Content uploaded by Antony D. Evans
Author content
All content in this area was uploaded by Antony D. Evans on May 22, 2015
Content may be subject to copyright.
Mitigating Aviation Carbon Dioxide Emissions: An Analysis for Europe
Lynnette Dray1*, Antony Evans1, Tom Reynolds1,2, Andreas Schäfer1
1Aviation Integrated Modelling Group
Institute for Aviation and the Environment
University of Cambridge
1-5 Scroope Terrace
Cambridge, CB2 1PX
UNITED KINGDOM
Tel: +44 (0) 1223 760124
Fax: +44 (0) 1223 332960
2Massachusetts Institute of Technology
Department of Aeronautics & Astronautics/MIT Lincoln Laboratory
Room 33-115
77 Massachusetts Avenue
Cambridge MA 02139
USA
Tel: +1 617 253 7422
Email:
Lynnette Dray: lmd21@cam.ac.uk
Antony Evans: ade26@cam.ac.uk
Tom Reynolds: tgr@mit.edu
Andreas Schäfer: as601@cam.ac.uk
*Corresponding author
Word count: 7467 (5717 words, 4 tables, 3 figures)
Dray, Evans, Reynolds, Schäfer
1
Mitigating Aviation Carbon Dioxide Emissions: An Analysis for Europe
Lynnette Dray, Antony Evans, Tom Reynolds, Andreas Schäfer
Institute for Aviation and the Environment, University of Cambridge, Cambridge, UK
ABSTRACT
This paper investigates the interaction between economic, technological and operational
measures intended to reduce air transport-related CO2 emissions. In particular, the introduction
of aviation to the European Emissions Trading Scheme (ETS) in 2012 may prompt increased
uptake of presently-available emission reduction options (e.g. retrofitting winglets, expanding
maintenance programs) by airlines operating in Europe. In the future, carbon prices may also
determine the usage of new options currently under development (e.g. open rotor engines,
second-generation biofuels and improved air traffic management (ATM)). We apply the results
of a number of studies analyzing the airline costs and emission reductions possible from different
mitigation options to a systems model of European aviation. Using a set of nine scenarios (three
internally-consistent projections for future population, gross domestic product, oil and carbon
prices, each run with three policy cases), we analyze technology uptake and the resulting effect
on fuel lifecycle CO2 emissions with and without an ETS. We find that some options are rapidly
taken up under all scenarios (e.g. improved ATM), others are taken up more slowly by specific
aircraft classes depending on the scenario (e.g. biofuels) and others have negligible impact in the
cases studied. High uptake of one mitigation option may also reduce the uptake of other options.
Finally, it is observed that European aviation fuel lifecycle emissions could be reduced below
2005 levels before 2050 if cellulosic biomass fuels are made available from 2020. However, the
land use requirements in this scenario may limit its practicality at currently-projected cellulosic
biomass yields.
Dray, Evans, Reynolds, Schäfer
2
INTRODUCTION
Global aviation demand, in terms of revenue passenger-kilometers (RPK), is predicted to grow at
a rate of around 5% per year over at least the next 20 years (e.g. 1, 2), with European domestic
aviation RPK growing at a rate of 2-4%. Since technology improvements typically deliver a 1-
1.5% decrease in fuel burn per RPK per year (e.g. 3), this suggests European aviation emissions
are likely to continue to increase. However, emissions targets typically envisage a lowering of
aviation emissions. For example, the UK Government has announced its intention to reduce UK
aviation emissions to below year-2005 levels by 2050 (4
In Europe, aviation is to be included in the EU Emissions Trading Scheme (ETS) from
2012 (
). Therefore, a number of policy options
have been proposed or are in the planning stage to lower emissions by speeding up technology
introduction, introducing operational changes or reducing RPK growth levels.
5) meaning that overall emissions will be capped at a given level that reduces year-on-
year. Participants in an ETS who emit more than their “free” quota under the cap can either
purchase permits from other sectors, reduce their emissions until they get back within their
quota, or accomplish a combination of the two. It is expected that aviation will primarily follow
the first course (6), as it is currently relatively expensive to reduce emissions from aviation in
comparison to many other sectors. However, a recent study (7) suggested that there do exist cost-
effective direct mitigation options which airlines can apply at present-day oil prices. In this case,
the higher effective fuel prices resulting from emissions trading will prompt airline actions such
as retrofitting winglets on older aircraft. The interaction between emissions trading and airline
responses (and passenger responses if airline costs are passed on to ticket prices) is potentially
complex and depends on airline costs and demand levels. These in turn depend on the
underlying trends in European population, gross domestic product (GDP) and fuel prices over the
time period considered (e.g. 8
Further promising mitigation options (each also with their own associated benefits, costs
and difficulties) are likely to become available over the next 20 years. Geared turbofan engines
are currently at the testing stage, and potentially offer a 10-15% improvement in fuel economy
(
).
7). Open rotor engines are expected to offer an even more significant decrease in fuel burn
compared to conventional turbofans, but may be unsuitable for long-haul flights because of the
slower cruise speeds at which they operate (9) and may require modifications to airport
infrastructure to ensure ground personnel safety. The introduction of improved European air
traffic management from the Single European Sky ATM Research (SESAR) project (10
Additional potentially large savings in lifecycle carbon dioxide emissions may be
achieved by introducing aviation-suitable biofuels. A range of biomass-derived fuels are
currently under development, each with different lifecycle emission, cost and yield
characteristics. Present-day aviation-suitable biofuels have been produced from feedstocks such
as canola, soybean and palm-kernel oils (
) could
reduce the extra fuel burn aircraft currently incur by flying non-optimal routes due to ATM
inefficiencies.
11). Cellulosic biomass fuels which do not compete for
land use with food crops (using feedstocks such as switchgrass) are also under development. In
the longer term, microalgae-based fuels may offer a higher-yield solution (12
These mitigation options may also interact with each other. For example, adopting
biofuels may lower carbon costs significantly, reducing the incentive for an airline to adopt open
rotor engines at a given carbon price. It is for this reason that a fully integrated model capable of
capturing the combined effects of different policies and mitigation options is desirable. This
).
Dray, Evans, Reynolds, Schäfer
3
paper applies such a model to examine how different mitigation options combine, what actions
they prompt by airlines and how this might affect fares and passenger demand, and what the
resulting effect on total carbon dioxide emissions is for a range of different future scenarios.
METHODOLOGY
Aviation Systems Model
An aviation systems model, the Aviation Integrated Model (13,14,15
15
), was used to capture the
interdependencies in the European aviation system. This is a UK NERC and EPSRC-funded
program, written in Java and Matlab, which has been in active development since 2006. It has
been used in analyses of the European air transport system for Omega ( ) and the UK Climate
Change Committee (16 14), and to study the US and Indian air transport systems ( ). The Aviation
Integrated Model consists of seven interacting modules as shown in Figure 1, each covering a
different component of the air transport and environment system. This architecture permits
important feedback and data flows between the key system elements to be captured and provides
natural input sites for policy measures to be imposed upon the system as shown. Detailed
descriptions of the modules and their interactions are given in (13). In this study the Aircraft
Technology & Cost, Air Transport Demand, Airport Activity and Aircraft Movement modules
were utilized. These modules are run iteratively to find an equilibrium solution for aviation
system demand, emission and technology characteristics for the given year, scenario and policy
variables. The set-up for these modules is briefly summarized below.
Aircraft Technology & Cost Module
The Aircraft Technology & Cost Module simulates fuel burn, key emissions and operating costs
as a function of stage length and load factor for airframe and engine technologies within the
forecast time horizon. The global fleet was represented by a set of six sample aircraft types by
size and technology age, shown in Table 1. Performance and emissions modeling for these
aircraft below 3,000 feet was based on the ICAO engine exhaust emission data (17) and the
ICAO reference Landing and Take-Off cycle (18), adjusted for airport-specific taxi-out delay
times from the Airport Activity Module. Above 3,000 feet, performance during climb, cruise,
descent, and airborne holding was modeled using the Eurocontrol Base of Aircraft Data (BADA)
model (19), adjusted for route-specific airborne delay and inefficiency from the Aircraft
Movement Module. The costs associated with owning and operating these aircraft were taken
from published US airline cost data (20), adjusted for global differences in operating costs (21).
European navigation charges were obtained from (22
The improvement in fleet fuel burn resulting from the retirement of older aircraft and the
introduction of new aircraft types was modeled based on historical fleet turnover behavior (
).
23
23
).
Existing aircraft were assumed to suffer an increase in fuel burn per RPK with age due to
airframe/engine deterioration, with a rate of 0.2% per year ( ). New models of aircraft were
assumed to take advantage of incremental improvements in technology and hence have lower
starting fuel burn than current models. However, the option of retrofits or introducing radical
new technologies (with associated changes in airline costs) is treated separately as an airline
choice, to avoid double-counting technological improvements. The rate of technology
development for future aircraft models is likely to be driven by future changes in fuel and carbon
Dray, Evans, Reynolds, Schäfer
4
costs. For this study it is assumed that fuel burn for the best available new aircraft technology,
excluding radical new technologies such as blended wing bodies or open rotor engines, improves
by 1%, 1.5% or 2% per year respectively for scenarios where the 2030 oil price plus associated
carbon trading costs is below $100/barrel (bbl), between $100/bbl and $150/bbl or over $150/bbl
in year 2005 dollars. These improvement rates and price thresholds represent, respectively, low,
medium and high values with respect to historical trends in fuel burn (3) and projected oil and
carbon prices (24
).
Air Transport Demand Module
The demand (D) for true origin-ultimate destination passenger air trips between cities i and j was
estimated by the Air Transportation Demand Module, using a simple one-equation gravity-type
model given in Equation 1.
τ
ωϕεδ
γα
ij
DFSBA
jijiij CeeeePPIID ijijijij
)()(=
(1)
The explanatory variables include base year metropolitan area population (P), associated
income (I), and generalized travel costs (C) consisting of fares, value of travel time and flight
delay. The binary variables A and B indicate whether one or both cities in the pair have qualities
which might increase visitor numbers (for example being a major tourist destination or capital
city), the binary variable S indicates whether road links exist between a given city pair, and the
binary variable DF indicates whether the flight is domestic.
Base year metropolitan area population and income data were obtained from individual
country censuses and household income surveys (e.g. 25, 26), with income converted to year
2005 dollars using market exchange rates. Base year fares and journey times were estimated
using published data on airline delays, yields with flight distance and business model (27, 28),
and schedules (29). Base year segmented passenger demand was obtained from (30
8
). As true
origin-ultimate destination demand data was not available, we used an assignment matrix
approach to estimate elasticities for short-, medium- and long-haul trips (14). Routing was
estimated using scheduled journey and available connection times (29) based on an analysis of
US routing used by ( ). Parameter estimates are given in Table 2; all parameter estimates are
significant at the 95% level and compare well to literature values (e.g. 31). The R2 obtained is
0.47. The future demand for air trips was estimated using scenario-based forecasts of the key
explanatory variables, with delay and airline cost values from the Aircraft Technology & Cost
and Airport Activity Modules. In particular, future fare trends depend on the change in operating
costs (most notably the oil price) and market economics. For simplicity and transparency, airline
rates of return are assumed to remain constant in all markets, as modeled by (32
). This means
that future fares between true origin-ultimate destination city pairs scale relative to base year
fares in the same way as average costs of carrying passengers between the respective cities,
accounting for flights serving both non-stop and connecting itineraries.
Airport Activity Module
Dray, Evans, Reynolds, Schäfer
5
The Airport Activity Module forecasts the global air traffic required to satisfy the demand
projected by the Air Transport Demand Module and estimates the resulting flight delay given
airport capacity constraints.
The flight routing network was assumed to remain unchanged from the base year, with
the proportion of different aircraft types used on the required flight segments estimated as a
function of projected passenger demand, segment length and network type (hub-hub, hub-spoke,
or point-to-point) according to a multinomial logit regression on historical data. Flight
frequencies were forecast by applying base year passenger load factors by segment to passenger
demand estimated by the Air Transport Demand Module (33), given average aircraft sizes
calculated by the multinomial logit model.
Flight delays, both on the ground and in airborne holding, were estimated as a function of
flight frequencies and airport capacity constraints. Published European airport capacities were
used where available. Where airport capacities were not available, they were estimated using
simplified runway capacity models (34) and standard capacity estimation charts corresponding to
different airport configurations (35). Delays due to airport capacity constraints were estimated
using queuing theory, applying the cumulative diagram approach and classical steady state
simplifications (36
37
).These were added to gate departure delays (due to mechanical failures and
late arrivals), which were assumed to remain at current levels. While actual delay values were
calculated using modeled European flight frequencies and airport capacities, the calculated
departure delays due to origin airport capacity constraints were distributed between the taxiway
and the gate according to a taxi-out threshold estimated from historical US data (37). Similarly,
delays due to destination airport capacity constraints were distributed between the air and ground
according to a US data-based airborne holding threshold ( ), above which delay was assumed
to be propagated upstream to the departure gate.
Future projections of airport capacity tend to be short-term and focused on capacity
expansions which are already in the planning or construction stage. Rather than use external
projections of capacity, we simulate future airport capacity expansion within the Aviation
Integrated Model by assuming that capacity will be increased as required to serve forecast
demand such that delays remain close to present-day levels. The majority of airports in the
scenarios explored in this paper do not reach their current capacity limits by 2050. However, a
small number of major hub airports do. For these airports it is likely that capacity expansion
would in reality come from more intensive use of runways and increased use of secondary
airports, as well as possible infrastructure expansion.
Aircraft Movement Module
The air traffic by flight segment generated by the Airport Activity Module was input to the
Aircraft Movement Module. This identified the amount and location of emissions released in
flight, accounting for inefficiencies introduced by the air traffic control system (some of which
will be addressed through SESAR) and constraints imposed by safety procedures (such as
separation requirements which cannot be completely removed from the system). These
inefficiencies manifest as extra distance flown beyond the shortest ground track distance or
excess fuel burnt above the theoretical optimum for different routes and aircraft types. These
extra distances and excess fuel burn in different flight phases were quantified for Europe by
using archived flight track and flight data recorder information from the region, as described in
(38,39).
Dray, Evans, Reynolds, Schäfer
6
Abatement Options
This study is intended to model airline and passenger responses to increasing costs (such as those
imposed by an ETS). A wide range of possible options to lower fuel use and emissions are
available to airlines, now and in the future. These include maintenance, operational changes and
retrofits in the short term and radical new technologies in the longer term. However, many of
these measures are not economic to adopt for most aircraft unless carbon prices significantly
exceed currently-projected levels. Others, for example increased use of turboprops, have
associated issues which are difficult to quantify, such as cabin noise (40
Each option has an associated upfront cost, change in the operating costs of a given
aircraft and change in the fuel burn of that aircraft (all of which may be a function of the aircraft
age, size or typical route). In addition, some measures are not applicable to the whole fleet. For
example, it is assumed that winglet retrofits are not applicable to aircraft types which already
have winglets, or to future models of aircraft which are assumed to be already fitted with
winglets if these can provide a cost-effective fuel burn advantage. Characteristics of these
options in terms of cost, applicability and fuel burn reductions are taken from (7) and (41). The
assumptions used here are significant simplifications and in many cases current information
about future costs and emissions is extremely uncertain (e.g. open rotor engines). However, the
general behavior of the interaction between options is unlikely to change significantly with more
accurate information.
). The combined effects
on emissions of any given two measures are not necessarily additive and can depend on adoption
order (e.g. applying an engine upgrade kit and then re-engining). For this paper, a range of
abatement options was chosen from those evaluated by (7). It should be emphasized that the
options studied here and listed in Table 3 are only a selection of those which may become
available, and that a full assessment of every abatement option available to airlines before 2050
would be significantly more complex.
Airlines are assumed to adopt measures based on a payback period of seven years, i.e., an
abatement option will be introduced only if the cost savings over the next seven years are
expected to be greater than the upfront and yearly costs of applying the measure over that time
period. Once a measure is adopted, the costs and fuel burn of the applicable cohort of aircraft are
adjusted accordingly. This then affects the choice of any further measures.
In the case of biofuels, it is assumed that costs under emissions trading are based on fuel
lifecycle ("well-to-wake") emissions rather than simply airborne emissions. We assume drop-in
cellulosic biomass biofuel is made available from 2020 in a 50/50 blend with Jet A, and that the
introduction of biofuels is gradual, with yearly production increases limited to historically-
observed rates from the Brazilian proEthanol program (42
12 ). Aviation biofuel prices were
assumed to be at least 70 US cents per liter ( ) or — following the profit-maximizing behavior
of the fuels industry — equivalent to the costs of Jet A, whichever value is higher. Lifecycle
emission characteristics are also derived from (12).
City Set and Scenarios
The global Aviation Integrated Model concentrates on a set of 700 cities for which airport-level,
demographic and socioeconomic data have been gathered, containing 1127 airports and
accounting for about 95% of global scheduled RPK. For the intraregional Europe model
Dray, Evans, Reynolds, Schäfer
7
presented here we use the corresponding European subset, which contains 173 cities and 337
airports. A full list is given in (15).
Underlying the projection of future aviation growth in the Aviation Integrated Model are
scenario-based projections of key variables such as population, GDP per capita and oil prices.
These factors are interdependent, with (for example) high oil or carbon prices affecting GDP.
Therefore any scenarios used need to incorporate integrated economic modeling which considers
these factors simultaneously. In this study we use a set of external scenarios from the US Climate
Change Science Program (24). These were developed using MIT’s Integrated Global Systems
Model (IGSM), Stanford’s Model for Evaluating the Regional and Global Effects of GHG
Reduction Policies (MERGE) and the Joint Global Change Research Institute’s MiniCAM
model. Scenario data for Western and Eastern European growth is summarized in Table 3.
The IGSM, MERGE and MiniCAM models each include a range of carbon trading sub-
scenarios. A near-term carbon price of around €20 per tonne of CO2 has been suggested by a
number of studies (e.g. 43), whether or not aviation is included (44
5
). Therefore, in this study the
carbon trading scenario for each model was chosen which most closely reproduced these prices
over the period to 2030. Although airlines will initially receive some free allowances in the EU
ETS ( ), a move to full auctioning has been suggested (e.g. 6). It is assumed here that airlines
pay in full for their allowances and do not receive a free allocation.
RESULTS
In order to assess the interaction between different mitigation measures, we ran three basic
policy scenarios for each of the IGSM, MERGE and MiniCAM models:
Base: In this scenario, no carbon price is applied and no abatement measures are made available
for adoption by airlines. Individual aircraft fuel burn is affected only by fleet turnover and
incremental improvements in the technology of new aircraft.
Technology: In this scenario, no carbon price is applied but all technological abatement
measures are made available to airlines, who will adopt them if they provide an overall cost
saving over a seven-year payback period.
Abatement: This scenario is similar to the Technology scenario, but in addition a carbon price is
imposed.
In Figure 2, the RPK and fuel lifecycle CO2 emissions from these three scenarios are
shown. The top, centre and bottom panels depict the IGSM, MERGE and MiniCAM background
models respectively. In addition, alternative RPK forecasts from Boeing and Airbus (1,2) and
historical data from ICAO (45) are shown. The yearly RPK growth rates we project for European
aviation are lower, at around 2%, than those from the Airbus and Boeing forecasts, although not
outside the range of those predicted for the European system (e.g. 46). A number of reasons may
be behind this difference, including the elasticities and background scenarios used in this study
(e.g. Eastern European GDP per capita growth rates are consistently below those used by Boeing
and Airbus). The airline rate of return assumptions used result in base case fares remaining
broadly constant over the time period studied, so RPK growth rates here will also typically be
lower than for models which use a declining trend in travel cost.
Dray, Evans, Reynolds, Schäfer
8
To help interpret Figure 2, Figure 3 shows the mitigation option uptake by scenario, in
terms of the number of aircraft in the fleet which adopt each measure compared to the total fleet
size. The Base case is omitted as its technology uptake is zero by assumption. The left-hand
panels indicate the Technology case in which a carbon price is not applied and the right-hand
panels the Abatement case, which includes carbon trading.
In the Technology case, as airlines are able to make fuel cost savings by adopting
abatement measures, they can lower fares slightly. Therefore demand is slightly increased in the
Technology case (blue lines in Figure 2) over the Base case (red lines). However, this effect is
minimal. Only low cost, low impact measures which do not have a strong effect on total
emissions are adopted before 2020, as shown in the right-hand panels of Figure 2 and the left-
hand panels of Figure 3. Increased engine maintenance is adopted by some of the fleet in all
applicable scenarios, with uptake increased by emissions trading. Improved air traffic control
(SESAR) is assumed to be introduced in 2020. For the purposes of this paper, we assume
compliance is optional, with complying aircraft gaining improved fuel burn if they pay
adaptation costs and increased navigation charges. In reality it is likely that SESAR compliance
will become mandatory either initially or after some threshold year. However, the adaptations
needed to take advantage of SESAR are economic for all or most of the fleet in all scenarios,
suggesting rapid adoption is likely. After 2020, therefore, the Technology scenarios have
approximately 10% lower emissions than the corresponding Base scenarios, which do not
include SESAR. However, without emissions trading neither open rotors nor biofuels are
adopted in any scenario.
Figure 2 also shows the corresponding RPK and fuel lifecycle emissions in the
Abatement case (when a carbon price is applied, green lines). The underlying uptake of
mitigation options by scenario is shown in the right-hand panels of Figure 3. RPK travelled is
consistently lower in the Abatement case than in the Base case (in 2020, 1.3% lower for IGSM,
2.6% lower for MERGE and 2.1% lower for MiniCAM). This indicates that airlines are choosing
to pass some of the costs of emissions trading on to passengers. However, airlines also take
action to reduce their emissions trading costs by investing in technology. The combined
fuel+carbon price burden on airlines is greatest in the IGSM Abatement scenario (see Table 4).
This makes it economic to purchase open rotor aircraft from soon after their assumed initial
availability in 2020, and adoption of biofuels occurs at a rate limited only by the assumed
production rate increases, as shown in Figure 3(b).
Because the combined fuel+carbon price development in the MERGE and MiniCAM
models is lower than for IGSM, open rotors are not cost-effective. However, the uptake of other
measures is increased over the Technology (no carbon price) case, and biofuels are used across
the fleet from 2020. Additional runs in which the biofuel option is not made available indicate
that, in the absence of biofuels, open rotors would be adopted in the MERGE small aircraft class
from 2030. This kind of interdependency is observed elsewhere in the simulations. For example,
there are two cases in which SESAR compliance is less than 100%. The first is the MiniCAM
Technology scenario, in which airline costs are low enough that SESAR compliance is not
economic for some of the fleet. The second is the IGSM Abatement scenario. In this case the
savings airlines have made from early adoption of one technology (open rotors) lower the cost-
effectiveness of adopting another (SESAR compliance).
The right-hand panels of Figure 2 show that, in the Abatement case, fuel lifecycle
emissions differ little from the Base case before 2020. After this point, the introduction of
SESAR and biofuels, and (for IGSM) open rotors reduces fuel lifecycle CO2 emissions
Dray, Evans, Reynolds, Schäfer
9
significantly. By 2040 all three Abatement scenarios have emissions below year-2005 levels,
even though RPK has increased. Most of this decrease in emissions is due to the lower lifecycle
emissions of biofuels. All three Abatement scenarios use biofuel (in a 50/50 blend with Jet A)
across the entire European fleet in 2050. This suggests that the UK’s target of reducing 2050 UK
aviation emissions below 2005 levels is potentially achievable in an ETS+biofuels scenario.
However, in the highest-growth scenario (IGSM) the biofuel usage for satisfying intra-
European air travel demand alone in 2040-2050 is around 18 billion gallons. To produce this
much cellulosic biomass, a land area of about 14 million hectares (roughly the size of England),
would be required. It is likely that such an extensive use of biofuels is not realizable unless a
higher-yield biofuel is developed.
CONCLUSIONS
This paper has explored the interaction between airline uptake of current and future CO2
emission mitigation measures and emissions trading, by applying the results of studies on
marginal abatement costs to an aviation systems model of the European air transport system.
Although not all abatement options which may be available to airlines before 2050 are studied,
the analysis in this paper demonstrates the general interaction of different options and the
emissions reductions which may potentially be achievable even when using a reduced selection
of measures. Whilst some abatement options (in particular winglet retrofits and increased engine
maintenance) are economic to adopt in the absence of an ETS, it is found that, under the
assumptions made in this paper, the widespread use of open rotor engines and biofuels only
occurs at higher oil and carbon prices within an ETS. In practical terms, this means that in a
future scenario with no ETS and low oil prices we would expect most airlines to opt to order
aircraft with traditional engine types and to use Jet A fuel even when an open rotor aircraft is
available to order and an aviation-suitable biofuel is widely available. It is also found that, even
when adaptation to take advantage of improved air traffic control is optional, its uptake by
airlines is at or near 100% in all applicable scenarios modeled here.
The interaction between different mitigation measures is potentially complex and
depends on the cost-effectiveness, availability and introduction order of each measure. The most
promising scenario for fuel lifecycle CO2 emissions reduction is one in which an ETS is applied
and cellulosic biomass fuels are made available. In this case, the results suggest that it could be
possible to reduce fuel lifecycle CO2 emissions from European aviation in 2040 to below 2005
levels. However, for this to be a feasible scenario in terms of land use, a higher-yield biofuel
would need to be developed.
ACKNOWLEDGEMENTS
AIM is funded by the UK Engineering and Physical Sciences Research Council (EPSRC) and the
Natural Environment Research Council (NERC). This study was funded by resources from the
UK Omega project. These sources of support are gratefully acknowledged.
REFERENCES
Dray, Evans, Reynolds, Schäfer
10
1. Airbus, Global Market Forecast 2007-2026, 2007.
www.airbus.com/fileadmin/documents/gmf/PDF_dl/00-all-gmf_2007.pdf. Accessed July 7,
2009.
2. Boeing, Current Market Outlook 2007, 2007.
www.boeing.com/commercial/cmo/pdf/Boeing_Current_Market_Outlook_2007.pdf.
Accessed July 7, 2009.
3. Lee, J. J., S. P. Lukachko, I. A. Waitz and A. Schäfer. Historical and Future Trends in Aircraft
Performance, Cost and Emissions. Annual Reviews in Energy and the Environment, Vol. 26,
2001, pp.167-200.
4. UK Committee for Climate Change, CCC to Report on New Aviation Emissions Cap, January
2009. theccc.org.uk/pdfs/AviationEmissionsPR15.01.09.pdf. Accessed July 7, 2009.
5. European Union. Directive 2008/101/EC of the European Parliament. Official Journal of the
European Union, January 13, 2009, pp. 3-21.
6. European Union. Impact Assessment of the Inclusion of Aviation Activities in the Scheme for
Greenhouse Gas Allowance Trading Within the Community. Commission Staff Working
Document, 2006. ec.europa.eu/environment/climat/pdf/aviation/sec_2006_1684_en.pdf.
Accessed July 7, 2009.
7. Morris, J., A. Rowbotham, A. Angus, M. Mann and I. Poll. A Framework for Estimating the
Marginal Costs of Environmental Abatement for the Aviation Sector. Omega Final Report,
2009.
www.omega.mmu.ac.uk/Events/omega%2014%20final%20draft%20v2%20020309.pdf.Acce
ssed July 7, 2009.
8. Jamin, S., A. Schäfer, M. E. Ben-Akiva and I. A. Waitz. Aviation Emissions and Abatement
Policies in the United States: A City Pair Analysis. Transportation Research D, Vol. 9, No. 4,
2004, pp. 294-314.
9. Lawrence, J., M. Kingan and R. Self. Integrated Study of Advanced Open Rotor Powered
Aircraft. Omega Final Report, 2009.
www.omega.mmu.ac.uk/Events/Study%208%20Report%20Revised%20FORMATTED-
%20300309.pdf. Accessed November 5, 2009.
10. Eurocontrol. Single European Sky ATM Research (SESAR) website.
www.eurocontrol.int/sesar/. Accessed July 21, 2009.
11. Hileman, J. et al. Near-Term Feasibility of Alternative Jet Fuels. Final Report, Partner
Project 15, 2009. web.mit.edu/aeroastro/partner/reports/proj17/altfuelfeasrpt.pdf. Accessed
November 6, 2009.
12. Schäfer, A., J. Heywood, H. Jacoby and I. Waitz. Transportation in a Climate-Constrained
World. MIT Press, Cambridge, 2009.
13. Reynolds, T. G. et al. Modelling Environmental & Economic Impacts of Aviation:
Introducing the Aviation Integrated Modelling Tool. AIAA-2007-7751, 7th AIAA ATIO
Conference, Belfast, 2007.
14. Dray, L. M., A. Evans, T. G. Reynolds, H. Rogers, A. Schäfer and M. Vera-Morales. Air
Transport Within An Emissions Trading Regime: A Network-Based Analysis of the United
States and India. TRB 88th Annual Meeting, Washington, D.C., 2009.
15. Dray, L. M., A. Evans, T. G. Reynolds, A. Schäfer and M. Vera-Morales. Opportunities for
Reducing Aviation-related GHG Emissions: A Systems Analysis for Europe. Omega Final
Report, 2009. www.omega.mmu.ac.uk/Events/FINAL%20-
Dray, Evans, Reynolds, Schäfer
11
%20Opportunities%20for%20Reducing%20Aviation-
related%20GHG%20Emmissions%20230209.pdf. Accessed July 7, 2009.
16. Dray, L. M., A. Evans, T. Reynolds, and A. Schäfer. A Comparison of Aviation Greenhouse
Gas Emission Mitigation Policies for Europe. AIAA-2009-7112, 9th AIAA ATIO
Conference, Hilton Head, SC, 2009.
17. ICAO. ICAO Engine Emissions Databank.
www.caa.co.uk/docs/702/080407%20%20ICAO_Engine_Emissions_Databank-Issue_15-
C.xls. Accessed June 27, 2009.
18. ICAO. ICAO Annex 16, Volume II: Aircraft Engine Emissions. ICAO Publications, 2006.
19. EUROCONTROL. Base of Aircraft Data (BADA). Version 3.6, July 2004.
20. US DOT, Bureau of Transportation Statistics. DB1B Survey, Form 41, T100 Traffic &
Financial Data. www.transtats.bts.gov. Accessed July 8, 2009.
21. ICAO. Regional Differences in International Airline Operating Economics: 2002 and 2003.
Clr 310 AT/132, 2006.
22. Reynolds, T. G., L. Budd, D. Gillingwater and R. Caves. Effects of Airspace Charging on
Airline Route Selection & Greenhouse Gas Emissions. AIAA-2009-7028, 9th AIAA ATIO
Conference, Hilton Head, SC, 2009.
23. Morrell, P. and L. M. Dray. Environmental Aspects of Fleet Turnover, Retirement and Life
Cycle. Omega Final Report, 2009.
www.omega.mmu.ac.uk/Events/Cranfield%20Cambridge%20Omega%20T2-
06%20draft%20FINAL4%20030409.pdf. Accessed July 8, 2009.
24. CCSP. Scenarios of Greenhouse Gas Emissions and Atmospheric Concentrations, Synthesis
and Assessment Product2.1a, 2007. www.climatescience.gov. Accessed July 9, 2009.
25. ONS. UK Census 2001, 2001. www.statistics.gov.uk/census2001/access_results.asp.
Accessed July 8, 2009.
26. ISER. British Household Panel Survey 2005, 2005. www.data-
archive.ac.uk/findingData/snDescription.asp?sn=5151. Accessed July 9, 2009.
27. AEA. Operating Economy of AEA Airlines 2006, 2006.
files.aea.be/RIG/Economics/DL/SumRep06.pdf. Accessed July 9, 2009.
28. AEA. AEA Yearbook 1997, 1997. www.aea.be/aeawebsite/datafiles/year97.pdf. Accessed
July 9, 2009.
29. OAG. Official Airline Guide Schedule Database. Back Aviation Solutions, 2005.
30. Eurostat. European Transport Statistics, 2008 epp.eurostat.ec.europa.eu/portal/. Accessed
July 9, 2009.
31. Oum T. H., W. G. Waters and J. S. Yong. A Survey of Recent Estimates of Price Elasticities
of Demand for Transport. The World Bank, Washington, D.C., 1990.
32. Waitz, I. et al. Architecture Study for the Aviation Environmental Portfolio Management
Tool. Partnership for AiR Transportation Noise and Emissions Reduction, Report No.
PARTNER-COE-2006-002, 2006. www.partner.aero. Accessed February 23, 2010.
33. Bhadra, D., J. Gentry, B. Hogan and M. Wells. CAASD’s Future Air Traffic Estimator: A
Micro-Econometric Approach. 13th Annual Federal Forecasters Conference, BLS Training
Center, 2003.
34. De Neufville, R. and A. Odoni. Airport Systems: Planning, Design, and Management.
McGraw-Hill, 2003.
Dray, Evans, Reynolds, Schäfer
12
35. Reynolds, T. G., D. Gillingwater, R. Caves and L. Budd. Climate Related Air Traffic
Management, Omega Final Report, 2009.
www.omega.mmu.ac.uk/Events/Climate_Related_ATM.pdf. Accessed July 9, 2009.
36. Evans, A. D. Rapid Modelling of Airport Delay. 12th ATRS World Conference, Paper 202,
Athens, 2008.
37. Federal Aviation Administration, Airline Service Performance Metrics Database, 2008.
www.apo.data.faa.gov/aspm/entryASPM.asp. Accessed June 27, 2009.
38. Reynolds, T. G. Analysis of Lateral Flight Inefficiency in Global Air Traffic Management.
AIAA-2008-8865. 8th AIAA ATIO Conference, Anchorage, AK, 2008.
39. Reynolds, T. G. Development of Flight Inefficiency Metrics for Environmental Performance
Assessment of ATM, 8th USA/Europe Air Traffic Management Research and Development
Seminar (ATM2009), Napa, CA, 2009. www.atmseminar.org. Accessed July 25, 2009.
40. Aerospace Engineering. Noise and vibration control: Three technologies are combined to
provide significant reduction of aircraft cabin noise and vibration levels. Aerospace
Engineering, vol. 14, 1994, no. 9, pp. 10–12.
41. Henderson, J. Controlling Carbon Dioxide Emissions in the Aviation Sector. Stratus
Consulting for the US EPA, 2005.
42. Moreira J. R. and J. Goldemberg. The Alcohol Program. Energy Policy, Vol. 27, 1999, pp.
229-245.
43. Ernst & Young. Analysis of the EC Proposal to Include Aviation Activities in the Emissions
Trading Scheme. Executive Summary. Ernst & Young, New York, 2007
44. DEFRA and DfT. Including Aviation into the EU ETS: Impact on EU allowance prices. Final
Report to the UK Government, 2006.
www.defra.gov.uk/environment/climatechange/trading/eu/pdf/including-aviation-icf.pdf.
Accessed July 25, 2009.
45. ICAO. ICAO Global Database, 2007.
46. ICAO. Outlook for Air Transport to the Year 2015. Clr 304 AT/127, 2004.
Dray, Evans, Reynolds, Schäfer
13
TABLE TITLES AND FIGURE CAPTIONS
TABLE 1 Reference Aircraft Types
TABLE 2 Elasticity Estimates and Standard Errors (in parentheses) for European Air
Passenger Demand
TABLE 3 Characteristics of Mitigation Options Considered, from (7, 12, 41)
TABLE 4 Main Scenario Data Used in this Study, Following the US Climate Change
Science Program Study (24)
FIGURE 1 University of Cambridge Aviation Integrated Model.
FIGURE 2 RPK flown and fuel lifecycle CO2 emitted in the Base (no abatement measures
adopted), Technology and Abatement scenarios. Panels (a) and (b) depict RPK and CO2 for
IGSM, (c) and (d) for MERGE and (e) and (f) for the MiniCAM background model.
FIGURE 3 Number of aircraft in the fleet adopting different emission mitigation measures
by time and background scenario in comparison to the total fleet. Panels (b), (d) and (f)
show the Technology scenario with no emissions trading; panels (a), (c) and (e) show the
Abatement scenario including emissions trading.
Dray, Evans, Reynolds, Schäfer
14
Local
Environment
Impacts
Local/National
Economic
Impacts
Global
Environment
Impacts
Aircraft
Technology & Cost
Aircraft
Movement
Airport
Activity
Air Transport
Demand
Global
Climate
Air Quality
& Noise
Regional
Economics
Airport Capacity
SESAR
Winglets, Open Rotors,
Engine Upgrades,
Aero/Engine Maintenance
EU ETS
Airport Capacity
SESAR
Winglets, Open Rotors,
Engine Upgrades,
Aero/Engine Maintenance
EU ETS
FIGURE 1 University of Cambridge Aviation Integrated Model.
Dray, Evans, Reynolds, Schäfer
15
TABLE 1 Reference Aircraft Types
Size Class
Age Classa
Airframe
Engine
<190 Seats
pre-1995
Boeing 737-300
CFM56-3-B1
post-1995
Airbus A319-131
V2511
190-299 Seats
pre-1995
Boeing 767-300ER
PW4060
post-1995
Airbus A330-300
CF6 80E1 A2
>299 Seats
pre-1995
Boeing 747-400
PW4056
post-1995
Boeing 777-300
Trent 895
a The 1995 threshold is chosen to be 10 years before the 2005 model base year, based on date of
first entry into the fleet.
Dray, Evans, Reynolds, Schäfer
16
TABLE 2 Elasticity Estimates and Standard Errors (in parentheses) for European Air
Passenger Demand
2α
2γ
δ
ε
φ
ω
τ
Short haul
(<500 statute
miles)
1.16
(0.04) 0.75
(0.05) 0.77
(0.10) -0.90
(0.07) 0.32
(0.07) 1.63
(0.06) -1.24
(0.09)
Medium haul
(500-1000 statute
miles)
1.09
(0.04) 0.85
(0.05) 0.70
(0.12) -0.88
(0.07) 0.24
(0.07) 2.19
(0.13) -1.27
(0.08)
Long Haul
(> 1000 statute
miles)
1.01
(0.03) 0.75
(0.03) 1.46
(0.19) -0.36
(0.07) 0.66
(0.07) 1.59
(0.14) -1.08
(0.05)
Dray, Evans, Reynolds, Schäfer
17
TABLE 3 Characteristics of Mitigation Options Considered, from (7, 12, 41)
Mitigation
Technology
Availability
(year,
proportion of
fleet)
Fuel Burn
Reduction
Potential
(% per aircraft)
b
Upfront
Costs
(2005$)
Yearly
Costs
(2005$)
Comment
Winglets
Base year, up to
25% depending
on aircraft size
1.2 - 2.4%
depending on route
flown
$740,000
$14,800 extra
maintenance costs
More frequent
engine
maintenance
Base year, all
Up to 2.5%
$0
85% increase in
engine maintenance
costs
Depends on
aircraft age
More frequent
airframe
maintenance
Base year, all
Up to 1%
$0
Function of MTOW
and fuel saving
achievable
Depends on
aircraft age
Engine
upgrades
Base year, up to
37.5%
depending on
size
1%
15% of
new engine
costs
5% reduction in
engine maintenance
costs
Open rotor
engines
2020, all new
<190-seat
aircraft
30% (relative to
conventional
aircraft with the
same year of
manufacture)
$7,400,000
extra on
purchase
price
Engine
maintenance cost
increase of
$740,000
Journey time
increase assumed
small
Improved air
traffic
management
2020, all
10.5%
$463,000
for
avionics
upgrade
$83,300 for
equipment and
training, 30%
increase in
navigation costs
Assumed
reduction
potential is about
half of the total
fuel-based
inefficiencies
observed in (39).
Cellulosic
biomass fuels
2020, all
(limited
availability
before 2040)
85% (lifecycle CO
2
emissions from
100% biofuel)
$0
Biofuel costs
Mit. pot. relative
to petroleum-
derived jet fuel on
lifecycle basis.
Reversible
b Where not otherwise noted, the fuel burn reduction quoted is relative to aircraft of the same age, type and route
network which do not adopt the measure.
Dray, Evans, Reynolds, Schäfer
18
TABLE 4 Main Scenario Data Used in this Study, Following the US Climate Change
Science Program Study (24)
2000
2020
2040
Population, millions
Western Europec
IGSM
390
388
368
MERGE
390
397
397
MiniCAM
457
486
481
Eastern Europe
IGSM
97
91
83
MERGE
411
393
393
MiniCAM
124
119
111
GDP per capita, $(2005)
Western Europe
IGSM
19437
33554
60457
MERGE
22163
31992
44211
MiniCAM
16598
15607
24387
Eastern Europe
IGSM
2548
5433
11913
MERGE
2145
4264
8079
MiniCAM
2845
5188
11124
World Oil Price, $/bbl
IGSM
33.1
88.8
125.5
MERGE
33.1
71.7
98.0
MiniCAM
33.1
62.3
77.8
Carbon Price, $/tCO2
IGSM
0
23.0
46.0
MERGE
0
33.7
112.5
MiniCAM
0
28.5
94.3
c Country lists for Western and Eastern Europe are given in (24) and references therein.
Note that the different scenarios use different country sets for Western and Eastern Europe.
Dray, Evans, Reynolds, Schäfer
19
FIGURE 2 RPK flown and fuel lifecycle CO2 emitted in the Base (no abatement measures
adopted), Technology and Abatement scenarios. Panels (a) and (b) depict RPK and CO2 for
IGSM, (c) and (d) for MERGE and (e) and (f) for the MiniCAM background model.
Dray, Evans, Reynolds, Schäfer
20
FIGURE 3 Number of aircraft in the fleet adopting different emission mitigation measures
by time and background scenario in comparison to the total fleet. Panels (b), (d) and (f)
show the Technology scenario with no emissions trading; panels (a), (c) and (e) show the
Abatement scenario including emissions trading.