Forecast of air traffic's CO2 and NOx emissions until ...

Int. J. Aviation Management, Vol. X, No. Y, xxxx

Forecast of air traffic’s CO2 and NOx emissions until 2030 Martin Schaefer German Aerospace Center (DLR), Institute of Propulsion Technology, Linder Höhe, 51147 Cologne, Germany Fax: +49-2203-64395 E-mail: [email protected] Abstract: This article presents both methodology and results of an emissions forecast for air traffic until the year 2030. Aircraft emissions include carbon dioxide (CO2) and nitric oxides (NOx), which influence atmospheric chemistry and may contribute to global warming. The forecast is based on a simulation model, which predicts future air traffic and its emissions using flight schedules for a base year in combination with traffic growth assumptions and a fleet rollover simulation. Aircraft performance software is applied to estimate fuel consumption, CO2 and NOx emissions. Models from the Base of Aircraft Data (BADA) are used to simulate today’s fleet of aircraft, supplemented by additional models representing aircraft of the near future. Forecast results include the number of flights, passenger-kilometres, fuel consumption, CO2 and NOx emissions for air traffic from 2010 until 2030. Keywords: air traffic; aviation; emissions; forecast; fuel consumption; carbon dioxide; CO2; nitric oxides; NOx. Reference to this paper should be made as follows: Schaefer, M. (xxxx) ‘Forecast of air traffic’s CO2 and NOx emissions until 2030’, Int. J. Aviation Management, Vol. X, No. Y, pp.xxx–xxx. Biographical notes: Martin Schaefer graduated from Technische Universität Berlin in 2006 as an Aeronautical Engineer with specialisation in air transport. From 2006 to 2013, he worked as a Research Engineer at the DLR Institute of Propulsion Technology with a focus on aircraft engine emissions and aviation scenarios. Since 2012, he holds a Doctoral in Engineering from Ruhr-Universität Bochum. In 2013, he joined the German Federal Ministry of Transport and Digital Infrastructure as a Technical Advisor on Aviation’s Environmental Impacts. This paper is a revised and expanded version of a paper entitled ‘Forecast of air traffic’s CO2 and NOx emissions until 2030’ presented at the 2013 Air Transport Research Society (ATRS) World Conference, Bergamo, Italy, 26–29 June 2013.

Copyright © 20XX Inderscience Enterprises Ltd.

1

2

1

M. Schaefer

Introduction

A number of studies predict air traffic’s future emissions, many of which use emissions calculated for a base year in combination with assumed traffic growth rates and assumptions about an annual improvement of fuel efficiency (e.g., Eyers et al., 2004; Owen and Lee, 2006; Owen et al., 2010). The model described here uses a more detailed approach: It enables researchers to simulate traffic growth, fleet composition and technology development simultaneously and in a consistent way using an automated chain of software modules. Unlike the approach by Apffelstaedt et al. (2009) and Nolte et al. (2011), which also estimate CO2 emissions from a fleet forecast, the bottom-up simulation of worldwide air traffic allows for the calculation of (flight-phase dependent) NOx emissions. In addition, for the first time in such studies, detailed engine models are applied to provide an estimate of the NOx output from the latest generation of engines with lean burn combustion.

2

Background

2.1 Air traffic growth and targets for aviation emissions According to the International Civil Aviation Organization (ICAO) (2010a), passenger-kilometres have increased by 5% per year on average in the past decades, which implies a doubling of air traffic every 15 years. Forecasts by Airbus and Boeing predict a similar traffic increase for the next 20 years, with average annual growth rates of 4.7–5.0% for passenger-kilometres and similar growth for freight (Airbus, 2014; Boeing, 2014). The growth of air traffic gives rise to concerns regarding the climate impact of aviation. Aircraft emissions of NOx indirectly affect climate and have a net warming effect resulting from their influence on atmospheric ozone and methane concentrations (Lee et al., 2009; IPCC, 1999). Greenhouse gas emissions of CO2, which are proportional to fuel consumption, directly contribute to global warming. Contrails and emissions of black carbon particles may influence cloud properties and are also believed to affect climate (Lee et al., 2009). In addition, NOx and black carbon emissions as well as unburned hydrocarbons and carbon monoxide emissions from aircraft negatively influence local air quality around airports. In order to reduce aviation’s environmental footprint, goals have been set by policy and by the aircraft and airline industries. The Advisory Council for Aeronautics Research in Europe (ACARE) aims at a 50% reduction of fuel consumption and an 80% reduction of NOx emissions for new aircraft between 2000 and 2020 (Argüelles et al., 2001). The ACARE targets can be regarded as ambitious, as they require major improvements in aircraft technology and air traffic management (ATM). More recently, the European Commission published their ‘Flight Path 2050’, aiming at a 75% reduction of CO2 and a 90% reduction of NOx emissions per passenger-kilometre in 2050 compared to the year 2000 (European Commission, 2011). The International Air Transport Association (IATA) aims at a 1.5% annual improvement of fuel-efficiency until 2020 and carbon-neutral growth after 2020. IATA’s (2009a) target for 2050 is a reduction of absolute greenhouse gas emissions to half the levels of 2005.

Forecast of air traffic’s CO2 and NOx emissions until 2030

3

2.2 Aircraft and engine technology to reduce emissions Past and ongoing improvements of aircraft fuel efficiency are impressive, but cannot fully compensate for a 5% annual growth of transport performance. Since the early days of jet airliners, fuel efficiency per available seat-kilometre (ASK) has improved by 50–70%, depending on the reference (Peeters et al., 2005). More than half of the improvement is attributable to engine technology (IPCC, 1999), while improved aerodynamics and weight reduction are also major contributors. As aircraft technology has matured, further improvement is more difficult to achieve. Current developments in aerodynamics include evolutionary improvements like advanced wingtips and wing design. Higher improvements are expected from the introduction of laminar flow control in order to reduce friction and form drag components. Laminar flow control is estimated to deliver fuel efficiency benefits of 5–15% and may be used for new designs with an entry into service in 2020 or later (IATA, 2009b). For the more distant future, alternative configurations like blended wing bodies are discussed. It remains to be seen whether their expected efficiency benefit in the order of 10–25% (IATA, 2009b) outweighs the disadvantages and the additional investment required for manufacturers when developing such aircraft. Reducing thrust-specific fuel consumption (TSFC) is a major goal for future aircraft engines. In the past decades, overall pressure ratios (OPR) of turbofan engines have increased, leading to improved thermal efficiency. In addition, increasing bypass-ratios (BPR) have led to higher propulsive efficiencies (IPCC, 1999). The trends towards higher OPR and BPR can be expected to continue. Besides conventional turbofans with ultrahigh BPR, geared turbofans and open rotor engines are being developed. A geared turbofan decouples the rotational speeds of the fan and the low-pressure turbine, enabling a larger fan with higher BPR and, at the same time, a comparably light and efficient lowpressure turbine with high rotational speed. For the PW1000G family, for example, the manufacturers state a TSFC-advantage in the order of 15% compared to engines from the previous generation (MTU Aero Engines, 2011). This engine family is one of two competing engine options for the A320 NEO family. As an alternative concept, open rotor engines have the potential for even lower fuel consumption. However, disadvantages of open rotors include less favourable noise emissions, lower optimal flight speeds and challenges with respect to aircraft-engine integration. A drawback of increasing pressure ratios in modern engines is the higher potential for NOx production. Higher combustion temperatures and pressures negatively affect NOx emissions (IPCC, 1999). Improved combustion technology is required in order to limit or reduce the NOx output without affecting flame stability and fuel efficiency. The majority of aircraft engines use the Rich-Burn/Quick-Mix/Lean-Burn (RQL) concept for reasons of stability and performance. An alternative has been introduced by General Electric with the GEnx engine powering the Boeing 787. Lean burn direct injection (LDI) or twin annular premixing swirler (TAPS) combustors like in the GEnx combine a fuel-rich pilot zone with a parallel fuel-lean main zone. The pilot zone is operated independently of the power setting, while the main zone is only fuelled at medium and high power. Lean combustion in the main zone enables low NOx emissions, while the pilot burner ensures flame stability. In cruise flight, LDI combustors are claimed to reduce NOx emissions by up to 50% compared to RQL technology (Ralph et al., 2009).

4

M. Schaefer

Apart from improvements in aerodynamics and engine technology, aircraft weight can be reduced by lightweight construction and new materials like composites. Besides the direct influence of weight reduction on fuel burn, advanced materials enable additional improvements in aerodynamics or engine technology. Further reduction of fuel consumption can be reached by more efficient ATM and aircraft operations.

3

Emissions forecast model

3.1 Model overview The model used to forecast air traffic emissions consists of a combination of software and database files covering aircraft and engine performance and emissions, flight simulation, worldwide aircraft movements and forecast scenarios (Schaefer, 2012). A bottom-up approach is followed to calculate fuel burn and emissions of air traffic, i.e. emissions are calculated individually for each flight from a flight movement’s database. The bottom-up approach is followed for historical traffic and for a year-by-year forecast. Simulation models of current and future aircraft types are used in combination with the database of flights and, for the forecast, assumptions about traffic growth, fleet rollover and ATM. A schematic of the model is shown in Figure 1. The main components and its input assumptions will be described below. Simulation results include emissions of air traffic for the year 2010 and a forecast of emissions until 2030. Figure 1

Schematic of the emissions forecast module (see online version for colours)

3.2 Base year movements and emissions The flight movements and emissions database for the year 2010 is based on monthly flight schedules for January until December obtained from the Official Airline Guide (OAG Aviation Solutions, 2010) and supplemented by additional information required for emissions calculations. The commercially available OAG data contain worldwide scheduled connections including departure and arrival airports, aircraft types, departure and arrival times as well as the days of operations. For this study, OAG data has been


5

supplemented by load factor information based on ICAO statistics by flight stage, world region or airline (ICAO, 2011). Amongst the aforementioned load factor sources, the most detailed source available for each flight is used. For emissions calculations, engine types from the ICAO engine emissions databank for jet engines (ICAO, 1995) or the emissions database for turboprop engines from the Swedish Defence Research Agency FOI (2007) need to be available. Engine types are assigned stochastically to each flight based on airline- and aircraft-type-specific fleet statistics obtained from a fleet’s database (ASCEND, 2011). Aircraft models from the BADA database (EUROCONTROL, 2011) are used to simulate each flight and calculate its fuel consumption. BADA models are used with DLR’s VarMission aircraft performance software (Schaefer, 2012), which also calculates NOx emissions along the trajectory using the DLR fuel flow correlation method (Deidewig et al., 1996). Generic trajectories between departure and arrival airports are assumed, using great-circle routes, atmospheric conditions according to the International Standard Atmosphere (ISA) and neglecting wind for simplicity. In order to speed up calculations, VarMission is not directly called for each flight. Instead, interpolation tables named ‘emission profiles’ have been produced, which specify generic flight profiles and corresponding emissions for each aircraft-engine combination as function of flight distance and load factor. Typical mission rules, cruise altitudes and reserve fuel quantities are assumed here. By use of the emission profiles, emissions along the trajectory of each flight are estimated. Results are stored in a 3D coordinate system around the globe. In a post-processing step, assumptions about inefficiencies attributable to ATM are made: 7% additional fuel burn and emissions due to ATM are assumed for the year 2010 in accordance with a study by the Civil Air Navigation Services Organisation (CANSO, 2008). Visualisation of results is optional and can be performed by the RACE module developed at the DLR Institute of Air Transport and Airport Research (see Figure 2). While the focus of this paper is on global emissions, the model also enables regional evaluations of historical and forecasted emissions (see Schaefer et al., 2013). Figure 2

Regional distribution of NOx emissions in 2010 and 2030 (see online version for colours)

6

M. Schaefer

3.3 Regional traffic growth In a year-by-year simulation, the database of flight movements and emissions for the base year is converted into a database for a forecast year. Traffic growth is considered first, while changes in the worldwide fleet of aircraft are simulated in a second step. Average annual growth rates are obtained from the Airbus Global Market Forecast (GMF) 2011–2030 (Airbus, 2011). The GMF is one of the two most influential market forecasts by manufacturers of transport aircraft – the other being Boeing’s Current Market Outlook (CMO). These studies, which aim at forecasting regional traffic growth and the demand for new aircraft over a 20 years period, can be regarded as a reference for such projections. Both publications are used by ICAO working groups, e.g., when assessing the potential effects of future emission standards. The GMF was chosen here for practical reasons, as its growth rates were made available to the author as function of time (e-mail conversation with Paul Bonnabau from Airbus, November 2011). Consequently, growth rates for revenue passenger-kilometres (RPK) can be specified separately for different time periods and regions, as shown in Table 1. Table 1

Excerpt of growth rates for passenger air traffic from the GMF

From

To

Dom/Int*

2011–2015

2016–2020

2021–2030

Western Europe

Western Europe

D

2.7%

2.9%

2.7%

Western Europe

Western Europe

I

3.8%

3.4%

2.8%

Western Europe

Central Europe

7.3%

6.2%

4.7%

Western Europe

Russia

5.8%

5.0%

4.4%

Western Europe

China

7.4%

6.5%

5.4%

Western Europe

Japan

2.0%

3.5%

2.8%

Western Europe

Indian Sub

7.9%

5.8%

5.7%

Western Europe

Asia

5.3%

4.5%

3.8%

Note: *Marker for domestic (D) or international (I) traffic

For the period 2011–2030, the average annual growth rate for passenger traffic amounts to 4.8% according to the GMF and is only slightly lower than the respective Boeing value of 5.1% (Boeing, 2011). Predicted growth rates for freight traffic from these sources are also of comparable magnitude (Airbus: 5.9%, Boeing: 5.6%). In the first step of the modelling, traffic growth rates are applied as factors onto the monthly frequencies of the base year flights, using OAG-provided regions and country codes to assign the most appropriate rate to each flight. It is obvious that this approach is simplified as, for example, no new city pairs are connected. However, the approach should still be sufficiently accurate for regional or global assessments and for the purpose of the emissions forecast. As the GMF 2011–2030 does not cover cargo flights, growth rates for freight have been obtained from the previous edition of the GMF (Airbus, 2010). The modelling is slightly different than for passenger flights, as given growth rates refer to total freight tonne-kilometres transported (FTKT), i.e. freight on cargo flights, but also belly cargo on passenger flights. Growth rates for FTKT are mostly similar or slightly larger than those for RPKs. Consequently, traffic growth is simulated for passenger traffic (including belly freight) first. Dedicated freight traffic is considered afterwards, when the frequencies of


7

cargo flights are increased by a factor that is required to match the yearly growth in terms of FTKT.

3.4 Aircraft fleet forecast model In the second step of the forecast, changes in the aircraft fleet are accounted for. Year by year, old aircraft are retired while new aircraft are introduced. The number of aircraft in service and the age structure of the base year fleet by aircraft type and airline are obtained from ASCEND (2011). The age distribution can be linked to the flight schedule entries by stochastically assigning an aircraft build year to each flight, such that the relative proportions of different build years correspond to statistics from ASCEND. In this assignment, proportionality is assumed between the number of a given aircraft type in the fleet statistics and the transport performance by this aircraft type in the movements database. In order to simulate aircraft retirements, assumptions about aircraft lifetimes are required. Retirement curves specify the ‘survival’ percentage of an aircraft as function of its age. Based on historical statistics, retirement curves have been produced for three categories of aircraft as shown in the left diagram of Figure 3. Given an aircraft type’s age distribution in a base year and a retirement curve, the number of active aircraft in the following year (and hence the number of retirements) can be estimated. A simplification is used for cargo aircraft: For wide body freighters, a constant lifetime of 30 years is assumed while for narrow body freighters a lifetime of 25 years was estimated. Figure 3

Retirement curves and deliveries by seat category (see online version for colours)

The year-by-year fleet rollover is based on the assumption that the available transport performance (measured in tonne-kilometres offered, TKO) provided by an aircraft type is proportional to the number in service of this particular type. After simulating traffic growth, the TKOs that can be covered by the base year fleet are determined for the forecast year, given the number of aircraft in the base year and the projected retirement number. The remaining flights in the forecast year need to have their aircraft replaced by newly delivered planes, assuming that no constraints exist regarding manufacturers’ production capacities. A flight frequency correction is applied if a replacement aircraft of a different seat category is chosen by the assignment algorithm. The frequency correction is chosen such that the transport performance of the flight schedule entry remains unchanged. The number of deliveries by size category, a result of the simulation, is shown in the right

8

M. Schaefer

hand diagram of Figure 3. The relative number of deliveries by seat category is specified as an input parameter: Assumptions chosen here are based on a combination of data from the GMF (Airbus, 2011) for large transport aircraft and the Embraer Market Outlook (Embraer, 2011) for regional aircraft. The GMF was chosen for consistency with the traffic growth assumptions described above. Amongst the forecasts for regional aircraft – a segment that is not covered in much detail by the GMF – the Embraer Market Outlook is used, as its seat categories are compatible to those used by Airbus. As a result, the size distribution of the simulated fleet is very similar to respective predictions by Airbus and Embraer. Incompatible seat categories in other publications make a comparison of results difficult: A difference in the Boeing CMO for 2030 (Boeing, 2011) compared to both the Airbus and Embraer publications is its higher expected market share for single-aisle aircraft with more than 90 seats, at the expense of lower shares for smaller aircraft. The share of wide body aircraft, however, that is predicted by Boeing was found to be close to the current forecast. Most aircraft families expected for the near future are considered for the forecast, including the new engine options for A320 and 737, the A350 family as well as an upgrade of the 777. Respective delivery periods correspond to plans by the manufacturers. Market shares between competing aircraft types within a seat category are in accordance with order backlogs from ASCEND (2011) or, for aircraft of the more distant future, are assumed to be equally split between manufacturers. Similar assumptions are used for the assignment of engine types.

3.5 Aircraft and engine models 117 representative aircraft models from BADA version 3.9 (EUROCONTROL, 2011) are used to calculate air traffic’s emissions for the year 2010. For the forecast, a selection of aircraft types of the near future were simulated by total energy models similar to those from BADA. These new models supplement BADA for the simulation of future traffic. By use of VarMission, emission profiles as described in chapter 3.2 were calculated for the new models and have been applied for the forecast. New simulation models were created for the 787–8, 747–8, A350–900, the A320 NEO family and a regional jet similar to the Mitsubishi MRJ90. These models consist of characteristic aircraft weights, aerodynamic drag polars and thermodynamic engine models. As an example, drag polars for the 787 models and fuel flow vs. thrust for the 787’s simulated engine are shown in Figure 4. As far as available, characteristic weights (including empty weight and max. take-off weight) were obtained from literature and mostly correspond to the manufacturers’ design targets. Aircraft aerodynamics is simulated by drag polars for different flap settings. Unlike in BADA, Mach-number dependent polars are used for the simulation, as far as plausible models were available. Parameters for which no data were available have been estimated by predesign methods described by Roskam (2003). DLR engine models are used for these simulated aircraft to calculate fuel burn and NOx emissions.

Forecast of air traffic’s CO2 and NOx emissions until 2030 Figure 4

9

L/D and engine fuel flow vs. thrust for generic 787 model (see online version for colours)

NOx emissions during cruise were estimated by the DLR fuel flow correlation method (Deidewig et al., 1996), as long as RQL combustors are concerned. Fuel flow correlations are not suitable for lean combustion like in the TAPS combustor of the GEnx. A staged p3-T3 correlation is used in order to predict cruise NOx emissions for the GEnx engine, which is one of two engine options simulated for the 787 model. Details about NOx calculation and the new aircraft models are found in Schaefer (2012). Given the large number of aircraft types that are newly introduced in the fleet forecast, simplified assumptions need to be used for those types, for which no detailed simulation models are available. Fuel efficiency and specific NOx emissions of these aircraft have been estimated relative to those of existing models (see Table 2). Table 2

Assumptions for passenger aircraft without simulation model

Aircraft type 737 MAX 7/8/9

Representative type

Assumptions for representation

A319/320/321 NEO*

Equal CO2 and NOx per kg payload

787–8


A350–800/1000

A350–900


777–200/300 Successor

A350–900

Equal CO2 per kg payload, –30% NOx per kg payload**

Generic Regional Jet (similar to MRJ-90ER)

Equal CO2 per kg payload, –30% NOx per kg payload**

CRJ & E-Jet Succ. (70 seats)

E170 (BADA)

–20% CO2 and NOx per kg payload

ERJ Successor (50 seats)

E145 (BADA)


ATR 72 (BADA)


ATR 42 / 72 (BADA)


787–9

E-Jet Successor (98 seats) CRJ Successor (90 seats)

Future Turboprop (90 seats) ATR and DHC Succ. (70 seats)

Notes: *LeapX engine option; **Assuming TAPS technology

Based on simulation results for the GEnx engine with TAPS combustor, a 30% improvement of NOx emissions on flight mission level has been estimated for this technology compared to RQL combustors of the same technology level. This estimation is assumed as the potential of lean-burn combustion and is reflected in the assumptions shown in Table 2. Modern engine types with less favourable NOx emissions are assumed

10

M. Schaefer

to be equipped with low-NOx combustors from 2020 onwards, as shown in Table 3. In the forecast, improved combustors are used only for new deliveries, i.e. no retrofits of the existing fleet are assumed. Table 3

Assumptions for major combustor revisions

Engine

Aircraft

EIS improved combustor

Assumed effect of improved combustor NOx as for GEnx

Trent 1000

B787

2020

Trent XWB

A350

2022

–30% NOx

Trent 900

A380

2024

–30% NOx

GP 7200

A380

2026

–30% NOx

3.6 ATM efficiency and load factors A minor increase of the average load factor is assumed until 2030. For 2010 the average seat load factor is 78% and the average weight load factor (including passenger and cargo) amounts to 65% (ICAO, 2011). These numbers are assumed to increase linearly to 81% and 68% respectively by 2030. The assumptions about the seat load factor are in line with the 80% value forecasted for 2025 in the ICAO Outlook for Air Transport (ICAO, 2007). ATM efficiency in the forecast is accounted for in the same way as described for the base year simulation: A certain percentage of inefficiency is assumed for the ATM system resulting in an equal percentage of additional fuel consumption. ATM efficiency is expected to increase with time, reflecting environmental goals described in CANSO (2008). An ATM efficiency of 93% is assumed for 2010 – corresponding to 7% additional fuel burn compared to ideal conditions – while efficiency is expected to rise to 95% by 2030.

4

Forecast of air traffic emissions until 2030

4.1 Overview on results Forecast results are shown in Table 4, including the number of flights, great-circle distance, revenue passenger-kilometres (RPK) and revenue tonne-kilometres (RTK, including passenger and freight transport). Growth rates of RPK and RTK with time reflect the input assumptions about traffic growth described earlier. The simulation for the year 2010 is based on flight schedules from OAG. As a consequence, only scheduled air traffic is considered here, including most passenger flights worldwide. Scheduled flights by cargo aircraft are included in the model, but are likely to cover only a fraction of actual cargo traffic. It can be shown that the coverage of traffic in terms of RPK and RTK for the year 2010 is within 3–6% of comparable ICAO statistics (ICAO, 2010a).

Forecast of air traffic’s CO2 and NOx emissions until 2030 Table 4

11

Forecast of movements, transport performance and emissions

Year

Flights [106]

Distance [109 km]

RPK [1012 Pkm]

RTK [109 tkm]

Fuel [109 kg]

CO2 [109 kg]

NOx [109 kg]

2010

30.5

37.5

4.95

603

188

593

2.47

2015

38.1

47.7

6.59

801

241

760

3.20

2020

46.2

58.8

8.38

1,032

293

924

3.89

2025

54.2

70.1

10.33

1,301

344

1085

4.43

2030

64.1

84.0

12.74

1,640

405

1278

4.96

Throughout the forecast, fuel burn and emissions grow slower than RPK and RTK, mainly due to the introduction of new aircraft. Fuel efficiency measured in fuel consumption per tonne-kilometre is forecasted to improve by 21% from 313 g/tkm in 2010 to 247 g/tkm in 2030. CO2 emissions are proportional to fuel burn. NOx emissions per tonne-kilometre are forecasted to decrease by 27% from 4.1 g/tkm to 3.0 g/tkm.

4.2 Forecast of fleet composition The development of the aircraft fleet has been simulated by considering retirements of old aircraft and the introduction of new aircraft. As was explained in Section 3.4, the relative split of newly delivered aircraft between seat categories is an input to the model. The resulting fleet composition by seat category is shown in Figure 5. Figure 5

Forecasted fleet compositions by seat category

It can be seen that a comparably large number of small aircraft is responsible for only a small contribution towards the ASKs. On the other hand, a small number of large and very large aircraft (VLA) contribute significantly towards the ASKs in 2030 due to their high seat capacity and the long-range routes they are mostly used on. A trend towards larger aircraft is visible in the diagrams. The forecasted fleet by aircraft family is shown in Figure 6. In the wide body segment, newly introduced aircraft include the 787 and A350 families. A major update of the 777 is assumed to be introduced around 2020. The narrow body segment is dominated by the A320 and 737, while re-engined versions of these aircraft are introduced in the second half of the current decade. Embraer, Bombardier and ATR are assumed to

12

M. Schaefer

dominate the regional aircraft segements, where new market entries include Suchoi, Misubishi and Comac. Figure 6

Forecasted fleet development by aircraft family

4.3 Forecast of fuel consumption and CO2 emissions Figure 7 presents the fuel consumption calculated for scheduled air traffic until the year 2030. The diagram highlights the individual effects that influence fuel consumption and fuel efficiency. Previous model results based on historical flight schedules for the years 2000 and 2002–2009 (OAG, 2010) are also shown in the diagram. Figure 7

Forecast of fuel consumption and influencing factors (see online version for colours)


13

An average increase of fuel consumption per tonne-kilometre of 1.18% per year is calculated between 2010 and 2030. Amongst the contributing factors, the introduction of new aircraft has the largest impact on fuel efficiency in the order of 0.76% per year. Load factor increase and ATM efficiency improvement have a considerably lower impact on fuel efficiency of 0.1–0.2% per year. As CO2 emissions are proportional to fuel burn, growth rates for CO2 are equal to those for fuel consumption. Figure 8 compares the results to third-party studies with a similar focus, i.e. the ICAO fuel burn forecast until 2036 (ICAO, 2010b), AERO2k (Eyers et al., 2004) as well as results from NASA (Sutkus et al., 2003) and AEDT (Wilkerson et al., 2010). The left diagram shows fuel consumption, the right diagram shows fuel efficiency defined as fuel burn per RTK. Historical fuel burn calculated by the model is about 9% lower than results from AEDT for 2004 and 2006, which is due to the inclusion of unscheduled flights in these inventories for commercial aviation. ICAO has specified five scenarios S1-S5 in its forecast for 2016, 2026 and 2036. The growth of fuel consumption from the current study is comparable to the ICAO S3 scenario. The fuel burn forecast from AERO2k for 2025 is lower than in other studies, which can be blamed (amongst other factors) on lower traffic growth than assumed here. Figure 8

Forecast of fuel consumption and fuel efficiency (see online version for colours)

Historical values of fuel efficiency are slightly above the 2006 value from AEDT, which is also the base year value for the ICAO scenarios. The deviation between both studies amounts to 4% and should be within the expected error bars of such simulations. In early years of the current forecast, the predicted efficiency improvement resembles the ICAO scenario S1, while from 2016 onwards the curve is more similar to ICAO scenarios S3 and S4. The delays in the introduction of new aircraft types in recent years (e.g., A380, 787) may explain this deviation, considering that assumptions for the ICAO forecast date back to the year 2007.

4.4 Forecast of NOX emissions Forecasted emissions of NOx are shown in the left diagram of Figure 9 while specific emissions in gram NOx per RTK are shown on the right. Reference values from AERO2k (Eyers et al., 2004), NASA (Sutkus et al., 2003) and three scenarios by the ICAO/Committee on Aviation Environmental Protection (CAEP) (2009) are also presented in the figure. As all studies use different data for flight movements, traffic

14

M. Schaefer

growth, flight simulation and NOx calculation, the comparison below must be seen as a plausibility check rather than a validation. As seen in Figure 9, the year 2006 value from AEDT is higher than historical NOx emissions calculated by the methodology described here. This partly results from different coverage of global air traffic, as explained in Section 4.3, but also from different NOx calculation methods: this study and AERO2k use the DLR fuel flow correlation (Deidewig et al., 1996) while the Boeing fuel flow correlation (Baughcum et al., 1996) is used by AEDT and for the ICAO forecast. While both methods share the same principles, the Boeing approach was found to deliver slightly higher emissions on average than obtained by the DLR method, if applied to global traffic scenarios (Schaefer, 2012). Until 2016, the relative increase in the forecasted NOx emissions is similar to the ICAO S1 scenario, while in later years it resembles more the ICAO S2 or S3 scenarios. Current forecast results are in between the NOx emissions predicted by NASA for 2020 and lower predictions for 2025 from AERO2k. More conservative technology assumptions are likely to cause the higher NASA value, while the AERO2k value can be attributed to lower traffic growth and more optimistic technology projections. Figure 9

Forecast of NOx emissions and specific NOx emissions (see online version for colours)

Reference data for specific NOx emissions are available from the AEDT and the ICAO CAEP scenarios: The deviation between the base year value from AEDT/ICAO and the current results is almost fully attributable to the differences in the NOx correlations. Similar to the absolute NOx results, the predicted development of specific NOx emissions with time resembles ICAO S1 until 2016 and the S2 and S3 scenarios in later years. No technology improvements regarding NOx emissions are assumed in the ICAO scenarios after 2026, which explains their lower improvement rates after this year. In the current study, the decrease in specific NOx emissions with time is due to better fuel efficiency combined with only a minor increase of the NOx emission index (EI) until 2020. After 2020, the EI for NOx measured in gram emissions per kg fuel is forecasted to decrease again, which is due to the large-scale introduction of low-NOx combustors assumed from 2020 onwards (see Section 3.5).

5

Outlook on biofuels and IATA targets

Alternative drop-in fuels are currently discussed in the airline community and may help to enable a more environmentally friendly air traffic system. Compared to jet fuel from fossil sources, biofuels promise significant reductions of greenhouse gas (GHG)


15

emissions from a life-cycle perspective, as CO2 is absorbed while the feedstock is grown. Equivalent CO2 emissions (CO2e emissions) are typically looked at when comparing life-cycle emissions of different fuels. These CO2e emissions include GHG emissions during the production, distribution and consumption phase of the fuel. Fischer-Tropsch fuel produced from biomass has the potential of an 80–90% reduction of CO2e emissions compared to fossil jet fuel, while a lower potential of 40–70% has been identified for hydro-processed esters and fatty acids (HEFA) (Stratton et al., 2010). The effects of biofuels on CO2e emissions from aviation are visualised in Figure 10. The left diagram shows the reduction of CO2e emissions obtained by hypothetical biofuel shares in 2030. An 80% average reduction of life-cycle CO2e emissions is assumed for biofuels compared to conventional fuel. GHG goals from IATA (2009a) are shown for comparison. A 40% share of biofuels would be required in 2030 in order to reach the goal of carbon-neutral growth from 2020 onwards. Potential market shares in the order of 5–30% for the year 2030 have been discussed in the literature (Bauen et al., 2009; EQ2, 2010). Cost-competitiveness, production capacities and feedstock resources are bottlenecks, which make a fast ramp-up of production challenging (Novelli, 2011). Figure 10 Comparison of forecast results to IATA targets (see online version for colours)

For 2050, IATA intends a reduction of GHG emissions to 50% of the emissions from 2005 (IATA, 2009a). As 2030 is the last year of the current forecast, the most optimistic S5 scenario by ICAO is used as a reference for fuel consumption in 2050 (ICAO, 2010b). Figure 10 indicates that IATA’s long-term target is ambitious – and cannot be achieved by biofuels alone. Predictions of biofuel shares for 2050 range from 30% to 100% (Bauen et al., 2009). The analysis shows, however, that additional offset mechanisms (e.g., emissions trading with other sectors) are likely to be required if the environmental goals are to be met.

6

Conclusions

The paper introduced a simulation tool to forecast future fleet composition, fuel consumption and emissions of air traffic using an integrated chain of simulations. It combines publicly available data and models in order to produce a consistent forecast of CO2 and NOx emissions. The model was shown to deliver plausible results, as the forecast has been validated against reference studies. Its results currently serve as input

16

M. Schaefer

data to ongoing research about the climate impact of aviation at the DLR Institute of Atmospheric Physics. Future use of the model includes the analysis of alternative scenarios regarding traffic growth and technology development. Particularly the simulation of constraints to growth, e.g., with respect to airport capacity, is an ongoing research topic, and respective functionality may be implemented in the future.

References Airbus S.A.S. (2010) Global Market Forecast 2010–2029, Presentation by John Leahy, Toulouse. Airbus S.A.S. (2011) Delivering The Future / Global Market Forecast 2011–2030, Blagnac. Airbus S.A.S. (2014) Global Market Forecast / Flying on demand 2014–2033, Blagnac. Apffelstaedt, A., Langhans, S. and Gollnick, V. (2009) ‘Identifying carbon dioxide reducing aircraft technologies and estimating their impact on global CO2 emissions’, Paper presented at the German Aerospace Congress 2009, 8–10 September 2009, Aachen. Argüelles, P. et al. (2001) European Aeronautics, A Vision For 2020 – Meeting Society’s Needs and Winning Global Leadership, Report of The Group of Personalities, Luxembourg. ASCEND (2011) ASCEND Online Fleets, Online Database Product obtained from 2008 to 2011 [online] http://www.ascendworldwide.com (accessed 12 January 2011). Bauen, A., Howes, J., Bertuccioli, L. and Chudziak, C. (2009) Review of the Potential for Biofuels in Aviation, E4tech, London. Baughcum, S.L., Tritz, T.G., Henderson, S.C. and Pickett, D.C. (1996) Scheduled Civil Aircraft Emission Inventories for 1992, Database Development and Analysis, NASA Contractor Report 4700, Hampton (VA). Boeing (2011) Current Market Outlook 2011–2030, Seattle. Boeing (2014) Current Market Outlook 2014–2033, Seattle. Civil Air Navigation Services Organisation (CANSO) (2008) ATM Global Environment Efficiency Goals for 2050, Hoofddorp. Deidewig, F., Döpelheuer, A. and Lecht, M. (1996) ‘Methods to assess aircraft engine emissions in flight’, ICAS Proceedings 1996, Vol. 1, Paper No. ICAS–96–4.1.2, Sorrento. Embraer S.A. (2011) Embraer Market Outlook 2011–203, São José dos Campos (SP). EQ2 (2010) Sustainable Flying – Biofuels as an Economic and Environmental Salve for the Airline Industry, London. EUROCONTROL (2011) User Manual for the Base of Aircraft Data (BADA) Revision 3.9, EEC Technical/Scientific Report No. 11/03/08–08, Brétigny-sur-Orge. European Commission (2011) Flight Path 2050 – Europe’s Vision for Aviation, Report of the High Level Group on Aviation Research, Luxembourg. Eyers, C.J., Norman, P., Middel, J., Plohr, M., Michot, S., Atkinson, K. and Christeou, R.A. (2004) AERO2k Global Aviation Emissions Inventories for 2002 and 2025, QinetiQ Ltd, Farnborough. Intergovernmental Panel on Climate Change (IPCC) (1999) Aviation and the Global Atmosphere, Cambridge University Press, Cambridge. International Air Transport Association (IATA) (2009a) Aviation and Climate Change/Pathway to Carbon-Neutral Growth in 2020, Geneva. International Air Transport Association (IATA) (2009b) The IATA Technology Roadmap Report, 3rd ed., Geneva. International Civil Aviation Organization (ICAO) (1995) ICAO Engine Exhaust Emissions Databank, ICAO Doc 9646-AN/943, 1st ed. (plus electronic revision issue 17), Montréal [online] http://www.caa.co.uk (accessed 8 March 2011).


17

International Civil Aviation Organization (ICAO) (2007) Outlook for Air Transport to the Year 2025, Montréal. International Civil Aviation Organization (ICAO) (2010a) Civil Aviation Statistics of the World, ICAO Statistical Yearbook, Various Editions, Montréal. International Civil Aviation Organization (ICAO) (2010b) Environmental Report 2010, Montréal. International Civil Aviation Organization (ICAO) (2011) ICAO Data [online] http://www.icaodata.com (accessed 8 June 2011). International Civil Aviation Organization (ICAO)/Committee on Aviation Environmental Protection (CAEP) (2009) Environmental Goals Assessment, CAEP/8-IP/8, Montréal. Lee, D.S., Fahey, D.W., Forster, P.M., Newton, P.J., Wit, R.C.N., Lim, L.L., Owen, B. and Sausen, R. (2009) ‘Aviation and global climate change in the 21st century’, Atmospheric Environment, Vol. 43, Nos. 22–23, pp.3520–3537. Lyssis Ltd (2010) PIANO-X Aircraft Performance and Emissions Software [online] http://www.piano.aero (accessed 11 November 2010). MTU Aero Engines (2011) PurePower PW 1000G Engine [online] http://www.mtu.de/de/ products_services/new_business_commercial/programs/pw1000g/PW1000G_brochure.pdf (accessed 8 April 2011). Nolte, P., Zill, T., Apffelstaedt, A., Stumpf, E., Dufresne, S., Briceno, S., Racuynksi, C. and Rötger, T. (2011) ‘Technology screening, selection and modelling for an environmentally sustainable aviation’, Proceedings of the 11th AIAA Aviation Technology, Integration, and Operations (ATIO) Conference, AIAA 2011–6968, Virginia Beach (VA). Novelli, P. (Ed.) (2011) SWAFEA/Sustainable Way for Alternative Fuels and Energy in Aviation, Final Report, Palaiseau. OAG Aviation Solutions (2010) OAG MAX, Flight Schedules on CD-ROM, Monthly Editions Covering the Years 2000 and 2003–2010, Luton. Owen, B. and Lee, D.S. (2006) Study on the Allocation of Emissions from International Aviation to the UK Inventory – CPEG7/Final Report to DEFRA Global Atmosphere Division, Report 3 of 3, Manchester Metropolitan University, Manchester. Owen, B., Lee, D.S. and Lim, L. (2010) ‘Flying into the future, aviation emissions scenarios to 2050’, Environmental Science & Technology, Vol. 44, No. 7, pp.2255–2260. Peeters, P.M., Middel, J. and Hoolhorst, A. (2005) Fuel Efficiency of Commercial Aircraft – An Overview of Historical and Future Trends, NLR-CR-2005-669, Amsterdam. Ralph, M., Kuentzmann, P., Maurice, L. and Tilston, J. (2009) Report of the Independent Experts to CAEP 8 on the Second NOx Review & Long Term Technology Goals, CAEP/8–WP/10, Montréal. Roskam, J. (2003) Airplane Design – Part I: Preliminary Sizing of Airplanes, 3rd ed., DARcorporation, Lawrence (KS). Schaefer, M. (2012) Development of a Forecast Model for Global Air Traffic Emissions, DLR Forschungsbericht 2012–08, Cologne/Bochum. Schaefer, M., Jung, M. and Pabst, H. (2013) ‘The regional distribution of air traffic emissions in the past, present and future’, Proceedings of the German Aerospace Congress 2013, DocumentID: 301154, Stuttgart. Stratton, R.W., Wong, H.M. and Hileman, J.I. (2010) Life Cycle Greenhouse Gas Emissions from Alternative Jet Fuels, PARTNER Project 28 Report, Ver. 1.2, PARTNER–COE–2010–001, Cambridge (MA). Sutkus Jr., D.J., Baughcum, S.L. and DuBois, D.P. (2003) Commercial Aircraft Emission Scenario for 2020: Database Development and Analysis, NASA/CR-2003-212331, Washington. Swedish Defence Research Agency FOI (2007) Confidential Database for Turboprop Engine Emissions [online] http://www.foi.se/en/Our-Knowledge/Aeronautics/FOIs-Confidentialdatabase-for-Turboprop-Engine-Emissions (accessed 22 April 2009).

18

M. Schaefer

Wilkerson, J.T., Jacobson, M.Z., Malwitz, A., Balasubramanian, S., Wayson, R., Fleming, G., Naiman, A.D. and Lele, S.J. (2010) ‘Analysis of emission data from global commercial aviation: 2004 and 2006’, Atmospheric Chemistry and Physics, Vol. 10, No. 13, pp.6391–6408.