Models, Data and Assumptions

This appendix provides an overview of the data sources, models and assumptions used to develop the information presented in Chapter 1 (Overview of Aviation Sector) and Chapter 2 (Technology and Design). These modelling capabilities have been developed and used to support various European initiatives, including SESAR and Clean Sky, as well as international policy assessments in ICAO CAEP.


The information in this report covers all flights from or to airports in the European Union (EU) and European Free Trade Association (EFTA). For consistency, regardless of the year, the EU here consists of the 28 member States: Austria, Belgium, Bulgaria, Croatia, Republic of Cyprus, Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Ireland, Italy, Latvia, Lithuania, Luxembourg, Malta, Netherlands, Poland, Portugal, Romania, Slovakia, Slovenia, Spain, Sweden and the United Kingdom. EFTA members are Iceland, Liechtenstein, Norway and Switzerland.

Data sources


Historical 2005-2017 flight operations were extracted from the EUROCONTROL database of filed flight plans called PRISME. PRISME covers all Instrument Flight Rules flights in Europe. Flight data are enriched with and validated against, for example, radar updates, billing data from the Central Route Charges Office and an internal database of global aircraft (PRISME Fleet).


European States collect statistics on air transport from their airports and airlines and provide these to Eurostat, which makes them public, although airline details are treated as confidential. Statistics on total activity (total passengers, total tonnes shipped, etc.) are as complete as possible. More detailed statistics, such as passengers and available seats for individual airport pairs, are focused on major flows. For example, we use these data to indicate trends in load factors, but we cannot calculate total available seat-kilometres solely from them. The estimates of total passenger kilometres flown in Chapter 1 are based on Eurostat directly, on analysis of other Eurostat flows and on data from PRISME. The great circle (i.e. shortest) distance between airport pairs is used when reporting passenger kilometres, while the actual flown distance is used when calculating the average fuel consumption.


The EUROCONTROL STATFOR 20-year forecast that was published in 2018 provided the traffic volumes from 2017 to 2040 used in this report. In this report, we focused on three of the four forecast scenarios: Regulation & Growth is the most- likely or ‘base’; Global Growth gives the ‘high’; and Fragmenting World gives the ‘low’. The forecast was prepared as part of the Challenges of Growth 201827 study. 112 airports provided future capacity plans to this study, and the forecast traffic respects the capacity constraints implied by these plans.


BADA (Base of Aircraft Data) is an Aircraft Performance Model developed and maintained by EUROCONTROL, in cooperation with aircraft manufacturers and operating airlines. BADA is based on a kinetic approach to aircraft performance modelling, which enables to accurately predict aircraft trajectories and the associated fuel consumption. BADA includes both model specifications which provide the theoretical fundamentals to calculate aircraft performance parameters, and the datasets containing aircraft-specific coefficients required to calculate their trajectories. The BADA 3 family is today’s industry standard for aircraft performance modelling in the nominal part of the flight envelope, and provides close to 100% coverage of aircraft types operating in the European region. The latest BADA 4 family provides increased levels of precision in aircraft performance parameters over the entire flight envelope, and covers 70% of aircraft types operating in the European region. This report uses BADA 4, complemented by BADA 3 for aircraft types not yet covered in BADA 4.

Aircraft Noise and Performance (ANP) Database

The Aircraft Noise and Performance (ANP) database is maintained by the US Department of Transportation, EUROCONTROL and EASA. It provides the noise and performance characteristics for over 150 civil aircraft types, which are required to compute noise contours around civil airports using the calculation method described in Annex II of European Directive 2002/49/EC relating to assessment and management of environmental noise, ECAC Doc 29 and ICAO Doc 9911 guidance documents. ANP datasets are supplied by aircraft manufacturers for specific airframe-engine types, in accordance with specifications developed by the ICAO and European bodies. EASA is responsible for collecting, verifying and publishing ANP data for aircraft which fall under the scope of Regulation (EU) 598/2014.

EASA Certification Noise Levels28

EASA maintains a database of all aircraft noise certification levels which the Agency has approved. The database provides certified noise levels for over 34,000 aircraft variants, including jet, heavy and light propeller aircraft as well as helicopters. In this report, the certified noise levels are used to assess the Noise Energy Index, to attribute an ANP airframe-engine type to each aircraft type in the fleet using the ECAC Doc 29 4th Edition recommended substitution method, as well as to create the noise charts in the Technology and Design chapter.

ICAO Aircraft Engine Emissions Databank (EEDB)29

The ICAO Aircraft Engine Emissions Databank (EEDB) hosted by EASA contains Landing and Take-Off (LTO) emissions data for NO , HC, CO as well as smoke number for over 400 jet engine types. The EEDB emission indices are used by the IMPACT model to compute NOx , HC, CO and PM, and to create the NOx charts in the Technology and Design chapter.

FOI Turboprop Emissions Database30

The Swedish Defence Research Agency (FOI) hosts a database of NO , HC and CO emission indices for turboprop engine types. The data was supplied by the turboprop engine manufacturers, originally for the purposes of calculating emissions- related landing charges. It is used to complement the ICAO EEDB for the NO , HC and CO estimates in this report.

CODA Taxi Times Database31

EUROCONTROL’s Central Office for Delay Analysis (CODA) collects flight-by-flight data from around 100 airlines and 130 airports, such as actual off-block and take-off times, and delay causes. Largely this is on a voluntary basis in return for performance and benchmarking information, but increasingly the data collection is influenced by the EU performance regulations [52]. CODA publishes aggregated performance statistics, such as on punctuality and all-causes delays from these data. The detailed actual taxi times from this source were used to assess taxi fuel burn and emissions.

Population Data32

The JRC Global Human Settlement population grid was used to calculate the number of people exposed to aircraft noise. This spatial dataset, developed in the European Copernicus Program, depicts the distribution and density of residential population. The dataset is generated using the 2011 censuses provided by Eurostat/GEOSTAT and the best available sources by country. The initial 1 km resolution has been further disaggregated to 100 m based on information from Corine Land Cover Refined 2006 and the European Settlement Map 2016.

Models and methods


IMPACT is a web-based modelling application used to assess the environmental impacts of aviation, and whose development, initiated in the context of the SESAR 1 programme, has since been steered and carried out by EUROCONTROL. It allows the consistent assessment of trade-offs between noise and full-flight gaseous emissions thanks to a common advanced aircraft performance-based trajectory model using a combination of the ANP database and the latest release of the BADA family. CO2, NOX, HC, CO and PM emissions are computed using the LTO emission indices in the ICAO EDB and FOI Turboprop Emissions database combined with the Boeing Fuel Flow Method 2 (BFFM2). PM emission indices of jet engines are estimated using the First Order Approximation (FOA3.0) method33. Both BFFM2 and FOA3.0 methods are detailed in the ICAO Airport Air Quality Manual (Doc 9889). The IMPACT methodology and data to assess fuel burn and emissions may differ from that used by Member States to report their emissions to UNFCCC or CLRTAP, hence the delta in estimates between these data sources.

System for Airport Noise Exposure Studies (STAPES)

STAPES is a multi-airport noise model jointly developed by the European Commission, EASA and EUROCONTROL. It consists of a software compliant with Annex II of Directive 2002/49/EC and ECAC Doc 29 modelling methodology, combined with a database of airports with information on runway and route layout, as well as the distribution of aircraft movements over these runways and routes. The 47 European airports within EU28 and EFTA modelled in STAPES are estimated to cover approximately three quarters of the total population exposed to aircraft noise levels of Lden 55 dB and above in this region.

Aircraft Assignment Tool (AAT)

AAT is a fleet and operations forecasting model jointly developed by the European Commission, EASA and EUROCONTROL. AAT converts a passenger demand forecast into detailed operations by aircraft type and airport pair for a given future year and scenario, taking into account aircraft retirement and the introduction of new aircraft into the fleet. It is now an integral part of the STATFOR 20-year forecast methodology. The forecast operations are processed through the IMPACT and STAPES models to assess the fuel burn, emissions and noise data in 2025, 2030, 2035 and 2040 presented in the Sector Overview chapter.


AERO is an application owned by EASA that is used to examine the impacts of different policies intended to reduce international and domestic aviation greenhouse gas emissions. AERO can assess the consequences of a wide range of policy measures aimed at reducing aviation emissions, including technological, operational and market-based measures. For this report, AERO was used to assess the impact of the EU Emissions Trading System in years 2018, 2019 and 2020 based on the STATFOR medium-term ‘Base’ forecast (see Sector Overview and Market-Based Measures chapters) and traffic data supplied by EUROCONTROL. No analysis was conducted beyond 2020 due to on-going discussions on ways to implement CORSIA in the EU through a revision of the ETS Directive, which would be consistent with the EU 2030 climate and energy framework.


Fuel burn, emissions and noise assessment

For consistency with other international emission inventories, full-flight emissions presented in this report are for all flights departing from EU28 or EFTA, i.e. flights coming from outside EU28 or EFTA are not included. In contrast, emissions below 3,000 feet and noise indicators include all departures and all arrivals. Historical fuel burn and emission calculations are based on the actual flight plans from PRISME, including the actual flight distance and cruise altitude by airport pair. Future year fuel burn and emissions are based on actual flight distances and cruise altitudes by airport pair in 2016. Future taxi times are assumed to be equal to the 2017 taxi times; where non available, ICAO default taxi times are applied. Helicopter operations are excluded from the assessment.

For the STAPES noise assessments, the number of airports, together with their respective runway and route layout, were assumed to be constant over the full analysis period – i.e. only the fleet, the number and time of operations vary. The standard take-off and landing profiles in the ANP database were applied. For historical noise, the day/evening/night flight distribution was based on actual local departure and landing times assuming the Environmental Noise Directive [6] default times for the three periods: day = 7:00 to 19:00, evening = 19:00 to 23:00, night = 23:00 to 7:00. For future years, the day/ evening/night flight distribution at each airport was assumed to remain unchanged. Population density around airports was also assumed to remain unchanged throughout the analysis period. The mapping of the fleet to the ANP aircraft follows the ECAC Doc 29 4th Edition recommended substitution method.

In addition to the noise contours at the 47 airports modelled in STAPES, the noise generated by aircraft take-offs and landings at all airports in the EU28 and EFTA area was estimated via the noise energy index, defined as:


Ndep and Narr are the numbers of departures and arrivals by aircraft type weighted for aircraft substitution;
LAT, FO and APP are the certified noise levels in EPNdB at the three certification points (lateral, flyover, approach) for
each aircraft type34.

Noise exposure-response curves

To estimate the total population highly annoyed (HA) and highly sleep disturbed (HSD) by aircraft noise, the following exposure-response regression curves recommended by WHO for the European region were used [16]:

Share of population highly annoyed (%HA) = -50.9693 + 1.0168 * Lden + 0.0072 * Lden2

Share of population highly sleep disturbed (%HSD) = 16.79 - 0.9293 * Lnight + 0.0198 * Lnight2

The total population at the 47 major airports in STAPES was assessed for Lden values between 45 and 75 dB and for Lnight values between 40 and 70 dB with one decibel increment, and then multiplied by the corresponding %HA and %HSD values. As the Lden and Lnight values represent outdoor noise levels, the annoyance and sleep disturbance estimates may not take into account the effect of local sound insulation campaigns for houses and buildings around airports.

Future fleet technology scenarios

Future noise and emissions in the Sector Overview chapter were assessed for different technology scenarios. The most conservative ‘frozen technology’ scenario assumes that the technology of new aircraft deliveries between 2017 and 2040 remains as it was in 2017. Under this scenario, the 2017 in-service fleet is progressively replaced with aircraft available for purchase in 2017. This includes the A320neo, B737 MAX, Airbus A220 (or Bombardier CSeries), Embraer E-Jet E2, etc.

On top of the fleet renewal, technology improvements for fuel burn (CO2), NOX and noise are applied on a year-by-year basis to all new aircraft deliveries from 2017 onwards following a single ‘advanced’ technology scenario. This technology scenario was derived from analyses performed by groups of Independent Experts for the ICAO CAEP, and is meant to represent the maximum noise and emission reductions that can be expected from aircraft and engine technology by 2040.

For noise, the advanced technology scenario modelled for this report assumes a reduction of 0.1 EPNdB per annum at each noise certification point for new aircraft deliveries. The previously reported 0.3 EPNdB reduction per annum scenario was considered too optimistic given the recent improvement in aircraft noise technology and the general trend towards heavier aircraft, and was therefore left aside.

For fuel burn and CO2, the advanced technology scenario assumes a 1.16% improvement per annum for new aircraft deliveries35. For NOX, the scenario assumes a 100% achievement of the CAEP/7 NOX Goals by 203636. No technology improvement was applied when estimating future HC, CO and PM emissions.

Lastly, the technology improvement assumptions do not take into account potential future aircraft designs like supersonic or counter-rotating open rotor powered aircraft.

Future ATM improvements

The existing ATM system efficiency is assumed to remain unchanged despite future increases in overall air traffic. As a first order approximation, fuel burn and emission gains can be directly deducted from the projected ATM-related fuel efficiency gains (e.g. a 3% fuel efficiency improvement can be assumed to generate a 3% reduction in total fuel burn and emissions).









32  European Commission, Joint Research Centre (JRC) [Dataset]

33 Due to the lack of smoke number data for turboprop engines, PM estimates currently exclude this category. As an indication, turboprop aircraft represented approximately 1.5% of the total fleet fuel burn in 2017.

34 For Chapter 6 and 10 aircraft (light propeller), the unique overflight or take-off level is used for the three values.

35 ICAO Environmental Report 2010 (p. 33).

36 ICAO Environmental Report 2010 (p. 29).