction for un-scheduled ights is made. 2.3. Aircraft emissions The LTO cycle is dened as all aircraft activity below a height of 3000 ft (z914 m) above eld elevation (AFE). For departures, the LTO cycle comprises taxiing out from the terminal to runway, hold on the taxiway, the takeoff roll, initial climb (up to 450 m) and climb-out to 3000 ft AFE. For arrivals this includes approach to runway from 3000 ft AFE, landing roll and taxi into the terminal (Watterson et al., 2004; ICAO, 2007). The 3000 ft AFE boundary for the LTO cycle is dictated by regulatory standards and is an approximation to a representative atmospheric mixing height (ICAO, 2007). This approximation is retained for the analysis given the regulatory framework, however observations suggest the mixing height can vary between 500 and 2000 m (Davies et al., 2007) with time of day and meteorological conditions, and the inuence of this on emissions estimates is explored in Section 3.2. Engine fuel ow factors and emission indices for jet and turboprop engines have been obtained from the ICAO Engine Emissions Databank (CAA, 2009) and the Emissions and Dispersion Modeling

M.E.J. Stettler et al. / Atmospheric Environment 45 (2011) 5415e5424 Table 1 Comparison between ICAO dened LTO cycle and that used within the emissions model developed in this study for the example of an A320 at LHR. Operation ICAO default LTO cycle Phase Arrival Approach Taxi in TIM (s) 240 420 F00 (%) 30 7 This study Phase Approach Landing roll Reverse thrust Taxi in Taxiway acceleration APU Taxi out Taxiway acceleration Hold Takeoff Initial climb Climb-out Mean TIM (s) 286 46 15 371 10 900 780 10e20 341 29 38 61 TIM uncertainty 10% for all

5417

F00 (%) 21e30 4e7 30 4e7 7e17% e 4e7 7e17% 4e7 75e100 75e100 75e85

Departure

Taxi out

1140

Takeoff Climb

42 132

100 85

System (EDMS) (FAA, 2010), respectively. Engine assignments from Kim et al. (2005) have been applied, where possible. In other cases, manufacturer specications were consulted. A complete list of aircraft-engine assignments is shown in the SI. Methods for each aircraft engine emission species are discussed in Sections 2.3.2e2.3.8 and summarised in Table 3. 2.3.1. Times-in-mode Emissions during a particular phase of the LTO cycle are proportional to the amount of time spent in that phase of operation e the time-in-mode (TIM). The standard ICAO certication LTO cycle is generally not representative of operations at airports (European Commission, 2001; Unique, 2004b; Patterson et al., 2009). As such, estimates of operation times for different phases of the LTO cycle are from the NAEI inventory (Watterson et al., 2004), shown in the SI. These TIMs vary by aircraft size category. Uncertainty in TIMs is estimated as 10% for airports for which empirical data existed (LHR, LGW and STN) and 20% for those where TIMs were estimated using the method described in Watterson et al. (2004). These are of similar order to variations in TIMs observed by Patterson et al. (2009) e deviations of 10e20% for takeoff and climb-out and 15e20% for approach. Thus, the modelled LTO cycle differs from the standard ICAO LTO as summarised in Table 1. 2.3.2. Thrust settings and fuel ow Fuel ow to the engine is approximately linearly proportional to engine thrust setting (Wey et al., 2006, p. 7), which is dened here as a percentage of maximum rated thrust (F00). It is likely that deviations in thrust setting are the main cause of the discrepancies between in situ measurements of emissions indices for a particular LTO phase and those tabulated in the ICAO Engine Emissions Databank corresponding to the default ICAO LTO cycle (Herndon et al., 2009, 2008; Mazaheri et al., 2009; Carslaw et al., 2008; Wood et al., 2008; Schrmann et al., 2007; Schfer et al., 2003). For instance, by comparing EI(NOx) observed during dedicated engine tests and in advected plumes from in-use aircraft during taxi operations, Wood et al. (2008) suggest that the thrust level used in the real taxi operations can be described by two discrete modes:

ground idle, with thrust settings resembling 4% instead of the 7% dened by the default ICAO LTO cycle; and taxiway acceleration with thrust settings up to 17%. British Airways operational reports (British Airways, 2006) conrm that increases in thrust, to approximately 10% and lasting approximately 10 s, do occur routinely when an aircraft is required to cross an active runway or make a sharp turn. Further evidence (King and Waitz, 2005; Underwood et al., 2004; Patterson et al., 2009) has been consulted to represent operational variability in thrust settings at other phases of the LTO model, which are summarised in Table 1. Commercial aircraft routinely make use of engine bleed ow to support air-conditioning and provide airframe power service, requiring a proportion of the engines power output (Baughcum et al., 1996; Herndon et al., 2009). Subsequently, fuel ow to the engine is increased for a given thrust output and correction factors suggested by Baughcum et al. (1996) have been applied. Engine ageing can also affect engine fuel ow and emissions by reducing the efciency of the engine, with estimates of increased fuel use ranging between 3 and 10% (Curran, 2006; Lukachko and Waitz, 1997). This, along with empirical evidence from Wey et al. (2006), suggests a suitable uncertainty range of 10% for the ICAO fuel ow indices. We note that correlations in aircraft ight performance variables (such as takeoff time and takeoff thrust) have not been accounted for here. These would be dependent on operational decisions by pilots as well as aircraft performance characteristics. 2.3.3. SOx emissions Sulphur emissions are proportional to fuel burn and depend on the fuel sulphur content (FSC), which varies by geographic region and uctuates over time (Hileman et al., 2010). In the UK, the mean FSC has uctuated between 360 ppm and 640 ppm over the past
Table 3 Summary of uncertainty estimates for aircraft engine emission species. Species Fuel burn Sulphur oxides NOx HC CO Black carbon Determining factor(s) Fuel ow rate FSC Parametric uncertainty 10% 500 300 ppm 2% nominal [0.5%, 5%] 30% (mode 10%) 90% 60% Uncertainty factors: 7% F00: [0.04, 37] 30% F00: [0.02, 2.89] 85, 100% F00: [0.27, 3.29] [1, 40] mg kg1 fuel (mode 20 mg kg1 fuel) [3148, 3173] g kg1 fuel

3
EI(NOx) EI(HC) EI(CO) EI(BC)

Table 2 Reference parameters for the alternative EI(BC) model. Parameter % F00 7 EI(NOx)ref SNMAx,ref EI(CO)ref 4 10 20 30 7 10 5 85 20 10 1 100 25 10 1

Organic carbon CO2

EI(OCIC) EI(CO2)complete

5418

M.E.J. Stettler et al. / Atmospheric Environment 45 (2011) 5415e5424

twenty years; in 2007, the mean was 483 ppm with a standard deviation of 600 ppm with skew towards lower values (Hileman et al., 2010; Rickard, 2008). The fuel sulphur content is estimated as 500 300 ppm for the current study. While the majority of elemental sulphur in jet fuel is emitted as SIV (SO2), some is emitted as SVI (SO3) at the engine exhaust plane. SO3 is rapidly converted to H2SO4, which is thermodynamically favoured to exist in liquid phase thereby resulting in volatile particulate matter (PM) emissions. Estimates for the SIV to SVI conversion efciency, 3, range between 0.08 0.01% (Timko et al., 2010b) and >10% (Schumann et al., 2002). However, the higher estimates are likely a result of measurement errors (Katragkou et al., 2004), or the (now known to be erroneous) interpretation of volatile organic material as SVI in earlier studies before organic PM emissions from aircraft were discovered. The lower bound only includes sulphate that coated soot particles and did not include nucleation or growth mode particles (Timko et al., 2010b). Another recent value of 0.13% (Onasch et al., 2009) included only condensed phase SVI. However, modelling the microphysical processes involved suggested this value is equivalent to a conversion efciency of 1e2% (Wong et al., 2008). Other estimates suggest a nominal value of 2% is suitable: 2 0.8% (Kiendler and Arnold, 2002); 2.3 1% (Sorokin et al., 2004); 2.3 1.2% (Katragkou et al., 2004). Curtius et al. (2002) estimate 3 to lie between 1.3 and 5.1% while Petzold et al. (2005) estimate the value to be within the range 2.5e6% and Kinsey et al. (2011) suggest a median value of 2.4%. Thus, a nominal value of 2% with a range 0.5e5% has been used to account for the range of estimates in the literature. 2.3.4. NOx emissions The NOx emissions indices for all aircraft engines are positively correlated to thrust setting, to which fuel ow is (approximately) directly proportional. The Boeing Fuel Flow Method 2 (BFFM2) (Baughcum et al. 1996) prescribes linear regression on a logelog plot of EI(NOx) against fuel ow to obtain EI(NOx) for intermediate thrust settings between certication data points. Comparing observed EIs from Timko et al. (2010a) for six engines with the corresponding EIs in the engine databank (shown in SI) suggests that EI(NOx) is w10% lower than the ICAO Engine Emissions Databank gure, as reported by Wood et al. (2008) and Timko et al. (2010a), and that the uncertainty bounds are approximately 30%. ICAO certication measures the total NOx emissions (NOx NO NO2). However the proportions of each component vary by engine and with thrust setting (Timko et al., 2010a). For most engines, NO is the predominant species at high thrust conditions (thrust settings of 65e100%) with NO2/NOx less than 10%. Analysis of APEX data for the CFM56-2C1 (Wey et al., 2006) engine suggests a suitable range to be 5e10%. At low power NO2 becomes dominant, with NO2/NOx ranging between 75 and 98% at 4% thrust (Timko et al., 2010a; Wood et al., 2008; Wormhoudt et al., 2007). Thus this is the assumed range for the taxi mode. For approach, analysis of the APEX data suggests that NO2/NOx z 12%, while charts in Timko et al. (2010a) and Wood et al. (2008) indicate that NO2/NOx lies within the range 12e20%. Another important emission is nitrous acid (HONO), a precursor to OH in the atmosphere and a component of total NOY emitted (NOy NO NO2 HONO HNO3 organic nitrates .). Wood et al. (2008) found that EI(NOx) and EI(NOY) agree within the experimental uncertainty (at most 6% discrepancy) indicating that HONO is likely included in the certication EI(NOx). Emissions of HONO are variable with respect to thrust setting and engine model. According to Wood et al. (2008), measured HONO/NOy ranges between 2e7% and 0.5e1% for low and high (65e100%) thrust settings respectively. Given that HONO/NOy is not known for the majority of engines, the 2e7% range has been applied to taxi and

approach, and the 0.5e1% range to climb-out and takeoff. The remaining NOx is assumed to be NO. 2.3.5. HC and CO emissions EI(HC) and EI(CO) decrease with increasing thrust. Therefore, the immediate effect of applying an idle thrust distribution below the ICAO specied 7% is to increase HC and CO emissions compared to previous approaches. BFFM2 (Baughcum et al., 1996) is employed to nd EI(HC) and EI(CO) at intermediate thrust settings. A bilinear t on a logelog plot of databank EI against fuel ow, consisting of a line between the two lower power setting points and a horizontal bisection of the two higher power setting points, has been implemented for each engine. Baughcum et al. (1996) recognised that some emissions data sets did not t the scheme described above; these cases have been overcome by implementing solutions described by Kim et al. (2005). Comparing observed EIs from Timko et al. (2010a) with certication EIs (shown in SI) indicates that the uncertainty in EI(CO) is 60%. EI(HC) (derived from EI(HCHO)) is more uncertain and is represented by a uniform distribution with bounds at 90%. HC emissions from aircraft engines are speciated into the constituent hydrocarbons according to U.S. EPA (2009). 2.3.6. Non-volatile PM emissions The primary form of non-volatile PM emitted by jet engines is soot (Timko et al., 2010b), which is primarily black carbon (BC) (Popovicheva et al., 2004; Petzold et al., 1999). Emissions indices can be estimated using the First-Order Approximation 3.0 methodology (FAO3) (Wayson et al., 2009). This calculation is dependent upon the mode-specic smoke number (SN) recorded in the engine databank. The SN is the current regulatory measure of emissions visibility and a dimensionless quantity related to the darkening of particle-loaded lters due to deposited particulate matter, rated on a scale from 0 to 100 (Sevcenco et al., 2009). Published measurement data from the Aircraft Particle Emission eXperiment (APEX1-3) (Wey et al., 2006; Timko et al., 2010b), the Delta-Atlanta Hartseld study (Lobo et al., 2007) and Agrawal et al. (2008) enables the performance of FOA3 e intended to be a rstorder approximation e to be assessed in relation to the accuracy of estimates of EI(BC) compared to observed values over a range of different engine types and engine operating modes. These studies have measured EI(BC) for a range of 12 separate engine models, which account for 30% of scheduled ATMs at the twenty airports in this study (see SI). Common thrust settings at which measurements are taken amongst these studies are 7%, 30% and 85% F00. Fig. 1 shows the comparison between estimated EIs obtained by implementing FOA3 using mode-specic SNs and the measured EIs. Note, while Wayson et al. (2009) recommend the use of mode-specic SNs, some other implementations of FOA3 may use the maximum SN where not all mode-specic SNs are reported. The method outlined in Calvert (2006) has been used to account for the one engine with incomplete SN data in the engine emissions databank. Fig. 1 indicates that a signicant number of data points lie on the horizontal axis (i.e. have very small/zero EI(BC)FOA3 but non-zero EI(BC)Measurement) and that there is an apparent systematic underestimation of EI(BC) by FOA3 at higher thrust e indeed 40% of FOA3 estimated are greater than an order of magnitude smaller than the mean measurement values. A possible reason for these discrepancies is that the measurement of SN could be susceptible to particle losses for diameters smaller than 20 nm (Sevcenco et al., 2009), such that the assumed correlation between SN and exhaust soot concentration index is inaccurate and not able to be applied across different engine technologies. More discussion of the FOA3 discrepancy is in the SI. Given the observed discrepancies, an alternative method of estimating EI(BC) is proposed using parameters available in the engine

M.E.J. Stettler et al. / Atmospheric Environment 45 (2011) 5415e5424

3 10 FOA3 7% FOA3 30% FOA3 85% New 7% New 30% New 85%

5419

Estimated EI(BC) (mg/kg fuel)

2 10

1 10

0 10

1 10 1 10 0 10 1 10 Measured EI(BC) (mg/kg fuel) 2 10 3 10

Fig. 1. Performance of FOA3 (hollow) and the proposed method (lled) at estimating EI(BC). Note that the zero estimated EIs for FOA3 are due to zero recorded SNs in the ICAO Engine Emissions Databank.

emissions databank (CAA, 2009). Dpelheuer and Lecht (1998) previously proposed a method of correcting engine exhaust soot concentration given a reference value, equivalence ratio, combustor inlet pressure and combustion ame temperature. Given that these parameters are proprietary, an alternative model is proposed using the EI(NOx), EI(CO), SNMAX and the engine Pressure Ratio (PR):

2.3.7. Volatile organic PM emissions Organic carbon (OC) PM emission indices, EI(OC), are typically estimated using FAO3 (Wayson et al., 2009; Ratliff et al., 2009), which is based upon an empirical relationship between the EI(HC) and EI(OC) as measured for the CFM56-2-C5 engine. However, emissions measurement data from Timko et al. (2010b) (7 engines), the Delta Atlanta Hartseld (4 engines) and APEX 1 (1 engine) measurement campaigns (supplied by Missouri S&T) suggest that this relationship does not hold for all engine models: Fig. 2 demonstrates the difference between measured EI(OC) and those calculated using FOA3. In 40% of cases, the discrepancy is greater than an order of magnitude. Timko et al. (2010b) identied that OC emissions comprise at least two primary exhaust gas components, products of incomplete combustion (OCIC) and lubrication oil (OCLO), and that EI(OCIC) lies with the range of 0.2e5 mg kg1 fuel for all engines at 7, 30 and 85% F00 once OCLO is accounted for, with no consistent trend with thrust setting. Lubrication oil emissions are highly dependent on engine technology and Timko et al. (2010b) estimate that the contribution of OCLO to the total mass of OC generally lies within the range 10e20% for low thrust and 50% for high thrust settings. Yu et al. (2010) suggest EI(OCLO) lies within the range of 2e12 mg kg1 fuel at idle and increases with engine thrust setting, suggesting a higher value at idle than that estimated by Timko et al. (2010b). In other studies, Kinsey et al. (2011) suggest that EI(OC) lies within the range 37e83 mg kg1 fuel and Agrawal et al. (2008) measure EI(OC) within the range 4e62 mg kg1 fuel, without discrimination between OCIC and OCLO. Thus, the EI(OCIC) is estimated to lie in the range 1e40 mg kg1 fuel with EI(OCLO) making up 10e20% and 40e60% of EI(OC) at low and high thrust settings, respectively. These discrepancies highlight the large uncertainty in OC emissions currently and the need for standardisation of measurement techniques, equipment and reporting

EINOx EIBCmg=kg fuel 4:57$ EINOx ref

!1:27

EICO $ EICOref

!0:4

SNMAX $ SNMAX;ref

!0:2   F 1:25 $ PR$ F00

where (F/F00) is the thrust setting as a fraction of full thrust, the multiplicative constant and exponents have been obtained by regression using the aforementioned measurement data and arbitrary reference values denoted by the subscript ref are detailed in Table 2. The model is implemented for each of the four standard thrust settings, EI(BC) for intermediate thrust settings are obtained by linear interpolation. Fig. 1 also shows the performance of this alternative model compared to FOA3. Uncertainties are estimated from the minimum and maximum difference between the estimated and measured EI(BC) and are shown in Table 3. The uncertainty range for EI(BC) at 85% is smaller than at lower thrust (a factor of 3), which is signicant given that EI(BC) and fuel ow are generally higher at higher thrust. This compares to maximum discrepancies of greater than an order of magnitude using FOA3. The R2 for FOA3 is 0.01 (meaning that the sum of residuals between the model and the observations is larger than the residuals between the mean of the observations and the observations) and for the proposed model the R2 is 0.58. This will be examined further in future work (Stettler et al., forthcoming). For the 12 engines used to derive this model, measured EIs were used in the emissions inventory developed in this paper with the proposed model lling-in for aircraft where measurements are not available.

Estimated EI(OC) (mg/kg fuel)

1 10

0 10

1 10 1 10 0 10 1 10

7% 30% 85%

Measured EI(OC) (mg/kg fuel)

Fig. 2. Performance of FOA3 at estimating organic PM emission indices at 7%, 30% and 85% F00.

5420

M.E.J. Stettler et al. / Atmospheric Environment 45 (2011) 5415e5424

before any large-scale measurement campaign of EI(OC) for regulatory purposes is embarked upon. 2.3.8. CO2 emissions Under complete combustion EI(CO2)complete depends on the ratio of carbon to hydrogen, for aviation fuel it lies within the interval [3148, 3173] (Hileman et al., 2010). At low thrust, incomplete combustion leads to a decrease in EI(CO2) relative to assuming complete combustion as EI(CO) and EI(HC) increase non-linearly (Wey et al., 2006). Thus, the following implementation of a carbon balance will be most signicant at low thrust settings:

are assumed to have made use of the APU, and airport stand characteristics have not been accounted for. The parametric uncertainties for APU emissions are summarised in Table 4. 2.5. Airside support equipment emissions Emissions arising from airside support equipment (also called ground support equipment (GSE)) have been derived using estimates of emissions factors (EFs) from Zrich Airport (ICAO, 2007; Unique, 2004a) as nominal values. Technology-specic default GSE emissions factors from EDMS combined with non-road emissions factors (EEA, 2009) provide an upper bound. These values are shown in the SI. Discrepancies between these two methods vary with emissions species, in the maximum case PM emissions per LTO cycle calculated using default assignments are greater by a factor of nine compared to those estimated from operations at Zrich. Airside operators have used ultra-low sulphur diesel at LHR since 2002 (Underwood et al., 2004) and this is assumed to be consistent across all airports in the study. The SIV to SVI conversion efciency for diesel engines is reported to be approximately 2% (Truex et al., 1980). Less than 1% of diesel engine measured NOy (as NO2) emissions are HONO, while around 90% are NO and the remaining 9% are NO2 (Kurtenbach et al., 2001). PM emissions from heavy-duty diesel engines are roughly half BC and half OC by mass (Miguel et al., 1998; Rogge et al., 1993). 3. Results and discussion 3.1. Emissions Emissions from UK airports in 2005 are detailed in Table 5. The median is presented with uncertainty bounds given by the 5th and 95th percentiles. In terms of primary PM, aircraft BC is the largest component of emitted PM2.5 (BC OC SO4), accounting for 47%. Aircraft LTOs are also the dominant source for sulphates (86%), however for OC the GSE contribution is over 66%. The GSE contribution to BC emissions is 20%, such that GSE are responsible for 28% of the total mass of emitted particulate matter. Low FSC of fuel used airside means that GSE emissions of SO2 and SO4 are negligible compared to those from aircraft and APUs. APUs contribute 6% to total PM2.5 emissions. The rationale for aggregating these distinct PM species into a total PM2.5 category is that there is no distinction in the concentrationeresponse functions used to derive health impacts in Part II (U.S. EPA, 2011). Similarly, PM from different airport sources is likely to have different size distributions and while there is evidence to suggest that the toxicity of PM increases as its size decreases, the epidemiological evidence base does not yet allow discrimination of PM size fractions below 2.5 mm (U.S. EPA, 2011). Emissions of CO2, SO2, NOx, CO and HC are also dominated by the aircraft LTO. The majority of NOy (84%) is emitted as NO, while 14% and 1.5% is emitted at NO2 and HONO respectively. Global Warming Potentials (GWPs) for a 100-year time horizon with uncertainty ranges from Fuglestvedt et al. (2010) are used to aggregate climate
Table 4 Summary of uncertainties applied to APU emissions. Species Fuel ow Sulphur derivatives (SO2, SIV) NOx HC CO PM10 CO2 Determining factor(s) Fuel ow rate [kg s1] FSC [ppm] 3 [%] EI(NOx) [g kg1 fuel] EI(HC) [g kg1 fuel] EI(CO) [g kg1 fuel] EI(PM10) [g kg1 fuel] EI(CO2) [g kg1 fuel] Parametric uncertainty 10% 500 300 ppm [0.5%, 5%] 30% 60% 90% Factor 10 [3148, 3173] g kg1 fuel

MWCO2 EICO2 EICO2 complete EICO$ MWCO MWCO2 $NHC EIHC$ MWHC MWCO2 $NC4 H7 EIOCIC $ MWC4 H7 MWCO2 EIBC$CBC $ MWC
where EI(X), MW(X), NX and CX are the emissions index, molecular weight, number of carbon atoms and carbon content by weight of species X. MW(HC) (82) and NHC (5) are weighted averages of the component species given by U.S. EPA (2009). For OC, Timko et al. (2010b) showed that a typical product of incomplete combustion of fuel is C4H7. Thus, MW(OC) is taken as 55, and NOC as 2. Aircraft soot, while mainly BC, is composed of two distinct fractions: a main fraction and a fraction of impurities; however the ratio of these fractions has not been quantied (Demirdjian et al., 2007). The carbon content of the main fraction is w98 wt%, and w40e60 wt% for the fraction of impurities (Popovicheva et al., 2004; Demirdjian et al., 2007). Given the unknown ratio, the total carbon content of BC is estimated to lie in the range 90e98% with a nominal value of 95%. Thus, the uncertainty of EI(CO2) is a function of the uncertainty in the carbon to hydrogen ratio and the uncertainties associated with all other carbon containing species. 2.4. APU emissions APUs are gas turbines used to generate electricity while the main engines are off and to provide bleed air to start the main engines (ICAO, 2007). Only one mode of operation has been assumed due to the available data. TIMs, fuel ows and emissions indices for NOx, HC, CO have been taken from Watterson et al. (2004) and are shown in the SI. Values for EI(NOx) and EI(CO) were within the same order of magnitude with values derived from measurement of APU exhaust plumes (Schfer et al., 2003). The NOx breakdown was assumed to be 90:9:1 for NO:NO2:HONO, consistent with measurements (Gerstle et al., 1999) and similar to an aircraft engine at high thrust setting. SOx, SVI and CO2 emissions were derived using the same methods as described above for aircraft engines. PM EIs measured by Gerstle et al. (1999) for two APU models suggest the total PM mass index lies in the range 0.4e0.8 g kg1 fuel. ICAO (2007) proposes a model to estimate EI(PM10) as a function of EI(NOx) applicable to a larger set of APU models. However, when this model is applied to two APU models tested by Gerstle et al. (1999), estimated EIs are a factor of 10 lower than those observed, indicating an upper uncertainty bound. APU PM is assumed to be similar in nature to aircraft engine PM, and thus PM2.5. Furthermore, it is estimated that once SVI emissions are subtracted, the PM is 95% soot by mass nominally (85e99%) given observations by Timko et al. (2010b) of other gas turbine engines at high thrust setting. The remaining mass is assigned as OC. All ights

M.E.J. Stettler et al. / Atmospheric Environment 45 (2011) 5415e5424 Table 5 UK airport emissions by source with (5the95th percentile) uncertainty ranges as percentage of the median. Species Units Source Aircraft LTO CO2 NOx NO NO2 HONO HC CO SO2 SO4 OC BC Total PM2.5 CO2-eq
a

5421

Total APU to to to to to to to to to to to to to 14] 36] 36] 50] 55] 173] 87] 33] 89] 72] 75] 58] 13] 3.23 7.16 6.57 5.30 5.36 7.31 9.50 10.00 3.62 9.34 1.38 1.86 3.27 108 105 105 104 103 104 105 104 103 102 104 104 108 [10 [22 [22 [28 [31 [64 [41 [29 [56 [85 [82 [64 [11 to to to to to to to to to to to to to 11] 24] 24] 37] 36] 62] 45] 31] 86] 240] 156] 123] 11] GSE 6.42 1.17 1.05 1.05 1.17 1.24 3.11 9.86 3.74 4.26 4.26 8.51 8.14 107 [63 to 113] 106 [59 to 105] 106 [59 to 105] 105 [59 to 105] 104 [59 to 105] 105 [61 to 108] 105 [46 to 91] 100 [44 to 57] 101 [54 to 102] 104 [63 to 117] 104 [63 to 117] 104 [63 to 117] 107 [56 to 89] 2.42 1.02 8.56 1.46 1.56 1.77 1.17 7.27 2.65 6.42 2.11 3.07 2.49 109 107 106 106 105 106 107 105 104 104 105 105 109 [13 [23 [24 [30 [35 [59 [42 [29 [56 [49 [41 [36 [12 to to to to to to to to to to to to to 12] 29] 30] 45] 49] 155] 77] 32] 89] 78] 54] 45] 12]

NAEI (2008)

kg kg kg kg kg kg kg kg kg kg kg kg kg

NO2 NO2 NO2 NO2 CH4

2.02 8.19 6.74 1.28 1.38 1.57 1.04 6.26 2.29 2.00 1.45 1.91 2.08

109 106 106 106 105 106 107 105 104 104 105 105 109

[14 [26 [27 [33 [38 [66 [47 [29 [56 [66 [49 [41 [13

1.27 107

1.44 106a 3.72 107 8.33 105

7.38 104

Derived from VOC using a factor of (1.15)1.

forcing species into an estimate of CO2-eq emissions (see SI). The largest contribution to total CO2-eq emissions is from CO2 emitted by aircraft, accounting for 81% and aircraft contribute to 84% of the CO2-eq emitted by UK airports. These estimates can be compared with estimates from the NAEI for the aviation sector in 2008. The NOx, HC and SO2 estimates agree within the uncertainty bounds. However, the NAEI CO and PM2.5 estimates are over a factor of 3 greater and 4 less than our estimates, respectively. The discrepancy in NOx estimates is likely due to reduced thrust at takeoff being applied to all airports instead of just LHR and LGW as in the NAEI methodology, while the largest contributor to our higher PM2.5 estimates is the new BC emissions methodology applied. CO emissions differ due to different thrust setting assumptions and the emissions index interpolation/extrapolation scheme used. Individual airports can be compared in terms of their emissions per service unit. Fig. 3 shows estimated total PM2.5 and CO2-eq and emissions per air trafc movement (ATM) and per passenger (PAX) for the 20 airports in the study. Total PM2.5 emissions per ATM are highest at LHR [157 g ATM1] however they are highest at LCY

[1.8 g PAX1] on a per passenger basis. CO2-eq emissions per ATM and per passenger (PAX) are highest at LHR [2500 kg ATM1; 17 kg PAX1] and LGW [1700 kg ATM1; 15 kg PAX1], the two busiest airports by ATMs and passengers in the UK. 3.2. LTO emissions Fig. 4 shows the relative proportion of emissions attributable to each phase of the LTO for various emissions species. It is clear that emissions of NOx and PM2.5 arise primarily from high thrust modes,
50 40 30 20 10 0 50 40 30 20

a
per ATM (g)

300 per ATM per PAX 200

3 per PAX (g)

Proportion of LTO emissions (%)

10 0 50 40 30 20 10 0 50 40 30 20

2.5

Total PM

100

Total PM

2.5

0
LH R LG W M AN ST N ED BH I X G LA LT N LC Y AB Z BR S N C SO L U LP L BF S BH D LB A EM A PI K C W L

b
CO eq per ATM (kg)

4000 per ATM per PAX

20 CO2eq per PAX (kg)

10 0 50 40 30 20 10 0
Sulfate OC BC

3000

15

2000

10

1000

Ta

Ac

LH R LG W M AN ST N ED BH I X G LA LT N LC Y AB Z BR S N C SO L U LP L BF S BH D LB A EM A PI K C W L

nd La

iti

Fig. 3. (a) Total PM2.5 and (b) CO2-eq per ATM and PAX for 20 UK airports in order of descending number of ATMs in 2005.

Fig. 4. Relative proportion of emissions attributable to each phase of the LTO for (a) CO2, (b) NOx, (c) CO, (d) HC and (e) PM2.5.

Ta

R ol ev l .T hr us t Ta Ta xi In xi Ac c. Ar r. R

ut

ep

ff

ut O b

lim

xi O

c. D

ke

ro Ap p

ac in

ol

xi

In

lim

Ta

al

5422

M.E.J. Stettler et al. / Atmospheric Environment 45 (2011) 5415e5424 Table 6 Total UK aircraft particulate matter emissions estimated using the proposed alternative methodology and FOA3. Species Aircraft OC (kg) Aircraft BC (kg) Total PM2.5 (kg) New method 2.00 104 [66 to 72] 1.45 105 [49 to 75] 1.91 105 [41 to 58] FOA3 1.39 104 [65 to 127] 1.83 104 [16 to 15] 5.67 104 [31 to 40]

conversely CO and HC emissions arise from low thrust modes reecting the non-linear prole of EIs with thrust settings described in Section 2.3. Mass of CO2 emitted is proportional to fuel ow which is approximately proportional to thrust setting and consequently the approach mode with greatest TIM is responsible for more CO2 emissions than any other LTO mode. These ndings also have implications for emissions mitigation potential at airports as some strategies may favour some species but not others. For example, introducing single-engine taxiing is not likely to signicantly reduce NOx or PM2.5 emissions, but would reduce CO, HC and CO2 emissions. The efcacy of this and other mitigation strategies is evaluated in Part II. As described above, the atmospheric mixing height is observed to vary between 500 and 2000 m (Davies et al., 2007). For these mixing heights [500 m, 2000 m] total emissions of CO2 and SOx are changed by [28%, 55%], NOx emissions by [36%, 72%] and total PM2.5 emissions are changed by [15%, 99%]. The changes to CO and HC emissions are less than 5%. This analysis suggests that the uncertainty in CO2, SOx, NOx and PM2.5 emissions resulting from mixing height variations are larger than the uncertainty resulting from the other parameters combined. 3.3. Uncertainty analysis Histograms depicting the probability density for UK airport emissions of CO2, CO2-eq, PM2.5 and HC are shown in the SI. 3.3.1. Model sensitivity The contribution from different model inputs to the total variance of the output emissions estimates are estimated using the Sobol method of global sensitivity analysis. Main-effect sensitivities, Si, capture the rst-order contribution of each input factor to the variance of the model output, such that the inputs can be ranked in order of importance (Saltelli et al., 2008; Allaire and Willcox,

EF(GSE PM) EI(BC) 85% EI(BC) 7% Thrust Takeoff EI(APU PM) A/C FF EI(OC) 0 0.1 0.2 0.3 Total PM 2.5 Si 0.4

a
0.5

2010). These Si have been estimated for all model outputs, with the most important factors for total PM2.5 and CO2-eq presented in Fig. 5. In the case of total PM2.5, the most important model inputs are the GSE EF(PM) (Si 0.39) and the uncertainty in aircraft EI(BC) at 85% F00 (Si 0.38) reecting the high degree of uncertainty in emissions arising from airside support vehicles and aircraft BC and their signicant contribution to total PM2.5 emissions. For CO2-eq, the most important factors are related to aircraft CO2 emissions (aircraft fuel ow indices, thrust settings and LTO TIMs) reecting the dominant contribution of aircraft to total CO2-eq emissions and in the case of the taxi thrust setting, the inuence of the carbon balance at low thrust due to EI(CO) and EI(HC). Indeed, aircraft HC emissions have the largest uncertainty and the range is positively skewed reecting the non-linear increase in EI at low thrust e the aircraft taxi thrust setting (Si 0.54) and EI(HC) (Si 0.36) are the most important uncertain parameters. Aircraft CO emissions are also positively skewed for the same reasons, with the similar inuence of aircraft taxi thrust setting (Si 0.53) and EI(CO) (Si 0.41). Another effect of the large uncertainties in CO and HC is that on the relative sizes of the LTO CO2 and CO2-eq uncertainty ranges: the CO2 uncertainty range is slightly larger (14%) compared to that of EI(CO2-eq) (13%) even though uncertainty is introduced through the GWPs. As EI(CO2) depends on other species through the carbon balance, it is high when EI(CO) and EI(HC) are low, and vice versa. However, CO2-eq is an aggregation of EI(CO2), EI(CO) and EI(HC) e thus the uncertainty is reduced. The evidence for taxi thrust setting <7% F00 and the importance of taxi thrust setting to CO and HC emissions estimates, where EIs are modelled using BFFM2, suggests that future measurement campaigns should explicitly measure and report emissions at w4% thrust. The most important model inputs with regards to aircraft NOx (largest source of airport NOx) are EI(NOx) (Si 0.49) and takeoff thrust setting (Si 0.27). For aircraft SO4, 3 is the most important model input (Si 0.75), however the importance of this factor to PM2.5 is reduced. All GSE emissions estimates are positively skewed as a result of the input EF uncertainties. Total-effect sensitivities for CO2-eq and PM2.5 are shown in the SI. 3.3.2. Sensitivity to PM models Estimates derived from FOA3 and the methodology outlined above for BC and OC particulate matter are shown in Table 6. For BC, the proposed alternative model leads to a factor 8 increase in estimated emissions over FOA3. For OC the methodology applied in this study leads to a 44% increase in the emissions estimate over FOA3. Sulphate emission estimates are unchanged, leading to an overall factor 3.4 increase in total PM2.5 emissions. The uncertainty in the FOA3 estimates results from uncertainty the other model parameters. 4. Conclusions

Thrust Taxi A/C FF Thrust Takeoff Thrust Approach EF(GSE CO2) TIM Taxi TIM Approach EI(BC) 85% 0 0.1 0.2 0.3 0.4 CO2eq Si
Fig. 5. Main-effect sensitivity indices (Si) for (a) total PM2.5 and (b) CO2-eq emissions.

b
0.5

A methodology to calculate gaseous and particulate matter emissions of air quality- and climate-concern from UK airports (including airside equipment and aircraft in the LTO cycle) has been developed and applied, with uncertainties explicitly estimated using a Monte Carlo approach. Sources of uncertainty considered span

M.E.J. Stettler et al. / Atmospheric Environment 45 (2011) 5415e5424

5423

operational (e.g. times-in-mode) and scientic (e.g. the SIV to SVI conversion efciency) factors, as well as structural modelling uncertainties (e.g. different PM emissions estimation methodologies). We estimate that in 2005, UK airports emitted 10.2 Gg [23 to 29%] of NOx, 0.73 Gg [29 to 32%] of SO2, 11.7 Gg [42 to 77%] of CO, 1.8 Gg [59 to 155%] of HC, 2.4 Tg [13 to 12%] of CO2, and 0.31 Gg [36 to 45%] of PM2.5, the latter estimated using revised methods. This translates to 2.5 Tg [12 to 12%] CO2-eq using Global Warming Potentials for a 100-year time horizon. CO2-eq emissions per ATM and per passenger are highest at LHR [2500 kg ATM1; 17 kg PAX1] and LGW [1700 kg ATM1; 15 kg PAX1]. Model sensitivity analysis showed that the model inputs with greatest contribution to the output variance of UK airport CO2-eq are related to aircraft CO2 emissions (fuel ow indices, thrust settings and LTO TIMs). For emissions of PM2.5 from UK airports, two most important input uncertainties are the PM emissions factor for airside vehicles and aircraft EI(BC) derived from the proposed alternative model at 85% F00. Such ndings may help prioritise future emissions measurement work. In Part II, emissions described here will be used in a regional chemistry-transport model, with nested dispersion modelling, to estimate the impact of UK airports on air quality and public health (in terms of premature mortalities). In addition, a scenario in which LHR is expanded will be assessed, along with possible mitigation strategies. Acknowledgements Funding was from UK EPSRC as part of the Energy Efcient Cities Initiative (www.eeci.cam.ac.uk) and the Airport Environmental Investment Toolkit. The authors are grateful to the Center of Excellence for Aerospace Particulate Emissions Reduction Research, Missouri University of Science and Technology (Missouri S&T), Rolla, Missouri for supplying aircraft emissions data. Appendix. Supporting information Supporting information associated with this article can be found, in the online version, at doi:10.1016/j.atmosenv.2011.07.012. References
Airports Council International (ACI), 2010. Airport Carbon Accreditation e Annual Report 2009-2010. Available at online: http://www.airportcarbonaccreditation. org/about.html. Agrawal, H., Sawant, A.A., Jansen, K., Miller, J.W., Cocker, D.R., 2008. Characterization of chemical and particulate emissions from aircraft engines. Atmospheric Environment 42 (18), 4380e4392. Allaire, D., Willcox, K., 2010. Surrogate modeling for uncertainty assessment with application to aviation environmental system models. AIAA Journal 48 (8), 1791e1803. Barrett, S.R.H., Britter, R.E., Waitz, I.A., 2010. Global mortality attributable to aircraft cruise emissions. Environmental Science and Technology 44 (19), 7736e7742. Baughcum, S.L., Tritz, T.G., Henderson, S.C., Pickett, D.C., 1996. Scheduled Civil Aircraft Emission Inventories for 1992: Database Development and Analysis. NASA. Appendix D. British Airways, 2006. British Airways Technical Documents Relating to Aircraft Operations Supporting the Project for the Sustainable Development of Heathrow. British Airways. CAA, 2005. UK Airport Statistics: 2005 e Annual. Available at online: http://www. caa.co.uk/default.aspx?catid80&pagetype88&sglid3&d2005Annual (accessed 27.07.10.). CAA, 2009. Aircraft Engine Emissions. Available at online: http://www.caa.co.uk/ default.aspx?catid702 (accessed 01.11.09.). Calvert, J.W., 2006. Revisions to Smoke Number Data in Emissions Databank. QinetiQ Ltd, Report to ICAO WG3 AEMTG: FOA Ad Hoc Group on Volatile PM. Carslaw, D.C., Ropkins, K., Laxen, D., Moorcroft, S., Marner, B., Williams, M.L., 2008. Near-eld commercial aircraft contribution to nitrogen oxides by engine, aircraft type and airline by individual sampling. Environmental Science and Technology 42 (6), 1871e1876. Curran, R., 2006. Correction to Engine Emission Data Resulting from Engine Deterioration. QinetiQ Ltd.

Curtius, J., Arnold, F., Schulte, P., 2002. Sulfuric acid measurements in the exhaust plume of a jet aircraft in ight: implications for the sulphuric acid formation efciency. Geophysical Research Letters 29 (7), 1113. Demirdjian, B., Ferry, D., Suzanne, J., Popovicheva, O.B., Persiantseva, N.M., Shonija, N.K., 2007. Heterogeneities in the microstructure and composition of aircraft engine combustor soot: impact on water uptake. Journal of Atmospheric Chemistry 56 (1), 83e103. Davies, F., Middleton, D.R., Bozier, K.E., 2007. Urban air pollution modelling and measurements of boundary layer height. Atmospheric Environment 41 (19), 4040e4049. Department for Transport (DfT), 2006. Project for the Sustainable Development of Heathrow. Report of the Air Quality Technical Panels. Available at online: http:// www.dft.gov.uk/pgr/aviation/environmentalissues/heathrowsustain/ (accessed 19.07.10.). Dpelheuer, A., Lecht, M., 1998. Inuence of Engine Performance on Emission Characteristics. Presented at the RTO AVT Symposium on Gas Turbine Engine Combustion, Emissions and Alternative Fuels. Lisbon, Portugal, 12e16 October 1998. European Environment Agency (EEA), 2009. EMEP/EEA Air Pollutant Emission Inventory Guidebook. EEA Technical Report No 9/2009. EEA. European Union (EU), 2008. Directive 2008/50/EC of the European Parliament and of the Council of 21 May 2008 on Ambient Air Quality and Cleaner Air for Europe. Ofcial Journal of the European Union. European Commission (EC), 2001. AEROCERT Final Report. EC. http://www. transport-research.info/web/projects/project_details.cfm?id19&pageresults. Federal Aviation Administration (FAA), 2010. Emissions and Dispersion Modeling System (EDMS). Available at online: http://www.faa.gov/about/ofce_org/ headquarters_ofces/aep/models/edms_model/ (accessed 20.07.10.). Fuglestvedt, J.S., Shine, K.P., Berntsen, T., Cook, J., Lee, D.S., Stenke, A., Skeie, R.B., Velders, G.J.M., Waitz, I.A., 2010. Transport impacts on atmosphere and climate: metrics. Atmospheric Environment 44 (37), 4648e4677. Gerstle, T., Virag, P., Wade, M., Kim, L., 1999. Aircraft Engine and Auxiliary Power Unit Emissions Testing. In: Particulate Matter Results, vol. 3. United States Air Force. Greater London Authority, 2011. A New Airport for London. Part 1 e The Case for New Capacity. Greater London Authority, London. Herndon, S.C., et al., 2008. Commercial aircraft engine emissions characterization of in-use aircraft at Hartseld-Jackson Atlanta International Airport. Environmental Science and Technology 42 (6), 1877e1883. Herndon, S.C., et al., 2009. Aircraft hydrocarbon emissions at Oakland International Airport. Environmental Science and Technology 43 (6), 1730e1736. Hileman, J.I., Stratton, R.W., Donohoo, P.E., 2010. Energy content and alternative jet fuel viability. Journal of Propulsion and Power 26 (6), 1184e1196. HM Government, 2010. The Coalition: Our Programme for Government. Cabinet Ofce, London. Internation Civil Aviation Organization (ICAO), 2006. World Air Trafc 1995e2006. www.icaodata.com. International Civil Aviation Organization (ICAO), 2007. Airport Air Quality Guidance Manual. International Civil Aviation Organization. Intergovernmental Panel on Climate Change (IPCC), 1999. Aviation and the Global Atmosphere. Cambridge University Press. Intergovernmental Panel on Climate Change (IPCC), 2006. 2006 IPCC Guidelines for National Greenhouse Gas Inventories. In: Energy, vol. 2. IPCC. Katragkou, E., Wilhelm, S., Arnold, F., Wilson, C., 2004. First gaseous sulfur (VI) measurements in the simulated internal ow of an aircraft gas turbine engine during project PartEmis. Geophysical Research Letters 31, L0217. Kiendler, A., Arnold, F., 2002. Unambiguous identication and measurement of sulfuric acid cluster chemiions in aircraft jet engine exhaust. Atmospheric Environment 36, 1757e1761. Kim, B., Fleming, G., Balasubramanian, S., Malwitz, A., Lee, J., Ruggiero, J., Waitz, I., Kilma, K., Stouffer, V., Long, D., Kostiuk, P., Locke, M., Holsclaw, C., Morales, A., McQueen, E., Gillette, W., 2005. System for Assessing Aviations Global Emissions (SAGE). Version 1.5 Technical Manual. Federal Aviation Administration, Washington DC. FAA-EE-2005-01. Kim, B., Waitz, I., Vigilante, M., Bassarab, R., 2009. ACRP Report 11: Guidebook on Preparing Airport Greenhouse Gas Emissions Inventories. Transportation Research Board, Washington DC. King, M., Waitz, I., 2005. Assessment of the Effects of Operational Procedures and Derated Thrust on American Airlines B777 Emissions from Londons Heathrow and Gatwick Airports. Partner, Cambridge, MA. Kinsey, J.S., Hays, M.D., Dong, Y., Williams, D.C., Logan, R., 2011. Chemical characterization of the ne particulate emissions from commercial aircraft engines during the Aircraft Particle Emissions eXperiment (APEX) 1 to 3. Environmental Science and Technology 45 (8), 3415e3421. Kurtenbach, R., Becker, K.H., Gomes, J.A.G., Kleffmann, J., Lrzer, J.C., Splitter, M., Wiesen, P., Ackermann, R., Geyer, A., Platt, U., 2001. Investigations of emissions and heterogeneous formation of HONO in a road trafc tunnel. Atmospheric Environment 35 (20), 3385e3394. Krzyzanowski, M., Cohen, A., 2008. Update of WHO air quality guidelines. Air Quality, Atmosphere, and Health (1), 7e13. Lee, D.S., Fahey, D.W., Forster, P.M., Newton, P.J., Wit, R.C.N., Lim, L.L., Owen, B., Sausen, R., 2009. Aviation and global climate change in the 21st century. Atmospheric Environment 43 (22e23), 3520e3537. 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., 2010. Transport impacts on atmosphere and climate: aviation. Atmospheric Environment 44 (37), 4678e4734.

5424

M.E.J. Stettler et al. / Atmospheric Environment 45 (2011) 5415e5424 Sorokin, A., Katragkou, W., Arnold, F., Busen, R., Schumann, U., 2004. Gaseous SO3 and H2SO4 in the exhaust of an aircraft gas turbine engine: measurements by CIMS and implications for fuel sulphur conversion to sulphur (VI) and conversion of SO3 to H2SO4. Atmospheric Environment 38, 449e456. Stettler, M.E.J., et al., forthcoming. Estimating Aircraft Particulate Matter Emissions: Assessment of FOA3.0 and a New Approach. Timko, M.T., Herndon, S.C., Wood, E.C., Onasch, T.B., Northway, M.J., Jayne, J.T., Canagaratna, M.R., Miake-Lye, R.C., Knighton, W.B., 2010a. Gas turbine engine emissions e Part I: volatile organic compounds and nitrogen oxides. Journal of Engineering for Gas Turbines and Power 132 (061504e1). Timko, M.T., Onasch, T.B., Northway, M.J., Jayne, J.T., Canagaratna, M.R., Herndon, S.C., Wood, E.C., Miake-Lye, R.C., Knighton, W.B., 2010b. Gas turbine emissions e Part II: chemical properties of particulate matter. Journal of Engineering for Gas Turbines and Power 132 (061505e1). Truex, T.J., Pierson, W.R., McKee, D.E., 1980. Sulfate in diesel exhaust. Environmental Science and Technology 14 (9), 1118e1121. Underwood, B.Y., Brightwell, S.M., Peirce, M.J., 2001. Air Quality at UK Regional Airports in 2005 and 2010. A Report Produced for DETR AEAT/ENV/R/0453. AEA Technology, Warrington. Underwood, B.Y., Walker, C.T., Peirce, M.J., 2004. Heathrow Emissions Inventory 2002: Part 1. Netcen, Warrington. Unique, 2004a. Aircraft Ground Handling Emissions at Zurich Airport. Unique Flughafen Zurich AG, Zurich. Unique, 2004b. Aircraft NOx Emissions within the Operational LTO Cycle. Unique Flughafen Zurich AG, Zurich. U.S. Environmental Protection Agency (U.S. EPA), 2009. Recommended Best Practice for Quantifying Speciated Organic Gas Emissions from Aircraft Equipped with Turbofan, Turbojet and Turboprop Engines. Version 1.0. Federal Aviation Administration Ofce of Environment and Energy and U.S. Environmental Protection Agency Ofce of Transportation and Air Quality. U.S. Environmental Protection Agency (U.S. EPA), 2011. The Benets and Costs of the Clean Air Act from 1990 to 2020. Final Report, March 2011. Ofce of Air and Radiation. Watterson, J., Walker, C., Eggleston, S., 2004. Revisions to the Method of Estimating Emissions from Aircraft in the UK Greenhouse Gas Inventory. Netcen, Oxford. Wayson, R.L., Fleming, G.G., Iovinelli, R., 2009. Methodology to estimate particulate matter emissions from certied commercial aircraft engines. Journal of the Air & Waste Management Association 59, 91e100. Wey, C.C., Anderson, B.E., Hudgins, C., Wey, C., Li-Jones, X., Winstead, E., Thornhill, L.K., Lobo, P., Hagen, D., Whiteeld, P., Yelvington, P.E., Herndon, S.C., Onasch, T.B., Miake-Lye, R.C., Wormhoudt, J., Knighton, W.B., Howard, R., Bryant, D., Corporan, E., Moses, C., Holve, D., Dodds, W., 2006. Aircraft Particle Emissions eXperiment (APEX). NASA. Wong, H.W., Yelvington, P.E., Timko, M.T., Onasch, T.B., Miake-Lye, R.C., Zhang, J., Waitz, I.A., 2008. Microphysical modeling of ground-level aircraft-emitted aerosol formation: roles of sulfur-containing species. Journal of Propulsion and Power 24 (3), 590e602. World Health Organization (WHO), 2006. WHO Air Quality Guidelines for Particulate Matter, Ozone, Nitrogen Dioxide and Sulfur Dioxide. Global Update 2005. Summary of Risk Assessment. World Health Organization. Wood, E.C., Herndon, S.C., Timko, M.T., Yelvington, P.E., Miake-Lye, R.C., 2008. Speciation and chemical evolution of nitrogen oxides in aircraft exhausts near airports. Environmental Science and Technology 42 (6), 1884e1891. Wormhoudt, J., Herndon, S.C., Yelvington, P.E., Miake-Lye, R.C., 2007. Nitrogen oxide (NO/NO2/HONO) emissions measurements in aircraft exhaust. Journal of Propulsion and Power 23 (5), 906e911. Yim, S.H.L., Stettler, M.E.J., Barrett, S.R.H. Air Quality and Public Health Impacts of UK Airports e Part II: Impacts, Assessment and Policy Analysis, forthcoming. Yu, Z., Liscinsky, D.S., Winstead, E.L., True, B.S., Timko, M.T., Bhargava, A., Herndon, S.C., Miake-Lye, R.C., Anderson, B.E., 2010. Characterization of lubrication oil emissions from aircraft engines. Environmental Science and Technology 44 (24), 9530e9534.

Lobo, D., Whiteeld, P.D., Hagen, D.E., Herndon, S.C., Jayne, J.T., Wood, E.C., Berk Knighton, W., Northway, M.J., Miake-Lye, R.C., Cocker, D., Sawant, A., Agrawal, H., Miller, J.W., 2007. The Development of Exhaust Speciation Proles for Commercial Jet Engines. Final Report. Prepared for the California air Resources Board and the California Environmental Protection Agency by the Center of Excellence for Aerospace Particulate Emissions Reduction Research, U. Missouri e Rolla. Aerodyne Research Inc, University of California e Riverside. London Assembly, 2010. Flights of Fancy: Can an Expanded Heathrow Meet its Environmental Targets? Greater London Authority, London. Lukachko, S.P., Waitz, I.A., 1997. Effects of Engine Aging on Aircraft NOx Emissions. In: International Gas Turbine & Aeroengine Congress & Exhibition. Orlando, Florida, 1997. The American Society of Mechanical Engineers. Mazaheri, M., Johnson, G.R., Morawska, L., 2009. Particle and gaseous emissions from commercial aircraft at each stage of the landing and takeoff cycle. Environmental Science and Technology 43 (2), 441e446. Miguel, A.H., Kirchstetter, T.W., Harley, R.A., 1998. On-road emissions of particulate polycyclic aromatic hydrocarbons and black carbon from gasoline and diesel vehicles. Environmental Science and Technology 32 (4), 450e455. National Atmospheric Emissions Inventory (NAEI), 2008. Data Request 29/11/10. OAG aviation, 2005. Available at online: http://www.oagaviation.com/Solutions/ AnalysisTools/Schedules/schedules.html (accessed 15.07.10.). Onasch, T.B., Jayne, J.T., Herndon, S., Worsnop, D.R., Miake-Lye, R.C., Mortimer, I.P., Anderson, B.E., 2009. Chemical properties of aircraft engine particulate exhaust emissions. Journal of Propulsion and Power 25 (5), 1121e1137. Patterson, J., Noel, G.J., Senzig, D.A., Roof, C.J., Fleming, G.G., 2009. Analysis of departure and arrival proles using real-time aircraft data. Journal of Aircraft 46 (4), 1094e1103. Petzold, A., Dpelheuer, A., Brock, C.A., Schder, F., 1999. In situ observations and model calculations of black carbon emission by aircraft at cruise altitude. Journal of Geophysical Research 104 (D18), 22171e22181. Petzold, A., Fiebig, M., Fritzsche, L., Stein, C., Schumann, U., Wilson, C.W., Hurley, C.D., Arnold, F., Katragkou, E., Baltensperger, U., Gysel, M., Nyeki, S., Hitzenberger, R., Giebl, H., Hughes, K.J., Kurtenbach, R., Wiesen, P., Madden, P., Puxbaum, H., Vrchoticky, S., Wahl, C., 2005. Particle emissions from aircraft engines e a survey of the European project PartEmis. Meteorologische Zeitschrift 14 (4), 465e476. Popovicheva, O.B., Persiantseva, N.M., Lukhovitskaya, E.E., Shonija, N.K., Zubareva, N.A., Demirdjian, B., Ferry, D., Suzanne, J., 2004. Aircraft engine soot as contrail nuclei. Geophysical Research Letters 31, L11104. Ratliff, G., Sequeira, C., Waitz, I., Ohsfeldt, M., Thrasher, T., Graham, M., Thompson, T., 2009. Aircraft Impacts on Local and Regional Air Qualty in the United States. PARTNER Project 15 nal report. PARTNER. Rickard, G., 2008. The Quality of Aviation Fuel Available in the United Kingdom Annual Survey 2007. QINETIQ/08/01656. QinetiQ Ltd. Rogge, W.F., Hildermann, L.M., Mazurek, M.A., Cass, G.R., Simonelt, B.R.T., 1993. Sources of ne organic aerosol. 2. Noncatalyst and catalyst-equipped automobiles and heavy-duty diesel trucks. Environmental Science and Technology 27 (4), 638e651. Saltelli, A., Ratto, M., Andres, T., Campolongo, F., Cariboni, J., Gatelli, D., Saisana, M., Tanrantola, S., 2008. Global Sensitivity Analysis. The Primer. John Wiley & Sons Ltd, Chichester, England. Schfer, K., Jahn, C., Sturm, P., Lechner, B., Bacher, M., 2003. Aircraft emission measurements by remote sensing methodologies at airports. Atmospheric Environment 37, 5261e5271. Schrmann, G., Schfer, K., Jahn, C., Hoffmann, H., Bauerfeind, M., Fleuti, E., Rappenglck, B., 2007. The impact of NOx, CO and VOC emissions on the air quality of Zurich airport. Atmospheric Environment 41, 103e118. Schumann, U., Arnold, F., Bunsen, R., Curtius, J., Krcher, B., Kiendler, A., Petzold, H., Schrder, F., Wohlfrom, K.-H., 2002. Inuence of fuel sulphur on the composition of aircraft exhaust plumes: the experiments SULFUR 1-7. Journal of Geophysical Research 107 (D15), 4247. Sevcenco, Y.A., Bowen, P.J., Johnson, M.P., Hilton, M., Welch, M., Miller, M.N., 2009. Mass and size distribution measurement of particulates from a gas turbine combustor using modern mobility analyzer and particle sizer. In: 45th AIAA/ ASME/SAE/ASEE Joint Propulsion Conference & Exhibit, 2e5 August 2009, Denver, Colorado. AIAA 2009e4827.