Aero Model

Download as pdf or txt
Download as pdf or txt
You are on page 1of 98

Aviation Emissions and Evaluation of Reduction Options

(AERO)
Main Report
Part I: Description of the AERO Modelling System
December 2000
Table of contents
1. Context and aims of AERO Project
1.1 Background and objectives
1.2 Applications of the AERO modelling system
1.3 Parties involved in project execution
1.4 Set-up of the overall AERO documentation
1.5 Purpose and overview of Main Report
2. Problems related to aircraft engine emissions
2.1 Nature, extent and development of aircraft engine emissions
2.2 Environmental problems related to aircraft engine emissions
3. Overview of AERO modelling approach
3.1 Computational steps
3.2 General modelling principles and modelling dynamics
3.2.1 Future developments with and without measures
3.2.2 Modelling dynamics and time aspects
3.3 Schematisation aspects
3.3.1 Air transport system specifications
3.3.2 Spatial detail in AERO
3.4 Overview of models in AERO modelling system
4. Aircraft technology model (ATEC)
4.1 Purpose of ATEC within the AERO modelling system
4.2 Fleet composition and aircraft characteristics
4.3 From aircraft purchases to fleet characteristics
4.4 Scenario developments considered in ATEC
4.5 Policy measures considered in ATEC
4.6 Aircraft new price and costs associated with measures
4.7 Calibration
4.8 Fleet matching from Base to Datum
5. The Unified Database
5.1 Introduction
5.2 Data content and structure of the Unified Database
5.3 Creation of the Unified Database
6 Aviation demand and air traffic model (ADEM)
6.1 Introduction
6.2 Computing the demand for air transport - the main principles
6.3 Response of demand to changes in cost through changes in fares
6.4 Treatment of policy measures in ADEM
6.5 Discriminatory and localised measures
7. Aviation cost model (ACOS)
7.1 Introduction
7.2 Structure of aircraft operating costs in ACOS
7.3 Processing of policy measures
8 Direct economic impacts model (DECI)
8.1 Introduction
8.2 Computations performed in DECI
8.3 DECI results
9 Macro-economic impact model for the Netherlands (MECI)
9.1 Computations performed by MECI
9.2 MECI results
10. Flights and emissions model (FLEM)
10.1 Overview of FLEM
10.2 Functionality of FLEM
10.2.1 The flight pre-processing module
10.2.2 The pre-flight fuel calculations module
10.2.3 The flight profile calculation module
10.2.4 The emission index calculation module
10.2.5 The grid calculation module
10.2.6 The military emission module
10.3. Validation and computation results of FLEM
10.3.1 FLEM specific results
10.3.2 A comparison between aircraft types
10.3.3 Comparison with data from Flight Data Recorder
11. Atmospheric impact models
11.1 Introduction and overview
11.2 Other Atmospheric Immissions Model (OATI)
11.2.1 Overview of OATI
11.2.2 OATI input data
11.3 Chemical Tracer Model of KNMI (CTMK)
11.4 Environmental Impact Model (ENVI)
11.4.1 Overview of ENVI
11.4.2 Processing steps in ENVI
11.4.3 ENVI input data and computational results of ENVI
12. Use of the integrated AERO modelling system (AERO-MS)
12.1 Overview of AERO modelling system
12.2 Model computations in Base, Datum and Policy case
12.3 Analysis capabilities of the AERO-MS
12.4 User interface and presentation options of the AERO-MS
12.4.1 Overview of main functions of the user interface
12.4.2 Specification and computation of cases
12.4.3 Analysis and evaluation of cases
List of tables
2.1 Emissions from commercial aircraft engines compared with global emissions from
human sources for the year 1992
2.2 Distribution of global NO
x
emissions from aircraft engines in 1992 by altitude layer
3.1 Generic aircraft types and numbers of aircraft in 1992 and 1997
3.2 Global 1992/1997 passenger and freight volumes
5.1 Validation of the Unified Database for Individual European Cities
6.1 Out-turn passenger fare elasticities by aggregated region pair
6.1 Demand, cost and revenue effects of a fuel levy of 20c per kg for two extreme
profitability criterion (PAF) settings
6.2 Demand, cost and revenue effects of a 50% tax credit measure for new aircraft of
latest technology for different profitability criterion settings
7.1 Cost components by physical unit
7.2 Dimensions of cost components
7.3 Variable cost components (1992 US$) of long haul aircraft between 300 and 499
seats between EU and North America
7.4 Impact of 10 US$ cents per kg fuel levy on unit costs (1992 US$) for long haul 300 to
499 seat aircraft
8.1 Selective DECI results for aggregated region pairs for base year 1992
8.2 Selective DECI results for aggregated region pairs for base year 1997
8.3 Selective DECI results for aggregated carrier regions for base year 1992
8.4 Selective DECI results for aggregated carrier regions for base year 1997
9.1 MECI results for 1992 and 1997
10.1 Fuel use and emissions comparison
12.1 Overview of AERO computations in Base, Datum and Policy case
12.2 Examples of assumption variables specifying the Base case
12.3 Variables to be potentially included in AERO scenario specification (Datum case)
List of figures
2.1 Contribution of changes in ozone content to global warming
2.2 Contribution of aircraft engine emissions of NO
x
to NO
x
concentrations at cruise level
(max 73%) in July 1997
2.3 Contribution of aircraft engine emissions of NO
x
to ozone concentrations at cruise
level (max 6%) in July 1997
2.3 Contribution of aircraft engine emissions of NO
x
to O
3
concentrations at cruise level
(max 3.5%) based on the situation of 1997
2.4 Vertical distribution of NO
x
emissions from aircraft engines
3.1 Overview of computational steps in AERO Modelling System
3.2 Effects of scenarios and measures
3.3 Computations in future situation without measures
3.4 Computations in future situation with measures
3.5 Location of IATA regions
4.1 Views considered in ATEC on fleet development in time
4.2 Aircraft retirement function used in ATEC
4.3 Scenario development of emission indices
4.4 Changes in aircraft new prices following imposed technology changes
4.5 ATEC prediction of future fleet characteristics
5.1 An illustration of Unified Database records
6.1 Structure of Unified Database manipulation by ADEM
6.2 Air traffic growth as a function of economic growth
6.3 Effect of surface competition on typical fare elasticities
6.4 Scheduled passenger flights of between 250 and 500 km
6.5 Scheduled passenger flights of between 500 and 1000 km
6.6 Linkages between the ADEM and ACOS models
6.7 The problem faced by the Cost to Fares mechanism for ADEM
7.1 Base cost composition for 300 to 499 seat long haul older technology aircraft for
route from Amsterdam to New York
7.2 Base cost composition for 300 to 499 seat long haul current technology aircraft for
route from Amsterdam to New York
8.1 DECI model structure
9.1 MECI model structure
10.1 FLEM computation scheme
10.2 FLEM flight stages
10.3 Typical short range flight profile
10.4 Typical long range flight profile
10-5 NO
x
emission model: Boeing-2 versus NLR P3T3 method
10.7 Emission indices by altitude layer
10.8 Global distribution of aircraft engine CO
2
emissions in 1992
10.9 Global distribution of aircraft engine CO
2
emissions in 2010
10.10 Distribution of flying time for small and large aircraft by altitude layer
10.11 FDR and FLEM fuel use comparison for B737-400 flights
10.12 FDR and FLEM flight time comparison for B737-400 flights
10.13 FDR and FLEM flight profile comparison for B747-400 flight
11.1 Atmospheric models in context
11.2 OATI computation scheme
12.1 Overview of AERO Modelling System
12.2 Main screen of the AERO-MS interface
12.3 Policy wizard, specification of a policy case
12.4 Run wizard: overview of models to include in a run
12.5 Graphical presentation of results in the AERO-MS
12.6 Presentation of cost-effectiveness of policies in the AERO-MS
1
1 Context and aims of AERO project
1.1 Background and objectives
In 1994, the Dutch Civil Aviation Authority started a major policy analysis called the AERO-
project (Aviation emissions and Evaluation of Reduction Options). The objectives of this project
were to assess the problems related to air pollution from aircraft engine emissions and to analyse
possible measures to reduce the impacts of air transport on the atmosphere, taking in account
the environmental benefits and the economic impacts of such measures. In order to achieve
these objectives, an extensive global information and modelling system was developed which is
referred to as the AERO modelling system (AERO-MS).
The AERO-MS covers a sequence of steps from the description and generation of air transport
demand to the assessment of the environmental and economic impacts of aircraft engine
emissions, providing a comprehensive integration of the relevant economic, commercial,
technological and environmental forces. In essence, the AERO-MS is a policy-testing tool to
evaluate the environmental and economic consequences of responses to emission-related
measures within the context of relevant future developments in the air transport sector.
Potentially, a great many possible measures and different future developments are relevant.
Consequently, the AERO-MS had to be capable of analysing a wide range of measures
(including economic, regulatory, technical and operational measures) within a variety of
autonomous (economic and technological) developments. The AERO-MS was therefore
designed to meet the following analysis requirements:
to provide an adequate description of the economic and environmental aspects of the air
transport system (in particular the extent and effects of aircraft engine emissions);
to adequately reflect the economic and technological developments in air transport; and
to assess the effects of a range of possible measures to reduce the environmental impact of
air transport, taking into account the responses of the major actors (airlines, consumers,
manufacturers) to such measures.
The design philosophy and architecture underlying the AERO-MS allow the user a large degree
of flexibility in analysing the effects of specific developments and measures in a "what-if" fashion.
This was implemented by creating a great many user options to change key assumptions,
schematisation aspects, (scenario) developments and possible measures (policy options).
The AERO project and the AERO-MS were principally developed to analyse the impacts of
aircraft engine emissions on a global scale and therefore provide a complete description of
world wide aviation activity. In addition, within the analysis capabilities of the modelling
system, specific options were provided to allow for a more detailed analysis of the aviation
sector in the European context (the EU-countries as of the end of the 20
th
century). Within the
European context, a number of specific options are provided for the analysis of the
Netherlands' aviation (in particular involving the aviation activity at Amsterdam airport). These
different spatial levels of analysis all reflect the areas of interest to the Dutch Civil Aviation
Authority, which as noted already - commissioned the AERO project.
1.2 Applications of the AERO modelling system
From the very start of the AERO project, activity has strongly focused on developing the
AERO-MS as a comprehensive tool for analysing the complex environmental and economic
effects of policy measures in different scenarios. This was ultimately achieved in five project
phases, though the system has been actively applied since the second phase.
From the second phase of the project onwards, versions of the AERO-MS have been
available for analysis. In subsequent phases, the modelling system was further expanded,
updated and improved. Based on earlier versions of the AERO-MS a number of rather
substantial analyses were carried out. In addition to the intermediate analyses directly carried
out for the Dutch Civil Aviation Authority, these include:
2
a global analysis of emission charges and taxes for the Focal Point on Charges (carried
out for CAEP/4) (FPC, 1998);
an analysis of the impacts of fuel taxation in the European context carried out for the
European Commission (Resource Analysis et al, 1999);
a study commissioned by the Dutch Civil Aviation Authority to facilitate the debate on the
national allocation of CO
2
between the Subsidiary Body for Scientific and Technological
Advice (SBSTA) of the UN Framework Convention on Climate Change (UNFCCC) and
ICAO's Committee on Environmental Protection (CAEP) (Resource Analysis / CE, 2000);
an analysis of market-based options for the reduction of CO
2
emissions from aviation for
the Forecast and Economic Support Group of ICAO (CAEP/5) (Pulles et al, 2000).
As described in the present documentation, the now-available version of the AERO-MS
provides a powerful and flexible tool to support the analysis of economic and environmental
(atmospheric) impacts on the aviation sector arising from a wide variety of possible
developments and measures.
1.3 Parties involved in project execution
Assigned by the Dutch Civil Aviation Authority (RLD), the AERO-MS has been developed in
association with the RLD by a core consortium of three parties:
Resource Analsyis of Delft, the Netherlands;
MVA Limited (MVA) of London, United Kingdom;
The Netherlands National Aerospace Laboratory (NLR) of Amsterdam, the Netherlands.
Important contributions to the AERO project were made by other parties, such as the Royal
Netherlands Meteorological Institute (KNMI) and the Netherlands National Institute of Public
Health and Environment and Protection (RIVM). KNMI was involved with the modelling of the
dispersion and impacts of the relevant emissions of aircraft engines and other (ground)
sources to the atmosphere. RIVM has provided the global inventory of emissions from ground
sources (human activities and selective natural sources) which was used in the atmospheric
impact assessment, together with the aircraft engine emissions.
In addition, a great many other parties have assisted in providing the required information for
the operationalisation, calibration and verification of the AERO-MS. Among the most
important suppliers of data and information are: the Royal Dutch Airline (KLM); the
International Civil Aviation Organisation (ICAO); the International Air Transport Association
(IATA); the Intergovernmental Panel on Climate Change (IPCC); and the Abatement of
Nuisance Caused by Air Traffic (ANCAT).
1.4 Set-up of the overall AERO documentation
The complete documentation of the AERO project includes the following parts:
A) Main Report
Part I: Description of the AERO modelling system
Part II: Analysis preparation, execution and results
B) Reports on individual models
General reports
- Aircraft Technology Model (ATEC)
- Air Transport Demand and Traffic Model (ADEM)
- Aviation Operating Cost Model (ACOS)
- Economic impact models: Direct Economic Impact Model (DECI) and Macro-
Economic Impact Model for the Netherlands (MECI)
- Flights and Emissions Model (FLEM)
3
- Atmospheric impact models: Other Atmospheric Immissions Model (OATI), Chemical
Tracer Model KNMI (CTMK), Environmental Impact Model (ENVI).
System documentation
- Aircraft Technology Model (ATEC)
- Air Transport Demand and Traffic Model (ADEM)
- Aviation Operating Cost Model (ACOS)
- Economic impact models: Direct Economic Impact Model (DECI) and Macro-
Economic Impact Model for the Netherlands (MECI)
- Flights and Emissions Model (FLEM)
- Atmospheric impact models: Other Atmospheric Immissions Model (OATI), Chemical
Tracer Model KNMI (CTMK), Environmental Impact Model (ENVI).
C) AERO Modelling System
Technical report
User manual
Scenario and policy variables
D) CD-Rom of AERO project
The CD-Rom of the AERO project is split into two parts. The first part contains an electronic
version of selective AERO documentation. The electronic documentation relates to both a
description of the AERO-MS and the main results of the AERO analysis. The second part is
involved with a demonstration version of the AERO-MS. It provides an insight in the
functionality and use of the AERO-MS.
1.5 Purpose and overview of Main Report
The Main Report describes the final version of the overall AERO-MS as it was completed
during the final phase of the AERO project, together with the integrated analysis which was
carried out in this project phase.
The Main Report is divided into two parts (Part I and Part II). Part I (Description of the AERO
Modelling System) provides the overall description of the methodology underlying the AERO-
MS, the basic features of the models included in the AERO-MS and the main implementation
aspects and use options of the integrated modelling system as a whole. Part II is involved
with the preparation, execution and results of the integrated analysis which were carried out
with the AERO-MS.
The document at hand includes Parts I and II of the Main Report. Chapter 2 of Part I is
involved with an overview of the problems related to aircraft engine emissions. In chapter 3 an
overview is provided of the approach followed in the AERO-MS . Subsequent chapters 4
through 11 are involved with an overview of the individual models which are integrated in the
AERO-MS. These chapters focus on the functionality, main mechanisms, inputs and
validation of the individual models. Also the role of the various individual models in the overall
modelling system is specified in these chapters. Finally chapter 12 describes the actual use of
the integrated AERO modelling system, including a description of analysis capabilities and the
user interface.
The first chapters of Part II of the Main report describe the calibration of the AERO modelling
system (chapter 13); the development of scenarios describing the future context (chapter 14);
and the specification of the analysis set-up (chapter 15). Further chapters are involved with
the description of analysis results, including the problem analysis (chapter 16); the analysis of
global individual measures (chapter 17); the analysis of specific individual measures (chapter
18); the analysis of side effects (chapter 19); the analysis of measure combinations (chapter
20); and selective sensitivity analyses (chapter 21). The final chapter 22 provides a summary
of analysis results and an overview of main conclusions.
Note: check with final set-up of Part II
4
References
FPC, 1998. Emission charges and taxes in aviation. Report of the Focal point on Charges
prepared for CAEP/4, The Hague, March 1998.
Resource Analysis and Centre for Energy Conservation and Environmental Technology (CE),
2000. National allocation of international aviation and marine CO
2
-emissions, RA/00-
429, Delft, September 2000.
Resource Analysis et al, 1999. Analysis of the taxation of aircraft fuel (VII/C/4-33/97).
Produced for European Commission by a consortium of Resource Analysis, MVA
Limited, Dutch National Aerospace Laboratory and International Institute of Air and
Space Law, Delft, January 1999.
Pulles et al, 2000. Analysis of Market_Based Options for the reduction of CO
2
emissions from
aviation with the AERO modelling system. Study for Forecast and Economic Support
Group (FESG) (CAEP/4) by Pulles, J.W. (Dutch Civil Aviation Authority); Velzen, A. van
(Resource Analysis; Baarse, G. (BB&C); Hancox, R. and others (MVA Limited),
November 2000.
5
2. Problems related to aircraft engine emissions
2.1 Nature, extent and development of aircraft engine emissions
Aviation has a number of impacts on the environment, including noise at and around airports;
local air quality in the vicinity of airports; and the contribution of aircraft engine emissions to
global problems such as climate change (global warming) and depletion of ozone in the
stratosphere, leading to an increase in UV-radiation. In addition, the risks for population
around airports (external safety) could also be considered an environmental impact.
The AERO-MS was specifically designed to consider the environmental impacts of global
aircraft engine emissions at cruise level. While focusing on the global perspective, in view of
the spatial detail considered, the modelling system is also able to assess the impacts of
emissions on a regional and local scale. Moreover, since the modelling system includes a
detailed description of the technical and economic features of air transport demand and
supply, it potentially provides an important basis for assessing other environmental impacts.
Nevertheless, the present applications of the AERO-MS are focused on global aircraft engine
emissions and their related problems.
Burning jet fuel (kerosene) in a turbofan or turboprop engine causes several waste products
to be released into the atmosphere. The most important emission types to be considered are:
carbon dioxide or CO
2
( 3.17 kg per kg fuel burnt); water (vapour) or H
2
O ( 1.25 kg per kg
fuel); nitrogen oxides or NO
x
; carbon monoxide or CO; hydrocarbons or C
x
H
y
; sulfur dioxide or
SO
2
; and soot. CO
2
and H
2
O are the products of complete burning of which fixed amounts are
formed with the burning of each kg of fuel. NO
x
is formed under high pressure and
temperature in the combustion chamber, the amounts of NO
x
produced depending on engine
working conditions and thrust setting. Substances resulting from incomplete burning, such as
CO, C
x
H
y
and soot are mainly produced when the engines are not operating at optimal
conditions, which particularly occurs during landing, taxiing, take-off and climb-out. Soot may
have some local effects at airports as it contains various kinds of PAC's (polycyclic aromatic
hydrocarbons), some of which are carcinogenic. SO
2
is produced as a consequence of the
small sulphur content in kerosene.
Table 2.1 provides an overview of global aircraft engine emissions in 1992 for the most
relevant substances, in comparison with total global emissions from human (and some
natural) sources.
Table 2.1 Emissions from commercial aircraft engines compared with global
emissions from human sources for the year 1992.
Substances Units Total emissions
from human sources
Aircraft engine
emissions
Contribution of
aircraft engines (%)
CO
2
10
6
ton 29,110 424 1.45%
NO
x
10
6
ton 101.91
1
1.69 1.66%
SO
x
10
6
ton 150.68 0.12 0.08%
C
x
H
y
10
6
ton 321.36 0.42 0.13%
CO 10
6
ton 973.45 1.01 0.10%
Source: EDGAR database (RIVM) and computations with AERO-modelling system
From table 2.1 it can be observed that relative contributions of aircraft engines to the global
emissions of CO
2
and NO
x
are by far the most substantial. This is particularly important, since
both substances contribute to global warming potential (climate change), while NO
x
emitted at
higher altitudes may also contribute to ozone depletion. Although these relative amounts of
the order of 1.5% may seem small, aviation belongs to the larger single activity contributors,

1
Includes about 16 Kton due to lightning
6
with a growing contribution of aviation in both relative and absolute terms. More importantly,
with respect to the possible effects of these emissions at higher altitudes it should be
considered that aircraft engines are the only anthropogenic source in the upper atmosphere.
Contributions of aircraft engines to the global emissions of SO
x
, C
x
H
y
and CO are rather
small. However these emissions may still cause local air quality problems around airports,
while SO
x
emitted at high altitude may also contribute to changes in global warming potential
and ozone depletion.
Over the years, relative to the growth in capacity offered, most emissions of aircraft engines
have reduced. In addition to the improvement of the overall fuel efficiency, especially the
emission indices of CO, C
x
H
y
and soot (emissions per kg of fuel burnt) have decreased
substantially. Since the increased fuel efficiencies were obtained by increasing the
temperature and pressure in the combustion chamber, the decrease in the NO
x
-emission
index was less spectacular. However, since 1988, increasing NO
x
stringencies have been
introduced by the International Civil Aviation Organisation (ICAO). Last strengthened in 1998,
these stringencies are expected to further reduce the future NO
x
emission index.
In spite of these technical improvements, due to the rather strong growth in actual and
expected air transport demand, aircraft engine emissions are expected to grow substantially
in the coming years, both in absolute terms and also relative to other sources of emissions.
Based on projected estimates following from the AERO-MS, the relative contributions from
aircraft engine emissions are expected to increase up to 2.1% of total emissions for CO
2
,

and
2.7% for NO
x
, in the year 2015. In the special IPCC report on aviation (IPCC, 1999), the
relative contribution from aviation with respect to all atmospheric impacts are determined
(CO
2
, NO
x
, H
2
O contrails, etc.). This report calculates a relative contribution from aviation to
global warming potential of 3.5% in 1992, which would increase to 5% in 2050.
2.2 Environmental problems related to aircraft engine emissions
The most important (global) problems to which aircraft engine emissions contribute are the
increase of global warming potential (inducing climate change) and the increase of UV-
radiation. In addition, aircraft engine emissions contribute to (local) air pollution.
Climate change
Aircraft engine emissions-related contributions to climate change may include:
direct contribution to global warming potential (through radiative forcing) of greenhouse
gases, in particular CO
2
;
the effects of NO
x
on radiative forcing by increasing tropospheric ozone and by decreasing
stratospheric ozone and tropospheric methane;
the effect on radiative forcing of water emissions, aerosol and particulate emissions of SO
x
and soot at high altitudes.
CO
2
is generally considered to be by far the most important greenhouse gas, contributing
more than 60% to the radiative forcing effect compared with pre-industrial values (IPCC,
1995). In view of the long lifetime of CO
2
(decades to centuries) it can be assumed that CO
2
is
well mixed in the global atmosphere, regardless of how and where CO
2
emissions take place.
Hence, the relative contribution of aircraft engine emissions to the climate change problem
directly follows from the relative contribution to CO
2
emissions.
The contribution of NO
x
to radiative forcing mainly follows from an increase in the
tropospheric ozone content, caused by NO
x
emissions. The troposphere is the layer of the
atmosphere between the earths surface and the stratosphere. It is the atmospheric layer
where temperature generally decreases with height, and where clouds and weather
phenomena occur. Its height varies between about 10 km at high latitudes and 18 km in
tropical areas, depending on temperature and pressure variability. The stratosphere is the
stratified and stable region of the atmosphere above the troposphere extending to about 50
7
km. The tropopause is the boundary between troposphere and stratosphere, usually
characterised by a rather strong vertical temperature gradient. Aviation activity at cruise levels
(typically between 9 and 12 km altitude) takes place in the upper parts of the troposphere and
(locally) in the lower parts of the stratosphere.
The contribution of aircraft engine emissions of NO
x
to radiative forcing by tropospheric ozone
formation may be of the same order as the aircraft engine emissions of CO
2
, although
uncertainties related to this assessment are high (Friedl et al., 1997). The potentially large
contribution is due to the sensitivity of radiative forcing to the ozone content at altitude bands
between 8 and 12 km, where most of the aviation activity takes place. This is illustrated in
figure 2.1, which shows the temperature change (in C) at the surface of the earth following
from a marginal increase in the ozone content of 1 Dobson unit. A Dobson unit corresponds
with the amount of ozone contained in a gaseous layer of 0.01 mm under a pressure of 1
atmosphere and a temperature of 273 K. Typical values of the amount of ozone contained in
the entire vertical atmospheric column are of the order several hundreds of Dobson units.
Figure 2.1 Contribution of changes in ozone content to global warming
From the graph in figure 2.1 it can be observed that the relative effect of an increase in the
ozone content on global warming is largest at cruising altitudes of subsonic aircraft. A
potential effect of NO
x
emissions at high altitudes is the decrease of stratospheric ozone.
From figure 2.1 it also follows that an increase in the ozone content at stratospheric levels
would effectively lead to a cooling effect (and consequently a decrease in stratospheric ozone
would also contribute to global warming). However, this effect would only occur in the upper
stratosphere, while NO
x
stratospheric emissions of aircraft engines - if at all - would merely
take place in the lower stratosphere.
In addition to increasing tropospheric ozone concentrations, aircraft NO
x
emissions are
expected to decrease methane concentrations. Since methane is also a greenhouse gas, this
will reduce the radiative forcing. This effect is expected to be smaller than the increase in
radiative forcing through ozone formation, however (IPCC, 1999).
Figure 2.2 provides an overview of the relative contribution of aircraft engine NO
x
emissions
to NO
x
concentrations at cruise level (around 35,000 feet). Figure 2.3 shows the relative
contribution to ozone concentrations at cruise level following from these aircraft engine NO
x
emissions.
Temperature change at earths surface
in C for each Dobson-unit
ozone increase
0
10
20
30
40
50
60
-0,01 0 0,01 0,02
Temprature change in C
H
e
i
g
h
t

i
n

k
m
8
0 max = 73%
Figure 2.2 Contribution of aircraft engine emissions of NO
x
to NO
x
concentrations at
cruise level (max 73%) in July 1997
0 max = 6%
Figure 2.3 Contribution of aircraft engine emissions of NO
x
to ozone concentrations at
cruise level (max 6%) in July 1997
9
From figure 2.2 it can be observed that aircraft engine emissions may contribute more than
70% to NO
x
concentrations in the Atlantic corridor (with remaining contributions coming from
lightning and emissions from ground sources). The related maximum contributions to ozone
concentrations, which particularly occur in the northern parts of the northern hemisphere,
amount to around 6% (figure 2.3).
Studies of the effects of water, including contrails and cirrus clouds at high altitudes have only
recently begun. Uncertainty is especially large with respect to the effects of cirrus clouds, but
the effect on radiative forcing of water vapour and contrails may in fact be of the same order
as the aircraft engine emissions of CO
2
, or even larger. The direct radiative forcing effects of
sulfate and soot aerosols are expected to be small compared with other aircraft engine
emissions, where sulfate tends to decrease, and soot tends to increase the warming potential.
Indirectly, these aerosols influence the formation of clouds and may change their radiative
properties, but as yet, these effects are basically unknown (IPCC, 1999).
UV radiation
Ozone, most of which resides in the stratosphere, provides a shield against the harmful
effects of solar UV-radiation. Hence, the preservation of the stratospheric ozone layer is of
vital importance. There is an ongoing discussion about the potential effects of aircraft engine
emissions in terms of NO
x
, aerosol and particulate emissions of SO
x
,

and soot, on
stratospheric ozone depletion. Earlier speculations (since the 1970s) about potentially very
significant contributions of aircraft engine emissions to ozone depletion, especially if
supersonic aircraft would be introduced, have raised serious concern. Based on the increased
understanding of the role of heterogeneous reactions and transport processes, it is currently
believed that subsonic aircraft flying in the upper troposphere do not affect stratospheric
ozone levels. Yet, given the fact that subsonic aircraft (locally) enter into the lower
stratosphere they are still of concern (NASA, 1996).
(Local) air pollution
The most important pollution problems experienced at ground level to which aircraft engine
emissions contribute are smog formation (due to emissions of NO
x
and C
x
H
y
) and acidification
(due to emissions of NO
x
and SO
x
). Smog formation is strongly enhanced by ozone
production, for which NO
x
and volatile organic compounds (VOC - C
x
H
y
) are important
precursor substances.
Potential aircraft engine contributions to air pollution mostly follow from the emission of NO
x
and to a much smaller extent of C
x
H
y
and SO
x
. Obviously, the contribution to concentrations
experienced at ground level clearly depend on location and altitude of the emission. Table 2.2
shows the relative and cumulative distribution of NO
x
emissions from aviation activities in
1992 across a range of altitude layers computed by the AERO-MS, while the relative
distribution is also shown graphically in figure 2.4. From table 2.2 it can be observed that, for
example, less than 30% of aircraft engine NO
x
emissions take place below 5000 m. In view of
the rather short NO
x
lifetime (of the order of days to weeks) it can therefore be concluded that
only a limited part of the aircraft engine NO
x
emissions effectively contributes to ground level
NO
x
and ozone concentrations. Similar observations hold for the other substances which may
contribute to air pollution problems at ground level.
10
Table 2.2 Distribution of global NO
x
emissions from aircraft engines in 1992 by
altitude layer (%)
Altitude band Relative distribution of NO
x
emissions
By altitude band Cumulative across altitude
bands
0 - 1000 m 13.71% 13.71%
1000 2000 m 3.28% 16.99%
2000 3000 m 2.73% 19.72%
3000 4000 m 2.32% 22.04%
4000 5000 m 2.05% 24.09%
5000 6000 m 2.08% 26.17%
6000 7000 m 2.22% 28.39%
7000 - 8000 m 2.29% 30.68%
8000 - 9000 m 3.85% 34.53%
9000 - 10000 m 13.62% 48.15%
10000 - 11000 m 38.03% 86.18%
11000 - 12000 m 13.48% 99.66%
> 12000 m 0.34% 100.00%
Source: Computations with the AERO modelling system
Figure 2.4 Vertical distribution of NO
x
emissions from aircraft engines
0% 5% 10% 15% 20% 25% 30% 35% 40%
0 - 1000 m
1000 2000 m
2000 3000 m
3000 4000 m
4000 5000 m
5000 6000 m
6000 7000 m
7000 - 8000 m
8000 - 9000 m
9000 - 10000 m
10000 - 11000 m
11000 - 12000 m
> 12000 m
A
l
t
i
t
u
d
e

l
a
y
e
r

(
m
)
Percentage of total No
x
emissions
11
References
Friedl, R. et al., 1997. Atmospheric Effects of Subsonic Aircraft: Interim Assessment Report of
the Advanced Subsonic Technology Program, NASA Reference Publication 1400, May
1997.
IPCC, 1995. The Second Assessment Report. Summary for Policymakers: The Science of
Climate Change, IPCC Working Group I (1995).
IPCC, 1999. Aviation and the Global Atmosphere, IPCC (WMO and UNEP), Cambridge
University Press (1999). ISBN 052166404.
NASA, 1996. Global Atmospheric Effects of Aviation, Report of the Proceedings of the
Symposium, Virginia Beach, Virginia, USA, 15-19 April 1996.
12
3. Overview of AERO modelling approach
3.1 Computational steps
The AERO-MS includes a sequence of logical steps from the description and generation of air
transport demand to the assessment of the environmental and economic impacts of aircraft
engine emissions. These steps cover the following computations:
1. Aircraft technology and fleet build-up
2. Air transport demand (passengers and freight), supply (capacity offered) and aircraft
flights.
3. Costs of air transport
4. Revenues of air transport
5. Direct economic effects of air transport
6. Aircraft flight paths, fuel-use and emissions
7. Atmospheric emissions from ground surface sources
8. Atmospheric concentrations of key substances and related environmental effects.
The result from the computation of one step feeds into another and this logical sequence defines
the relationships between them. A graphical overview of these steps is provided in figure 3.1.
Figure 3.1 Overview of computational steps in AERO Modelling System
The following briefly describes the computational steps shown in figure 3.1.
1. Aircraft technology particularly applies to the fuel-use and emission characteristics of
different aircraft types. The fuel-use and emission characteristics of these aircraft types are
expressed as a function of the technology age of the aircraft type (following from the year in
which the aircraft type is certified), which defines the technology 'level'. The technology age
1. Aircraft
technology and
fleet build-up
2b. Air transport
supply (capacity)
2a. Air transport
demand (pax and
freight, fares and
freight rates)
2c. Flights by
aircraft type and
technology
6. Aircraft flight
paths, fuel use and
emissions
7. Emissions
from ground
sources
8. Atmospheric
concentrations and
environmental effects
3. Costs of air
transport
4. Revenues of
air transport
5. Direct economic effects
Airlines:
- operating costs, revenues, results
- employment
- contribution to gross value added
Government income
Change in consumer surplus
and expenses
13
distribution is determined by the fleet build-up which depends on the development in time of
aircraft sales (following air transport demand) and aircraft retirement.
2. The AERO-MS explicitly considers, and matches, the supply and demand side of air
transport, i.e. the air transport demand in terms of passengers and freight and the frequency
and capacity of air transport services offered. Passenger fares and freight rates are
determined in the balancing of supply and demand. The number of aircraft flights then
follows from the volumes of passengers and freight transported, corresponding with the
balance of supply and demand. In addition to the geographical component (origin-
destination), aircraft flights are expressed in terms of aircraft types and technology levels, in
accordance with available fleets.
3. Aircraft operating costs are associated with flights by aircraft type and technology level.
Among the variable (flight-related) cost components considered are: fuel costs, route
charges, airport charges, flight and cabin crew costs, maintenance costs, capital
(depreciation) and finance costs. In addition, total operating costs include a number of
other, volume-related, costs such as the costs of ground-handling, sales, ground facilities
(buildings) and overhead.
4. Revenues of air transport directly follow from the numbers of passengers and freight
transported and passenger fares and freight rates applied.
5. The direct economic effects of air transport include the effects on airlines in terms of
operating costs, revenues and results, and a number of macro-economic effects such as
airline-related employment and contributions to gross value added. In addition, the effects
of imposed measures on some other actors involved in air transport are determined,
including the 'government' (changes in government income) and air transport 'consumers'
(changes in consumer expenses and consumer surplus). Effects on manufacturers follow
from changes in required fleet capacity and related aircraft purchases.
6. Fuel-use and emissions for each flight are computed in three-dimensional space based
on a specification of flight paths, taking into account the geographical flight specification
and the technical characteristics by aircraft type and technology level. The emissions
considered include CO
2
, NO
x
, SO
2
, C
x
H
y
, CO and H
2
O. In addition, the amount of oxygen
consumed is calculated.
7. The effects of aircraft engine emissions are considered within the perspective of
emissions from ground sources related to other human activities. For this reason, a global
inventory is made of the relevant emissions from these other sources.
8. Effects of aircraft engine emissions on atmospheric concentrations are determined for a
limited number of key substances including CO
2
, NO
x
and ozone. The impacts of
(changes in) atmospheric concentrations are assessed in terms of effective UV doses
and changes in radiative forcing (global warming potential).
3.2 General modelling principles and modelling dynamics
3.2.1 Future developments with and without measures
The objective of the AERO-MS is specifically to quantify the economic and environmental
impacts of possible measures related to the air transport system, under different future
developments. For this purpose the modelling approach very clearly distinguishes between
the following three situations (see figure 3.2).
1. A Base situation (or Base case), providing the best-possible description of today's world
(the existing situation).
2. A user-defined specification of a future situation, without specific measures taken in the
air transport sector (the Datum case).
14
3. The user-defined future situation, with specific measures taken in the air transport sector
(the Policy case).
The second (Datum) case may involve a variety of future situations, reflecting different sets of
assumptions related to autonomous economic and technological developments, providing
alternative future contexts within which the air transport system operates. These alternative
future contexts without measures are referred to as scenarios. In this respect it should be
noted that in other studies the word scenario may be used in a different meaning. For
example in CAEP, the word scenario may be used to indicate a future situation with
measures, which according to the definitions used in the present analysis is referred to as a
Policy. The effects of different scenarios follow from a comparison of the alternative future
contexts with the Base situation (situations 2 with 1).
Within a specified future, the effects of an air transport measure or set of measures can be
tested by comparing situations 3 and 2 (i.e. the Policy case versus the Datum case). This can
be repeated for different measures or sets of measures, allowing for a comparison across
different measures within the same scenario context. Figure 3.2 illustrates these principles.
Figure 3.2 Effects of scenarios and measures
The computational steps in the AERO-MS (see figure 3.1) apply to all three situations shown
in figure 3.2, although the meaning and purpose of some of these computations varies
between the situations.
The computations in the Base situation provide a coherent description of the existing air
transport system in terms of passenger and freight transported; aircraft flights; airline
operating costs, revenues and results; fuel-use and emissions; and related environmental
effects (see figure 3.1). These results take into account all relevant details according to the
various dimensions and levels of geographical detail considered in AERO (see section 3.3).
The primary base year used in the AERO-MS is 1992, for which a full set of data was
available at the start of the AERO project. An extensive processing of raw data sources has
taken place to create the basic data for 1992. In order to reflect the actual (known)
developments after 1992, an update of the base year 1992 was made for the most recent
year for which data have become available (presently 1997).This provides a 'secondary',
more recent base year to be used in the AERO analysis.
1. Base situation
2. Future (I) without
measures
Future (II)
Future (III)
3. Future (I) With
measures (a)
Future (I), measures (b)
Future (I), measures (c)
Effects of
scenarios
Effects of measures
within a specified
scenario
15
The future situation without measures (the Datum case), is driven by the effects of
autonomous macro-economic, demographic and technological developments, as shown in
figure 3.3. The effects of such developments become manifest in changes in the fuel-use and
emission characteristics and purchase prices of aircraft; air transport demand (passenger and
freight), fares and freight rates; and aircraft flights by flight stage, aircraft type and technology
level. Consequently, the computation of all economic and environmental effects needs to be
updated. These computations take essentially the same form as for the Base situation.
Figure 3.3 Computations in future situation without measures
The future situation with measures (the Policy case) is reflected in figure 3.4. The Policy case
is based on the scenario situation as in figure 3.3 with the addition of the one or more
measures (or policies) to be evaluated. The comparison of the situations shown in figures 3.3
and 3.4 represents the essence of the AERO approach, where two distinct lines of future
development from the Base position are considered: one without measures and the other
with.
A variety of measures can be introduced, including regulatory, economic, operational and
other measures. An important feature of these measures is that they generally lead to an
increase in the cost of air transport. For example, regulatory measures to be tested, such as
operation restrictions, may affect aircraft technology (fuel-use and emission characteristics)
and aircraft purchase prices, which in turn will affect the available fleet capacity and the
computation of aircraft operating costs. Economic measures, such as fuel, ticket, airport and
route charges, have a direct impact on one or more of the cost components considered in the
computation of aircraft operating costs.
Macro-economic,
demographic and
transport market
development
Tranport market
and technological
develeopment
Technological
development
Aircraft
technology and
fleet build-up
Air transport
supply (capacity)
Air transport
demand
(pax/ freight)
Flights by aircraft
type and technology
Aircraft flight
paths, fuel use
and emissions
Emissions
from ground
sources
Atmospheric
concentrations and
environmental effects
Costs of air
transport
Revenues of
air transport
Pax fares.
Freight rates
Airlines:
- operating costs, revenues, results
- employment
- contribution to gross value added
Direct economic effects
Government income
Change in consumer surplus and
expenses
Scenario
16
Increases in the operating cost of air transport imposed by measures generally lead to
changes in fares, and therefore to changes in the growth of air transport demand. The AERO-
MS takes account of these measure-induced changes in costs, fares and demands, and
updates the computation of all economic and environmental effects. In the situation with
measures, in addition to the effects on airlines, the direct economic effects also include a
number of specific effects of the measures imposed on other actors, such as changes in
government income and changes in consumer surplus and consumer expenditure.
Figure 3.4 Computations in future situation with measures
3.2.2 Modelling dynamics and time aspects
The AERO-MS is first and foremost a tool for testing and comparing alternative policy options
and, therefore, its objective is to quantify the effects of policy options relative to the situation
without policy measures (the scenario). To achieve this it focuses directly on estimating the
differences in (especially) aircraft operating costs between the Policy case and the Datum
case, from which it infers the corresponding differences in airline and consumer behaviour,
and ultimately in aircraft emissions.
Thus, the fundamental question addressed by AERO is: Given a scenario and a policy option
to be tested, in what ways and to what extent would the results in the Policy case differ from
the Datum case? The model can answer this question for any scenario in any (forecasting)
year specified by the model-user. For a given scenario in the users selected year, therefore,
the effects of alternative policy options can be quantified in relation to a common benchmark.
This produces a snapshot of each policy option (i.e. with measures) against the same
scenario in the same year, which allows the model-user to carry out a comparative evaluation
of the policy options on a completely consistent basis.
Macro-economic,
demographic and
transport market
development
Tranport market
and technological
develeopment
Technological
development
Aircraft
technology and
fleet build-up
Air transport
supply (capacity)
Air transport
demand
(pax/ freight)
Flights by aircraft
type and technology
Aircraft flight
paths, fuel use
and emissions
Emissions
from ground
sources
Atmospheric
concentrations and
environmental effects
Costs of air
transport
Revenues of
air transport
Regulatory
measures
Operational
and other
measures
Economic
and other
measures Pax fares.
Freight rates
Airlines:
- operating costs, revenues, results
- employment
- contribution to gross value added
Direct economic effects
Government income
Change in consumer surplus and
expenses
Scenario
Measure
17
In order to inform decisions on the long-term comparative merits and costs of different policy
options, it is preferable to consider the effects of measures when they have matured: that is,
when the adaptations of carriers, consumers and other affected parties to the measures can
be regarded as reasonably complete. With a time-series model, it is necessary to estimate the
process of adaptation to each measure through time, and the years for which the criterion of
reasonably-complete adaptation is satisfied may be far into the future, by which time the
scenario itself has become highly speculative.
The snapshot approach permits the model-user to compress this time scale. In effect, for the
purpose of comparative evaluation of policy options, the user can assume that the measures
were introduced sufficiently long before the selected forecast year so that, by that year,
reasonable adaptation will have been achieved. The implicit assumption is that a comparison
of measures in their mature state will give rise to essentially the same conclusions (such as
cost-effectiveness ranking) irrespective of the year for which the comparison is made. This
further eliminates the need to carry out the comparison through time, using time-series
models, though it is straightforward to confirm the assumption by taking a further snapshot for
a different (in between) forecasting year.
As noted above, AERO focuses on estimating the differences between the Policy case and
the Datum case. This requires in turn estimates of responsiveness of (for example) the choice
between older and newer technology aircraft to policy-induced differences in operating costs.
The best evidence for this is the observable current mix of aircraft types, and how that is
related to differences in operating costs. The current mix is of course the cumulative outcome
of many acquisition and operating decisions over a long period of time. Thus, if systematic
relationships between fleet mix and costs can be identified from the current situation (which
they can be), these will provide appropriate estimates of the ultimate, mature responsiveness
to differences in costs that are needed for the snapshot approach.
In contrast, time-series models depend upon the mature relationships being identifiable from
time-series data. In an industry for which the historical record is dominated by trend growth,
economic cycles and other short-term factors, it is especially difficult to differentiate the
required (mature, long-term) relationships from the effects of those more dominant influences.
The snapshot approach is not devoid of a temporal dimension, however. For this it depends
on the Datum case scenario. In AERO, the model-user is provided with considerable freedom
to specify his preferred scenario(s), in terms of future global and regional economic growth,
technology development, trends in costs and fares, the competitive environment, etc. These
business-as-usual factors are predominantly exogenous to the industry, and thus are givens
for analysing policy options. The snapshot approach ensures that the estimated effects of
policy options are entirely consistent with these givens. Consequently, the relative benefits
and costs of measures can be explored rapidly and flexibly for different scenarios, such as the
extent to which the competitive environment will permit increased operating costs to be
reflected in higher (compared with the scenario) fares.
It has already been emphasised that the AERO approach estimates the mature outcome of
policy options, to permit consistent comparison of the long-term merits and costs of
alternative measures. Thus, though cases with measures are modelled as differences from a
scenario, it is important to appreciate that they represent the outcome (by the forecast year)
of a different course of events through time compared with the scenario. This reflects that in
reality either the case without measures (scenario) or a case with measures could occur, but
not both, and certainly not a sudden shift from one to the other when a measure is introduced.
Rather, if policy options are applied in reality, there will be a period of transition from the
historical course of development to the mature '"with measures" course of the future. AERO
provides for this. For some policy options (such as certification measures) the model-user is
required to specify an announcement year in addition to the forecast year; the snapshot
taken in the forecast year then takes into account that not necessarily all of the fleet will by
that time have become subject to the measure.
18
For other measures, the model-user can decide whether the time horizon between now and
his selected forecast year is too short for the mature position with measures to have been
reached. The AERO-MS then applies a transition mechanism to acknowledge that adaptation
to the mature position will in reality not be complete by the forecast year. There are two ways
in which it is possible to model airlines not being able to adapt in the short-term. Firstly, it is
possible to show the impact of airlines being faced with additional capacity above that
required and therefore unable to completely reduce some components of cost in line with
supply requirements. These additional costs may then be reflected in terms of higher fares or
reduced airline financial results according to the scenario specification. For example, airlines
might have long-term aircraft leasing agreements or purchase commitments which involve
costs that cannot be avoided.
Secondly, the model assumes that the airlines ability to adapt the mix of aircraft types
operated to minimise their costs following the imposition of a measure is constrained. This
has the effect of increasing airline costs and therefore potentially the short-term impact of the
measure on consumers and airlines. Thus, the AERO-MS can estimate more accurately the
true though transitory - position concerning the effects of a measure in a specified future
year, as well as allowing the user to compare alternative measures in terms of their long-term
mature benefits and costs.
The original base year considered is 1992. Based on projections from the year 1992, any future
year can be simulated. For each future year a separate computation is made. Different
assumptions about specific growth rates (for example passenger demand) can be made for
different time intervals, allowing for the creation of a consistent scenario projection across a
number of subsequent scenario years. In this way, based on the latest available actual
information, an intermediate base year was created (1997) to be used as a reference for
comparison.
3.3 Schematisation aspects
3.3.1 Air transport system specifications
In the AERO-MS, aircraft flights are specified by nine generic aircraft types following from
relevant combinations of range and capacity. Within each aircraft type, two different
technology levels are distinguished, which are defined in terms of technology age. Aircraft for
which technology age is less than or equal to 12 years belong to the group of current aircraft,
while aircraft with a technology age in excess of 12 years are identified as the "older"
technology aircraft. Note that the concept of dividing aircraft by technology age should always
be considered in relation to a particular analysis year. In the AERO analysis, this would
typically be the year for which the forecasts on the effects of future developments (with and
without measures) are required. Table 3.1 provides an overview of the aircraft types
distinguished in the AERO-MS, together with the actual numbers of aircraft determined as
computed by the AERO-MS for the years 1992 and 1997.
The year 1992 is the 'primary' base year considered in the AERO-MS, while 1997 is
considered as a 'secondary' base year, reflecting the latest available information on the
existing situation (see section 3.2).
19
Table 3.1 Generic aircraft types and numbers of aircraft in 1992 and 1997
Nr of aircraft 1992 Nr of aircraft 1997 Aircraft
Type
Range (Km) Seat Band
ta=<12 *) ta>12 ta=<12 ta>12
0 < 4000 < 20 1518 925 1449 1629
1 < 4000 20 to 79 1350 2103 2338 2399
2 < 4000 80 to 124 1398 880 1330 1785
3 < 4000 125 to 179 656 516 617 739
4 4000 to 8000 80 to 124 1312 132 836 869
5 4000 to 8000 125 to 179 682 1101 1007 1224
6 4000 to 8000 180 to 299 435 966 766 1071
7 > 8000 180 to 299 35 207 116 204
8 > 8000 300 to 499 849 416 747 1026
9 > 8000 > 500 - - - -
*) ta = technology age
Source: computations with AERO modelling system
In addition to aircraft types and technology levels, two aircraft purposes are considered, i.e.
passenger/combi and freighter aircraft.
A number of traffic categories are distinguished within air transport demand, as follows.
Commercial flights are split into passengers and freight, both scheduled and non-scheduled.
Scheduled passengers are further divided into three classes: first/business, economy and
discount. Passengers are only carried by passenger/combi aircraft (which can be scheduled
or non-scheduled). Cargo is carried by both passenger/combi and freighter aircraft, scheduled
as well as non-scheduled. Table 3.2 provides an overview of the global volumes of scheduled
and non-scheduled passengers and freight transported in the base years 1992 and 1997.
Table 3.2 Global 1992/1997 passenger and freight volumes
Air transport volumes Units 1992 1997 Units 1992 1997
Scheduled pax 10
6
pax 1181.2 1728.7 10
9
pax-km 1836.5 2694.9
Non-scheduled pax 10
6
pax 132.8 198.5 10
9
pax-km 243.4 370.6
Scheduled freight 10
6
tonne 22.6 39.5 10
9
tonne-km 61.2 95.3
Non-scheduled freight 10
6
tonne 4.0 5.5 10
9
tonne-km 9.0 13.3
3.3.2 Spatial detail in AERO
In addition to the specifications related to aircraft characteristics and traffic categories, a lot of
spatial detail is included in the AERO model schematisation. More than 50,000 city pairs are
covered in the original data which, after grouping of minor city pairs, have resulted in some
19,000 'flight stages' to be explicitly distinguished in the model schematisation. Results for
individual flight stages can be aggregated to 196 IATA region pairs, based on 14 regions
derived from IATA definitions.
Figure 3.5 provides a map of the geographical locations of the IATA regions considered in the
AERO-MS. In addition to the geographical meaning of the IATA region, it also forms the basis
for the assessment of impacts on the various actors, such as the airlines, the government,
and the airline clients (consumers). In this respect, the airlines are considered as clusters of
carrier groups which are based in the various IATA regions. The actor 'government'
represents the authority collecting the income from charges in each IATA region. Effects on
consumers (changes in consumer surplus and expenses) are separately considered for the
groups of consumers transported by IATA region carrier groups.
20
Figure 3.5 Location of IATA regions
Emissions and concentrations are computed by the AERO-MS in three-dimensional space,
based on a global geographical grid of 5 by 5 (longitude/latitude) and 15 equidistant altitude
bands of 1 km. The inventory of emissions from ground sources is based on a global
geographical grid of 1 by 1.
3.4 Overview of models in AERO modelling system
The AERO modelling system was implemented as an integrated system of models which
more or less coincide with the computational steps described in section 3.1.
A total of nine separate models is distinguished as follows.
1. Aircraft technology model (ATEC)
2. Air transport demand and traffic model (ADEM)
3. Aviation operating cost model (ACOS)
4. Direct economic impacts model (DECI)
5. Macro-economic impact model for the Netherlands (MECI)
6. Flights and emission model (FLEM)
7. Other atmospheric immissions model (OATI)
8. Chemical Tracer Model KNMI (CTMK)
9. Environmental impact model (ENVI)
The model ATEC is involved with the computation of technical characteristics by aircraft type and
technology level (including fuel-use, emission coefficients and purchase prices of new aircraft)
based on a modelling of fleet build-up over time, driven by air transport capacity demand.
The model ADEM matches the supply and demand side of air transport and determines the
volumes of passengers and cargo transported; the number of aircraft flights by flight stage,
aircraft type and technology level; and the resulting fares and freight rates. The starting point
for the modelling of air transport demand and aircraft flights is provided by the Unified
Database, which is a computerised description of the volume and pattern of global air transport
activity in the base year (1992). The Unified Database was compiled from a number of major
existing databases and contains a detailed description (on the level of individual flight stage) of
passenger and cargo demand and actual flights by aircraft type and technology level.
21
The model ACOS computes all relevant variable cost components and the unit costs (per pax
and kg of cargo transported) of air transport by aircraft type, technology level and IATA region
pair. Moreover, the model ACOS converts the costs of possible measures in the air transport
sector to changes in unit operating costs.
Direct economic impacts on a global level are computed by the model DECI. These primarily
include the impacts on airlines in terms of operating costs, revenues and results. In addition, the
economic impacts of possible measures in the air transport sector are computed for other
relevant actors, including the government (changes in government income or expenditure),
consumers (changes in expenses and consumer surplus) and the airline industry (changes in
required fleet capacity).
The model MECI specifically applies to the Netherlands and computes the macro-economic
impacts in terms of employment and contribution to GNP), based on passenger and freight
volumes transported on flight stages to and from the Netherlands.
The model FLEM provides a detailed description of the actual flight profiles of individual aircraft
flights and computes the resulting fuel-use and emissions (including CO
2
, NO
x
, SO
2
, C
x
H
y
, CO
and H
2
O) on a three-dimensional spatial grid.
The model OATI computes the atmospheric emissions of other activities than air transport
(ground sources). OATI also computes the changes in land transport emissions due to
substitution effects between air and land transport modes, as estimated by the AERO-MS,
induced by scenario developments and/or measures in the air transport sector.
The emissions from air transport and ground sources are translated to a change in atmospheric
concentrations for a number of relevant substances (CO
2
, NO
x
and ozone) by the Chemical
Tracer Model (CTMK) from the Royal Netherlands Meteorological Institute.
The model ENVI computes a number of environmental impacts, including the effective UV-
radiation and changes in radiative forcing, following from changes in atmospheric concentrations
of CO
2
and ozone.
The following chapters 4 through 11 provide a further elaboration on the individual models
included in the AERO modelling system. Chapter 12 provides an overview of the analysis
capabilities and use options of the integrated modelling system as a whole.
22
4. Aircraft technology model (ATEC)
4.1 Purpose of ATEC within the AERO modelling system
The aircraft fleet operated by the airlines is continuously changing as a consequence of the
introduction of new aircraft to the fleet and retirement from active service of the older ones.
The process of advancing technology implies that the new aircraft taken into service have
different characteristics from the aircraft that they replace. In addition, because of changing
demands for air transport, the number of newly-introduced aircraft will be different from the
number of aircraft removed. The aircraft technology model ATEC was included in the AERO
modelling system (AERO-MS) to evaluate the fleet properties based on aircraft fleet
composition and technological advances.
ATEC focuses on the development over time of fuel-use, emission characteristics and new
aircraft prices of the aircraft types as modelled in the AERO-MS. These characteristics follow
from ongoing developments in technology, fleet sales, aircraft prices and active-service lives
of aircraft. These ongoing developments are defined using scenarios and can be influenced
by policy measures.
To facilitate the prediction of future fleet characteristics, ATEC is built on the concept of
generic aircraft types, which are specified in terms of seat and range bands. Within those
aircraft types, further distinctions are made based on aircraft purpose (passenger/combi
versus freighter aircraft) and technology level.
The relevant types of information to be assessed or predicted by ATEC are:
Aircraft fleet composition in terms of number of aircraft by IATA region, by aircraft type,
aircraft purpose, and technology level. The world fleet size in a future year can be
estimated assuming continuous sales-growth trends or be specified if such predictions
are already available from air transport demand calculations. In the latter case the fleet
growth per year from the base year onwards (1992) is determined from the fleet sizes
predicted by the AERO-MS, which are computed both at the IATA region and the global
level.
Aircraft characteristics such as fuel-use, emissions and new aircraft price of the aircraft in
service by aircraft type, aircraft purpose, and technology level. These characteristics are
expressed in terms of fuel-use and emission levels of a generic aircraft type reflecting the
average characteristics of all aircraft within the same combination of aircraft type, aircraft
purpose and technology level. The calculations of these properties take into account the
aircraft purchase rates over time and the technology characteristics associated with these
purchases on a year-by-year basis. Regional differences in technology characteristics are
considered based on the regional differences in fleet ages.
In accordance with the overall modelling principle underlying the AERO-MS, ATEC
computations are made for three situations: the Base, Datum and Policy case. The Base case
was used for the calibration ATEC, using recorded 1992 fleet and traffic data. The Datum
case reflects the situation where a future fleet is specified based on a number of scenario
specifications, including the autonomous advances in aircraft technology and aircraft prices
as a function of time. Policy cases describe the properties of a future fleet subject to certain
technological measures, from a specified measure year onwards.
4.2 Fleet composition and aircraft characteristics
Since aircraft may remain in service for 30 or more years, there may be large differences in
technical, economical and operational characteristics of individual aircraft. However, the
characteristics of a future aircraft fleet are difficult to predict solely on the basis of individual
aircraft type assessments.
23
For these reasons, within the AERO-MS the aircraft and their characteristics are grouped into
10 generic aircraft types, which are distinguished by seat band and range band (see table
3.1). Within the 10 generic aircraft types a further distinction was made on the basis of aircraft
purpose: passenger/combi and freighter aircraft resulting in 20 generic aircraft types. Each
of these 20 types can be associated with its own set of scenario and policy assumptions that
describes its development in time.
Due to the large differences in aircraft ages apparent within a fleet, a further distinction in
aircraft characteristics was made based on the technology age of the aircraft. Two technology
levels were defined: 'current' aircraft with technology age <= 12 years; and 'older' aircraft with
technology age > 12 years). This distinction allows the exploration of the consequences on
fleet characteristics of advancing technology.
Consequently, at the global level, the AERO-MS distinguishes 40 generic aircraft types.
Although the aircraft market is a global one, there is considerable variation in the
technological characteristics from one region to another, within the same aircraft type and
purpose. This is reflected in the regional differences in the average age of aircraft. In general,
aircraft of the same age of the same size and range have similar characteristics, irrespective
of their region of operation. In addition, the number of aircraft manufacturers is limited and
each serves world-wide and transparent markets. This allows for a further regional refinement
of generic aircraft type characteristics within the AERO-MS, which is based on the following
assumptions:
regional fleet differences in technological characteristics follow from differences in aircraft
age on a per region basis;
at purchase or manufacturing level, the aircraft business is considered a global market,
where the properties of aircraft solely depend on the technology age of the aircraft.
4.3 From aircraft purchases to fleet characteristics
To evaluate the fleet characteristics for a given aircraft type and purpose, and for a specific
moment in time (the scenario year), ATEC considers three aspects:
the aircraft purchase levels on a per year basis;
the technology age of the aircraft bought as a function of time;
the number of years elapsed in operational (active) service.
ATEC is based on these three aspects that are in itself a time-based series. Each aircraft in
the fleet is coupled to a purchase year, a technology year and the number of years in service,
being the difference between scenario year and purchase year. This coupling is best
explained based on the three views, as shown in figure 4.1:
The Purchase View shows the number of aircraft purchased on a per-year basis; with a
fixed timescale purchase, ranging from some 40 years before the base year up to the
scenario year.
An Active Fleet View with a timescale that is fixed relative to the scenario year. This view
shows the number of aircraft still in operation at the scenario year by purchase year;
A Technology View shows the aircraft produced as a function of technology year, and
thus represents the link between aircraft production and aircraft technology level.
The link between these views follows from the purchase behaviour of airlines and the
operational life evolution of aircraft. In each (purchase) year, new aircraft are bought by the
airlines. As aircraft production lines last a number of years with relatively constant technology
level (around 8 years according to AVMARK studies), this purchase volume of new aircraft
consists of a mix of technology standards of the last (typically 8) years. This implies that
aircraft with the same technology standard are apparent in only one technology year but
appear in more than one purchase year. Likewise, aircraft bought in one purchase year stem
from more than one technology year.
24
Figure 4.1 Views considered in ATEC on fleet development in time
These links are reflected in figure 4.1. The smallest entity that is shown here (and used within
ATEC) is a batch of aircraft of the same aircraft type and purpose that are bought in the same
year, being from a technology level of the same technology year. Figure 4.1 shows some
batches of aircraft that have the same technology standards (for one technology year) or the
same purchase year (for one selected year) in all three views.
The link between the purchase view and active fleet view follows from the operational life
evolution of aircraft. Once aircraft are purchased, they are operated for a number of years and
finally removed from active service. From statistics gathered by Airbus (market forecasts), the
rate of retirement from active service is known and is reflected in figure 4.2. This graph
reflects the fraction of aircraft still in active service as a function of operational life span. This
withdrawal scheme takes into account premature withdrawals due to accidents, etc.
The average service life varies between aircraft purpose and aircraft size. According to
Boeing and Airbus data related to the 1992 fleet, narrow body aircraft last on average 23
years (indicated as average scrap age), while wide body aircraft last approximately 19 years
on average. Freighter aircraft (not indicated) are retired on average after 27 years,
irrespective of their size. The retirement scheme relates the number of aircraft, purchased in a
particular year, to the number of aircraft still in service after a number of operational years.
The number of years in service is in turn determined by the scenario year and the year the
aircraft was purchased. As part of the scenario specification, it is assumed that the average
scrap age of aircraft increases over time. This makes the retirement curve more in line with
assumptions used in the analysis of Market Based Options for CAEP/FESG (Pulles et al,
Active fleet view
P
r
o
d
u
c
t
i
o
n
Technology view
P
u
r
c
h
a
s
e
s
Purchase view
A
c
t
i
v
e

f
l
e
e
t
Purchase Year
Technology Year
Purchases
Purchase Year
Retirements
Identical technology
Same production year
Aircraft in service
Aircraft by
technology
25
2000) where an average retirement age of 31 years for passenger aircraft and 39 years for
freighter aircraft was used.
Figure 4.2 Aircraft retirement function used in ATEC
Consequently, each batch of aircraft (with a given purchase year and technology year) can be
considered within three different views (see figure 4.1). The fleet characteristics are
determined by the appearance of aircraft in the active fleet, in itself a function of purchase
volumes and technology advances in time, as specified by scenarios and policies. Therefore,
the various scenarios and measures affecting fleet properties are effective in all views, but
usually are best presented within one of these views. Hence, whenever a policy or scenario is
developed and tested, there is a view that is best suited for the analysis. Some of the
measures are best reflected in the active fleet view, such as banning or scrapping of aircraft
that are bought more than X years ago. Likewise the possibility of a tax credit on new aircraft
purchases would be best reflected in the purchase view, while the scrapping of aircraft that do
not meet certain technology standards is best defined in a technology view.
In order to assess the technical properties of a future fleet, a scenario needs to be
established. Scenarios express different sets of assumptions regarding the ongoing,
autonomous and continuous developments of technology and aircraft prices in time, and are
best defined in the technology view. Once the technical properties (by technology year) have
been determined, the distribution of aircraft in the purchase view can be traced to assess the
technology levels of the aircraft purchased and the aircraft still in service. An age-weighted
summation across the technology levels of the aircraft in the active fleet (in itself a series of
purchase years) yields the fleet properties by region.
These views also show the (relatively) slow rate of introduction of new technology into the
100
A
i
r
c
r
a
f
t

i
n

s
e
r
v
i
c
e


[
%
]
0
10
20
30
40
50
60
70
80
90
Scenario
Year
5 10 15 20 25 30 35
Years in active service
Average
scrap age
Wide body
Narrow body
26
fleet. It takes approximately 8 years after introduction of new technology before all aircraft,
being purchased in a single year, will reflect this new technology. It takes approximately
another 30 years before all aircraft in the active fleet have at least this technology standard.
4.4 Scenario developments considered in ATEC
For the aircraft technology model, scenarios are defined that describe the future and past
(changes in) technology levels. In this respect, scenarios are considered as autonomous and
continuous developments over time, driven by economic and market forces. These scenarios
define the (change in) technology level on an annual basis. The scenarios are calibrated
using the data of the base year (1992) to model various technological and economic
developments. Additional data of 1997 are used for validation. The AERO-MS Base
(calibration) runs ensure that these scenarios are well-rooted and well-balanced. Within
ATEC, the relevant scenario developments relate to fuel-use, emissions and aircraft new
prices.
Fuel-use
Ongoing technology improvements indicate a reduction of fuel-burn rates over time. The fuel-
burn is a function of the technology year and is calibrated using the ICAO emission database,
1992 ANCAT data, as well as miscellaneous other data available in the public domain.
Roughly speaking, it can be assumed that before 1992 the characteristic fuel savings are 2%
per year. After 1992, these fuel savings are reduced to 1% per year.
Emissions
Emission developments over time depend on the type of emission. Unburned hydrocarbons
(C
x
H
y
), carbon-monoxide (CO), nitrogen oxides (NO
x
) and soot emissions are strongly related
to engine technology standards, both in terms of fuel-use per unit thrust and emission
characteristics per unit fuel-use. Historically these developments have differed across the
various types of emissions; hence a separate scenario was developed for each of them (see
figure 4.3). For calibration purposes, these scenarios need to be defined up to 1992. For
calculations of a future year, the scenarios from 1992 onwards allow to reflect the user's view.
Each scenario expresses the mass of the emission type formed per kg of fuel burnt (emission
index) as a function of time.
Figure 4.3 Scenario development of emission indices
Certification year
1992
CxHy
1980
NOx
20XX
CO
Emission index
27
The scenarios for C
x
H
y
and CO assume a constant annual development rate up to 1980 and
a zero change from 1980 to 1992, reflecting the historical pattern of introduction of
regulations. The NO
x
scenario assumes a constant annual development rate up to 1992.
Water vapour (H
2
O), carbon dioxide (CO
2
) and sulphur dioxides (SO
2
) emissions per kg of
fuel are directly dependent on the fuel chemical composition and not a function of engine
technology developments.
Aircraft new prices
Due to the ongoing economic and technological developments, aircraft new prices have been
rising at an approximately steady rate. This process is extrapolated into the future, assuming
the constant rate is unchanged if aircraft technology developments keep on following the
historical trend. If a policy measure or scenario is applied that affects the aircraft technology
rate of change beyond this trend, an associated increase in aircraft new prices is assumed.
4.5 Policy measures considered in ATEC
Policy measures of various kinds may be defined to simulate actions by the air transport
industry and the government. A policy measure typically acts on a scenario, starting at a
given year of introduction: the measure year. The impact of such a policy may vary in time, as
specified by the user. These measures (may) break the trends described by the underlying
scenario. Policy measures may have a direct, straightforward impact on a scenario, e.g. in
terms of enforced emission reductions following from government action. Other policies, such
as aircraft manufacturers' or airlines' responses to enhance technology improvement might
follow from changes in aircraft operating costs as induced by governmental policies. In the
latter case, the introduction of such policy measures (changing the technical characteristics of
aircraft and the resulting fleet mix) would take place some time after the initial measures have
been implemented. The implementation of a policy might also lead to aircraft purchase rates
breaking trends, both in aircraft purchase volume as well as in technology composition within
a purchase year. When using the AERO-MS, it is the responsibility of the user to specify a
coherent set of policies and scenarios.
Within ATEC, the following policy measures can be distinguished:
Fuel efficiency improvement. A fuel efficiency improvement measure is reflected by an
change in fuel technology development over time (in addition to the assumed scenario
development). In case of (almost) readily-available technology such an enforced fuel
efficiency improvement measure could be introduced promptly. A fuel efficiency
improvement will affect both fuel-use and emission levels, assuming invariant emission
indices. Also, the additional fuel efficiency improvement beyond the assumed scenario
development will have an impact on aircraft new prices.
Emission stringency. The emission stringency acts on a user-specified emission index
only. Fuel-use and other types of emissions are not affected. As with the fuel efficiency
improvement, aircraft new prices will be affected, depending on the extent to which the
specified stringency is more severe than the assumed scenario development.
Regional operating ban. The regional operating ban forbids aircraft that do not comply
with a given technology standard to be operated in (a) specified region(s). In assessing
the effects of this policy measure on regional fleet operation it is assumed that aircraft are
allowed to escape such a ban and be operated in the non-banned regions.
Technology and purchase scrap. These policies will remove and scrap the oldest aircraft
in the fleet. In the case of technology scrap, aircraft are removed based on technology
age; in the case of purchase scrap, aircraft are removed on the basis of the number of
years in active service. The scrapping measures remove the oldest and thereby most
polluting aircraft from the fleet.
28
Tax credit measure. This measure provides a means to increase the proportion of the
latest technology in the annual purchases. This increase is triggered by a tax credit and
will effectively reduce the market price of the aircraft with the latest available technology.
Consequently the market share of these aircraft will increase. It might also imply that
demand for new aircraft and hence the fleet size increase. However, manufacturers of
aircraft not eligible for the tax credit usually try to regain their market share. A possible
price response to reflect this behaviour is included in the measure specification and can
be controlled by the user.
The application of measures is not restricted to a single type of measure. In general, a well-
balanced set of various measures can be introduced simultaneously to reflect the impacts of
policies imposed by the government and the impacts of resulting manufacturers' and airlines'
responses on fleet technology developments.
4.6 Aircraft new price and costs associated with measures
Aircraft manufacturers may be faced with market forces or new technology standards that
enforce accelerated technological developments, i.e. developments beyond a standard rate of
improvement in aircraft technology. If the change in technology is smaller than the assumed
changes in the underlying scenario, there is no additional price change beyond the standard
change as reflected in the assumed scenario development of aircraft new prices. If
technology developments exceed the assumed scenario development, the additional
development costs to make new aircraft comply with the new technology standards will be
reflected by increasing new aircraft prices (see figure 4.4).
Figure 4.4 Changes in aircraft new prices following imposed technology changes
4.7 Calibration
In the Base run, historical technological and fleet size developments are matched to recorded
real world data. The aircraft technological properties are established within ATEC as a
function of historical scenario developments expressed in scenario parameters, by technology
and/or purchase year. These properties have been aggregated into the two technology levels
(technology age <= 12 years and technology age > 12 years) for the base year 1992 and
expressed in the scenario parameters to be determined.
0
Technol ogy change
Def aul t
t echnol ogy
change
Def aul t
pr i ce
c hange
Addi t i onal change i n
ai rcraf t new pri ce
29
For calibration purposes, the 1992 ANCAT database was investigated and the technological
properties of the ANCAT-recorded aircraft types were coupled to the ICAO emissions
database. All recorded aircraft types within the ANCAT database were classified according to
the AERO-MS aircraft types and the two technology levels. Using ANCAT recorded aircraft
types and flight data, and aggregating across all flights within the ANCAT database per
AERO-MS aircraft type and technology level, fleet average values for the various
technological characteristics were obtained. These values were then used to determine the
required scenario parameters by comparison of the recorded characteristics (from the ANCAT
database) and the AERO-MS predictions.
4.8 Fleet matching from Base to Datum
The ATEC Base run follows from recorded data up to 1992. The data related to the period up
to 1992 follow from the calibration process and cannot be changed by the user. Whenever a
Datum or Policy run is made, ATEC requires an estimate for the future fleet size, and a set of
coherent scenario specifications which describe the development of technology
characteristics in time. The estimate of the future fleet size may either be user-specified or
obtained from other models within the AERO-MS (ADEM and DECI). Based on the estimate
of the future fleet size ATEC determines the growth in purchases from 1992 onwards. Once
the active fleet has been determined, the fleet characteristics are determined based on the
aircraft fleet composition in terms of technology year, and the development of the technology
characteristics over time as specified in the technology scenario. This procedure is illustrated
in figure 4.5.
Figure 4.5 ATEC prediction of future fleet characteristics
At the regional level, the regional fleet characteristics follow from the technology age
distribution of the regional fleets compared with the global fleet, as specified by the average
ages by aircraft type. If not specified, it is assumed that the historical differences in relative
ages are preserved. Optionally, the user may specify a change in regional aircraft ages to
enable the simulation of a technological catch up in some regions.
T
e
c
h
n
o
l
o
g
y
C
h
a
r
a
c
t
e
r
i
s
t
i
c
s
A
i
r
c
r
a
f
t

s
a
l
e
s
Scenari o Year
Scenario
Datum run
Base run
Cal i brated
scenari o
1992
Fi xed by cal i brati on proces Further devel opment of fl eet
Ti me
30
5 The Unified Database
5.1 Introduction
The Unified Database holds data on civil aircraft flights and air passenger and cargo demand
for the base year of the AERO model (1992). The term "Unified" is expresses that the
database was created by the processing (unification) of several existing databases. The
Unified Database gives world-wide coverage, though data are held at the level of individual
major city-pairs and groups of minor city-pairs. In spatial coverage and detail, and in holding
passenger and cargo demand data as well as aircraft flights, the Unified Database provides
the foundation of the AERO models comprehensive forecasting capabilities, in particular for
the aviation demand and air traffic model ADEM and the flight and emissions model FLEM.
The data structure of the Unified Database is maintained throughout the forecasting process.
It allows the economic and environmental effects on aviation activity of applying policy
measures to be identified in detail by direct comparison of the base year and (alternative)
forecast databases.
The Unified Database was created by merging the content of four other major aviation
databases. These are:
the ICAO 1992 "Traffic by Flight Stage" (TFS) data for international scheduled flights;
the US Department of Transport 1992 "T-100" data for US domestic scheduled flights;
the July 1992 ABC (now OAG) timetable for scheduled flights; and
the ANCAT (Abatement of Nuisances Caused by Air Transport) database for April 1992.
In consequence, the Unified Database is more comprehensive than any one of these
individually in its coverage of:
international and domestic flights;
scheduled and non-scheduled operations;
passenger and freighter aircraft flights; and
passenger and cargo demand.
The availability of demand data in the Unified Database is particularly distinctive among
aviation databases. This is fundamental to the forecasting processes of the AERO-MS, in
which the future volume of aircraft flights is largely determined by the forecasts of passenger
and cargo demand for air services.
The magnitude of the Unified Database can be appreciated from the following annual
quantities for 1992:
more than 53500 city-pairs (about 19000 after grouping of minor city-pairs);
24 million flights;
over 1300 million passenger trips;
over 2000 billion passenger-km;
close to 27 million tonnes of cargo lifted; and
70 billion tonne-km of cargo.
5.2 Data content and structure of the Unified Database
As mentioned above, the primary spatial concept in the Unified Database is each pair of cities
for which direct (non-stop) aircraft flights were recorded in 1992. For each city-pair or "flight
stage", the Unified Database contains data for:
the number of civil aircraft flights;
the volume of passengers carried; and
the volume of cargo carried.
Aircraft flights are categorised by the different generic aircraft types (defined in terms of
31
capacity and range capability) as shown in table 3.1. A further distinction is made based on
the two technology levels current (technology age <= 12 years) and older (technology age >
12 years) aircraft. Flights are also distinguished by different categories of operation (flight
types) as follows:
scheduled passenger/combi aircraft;
scheduled freighter;
charter passenger/combi aircraft;
charter freighter;
general and other non-commercial aviation.
Passenger demand is classified by scheduled or charter flight, and scheduled passengers are
further subdivided into three classes (first/business, full economy and discount). Cargo
demand is split between belly-hold/combi capacity and full-freighters.
On each flight stage (separately by scheduled and charter flight and by aircraft function), the
proportion of capacity provided by carriers based in the Netherlands, in the EU as a whole,
and in other world regions is recorded. This information permits the AERO-MS to report the
economic impacts on carriers according to the region in which they are based.
The Unified Database is structured in the form of three linked tables:
Flight Stage Data: information applicable to all flights on each flight stage, including origin
and destination cities, distance, region pair.
Zone Data: information applicable to each of the origin and destination cities referenced
by the flight stage data, including location and major/minor city type (where major cities
are considered individually and minor cities are considered in clusters).
Traffic Line Data: for each flight stage a separate "traffic line" for each category of
operation (flight type), recording the volume of demand, and the numbers of flights by
aircraft type and technology level.
An example of the Unified Database records and their links is presented in figure 5.1, showing
the entries in the Unified Database relating to the Zurich-New York city-pair. The central
component is the flight stage record from which the city of arrival and city of departure are
used to link to the appropriate city records. This particular flight stage links major cities and is
thus recorded as a major flight stage. A set of physical information is recorded such as the
flight distance (great-circle distance), Route Group and Region Pair; other indicators record
that it has no surface (ground transport) competition and is international.
For each of the origin and destination cities, location information is recorded (latitude and
longitude), as well as information used to relate cities (and thus flight stages) to flight
information regions, countries and world regions.
As noted above, for each flight stage there may be up to five different categories of air traffic,
and these give rise to different traffic line records. In this example there are scheduled
passenger and non-commercial records. The scheduled passenger flights are operated by the
largest aircraft types 6, 7 and 8. The non-commercial flights are solely operated by business
jets in AERO aircraft category 0.
At the traffic line level the demand is also recorded. For scheduled passenger aircraft flights
there is both passenger and cargo demand, the latter being belly-hold or combi traffic. The
relative sizes of the markets by scheduled passenger class can also be seen.
In this example 54.5% of scheduled passenger capacity is provided by carriers from non-
aligned Europe and the remaining 45.5% is supplied by carriers based in the destination
region (North America in this case). Where appropriate, the database also separately records
the capacity provided by EU carriers and Dutch carriers.
32
City: Zurich
Ci t y Zurich
Flight Information Region 153
Country Switzerland
Region Non-Aligned Europe
Latitude 47.4594
Longitude 8.5492
CPB Region West Europe
Maj or / Mi nor Maj or
City: New York City
City New York City
Flight Information Region 116
Country United States
Region North America
Latitude 40.85
Longitude -74.061
CPB Region North America
Maj or / Mi nor Maj or
Flight Stage: Zurich to New York City
Flight Stage Number 12626
Ori gi n Ci ty New York City
Destination Zurich
Distance (km) 6,322
Route Group North Atlantic
Region Pair North America to Non-Aligned Europe
Maj or / Mi nor Maj or
Scheduled passenger traffic index 19899
Scheduled freighter traffic index - 1
Charter passenger traffic index - 1
Charter freighter traffic index - 1
Non-commercial traffic index 19900
Surface Competition identifier 0
International 0
Traffic Line: Non Commercial
Traffic Line Number 19900
Flight Stage Number 12626
Info Source Non-Scheduled
Function Passenger
Number of Sectors 1
Number of Carriers 1
Proportion of capacity met by
carriers from the EU 0
Proportion of capacity met by
carriers from the origin region 1
Proportion of capacity met by
carriers from the destination
region 0
Proportion of capacity met by
carriers from the Netherlands 0
First Class Passengers -
Economy Class Passengers -
Discount Class Passengers -
Cargo (kg) -
Movement s Older Technology Current Technology
Aircraft Type 0 12 0
Traffic Line: Scheduled Passenger
Traffic Line Number 19899
Flight Stage Number 12626
Info Source Scheduled
Function Passenger
Number of Sectors 1
Number of Carriers 4
Proportion of capacity met by
carriers from the EU 0
Proportion of capacity met by
carriers from the origin region 0.545
Proportion of capacity met by
carriers from the destination
region 0.445
Proportion of capacity met by
carriers from the Netherlands 0
First Class Passengers 6,987
Economy Class Passengers 22,445
Discount Class Passengers 109,219
Cargo (kg) 6,108,062
Movement s Older Technology Current Technology
Aircraft type 6 0 343
Aircraft type 7 0 325
Aircraft type 8 36 298
Figure 5.1 An illustration of Unified Database records
33
5.3 Creation of the Unified Database
For each of the four databases (ICAO TFS database, USDOT T-100 database, ABC
timetable, and the ANCAT database) from which the Unified Database was created, their
respective classifications of aircraft types were made equivalent to the aircraft type
classification underlying the AERO-MS and the Unified Database (see table 3.1).
Next, the three data sources dealing specifically with scheduled traffic (TFS, T-100 and ABC)
were merged. Double-counting was eliminated, based on matching origin city, destination city,
carrier and aircraft function (passenger/combi or freighter) between the sources. The ANCAT
database was compared with the total information on scheduled flights derived from TFS, T-
100 and ABC sources and where a match was not found, the flights recorded in ANCAT were
considered for inclusion in the Unified Database as non-scheduled flights.
Given the dependency of the AERO model on aviation demand forecasting, a major
advantage of the TFS and T-100 data sources is that they include data on passenger and
cargo demand, as well as aircraft flights, for the flight stages which they cover. For the
scheduled flight stages drawn from the ABC timetable only data on flights are available, but
passenger and cargo demand can reasonably be synthesised from the load factors for similar
stages (by region-pair, length of haul, etc) implied by the TFS and T-100 sources. Demand
on charter flights was estimated by applying load factors (drawn from reports of charter
operations) to non-scheduled flights, derived mainly from the ANCAT database.
Data giving the scheduled passenger class breakdown were provided by IATA for a selection
of city pairs. These were the basis of the passenger class split for scheduled flight stages;
where no data were available for particular flight stages, estimates were synthesised from the
IATA information (again by region-pair, length of haul, etc). All charter passenger demand is
treated as discount class.
The ABC timetable gives information on planned flights, which tend slightly to exceed actual
flights. Thus an adjustment was made to flight levels taken from the ABC source, based on
flight stages present in both the TFS/T-100 and the ABC data, prior to the elimination of
double-counting. This adjustment also compensated for the fact that the ABC data used were
taken from July 1992, a month with above average traffic due to seasonal variation.
To speed processing, the version of the Unified Database used in the AERO model is
aggregated over minor flight stages. For this purpose, minor origin and destination cities are
grouped by countries (within the EU) or world region (elsewhere), hub airport to which
connected, and stage length. As a result, the initial more than 53500 flight stages were
aggregated to about 19000. No flights or demand flows were omitted in this process, and it
has been verified that the process of aggregation itself has not distorted the modelled
emissions or economic (operating cost and revenue) quantities.
Within the aggregation process, some 12300 flight stages connecting about 350 major world
cities continue to be recorded individually; these stages account for over 90% of the available
seat kilometres (ASKs) flown. This permits demand to be modelled by true origin and
destination, allowing for transfers between flight stages at the major cities.
When aircraft movements and passenger and cargo volumes were aggregated to the global
level, there was close agreement between the Unified Database and independent data
sources, but this could conceal differences at a more detailed level. Thus comparisons were
made of numbers of movements and volumes of passengers with independent data for large
airports (aggregated to the city level as necessary) around the world. The accuracy with
which the latter reproduced the quantities of movements and passengers for individual cities
was very good, as table 5.1 shows for some of the major European centres.
34
Table 5.1 Validation of the Unified Database for Individual European Cities
Commercial Flights Commercial Passenger Demand (10
6
)
City Independent
Source
Unified
Database
Difference Independent
Source
Unified
Database
Difference
Amsterdam (sched) 213 574 213 024 0% 15.6 15.8 0%
London (sched) 546 824 546 060 0% 58.6 54.6 +7%
Paris (total) 510 685 511 157 0% 50.3 40.8 -23%
Munich (sched) 140 925 139 804 0% 8.8 8.6 -2%
Berlin (sched) 137 524 125 930 -9% 7.1 6.6 -7%
35
6 Aviation demand and air traffic model (ADEM)
6.1 Introduction
A fundamental assumption of the AERO-MS is that the volume and spatial distribution of
aircraft flights is predominantly determined by the level and location of demand for air
services by passengers and cargo shippers. This assumption is central to the operation of
the aviation demand and air traffic model (ADEM), which forecasts the demand for air
services as an essential step in forecasting aircraft flights. ADEMs forecasts of aircraft flights
are fed to the model FLEM (the Flights and Emissions model) which then estimates the
emissions that would arise from those flights. ADEMs forecasts of demand, and associated
airline operating revenue, are meanwhile available to the model DECI (the Direct Economic
Impact Model) for computing the economic implications of the forecasts.
According to the principles illustrated in figure 3.2, ADEM prepares air demand and capacity
forecasts for two distinct future situations, i.e. a situation without emission-related policy
measures (the Datum case) and a situation where, for the same future year, one or more
policy measures have been introduced (the Policy case). It is the comparison of the Policy
case with the Datum case which establishes the economic and environmental effects of the
policy measures.
ADEMs forecasts of passenger and cargo demand are made for a prevailing level of fares
and freight rates. ADEM recognises that, in practice, the capacity offered on any route is an
interplay of the forces of supply and demand, including the costs of operation, the fares and
freight rates charged, the sensitivity of passenger and cargo demand to these, and
competition between carriers. If costs are increased (through a carbon tax, for example), fares
may be increased. Demand will then be lower than it otherwise would have been (in the
Datum case), as will the capacity offered. ADEM, in association with the model ACOS (the
Aviation Cost Model), emulates the final position that would be reached (in the Policy case) as
a result of these complex real-world interactions.
In response to its demand forecasts, ADEM forecasts scheduled and charter aircraft flights
(plus longer-distance non-commercial civil traffic)
at a high level of spatial disaggregation (individual city-pairs for the majority of seat-km), and
also by (generic) aircraft type, distinguishing range, capacity and technology level. This
permits ADEM to output its forecasts of passenger and cargo demand and aircraft flights to a
database of identical structure to the Unified Database, so that direct comparisons can be
made between the Datum and Policy forecasts and the base year situation. ADEMs
forecasting procedures are described in more detail in the following sections.
6.2 Computing the demand for air transport - the main principles
Updating the Unified Database
ADEM is an incremental forecasting model. Its fundamental input is the Unified Database
2
of
passenger and cargo demand, and of aircraft flights, by city-pair in the AERO base year. It
then estimates changes in demand and flights arising from the Datum scenario specified by
the AERO-User. The outputs of this Datum case then become the input database for a Policy
case in which the further differences in demand and flights due to Policy measures are
estimated.
The interaction between ADEM and the Unified Database during the forecasting process is
illustrated in figure 6.1. The Datum and Policy databases are copies of the Base Unified
Database where a common structure of flight stage records are maintained and ADEM
updates the air transport demand and flights pertaining to each flight stage.

2
Section 5 describes the Unified Database.
36
Flight Stage
Characteristics
DATUM
RUN
UNIFIED
DATABASE
(BASE)
UNIFIED
DATABASE
(DATUM)
POLICY
RUN
UNIFIED
DATABASE
(POLICY)
Base Demand and
Movements
Datum Demand and
Movements
Datum Demand and
Movements
Policy Demand and
Movements
Flight Stage
Characteristics
Figure 6.1 Structure of Unified Database manipulation by ADEM
Air traffic growth driven by economic growth
In a Datum scenario, economic and demographic growth rates around the world are assumed
to be the main drivers of growth in air transport demand. Air transport demand is forecast by
means of elasticities to these growth rates, calibrated from historical data. Separate growth
37
forecasts are made for business and non-business passenger demand, and for cargo
demand, by utilising different elasticities with respect to economic and demographic growth of
these categories of demand. As figure 6.2 illustrates, there is strong empirical evidence
supporting the relationship between economic factors, in this case GDP, and air traffic growth.
The points on this graph represent region pairs.
Figure 6.2 Air traffic growth as a function of economic growth
The response of air transport demand to economic and demographic growth is computed at a
country-pair level, by mapping from AERO-User-supplied inputs at the level of IATA regions.
Price elasticity of air transport demand and effect of other transport modes
Air transport demand is sensitive to changes in real fare levels and freight rates. In ADEM the
modelled response to changes in fares and freight rates is governed by elasticities which vary
in ADEM by region of travel and type of demand (business, leisure and cargo). The elasticity
values used in ADEM have been distilled from an extensive literature search of empirical
demand studies. Table 6.1 shows the elasticity values for two situations: without and with the
effect of competition from existing (surface) transport modes, where the effects of surface
competition may apply to overland flight stages.
Table 6.1 Out-turn passenger fare elasticities by aggregated region pair
Aggregated region pair No surface competition Existing surface
competition
Intra North America -0.6 -0.6
Intra EU -0.6 -0.8
North America EU -0.7 -0.7
Intra Asia -0.6 -0.7
North America Asia -0.8 -0.8
EU Asia -0.9 -0.9
EU Other Europe -0.7 -0.7
EU Middle East, Africa -0.8 -0.8
All Other -0.7 -0.7
Global average -0.7 -0.7
2
3
4
5
6
1,2 1, 6 2 2, 4 2,8 3, 2
GDP Growth/head (% growt h pa)
T
r
a
f
f
i
c

G
r
o
w
t
h
/
h
e
a
d
(
R
P
M
s

%

g
r
o
w
t
h

p
a
)
Actual Estimated
38
Absolute elasticity values increase with increasing surface competition, as can be observed
from the values in table 6.1 for EU and intra Asia, where surface competition is most relevant.
In this respect it should be noted that table 6.1 shows the elasticity values for aggregate
region pairs, while surface competition may be particularly significant for selective individual
flight stages.
The elasticities are applied to fare and freight rate changes specified by the AERO-User for
the Datum scenario. They are also applied in the Policy case, when the model itself
estimates the fare and freight rate changes that would arise from policy-induced changes in
airline operating costs, as computed by the model ACOS (a further description of this
interaction is presented below).
In ADEM the effects of air services being faced with competition from ground surface
transport modes are modelled in two respects. Firstly, air passenger and cargo demand
elasticities are higher (more sensitive) to reflect that, if air transport fares and freight rates are
increased, there will be loss of air traffic to the surface modes in addition to the outright
suppression of demand. Secondly, the AERO-User is permitted (as part of the Datum
scenario) to assume that different extents of high-speed rail (HSR) networks are
implemented, which increases surface competition on parallel flight stages and renders the air
traffic forecasts even more sensitive to changes in air fares. Figure 6.3 below illustrates these
effects. This implies, for example, that, on a flight stage of 1000 km a 5% fare increase would
reduce demand by 3% without a parallel high-speed rail service but by 4% if a parallel HSR
existed. The same fare increase on a 500 km flight stage would reduce demand by 5% or 7%
respectively.
- 2
-1.5
- 1
-0.5
0
300 500 700 900 1100 1300 1500
Stage Length (km)
Passenger
Fare
Elasticity
Elasticity -
No HSR
Elasticity-
With HSR
Figure 6.3 Effect of surface competition on typical fare elasticities
Consistently with the more sensitive air fare elasticities in the presence of surface
competition, ADEM applies cross-elasticities to estimate that part of the change in air
transport demand (due to fare changes) that would transfer to or from surface modes. The
resulting change in demand on surface modes is subsequently used by other models to
forecast the effect of air transport emission-related policy measures on surface emissions.
Forecasting aircraft flights
As already noted, capacity on each flight stage is assumed to match the forecast demand.
Empirical data show that, for a given length of route, both the service frequency and the
distribution of flights by size of aircraft vary systematically with the level of demand.
Examples are shown in figures 6.4 and 6.5. Various demand levels are expressed in different
density bands (ranging from 0.00-0.01 to >0.50), where density is expressed in million
39
passengers per annum (mppa). Figure 6.4 shows that, as the density, or demand, of a flight
stage increases so the proportion of larger aircraft increases. Figure 6.5 shows a similar
pattern for the next longer distance band, but the proportion of the larger aircraft is higher
than in figure 6.4 for any given density band. The pattern of increasing proportions of larger
aircraft as density increases is repeated for other distance bands, as is the pattern of a higher
proportion of larger aircraft for a given density band as distance increases.
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0.00 - 0.01 0.01 - 0.02 0.02 - 0.03 0.03 - 0.04 0.04 - 0.05 0.05 - 0.10 0.10 - 0.25 0.25 - 0.50 0.50 +
Density Band (million passengers pa on flight stage)
P
r
o
p
o
r
t
i
o
n

o
f

F
l
i
g
h
t
s
Up to 20 seats
20 to 79 seats
80 to 124 seats
125 to 179 seats
180 to 299 seats
300 to 499 seats
Figure 6.4 Scheduled passenger flights of between 250 and 500 km
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0.00 - 0.01 0.01 - 0.02 0.02 - 0.03 0.03 - 0.04 0.04 - 0.05 0.05 - 0.10 0.10 - 0.25 0.25 - 0.50 0.50 +
Density Band (million passengers pa on flight stage)
P
r
o
p
o
r
t
i
o
n

o
f

F
l
i
g
h
t
s
Up to 20 seats
20 to 79 seats
80 to 124 seats
125 to 179 seats
180 to 299 seats
300 to 499 seats
Figure 6.5 Scheduled passenger flights of between 500 and 1000 km
Using these clear empirical patterns, ADEM has been calibrated to forecast the change in
fleet mix as demand grows. On each flight stage, the typical effect is that the proportion of
flights operated by higher-capacity aircraft increases, while the proportion operated by lower-
capacity aircraft declines, subject to constraints on the (generic) aircraft types that have the
range capability. At the model-users discretion, ADEM introduces a new large aircraft type
where demand growth would justify it. ADEM also estimates the increase in flight frequency.
40
Thus ADEM forecasts growth in capacity by adjusting both the aircraft size mix and service
frequency on each flight stage in accordance with the growth in demand. In this way,
capacity is estimated to develop so that it satisfies the revised demand with realistic load
factors, thus reflecting the reasonable actions of carriers in their choice of aircraft types and
schedule frequencies.
As well as being responsive to demand changes, the mix of aircraft (in terms of size and
technology level) on a route is also modified by ADEM in response to unit cost changes
resulting from measures, as estimated for each aircraft type by the model ACOS. As
explained in connection with ACOS (see section 6.4), it is in effect assumed that carriers
would choose (so far as they could) to provide an increasing proportion of capacity by those
aircraft types for which cost per unit of capacity increases least (or falls most) in the Policy
case compared with the Datum case.
The response of aircraft mix to cost changes assumes that the total capacity (of seats and
cargo space) provided per flight stage does not change from that which arises from the
adjustment of the mix due to demand changes. Thus demand-led changes in aircraft mix are
calculated first, and the effect of the cost changes is then to modify the proportions of the
resulting capacity that would be provided by the different aircraft types. The linkages between
the ADEM and ACOS models are illustrated in figure 6.6.
ADEM
Without Measures
Forecasts of Aircraft
Movements by Flight
Type, Aircraft Type and
Flight Stage
ADEM
Without Measures Forecasts
of Aircraft Movements by
Type
Datum Case
Policy Case
ADEM
Computes Fares and
Demand With Measures
ACOS
Computes Unit Costs by
Aircraft Type and Flight
Stage
ACOS
ComputesChangein
Unit Costs by Aircraft
Type and Flight Stage
With Measures Forecasts
of Aircraft Movements by
Type
Computes Changein
Unit Costs by Flight
Stage
Figure 6.6 Linkages between the ADEM and ACOS models
In a Datum run, ADEM adjusts the technology mix of aircraft of each type operating on a flight
stage in accordance with the trends in technological development of the fleet and aircraft
purchase patterns as forecast for each region by ATEC (the Aviation Technology Model). In a
Policy run, ADEM computes adjustments to the mix of technologies required to minimise the
overall impact of measures affecting aircraft costs. This takes account of the relative
substitutability of generic aircraft types to provide the required capacity (on each flight stage)
over the stage length involved.
41
ADEM recognises that, though a particular aircraft type may achieve a cost advantage over
other types as a result of a policy measure, the extent to which it could actually be substituted
for the other types will be constrained, especially by its capacity and range capability. Thus,
aircraft of different technologies but having the same capacity and range characteristics will
be highly substitutable, and ADEM consequently forecasts a relatively strong reaction in
terms of the use of (say) current aircraft as the costs of older aircraft increase. On the
other hand, even a long time horizon does not facilitate the substitution of aircraft with
different range capabilities, so higher costs for one type of aircraft are forecast to cause very
little increase in the use of aircraft from a different range capability. The substitutability of
aircraft of different capacities lies between these two extremes.
In the model, the different levels of substitutability are represented by parameters calibrated
by relating the base year mix of aircraft types on flight stages to their differences in operating
costs. Thus the forecast aircraft mix on each flight stage adapts as operating costs change in
response to policy measures.
6.3 Response of demand to changes in cost through changes in fares
Most policy measures will increase airline operating costs. In AERO these cost changes are
computed by ACOS and passed to ADEM as a cost change per unit of capacity for each
aircraft type. As described above, ADEM estimates the extent to which there will be
substitution of relatively cost-advantageous types for other types. Another key response of
airlines, however, will be to seek to recoup the additional costs of policy measures from
higher fares and freight rates. The greater the degree to which they are able to do this, the
smaller will be the impact of the measures on their profitability. At the same time, the higher
fares and freight rates will restrain the growth in demand compared with the Datum case.
Thus the effect of the measures will depend crucially on how airlines respond to the increase
in costs, which in turn will depend on the scope that carriers have - within the constraints of
industry business cycles and competition - to pass on higher costs in increased fares and
freight rates. This is allowed for in AERO. As part of setting the Datum scenario, the AERO-
User can select a profitability criterion which determines the extent to which carriers would
be able to pass on higher costs.
At one extreme, the AERO-User can stipulate that carriers would be able to pass on cost
increases in higher fares and freight rates to achieve (to the maximum extent possible) the
profitability per unit of capacity that is estimated to be attainable in the Datum case. In this
situation, the policy measure would curtail growth in demand and capacity. Together with the
shift in aircraft mix towards cost-advantageous types, this would have an appreciable impact
on emissions. At the opposite extreme, the AERO-User can stipulate that carriers might have
to absorb the additional costs, so that fares would be unchanged (from the Datum level) and
demand would consequently be unaffected. Capacity also would be unchanged from the
Datum, apart from any mitigating effects of a shift in the aircraft fleet mix. Thus the policy
measure would be predicted to have only a limited impact on emissions (through the change
in aircraft mix), while the profitability of airlines would deteriorate. The range of outcomes is
illustrated in table 6.1, which shows AERO output for a fuel levy, with the two extreme
profitability criterion (PAF) settings, PAF=1 where the levy is passed on in fares and PAF=0
where the airline absorbs the levy.
Table 6.1 Demand, cost and revenue effects of a fuel levy of 20c per kg for two
extreme profitability criterion (PAF) settings
Scenario PAF=1 PAF=0
Absolute Increase Absolute Increase
Demand effects
Passenger Billion pax-km pa 5050 4670 -7.5% 5050 0.0%
Cargo Billion freight-km pa 236 217 -8.2% 236 0.0%
Effects on airlines
Operating costs Billion 1992 US$ 569 568 -0.2% 611 7.4%
Operating revenues Billion 1992 US$ 585 583 -0.3% 585 0.0%
42
Where a measure to provide a tax credit to the purchase of new aircraft is modelled, the
profitability criterion is also important in determining the effects of the measure. If the AERO-
User believes that carriers would be able to maximise the level of unit profitability from aircraft
operations, the model assumes that none of the cost reductions of a tax credit would be
passed on to consumers through lower fares and freight rates. In this situation the carriers
would benefit from enhanced profit levels, while demand - and hence flights and emissions -
would not grow faster than in the Datum case. (If low-emission aircraft types only were to be
subject to the tax credit, emissions would in fact decrease relative to the Datum.) On the
other hand, the AERO-User could stipulate that all cost reductions from the tax credit would
be passed on to fares and freight rates. This would increase the level of demand, thus
reducing the benefits arising from purchase of more aircraft of the newest technology. Again,
the two extremes are illustrated in table 6.2, which shows AERO output for a typical tax credit
measure, with the two extreme profitability settings, PAF=1 where the subsidy is not passed
on to consumers through lower fares and PAF=0 where the airline does pass it on. In this
example a tax credit of 50% has been applied to new aircraft of the latest technology year.
Table 6.2 Demand, cost and revenue effects of a 50% tax credit measure for new
aircraft of latest technology for different profitability criterion settings
Scenario PAF=1 PAF=0
Absolute Increase Absolute Increase
Demand effects
Passenger Billion pax-km pa 5050 5050 0.0% 5100 1.8%
Cargo Billion freight-km pa 236 236 0.0% 238 1.0%
Effects on airlines
Operating costs Billion 1992 US$ 569 557 -2.1% 566 -0.5%
Operating revenues Billion 1992 US$ 585 585 0.0% 586 0.2%
Figure 6.7 illustrates the complex real world process that ADEM is emulating if the AERO-
User assumes the greatest-possible pass-through of policy-induced cost increases in higher
fares and freight rates. Essentially, ADEM identifies the proportional fare and freight rate
change for each flight stage that (from a change in unit costs) will satisfy the AERO-Users
profitability criterion within the range of possibilities described above.
Figure 6.7 The problem faced by the Cost to Fares mechanism for ADEM
43
6.4 Treatment of policy measures in ADEM
Three types of policy measures influence the forecasts of ADEM as shown in figure 6.7.
These are:
policy measures resulting in unit aircraft operating cost changes;
policy measures imposing levies on demand (passenger and/or cargo);
policy measures resulting in fleet changes.
In this section it is assumed that measures are applied globally, i.e. on any particular flight
stage a measure would be imposed on all carriers equally. Section 6.5 discusses applying
local and discriminatory measures.
Policy measures resulting in unit aircraft operating cost changes
The majority of the policy measures affecting ADEM are due to cost changes computed by
the model ACOS. These measures generate changes in unit operating costs which will
influence the mix of aircraft flights supplying capacity. For example, a fuel levy will result in
changes in unit costs for each aircraft type for which the flight mix adjustment sub-model (in
ACOS) will determine the change in aircraft mix which minimises the impact of these
additional costs on carriers. Dependent upon the AERO-Users setting of the profitability
criterion, the cost changes will be translated into fare and freight rate changes, which then
affect demand. The unit cost changes by aircraft type are used to adjust the final mix of
aircraft required to provide sufficient capacity to accommodate the revised level of demand.
As an illustration of the translation of cost changes into changes in passenger fares and
freight rates, consider the effect on Amsterdam-New York scheduled fares of imposing a 10
US$ cents per kg fuel levy. Such a levy increases fuel costs by almost 50%, and unit
operating costs on this route by 10.2%. Assume that the profitability criterion is that airlines
will seek to maximise their profitability by passing on as much as possible of the unit cost
increase. The fare and freight rate elasticities on this route imply that the increased yield from
higher prices is offset by a reduction in demand (relative to the Datum case), so that it is not
possible to achieve the level of unit profit that would have been attainable in the Datum. In
fact, carriers can only achieve two-thirds of the Datum unit profit for this flight stage; this
maximum occurs when a 5% increase in passenger fares and freight rates is imposed.
Policy measures imposing levies on demand (passenger and/or cargo)
If the policy measure is a levy applied by the authorities directly to passenger tickets and
freight rates, ADEM treats this as an additional component of cost to be met by airlines for
each passenger or unit of cargo that they carry. The model then assumes that carriers pass
on these cost increases in fares and freight rates only to the extent implied by the AERO-
Users profitability criterion. Therefore, fares and freight rates are not necessarily increased
by the magnitude of the levy; in effect, the airlines may have to pay some of the levy
themselves. It follows also that, despite the levy being on fares and freight rates, demand will
only be affected to the extent that the levy is actually passed on to passengers and cargo
shippers.
While the ticket levy affects demand and hence capacity required, it does not alter the unit
operating costs of each aircraft type. Thus, with a ticket levy, changes in aircraft mix are
solely due to the impact of levied costs on fares and freight rates and ultimately demand.
Policy measures resulting in fleet changes
Non-fiscal measures can also be tested in ADEM. These do not impose cost effects directly,
but there are likely to be cost implications for airlines nonetheless. One such policy is to
enforce the retirement of the oldest aircraft. In this case, ADEM works with ACOS to compute
both the enforced changes in aircraft operations and the additional cost imposed on carriers
in adjusting their operations to comply with this imposed change to the aircraft fleet. From the
44
point of view of determining the effect on fares, demand and capacity, the higher operating
costs are treated in the same way as the fuel and ticket levies. The difference here, however,
is that the forecast adjustment in the mix of aircraft flights is constrained by the enforced
retirement of part of the fleet.
6.5 Discriminatory and localised measures
The description of ADEM thus far presumes that, on any particular flight stage, a measure
would be imposed on all carriers equally. For European policy-makers, however, it is also
important to understand the effects of imposing measures only on carriers based in the
European Union (EU). The AERO-MS allows for such tests. Where EU carriers compete with
non-EU carriers on a flight stage, ADEM applies the theory of Cournot oligopoly. This
presumes that demand is ultimately not loyal to particular carriers, so that any carrier
seeking to raise fares above those of competitors on the flight stage would rapidly lose traffic.
Hence if only EU carriers had (for example) to pay a fuel levy, they would not be in a position
to pass on the cost increase in the same way as if all carriers were levied. On the other hand,
there would be advantage to other carriers in raising fares if higher costs were inducing EU
carriers to do so, since the non-EU carriers would thereby achieve a windfall gain in revenue
(and profitability, since their costs would not be higher). The outcome is that some increase
in fares occurs, which is common to all carriers on the flight stage, but this will be less than in
the case where all carriers are subject to the measure.
As a result of discriminatory measures, therefore, the financial position of EU carriers is
forecast to deteriorate, while that of other carriers (which have the benefit of the fare increase
without the policy-induced costs) improves. Moreover, given the higher operating costs for
EU carriers, the equilibrium market shares also shift, it becoming more profitable for non-EU
carriers to expand their share and less unprofitable for EU carriers to diminish theirs. Thus
ADEM fully emulates the impact on competition of discriminatory measures.
For example, if a 10 US$ cents per kg fuel levy were to be applied to EU carriers only, the
impact on scheduled fares between Amsterdam and New York would be much less than
when all carriers are affected equally. There would be only a 2.7% increase in passenger
fares and freight rates (for all carriers on the route), compared with 5% in the earlier example.
Since the EU carriers have to bear a 10.2% cost increase they would see a decline in unit
profits of 112%, implying an operating loss from this flight stage. The non-EU carriers, on the
other hand, would benefit directly from the fare increase and thus would see a small rise in
their unit profits roughly in line with the increase in fares. In this instance the proportion of
capacity provided by EU carriers is estimated to decline from 52.5% to 50.9%.
ADEM can also estimate the effect of measures imposed specifically on The Netherlands.
Price elasticities on flight stages to and from Amsterdam are increased to reflect the
additional deterrence that a locally-applied policy measure would have on transfer
passengers at Amsterdam. A large proportion of this segment of passengers could travel via
alternative hubs, and suitable cross-elasticities ensure that demand on routes via other hubs
is correspondingly increased. For example, a 5% increase applied to all passenger fares
departing from Amsterdam only, reduces the scheduled passengers at Amsterdam by 16%
compared with only a 4% reduction when this fare increase is applied to all flight stages from
all airports.
45
7 Aviation cost model (ACOS)
7.1 Introduction
The cost of providing a given level of capacity (seats, cargo space) is likely to vary over time
and as a result of the imposition of policy measures. In part this will arise because the relative
cost of operating different aircraft types will probably change. The main function of the
aviation cost model (ACOS) is to estimate the changes in overall operating cost and the
relative cost of operating different aircraft types, because they are fundamental to the
forecasting of passenger and cargo demand, and aircraft flights.
On the basis of the cost changes estimated by ACOS, the model ADEM (the aviation demand
and air traffic model) adjusts fares and freight rates. These influence ADEM's forecasts of
passenger and cargo demand, which in turn lead to estimates of the capacity to be offered.
ADEM uses the relative cost changes between aircraft types to modify the mix of aircraft (by
size and technology level) through which the forecast capacity will be met. ADEM is then able
to forecast aircraft flights, for which the model FLEM calculates the volume and spatial
distribution of emissions.
The cost changes from ACOS also contribute to the synthesis by the Direct Economic
Impacts Model (DECI) of the effects of policy measures on the financial position of airlines. .
DECI acts as a post-processor of ACOS and ADEM outputs. Therefore, whilst unit cost data
has been drawn from a number of sources and validated internally, ACOS results are
effectively validated and calibrated in the DECI validation process.
ACOS estimates the operating costs for each of the generic aircraft types shown in table 3.1.
As well as distinguishing by size (capacity) and range capability, costs vary by type of
operation (scheduled versus non-scheduled and passenger/combi versus freighter aircraft)
and technology level. The operating costs are estimated for each flight stage recorded in the
Unified Database. However, in practice only changes in operating costs through time or as a
result of policy measures are required by the AERO forecasting system. Thus ACOS only
has to be concerned with those categories of cost which will vary over time differentially by
generic aircraft type and/or which will be directly modified by the imposition of policy
measures. Of particular interest are cost changes, which may differ between the older and
current technology levels. The cost categories treated by ACOS are: i) flight crew; ii) cabin
staff; iii) maintenance; iv) fuel; v) airport fees and navigation (en-route) charges; and vi)
capital and finance costs.
Costs which do not vary with either aircraft type or policy measure, such as general and
administrative or ticket and sales costs, are excluded from ACOS, though these are
subsequently treated as "volume-related costs" by DECI.
ACOS estimates the differences in costs in the forecast year with policy measures in
operation (the Policy case) compared with the same year without measures (the Datum
case). Since carriers are assumed to have had the opportunity to adapt to the impact of policy
measures, ACOS is not constrained to the short-run variability of costs. In particular, the mix
of aircraft types and technology levels that are forecasted in the Policy case may develop
differently since the base year, compared with the Datum case.
7.2 Structure of aircraft operating costs in ACOS
The ACOS computations focus on the calculation of unit variable costs. The cost structure for
a particular type of aircraft and technology level is built up from "physical" aspects and costs
per physical unit. The physical aspects relate to the rate of fuel consumption by aircraft, their
speed, and flight stage distance. From these, block time and fuel consumption per flight are
derived for each combination of generic aircraft type and technology level, on each flight
stage (to the extent that aircraft range is sufficient for the flight stage distance). Costs per kg
46
of fuel and per crew-hour, etc. are then applied to the physical quantities to generate the
financial cost of operation of each aircraft type and technology level combination on each
flight stage.
Initially, the cost per physical unit of each component of cost is calculated at a region pair
level of spatial aggregation and relates to cycle (a single flight), distance or block time,
depending upon the component of cost, as set out in table 7.1 below.
Table 7.1 Cost components by physical unit
Cost component Cycle-related Distance-related Block hour-related
Block time
* *
Fuel
* *
Maintenance
*
Flight crew
*
Cabin crew
*
Landing
*
En-route/navigation
*
Depreciation
*
Finance
*
Such cost components are based on empirical data at different levels of aggregation. They
are all estimated on a region pair level for each aircraft type. Some also have additional
dimensions as outlined in table 7.2.
Table 7.2 Dimensions of cost components
Cost
component
Physical
Unit
Region
Pair
Aircraft
type
Techno.
level
Aircraft
purpose
Distance
band
Movement
type
Block time Cycle
* *
Block time Distance
* * * * *
Fuel Cycle
* * * *
Fuel Distance
* * * * *
Maintenance Block-hour
* * *
Flight crew Block-hour
* * * *
Cabin crew Block-hour
* * * * *
Landing Cycle
* * *
En-route Distance
* * *
Depreciation Block-hour
* * * * *
Finance Block-hour
* * * * *
Estimation of fuel consumption in ACOS is based on a sub-model which mimics the fuel-use
processes in the flights and emissions model (the model FLEM, see chapter 10). This sub-
model and the resultant ACOS outputs have been calibrated against FLEM outputs.
Depreciation charges represent the loss in value of aircraft through ageing and
obsolescence. The charges are based on empirical information (though cyclical effects are
excluded, as far as possible). Analysis of market values for different ages of aircraft of the
same type indicated that aircraft typically depreciate at about 6% of their residual value (in
real terms) per annum. This was incorporated in an annual depreciation function, which
varies with age around the 6% figure. Coupled with the age distribution and the initial
purchase price of each aircraft type in the forecast year (estimated by the model ATEC), the
depreciation function allows a weighted average annual depreciation cost to be calculated for
each generic aircraft type/technology level. This is divided by the average annual aircraft
utilisation for the type and technology level (also obtained from empirical data) to estimate
depreciation charges per block hour.
Finance charges are the opportunity cost of capital tied up in the purchase of aircraft. While
in practice this can take several forms, including retained profit and aircraft leasing costs,
finance charges are represented in the AERO system by interest on the residual (non-
47
depreciated) value of the aircraft fleet in the forecast year. They are estimated by multiplying
the average residual value of each aircraft type/technology level in the forecast year by a real
interest rate, which is known by IATA region as part of the AERO scenario specification. The
resulting interest burden for each generic aircraft type/technology level for each region is
divided by the annual utilisation to obtain a finance charge per block hour. Values for each
region-pair are then obtained as the simple average of the charges in the origin and
destination regions concerned.
For example, scheduled aircraft flights between the EU and North America by long-range
aircraft with between 300 and 499 seats would typically give rise to the cost components
shown in table 7.3, below. Costs are shown separately for the older and current
technology levels, to emphasise the important distinction between the AERO technology
levels in determining aircraft operating costs.
Table 7.3 Variable cost components (1992 US$) of long haul aircraft between 300 and
499 seats between EU and North America
Component Technology age > 12
years (older)
Technology age <= 12
years (current)
Cost per cycle 4368 4216
fuel cost 699 548
route and landing costs 3668 3668
Cost per km 4.15 3.35
fuel cost 3.43 2.63
en-route costs 0.72 0.72
Cost per block-hour 5711 5507
cabin crew 675 630
flight crew 1182 1182
Maintenance 2245 1461
capital depreciation 952 1139
Finance charges 656 1095
This example demonstrates typical differences between the costs of current and older
technology aircraft. As can be seen, depreciation and finance charges are higher for current
technology aircraft than for older; this is because they are younger on average, and thus
less-depreciated, and they also have higher initial purchase prices (in real terms) than older
technology aircraft had. However, current technology aircraft have advantages over older
technology in terms of running costs, such as lower fuel consumption and maintenance.
Figures 7.1 and 7.2 clearly demonstrate the trade-off between capital charges (depreciation
and financing) and running costs (fuel and maintenance) within the total costs of the two
technology levels.
These costs per cycle, kilometre and block-hour, having been established for a region pair,
can be used to compute the cost specifically for any city pair within this region pair, for the
type of aircraft concerned. For example, the Amsterdam-New York route has a (great circle)
distance of 5857 km and takes 8.17 block hours. Thus the cost per flight on this flight stage
for the type of aircraft in table 7.3 is:
older technology: 4368 + 4.15 x 5857 + 5711 x 8.17 = $ 75347
current technology: 4216 + 3.35 x 5857 + 5507 x 8.17 = $ 68861
The final step to obtain unit costs in a comparable form between aircraft types is to divide the
per-flight cost by aircraft size (in terms of seats for passenger aircraft and cargo capacity for
dedicated freighters) to give the cost per unit of capacity for each aircraft type and technology
level
48
Fuel Costs
29%
Maintenance
Costs
24%
Route and
Landing Costs
10%
Crew costs
20%
Capital and
Finance Costs
17%
Figure 7.1 Base cost composition for 300 to 499 seat long haul older technology
aircraft for route from Amsterdam to New York
Fuel Costs
23%
Maintenance
Costs
17%
Route and
Landing Costs
11%
Crew costs
21%
Capital and
Finance Costs
28%
Figure 7.2 Base cost composition for 300 to 499 seat long haul current technology
aircraft for route from Amsterdam to New York
49
7.3 Processing of policy measures
The difference in unit costs between Datum and Policy cases (due to measures) for each
aircraft type is entered into the fleet mix adjustment sub-model in ACOS. This sub-model
forecasts, for each flight stage, an increase in the proportion of capacity supplied by the
aircraft type(s) with the least unfavourable change in costs. The degree of adjustment in the
aircraft mix depends upon the differential changes in cost between airicraft types, and their
relative substitutability
3
in terms of capacity required and length of haul. The upshot is that the
average change in costs for each flight stage will be less than if the aircraft mix had remained
the same. This average cost change is passed to ADEM to calculate changes in fares, with
consequential effects on ADEM's forecasts of demand and capacity for each stage.
The cost changes by aircraft type are subsequently transmitted to ADEM, where the flight mix
adjustment sub-model is applied again, though after the final capacity forecast for each stage
has been estimated. This further adjusts the mix of aircraft types to be consistent with ACOS-
estimated unit cost changes and the final estimate of required capacity.
A variety of policy measures can be considered in the AERO-MS which will change the costs
computed in ACOS, including economic and regulatory measures (see also chapter 12). The
economic measures generally lead to a direct change (increase) in one or more of the
variable cost components such as fuel or landing costs. Such economic measures include:
fuel taxation;
airport charges;
route charges.
Another example of an economic measure is to provide a tax credit to the purchase cost of
the latest technology aircraft. The effects of such a measure are reflected in the aircraft
characteristics modelled by ATEC, including the revised purchase cost of new aircraft by
purchase year and a modified age distribution. Within ACOS, these changes feed through into
the calculations of aircraft depreciation and finance charges.
Examples of regulatory measures are the scrapping measure and the operating ban, under
which aircraft are forced to be prematurely retired or taken out of service on specific routes.
The primary effect of this policy is to alter the age distribution of aircraft (as modelled by
ATEC) and as a consequence, the weighted average depreciation and finance charges are
altered in ACOS (in effect to exclude the oldest aircraft from the calculations). In addition, the
fuel consumption and emissions characteristics are altered (also modelled by ATEC), and
these are reflected in cost changes estimated by ACOS.
To illustrate the process of computing the overall change in operating costs for a traffic line,
the earlier example of the Amsterdam to New York flight stage for scheduled traffic is again
considered. Suppose that a fuel levy is to be introduced of 10 US$ cents/kg. This is a non-
discriminatory measure as it will give rise to changes in costs for both older and current
aircraft, as shown in table 7.4.
As the previous figures have shown, in the Base case the proportion of unit cost related to
fuel is significantly higher for aircraft of older technology and this pattern is repeated in the
Datum case. While the proportion of operating costs due to fuel increases considerably,
irrespective of technology standard from 25% to 33% for older aircraft and 20% to 27% for
current the advantage of current aircraft is clear from table 7.4. Even so, the increase in
overall costs will have a significant impact on demand and hence the overall number of flights,
when these effects are computed by ADEM.
Unit costs change by aircraft type as well as by technology level. The flight mix sub-model
minimises the overall cost impact of these relative unit cost changes and the 10 cents per kg
fuel levy results in increases in overall unit costs per seat of $20.40 and $15.30 for older and
current aircraft respectively (table 7.4). The change in unit cost which is passed to the

3
The model ADEM also uses the fleet mix sub-model. This sub-model and the term substitutability was explained within the
description of ADEM, section 6.2.
50
model ADEM from this measure, and which will influence fares, is a composite of these
increases weighted by the shares of capacity provided by different aircraft types on the flight
stage in the Datum case. This composite value is $17.40.
Table 7.4 Impact of 10 US$ cents per kg fuel levy on unit costs (1992 US$) for long
haul 300 to 499 seat aircraft
Cost Item Technology age > 12
years (older)
Technology age <=
12 years (current)
Datum cost per flight (US$)
78695 77233
Datum unit cost per seat (US$)
216.8 212.8
Policy cost per flight (US$)
86084 82809
Policy unit cost per seat (US$)
237.2 228.1
Change in unit cost (US$)
20.4 15.3
Percentage change in unit cost
9.4% 7.2%
51
8 Direct economic impacts model (DECI)
8.1 Introduction
The Direct Economics Impact model DECI has three main functions:
Post-processing of ACOS and ADEM results in terms of: traffic volumes, operating costs
and revenues.
Computation of other direct socio-economic impacts on airlines, such as: airline operating
results, employment, and contribution to (global) gross value added.
Computation of direct policy impacts on other actors: airlines' clients (consumers) and
governments.
Most of the information on air transport activity and related financial information for airlines
processed in DECI directly follows from the results of the models ACOS and ADEM. In
addition, more selective impacts are computed for two other important actors, i.e. the airlines'
clients (the consumers) and the 'government'. The 'government' is introduced as a third party
which may interact with airlines and consumers by means of taxation and tax credit
measures, introducing and affecting flows of money between each of the three parties. All
quantities computed in DECI are to be considered as 'direct' effects of air transport activity.
Indirect effects, i.e. effects to the wider economy, are not computed in DECI.
The most important role of DECI is to provide a comprehensive report of the most relevant
information required for the interpretation of the socio-economic effects of different scenario
developments and policies. As such, DECI also has a very important role in presenting the final
results of ACOS and ADEM.
Detailed information from ACOS and ADEM generally exists on the level of flight stage (from
origin to destination city) which is aggregated to the level of IATA region pairs. From IATA region
pairs, the relevant information is transformed in DECI to IATA regions by using IATA region
proportions of air transport on each IATA region pair. In this respect, the IATA regions represent
the total of the carriers (airlines) which are based in the IATA region. Hence in DECI, the IATA
regions represent both a geographical entity and the group of carriers based in the IATA region.
More specifically, the following interpretations apply:
airline impacts by IATA (carrier) region: impacts to the total of the carriers based in the IATA
region;
consumer impacts by IATA (carrier) region: impacts to the clients of the carriers based in the
IATA region;
government impacts by IATA region: impacts to the total of the governments within the IATA
region.
In addition to the regional dimension, where relevant the impacts computed in DECI distinguish
between scheduled and non-scheduled passengers and freight; and/or different flight types in
terms of scheduled and non-scheduled passenger/combi and freighter flights.
8.2 Computations performed in DECI
Figure 8.1 provides an overview of the computations made in DECI. Most computations apply to
each of the three different situations considered in the AERO-MS, that is the Base, Datum and
Policy case. These computations are referred to as the 'generic' computations. In addition, a
number of computations are made in DECI to assess the impacts on the various actors involved
with air transport. These computations apply to the Policy case only, and are referred to as the
policy-specific computations.
52
Figure 8.1 DECI model structure
The generic computations include:
Air-transport-related quantities.
Airline operating costs.
Airline operating revenues.
Airline operating results (difference between operating revenues and operating costs).
Air transport efficiency indicators.
Computation of required fleet.
Variable and total employment.
Contribution of airlines to global gross value added.
Policy-specific computations include:
Change in consumer surplus and consumer expenditure.
Change in airline operating revenues, costs and results.
Government income from charges.
Government expenditure on measures.
The air transport-related-quantities include a number of air transport volume parameters such
as the number of flights, passengers and freight transported, block-hours flown and air
transport volume expressed in revenue tonne-km. The number of revenue tonne-km is equal to
the summation of the number of passenger-km, multiplied by the average weight of one
passenger plus baggage (typically of the order of 100 kg), and the number of tonne-km of cargo
transported. All basic information is supplied to DECI by the model ADEM. In DECI, this
information is processed to serve as a basis for the further computation of costs and other socio-
economic impacts. Included in this processing is the computation of IATA carrier region
proportions and the conversion of air transport quantities to IATA carrier regions.
ACOS
cost information
per unit of transport:
- aircraft movements
- aircraft kilometres
- block hours
(by IATA region pair)
ADEM
(Unified Database)
- Air transport demand
(pax, cargo)
- carrier proportions
- pax fares and freight
rates
(by flight stage)
Airline operating
revenues
Airline operating
revenues
Impacts of policies
on various actors:
- change in consumer
surplus and expenses
- change in airline
operating costs,
revenues and results
- government income
from charges
- government cost
of tax credits and
scrapping
Impacts of policies
on various actors:
- change in consumer
surplus and expenses
- change in airline
operating costs,
revenues and results
- government income
from charges
- government cost
of tax credits and
scrapping
Airlines operating
results
Airlines operating
results
Contribution to
gross value
added
Contribution to
gross value
added
Employment:
- flight crew
- cabin crew
- maintenance staff
- other (volume related)
Employment:
- flight crew
- cabin crew
- maintenance staff
- other (volume related)
Airline operating
costs (by IATA region
and region pair)
Variable costs:
- flight crew costs
- cabin crew costs
- maintenance costs
- fuel costs
- capital and finance
charges
- route and landing
charges
Volume related costs
Total operating costs
Airline operating
costs (by IATA region
and region pair)
Variable costs:
- flight crew costs
- cabin crew costs
- maintenance costs
- fuel costs
- capital and finance
charges
- route and landing
charges
Volume related costs
Total operating costs
Base Base, , Datum Datum & &
Policy case Policy case
only only Policy case Policy case
Air transport related
quantities (by IATA
region pair)
- carrier(IATA Region)
proportions
- aircraft flights
- pax transported
- freight transported
- block hours flown
- revenue tonne
kilometres (RTK)
Air transport related
quantities (by IATA
region pair)
- carrier(IATA Region)
proportions
- aircraft flights
- pax transported
- freight transported
- block hours flown
- revenue tonne
kilometres (RTK)
Air transport
efficiency
indicators
Required fleet
53
Airline operating costs include variable operating costs and volume-related costs. The variable
operating costs are defined as the annual airline operating costs which may be directly affected
by the measures considered in AERO. They also reflect the various types of operating costs as
reported by ICAO, including: flight crew costs (salaries, expenses and bonuses); cabin crew
costs; maintenance costs; fuel costs; capital costs and finance charges; en route charges (costs
related to distance flown such as Air Traffic Control); and airport charges (direct station expenses
related to aircraft flights). These costs coincide with the cost components considered in ACOS
and are directly obtained from the processing of ACOS results. Volume-related airline operating
costs include all other annual airline operating costs, which are not directly related to aircraft
operation, such as aircraft standing charges, passenger services, station and ground expenses,
ticketing and promotion, etc. Volume-related costs are assumed to vary with total air transport
volume. Total airline operating costs equal the summation of the variable and volume-related
annual airline operating costs.
Airline operating revenues include all (annual) revenues from aviation operation, both passenger
and freight, and both scheduled and non-scheduled operation. They are computed by multiplying
the quantities of passengers and freight transported on the flight stage level with the passenger
fares and freight rates applying to the specific flight stage, based on the information obtained
from ADEM. Airline operating results equal the difference of total airline operating revenues and
total airline operating costs.
Based on the various air transport quantities, a number of different air transport efficiency
indicators are computed in DECI. These efficiency indicators include various types of load
factors as well as the specific cost and fuel-use per unit of air transport achieved (see section
8.3). In addition, a computation is made of the fleet required by carrier region to perform the
actual and projected air transport activities.
Variable employment includes the employment associated with the variable cost components
such as flight crew, cabin crew and maintenance. Volume-related employment includes all other
direct employment categories associated with the volume-related costs. In accordance with the
concept of volume-related costs, the volume-related employment is assumed to vary with the
total air transport volume.
The contribution to gross value added includes the summation of wages, profit (before capital
charges) and finance charges. Profit before capital charges is approximated by the summation of
airline results and capital costs. The total sum of wages, based on the computation of
employment, and finance charges are then added to this.
Consumer surplus is defined as the sum of the differences for each consumer of air transport
between the price he is willing to pay and the price he actually pays. The consumer surplus can
be interpreted as one of the measures of societal benefits, accruing from air transportation.
Consumer expenditure reflects the total amount actually paid by the consumers of air transport.
Government income from charges consists of all revenues derived from the application of
taxation measures. The government expenditure on measures follows from the application of
measures involving a tax credit or compensation payment to airlines or consumers (e.g. related
to the loss in asset value associated with the premature retirement of old aircraft). By definition,
DECI is only involved with the direct incremental effects on government income and expenses
following from the application of measures. Government income from regular taxation of the
airline industry, and indirect economic effects of global aviation activity, are not considered in
DECI.
In accordance with the results of the other models in the AERO-MS, all of the above impacts
computed in DECI (with the exception of the required fleet) are expressed in annual quantities.
54
8.3 DECI results
DECI produces a great many output variables on different geographical levels, i.e., on the
level of carrier regions, IATA region pairs, and the global level. In addition, results can be
produced on a more aggregate level, only separately considering the most important IATA
region pairs, while aggregating the other region pairs into a limited number of clusters.
In order to facilitate the presentation of results, the DECI output variables have been
structured in a number of output categories, including:
Air transport quantities;
Aircraft operating costs, revenues and results;
Employment and gross value added (GVA);
Fuel consumption;
Air transport efficiency.
In addition, for the Policy case only, a number of specific impacts of policy measures are
computed such as the change in consumer surplus and the government income and
expenditure related to measures.
The DECI results provide the basis for the interpretation and evaluation of the socio-economic
effects of policy measures. In order to support this evaluation process, a number of facilities
have been provided in the User Interface of the AERO-MS for the processing and
presentation of outputs, both in terms of tables/reports and in graphical form (see section 12.4
of this report). In addition, within the available geographical and air transport-related
dimensions, results can be presented in considerable detail, if so desired.
Since the outcomes of DECI also reflect most of the outputs from ADEM and ACOS, the
results of DECI have been extensively used in the calibration and verification of these models.
The set-up and results of the calibration and verification procedure are further described in
chapter 13 of this report.
Tables 8.1 through 8.4 provides a number of selective results within the main output
categories of DECI, for aggregated region pairs and aggregated carrier regions. These results
are presented for the base years 1992 and 1997 (wtth all financial quantities being expressed
in 1992 US$) and include:
Aggregated region pairs:air transport demand in billion revenue tonne-km (RTK) per
annum;
- aircraft flights in million flights per annum;
- aircraft-km in billion per annum;
- operating costs in billion 1992 US$ per annum;
- operating revenues in billion 1992 US$ per annum;
- operating results in billion 1992 US$ per annum;
- fuel consumption in billion kg per annum.
Aggregated carrier regions
- air transport demand in billion RTK per annum;
- commercial fleet in number of aircraft;
- operating costs in billion 1992 US$ per annum;
- operating revenues in billion 1992 US$ per annum;
- operating results in billion 1992 US$ per annum;
- airline-related employment in thousand employees;
- airlines contribution to gross value added in billion 1992 US$ per annum;
- fuel consumption in billion kg per annum;
- costs per RTK;
- fuel-use per RTK;
- RTK per ATK (air transport capacity offered in terms available tonne-km);
- RTK per aircraft-km;
- revenues per RTK;
- fuel consumed per aircraft-km.
55
Table 8.1 Selective DECI results for aggregated region pairs for base year 1992
Aggregated region pairs Total
Intra NA Intra EU NA-EU Intra Asia NA-Asia EU-Asia Other
Air transport quantities
Revenue tonne-km (10
9
) 73.6 18.8 34.6 27.9 22.4 15.1 85.8 278.2
Aircraft flights (million)
- technology age > 12 years 5.07 1.72 0.09 0.97 0.04 0.02 3.46 11.37
- technology age <= 12 years 6.93 1.66 0.11 0.95 0.03 0.03 1.94 11.66
- total 12.00 3.39 0.19 1.92 0.08 0.05 5.40 23.03
Aircraft-km (billion)
- technology age > 12 years 3.96 1.32 0.63 0.75 0.37 0.15 3.65 10.83
- technology age <= 12 years 5.05 1.16 0.75 0.86 0.31 0.31 2.77 11.23
- total 9.01 2.49 1.38 1.62 0.68 0.47 6.42 22.06
Operating costs/revenues
Operating costs (10
9
1992 US$) 59.6 32.4 19.1 25.6 12.5 9.5 75.7 234.3
Oper. revenues (10
9
1992 US$) 54.8 31.9 17.5 28.8 12.7 9.9 79.0 234.6
Operating result (10
9
1992 US$) -4.8 -0.5 -1.6 +3.2 +0.2 +0.4 +3.3 +0.3
Fuel consumption
Aircraft fuel use (10
9
kg) 38.1 9.7 12.8 11.1 8.7 5.5 38.6 124.6
Table 8.2 Selective DECI results for aggregated region pairs for base year 1997
Aggregated region pairs Total
Intra NA Intra EU NA-EU Intra Asia NA-Asia EU-Asia Other
Air transport quantities
Revenue tonne-km (10
9
)
96.3 28.4 48.0 56.7 33.8 28.6 123.4 415.1
Aircraft flights (million)
- technology age > 12 years
6.3 1.9 0.1 1.0 0.0 0.0 3.2 12.6
- technology age <= 12 years 8.7 2.9 0.1 3.0 0.1 0.1 3.9 18.7
- total
15.0 4.8 0.3 4.0 0.1 0.1 7.1 31.3
Aircraft-km (billion)
- technology age > 12 years
4.9 1.4 0.8 0.9 0.4 0.2 3.7 12.3
- technology age <= 12 years 6.2 2.0 1.1 2.5 0.6 0.6 4.8 17.7
- total
11.1 3.4 1.8 3.3 0.9 0.9 8.5 30.0
Operating costs/revenues
Operating costs (10
9
1992 US$)
65.7 35.7 19.9 42.8 15.6 14.2 90.3 284.2
Oper. revenues (10
9
1992 US$)
65.7 37.0 20.6 44.7 16.2 14.7 94.0 293.0
Operating result (10
9
1992 US$)
0.0 1.3 0.7 1.9 0.6 0.5 3.8 8.8
Fuel consumption
Aircraft fuel use (10
9
kg)
42.0 12.5 15.0 19.7 10.7 8.9 46.0 154.8
56
Table 8.3 Selective DECI results for aggregated carrier regions for base year 1992
Aggregated carrier regions Total
North America EU Asia Other
Air transport quantities
Revenue tonne-km (10
9
) 112.6 62.2 52.4 51.0 278.2
Aircraft fleet (# of aircraft)
- technology age > 12 years 3818 1561 761 2095 8235
- technology age <= 12 years 4316 1274 626 1029 7245
- total 8134 2835 1387 3124 15481
Operating costs/revenues
Operating costs (10
9
1992 US$) 83.2 63.7 40.0 47.4 234.3
Oper. revenues (10
9
1992 US$) 80.6 62.3 43.0 48.6 234.6
Operating result (10
9
1992 US$) -2.6 -1.4 +3.0 +1.2 +0.3
Employment and GVA
Employment (thousands) 768 377 296 382 1824
Contr. to GVA (10
9
1992 US$) 41.8 25.4 21.6 19.5 108.4
Fuel consumption
Aircraft fuel use (10
9
kg) 53.6 26.5 20.2 24.4 124.6
Air transport efficiency
Costs/RTK (US$/tonne-km) 0.74 1.02 0.76 0.93 0.84
Fuel/RTK (kg/tonne-km) 0.48 0.43 0.38 0.48 0.45
RTK/ATK (factor) 0.59 0.65 0.67 0.59 0.62
RTK/aircraft-km (tonne-km/ac-km) 10.45 13.72 21.33 11.86 12.61
Revenues/RTK (US$/tonne-km) 0.72 1.00 0.82 0.95 0.84
Fuel/aircraft-km (kg/ac-km) 4.97 5.84 8.2 5.66 5.65
Table 8.4 Selective DECI results for aggregated carrier regions for base year 1997
Aggregated carrier regions Total
North America EU Asia Other
Air transport quantities
Revenue tonne-km (billion) 152.5 94.0 95.5 73.1 415.1
Aircraft fleet (# of aircraft)
- technology age > 12 years 4765.9 1720.3 797.1 1922.9 9206.2
- technology age <= 12 years
5370.9 2042.0 1615.1 1918.7 10946.7
- total 10136.8 3762.3 2412.2 3841.7 20152.9
Operating costs/revenues
Operating costs (10
9
1992 US$) 92.9 73.9 61.6 55.8 284.2
Oper. revenues (10
9
1992 US$)
97.3 74.7 63.3 57.7 293.0
Operating result (10
9
1992 US$) 4.4 0.8 1.7 1.9 8.8
Employment and GVA
Employment (thousands)
799.6 448.1 445.6 417.6 2110.8
Contr. to GVA (10
9
1992 US$)
57.1 34.3 32.4 24.5 148.3
Fuel consumption
Aircraft fuel use (10
9
kg)
60.4 33.8 31.7 28.9 154.8
Air transport efficiency
Costs/RTK (US$/tonne-km) 0.61 0.79 0.64 0.76 0.68
Fuel/RTK (kg/tonne-km)
0.40 0.36 0.33 0.40 0.37
RTK/ATK (factor) 0.63 0.68 0.67 0.62 0.65
RTK/aircraft-km (tonne-km/ac-km)
11.35 14.79 20.96 12.90 13.83
Revenues/RTK (US$/tonne-km) 0.64 0.80 0.66 0.79 0.71
Fuel/aircraft-km (kg/ac-km)
4.50 5.32 6.95 5.10 5.16
57
9 Macro-economic impact model for the Netherlands (MECI)
9.1 Computations performed by MECI
The Netherlands Macro Economic Impact Model (MECI) is concerned with air transport activities
at Schiphol Airport and their socio-economic implications for The Netherlands' economy. These
include impacts in the air transport sector proper, and in other sectors which are directly and
indirectly related to the volume of air transport activity.
The impacts considered are employment and income, measured in number of employees and
million US$ of gross value added (or GDP), respectively. Employment is subdivided into three
components, viz. direct, indirect backward and indirect forward employment.
Direct air transport employment is split into:
passenger-related employment;
cargo-related employment;
aircraft movement-related employment;
other direct employment.
The indirect air transport is split into:
indirect backward employment;
indirect forward employment.

The indirect backward employment is related to those organisations that are suppliers to the air
transport activities at Schiphol Airport. The indirect forward employment is related to those
organisations for which Schiphol Airport is the main reason for establishment. In MECI the
indirect forward employment is subdivided into:
employment related to commercial activities;
employment related to hotels and catering;
employment related to transport and distribution.
In figure 9.1 the model structure of MECI is presented. The main inputs for MECI are a number of
specific ADEM outputs, related to volume of air transport at Schiphol Airport.
All functions used in the MECI computations are derived from an existing model which was
developed for the 'Inventory of Economic Effects Schiphol' (IEE) project'. This IEE model is
based on the following assumptions:
there is a direct relation between the volume of demand and the volume of employment at
Schiphol Airport;
there is a direct relation between the volume of employment at Schiphol Airport and the
volume of employment at the suppliers;
there is a direct relation between and the volume of demand and the volume of
employment at the users of aviation facilities and products.
The different components of the direct employment are computed using a set of exponential
relationships for each of the above demand categories, as developed in the IEE project,
based on a time series over the period 1980-1992. The backward-related employment is
computed on the basis of a regional input-output table used in the IEE project. The application
of the input-output table has resulted in the determination of a linear relationship of indirect
backward employment with the total direct employment. The forward employment is computed
by a linear relationship with either the number of passengers flying to/from Amsterdam (in the
case of the commercial services and the hotel and catering industries), or the tonnes of freight
transported (in the case of the transport and distribution-sector employment). The linear
relationships were based on data from the IEE project. The total air transport-related employment
simply follows from the summation of the direct and indirect employment.
58
Figure 9.1 MECI model structure
In the Base case, the contribution to GDP of the air transport sector is computed. by
multiplying the gross value added per employee by the total employment. The contribution to
gross value added (which is equivalent to GDP) includes wages, interest, rent, profit and
depreciation. For the Datum and Policy cases, growth in the gross value added per employee
is also taken into account, consistent with the scenario specification.
9.2 MECI results
Table 9.1 provides an overview of the MECI results computed for the base years 1992 and
1997, compared with the original results of the IEE model (valid for 1990).
Table 9.1 MECI results for 1992 and 1997
IEE model (1990) AERO-MS (1992) AERO-MS (1997)
Employment (employees)
Direct employment
- Passenger related
- Cargo related
- Movement related
- Other
37900
8800
3600
9000
16500
38187
9226
3422
8870
16669
54554
14023
5192
10405
24934
Indirect employment
- Backward
- Forward
26400
17700
8700
26827
18139
8688
39888
25913
13975
Total 64300 65013 94441
Contribution to GDP (10
9
US$) 3.69 3.90 6.32
Contribution to
national income
(GDP)
Indirect forward
employment:
- commercial services
- transport & distribution
- hotel/catering
ADEM
Air transport activity
Amsterdam:
- aircraft movements
- pax transported
- cargo transported
Scenario variable:
gross value added
per employee
input/output input/output
analysis analysis
regression regression
analysis analysis
Indirect backward
employment
Direct employment
related to:
- aircraft movements
- passengers
- cargo
- other
59
60
10 Flights and Emissions Model (FLEM)
10.1 Overview of FLEM
The purpose of the Flights and Emissions Model (FLEM) is to calculate aircraft engine
emissions and the distribution of these emissions in the global atmosphere. The model
calculates the three dimensional flight paths of aircraft flying from flight origin to destination.
Along these generated flight profiles the emissions during the flight are calculated based on
local atmospheric parameters and momentary throttle settings. These parameters and
settings are in turn dependent on flight parameters such as speed, weight and altitude.
The calculated emissions are distributed across a three-dimensional (3D) grid that spans the
global atmosphere. Summation across all flights yields total emissions per grid cell, as a
result of aviation activity. These 3D emissions can be further processed and presented in
various ways, e.g. as global totals, by altitude band, per LTO cycle, or by user-selected
airport.
For the computation of total commercial aircraft engine emissions, FLEM uses inputs from
other models: a description of aviation activity and a description of (future) aircraft fleet
properties. The aviation activity in the base year is input from the Unified Database. For future
years, the traffic distribution is determined by the model ADEM.
The information contained in the Unified Database is used to select one of the AERO-MS
aircraft types and to map the flight track on the globe from origin to destination. Other required
inputs relate to aircraft and engine data (reflecting future fleet characteristics), which are
partly calculated by ATEC and partly supplied by the AERO-MS database. This information is
used to calculate flight profiles (altitude, speed, distance), fuel-use and emissions. The flight
profiles in turn are mapped onto the flight track between origin and destination. Figure 10.1
provides an overview of the FLEM computation scheme.
Figure 10.1 FLEM computation scheme
In addition to commercial aircraft engine emissions, FLEM is also able to estimate military
emissions. The ANCAT military traffic inventory for 1992 is used for the calculation of the
military contribution to fuel-use and (NO
x
) emissions.
Spati al emi ssi ons
di stri buti on
Present at i on of
emi ssi ons
Ai rcraft and fl eet
characteri sti cs
Dat a from ot her model s
FLEM speci f i c
Scenari os
Pol i ci es
Scenari os
Pol i ci es
Avi ati on acti vi ty
dat abase
FLEM:
Ai rcraf t wei ght
Ai rcraft 3D fl i ght path
Ai rcraft fuel fl ow
Ai rcraft emi ssi ons
61
Some basic features of FLEM are:
Several emission models can be selected; for example, NO
x
emissions can be computed
using two different methods, the 'P3T3' or the 'Boeing-2' method (see section 10.2.4).
The flight profile computation is based on actual (fuel) weight estimates and includes the
phenomenon of 'step climb', which can be controlled by user specifications. The flight
profiles reflect actual aircraft operating procedures.
Fuel-use and emissions are based on aircraft technology characteristics of a future fleet
that are driven by scenarios and optionally subjected to one or more policies (such as the
scrapping of older aircraft or flight altitude constraints).
Flexible grid definition options are available to allow for a user defined resolution of grid
cell sizes; several options exist for the aggregation and presentation of emissions, e.g. by
specified origin or destination; by altitude band or by aircraft type.
These basic features are further discussed in the next sections with respect to the
functionality of FLEM. The main output of FLEM are global aircraft engine emissions (CO
2
,
C
x
H
y
, H
2
O, SO
2
, NO
x
, CO) and fuel-use distributed across the 3D grid. Furthermore, the
model computes the use of oxygen and the flight time associated with aviation activity. The
CO
2
and NO
x
aircraft emissions are an input to the CTMK model in order to estimate changes
in concentrations of CO
2
, ozone and NO
x
as a result of aviation activities (see chapter 11).
FLEM computes emission levels for the three situations: the Base, Datum and Policy cases.
With respect to the Policy case, FLEM is able to compute the consequences in terms of a
change in fuel-use and emissions because of policies that affect the volume of aviation
activity (following from ADEM) or technology parameters (following from ATEC). Moreover,
FLEM itself contains a number of options for the simulation of policy measures. These
measures relate to (changes in) aircraft operations that affect the flight profiling, and thereby
the aviation fuel-use and emission levels. The available operational policies in FLEM which
affect the flight profiling are:
Flight level limits to simulate altitude restrictions to the flight profile (affecting the flight
level at cruise altitude), forcing aircraft to fly at a lower than optimal altitude.
Air speed restrictions.
Changes in climb and descent angles to reflect e.g. noise abatement procedures.
Changes in the extent of detouring, reflecting changes in air traffic congestion and
efficiency improvements in the ATC system.
Tankering: the behaviour of airlines to minimise fuel costs by taking extra (lower priced)
fuel onboard for the subsequent flight .
In addition to the operational policies that affect flight profiling, there is a policy option
available in FLEM that adds functionality to the aircraft technology (ATEC) policies. This
policy option allows changes to be made in aircraft aerodynamics, to reflect the impact of
technological advances or constraints. The effect of such measures is similar to a fuel
technology improvement measure.
10.2 Functionality of FLEM
FLEM can be divided into six functional parts:
a. The flight pre-processing module that gathers the flight information from the aviation
activity database (Unified Database or ADEM) and the aircraft technology
characteristics (from ATEC).
b. The pre-flight fuel calculations module to estimate the aircraft take-off weight based on
payload factor and flight distance on a per flight basis.
c. The flight profile calculation module, which calculates the flight profiles (in terms of
altitude, speed, weight and fuel-flow as a function of distance flown), flying from origin
to destination.
d. The emission index calculation module, computing emission indices along the flight
profile.
62
e. The grid calculation module, which converts the speed, fuel-flow and emission index
profiles into total fuel used and emissions per grid cell.
f. The military emission module, which calculates the military fuel used and NO
x
emissions.
10.2.1 The flight pre-processing module
Before the actual calculations of the flight profile and emissions are started, data need to be
retrieved from various models/databases. In order to reduce the amount of computation, in
the flight pre-processing module a considerable effort was spent to avoid repetitive
processing of data.
10.2.2 The pre-flight fuel calculations module
The pre-flight profile calculation module starts with a semi-empirical computation of the
aircraft's take-off and landing weight, taking into account the advance of aircraft technology as
computed by ATEC. This (fuel) weight estimate is provided by a separate model routine which
allows estimation of fuel-use throughout the AERO-MS. The take-off weight calculations are
based on aircraft data for a representative aircraft and on fuel-use parameters as calculated
by ATEC. The take-off weight is equal to the summation of:
aircraft empty weight;
the payload described in terms of maximum payload and a payload factor derived from
ADEM;
the fuel weight needed for the flight (including reserve fuel weight and the possible effects
of tankering).
The fuel weight is computed as a function of the distance to fly and determined by lift and
drag characteristics and engine fuel consumption, taking into account changing fuel weight
during the flight. This calculation procedure is based on typical cruise conditions (speed and
altitude), where the distance to fly is corrected to take into account the fuel consumption
during taxiing, take-off, climb and descent. It also takes care of aircraft operational constraints
with respect to the maximum take-off weight, maximum landing weight, the maximum fuel
weight and the maximum payload weight.
The FLEM calculations are based on the average aircraft characteristics obtained from ATEC.
Within the AERO-MS, generic aircraft types are defined that reflect the average
characteristics of groups of aircraft defined by seat and range band, aircraft purpose and
technology level. By definition, the use of average aircraft characteristics implies that there
are aircraft within the AERO-MS that have better and aircraft that have worse characteristics.
As a consequence, FLEM needs to process flights that cannot be operated with aircraft
having average characteristics, but are in fact operated with better-than-average aircraft
characteristics. This is especially applicable to flight distance where, within the seat and range
band definitions, there are shorter and longer range aircraft.
A special rubberising procedure was implemented to be able to handle the above situation.
This procedure basically allows scaling (re-sizing) of the average aircraft for the given
mission, while at the same time preserving the primary average technology characteristics
such as weight-to-thrust ratio, fuel-use per unit thrust, lift-to-drag ratios and wing loading.
Differences in payload characteristics are avoided using a concept based on load factors
rather than absolute payload weight. This procedure ensures that no flight obtained from the
aviation activity database is ignored because of aircraft type flight distance conflicts.
Tankering is a policy to take extra fuel on board that is not used for the current flight. Reasons
to do so are the opportunity to take advantage of differences in fuel prices or a reduced turn
around time in between flights. Tankering might be part of an airline response to avoid
increased costs implied by certain policies (such as fuel taxation). Because of the extra
onboard fuel, the aircraft weight during the flight will be higher than without tankering, which
implies an increase in fuel-use and emissions. The aircraft characteristics and the distance to
63
be flown limit the potential amount of tankering. Within FLEM, tankering is subjected to the
following constraints:
Aircraft maximum take-off weight of the tankered flight cannot not be exceeded.
The amount of fuel including tankering is limited to the aircraft maximum fuel capacity.
The aircraft landing weight including the tankered fuel should not exceed the aircraft
maximum landing weight.
Depending on the user specification of the tankering option, one of two additional constraints
is active:
The amount of tankered fuel is maximised, within the above mentioned constraints (full
tankering) regardless of the distance to fly on the next flight.
The tankered fuel is further limited to the amount of fuel that is required for the return trip
(return flight tankering).
10.2.3 The flight profile calculation module
Once the aircraft weights and characteristics are determined prior to a flight, the actual flight
profile is calculated. The flight profile is divided into the following flight phases:
Taxi-out from the airport gate to the runway in use.
Take-off, at full engine thrust.
Climb out up to 3000 ft at 80% engine thrust.
Continued climb from 3000 ft to cruise level at variable engine thrust setting.
Cruise phase (including stepped climb) with a prescribed speed schedule.
Initial descent, where the aircraft descents from the its last cruise segment down to 3000
ft in an aerodynamically clean configuration.
Final approach, down from 3000 ft to the flare, with the aircraft in landing configuration.
Landing.
Taxi-in from the runway to the airport gate.
The various phases of the flight profile as considered in FLEM are illustrated in figure 10.2.
Figure 10.2 FLEM flight stages
taxiout taxi in
Origin Destination
Takeoff Touchdown
cruise
c
l
i
m
b
o
u
t
c
o
n
t
i
n
u
e
d

c
l
i
m
b
i
n
i
t
i
a
l

d
e
s
c
e
n
t
a
p
p
r
o
a
c
h
s
te
p
c
lim
b
3000 ft
64
The flight phases taxi-out, takeoff, climb out, final approach, landing and taxi-in are in
accordance with the standard ICAO landing and take-off (LTO) cycle. In other flight phases
the flight profiles in terms of altitude, speed, thrust and fuel-flow as a function of actual flight
distance are calculated based on the following assumptions and data:
Stationary, symmetrical flight.
International standard atmosphere circumstances (ISA).
Distance to fly equals the great circle distance from origin to destination corrected by a
'detour' factor. Detouring accounts for deviations from the great circle between origin and
destination, due to airspace routing and ATC (air traffic control) procedures and
restrictions.
The computed take-off weight, based on (future) aircraft characteristics.
The computed landing weight based on (future) aircraft characteristics.
A step-climb table that couples the actual aircraft weight to the aircraft preferred flight
levels and ATC vertical separation procedures.
Special measures are taken to properly calculate representative profiles of short flights. Here
short refers to those flights that consists only of a take-off, climb, descent and a landing (and
of course taxiing) and that do not reach the (initial) cruise altitude. Figures 10.3 and 10.4
show typical examples of a short range and long range flight profile.
During the cruise part of the flight, the aircraft weight is monitored. If sufficient fuel is burnt
and weight lost during the preceding cruise segment, an aircraft is technically able to climb to
a new cruise level where the fuel economy is better. This is indicated as a "step climb". The
step climb will not be made if the remaining cruise part of the flight is less than a user-
specified distance. Whether or not to include 'step climb' in the flight profile calculation at all is
an option to be decided by the user.
The flight profiles resulting from the flight profile module are defined in terms of altitude,
speed, weight, fuel-flow and engine shaft speeds of the high and low pressure
compressors/turbines, as a function of the distance flown (see figure 10.3 and 10.4).
Figure 10.3 Typical short range flight profile
Distance [km]
41
43
45
0 20 40 60 80 100 120
W
e
i
g
h
t

[
M
g
]
20
60
100
140
180
S
p
e
e
d

[
m
/
s
]
0
1
2
3
4
A
l
t
i
d
u
t
e

[
k
m
]
1.0
2.0
F
u
e
l
f
l
o
w

[
k
g
/
s
]
height
weight
speed
fuelflow
65
Figure 10.4 Typical long range flight profile
10.2.4 The emission index calculation module
Based on a mapping of flight profiles on the global 3D grid structure, emission indices of all
aircraft engine exhaust gases are computed along the flight profile and distributed in the
worlds atmosphere. For a number of emissions (CO
2
, SO
2
and H
2
O) the level of emissions is
directly related to fuel-use. For other emissions (C
x
H
y
, CO and NO
x
) this is not the case,
however. For the latter emissions there are different ways to calculate the emission as a
function of altitude, speed and other engine parameters. These calculations have been
based, or calibrated, on the ICAO emissions database that contains engine emissions and
fuel-use data for four different thrust settings at zero altitude and zero speed. It should be
noted that the various existing calculation algorithms share a general lack of validation data
when considering actual flight conditions, i.e. non-zero altitude and non-zero speeds.
The main substance emitted by aircraft, which is not directly related to fuel-use, is nitrogen
oxide (NO
x
). For the computation of NO
x
-emission indices, two alternative methods are
available in FLEM for selection by the user. The alternative methods available are the so-
called P3T3-method, based on a characteristic engine combustion temperature and
pressure, and the Boeing-2 method, based on fuel-flow and atmospheric properties. The first
method allows for a detailed calculation of NO
x
emissions, but requires specific knowledge of
the engine characteristics and does not allow for an atmospheric humidity correction. The
Boeing-2 method does include a humidity correction that is dependent on the atmospheric
properties. Conceptually, the Boeing method is better suited for generic engines (of future
aircraft) while the P3T3 method yields results that are more accurate if based on accurate
engine characteristics. Figure 10.5 shows the differences in size and distribution of NOx
Taxi, takeoff
and climb
Cruise, including
stepclimb
Descent, landing
and taxi-in
0
2
.
4
4
.
8
7
.
2
9
.
6
1
2
.
0
A
l
t
i
t
u
d
e
0
.
0
2
.
0
4
.
0
6
.
0
8
.
0
1
0
F
u
e
l

f
l
o
w
2
0
0
2
2
4
2
4
8
2
7
2
2
9
6
3
2
0
W
e
i
g
h
t
0
6
0
1
2
0
1
8
0
2
4
0
3
0
0
S
p
e
e
d
height weight
speed fuelflow
Note: distance not to scale
0 100
200 250 2250 4250 5500 5700 5900
66
emission indices (in gram NO
x
per kg of fuel burnt) computed with the P3T3 and the Boeing-
2 methods based on a representative traffic volume. Up to 60% of the apparent differences is
known to be caused by the humidity correction that is part of the Boeing-2 method but not
part of the P3T3 method.
Figure 10-5 NO
x
emission model: Boeing-2 versus NLR P3T3 method
10.2.5 The grid calculation module
A three-dimensional grid spanning the globe's atmosphere is used for the spatial allocation of
emissions and fuel-use. The user can specify the grid size, but default values used are 5
o
x 5
o
x 1 km (latitude x longitude x altitude). The fuel-use and emissions in the grid cells are
calculated by integration along the flight path within the grid cell. In addition, the time spent
within the grid cell by the aircraft is calculated, together with the CO
2
, SO
2
and H
2
O
emissions, as well as oxygen consumption, which can easily be derived from the calculated
total fuel-use by grid cell. The result of these integrations and summations is a grid cell
database including all emissions and fuel-use from the flights according to the aviation activity
database, as specified by the model ADEM.
The fuel-use and emissions per grid cell can be further processed and aggregated by FLEM
to obtain values per:
airport and/or IATA region;
aircraft type and technology level;
altitude band.
10.2.6 The military emissi on module
The AERO-MS model focuses on the civil aviation fuel-use and emissions but has a facility for
processing military data as well. The military data are processed in a much simpler way
however, as these data were obtained from existing emission and fuel-use calculation results.
These existing results are based on military aircraft flights according to the ANCAT emission
NOx in g/kg fuel
0.0 4.0 8.0 12.0 16.0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
A
l
t
i
t
u
d
e

l
a
y
e
r

[
k
m
]
EI_NOx (Boeing-2)
EI_NOx (P3T3)
67
inventory subgroup. From this subgroup, two sets of results are available, one for 1992 and
one for 2015, where the latter is derived from the first one. The AERO-MS is capable of
interpolating over time in between these two (military) databases, taking into account the
anticipated growth in military traffic up to the year of calculation. FLEM then converts the
ANCAT data to the AERO-MS grid definition. The result of this conversion is a grid
representation of fuel-use and nitrogen, carbon dioxide and water vapour emissions, resulting
from military flights according to the ANCAT emission inventory subgroup.
10.3 Validation and computation results of FLEM
Table 10.1 shows a comparison of total fuel-use and NO
x
emissions computed by the AERO-
MS with information obtained from other sources. This information includes calculated
aviation fuel burn in 1992 from NASA, ANCAT and DLR, as published in the IPPC report on
aviation and the global atmosphere (IPCC, 1999). From the comparison of the AERO-MS
computational results with these other sources, it can be inferred that the computational result
of the AERO-MS is about 17% higher. One obvious reason for this is that the Unified
Database, underlying the AERO-MS computation, is more comprehensive compared with the
traffic databases underlying the other computational results. The ANCAT computation is for
example based on the ANCAT database, which is only one of the four sources used in
creating the Unified Database (see section 5.1 of this report).
Table 10.1 Fuel use and emissions comparison
Unit NASA ANCAT DLR AERO-MS Fuel use and
emissions *) P3T3 Boeing
Fuel use 10
9
kg pa 113.8 114.2 112.2 134.2 134.2
NO
x
emissions 10
9
kg pa 1.44 1.60 1.60 1.43 1.69
NO
x
emission
index
Gram/kg
fuel
12.6 14.0 14.2 10.6 12.6
*) In all numbers presented in table 10.1, fuel-use and emissions related to military air traffic
are excluded.
Further comparisons were made with a number of statistical sources. The International Energy
Annual (IEA) of the U.S. Department of Energy is one of the institutions which collects data
with respect to global aviation fuel-use. For 1992, the IEA reported 171.4 billion kg. This
includes the fuel-use related to military traffic. If military fuel-use is subtracted from the IEA
number, the fuel-use for civil aviation in 1992 is of the order of 154 billion kg. This is thus
significantly higher compared with the results of all computational sources, the difference with
the AERO-MS being the smallest (IEA data suggest that global aviation fuel-use is about 15%
higher compared with the AERO computation). Other statistical sources show even higher
global fuel-use numbers. A likely explanation for these differences is that statistical sources
on global aviation fuel-use include the use of kerosene in other sectors.
The FLEM fuel computations are also used in the AERO-MS to determine the costs of fuel,
which is an important component of the overall aircraft operating costs. From the comparison
of the fuel cost computations with the AERO-MS and actual ICAO data for 1992, as provided
in chapter 13 (table 13.7), it can be verified that a rather close match is obtained which also
confirms that the fuel-use computed by FLEM should be about right.
The aircraft emissions computed in the AERO-MS include NO
x
, CO
2
, C
x
H
y
, SO
2
, CO and H
2
O.
For the substances CO
2
, SO
2
and H
2
O there is a direct relationship between the extent of the
emission and the amount of fuel used. Hence, the reliability of these emission computations
equals the reliability of the fuel computation. For CO and C
x
H
y
emissions, no information from
other data is available for comparison. However, for these substances aviation emissions only
account for about 0.1% of total anthropogenic emissions. For CO
2
and NO
x
this percentage
was about 1.5% in 1992.
68
With respect to the of NO
x
emissions, the computational result of the AERO-MS for 1992 is
comparable with the result of NASA, in terms of NO
x
emission per kg of fuel (see table 10.1).
However, as the fuel-use computed by NASA is significantly lower, different absolute NO
x
emissions are computed by NASA and the AERO-MS. ANCAT and DLR compute NO
x
emission indices that are higher. There are many reasons that may explain the differences
between ANCAT, DLR, NASA and AERO, since NO
x
emissions are very sensitive to throttle
settings, specific engine types, and hence traffic composition. For specific flights, the AERO
NO
x
computations have been checked with a limited amount of available FDR data. From
these comparisons it followed that recorded and computed data showed good agreement.
10.3.1 FLEM specific results
FLEM results for the Base run (1992) across all flights by altitude band are depicted in figures
10.6 and 10.7. Figure 10.6 shows the flight time (in aircraft-minutes), the fuel-use (in 100 kg)
and the NO
x
emission (in kg) as a function of altitude. Summation across all altitude bands for
each quantity yields the overall totals. As can be expected, fuel-use and NO
x
emissions peak
at an altitude around 10,000 meters, where considerable time is spent at relatively high
(cruise) thrust settings. Between 1000 and 8000 meter altitude, relatively little fuel is burned
and emissions (in kg) are low, even when compared with the time spent at these altitudes.
This implies that aircraft spend more time at these altitudes in descent at relatively low thrust
settings and speed, than in climb (at relatively high thrust settings and speed). This is
confirmed by the examination of FLEM flight profiles.
Figure 10.7 shows the emission indices for NO
x
, CO and C
x
H
y
. This figure shows that most of
the C
x
H
y
and CO is emitted at low altitude, during taxi-out at low thrust settings. At these low
thrust settings the combustion process is incomplete. When observing the NO
x
emissions
relative to the fuel-use (emission index) in figure 10.7, the engines appear to produce
minimum NO
x
emissions at approximately cruise altitude (10,000 meters). At this altitude, all
emission indices show relatively low values. Furthermore, at altitudes between 2000 and
8000 meters, emissions indices of all substances (CO, C
x
H
y
and NO
x
) show relatively high
values. The engines are either operating at climb thrust (high NO
x
,) or at descent thrust
settings (high CO and C
x
H
y
) but never at design (low emissions) operating conditions.
Figure 10.6 Traffic characteristics by altitude layer
200*10
6
400*10
6
600*10
6
800*10
6
0
1
2
3
4
5
6
7
8
9
10
11
12
13
A
l
t
i
t
u
d
e

l
a
y
e
r
Ti me [mi n], Mass [100kg] (fuel ) and [kg] (NO
x
)
flight time (min)
Fuel use (100 kg)
NO
x
(kg) (Boei ng-2)
0
69
Figure 10.7 Emission indices by altitude layer
With respect to the horizontal distribution of fuel-use, figures 10.8 and 10.9 provide an
overview of global CO
2
emission distributions for the years 1992 and 2010, respectively.
Figure 10.8 Global distribution of aircraft engine CO
2
emissions in 1992
0 5 10 15 20 25
0
2
4
6
8
10
12
A
l
t
i
t
u
d
e

l
a
y
e
r
g/kg fuel per altitude layer
EI_NOx
EI_CO
EI_CxHy
70
0 maximum 2010
Figure 10.9 Global distribution of aircraft engine CO
2
emissions in 2010
Both figures 10.8 and 10.9 clearly show the world's dominant flying routes. The biggest
concentrations in CO
2
emissions (fuel-use) are in the US, Western Europe and to a more
limited extent in the Far East. There is a considerable increase in fuel-use between 1992 and
2010, with contributions of Western Europe and especially the Far East becoming relatively
more important.
10.3.2 A comparison between aircraft types
The graph of figure 10.10 reflects the differences in operation between aircraft types 0 (short
haul, less than 20 seats) and 8 (300500 seats, long haul). For short haul flights a significant
portion of time is spent at low altitudes, while the long haul aircraft spend most of their time at
cruise altitude. Consequently, the distribution of the amounts of fuel-use and related
emissions across altitude bands will be very different for these two aircraft types.
71
Figure 10.10 Distribution of flying time for small and large aircraft by altitude layer
10.3.3 Comparison with data from Flight Data Recorder
The computation of emissions is closely related to the computations of aircraft fuel-use and
flight profiles. For specific aircraft types, the flight profiles that form the basis for the FLEM
emissions and fuel-use calculations, can be compared with data from the Flight Management
System (FMS) or Flight Data Recorder (FDR). Such FLEM flight profiles and FDR data are
available for a number of B747-400 and B737-400 flights, though, due to the limited
availability of FDR data, it is not completely clear to what extent the FDR recordings are
actually representative.
Fuel-use and flight times on a limited number of city pairs were obtained for comparison with
the FLEM flight profiles. Figures 10.11 and 10.12 show a fuel-use comparison and flight time
comparison, respectively, between FLEM and FDR data for a number of Boeing 737-400
flights. From the recorded data, it turns out that there are relatively large variations from flight
to flight.These variations follow from day-to-day variations in weather conditions (wind), air
traffic control, use of runways, payload, etc.
From figures 10.11 and 10.12 it can be observed that the FLEM results appear to fit nicely
within the recorded variations. It should be noted that in these comparisons, FLEM flight
profiling has not been tuned to the specific airport departure and arrival routings according to
Standard Arrival Routes (STARs) and Standard Instrument Departures (SIDs).
1
2
3
4
5
6
7
8
9
10
11
12
13
14
A
l
t
i
t
u
d
e

L
a
y
e
r

[
k
m
]
0.0 10.0 20.0 30.0 40.0 50.0 60.0
Percentage of flying time
Long haul, 300 to 499 seats
Short haul, less than 20 seats
72
Figure 10.11 FDR and FLEM fuel use comparison for B737-400 flights
Figure 10.12 FDR and FLEM flight time comparison for B737-400 flights
Flight time Comparison B737-400
0
100
200
300
0 500 1000 1500 2000 2500 3000
Distance [km]
F
l
i
g
h
t

t
i
m
e

[
m
i
n
u
t
e
s
]
outbound
inbound
FLEM
Fuel use Comparison B737-400
0
2000
4000
6000
8000
10000
0 500 1000 1500 2000 2500 3000
Distance [km]
F
u
e
l

u
s
e

[
k
g
]
outbound
inbound
FLEM
73
For a single long-haul B747-400 flight, the FDR (altitude versus distance) data has been
recorded and can be compared with a FLEM generated flight profile. The results are depicted
in figure 10.13. Since the FDR payload (factor) is not known, a comparison has been made
with FLEM results for two different payload factors. From figure 10.13 it follows that, taken
into account the uncertainty in the load factor, the FLEM computation of the flight profile
matches recorded data rather well.
Figure 10.13 FDR and FLEM flight profile comparison for B747-400 flight
Note: I am not able to properly handle figure 10.13! A version is required that can be normally
handled as a powerpoint picture!
0
4
8
12
0 100 200 300 350 1350 2350 3350 4350 5400 5600 5800
47-400 FDRB7
FLEM-5 loadfactor 0. 7
FLEM-5 loadfactor 0.45
Distance [km] (not to scale)
A
l
t
i
t
u
d
e
[
k
m
]
74
11 Atmospheric impact models (OATI, CTMK, ENVI)
11.1 Introduction and overview
The assessment of atmospheric impacts in the AERO-MS involves three main steps, which
are related to the computation of:
emissions;
concentrations;
environmental impacts.
In relation to the computation of emissions, in the AERO-MS a distinction is made between
the computation of aircraft emissions and emission from ground sources. Aircraft emissions
are computed by the Flights and Emission Model (FLEM), as described in the previous
chapter. Ground emissions are computed by the Other ATmospheric Immissions model
(OATI). The purpose of the OATI model in the overall AERO-MS is twofold. First of all the
consideration of ground emissions allows for the development of aircraft emissions to be put
into perspective (i.e. how does the development of aircraft emissions relate to the overall
development of emission levels). Secondly, it allows the overall effects of emissions on
concentrations and environmental impacts to be estimated.
The effect on concentrations is computed through the Chemical Tracer Model of the Royal
Netherlands Meteorological Institute (KNMI), which is referred to as CTMK. The main output
of CTMK is CO
2
, NO
x
and ozone concentrations. Previously a simple representation of CTMK
was used as the original model was too large and took too long to run. However, modern high
speed PCs and software allowed for the embedding of the complete CTMK-model (version
1999) in the AERO-MS.
The environmental impacts computed are UV-radiation and radiative forcing (global warming
potential). These impacts are computed through the Environmental Impact Model (ENVI).
An overview of the atmospheric models in context is provided in figure 11.1.
Figure 11.1 Atmospheric models in context
As indicated in figure 11.1, the aircraft emissions computed in FLEM include: NO
x
, CO
2
, C
x
H
y
,
SO
2
, CO and H
2
O. The aircraft emissions are computed on an annual basis in kg per grid cell.
The default schematisation of FLEM is based on a three-dimensional grid structure,
CO2
NOX
(CxHy)
(SO2)
(CO)
(N2O)
UV-
radiation
Radiative
forcing
Other Atmospheric
Immisions Model
(OATI)
Other Atmospheric
Immisions Model
(OATI)
Flights and
Emission Model
(FLEM)
Flights and
Emission Model
(FLEM)
Chemical Tracer
Model KNMI
(CTMK)
Chemical Tracer
Model KNMI
(CTMK)
Environmental
Impact Model
(ENVI)
Environmental
Impact Model
(ENVI)
Emissions Concentrations Impacts
Surface emissions
Aircraft emissions
CO2
NOX
(CxHy)
(SO2)
(CO)
(H2O)
grid
conversion
NOX
O3
CO2
AERO-MS
75
consisting of a horizontal grid of 5 by 5 latitude/longitude and 15 vertical layers of 1 km.
The emissions computed in OATI include: NO
x
, CO
2
, C
x
H
y
, SO
2
, CO and N
2
O. All emissions
are also computed on an annual basis in kg per grid cell. The OATI model is based on a
horizontal grid of 1 by 1 latitude and longitude. Emissions computed in OATI are converted
to the first vertical layer of the FLEM schematisation (between 0 and 1 km from ground
surface).
Concentrations are computed in the model CTMK, which computes the concentrations of
NO
x
, ozone and CO
2
in ppmv (parts per million volume of air) per grid cell. Due to the long
residence time of CO
2
(of the order of 100 years), the computation of the CO
2
concentrations
is based on complete mixing, leading to a relatively simple computation. Inputs to CTMK are
the emissions of NO
x
, CO and CO
2
provided by OATI and the aircraft emissions NO
x
and CO
2
provided by FLEM.
CTMK operates on a three dimensional grid structure; horizontally, 36 cells are distinguished
in longitude direction and 24 cells in latitude direction, from pole to pole, whereby cells at the
poles have half the latitude of a regular cell. The vertical distribution is in 19 layers which are
defined by pressure-levels, covering the same range as the 15 layers of FLEM, but with a
different distribution.
ENVI uses the CTMK results in terms of ozone and CO
2
concentrations and computes two
different types of environmental impacts, i.e. the differences in radiative forcing (due to
differences in both ozone and CO
2
concentrations) and the differences of the content of the
ozone column (in Dobson units) and resulting differences in effective UV-radiation. Both
radiative forcing and effective UV-radiation are computed by horizontal CTMK grid cell (i.e.
the entire vertical CTMK column by horizontal grid cell). Therefore, ENVI only uses a two
dimensional grid which coincides with the horizontal grid of CTMK.
The following sections 11.2 through 11.4 further describe the computation principles and
implementation aspects of each of the models OATI, CTMK and ENVI, respectively.
11.2 Other Atmospheric Immissions Model (OATI)
11.2.1 Overview of OATI
OATI computes the atmospheric emissions from ground sources. By definition, these
emissions take place in the first (ground) layer of the schematisation of the atmospheric
system. The purpose of OATI is to determine the present and future emissions of all sources
other than aviation in order to put the aviation emissions into perspective and to provide a
complete set of inputs to CTMK. Figure 11.2 provides an overview of the OATI computation
scheme in relation to other models and databases.
Source data for OATI were obtained from the EDGAR database (based on 1990). EDGAR
stands for Emission Database for Global Atmospheric Research. The database was
developed by the Netherlands National Institute of Public Health and Environmental
Protection (RIVM) and contains a global overview of emissions from human sources and
selected natural sources. EDGAR data are available for various substances, 4 source types
and 11 sub-sectors on a 1 by 1 horizontal grid.
The EDGAR data were processed to obtain the input data for the OATI substances: NO
x
,
CO
2
, C
x
H
y
, SO
2
, CO and N
2
O and seven combinations of aggregated sectors and source
types. In the OATI model the emission data from EDGAR can be modified by scenario
assumptions to reflect future developments. Other inputs are obtained from the surface
competition module in model ADEM, which determines changes in land transport as a result
of possible scenario developments and/or measures affecting the transport volume of
aviation. Such changes in land transport (expressed in respectively rail passenger-kilometres,
76
car vehicle-kilometres and truck-kilometres) are converted by OATI to changes in emissions
by the land transport sector.
Figure 11.2 OATI computation scheme
Outputs of OATI in terms of emissions are converted to the FLEM grid (5 by 5) for
presentation purposes. The outputs in terms of NO
x
, CO and CO
2
emissions are further
processed by CTMK in terms of concentrations (NO
x
, ozone and CO
2
) to be used by ENVI for
the assessment of a number of environmental effects.
The main processing steps considered in OATI are:
1. One-time processing of EDGAR data in order to generate Base case emissions by
substance and sector/source for each OATI grid cell (1 by 1).
2. Aggregation of OATI grid cell emissions by substance and sector/source for the base year,
by IATA region.
3. Generation of changes in emissions from ground sources for future years, by IATA region.
4. Computation of surface competition factors to be used for updating future land transport
emissions by OATI grid cell (Note: the changes from aviation to surface transport are
determined by ADEM).
5. Update of future OATI emissions by OATI grid cell and aggregation across sector/source.
6. Conversion of OATI emissions from OATI to FLEM grid.
11.2.2 OATI input data
In addition to spatial information to convert the OATI grid schematisation to the FLEM grid
and to IATA regions, the main input data required for OATI include:
emissions from ground sources by substance, sector/source and IATA region;
emission factors by substance, related to various land transportation types.
All of these inputs are set up as scenario variables in the AERO-MS. This means that, in
Presentation of
emissions:
- tables
- FLEM grid
Presentation of
emissions:
- tables
- FLEM grid
ADEM
Changes of:
- rail pax kilometres
- car vehicle kilometres
- truck kilometres
ADEM
Changes of:
- rail pax kilometres
- car vehicle kilometres
- truck kilometres
Volume growth
surface competition
OATI
- present and future
emissions from
other sources
- changes of land
transport emissions
due to substitution
- conversion of
emissions to
FLEM grid
OATI
- present and future
emissions from
other sources
- changes of land
transport emissions
due to substitution
- conversion of
emissions to
FLEM grid
Inputs to CTMK
- NO
x
emissions
- CO
2
emissions
Inputs to CTMK
- NO
x
emissions
- CO
2
emissions
EDGAR
Emissions (1990) by:
- substance (N
2
O, CO
2
,
SO
2
, CH
4
, CO, NO
x
)
- four source types
- ten sub-sectors
EDGAR
Emissions (1990) by:
- substance (N
2
O, CO
2
,
SO
2
, CH
4
, CO, NO
x
)
- four source types
- ten sub-sectors
Measures and
Strategies
Measures and
Strategies
Scenarios:
- demographic
- economic
- technical
Scenarios:
- demographic
- economic
- technical
Volume growth
emission factor
77
addition to the data values for the Base case, further specifications are required to reflect
future situations. Base values for the emissions from ground sources directly follow from the
processing of information from the EDGAR database. From the base information, scenarios
have been developed for a number of different future situations. Global trends by substance
and sources were derived from scenarios developed by the Intergovernmental Panel on
Climate Change (IPCC, 1992). Regional variations (across IATA regions) were based on
existing scenarios of the Netherlands Central Planning Bureau (CPB, 1992).
Base and future data on emission factors by substance and land transportation type were
mainly based on Netherlands sources, which are considered to be representative for existing
and future western technology. From this information estimates were made for other, less
developed, areas in the world.
11.3 Chemical Tracer Model of KNMI (CTMK)
The global three-dimensional chemistry transport model (CTMK) was developed by the Royal
Netherlands Meteorological Institute (KNMI) to study the effects of aircraft emissions on
tropospheric NO
x
and ozone (O
3
) concentrations. In earlier versions of the AERO-MS, a
simplified version of the CTMK model was made and included in the overall system, since
initially it was not possible to run CTMK in a PC-environment. Because of the availability of
modern high-speed PCs, in the final phase of the AERO project it was decided to integrate
the full version of CTMK, as developed by KNMI, in the AERO-MS.
The CTMK model as it is integrated in the AERO-MS includes a conversion step from the
FLEM grid to CTMK grid. Furthermore a relatively simple computation of CO
2
concentration is
included in the model. The CTMK version in the AERO-MS is therefore executed through the
following steps.
1 Conversion of the FLEM (aviation) and OATI (ground sources) emissions (CO
2
and NO
x
)
from FLEM grid to CTMK grid (horizontal and vertical conversion).
2 Computation of the CO
2
concentration as an average over the year and homogeneous in
all CTMK grid cells.
3 Computation of NO
x
and ozone concentrations for the winter and summer (represented by
values for the months January and July) by CTMK grid cell.
Emissions from FLEM and OATI are available to CTMK on the default 3D FLEM
schematisation: a 5 by 5 horizontal grid and 15 uniform layers of 1 km in vertical direction.
The CTMK-model operates on a horizontal grid of 10 (longitude) by 7.83 (latitude). The cell
schematisation at the pole consists of a single 'cap' of half the size of a regular latitude cell
and no distinction in longitude direction. This leads to a latitude cell dimension of 180 divided
by 23 = 7.83 (22 cells of regular latitude size and 2 cells of half latitude size). The vertical
schematisation consists of 19 layers of which the boundaries are expressed in pressure
levels. The boundaries of cells in a vertical column are dependent on the pressure level at
ground surface, which may be different for each horizontal grid cell, as the variation in ground
surface elevation is taken into account.
The conversion of emissions on FLEM-grid to CTMK-grid involves 2 steps: a horizontal
conversion and a vertical conversion. In the horizontal conversion step, the contents of FLEM-
format grid cells in terms of emissions by substance (originating from FLEM or OATI) are
allocated to CTMK-grid cells by ratio of the overlapping areas. In vertical direction, emissions
are allocated from FLEM to CTMK layers by ratio of altitude levels (in m). This requires a
'translation' of altitude levels into the CTMK vertical pressure levels (which are different for
each horizontal grid cell, depending on its surface elevation). This translation is based on the
78
ICAO standard atmosphere as used in FLEM.
Changes of CO
2
-concentrations between the base year (1992) and a future year are based
on the (changing) CO
2
-emissions as computed by FLEM (aviation CO
2
-emissions) and OATI
(CO
2
-emissions from ground sources). The modelling of the changes of CO
2
-concentrations
in CTMK is based on the assumption that, since CO
2
has a very long residence time
(estimates range from 50 to 200 years), there is a complete mixing resulting in a globally
homogeneous CO
2
distribution. This implies that a single concentration value is computed for
all CTMK-grid cells. Furthermore it is assumed that, given the emissions of CO
2
in two
different years (base year and the forecast year), the annual emission change rate over the
period between these two years is a constant.
When the future concentration of CO
2
is to be computed, an assumption has to be made with
respect to the residence time of CO
2
. Within realistic bounds, the residence time of CO
2
used
in CTMK can be specified by the user (providing the homogeneity assumption, which is
associated with a relatively long residence time, is not violated). The default value in CTMK is
set to 125 year. In 1992 the concentration CO
2
was 355 parts per million volume parts of air
(IPCC, 1992). On the basis of the scenario used in the AERO analysis (scenario AERO-M
2010), for the year 2010 a CO
2
concentration of 380 parts per million volume of air (ppmv) is
computed. The concentration of 380 ppmv per year for CO
2
in the year 2010 means an
average increase of 1.4 ppmv per year over the period 1992 - 2010. Compared with the
IPCC-scenario of 1.5 ppmv per year (IPPC, 1992) this is slightly lower. Hereby it has to be
emphasised that the results are very sensitive to the chosen value for the residence time.
The processes regarding NO
x
and ozone are much more complex than those for CO
2
. Since
the residence times are much shorter compared with CO
2
(10-20 days for NO
x
and about 60
days for ozone), homogeneity cannot be assumed and transport phenomena have to be
considered. Also these substances have very complex chemical pathways associated with
them.
The CTMK model, as developed by KNMI and integrated in the AERO-MS, takes these
aspects into account, and uses databases with climatological data, including e.g. prevailing
wind directions, information on lightning (which generates NO
x
and ozone) and cloud
information (Wauben et al, 1995). CTMK uses analysed meteorological data to drive the
transport processes in combination with a computationally-efficient chemistry module, to
perform model simulations across several years. Therefore, the seasonal and spatial
dependence of the NO
x
and ozone perturbation caused by the aircraft NO
x
emissions can be
studied in detail. The natural variability of the concentrations of NO
x
and ozone are also
considered. The model results make it possible to judge whether the effect of aircraft NO
x
emissions on atmospheric NO
x
and ozone are distinguishable from the natural variations. For
a more elaborate description of the complex interactions modelled in CTMK, reference is
made to the KNMI documentation (Wauben et al, 1997).
Three emission sources are an input to CTMK for the computation of NO
x
and ozone
concentrations: i) aircraft emissions; ii) surface emissions; and iii) lightning. The first two are
depending on scenarios and measures, and are computed by respectively FLEM and OATI.
Annual emissions from lightning are assumed to remain constant over time. For the
computation of the effect on ozone concentrations, the emissions of NO
x
and CO need to be
specified, as these substances play an important role in the ozone chemistry. CTMK uses
calculated NO
x
emissions from aviation (FLEM) and ground sources (OATI). With respect to
CO emissions only ground emissions (based on OATI results) are taken into account in
CTMK as the total amount of CO emissions from aircraft engines during flight is assumed to
be negligible compared to the emissions from ground sources (see also table 2.1 in chapter
2).
CTMK is run based on the NO
x
and CO emissions for a period of 2 years. The first year is
used to initialise the model. The chemical initialisation of the transported tracers has been
derived from a CTMK run with an integration time of several years using a horizontal
79
resolution of 8 x 10, which in turn has been initialised with distributions obtained from a 2D
model. The results of the second year, in terms of concentrations of NO
x
and ozone per
CTMK grid cell, are computed for January and July. Both January and July are considered
because the seasonal differences in transport and photochemistry are reflected in large
differences between the results for these months.
11.4 Environmental Impact Model (ENVI)
11.4.1 Overview of ENVI
An overview of the ENVI computation scheme in relation to other models is provided in figure
11.3.
Figure 11.3 ENVI computation scheme
ENVI computes effective ultraviolet (EUV) doses and changes in radiative forcing (RF)
intensities for vertical columns of the CTMK schematisation, using the CO
2
concentrations
and ozone distributions generated by CTMK. Hence, results are produced for the horizontal
CTMK grid.
The main inputs are the CTMK ozone concentrations by CTMK grid cell, the global CTMK
CO
2
concentration and the bottom pressure at surface level by horizontal grid cell (taking into
account the actual average surface elevation by horizontal grid cell).
In the computation of both EUV and the changes in RF, seasonal effects are included. These
seasonal effects are taken into account by considering a typical winter (January) and summer
(July) situation.
11.4.2 Processing steps in ENVI
The model ENVI performs its tasks in three steps:
1 Computation of air mass per CTMK grid cell.
2 Computation of annual EUV and ozone mass per CTMK vertical column (horizontal grid
cell) for the base year.
3 Computation of annual EUV and changes in radiative forcing intensities per CTMK vertical
column for future situations.
ENVI
1 computation of air
mass per CTMK grid cell
2 Base case: computation
of annual EUV and ozone
mass per CTMK column
3 Datum/Forecast:
computation of annual
EUV and RF per CTMK
column
ENVI
1 computation of air
mass per CTMK grid cell
2 Base case: computation
of annual EUV and ozone
mass per CTMK column
3 Datum/Forecast:
computation of annual
EUV and RF per CTMK
column
CTMK
- O
3
concentrations
- CO
2
concentrations
- bottom pressure
CTMK
- O
3
concentrations
- CO
2
concentrations
- bottom pressure
External data
- monthly EUV data
- EUV rates of change
- conversion factors
- sensitivity distribution
arrays
External data
- monthly EUV data
- EUV rates of change
- conversion factors
- sensitivity distribution
arrays
Outputs
effective UV
radiative forcing
Outputs
effective UV
radiative forcing
80
(1) Computation of air mass
For the purposes of ENVI, the mass of air is defined as the amount of air in kilograms per
square metre of surface area. It is calculated for every CTMK grid cell based on bottom
pressure and the fraction of the entire vertical air column present in the associated grid cell.
The bottom pressure is made available by CTMK for all grid cells at the lowest CTMK level.
This is a one-time computation (for the base year only).
(2) Computation of reference amounts of annual EUV and ozone mass
Reference amounts of annual EUV and ozone mass are calculated for the base year. ENVI
uses monthly reference EUV values available as external data from satellite observations for
the winter (January) and summer season (July). These data only show significant variation
with latitude. Annual reference amounts follow from the two available monthly EUV values (for
summer and winter), assuming a fixed ratio between the annual EUV value and the sum of
the two monthly values. This ratio also varies with latitude and is available to ENVI as external
data.
Reference ozone mass follows from the CTMK ozone concentrations computed for the base
year. Ozone concentrations are converted into mass fractions (yielding kilograms of ozone
per kilograms of air) and multiplied by the air mass calculated in step 1. This is done for all
CTMK grid cells and for winter and summer, as CTMK ozone concentrations also vary with
season. Ozone masses are summed per CTMK column and converted to Dobson units, using
a conversion factor. One Dobson unit of ozone is defined as the amount of ozone present in a
gaseous layer of thickness 0.01 mm, under a pressure of 1 atmosphere and at a temperature
of 273 K.
(3) Computation of annual EUV and RF for future situations
Based on computed CTMK ozone concentrations for future situations, future ozone mass in
Dobson units is found in the same way as the reference amount, described above. For the
winter and summer season, the change of ozone mass can then be calculated by subtracting
the reference amount of ozone from the future amounts. The changes of ozone mass are
converted to a change in EUV for both seasons, again using external data. This change is
added to the reference values of EUV computed in step 2. Annual future EUV amounts (in
Joules per square metre) are derived from the two seasonal monthly values in the same way
as described in step 2.
Radiative forcing (RF) intensities, expressed in Watt/m
2
, can be interpreted as the net
downward radiation at the tropopause. As radiative forcing is affected by a great number of
variables, the computations in ENVI only reflect the change in radiative forcing as a result of
the changes in ozone mass and CO
2
concentrations. Computations of the change in RF are
made for future situations, relative to the base year, for the summer and winter situation
separately. Changes of RF are determined using so-called sensitivity curves. Sensitivity
curves represent a set of ratios by CTMK grid cell and season that provide RF changes per
unit of concentration change, both for CO
2
and ozone. Such sensitivity curves are available to
ENVI as external data. The changes of RF are simply computed by taking the difference
between the CTMK concentrations in the future situation and the base year and multiplying
these differences by the corresponding sensitivity ratio. For CO
2
this is done for the CTMK
grid cells at the lowest CTMK level only (i.e. the horizontal grid cells). For ozone, this is done
for all vertical layers of the CTMK grid.
11.4.3 ENVI input data and computational results of ENVI
The relevant ENVI input variables are:
Monthly EUV values for the summer and winter situation as a function of latitude.
Change in EUV per ozone mass change of one Dobson unit for the summer and winter
situation as a function of latitude.
Ratios of EUV by season (summer and winter) to year as a function of latitude.
81
Fraction of total air column by altitude layer.
Sensitivity curves relating changes in radiative forcing to changes in ozone and CO
2
concentrations.
All input variables were supplied by the Royal Netherlands Meteorological Institute (KNMI). In
fact ENVI reflects the results of a number of atmospheric data and relationships based on
extensive meteorological research by KNMI. The above, existing, databases and models
represent state-of-the-art knowledge in the respective fields. Hence, within the context of
AERO, the validity of the results provided by these 'source' databases and models proper was
taken as a starting point and no specific attempts were made to further calibrate or verify
these results. Validation of ENVI against these results was achieved by extensively testing
the model and by checking its outputs against the outcomes of KNMI computations.
Note: Still to be included figures to show the types of outputs of ENVI (this to be done if final
scenario is available):
A figure (world map) with annual EUV amounts (in Joules per square meter) for
both 1992 and 2010;
A figure (world map) changes in RF (in Watt/m
2
) between 1992 and 2010
because of changes in CO2 concentrations for both January and July.
References
Central Planning Bureau (CPB) (1992). Scanning the future - A long-term scenario study of
the world economy 1990-2015, Sdu Publishers, The Hague, 1992.
IPCC (WMO/UNEP) (1992). Climate Change 1992 - The supplementary report to the IPCC
scientific assessment.
Wauben, W.M.F, P.F.J. van Velthoven and H. Kelder (1995). Changes in tropospheric NO
x
and O
3
due to subsonic aircraft emissions, Koninklijk Nederlands Meteorologisch Instituut,
Scientific Report, WR 95-04.
Wauben, W.M.F, P.F.J. van Velthoven and H. Kelder (1997). A 3D Chemistry Transport
Model study of changes in atmospheric ozone due to aircraft NO
x
emissions.
82
12 The AERO integrated modelling system
12.1 Overview of AERO modelling system
The AERO-MS is an integrated system of interlinked models as described in the previous
sections. Figure 12.1 provides an overview of the overall system showing the various AERO
models in context.
Within the diagram, models are distinguished within three categories, i.e.:
human activities;
natural system;
environmental impacts.
From figure 12.1 it follows that most models in the AERO-MS relate to the description of the
human activities within the air transport sector (i.e. the models ATEC, ADEM, ACOS, DECI,
MECI, FLEM and OATI, including the Unified Database). The model CTMK describes the
natural (atmospheric) system, whereas the model ENVI describes a number of environmental
impacts following from changes to the atmospheric system.
Scenario's
demographic
economic
technical
UNIFIED
DATABASE
air transport
demand and
capacity
AVIATION
TECHNOLOGY
MODEL
(ATEC)
AIR TRANSPORT
DEMAND AND
TRAFFIC MODEL
(ADEM)
global demand
module
airline services
module
AVIATION
OPERATING
COST MODEL
(ACOS)
flight related
costs
investment
costs
OTHER
ATMOSPHERIC
IMMISSIONS
MODEL
(OATI)
DIRECT
(GLOBAL)
ECONOMIC
IMPACTS
MODEL
(DECI)
(NETHERLANDS)
MACRO
ECONOMIC
IMPACTS
MODEL
(MECI)
ECONOMIC EFFECTS
FLIGHTS AND
EMISSIONS
MODEL
(FLEM)
flight profile
calculation
module
emission
calculation
module
military
emission
module
ATMOSPHERIC
PROCESSES
DISPERSION
MODEL
(APDI)
3D dispersion of
substances
ozone depletion
and formation
ENVIRONMENTAL
IMPACTS
MODEL
(ENVI)
radiative forcing
UV radiation
NATURAL SYSTEM IMPACTS
1 2 3
EVALUATION OF COST EFFECTIVENESS
Measures
and
strategies
HUMAN ACTIVITIES
Figure 12.1 Overview of AERO Modelling System
The AERO-MS was designed to facilitate the evaluation of possible measures in the air
transport sector, within the context of different autonomous developments, to be reflected in
different scenarios. As shown in the diagram of figure 12.1, three different types of trade-offs
could be considered in the evaluation of measures, i.e. an assessment of the cost-
effectiveness based on:
83
1. the economic impacts versus emission reductions achieved;
2. the economic impacts versus changes in atmospheric concentrations;
3. the economic impacts versus changes in environmental effects.
The type of trade-off to be considered depends on the desired depth and scope of the
evaluation. In general, the model results become more complex and uncertain when moving
from the first to the third type of evaluation. Depending on the type of evaluation, the analysis
could be limited to using only a subset of the models included in the AERO-MS.
For example, if one is only interested in the economic impacts and impacts on fuel-use on an
international level, a calculation can be made without running the models MECI, FLEM, OATI,
CTMK and ENVI. In that case the relevant outputs can be obtained by running only the ATEC,
ACOS, ADEM and DECI models. In fact, running these four models is required for every
evaluation. If for example an analysis is to be made of environmental effects achieved, and no
specific economic information is required for The Netherlands, the model MECI can be
switched off before running the AERO-MS.
12.2 Model computations in Base, Datum and Policy case
The AERO-MS facilitates the model computations within each of the following three situations
(see figure 3.2):
1. the Base case, providing the best possible description of today's world;
2. a user-defined specification of a future situation, without emission reduction measures
taken in the air transport sector (the Datum case);
3. the user-defined future situation, with specific emission reduction measures taken in the
air transport sector (the Policy case).
The nature and purpose of some of the individual model computation varies, depending on
the above situation considered. An overview of the specific computations and main results of
the individual AERO models for each of the three situations (Base case, Datum case and
Policy case) is provided in table 12.1.
The results of the Base case follow from a once-through computation of all models in the
order as indicated in table 12.1, providing a unique reference situation for the Datum and
Policy case.
The Datum case is driven by the effects of autonomous developments, which affect fuel-use
and emission characteristics of aircraft, air transport demand, passenger fares and freight
rates. Consequently, the computation of all economic and environmental effects needs to be
updated. The order and nature of these computations is similar to the Base computations,
with certain additions, such as the price development of new aircraft in ATEC, the
computation of unit operating cost (per passenger or kg of freight transported) in ACOS, and
the change in global warming potential in ENVI.
More significant changes occur if measures are applied (the Policy case). An important
feature of the measures considered is that they generally lead to an increase in the cost of air
transport. The need to reflect the effects of these cost increases has a number of important
consequences, not only for the types of computations required, but also for the order in which
the computations take place. In table 12.1 the changes in the computation order, pertaining to
the models ADEM and ACOS, are indicated in the italic descriptions.
Following the computation of fuel-use, emission characteristics and aircraft purchase prices in
ATEC (see table 12.1), the next model to run in the Policy case is the model ACOS which
takes care of the computation of the changes in air transport costs brought about by the
imposed measures. This is done by re-computing the unit operating costs (the costs per
passenger and/or kg of freight transported) by flight stage, aircraft type and technology level
and by comparing these with the unit operating costs computed in the Datum situation.
84
Changes in unit costs across aircraft types and technology levels by flight stage are then used
to update the mix of aircraft used on the flight stage from the projected mix to be used without
the measure in place. This leads to a preliminary forecast of the flight mix operated on the
flight stage. In addition, a change in unit composite costs is computed for each flight stage,
where the unit composite costs reflect the unit costs (per passenger and/or kg of freight) as a
weighted average across the aircraft types and technology levels applied on the flight stage.
In the next step, the changes in aircraft operating costs, as reflected in the changes of unit
composite costs by flight stage, are used in the model ADEM to predict changes in fares, from
which a revised air transport demand is computed. In view of the important consequences of
this 'fare adjustment', and the relative uncertainty in the possible responses of airlines to
adjust fares (depending on the specific market situation), this computation is driven by
scenario assumptions which can be modified by the AERO-MS user. Aircraft flights by aircraft
type and technology level are computed in ADEM, based on revised demand, leading to an
update (final computation) of the preliminary flight mix operated by flight stage and the
change in unit composite costs, as originally computed by ACOS.
In addition to the computations in the Base and Datum case, the model DECI computes a
number of additional economic effects (on other actors), resulting from the changes between
the future situations without and with measures. These effects include the income from
charges generated by the 'government' (if charging measures apply) and the effects on
consumers in terms of changes in air-transport-related consumer surplus and consumer
expenditure.
Policy case computations in the remaining models (MECI, FLEM, OATI, CNMK, ENVI)
basically provide an update of the computations of the economic and environmental effects,
using the same principles as applied in the computations without measures.
12.3 Analysis capabilities of the AERO-MS
All modelling components, input data, user inputs and modelled outputs are embedded in a
software shell. The main functions of this shell are to facilitate communication and interaction
between the various models, and between the user and the AERO-MS. User interaction with
the AERO-MS is facilitated by allowing the user to define or modify the run specifications for
each of the three model situations, i.e. the Base case, the Datum case and the Policy case.
In relation to the specification of the Base case, the user has control over a variety of system
assumptions related to main modelling principles and schematisation aspects. These allow
the user to reflect best-available knowledge of the variables concerned, such as economic
system parameters (various elasticities, discount rate, depreciation rate) and technical system
parameters (specific modelling parameters and constants). This category of user controls is
referred to as 'assumption variables'. Table 12.2 shows a number of examples of important
assumption variables. Some of the assumption variables are factors and proportions, which
are applied in general, while some can be specified according to various distinctions like
passenger purpose or aircraft type. For a limited number of variables, distinctions on a
geographical level can also be made.
A great deal of flexibility is provided in allowing for a user-defined description of the Datum
case, to be expressed in different scenario specifications. Table 12.3 provides an overview of
the main scenario variables which can be specified by the user, falling within four major
categories, i.e. macro-economic, demographic, transport market and technological
development. As indicated in table 12.3, most of the scenario variables allow for a further
specification according to aircraft characteristics, traffic categories or other relevant
distinctions. In addition, in most cases there is the option to make geographical distinctions
according to IATA region or IATA region-pair (and in a few instances even at the flight stage
level).
85
Table 12.1 Overview of AERO computations in Base, Datum and Policy case
Models in AERO-MS Specific computations and main results
Base case Datum case Policy case
ATEC
Aircraft technology
1. Fuel use and emission characteristics by
aircraft types and technology level
1. Fuel use and emission characteristics
by aircraft type and technology level
Aircraft purchase prices because of price
developments
1. Fuel use and emission characteristics by aircraft type
and technology level
Aircraft purchase prices because of possible measures
ADEM
Air transport demand and
traffic
2. Processing existing data on air transport
demand and flights (unified Database)
Assessment of passengers/ freight
transported, flights by aircraft type and
technology level, fares and freight rates
2. Translation of economic, demographic
and fare developments into air transport
demand (passengers/freight) forecast
Aircraft flights by flight stage, aircraft
type and technology level based on
projected demand
3. Translation of changes in costs into changes in fares to
produce revised air transport demand forecast
Aircraft flights based on revised demands
Final changes in unit composite costs
Final forecast of aircraft flights by flight stage, aircraft type
and technology level
ACOS
Aviation operating cost
3. Aircraft operating costs by aircraft type,
technology level and flight stage
3. Aircraft operating costs by aircraft
type, technology level and region pair
Unit operating costs (per passengers
and kg freight) by aircraft type,
technology level and flight stage
2. Aircraft operating costs by aircraft type, technology level
and region pair
Unit operating costs (per passengers and kg freight) by
aircraft type, technology level and flight stage
Change in unit operating costs (because of possible
measures)
Change in unit composite costs by flight stage
Preliminary forecast of aircraft flights by flight stage
DECI
Direct economic impacts
(global)
4. Airline operating costs, revenues, results
Airline related employment
Airline contribution to gross value added
Extent and composition of airline fleets
(by carrier group/IATA region)
4. As in Base case 4. Airline operating costs, revenues and results
Airline related employment
Airline contribution to gross value added
Extent and composition of airline fleets
Government income from charges (if applicable)
Changes in consumer surplus and consumer expenses
(by carrier group/IATA region)
MECI
Macro-economic impacts
(Netherlands only)
5. Contribution to employment and gross
value added related to air transport to and
from the Netherlands
5. As in Base case 5. As in Base case
FLEM
Flights and emissions
6. Fuel use and emissions (CO2, NOx, SO2,
CxH, CO, H2O) in 3-dimensional space (5
by 5 horizontal grid cells and 15 equidistant
altitude bands of 1 km)
6. As in Base case 6. As in Base case
OATI
Other atmospheric immissions
7. Emissions of CO2, NOx, SO2, CxH, CO,
N2O by horizontal (1 by 1) grid
7. As in Base case 7. As in Base case
CTMK
Chemical Tracer Model KNMI
8. Concentrations of CO2, NOx and O3
(36x24 horizontal grid cells and 19 layers)
8. As in Base case 8. As in Base case
ENVI
Atmospheric (environmental)
effects
9. Effective UV radiation 9. Effective UV radiation
Change in global warming potential
9. As in Datum case
86
This flexibility allows the 'construction' of coherent scenarios based on a thorough understanding
of global developments in general and the air transport sector in particular. In practice, the usual
approach is that a limited number of - generally agreed-upon - scenarios are developed
beforehand, which are then consistently used in the analysis. The range of scenario specification
options provided in the AERO-MS will ensure that agreed-upon specifications can be actually
implemented, which can be achieved in various ways. For example, if sufficient information is
available, air transport demand could be derived from a detailed specification of population, GNP,
export growth, etc. However, if only a total projection of air transport demand is available this
could also be directly reflected in the scenario specification. In the latter case, some of the more
detailed scenario variables would not be used. During analysis execution, the flexibility in defining
scenario variables offered can be used to test the effects of specific scenario assumptions, or to
make selective amendments to pre-defined scenarios, if the user feels that there are good
reasons to do so.
Table 12.4 provides an overview of emission-reduction measures that can be considered in
defining the Policy case in the AERO-MS. Measures are considered within a number of
categories, including economic, regulatory, operational and other measures. Similar to the
options provided in the scenario specification, most measures allow for a further specification
according to aircraft characteristics, traffic category or other relevant distinctions. In addition,
for a considerable number of measures (including nearly all economic measures) there is the
option to make geographical distinctions according to IATA region or IATA region-pair. This
allows for a variety of possibilities with respect to the regional application of measures.
Possibilities (for example) are to apply measures only to flights departing from Annex B
countries (i.e. countries who signed up to the reduction of national emissions under the Kyoto
Protocol) or flights departing from the European Union (EU). With respect to a fuel taxation,
the AERO-MS also allows the user to specify a taxation on only domestic traffic or only
international traffic.
Furthermore, for a number of measures the following two types of discrimination can be
activated:
Apply only to EU carriers;
Apply only to flight stages to and from Amsterdam.
In addition to the analysis of individual measures, the AERO-MS allows for the analysis of
combinations of measures. This also creates the opportunity to combine certain measures
with logical responses of the various actors involved. For example, in response to a fuel
taxation applying to specific geographical regions, the tankering behaviour of airlines is likely
to change. If more tankering takes place, part of the cost increases imposed by the measure
(potentially leading to adjustments of aircraft technology and reduction of air transport
demand growth) would be circumvented, while fuel-use and aircraft engine emissions would
increase due to the extra fuel weight carried. This would in fact considerably decrease the
effectiveness of a regional fuel taxation.
Another example is the application of technology-related charges (charging only older-
technology aircraft) in order to induce airlines and manufacturers responses, leading to
accelerated technology improvement. Such responses could be simulated by the addition of
specific measures, such as fuel efficiency improvement or emission stringency measures, or
the premature retirement (scrapping) of old aircraft.
Finally, within the measures considered in the AERO-MS, various possibilities are included to
specify revenue neutral charging systems or to 'redistribute' any revenues from charges within
the air transport sector (the so-called re-channelling options). Revenue neutrality could be
achieved by increasing charges for aircraft of relatively old technology while at the same time
reducing charges for aircraft of relatively modern technology. Re-channelling could for
example be achieved by providing tax credits, buying out old aircraft by government (as
reflected in the scrapping measure), or by different forms of 'negative' taxation.
87
Table 12.2 Examples of assumption variables specifying the Base case
Description of assumption variables by category Unit Specification Spatial detail
Schematisation parameters
Size of the grid cells in altitude direction Meters
Size of the grid cells in latitude direction Degrees
Size of the grid cells in longitude direction Degrees
Minimum number of annual flights required for processing Flights pa
Economic system parameters
Passenger fare elasticity Elasticity
Passenger purpose IATA region pair
Elasticity of demand for aircraft with respect to capital cost Elasticity
Aircraft type, aircraft purpose
Proportion of labour costs in total maintenance costs Proportion
IATA region pair
Average fare per kilometre paid by passengers by purpose and mode 1992US$/km Business / non-business passengers, transport mode
Factor to adjust airline revenue computation for scheduled passengers fares Factor
Technical system parameters
Reference emission factor change Factor Aircraft type, emission type
Annual number of block hours worked by flight crew and cabin crew staff Hours pa IATA region
Realised market share of cargo demand transported by truck mode Proportion
Average speed for the entire journey km/hour Transport mode
Relative humidity Factor
Factor on emission index to make specific aircraft type representative for all aircraft
with the aircraft class
Factor Aircraft type, thrust, emission type
Spread parameter for choice between air mode and rail modes for surface
competition
Utility units/hour Business and non-business passengers
88
Table 12.3 Variables to be potentially included in AERO scenario specification (Datum case)
Description of scenario variables by category Unit Specification Spatial detail
Macro-economic development
Crude oil price (for aviation fuel computation) US$/kg IATA region
Real interest rate used in computation of finance charges % per annum IATA region
Autonomous passenger demand growth % per annum Business and non-business passengers IATA region pair
Autonomous cargo demand growth % per annum IATA region pair
Growth in general aviation flights % per annum IATA region pair
Per capita GNP growth % per annum IATA region
Per capita export growth % per annum IATA region
Passenger fares US$/seat Passenger class (first/business, economy, discount) Flight stage/IATA region pair
Freight rates US$/kg Flight stage/IATA region pair
Demographic development
Population growth % per annum IATA region
Transport market development
Aircraft utilisation rates Hours per annum Aircraft type, aircraft purpose (passengers/combi and freighter),
technology level
Cabin and flight crew salaries US$/hour IATA region
Airport charges US$/cycle Aircraft type, technology level IATA region
Route charges US$/km Aircraft type, technology level IATA region pair
Maintenance salary costs US$/hour Aircraft type, technology level IATA region
Profit adjustment factor Proportion Proportion of cost increase passed on to fares/freight rates IATA region pair
Ground transport competition Flag 0: no surface competition (testing purposes)
1: surface competition as present, not intensified by HSR
2: HSR competition for flight stages located outside the EU
3: HSR competition for flight stages within the EU
4: HSR competition for HSR Amsterdam, Paris, Frankfurt
Detour factor determining actual flight distance Factor Aircraft type IATA region pair
Volume related cost development Factor Scheduled/non-scheduled pax/combi aircraft and freighters IATA region pair
Technological development
Cabin crew needed Employees Aircraft type, technology level IATA region
Flight crew needed Employees Aircraft type, technology level
Maintenance requirements Hours per cycle and
flight hour
Aircraft type, technology level
Availability of new large aircraft (NLA) Flag 0: introduction of NLA; 1: no introduction of NLA
Specific aircraft fuel use % per annum Aircraft type
Change in specific aircraft engine emissions % per annum Aircraft type
Substance: CO2,NOx,SO2,CxHy,CO,H2O
Emission factors land transport kg/km and
kg/passengers
Transportation type
Substance: CO2,NOx,SO2,CxHy,CO,N2O
IATA region
Emissions from ground sources kg per annum Ground activity/sector
Substance: CO2,NOx,SO2,CxHy,CO,N2O
IATA region
89
Table 12.4 Potential measures to be considered in the AERO-MS (Policy case)
Description of measure variables by category Unit Specification Spatial detail
Economic measures
Fuel taxation by IATA region US$/kg IATA region
Fuel taxation by IATA region pair US$/kg IATA region pair
Fuel taxation for domestic flights US$/kg IATA region
Airport charges US$/cycle Aircraft type, technology level IATA region
Route charges US$/km Aircraft type, technology level IATA region pair
Cargo demand taxation US$/kg Taxation of cargo carried charged to airlines IATA region pair
Passenger demand taxation US$/seat Taxation of passenger carried charged to airlines
(by passenger class: first/business, economy, discount)
IATA region pair
Freight rate taxation US$/kg Direct freight rate taxation (not through airlines) Flight stage/IATA region pair
Passenger fare taxation US$/seat Direct passenger fare taxation (not through airlines) Flight stage/IATA region pair
Subsidising purchase of new aircraft % of new aircraft
price
Aircraft type, aircraft purpose, certification age
Regulatory measures
Non-operation ban of (old) aircraft Flag by IATA
region
Aircraft type, aircraft purpose, purchase or certification age
beyond which aircraft are banned
IATA region
Emission stringency Factor Increase in emission stringency by aircraft type, substance
(CO2,NOx,SO2,CxHy,CO,H2O) and (measure) year
Fuel technology improvement Factor Increase in fuel efficiency by aircraft type and year
Scrapping old technology aircraft Purchase or
certification year
Specification of purchase or certification age by aircraft type
and purpose beyond which aircraft are removed from fleet
Operational measures
Climb speed adjustment Factor Speed factor in climb phase by aircraft type and technology
level
Cruise speed adjustment Factor Speed factor in cruise phase (speed restrictions) by aircraft
type and technology level
Descent angle adjustment Factor Factor to change descent angle by aircraft type and
technology level
Flight altitude limitation 100 feet Maximum flight level IATA region pair
Lift-drag property adjustment Factor Factor to change lift-drag properties by aircraft type and
technology level
Other measures
Decrease in detour factor (air traffic congestion) Factor Factor to change actual flight distance by aircraft type IATA region pair
Changes in tankering behaviour Flag 0: no tankering
1: tankering for return flight in origin region
2: maximum tankering in origin region
-1: tankering for return flight in previous origin region
-2: maximum tankering in previous origin region
IATA region pair
12.4 User interface and presentation options of the AERO-MS
12.4.1 Overview of main functions of user interface
The main function of the user interface of the AERO-MS is to facilitate the communication and
interaction between the user and the various models. The user facilities which are
distinguished are listed below.
Specification of cases:
a Base case: base year data and assumptions;
a Datum case: the autonomous future development based on a scenario;
a Policy case: the future development including measures.
Computation of cases:
defining which cases to run;
defining which models to run;
running the models.
Analysis of an individual case:
select a case to analyse;
make a report;
show specific case results.
Evaluating cases:
select cases to evaluate;
make a scorecard of cases;
compare cases;
show cost effectiveness of cases.
A complete description of the user interface and its functionalities is provided in the User
Manual of the AERO-MS. In the following, a brief overview of the user-interface is given.
Figure 12.2 shows the main screen of the AERO-MS. The user-interface consists of the menu
bar at the top and the tabs at the left side of the main screen. These tabs relate to the main
user facilities as listed above.
Figure 12.2 Main screen of the AERO-MS interface
91
12.4.2 Specification and computation of cases
As can be seen in figure 12.2, the specification of cases tab consists of Define and View
buttons for assumptions, scenarios and policies. Assumptions relate to the Base case,
scenarios to the Datum case and policies to the Policy case. Furthermore, two buttons are
available for batch processing, which relates to the defining and starting of computational
runs.
The user-interface allows for a fast and controlled way of specifying and computing the effects
of many cases. In figure 12.3 an example is given of the specification of a policy. After using
the policy define button of the main screen of the AERO user-interface, the user enters the
policy wizard. In the left window of figure 12.3 the available policy variables are listed. The
right window shows the policy variables included in the existing policy. A policy variable can
be included in the new policy by dragging and dropping it to the right window. When a
variable is dropped in the right window, the user enters the so-called variable editor, and
values for the policy variable can be specified.
Figure 12.3 Policy wizard, specification of a policy case.
Any case specification can be verified by making use of the view button. If an incorrect
specification has been made, the specification can be easily revised.
Once a case is specified, the user can make a run. After pressing the New run button the
user enters the run wizard. In the run wizard the user can subsequently select one or more:
assumption sets (Base cases);
scenarios (Datum cases);
policies (Policy cases).
92
After the selection of cases, the user can specify which models to run. Figure 12.4 contains
the complete modelling system in a highly schematic form as presented through the user-
interface. The figure illustrates that the models ATEC, ACOS, ADEM and DECI are always
included in a run. The other models are optional. Some models need results of other models
as input, and can therefore not be selected separately. If, for example, the user selects the
model ENVI, the models FLEM, OATI and CTMK (APDI) are also automatically selected. In
this respect it should be noted that in the user interface representation, the model CTMK is
still referred to by the 'old' name APDI (standing for "Atmospheric Processes and Dispersion
Model") which was used to indicate an earlier repro-version of CTMK.
Figure 12.4 Run wizard: overview of models to include in a run
12.4.3 Analysis and evaluation of cases.
Before the analysis and evaluation of cases can start, the user has to select one or more
cases. Under the analysis button, the user can look at the detailed results of an individual
case. For this purpose, a standard output report is available. In addition, the user can design
his personal report based on a selection from available output variables. All output reports
can be exported to a spreadsheet for further processing. The standard output report contains
computational results in the following main categories:
aviation activity;
passenger activity;
cargo activity;
Schiphol-related demand;
direct economic effects;
required fleet;
effects on Dutch economy;
aircraft emissions;
ground emission;
land transport.
93
If the user wants to have a more detailed look at the computational results of individual cases,
he can inspect all the raw model outputs. In this respect it is noted that also intermediate
results can be viewed so the results of the subsequent computational steps can always be
verified. The data inspect option allows for the aggregation of data across any or all of the
dimensions of an output variable. This implies that the user can for example inspect the
number of flights on an individual flight stage, and also on various intermediate levels of
aggregation (IATA region-pairs or clusters of region-pairs) up to the total number of flights (for
all flight stages).
With regard to looking at results of individual runs, the user-interface allows for the
(geo)graphical presentation of computational results. A number of different types of graphs
that show results of variables are available in the AERO-MS. An example of the graphical
presentation of results is given in figure 12.5.
Figure 12.5 Graphical presentation of results in the AERO-MS
Evaluation refers to the comparison of computation results of various cases. One option is to
compare the effects of different scenarios (Datum cases), reflecting various possible future
developments without measures. Another option is to compare the effects of different policies
(Policy cases) relative to one scenario (Datum case). For the cases selected a standard
scorecard can be made. The standard scorecard is based on a further selection and
processing of the most policy-relevant outputs of the standard report. In addition, the user can
also design his personal scorecard layout. Both the standard and user specific scorecards
94
can be exported to a spreadsheet for further processing. The standard scorecard contains
computational results in the following main categories:
air transport and aircraft operation;
effects for airlines;
economic effects for other actors;
effects for the Netherlands;
fuel-use and emissions of aviation;
operating efficiency;
ground emissions.
The user-interface also allows for the presentation of the cost-effectiveness of various
policies. An example of this is presented in figure 12.6.
Figure 12.6 Presentation of cost-effectiveness of policies in the AERO-MS

You might also like