You are on page 1of 65

The CREDOS Project

Madrid Barajas Case Study D5-7



Abstract for Deliverable: Madrid Barajas Case Study



Contract Number: AST5-CT-2006-030837 Proposal Number: 30837
Project Acronym: CREDOS
Deliverable Title: Madrid Barajas Case Study
Delivery Date:
Responsible: INECO
Nature of Deliverable: Report
Dissemination level: Public
File Id N: CREDOS_410_ECTL_Ineco_D57_Madrid Barajas Case Study
Status: Approved version Version: 1.0 Date: December 2009

Approval Status
Document Manager Verification Authority Project Approval
Ineco EEC PMC
Alvaro Urech Andrew Harvey PMC members
L.Ballesteros, D.Prez,
L.Garca
WP5 Leader


1


Acronyms and definitions

ACRONYM

DEFINITION
Aena
Aeropuertos Espaoles y Navegacin Area
ANSP
Air Traffic Service Provider
ATM
Air Traffic Management
ATC
Air Traffic Control
ATCO
Air Traffic Controller
ATS
Air Traffic Service
CBA
Cost Benefit Analysis
CFMU
Central Flow Management Unit
CREDOS
Crosswind-Reduced Separation for Departure Operations
CROSSWIND
Crosswind in this document should be understood as the
crosswind component of a specific runway direction.
CTOT
Calculated Take Off Time (CFMU)
EMOSIA
European Model for Strategic ATM Investment Analysis
STAFOR
Air Traffic Statistics and Forecasts
ICAO
International Civil Aviation Organisation
IFR

Instrumental Flight Rules

METAR/SPECI
Meteorological Aerodrome Report
METAR is a format for reporting weather information. METAR
means "aviation routine weather report" and is predominantly
used by pilots in fulfilment of a part of a pre-flight weather
briefing.
It is also used by meteorologists, who use aggregated
METAR information to help forecast the weather. METAR
reports usually come from airports. Typically, reports are
2


DEFINITION ACRONYM

generated once an hour; however, if conditions change
significantly, they may be updated in special reports called
SPECI's.
NM
Nautical Miles
A measure used in navigation. The unit is equal to 1852 m.
NPV
Net Present Value
SID
Standard Instrument Departure (Route)
SID standard ATS routes identified in an instrument departure
procedure by which aircraft should proceed from take-off
phase to the en-route phase.
TMA Terminal Manoeuvring Area
A control area normally established at the confluence of ATS
routes in the vicinity of one or more major aerodromes.
WV
Wake Vortex
Also known preferably by ICAO as wake turbulence. Wake
turbulence is turbulence that forms behind an aircraft as it
passes through the air. This turbulence includes various
components, the most important of which are wingtip vortices
and jetwash. Jetwash refers simply to the rapidly moving air
expelled from a jet engine; it is extremely turbulent, but of
short duration. Wingtip vortices, on the other hand, are much
more stable and can remain in the air for up to two minutes
after the passage of an aircraft. Wingtip vortices make up the
primary and most dangerous component of wake turbulence.
Wake turbulence is especially hazardous during the landing
and take-off phases of flight, for two reasons. The first is that
during take-off and landing, aircraft operate at low speeds and
high angle of attack. This flight attitude maximises the
formation of dangerous wingtip vortices. Secondly, takeoff and
landing are the times when a plane is operating closest to its
stall speed and to the ground - meaning there is little margin
for recovery in the event of encountering a different aircraft's
wake turbulence




3

Executive Summary

The CREDOS concept consists in authorising a suspended wake turbulence time and
distance separation between certain pairs of departures depending on the strength and
direction of the wind. Compared to the actual wake turbulence separations of 2 or 3
minutes and 4NM, 5NM or 6 NM, the optimal separation minima target for the
crosswind application is then the corresponding minimum radar separation
This report is aimed at analysing CREDOS Concept implementation in Madrid Barajas
airport (LEMD), with reliable and real data as far as possible taking the current level of
development of the project, through expert experience, fast time simulations and data
offered from other similar studies. This deliverable provides a specific case study of the
CREDOS implementation from the point of view of Operation, Human Factor (HM) and
Cost Benefit Analysis (CBA).
The Operational study analyses the scenario at LEMD, including current traffic, wind,
and operational restrictions. A set of Fast Time Simulations were performed in order to
provide an estimation of the improvement in reduction of delay that the use of
CREDOS separations could bring to Madrid-Barajas. The main conclusion is that
current traffic figures do not account for a large improvement due to CREDOS as the
target runway is not saturated, although the reduction of delay during peak periods
could be enough to justify its implementation.
From the Human Factors point of view, the case study has analysed the possible
impact that the implementation of CREDOS could have on Madrid-Barajas, allowing
the identification of critical areas and the efforts needed for this integration. The main
conclusion of the study is that the main position affected by the implementation of
CREDOS would be the LCL (Local Controller), but without a significant impact.
The Cost Benefit Analysis (based on EUROCONTROLs EMOSIA methodology) takes
the inputs of the previous analysis and together with some assumptions in terms of
traffic evolution, costs, etc, provides some economic results like the net-present value
(NPV) or a sensitivity analysis. The main conclusion of the CBA is that with the
assumptions established, the NPV of the implementation of CREDOS separation in a
single runway at Madrid-Barajas is positive; also important is the identification of the
most sensitive variables like the demand growth expected.
Finally the methodology used in the case study is summarised in order to serve as
guidelines for other possible case studies of the implementation of CREDOS at other
airports who would wish to follow.

4

Table of Contents
1 INTRODUCTION 8
1.1 BACKGROUND 8
1.2 PURPOSE OF THE DOCUMENT 8
1.3 STRUCTURE OF THE DOCUMENT 8
2 CREDOS CONCEPT AND ARCHITECTURE 9
2.1 SYSTEM ARCHITECTURE NEEDED 9
3 OPERATIONAL STUDY 12
3.1 MADRID BARAJ AS: GENERAL OPERATION 12
3.2 CONFIGURATION UNDER STUDY 15
3.3 LOCAL WIND ANALISIS 17
3.4 SIMULATION TOOL 19
3.5 HYPOTHESIS 20
3.6 SIMULATION RESULTS 24
3.7 CONCLUSION 26
4 INTEROPERABILITY WITH CURRENT LEMD INFRASTRUCTURE 27
5 HUMAN FACTORS STUDY 29
5.1 INTRODUCTION 29
5.2 OBJ ECTIVES 29
5.3 APPROACH, METHODOLOGY, AND ACTIVITIES 29
5.4 MADRID-BARAJ AS TOWER: NEW ROLES AND RESPONSIBILITIES 30
5.5 LCL WORKLOAD ANALYSIS 38
5.6 HUMAN ERROR ANALYSIS 41
5.7 IMPLEMENTATION NEED ANALYSIS 46
5.8 CONCLUSIONS AND RECOMMENDATIONS 47
6 COST BENEFIT STUDY 49
6.1 METHODOLOGY APPROACH 49
6.2 STAKEHOLDERS 50
6.3 GEOGRAPHIC SCOPE 52
5

6.4 TIME PERIOD 52
6.5 BASELINE DO NOTHING OPTION. 53
6.6 APPROACH TO THE COST ANALYSIS 54
6.7 APPROACH TO BENEFIT ANALYSIS 55
6.8 ECONOMICAL STUDY 56
6.9 CBA CONCLUSIONS AND RECOMMENDATIONS 62
7 METHODOLOGY USED AND GUIDELINES 64
8 REFERENCES 65

List of tables
Table 1: Wind probabilities ........................................................................................... 17
Table 2: Local wind year 2007...................................................................................... 18
Table 3: Wind summary................................................................................................ 19
Table 4: Aircraft separations based on speed.............................................................. 22
Table 5: Wake Turbulence Separations ....................................................................... 23
Table 6: Specific Madrid Separations........................................................................... 23
Table 7: Actors and changes in role when using the CREDOS................................... 31
Table 8. Mental workload impact by CREDOS implementation. ................................. 40
Table 9: Human Error Analysis.................................................................................... 43
Table 10: HEART General Task Types Descriptions ................................................... 44
Table 11. HEART Error Producing Conditions Effects. ............................................... 45
Table 12: Stakeholders................................................................................................ 52
Table 13: Fuel burn and CO
2
Emissions....................................................................... 56
Table 14: Cost and benefits in economical terms......................................................... 57
Table 15: Parameter ranges for sensitivity analysis ..................................................... 60




6


List of Figures
Figure 1: CREDOS initial functional system architecture [4] ........................................ 10
Figure 2: Madrid Barajas layout.................................................................................... 12
Figure 3: Interferences due to missed approaches ...................................................... 13
Figure 4: Madrid Barajas traffic evolution..................................................................... 14
Figure 5: Madrid Barajas demand along the day.......................................................... 14
Figure 6: One day traffic mix in Madrid Barajas............................................................ 15
Figure 7: Madrid Barajas wind rose.............................................................................. 15
Figure 8: Missed approach limitation on runway 15R................................................... 16
Figure 9: Analysis scenarios......................................................................................... 21
Figure 10: Departure daily delay sensitivity.................................................................. 22
Figure 11: SID separation............................................................................................. 23
Figure 12: Average Delay per hour............................................................................... 24
Figure 13: Maximum Runway Throughput.................................................................... 25
Figure 14: Delay vs. Number of departures.................................................................. 25
Figure 15: Delay Reduction wrt Departures.................................................................. 26
Figure 16: LEMD Tower................................................................................................ 27
Figure 17: SACTA controller positions.......................................................................... 28
Figure 18: Departure operations under CREDOS ........................................................ 32
Figure 19. Local Controller Hierarchical Task Analysis under CREDOS...................... 37
Figure 20: EMOSIAs six steps..................................................................................... 49
Figure 21: CBA Timeline............................................................................................... 52
Figure 22: Madrid Barajas Traffic demand................................................................... 53
Figure 23: Net Cash Flow............................................................................................. 59
Figure 24: Sensitivity Analysis. Tornado diagram......................................................... 61
Figure 25: Probability Curve. Risk analysis .................................................................. 62
Figure 26: Case study methodology............................................................................. 64

7

1 Introduction
1.1 Background
The need to increase airport capacity is one of the major challenges facing ATM
research today. Such an increase could be achieved by reducing the current
separation minima while maintaining levels of safety.
ICAO separation standards for landing and take-off were implemented in the 1970s to
protect an aircraft from the wake turbulence of a preceding aircraft. However, research
has shown that the transport and persistence of wake vortices are highly dependent on
meteorological conditions, so that in many cases the ICAO standards are over-
conservative. By developing a full understanding of wake vortex behaviour in all
weather categories, separations could be reduced under certain suitable conditions.
The CREDOS project has studied the operational feasibility of this approach by
focusing on crosswind conditions. The aim of CREDOS is to prove that this approach
to reducing separations is valid, feasible and safe.
The main targets of CREDOS project include:
To evaluate the feasibility of a Concept of Operations allowing reduced
separations for Single Runway Departures under crosswind. This objective will
be reached through:
o A Safety Case (Functional Hazard Assessment to Preliminary System
Safety Assessment)
o A Human Factors Case
Analysis of expected benefits at selected European & North American airports
1.2 Purpose of the Document
The purpose of this document is to present the results of the analysis carried out by
INECO in the frame of the CREDOS project of the possible implementation of the
CREDOS Concept (CONOPS B [1]) at the Madrid Barajas airport
1.3 Structure of the Document
The document is organised in the following sections:
1. Introduction. This section, which offers general information about the document.
2. CREDOS concept and architecture. Brief introduction to the main lines of the
CREDOS concept applied in the analysis as well as the main elements that
would compose the CREDOS system at an airport.
3. Operational Study. General description of the Madrid-Barajas scenario to be
analysed as well as the results, in operational terms, that the implementation of
CREDOS would obtain.
4. Interoperability with current LEMD infrastructure. Some considerations on how
the implementation of CREDOS could impact the systems at Madrid Barajas.
5. Human Factors study. Analysis of the impact of CREDOS at Madrid Barajas
from the Human Factors point of view.
6. CBA. Cost Benefit Assessment and final conclusions of the study
7. Lessons learned: Guidelines for the elaboration of a similar study at another
airport
8. References
8

2 CREDOS Concept and Architecture
This section does not pretend to be a thorough description of CREDOS, but it is
included in the document for the sake of completeness. The full description of the
Concept of operation is described in its final version ConOps E, however, due to
schedule constraints motivated by the present study running in parallel to the actual
development of the concept of operations, ConOps B [1] was the version used for this
work.
The CREDOS concept consists in authorising a suspension of the wake turbulence
time and distance separation between certain pairs of departures depending on the
strength and direction of the wind. CREDOS has one high-level objective:
.
To temporarily increase runway throughput by safely suspending WV related
separation minima
The CREDOS concept is crosswind dependant. The application of reduced
separations is subject to being able to observe sufficient crosswind conditions which
should ensure that the wake turbulence will be carried away from the necessary part of
the initial climb path of the following aircraft.
The CREDOS concept is most beneficial when a runway is used in segregated mode
i.e. one runway is used for departures and another for arrivals. However CREDOS can
be applied between two consecutive departing aircraft if all CREDOS requirements are
met. CREDOS assumes as an example that when a MEDIUM aircraft departs behind a
HEAVY aircraft no additional wake turbulence separation in time or distance is needed
when crosswinds are supposed to transport any hazardous turbulence out of the track
of the following aircraft. Compared to the current wake turbulence separations of 2
minutes, the optimal separation minima target for the crosswind application is then
approximately 1 minute or the equivalent radar distance separation of 3 NM.
2.1 System Architecture needed
The CREDOS project requires an architecture which would consist in five new
elements and in the necessary interfaces to link these elements with the current ATC
system. The actual implementation will be airport-dependant, but in general terms all or
most of these elements will be included. These five new components [4] are:
CREDOS Separation Mode Advisory: It provides advice concerning the possible
suspension of wake turbulence separations (i.e. CREDOS wind conditions available or
not available), including the expected time for future mode transitions for each runway.
Such advice is based on meteorological forecast information (e.g. wind profile pictures,
METAR, TAF).
CREDOS Departure Planning Advisory: It provides up-to-date departure planning
information and advisory on how to optimise the departure sequence while CREDOS
wake turbulence separation suspension is applied, possibly using a Departure
Manager (DMAN).
CREDOS SID Wind Forecast Service: It estimates the expected wind conditions for the
departing aircraft along the applicable SID (including up or downwind indication) up to a
certain altitude. No HMI is foreseen; the output data is used for the CREDOS Go/No-go
indicator.
CREDOS SID Wind Monitoring Service: It monitors the wind conditions as experienced
by the departing aircraft along the entire SID. Information regarding the actual aircraft
positions is derived from surveillance radars and possibly flight data processing. The
9

monitoring information is updated in short intervals (e.g. 1 or 2 minutes). The output
data is used for the CREDOS Go/No-go indicator and the Departure Controller's HMI.
CREDOS Go/No-Go Advisory: It indicates whether or not for the next take off wake
turbulence separation suspension (CREDOS mode) is possible. Go or NO-GO
indication will appear on the HMI display depending if CREDOS is or not
recommended. The HMI display also needs to include the actual now-casted wind,
enabling the runway controller to cross check whether the crosswind conditions for
CREDOS are still met at take off clearance time. Based on wind now-cast data and
data from the SID Wind Forecast Service and Departure Planning Advisor, it provides:
A go/no-go indication for each aircraft ready to depart (ie with or without
CREDOS).
The actual surface wind for cross checking go/go-no indication against
crosswind.
These five elements appear in the yellow boxes below,

Figure 1: CREDOS initial functional system architecture [4]
To operate CREDOS it will be necessary to enhance part of the current Human
Infrastructure System (boxes in green in the figure above) which are described as:
TOWER AND APPROACH SUPERVISOR: He is responsible for obtaining and
monitoring weather forecast service. The tower Supervisor, in coordination with the
approach control Supervisor, would be the role who decides when CREDOS is in
operation or not. The Supervisors shall be aware of when CREDOS will be applied by
the Runway controller
GROUND CONTROLLER: He is responsible for executing the sequencing of departing
traffic. This task is based on knowledge about CFMU issued constraints per aircraft
(i.e. CTOT etc), present runway capacity and configuration and local constraining
factors such as closed taxiways, other traffic etc. The Ground controller shall know
10

when CREDOS is in use in order to integrate the potential benefit in the sequence
preparation work.
RUNWAY CONTROLLER: He is responsible for applying a safe and efficient departure
flow whereby he has to consider runway separation rules and local spacing
methodology in order to deliver aircraft to the radar Departure controller.
DEPARTURE CONTROLLER IN TMA: The management of the traffic transferred to
the Departure Controller using CREDOS separation is not significantly different as all
aircraft are still delivered applying standard radar or vertical separation as if they where
all MEDIUM.
The Meteo System will be enhanced as well. Three of its parts will be modified:
WIND MEASUREMENTS
Wind (speed, direction, gusts) could be measured with several anemometers, installed
along the runways at e.g. 10 or 15 m height. Different types of anemometers exist. Cup
anemometers, which are often used, consist of a mast (vertical axis) and three cups (or
propellers) which capture the wind
A Wind and Temperature Radar (WTR), is used for measurement of wind (from 100 m
to 1500 m height, with a precision of 0.5 m/s) and temperature (from 100 m to 1000 m
height, with a precision of 0.5 C/100 m).
AIRPORT WIND NOWCASTING
Accurate wind predictions for the whole airport area in the very short term are needed.
AIRPORT WIND FORECASTING
Wind forecasting information consists of METARs, TAFs, and basic forecasts of the
wind vector for a time period of six to twelve hours after generation. The COSMO-DE
model of the German Weather Service (DWD) could e.g. be used as basic forecasting
tool, if adapted to the airport area considered.
The architecture foreseen to implement CREDOS is described at a high level (it will
depend on local conditions) in deliverable D4.7 [4], from where the information above
has been extracted.

11

3 Operational Study
3.1 MADRID BARAJAS: GENERAL OPERATION
3.1.1 Layout
Madrid Barajas airport layout is composed of 4 runways used in single operation mode:
2 runways for departures and 2 runways for arrivals.
The airport operates in 2 possible configurations, North and South in which the
runways are used as follows:
North Configuration:
- departures: 36L and 36R
- arrivals: 33L and 33R
South Configuration:
- departures: 15L and 15R
- arrivals: 18L and 18R
These possible configurations are summarised in the following graph:


Figure 2: Madrid Barajas layout

3.1.2 Operation
Arrivals to the parallel runways are conducted in a dependent mode and departures on
the parallel runways are independent.
There are dependencies between departure runway 36L and arrival runway 33R due to
the interference of a possible missed approach on 33R whose path crosses departure
runway 36L.
The interference is controlled with a departure blocking area on the arrival runway 33R.
Any arrival within the defined departure blocking area inhibits any departure on runway
36L.
12


This procedure is also implemented in south configuration with runways 15R and 18L.



Figure 3: Interferences due to missed approaches
3.1.3 Traffic
Figure 4 below shows a sample of the traffic evolution in Madrid from the year 2001 to
year 2006. There is a significant increase of the number of operations through the past
years. This graph shows that the airport is suitable for the study given the important
number of operations.
13


Figure 4: Madrid Barajas traffic evolution
Figure 5 (below) shows a mix of capacity and demand at Madrid Barajas airport. As it
can be seen in the graph, there is a high demand during daytime.


Figure 5: Madrid Barajas demand along the day
The following figure shows the traffic mix in operation in Madrid Barajas at the moment
of the study. As it was decided in the beginning of the CREDOS project, the fleet mix is
used as is in the moment of the study without taking into consideration future
changes. That is why there is a category of 757 and 737-800 considered separately.

14

Heavy 11.85%
TurboProp
4.22%
B757;737-800
9.13%
Medium Jet
74.75%

Figure 6: One day traffic mix in Madrid Barajas
3.2 CONFIGURATION UNDER STUDY
As mentioned before, Madrid Barajas airport can operate 4 different runways (2 in each
configuration) in departure. The runway most exposed to strong crosswinds is
determined by analysing the wind rose of the airports wind conditions over the year
2007.

Figure 7: Madrid Barajas wind rose
15

3.2.1 Hypothesis
3.2.1.1 Operational
According to the wind conditions in Madrid, studied in the previous paragraph, the
runway most exposed to crosswinds is runway 15L operated in South configuration
only in single mode departure operation.
3.2.1.2 Credos
The study has assumed that 10kt crosswind is necessary to apply CREDOS
separations.
Reduce H/M and H/H from 120s to 60s when no procedural restriction.
Reduce H/M and H/H from 120s to 100s when SID restriction (for aircraft following on
the same or downwind SID and where 5 NM have to be applied either because of en-
route separations or because of the 5 NM separation required at a later phase than
initial climb phase).
3.2.1.3 Conclusion
Runway 15L will be studied as a single mode operation with no limitation due to
operations on other runways.
The data that will be used for the study is the real traffic on this runway, the traffic mix
and the separation applied (except blocking) are the particular SID restrictions.


Figure 8: Missed approach limitation on runway 15R

16


3.3 LOCAL WIND ANALISIS
3.3.1 Input data
To perform the case study of Madrid, it is necessary to evaluate the average frequency
of CREDOS wind conditions (10kt crosswind) in order to perform simulations that
recreate as closely as possible the real conditions in Barajas, and therefore evaluate
the possible benefits of CREDOS.
The most accurate information available is the METAR reports that are published every
30 min approximately.
From this information it is possible to make a gross estimation of the frequency of
CREDOS wind conditions.
The objective is to evaluate the frequency of CREDOS wind conditions when the RWY
15 is in use (South Configuration) using a probabilistic approach.

3.3.2 Statistical Analysis
A statistical analysis was performed on the wind conditions in order to determine the
wind probabilities. The following table shows the results of the analysis.


Wind
Direction
Wind
Speed
TAILwind on
360
CROSSwind on
150
10kt 0 8.6
West 270
20kt 0 17.2
10kt 5 10
WSW 240
20kt 10 20
10kt 7 9.6
SW 225
20kt 14 19.3
10kt 9 7.8
SSW 202
20kt 18 15.7
10kt 10 5
South 180
20kt 20 10
10kt 9 1.4
SSE 158
20kt 18 2.8
10kt 7 4.2
SE 135
20kt 14 8.4
10kt 5 6
ESE 112
20kt 10 12
Table 1: Wind probabilities
The criteria used for the analysis are as follows:
Change from NORTH configuration to SOUTH configuration when tail wind
exceeds 10kt.
17

CREDOS conditions when crosswind on 15L exceeds 10kt.
Probability of South Configuration: The preferential configuration in Madrid Barajas is
the NORTH configuration. The Tower controllers change to SOUTH configuration when
tail wind exceeds 10kt.
The wind conditions (direction and speed) that give South configuration wind condition
are coloured in blue in Table 2.
Probability of CREDOS wind conditions: The wind conditions (direction and speed) that
give CREDOS wind condition on runway 15L are circled in red in Table 2.
Probability of CREDOS wind conditions when 15L is in use (South configuration):
The wind conditions (direction and speed) that give CREDOS wind condition on runway
15L are circled in red and coloured in blue in Table 2.
The following table gives for each wind direction and speed, the number of METAR
published for the year 2007 in Madrid Barajas.


Wind
Direction
<> 5 10 20 >20 TOTAL
Calm 669 669

East 372 70 4 0 446
ENE 294 195 53 0 542
ESE 302 75 6 0 383
NE 365 318 283 8 974
NNE 551 317 175 18 1061
NNW 898 514 71 12 1495
North 2261 1013 534 54 3862
NW 259 158 63 20 500
SE 353 161 57 10 581
South 679 534 129 5 1347
SSE 368 264 42 3 677
SSW 316 282 263 12 873
SW 189 254 565 33 1041
Variable 2032 56 5 0 2093
West 190 221 321 36 768
WNW 131 112 143 35 421
WSW 151 318 453 95 1017

9711 4862 3167 341 18750
Table 2: Local wind year 2007





18


The following table summarises the results of the probabilities:

Number of METAR % TOTAL
% TOTAL South
Conf
TOTAL 18750 100
TOTAL South Conf: 1937 10.3
TOTAL CREDOS: 2141 11.4
TOTAL CREDOS/South Conf: 1472 7.9 75
Table 3: Wind summary
According to the table, when Madrid Barajas is operating in South configuration, the
probability that the wind condition allow CREDOS separation is of 75%.
Over the year this would mean about 10% of CREDOS application since the probability
of South configuration is 15%.
3.4 SIMULATION TOOL
Background: Aena, in collaboration with INECO, initiated in 1997 the development if
the Airport ATC Capacity Research Programme (PICAP), characterised by a
transparent project execution process open to the participation of all actors involved in
airport operations.
3.4.1 Runway Occupancy Time Parameters
Precise knowledge of those factors influencing runway capacity at an airport is
fundamental if capacity is to be determined with accuracy.
The following parameters describe the time intervals to be collected for departure
studies:
FRLC Flight crew reaction time to line up clearance. The time interval
between the ATC clearance to line up or take off and the commencement of taxi
to line up. Note that this will not be considered when the clearance is given
whilst the aircraft is moving.
LUPT Line up time. The time interval between crossing the stop bar and the
moment the aircraft is fully lined up.
FRTT Flight crew reaction time to ATC take off clearance. The time interval
between the ATC clearance, and the commencement of take off roll. Note that
this will not be considered when clearance is given whilst the aircraft is moving.
TOFT Take Off roll time. The time interval between the moment the take off
roll commences and the moment the main gear lifts off the runway.
3.4.2 Runway capacity indicators
RMP (Spanish acronym of Maximum Runway Throughput) The maximum number
of operations that can be reached, on a given runway, in a period of time, according to
the rules of operation and taking no account of delays. In consequence of the different
factors affecting runway operations there will be more than one value for the maximum
number of movements (RMP) and the solution to the problem will be, in general, a
range or interval of values.
19

PGS (Spanish acronym of Separation assurance percentage) It relates the aircraft
approaching behaviour and the runway occupancy time, with the theoretical
separations estimated. In other words, it indicates the probability of an aeroplane being
within the last mile prior to the runway threshold, while the previous one still has not left
the runway vacant. So, PGS is also an appropriate indicator to quantify missed
approach probabilities due to a lack of aerodrome separation.
3.4.3 Quality Service Results
They are basically delay parameters, which are the best indicators of the systems
efficiency and a way of checking the systems capability to cope with a traffic demand.
It is important to notice the difference between the maximum traffic volume and the
optimum traffic level. The maximum traffic volume is the maximum that the system
could manage due to procedures and infrastructures. The optimum traffic level results
from establishing the offset between the maximum output of the system and the
adjustment to the users demand requirements and stakeholders.
Delay It is the elapsed time from the moment that an operation was scheduled until it
is effectively performed. Depending on the study to carry out, the time range can vary
from the average delay per day to the delay of a single operation.
Sensitivity It is the representation of different pairs of RMP-average daily delay
obtained from the simulation process. The utility of this metric is, once an acceptable
average daily delay is set, procuring the maximum number of operations that an airport
can accommodate.
Punctuality That is the result of pairing all the different times when a movement has
occurred during the simulation with their respective delay. Unlike the sensitivity index,
the punctuality offers the percentage of operations over the total number of operations
of the same type (arrivals, departures, or total), that are performed within an agreed or
assumed delay limits.
Data sources used:
GESLOT Scheduling Database
CONOPER Operation Database
Radar Data
Flight Progress Strips (FPS) and Operational Information provided by TWR
PICAP Data
3.5 HYPOTHESIS
3.5.1 Simulation Plan
The objective of this Case study is to evaluate the impact of applying CREDOS
separation on delay reduction and instantaneous throughput of a departure runway.
The previous analysis on CREDOS conditions in Madrid show that when South
Configuration is used, the minimum wind condition for CREDOS application (10knots in
this study) is fulfilled 75% of the time
It is also interesting to evaluate intermediate values of CREDOS application.
The proposed simulation plan is to perform 6 simulations:
- NOMINAL: 0% application of CREDOS separations
- CREDOS_30%: 30% of the time is under CREDOS separations
- CREDOS_67%: 67% of the time is under CREDOS separations
20

- CREDOS_75%: 75% of the time is under CREDOS separations
- CREDOS_85%: 85% of the time is under CREDOS separations
- CREDOS_100%: 100% of the time is under CREDOS separations


Figure 9: Analysis scenarios
3.5.2 Schedule
The schedule used in the simulation comes from Madrid Barajas airport CONOPER
database which contains all the scheduled flights for one day.
According to the simulation needs, the day selected was October 3rd 2007, day of
South configuration with 243 scheduled departures from runway 15L.
The structure of the schedule is the following:
Call sign Scheduled time A/C category SID Runway Type of operation

The initial database has been processed in order to select only scheduled flights taking
off from runway 15L: DEP 15L
It is important to note that in order to evaluate the benefits of CREDOS in term of delay
reduction, the runway must be saturated. The criterion used in this case is the
number of movements that would create an average daily delay of 5min per aircraft. It
is the result of the total delay divided by the total runway movements.
Even when the runway is not saturated, some improvement could also be provided by
CREDOS for those moments of transitory saturation where some delays would
appear at peak times.
This criterion is used for an entire airport, which includes taxi delay and runway delay.
In the study we will consider only the runway delay.
The following graph relates average daily delay with the number of movements in the
case of a single departure runway.
21

DEPARTURE DAILY DELAY SENSITIVITY TO THE NUMBER OF
OPERATIONS
y =1E-12x
4.7554
R
2
=0.9825
0.0
2.0
4.0
6.0
8.0
10.0
12.0
240 290 340 390 440 490
Number of Total Operati ons
A
v
e
r
a
g
e

D
a
i
l
y

D
e
l
a
y

(
m
i
n
)

Figure 10: Departure daily delay sensitivity
According to the graph the initial traffic (243 departures) has to be augmented to
435 departures to reach a saturation level of the runway, which corresponds to an
80% increase of the real traffic.
The CREDOS separation has been associated to a category of aircraft (Light for
PICAP) that has HEAVY performance characteristics but no heavy wake vortex.
Behind such aircraft are applied the CREDOS separations to Heavy and Mediums.
The CREDOS Heavys are randomly distributed over the day following a simple
algorithm:
- wind conditions are known every 30min (METAR)
- we suppose wind conditions are randomly distributed over the day
- x% of the hours are considered with CREDOS wind conditions
- the Heavys scheduled for these CREDOS hours become CREDOS Heavys
3.5.3 Separations
Speed Separation
The table below shows the speed separations used for the Madrid study.
PRECEEDING FOLLOWING
Speed Profile =SID <>SID
Equal or Faster 60 seconds 60 seconds
Slower 180 seconds 180 seconds
Table 4: Aircraft separations based on speed
22

Wake Turbulence Separation
The table below shows the wake Turbulence separations used for the Madrid study.

Table 5: Wake Turbulence Separations
Specific Madrid Separation

PRECEDING FOLLOWING EQUAL SID DIFFERENT
JET
JET 5 NM / 1.5 3 NM / 1 minute
JET
Not J ET 3 NM / 1 minute 3 NM / 1 minute
Not J ET
TAS>210
Not J ET
TAS>210
5 NM / 1.5 3 NM / 1 minute
Not
TAS>210 KTS
Not J ET
TA 210 KTS
3 NM / 1 minute 3 NM / 1 minute
Not J ET
TAS 210 KTS
Not J ET
5 NM / 1.5 3 NM / 1 minute
TAS 210 KTS
Table 6: Specific Madrid Separations
Up/Down wind SID separation (CREDOS)
In order to create the scenarios for the simulation the SIDs were grouped in the
following manner. Each colour marks a different group of SIDs:

Figure 11: SID separation
23

3.6 SIMULATION RESULTS
The objective of CREDOS is to be able, when wind conditions are adequate, to reduce
the separation due to wake vortex between departures from a same runway. Due to
variability and the unpredictability of wind conditions, the results will be given in terms
of delay reduction and instantaneous capacity to better picture the delay reduction.
3.6.1 Delay Results
As mentioned before, the probability of CREDOS wind conditions, when in South
configuration, is 75% percent. Below are the results from the simulation indicating the
possible delay reductions in the before mentioned conditions with saturated runway
(Section 3.5.2)
3.6.1.1 Average Delay per Aircraft
Without CREDOS (0% CREDOS), in the simulated conditions, there would be 5.2
minutes average delay per aircraft. This result is obtained by calculating the overall
delay during the 12 hour simulation, divided by the total number of departures on the
runway.
With CREDOS at 75% there would be and average delay per hour of 2.9 minutes per
aircraft
On a given hour and using CREDOS 75% percent of the time, the delay reduction
could reach 2.3 min. per aircraft.
AVG DELAY PER HOUR
5.2
3.7
3.2
2.9 2.9
2.2
0
1
2
3
4
5
6
0% 20% 40% 60% 80% 100% 120%
% CREDOS
M
i
n
AVG DELAY

Figure 12: Average Delay per hour
3.6.1.2 Runway Throughput
The following graph shows how an increase of 3 operations would be possible if
CREDOS would operate at 75%. On the vertical axis the number of operations per
hour is shown as RMP (see section 3.4.2). On the horizontal axis the percentage of
CREDOS used is shown.
24

RMP
44
46
47 47
49 49
40
42
44
46
48
50
52
54
56
58
60
0.00 0.20 0.40 0.60 0.80 1.00 1.20
%
R
M
P
RMP_AL Lineal (RMP_AL)

Figure 13: Maximum Runway Throughput
3.6.1.3 Sensitivity of the Reduction of Delay to the Number of
Operations
As it can be ascertained from the results, the benefit from CREDOS is heavily
dependent on the demand.
On the following graph, the average delay per aircraft is measured up against the
number of operations, both in the baseline and 75% of CREDOS.
The reduction in delay is obtained as the difference between both values.
Delay vs Number of Departures
0.0
1.0
2.0
3.0
4.0
5.0
0 200 400 600
Departures per day
A
v
e
r
a
g
e

d
e
l
a
y

(
m
i
n
)
Baseline (0%CREDOS)
75% CREDOS

Figure 14: Delay vs. Number of departures
With CREDOS at 75% the graph below shows how the delay reduction increases as
the traffic demand increases.
25

Delay Reduction per ACFT
0.0
1.0
2.0
3.0
4.0
5.0
300 350 400 450 500 550
Departures per day
A
v
e
r
a
g
e

d
e
l
a
y

(
m
i
n
)

Figure 15: Delay Reduction wrt Departures
3.7 CONCLUSION
After having studied Barajas current situation we can conclude that CREDOS is not
fully applicable with todays operations as Barajas is not saturated at this moment.
Nevertheless, this project could be useful in any saturated airport with the necessary
conditions in order to optimise its capacity in certain situations.
If the traffic conditions increase in the future, as it is expected, the use of CREDOS for
the reduction of delay is certainly a good candidate in operational terms.
These conclusions are realised when the analysis is performed in average terms. It is
also true, as it shall be used in the Cost Benefit Analysis (section 6), that if CREDOS is
present at a runway which is not saturated in average terms, for those peak times
when a transitory saturation occurs, some benefit in terms of reduction of delay can
also be obtained. It is true that those delays would be absorbed as the schedule is not
saturated, but as some aircraft would have suffered the punctual delay, this could also
be ameliorated with CREDOS.
26

4 Interoperability with current LEMD infrastructure
To evaluate the interoperability of the new CREDOS components with the current
systems present at Madrid Barajas, there are two main points to consider:
Difficulty obtaining meteo data from AEMET (Spanish provider): CREDOS
contemplates the need of Meteorological information for the proposed HMI
(Conops B). The Spanish service provider has a policy of not providing data
flow but instead providing its own display to mount on the towers.
Overcrowded displays and towers: There are many systems at use in a tower
and each of them has its own display and accompanying hardware. Also
through the years all the systems have evolved to provide more and more
information to the controller thus projecting more information on its relative
display screens.
This situation makes for overcrowded displays and overcrowded towers, making it
difficult to include yet another display screen or hardware to the already crowded
towers. Notice on the image below how there are screens already overlapping.


Figure 16: LEMD Tower
As a possible solution it is proposed that the CREDOS display would be included in an
evolution of the current Spanish ATC control system, the SACTA System, although it is
our belief that this solution could have important cost and time impact (not analysed in
detail).
The SACTA system through the years has been working for the modernization of the
display systems for the display positions in the towers of the main Spanish Airports.
The VICTOR System, part of the SACTA system, constitutes one of the first steps to
achieve an advanced surface movement and guidance system and it includes ICAOs
recommendation for this type of system.
27

The objective of the system is to provide the necessary tools to guarantee the safe and
efficient movement of aircraft and vehicles in the airport environment and all visibility
conditions. To achieve its purpose it receives information from the following sources:
Approach radars
Surface radars
Closed Circuit Television Cameras
Air Traffic Control Centre Flight Plan information.
Aeronautical and Meteorological information
Stand assignment information
ATIS messages.
Information on critical nav-aids on the airport.


Figure 17: SACTA controller positions
28

5 Human Factors Study
5.1 Introduction
The role of Human Factors in ATM system design, evaluation and implementation is
critical. With increasing automation and advanced technology it is important to identify
and manage human performance related issues as early as possible. This will result in
optimised relationships between people, tasks, technologies and the working
environment aiming to ensure safe and efficient human performance.
CREDOS implementation at Madrid-Barajas Airport may mean some variations in
controller working methods. Thus, its impact on Human Factors has to be addressed at
an early stage of the Project, in order to efficiently fit the task with the human element.
5.2 Objectives
The main objective of the Human Factors Analysis is to assess CREDOS
implementation feasibility at Madrid-Barajas from a Human Factors perspective.
This will allow identifying critical areas and efforts needed to better integrate the new
operational concept at a particular airport, summarised in a set of conclusions and
recommendations for future implementation.
5.3 Approach, methodology, and activities
Due to the early stage of the project development, an implementation model without
carrying out a human factors assessment of specific issues at Madrid-Barajas airport
would not be feasible.
Based on the CREDOS HF Plan, Conops B and the Madrid-Barajas operational
context, a series of HF issues have been identified. Some of these issues are:
- New roles and responsibilities;
- cognitive processes;
- mental workload and
- human error
For each HF issue identified, the tasks and activities required in order to measure the
impact of human factors on the concept have been identified.
1. Madrid-Barajas Tower: New Roles and Responsibilities. It considers not only the
description of new roles and responsibilities of all actors involved in departure
operations but also a LCL controller Hierarchical Task Analysis based on previous
EUROCONTROL work.
2. Mental Workload Analysis. It is about analysing CREDOS foreseeable impact on
the Local Controllers mental workload at Madrid-Barajas North Tower. Thus, by
using mental workload defined by the SWAT technique, it has been assessed how
CREDOS implementation would impact on each LCL controllers main tasks.
3. Human Error Analysis. The identification of critical human errors from the Local
Controller perspective by analysing possible task deviations using a set of
guidewords (Human Hazop). One of these errors has been quantified using Human
Error Assessment and Reduction Technique (HEART) and considering specific
Madrid-Barajas working environment.
The analysis concerning these different issues has allowed proposing training
recommendations in order to cope with implementation needs by enhancing
29

acceptance and facilitating future integration of CREDOS in this particular operational
context.
5.4 Madrid-Barajas Tower: New Roles and Responsibilities
5.4.1 Introduction
CREDOS concept implementation at Madrid-Barajas would imply some changes in the
roles and responsibilities of operational personnel that must be taken into account.
This section provides a brief description of new roles and responsibilities assumed by
the main actors involved in CREDOS operation based on CONOPS version B,
CREDOS Preliminary System Safety Assessment (NLR) and Barajas specific scenario.
5.4.2 Actors
The following table presents an overview of the actors with their current responsibility
and their specific additional role in the CREDOS single runway departure operation
with some specific features regarding Madrid-Barajas operational context.
Actor
Current
Responsibility
Specific/additional Role in CREDOS
North Tower
ATCO
Supervisor
(SUP)
Monitors ATC tower
and ground
operations. Decision
on arrival and
departure rate to be
applied. Propose
north or south
runway configuration
depending on
forecasted weather.
Decides when to start the CREDOS system
and apply CREDOS based on traffic demand,
runway configuration, forecast, now cast and
wind information.
Informs the air traffic controllers that
CREDOS is active, and wake turbulence
separation may be suspended.
Termination of CREDOS operation.
Local
Controller
(LCL)
In charge of take-off
phases.
Put in place and monitor safe separations and
efficient spacing and sequence using the
CREDOS suspension of wake turbulence
separations for each individual aircraft under
South configuration (LCL DEP 15L y LCL DEP
15R). In case display with go/no-go indicates
GO, LCL cross checks the validity of the GO
indication against the actual crosswind. LCL
decides to suspend the wake turbulence
separations or not.
Ground
Controller
(GMC)
Monitor Surface
Ground Movement
and surroundings.
Informs based on CREDOS or adjusts
manually to the CREDOS capacity and
sequencing options. GMC prepares the
sequence of taxiing aircraft such that most
benefit from suspended wake turbulence
separation takeoffs is obtained.
GMC will when necessary inform crew about
CREDOS and check whether crew read-back
CREDOS.
30

Departure
Radar
Controller
(DEP-
Madrid
TMA)
To efficiently
separate aircraft on
radar after departure.
Monitors the CREDOS availability and
application per flight. Receive and disseminate
CREDOS critical wake vortex and weather
information. Monitors flight track adherence,
and vectors aircraft to the right position in
case of insufficient accuracy. Monitors and
ensures ICAO radar separation under
CREDOS operation.
Flight Crew
(FC)
Navigates aircraft
safely.
Are aware of the CREDOS operation and the
suspension of wake turbulence separations.
Report CREDOS critical information.
Pilot decides whether or not to accept
CREDOS, if not already done so, informs the
Local controller (via read-back) and takes off
at an appropriate time.
Table 7: Actors and changes in role when using the CREDOS
In the following flow-chart for the Barajas Tower departure operations, the new tasks
introduced by CREDOS implementation are identified with red-filled appearance and
with red frames those tasks that would need to be changed.
The analysis of new roles and responsibilities proposed by CREDOS shows at first that
a LCL Controller would experience the most impact by the implementation. This is
because apart from the pilot, an LCL controller is the final actor who decides
suspending wake-vortex separation minima under CREDOS operation.
31

32

Figure 18: Departure operations under CREDOS

33
5.4.3 Local Controller Task Analysis
As it was mentioned before, the LCL Controller is the main actor who would undergo
the most impact by CREDOS implementation.
Therefore, the purpose of this section is to carry out a LCL controller task analysis,
which will allow to study in depth cognitive processes derived from changed tasks and
new tasks under CREDOS utilization.
The LCL controller task analysis is based on EUROCONTROLs previous work and
particularised and validated for a specific Madrid-Barajas Tower operational
environment.
The main objective of a departure LCL controller is to efficiently and safely manage
aircraft take-off operations. In this case, for Madrid-Barajas, as it is considered that
CREDOS would be applied in south configuration, the departure LCL controller would
be in charge of 15L.
In order to manage take-off operations a LCL controller carries out, in order, 5 main
tasks:
- Assume control of aircraft
- Select sequence and holding point
- Execute take-off
- Supervise take-off
- Handover to DEP
5.4.3.1 Assume control of aircraft
This is where the LCL controller assumes control of the aircraft.
It was cleared that CREDOS operations will likely not have much varied influence on
this task depending on the procedural variations.
In order to assume control of aircraft the LCL controller carries out 4 subtasks (do 1-2
in order; 3 and 4 at any moment)
1.1. Handover from GND controller
1.2. Confirm flight crew awareness of CREDOS status. This is a new task introduced
by CREDOS.
1.3. Identify aircraft. Strip is received before aircraft establishes communication at
holding point. Visual contact takes place checking aircraft type and position. If
visual conditions do not allow establishing this contact then surface radar is used.
1.4. Verify GND original sequence is right. In the nominal scenario the strip will
already be organised in reference to take-off sequence. This will have been done
by previous GND controller, but the LCL controller has to verify that the
sequencing is ok in reference to other traffic. The LCL controller does not, at this
point, consider any constraints. It is just a matter of getting a mental picture of
where the aircraft is in reference to other traffic.
5.4.3.2 Select sequence and holding point
The LCL controller considers all constraints for departure in order to organise flight
strips in sequence, keeping aircraft under surveillance. Even though the introduction of
new tasks is not expected, some changes in task content are taken into account.
With this objective the LCL controller carries out the following 3 subtasks:

3.1. Consider all constraints for departure. This is at a planning state. The LCL
controller considers all constraints that can affect not only the independent A/C
but also the overall traffic flow. The controller develops and changes the strategy
and mental model continuously. It is a task slightly affected by constraints derived
from CREDOS application.
3.2. Organise strips in sequence. This is where the LCL controller decides on the
most appropriate departure sequencing after considering all the constraints.
Normally, the previous GND controllers will have arranged a foundation for the
sequencing and given aircraft taxi clearances accordingly. Runway entry has to
be decided and instructed to aircraft, and this can vary depending on where
aircraft are and where they should be in the sequence. It represents the first task
where LCL controller can mentally assess how the aircraft is in reference with
other traffic and when it should be to facilitate operations under CREDOS.
3.3. Keep aircraft under surveillance. Controller is continuously aware of aircraft at
holding point performing visual surveillance from the tower in order to ensure
absence of possible manoeuvring conflicts.
5.4.3.3 Execute take-off
These tasks carried out by LCL controller include giving runway entry, line up and
take-off clearance and finishes when the aircraft has become airborne.
With the aim of executing take-off operations the LCL controller carries out 14 subtasks
(do 1-10 in order; 11-13 at any moment; 14):
3.1. Assign departure and runway entry sequence. As the first step in executing take-
off sequence, the controller mentally assigns information to the aircraft with the
aim of organising traffic. Then, it is possible to deliver corresponding clearances
in order to execute assigned solution.
3.2. Ensure runway clear for line-up. A task where the LCL controller ensures that the
runway entry to be used and the part of the runway needed for line-up is clear.
The runway might be clear to line-up but not takeoff. It means that an aircraft is
can be cleared to a stationary position on a specific position of the runway.
3.3. Decide when to issue line-up clearance. The controller makes a decision when to
give a line-up clearance. This can be done more or less at the same time the
clearance is issued, but also further in advance based on anticipation and
expectations. If CREDOS operations are active it is possible to anticipate line-up
clearance.
3.4. Radio transmission: Line-up clearance. The controller issues a line-up clearance
to the A/C.
3.5. Verify line-up executed. The LCL controller verifies line-up has been executed
correctly alter clearance transmission and the pilot confirms it.
3.6. Ensure runway clear in departure direction. This is a task where the LCL
controller ensures that the runway to be used for take-off is clear.
3.7. Choose runway separation and airborne spacing to apply based on rules and
regulations. Based on rules and regulations, the LCL controller has to choose
what separations to apply. The decision is made after considering all the
constraints in task 2.1. It is here assumed that the decision is correct in reference
to the outcome of the constraints considered. It could be interesting to consider
the process of making a decision when constraints are not coherent because this
occurs sometimes.
34

3.8. Ensure separation. The LCL controller has to ensure that adequate separation
will be kept between aircraft prior to giving a take-off clearance. There can be
cases where only runway separation is used. If there is only one aircraft,
separation is not applicable. There can also be requests from other sectors to
provide spacing for later airborne separation. This task is performed in the wake
of the constraints considered in task 2.1. Based on the outcome of the
considerations and the decision needed to ensure separation, a take-off
clearance can be given. Complexity of the task may be affected by the fact that
the controller has an additional objective regarding runway, wake-vortex and
airborne separation.
3.9. Decide when to give take-off clearance. The LCL controller makes a decision
when to give a take-off clearance. If CREDOS operations are active, the take-off
clearance can be issued earlier due to reduced separation. In line with this task,
the controller will have to decide whether to apply this reduction or not. At Barajas
Tower a 2 minute separation is established using a conventional clock. For
CREDOS it is suggested to have a timer tool integrated into the working position
to support this task.
3.10. Radio transmission: Issue take-off clearance. The controller issues a take-off
clearance to the aircraft after informing about wind conditions to the pilot.
3.11. Inform aircrew about CREDOS status. The LCL controller transmits information to
the Flight Crew on CREDOS operation status. Either CREDOS operations will be
active or they will be inactive. The controller determines this by referring to the
current conditions and by the green CREDOS dial on the display panel. The
Flight Crew will have to be informed that CREDOS operations are active, if not
done previously.
3.12. Receive CREDOS confirmation from flight crew. The Flight Crew reads back
information on CREDOS operation status. The LCL controller knows that the
Flight Crew is aware and accepts CREDOS operations. The controller can also
assume that the Flight Crew knows reduced separation may be used.
3.13. Update flight strip. The flight strip is updated with relevant information. ex.:
airborne time.
3.14. Verify take-off executed. The controller has to verify that the take-off is executed
before he/she can relay the responsibility to the next controller. Different factors,
such as time, weather and aircraft type determine when the following aircraft can
be cleared for take-off. The process of doing so will depend on where the
preceding aircraft in its take-off phase. This will also affect other movements on
the runway, such as maintenance. I.e. as soon as the preceding aircraft has
commenced its take-off roll, the following aircraft can line-up.
5.4.3.4 Supervise take-off
This task includes supervision and correction functions during take-off and climbing
phase.
In order to assume control of aircraft the LCL controller carries out 4 subtasks (do 1-3
in order; do 4 if necessary):
4.1. Monitor squawk and call-sign correlation. The controller monitors that the squawk
code received from a specific aircraft is correct in reference to the assigned
squawk code, and that it is correct in reference to the aircraft call sign. If the
controller is uncertain of which aircraft is squawking what and it turns out that the
squawk does not correlate to the call-sign, the controller can take correcting
35

measures. E.g. ask aircraft to first squawk ident, secondly to change squawk
code.
4.2. Supervise aircraft altitude for separation and conformance. The controller
monitors aircraft and confirms that they are climbing according to procedures,
ensuring that separation is maintained between aircraft.
4.3. Supervise aircraft conformance with cleared SID. It is important that departing
traffic follow the expected and cleared SID. If so, it will be easier for the LCL
controller to ahead the traffic and better facilitate an efficient departure flow
while maintaining minimum separation. The controller may need accurate
visualization of aircraft conformance with SID if CREDOS operations are active.
4.4. Take action to correct a deviation and or mitigate a conflict. If an aircraft does not
perform or follow controller expectations, an action may be required to solve the
deviation. This is, however, in reference to current traffic conditions and controller
workload. If the deviation is small and it doesnt affect other traffic, the controller
can choose to not act and accept the deviation, or simply give the Flight Crew a
chance to correct it themselves. SID deviations and non-normal occurrences in
active CREDOS operations might necessitate an extended sector authority for the
LCL controller. An increased knowledge requirement of wind behaviour with
altitude, SID structure (upwind/downwind) and A/C conformance with SID may
constitute such authority. Consequently, the LCL controller will have more time to
try to solve the deviation or possible conflict. As current procedures for how,
where and when the handover of control and associated frequency transfer shall
occur, varies greatly with local procedures, the time when to delegate a task to
the next sector varies accordingly. Note that situation criticality will determine how
important the time factor is in taking corrective measures.
5.4.3.5 Handover to DEP
Once the controller has checked that the take-off has been carried out properly, control
is transferred to the DEP controller in the TMA. Pilot is informed about the new
frequency on which to establish communication with the DEP controller and finally the
flight strip is filed.
Thus, this task can be summarised in 3 main subtasks (Do 1-2 in order; Do 3 at any
moment):
5.1. Update flight strip. The LCL controller notes on the flight strip that A/C departed
with CREDOS reduced separation.
5.2. Transfer strip and communicate. The LCL controller transfers the flight control to
the DEPARTURE controller. CREDOS operations may require a more coherent
procedure for how, where and when the handover shall occur. The reduced
separation will, consequently, give a smaller opportunity window to take
corrective actions in case of a non-normal or unforeseen deviation by the
preceding aircraft.
5.3. Transfer of frequency. The aircraft is transferred from the LCL controller to the
DEPARTURE controllers frequency. Local procedures can vary greatly, but it is
most common that the LCL controller issues a clearance to change frequency.
Other procedures include directions to change frequency directly when airborne,
or at a specific altitude.


36

5.4.3.6 Task Analysis Representation
Figure 19 shows overall result from previous analysis based on EUROCONTROLs
work.


Figure 19. Local Controller Hierarchical Task Analysis under CREDOS.
From the task analysis it can be preliminarily concluded that CREDOS would not
suppose a significant impact on a LCL controllers work and thus on its performance.
Nevertheless, a set of specific studies have been done based on present task analysis
in order to justify this previous affirmation.
Firstly, a basic mental workload analysis has been done, and secondly, possible
human errors derived from CREDOS utilisation have been identified, both applied to
the LCL controller.
37

5.5 LCL Workload Analysis
Before making some considerations concerning CREDOS implementation operational
impact on controllers mental workload, it is advisable to define what is meant by
mental workload and how it can be assessed.
5.5.1 Concept and assessment of mental workload
Henceforth we will use the following definition of workload proposed by Stein and
Rosenberg (1983): amount of effort that an operator must develop in carrying out a
task to get a specific result
In accordance with ISO 10075:1991 Ergonomic principles related to mental work-
load, assessment and measurement of mental workload has to be related with the
following stages:
Assessment of working conditions producing mental stress, in a similar way
when designing and evaluating working system models.
Assessment and measure of mental strain produced by mental stress in order
to establish, for example, stress tolerability.
Measure of strain effects on the operator, for example, fatigue, monotony,
overload or reduced surveillance.
With the aim of assessing mental workload different techniques can be used which
suitability depends on the measure context. In particular, the following can be applied:
Physiological measures. These methods provide information about workers
physiological states under specific working conditions.
Subjective allocation. These methods provide information about how workers
assess, in a subjective way, different issues of mental workload and how they
perceive its own working conditions, using for example, psychometrical scales.
Performance assessment. These methods offer the possibility of assessing
mental and psychomotor performance under specific working conditions, with
the aim, for example, to establish performance deviations due to mental
workload effects.
Task and work analysis. These methods assess elements of the task as
sources of mental workload. Physical and psychosocial working conditions, as
well as environment conditions and organization of working process.
5.5.2 Methodology
For the Madrid-Barajas Case study, a preliminary analysis corresponding to the last
group (task and work analysis) has been carried out. In this case, LCL controller task
analysis is used as a baseline, which allows initiating a mental workload assessment
process with a qualitative approach.
As it can be seen on LCL controller task analysis, the general objective of managing
take-off operations is made up of 5 main tasks:
1. Assume control of aircraft
2. Select sequence and holding point
3. Execute take-off
4. Supervise take-off
5. Handover to DEP
38

For each one of these tasks, it has been assessed the degree of additional workload
imposed by CREDOS implementation. For that, it is assumed that mental workload is
multidimensional and is composed by three variables extracted from specialised
literature (Reis et all, 1982).
Time Load. Time Load is the amount of time pressure experienced in
performing the task. This includes the fraction of total available time that
controller is busy and the degree to which different aspects of the task overlap
or interfere with one another. Under high Time Load, controller is unable to
complete the task due to a shortage of time or interference created by overlap
of activities.
Mental Effort Load. Effort Load is the amount of attention and/or concentration
required to perform a task. Things that are considered mental effort include
recalling things from long-term memory, decision making, performing
calculations, storing and retrieving things from short-term memory, and problem
solving. High levels of Mental Effort Load are required in a situation which
demands total concentration to perform, while during lower levels of mental
effort, the mind may wander or attention may easily be shared with several
relatively easy tasks.
Psychological Stress Load. Psychological Stress Load refers to the presence of
confusion, frustration, and/or anxiety which hinders completion of the task.
Psychological stress refers to the feelings of apprehension and tension one
usually thinks of when controller hears the term stress. In addition, other factors
such as fatigue, motivation, and low levels of physical stressors may also
contribute to the feeling of psychological stress load.
5.5.3 Results
It is possible to evaluate CREDOS implementation impact on LCL controller mental
workload by considering the introduction of new and changed tasks.
The following summary table shows the main results extracted from additional mental
workload preliminary analysis.

39

40
LCL ATCO Time Load Impact Mental Effort Load Impact
Psychologic
al Stress
Load Impact
Overall
Impact
Recommendations
1.
Assume
control of
aircraft

YES
The fact of including a new task for
confirming that aircrew is aware of
CREDOS reduces the time to perform the
rest of subtasks generating a slight impact
on time load.
NO NO LOW

Define appropriate taxonomy;
Boost teamwork;
Develop silent confirmation
procedure.
2.
Select
sequence and
holding point
NO
YES
It means to add an additional
constraint for decision process about
organising departure sequence in
order to optimise delay reduction
derived from CREDOS utilization.
NO LOW
Support to optimal departure
planning.
Training.
3.
Execute take-
off
YES
The fact of including a new task for
informing aircrew about CREDOS reduces
the time to perform the rest of subtasks
producing a slight impact on time load.
YES
Task complexity is affected by the
introduction of a new objective for the
LCL controller concerning wake-vortex
and airborne separation. It is foreseen
that controller has a new table to apply
separations under CREDOS operation.
NO
MEDIUM-
LOW
Training; Avoid ambiguous
information; Procedural
improvement for timing take-
offs between each pair of
aircraft. Human error tolerance;
Avoid focusing on objective of
efficiency instead of safety.
4.
Supervise
take-off
NO
YES
Requires higher effort the need of
having more knowledge about wind
behaviour with altitude, SID structure
and aircraft conformance with SID
NO LOW
Technical support to check SID
conformance with appropriate
information presentation.
5.
Handover to
DEP
YES
Reduced separations give a smaller
opportunity window to take corrective
actions in case of a non-normal or
unforeseen deviation by the preceding
aircraft.
NO NO LOW Training
Table 8. Mental workload impact by CREDOS implementation.

41





Results show that maximum mental workload increase would be Medium-Low for the
task 3 Execute take-off. Rest of tasks would experience probably a Low impact. This
implies that it would not be necessary an additional control position with regard to
actual configuration.
Although from the analysis it can be foreseen a low mental workload increase for LCL
controller, it is proposed a set of recommendations in order to efficiently facilitate
CREDOS new concept implementation at Madrid-Barajas.
Next, an analysis of possible human errors has been carried out derived from
CREDOS implementation and again in the specific Madrid-Barajas context.
5.6 Human Error Analysis
Human Error analysis responds to the need of assessing level of departure operational
safety after CREDOS implementation.



Results of the analysis include a set of possible human errors to be included in an
operational safety assessment in order to validate the new concept and define risk
mitigation and control measures.
5.6.1 Human Error Concept and Assessment
The term human error has been pragmatically defined by Swain (1989) as follows: any
member of a set of human actions or activities that exceeds some limit of acceptability,
i.e. an out of tolerance action where limits of performance are defined by a system.






Human errors are not intrinsically different from any other form of human behaviour;
rather they are actions which are misplaced. Reason (1990) usefully distinguishes
between slips and lapses (e.g. pressing the wrong push-button, or forgetting a step in a
long procedure), which are unintentional or inadvertent errors, mistakes, which are
errors of intention (e.g. misdiagnosis), and violations (e.g. taking an illegal short-cut
during a procedure).
5.6.2 Methodology
The objective of actual analysis is to identify possible human errors that in Madrid-
Barajas context would suppose a possible operational safety risk for CREDOS
application.





With the support of information extracted from CONOPS version B, Functional Hazard
Assessment and LCL controller Task Analysis, there have been identified new
functions of human element through a sequential description of its activities in
departure operations.
In order to facilitate the identification of possible tasks deviations produced by human
behaviour it is used a set of Guidewords (Human Hazop) applied one by one to each of
controller tasks.
No. A complete negation of the intention
Reverse. The clear opposite of the intention.
Less of. A quantitative decrease.
More of. A quantitative increase.
As well as. A qualitative increase.
Part of. A qualitative decrease.
Other. Complete substitution.
Sooner than. Intention done sooner than required.

Later than. Intention done later than required.
Result of this process is a human error list that subsequently can be quantitatively
assessed by determination of occurrence probability through human reliability
techniques. This allows analyzing the grade of impact of each error and justifies a set
of mitigation and reduction measures to guarantee that human tasks during CREDOS
operations do not suppose an operational safety decrease.
To carry out a quantitative assessment it is proposed to use HEART methodology
(Human Error Assessment and Reduction Technique, Williams, 1986). In this case, as
a first approach and as an example, it will be only applied to the assessment of the
most critical error identified.
HEART, which is a relatively quick technique to use, is based on a review, by its
author, both of the literature on human factors and, in particular, of experimental
evidence showing the effects of various parameters on human performance. The
technique defines a set of generic error probabilities for different types of task; these
are the starting point for a HEART quantification. Once a task is thus classified, it is
then determined by the analyst whether any error-producing conditions (EPC) as
defined in HEART are evident in the scenario under consideration. For each EPC
evident, the generic HEP (Human Error Probability) is multiplied by the EPC multiplier,
which thus increases the HEP. The technique also has a set of practical error-reduction
strategies which can be used to reduce the impact of errors on the system, or even to
prevent them entirely.
5.6.3 Results
5.6.3.1 Human Error Identification
Next to possible deviations analysis in execution of new and changed task by LCL
controller under CREDOS operations, human errors have been identified. Thus, the
following table shows Guidewords used, possible Performance Shaping Factors
causes, severity class for error consequences as well as recommendations derived
from human error analysis.
42

43
LCL ATCO Subtask Guideword External Error PSF Causes Severity Recommendations
1.
Assume
control of
aircraft
1.2. Confirm
aircrew aware of
CREDOS status
No Not confirm aircrew aware of CREDOS active status
Memory: forget
action or information
Low
3.8. Ensure
separation.
No
Not ensure separation between aircraft previous to
take-off clearance due to the increase of task
complexity during CREDOS operation
Violation: Situational
violation.
Medium
Sooner than
Decide sooner to give take-off clearance breaking
minimum separations needed under CREDOS
operations. A reason could be to do harm timing,
issuing early take-off clearance.
Decision: lack of
projection; Violation;
Memory: forget
information.
Medium-
High
Training; Procedure: improvement
of timing take-off between aircraft.
3.9. Decide
when to give
take-off
clearance.
Other
Apply CREDOS while conditions are not met.
Reasons could to misunderstand traffic mix or
compatible SID.
Action: not clear
information
High
Training: Theory and practice;
Information Design: Avoid
ambiguous information.
No Not inform aircrew about CREDOS status.
Memory: forget
action
Medium
Later than Inform late about CREDOS status
Decision: late
decision.
Low
Reverse Inform CREDOS not active while it is.
Action: wrong
information
Low
Other Inform CREDOS active while it is not.
Action: wrong
information
Medium-
High
Procedure: Silent confirmation
procedure; Engineering devices:
not possible to confirm active
while not.
3.
Execute
take-off
3.11.Inform
aircrew about
CREDOS status
Part of Suppress information about CREDOS operation
Action: not clear
information
Low
No
Omission of aircraft SID conformance checking. In
Barajas it is not critical because of common and
straight initial flight path.
Memory: forget
action
Low
4.
Supervise
take-off
4.3.Supervise
aircraft
conformance
with cleared SID Later than
Deviation or conflict correction delayed under
CREDOS operations.
Decision: late
decision.
Medium
Low
Procedure: establish procedure
on Letter of Agreement.
Table 9: Human Error Analysis


Page 44



After main human errors identification it could be conclude that the most critical error
refers to task 3.9: Decide when to give take-off clearance.
It seems reasonable because CREDOS affects manly to wake-vortex separation
between aircraft. Likewise, there is human error risks derived from supervising aircraft
SID conformance and informing aircrew about CREDOS status by the LCL controller.
Next, attention is paid on quantitative assessment of critical human errors previously
identified. It is an example of what it should be done with support of operational
personnel from Barajas North Tower.
5.6.3.2 Human Error Quantification
With the purpose of evaluating quantitatively human error it is suggested to use HEART
technique.
For this case, it has been assessed the probability of most critical human error
corresponding to task 3.9: Apply CREDOS while conditions are not met
With this objective, first of all it is allocated the most appropriate General Task Type for
the task corresponding to the error (Decide when to give take-off clearance) by using the
following table.

For task 3.9., and assuming type E - Routine highly-practised, rapid task involving
relatively low level of skill it is assigned a Nominal Human Error Probability of 0.02.
E - Routine highly-practised, rapid task involving relatively low level
of skill
D - Fairly routine task performed rapidly or given scant attention
C - Complex task requiring high level of comprehension and skill
B - Shift or restore system to new or original state on a single
attempt without supervision or procedures
A - Totally unfamiliar, performed at speed with no idea of likely
consequences
H - Respond correctly to system command even when there is an
augmented or automated supervisory system
G - Completely familiar, well designed, highly practised routine task
occurring several times per hour
0.0004
F - Restore or shift a system to original or new state following
procedures with some checking

GTT Description
Table 10: HEART General Task Types Descriptions
Nominal
Unreliability
0.00002
0.003
0.16
0.26
0.55
0.02
0.09


Page 45



After, there were identified possible Error Producing Conditions that could exist in
Barajas scenario under CREDOS operations. This is indicated in the following table.
Error Producing Conditions (EPC) Total Effect
1 Unfamiliarity x 17
2 Time pressure x11
3 - Low Signal to Noise Ratio x 10
4 - Ease of Information Suppression x 9
5 - Ease of Information Assimilation x 8
6 - Model Mismatch (Operator / Designer) x 8
7 - Reversing Unintended Actions x 8
8 - Channel Capacity Overload x 6
9 - Technique Unlearning x 6
10 - Transfer of Knowledge x 5.5
11 - Performance Standard Ambiguity x 5
12 - Mismatch between Perceived / Real Risk x 4
13 Poor, ambiguous, ill-matched system feedback x4
14 No clear, direct or timely confirmation x 4
15 Operator inexperience x 3
16 Impoverished quality of information x 3
17 Little or no independent checking x 3
18 Conflict among immediate and long term x 2.5
19 No diversity of input information for checks x 2.5
20 Mismatch between education and task demands x 2
21 Use other more dangerous procedure x 2
22 Little mind & body exercise outside work x 1.8
23 Unreliable instrumentation x 1.6
24 Absolute judgements beyond experience x 1.6
25 Unclear allocation of function / responsibility x 1.6
26 No obvious way to keep track of an activity x 1.4
Table 11. HEART Error Producing Conditions Effects.
It is assumed the possibility of channel capacity overload (n 8) presence and use of
other more dangerous procedure at Madrid-Barajas tower under CREDOS application.
For each EPC identified in previous step, the analyst makes a judgement on how much it
influences the overall unreliability of the task. This is known as the Assessed Proportion
of Affect (APOA) for the EPC. A value between 0.1 (weak affect) and 1 (full affect) is
selected.


Page 46



In this case:
Channel Capacity overload: APOA=0.1
Use other more dangerous procedure: APOA=0.3
For each EPC, Assessed Affect (AA) is calculated with the following operation:
AA = (PF Total Affect 1) x APOA + 1
Therefore,
AA1 (Channel Capacity overload) =(6 1) x 0.1 +1 =1.5
AA2 (Use other more dangerous procedure) =(2 1) x 0.3 +1 =1.3
Finally, the total resulting human error probability (HEP) is calculated through the
following expression:
HEP = Nominal HEP x AA1 x AA2 x AAn
Hence, the probability of an error during task execution is:
HEP= 0.02x1.5x1.3= 0.039
This result itself does not offer safety information about operations under CREDOS.
Therefore, it would be necessary to integrate it into an operational safety analysis
specific for Barajas scenario, which is not within the scope of this Case Study.
With all analysis carried out at present, it is not foreseeable that CREDOS
implementation supposes a significant impact from a Human Factors perspective.
However, from here forth it has been gathered a set of information and training issues
necessary to effectively integrate CREDOS at Barajas, and thus count on controllers
trust and acceptance of new operational concept.
5.7 Implementation Need Analysis
From previous analysis it is possible to justify a set of activities in order to support
implementing CREDOS concept at Madrid-Barajas airport from a Human Factors
perspective.
5.7.1 Objectives
Implementation Need Analysis aims to identify training and information needs to
integrate CREDOS at Barajas.
It implies the selection of the most appropriate training methodologies so that the
controller has the necessary knowledge and skills to efficiently and safely apply
CREDOS. Likewise, information and training is in itself an effective tool to manage
change, trust and acceptance of the new concept.
5.7.2 Training and Acceptance Plan
Based on results from previous analysis, the following training activities necessary to
implement CREDOS at Barajas Tower have been considered.
First of all, it would be necessary to carry out a detailed training organization, with the
objective of establishing by consensus minimum training needs given the amount of


Page 47



information generated by the project. It would be beneficial to involve controllers in the
training, as trainers. (Guideline: involve controller in the training, as trainers).
Secondly, it has been agreed that minimum training would correspond to 6 hours (Aena:
1 FAENT unit) including theory and practice relative to operational concept, different
roles of actors involved, activation and application procedure, as well as issues related to
main tools to be used during wake vortex separation suspension.
Contents of this training plan would include:
Theory: 2 hours
Practice: On J ob Training. 4 hours
Approximately 200 tower personnel would be target audience for CREDOS training and
acceptance plan.
It is important that controllers understand why, and under what conditions, CREDOS
architecture might have errors. Leaving aside the problems of testing a prototype tool in
a simulation, even if CREDOS is working as intended it is unrealistic to expect it to be
perfect. Thus, trust and acceptance will grow if operators find compensating strategies
for the consequences of an error.
Likewise, so as to increase the likelihood that the new system will fit most controllers
and their tasks, a representative number of controllers should be involved to provide
advice and feedback in the development and implementation of CREDOS. Not only
should this help with system development, but it should also give a reasonable number
of controllers a feeling of ownership which they can transmit to their colleagues,
thereby helping to facilitate the development of trust.
Training and transition should take place over a reasonably long period to gain trust.
This extended timeframe also gives system developers the space to more fully test and
de-bug the system before it goes operational.
Furthermore, communication will be a key means of keeping controller up-to-date with
developments and timings, and will generate confidence in those who are bringing in the
new operational concept. It is also useful for controllers to know such details as the
rotation in which they will all be involved as they near final transition and implementation.
Such communication will not cause trust its own, but may prevent problems arising if
controllers feel uninvolved or uninformed, or feel surprise by the course of events.
5.8 Conclusions and Recommendations
After studying human performance foreseeable impact due to CREDOS implementation
at Madrid-Barajas Tower context, the following set of conclusions and recommendations
have been drafted.
5.8.1 Conclusions
The study concerning human factors issues has not shown a significant impact derived
from CREDOS implementation at Madrid-Barajas.
Identification of different tower control positions, analysis of its roles and responsibilities,
as well as new tasks and procedures during CREDOS application have shown that LCL
controller would be the actor who would undergo the most relative impact.
This circumstance has justified to study in more depth those issues related with
performance and reliability during take-off management. These issues cover a


Page 48



preliminary additional workload analysis and a human error analysis in Barajas specific
context which have result into several considerations with a view to take into account
acceptance and transition related issues.
Main conclusions derived from the study can be summarised by the following:
With all analysis carried out at present, it is not foreseeable that CREDOS
implementation would imply a significant impact in Madrid Barajas from a Human
Factors perspective.
Maximum mental workload increase would be Medium-Low for executing take-off
by LCL controller. Rest of tasks would experience probably a Low impact.
It would not be necessary any additional control position with regard to actual
configuration.
Results of the analysis include a set of possible human errors to be included in
an operational safety assessment in order to validate the new concept and define
risk mitigation and control measures.
After the main human errors identification it can be conclude that the most critical
error refers to task 3.9 Decide when to give take-off clearance.
It has been agreed that minimum training would correspond to 6 hours (Aena: 1
FAENT unit) including theory and practice relative to operational concept,
different roles of actors involved, activation and application procedure, as well as
issues related to main tools to be used during wake vortex separation
suspension.
5.8.2 Recommendations
With the purpose of facilitating CREDOS integration at Madrid-Barajas and during this
study, the following recommendations have been identified:
Engineering devices: Introduce a timer (at present a regular clock is used) to
apply take-off separations; Technical support to check SID conformance with
appropriate information presentation; Not physically possible to confirm CREDOS
active while not.
Procedures: Define appropriate taxonomy; Consider possibility of developing
silent CREDOS confirmation procedure; Procedural improvement for timing take-
offs between each pair of aircraft; Human error tolerance; Establish procedure on
Letter of Agreement (LoA).
Organization: Boost teamwork.
Training: Develop information and training plan (theory and practice);
Understanding possible errors of CREDOS by controllers; Involve a substantial
proportion of the controllers who will inherit the system;
Information Design: Avoid ambiguous information; Support to optimal departure
planning; It would not be advisable that the CREDOS activation advisory system
changed frequently so not to overload controllers and avoid possible errors.
Duty readiness: Avoid focusing on objective of efficiency instead of safety;
Expert controllers take part of CREDOS design and implementation process;
Allow adequate time for transitions; Inform the controller population of the
implementation schedule;


Page 49



6 Cost Benefit Study
6.1 Methodology approach
In this chapter we analyse, as far as possible with the available information, the costs
and benefits of the CREDOS implementation and operation in the period of the study.
A more analytic CBA should be made when specific and quantitative CREDOS
infrastructure costs would be described. Those costs would be provided by
Meteorological and Aeronautical industry.
We will use EMOSIA general guidelines to analyse this project, considering the
information available.
EMOSIA is the European Model for Strategic ATM Investment Analysis. Its objective is
to facilitate decision making, by understanding the global impact on ATM performance of
any proposed change, thus reducing risk
EMOSIA has a step by step approach that follows the project development. This will
enable us to monitor the value of the project and to focus on the most significant issues.
It has been designed to take into account uncertainty as making investment decisions
early on in a project means basing your decision on limited information. If is noted that
these variables are very significant, then the decision maker could repeat the exercise
entering exact values.
EMOSIA six steps are shown in the diagram below:


Figure 20: EMOSIAs six steps



Page 50



These six steps are:
1. Define decision criteria, identify the Stakeholders involved and collect data
(costs, flight improvements, delay avoidance, etc). The outputs of this stage are
the data inputs to use in the model.
2. The data inputs obtained in step 1 are used in this step to create spreadsheet
models. The output of this stage is a new output which involves economic
indicators.
3. Analyse Sensitivity. This identifies the most critical variables to the Net Present
Value of the project. Tornado diagrams are produced by the model to enable the
sensitivity analysis.
4. The most critical variables are assessed in probabilistic terms to analyse the risk
of the project. Accumulative probability distribution curve is produced.
5. Conclusions and recommendations.
6. All of the above steps can be iterated if required (taking into account the most
sensible variables)
In order to make a complete analysis, the information needed is as follows:
Previous analysis of the current situation in Madrid- Barajas airport
Infrastructure of the CREDOS implementation
Previous analysis of quantitative benefits obtained from CREDOS operation
6.2 Stakeholders
This section aims to identify the different stakeholders or groups affected by the
CREDOS implementation in the airport of Madrid-Barajas.
Following the EMOSIA methodology, the following groups have been identified:
Air Navigation Service Provider (ANSP): He will be one of the most relevant
actors of this project. CREDOS implementation and operation will depend on the
service provider activity (AENA in this case).
Airport (Madrid, in this study it is Madrid-Barajas): The airport performance,
during the period study, is passive. The airport is not going to invest and,
moreover the airport is not going to receive any benefits.
Airspace users:
o Commercial airlines
o General aviation
o Charge aviation
o Military aviation
Additionally we can identify other groups affected by the new project:
Civil Aviation Authorities and Regulators: as soon as that they participate in the
approval, certification, etc. of the system.


Page 51



Industry: manufacturers and all the industry overall affected by a new project of
the character of the implementation of CREDOS (e.g. Industry related to MET
services)
Passengers: people that make use of some type of air transport service and pay
an amount. They will be affected in time saving when there is a decrease of the
delay
Society: groups of people, that without doing a direct use of air transport
services, are seen affected from temporary or permanent way for the project that
is being analysed.
The following table presents a brief summary of the stakeholders communities to be
considered in the CREDOS CBA approach. It has been noted that each stakeholder can
act as a contributor to cost, as a beneficiary or as both.

Stakeholder Actor CREDOS
consequences
Changes
(costs or benefits)
Airport / ATC
Supervisor
Decides when to start
the CREDOS system
and apply CREDOS
based on forecast
now cast and wind
information.
Training Costs
A
I
R
P
O
R
T

/

A
T
C

Runway Controller
(TWR)
Allocate and monitor
safe separations,
efficient spacing and
sequence using the
CREDOS suspension
of wake turbulence
separations.
Training Cost
ANSP
Infrastructure and
equipment needed
Infrastructure and
equipment costs
Ground Controller
(GND)
Use DMAN (or similar)
information based on
CREDOS or adjusts
manually to the
CREDOS capacity
and sequencing
options.
Training
A
N
S
P
/

A
T
C

Departure Radar
Controller
Monitors the
CREDOS availability
and application per
flight. Receive and
disseminate CREDOS
critical wake vortex
and weather
information.
Training


Page 52



Stakeholder Actor CREDOS Changes
consequences (costs or benefits)
Flight crew
Are aware of the
CREDOS operation
and the suspension of
wake turbulence
separations. Report
CREDOS critical
information
Not considered
A
I
R
L
I
N
E
S

Airline companies
Reduction of delay.
Reduction of CO2
emissions
Benefits from:
- Saving Delay cost
- Decrease of the CO2
emission right payment
Table 12: Stakeholders
6.3 Geographic scope
The geographic scenario considered in this study is the Madrid-Barajas Airport. This
scenario was described in section 3.1 of this document.
6.4 Time Period
The period of the time of the CBA study is from 2009 until 2028 as we can see in the
figure below.
2028

2020 2019 2018 2017 2016 2015 2014 2013 2012 2011 2010 2009 2028

2020 2019 2018 2017 2016 2015 2014 2013 2012 2011 2010 2009
Set off
CREDOS project
implementation
Operation Costs
........
Operational Benefits

Figure 21: CBA Timeline
As we can see in the figure shown above, it was assumed for the study that the
implementation of CREDOS project would be carried out during the year 2009. The new
system will begin to be operative in the year 2010 and will be operational in the following
years. The CBA study will analyse the period highlighted (from 2009 to 2028).
Although this is clearly a non-realistic timeline in terms of start of operations, we have
preferred to make this assumption in order to use as precise as possible starting data
(traffic, etc). We understand that this assumption is acceptable in the current level of
development of the CREDOS project.


Page 53



6.5 Baseline do nothing option.
6.5.1 Scenario description
Madrid-Barajas Airport has four runways which are working in single operation mode.
That is, there are two runways for departures and two runways for arrivals.
The airport operates in two possible configurations; North and South (see section 3.1).
6.5.2 Demand analysis
Looking at the European situation, and before the analysis of the traffic demand
prediction or the future traffic demand, we know through some reports that more airport
capacity is needed.
The EUROCONTROL Performance Review Commission Report (PRRC7) published in
mid-2004 recorded that the cost and impact of airport-related delays had reached parity
with those for en-route and noted that airports were now a major constraint to growth.
The EUROCONTROL Constraints to Growth study reached the same conclusion. It
predicts that if the traffic demands continue to grow, even at a conservatively predicted
rate, and airport throughput capacity problems are not resolved, some thirty percent of
traffic demand will not be accommodated. Interest in the potential capacity gains from
Wake Vortex (WV) research has increased in recent years, as part of the general need
to increase airport capacity and resilience without extension of the existing airport
infrastructure.
The evolution of the number of movements affected by CREDOS has been assumed to
be equal to the evolution of the demand in the period of the study (2009-2029), extracted
from Aena and medium and long term STATFOR forecast [1], [3]):

Madrid-Barajas Airport.
Annual Operations
0
50000
100000
150000
200000
250000
300000
350000
400000
450000
500000
550000
600000
650000
700000
750000
800000
850000
900000
2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029
Year
O
p
e
r
a
t
i
o
n
s

Figure 22: Madrid Barajas Traffic demand


Page 54



As we can see from the bar chart above, the number of operations per year at Madrid-
Barajas airport is expected to increase with time. It can be noted that this increase is
expected to be very important and almost linear in time (there are no abrupt changes in
the number of operations), moving from almost 550.000 operations in 2009 to more than
900.000 in 2029 which means almost doubling the number of operations which need to
be managed by the airport in twenty years.
The traffic demand applicable to the CBA study is only the movements associated to
operations in a single departure runway when in South configuration. That is, only one
quarter of the total traffic of the Madrid Barajas when in South Configuration (15% of the
time) is taken into account in the CBA performance, because only this part of the traffic
is really affected by the CREDOS project operation.
6.6 Approach to the cost analysis
6.6.1 Investment Cost
This cost includes all non-recurring cost from the early equipment acquisition up to the
operational readiness of the system. This group of cost covers:
R&D costs: Costs incurred during the process of Research and Development. It
would be recommended to consider the cost of this study (CREDOS) as part of
R&D costs in future analyses. The cost of this project is co-financed by the
European Commission within the Sixth Framework Programme
Acquisition cost: procurement cost including acquisition of spares
Integration and installation cost
Cost of testing and validation
Commissioning and certification cost
At the present level of maturity of the CREDOS project, it is very difficult to obtain some
of the above costs as, for instance, there is no clear architecture defined for the system
yet. In order to solve this problem, a global value for infrastructure costs (including
replacement) was provided by Expert J udgement within the project team in WP4; this
figure was reused for the present study (see Table 14).
6.6.2 Operating Cost
Expenditure associated to the operating phase of the Operational Improvement
including:
Staff (operating and support staff, internal or external).
Operations (maintenance cost, repair, materials, leasing, etc
Overhead: Other costs like administration, repayment, loans.
In this study it is considered that there is no new staff cost from the implementation and
operation of CREDOS. Other operating cost linked to the Operation activity, such as
maintenance, repair, reposition, etcare included in the study.


Page 55




6.6.3 Transition Cost
The main transition costs are:
Cost which occur when it is necessary to maintain parts of the current system
during the transition period to a new system.
Training cost: Cost of the training for each stakeholder if it were necessary:
In this case only controllers need training. The analysis of these training needs is done in
section 5. Training includes the cost of a plan which will include 2 hours of theory and 4
hours of practice. 200 controllers will receive this training.
In the present study, it is not necessary to keep parts of the initial system during the
transition period. Only training cost would be included.
6.7 Approach to benefit analysis
The first purpose of the CREDOS concept (Conops B) is to suspend the wake
turbulence time and distance separations up to ideally 60 sec while maintaining
minimum 3 NM (for following on a diverging upwind SID) radar separation and 100 sec
while maintaining 5 NM radar separation (for aircraft following on the same or downwind
SID and where 5 NM have to be applied either because of en-route separations or
because of the 5 NM separation required at a later phase than initial climb phase).
The application of reduced or suspended wake vortex separations has the potential to
significantly increase the efficiency of departure movements by the reduction of
(intermediate) delays and to increase the maximum number of movements per runway.
In order to obtain the benefits explained above, from a safety point of view, a reduction
of wake turbulence separations can be envisaged provided that the crosswind
component transports the wake vortices out of the runway in a shorter lapse of time than
the one applied between two departures. In todays operations on a larger airports the
minimum separation for aircraft on the same track is imposed by the radar constraints (3
NM) which represents approximately a 1 minute time spacing when departing from a
runway. According to the ICAO wake turbulence rules the potential improvement brought
by a CREDOS concept will mainly focus on the suspension of the wake turbulence
separations between:
MEDIUM and LIGHT behind HEAVY (today 2 min and 5 NM)
LIGHT behind MEDIUM (today 2 min and 3 NM)
In this section we identify two main benefits which came from the simulations made in
this deliverable, always under the assumptions and hypothesis made in the Operational
Study (section 3).
6.7.1 Delay decrease
The standard figure for cost per minute of delay includes cost as: fuel, maintenance,
crew, passenger compensation and passenger opportunity costs; it is estimated as 51
[6]. Delay decrease of the aircrafts in queue is the main intuitive benefits from the
CREDOS operation. We must not forget that CREDOS offers a punctual increase of
capacity of the system which can be translated into a delay reduction. This reduction of
the delay has been analysed in the Operational Study trough the simulations described
in section 3.6.


Page 56



6.7.2 Environmental benefits: Decrease of the CO2 emissions
The reduction of time in queue due to the implementation of CREDOS will mean a
reduction on the emission of contaminants from the airport, as well as savings for the
airlines that will need to pay fewer rights to contaminate.
Taking into account the agreement reached by the members of the EU, from 2012 on,
civil aviation will be included in the community system of the CO2 emissions trade. The
normative will be applied on all the flights that land or take off in EU airports.
The agreement mentioned above states that airlines will have to pay for the right to
contaminate, as other industries do. In case of surpassing their limit of emissions, they
will be obliged to buy supplementary rights in other industries markets.
The table below shows the aircrafts fuel consumption during queuing and therefore with
their engines on. Taking into account Madrid-Barajas eight most representative aircraft,
we can calculate the mean CO2 emitted per aircraft. In order to do this, we obtain the
corresponding CO2 emissions from these aircrafts fuel consumption: the combustion of
a ton of fuel causes the emission of approximately 3.15 tons of CO2 [5], values which
are indicated in the table below.

Type of aircraft
[7]
Fuel burn (kg/min)
[6]
CO
2
Emissions
(Tn/min)
AIRBUS A320 10,50 0.023
AIRBUS A319 10,50 0.035
AIRBUS A321 11,17 0.036
BOEING 737/800 11,50 0.040
BOEING 737-800 11,50 0.053
MCDONNELL
DOUGLAS MD87
16,70
0,05
AIRBUS A340-300 12.40 0,04
BOEING B757/200 13,67 0,04
WEIGHT
AVERAGE
11.20 0.035
Table 13: Fuel burn and CO
2
Emissions

6.8 Economical Study
6.8.1 Costs and benefits in economical terms
Not all the costs and benefits listed in the previous sections were included in the
economical study. Some of them were eliminated for the lack of information or for being
irrelevant.
In this section, we have analysed the costs and benefits which have been taken into
account in the economical model.


Page 57




Stakeholder Cost / Benefit Base Value Units
Training Cost 0.4 K/training hour
1
Thousand of euros
Infrastructure Cost 5000
2
Thousand of euros
Operating Costs 140 (yearly) Thousand of euros
A
N
S
P
/

A
T
C

Replacement
750 (every seven
years)
Thousand of euros
Saving delay costs 0.051 Thousand of euros/min
A
I
R
L
I
N
E
S

Fuel burnt reduction
(CO2 reduction)

0.05918

Thousand of euros / Ton
A
I
R
P
O
R
T

None 0 0
Table 14: Cost and benefits in economical terms

6.8.2 Assumptions
In order to carry out the CBA analysis the following assumptions have been taken into
account:
6.8.2.1 Economical assumptions
The monetary unit used is the Euro ().
Constant prices referred to the year 2009 have been used, inflation and discount
tax were not be included;
Following EMOSIAs recommendations, a 8% discount rate was applied;
No external financing was assumed to exist, and therefore all resources would be
the companys own resources;

1
Expert judgment from Aena/INECO
2
Expert judgment from CREDOS partners


Page 58



6.8.2.2 Technical assumptions
Incremental costs and benefits were taken into account, this is, all which
appeared when compared to the Base Line;
A global model was developed considering costs and benefits as part of the
whole system, without making any distinction between the two most important
parts involved (stakeholders): air space users and air navigation service
providers;
Due to the nature of the air navigation service provider, it will not be able to
provide benefits. All the costs incurred by the service provider will be made up for
by a fee system (full cost recovery). The calculation of the fee, which represents
a cost for the airspace users, has not been included in this study;
The period ranging from 2009 to 2028 was established as time horizon for the
economical study;
The costs derived from the system operation (e.g.: maintenance) were
considered as 20% of the total costs of the implementation of the system;
The mean price for CO2 was considered as the value published on the 23rd of
J une 2009.
This studys Operational assumptions were included:
o For different reasons such as meteorological ones, CREDOS will only be
used 75% of the total time and in the South configuration (runway 15L)
o The delay will be reduced 0.6 minutes/aircraft during the first year and will
increase 21% during the following years
o CREDOS will have a go/no-go display that shows when CREDOS can be
applied. Pilots always decide go. CREDOS will produce benefits only
when is active. It was estimated that CREDOS separations will be only be
applicable 15% of the time (see section 0)
6.8.3 Net Present Value and Net Cash Flow

The NPV is a measure of assets based on the discounted cash flow; future cash flows
are discounted to their value now.
Otherwise, it is the difference between the present value of cash inflows and the present
value of cash outflows.
The NPV is obtained through the formula below:



The Net Present Value of the implementation and operation of CREDOS project in
Madrid-Barajas, calculated for the period of the study 2009-2028 is 42.02 M



Page 59




The cash flows in this period are shown in the figure below.

Figure 23: Net Cash Flow
As we can see in the figure, the most negative cash flow is in year 2009, because in this
year the ANSP incurs in the implementation costs: acquisition of the necessary
elements, installation, trials, etc.., of the CREDOS system.
From 2010, when the benefits appear from the delay cost reduction, the benefits start to
appear. From 2012, as another benefit appears (saving in the payment of the CO2
emission rights), the cash flows start be more and more positive and keep increasing as
the demand is also growing. Each year has higher cash flows than the previous year
because cash flows depend on the traffic demand: If there are more flights, the delay
reduction is bigger for each flight. Although the saving on the payment of CO
2
per aircraft
is constant, as the demand is increasing the total saving from this concept also
increases. There are only two exceptions: years 2016 and 2023. Here, we can see a
smaller increase of the benefits than in the other years because the model takes into
account the cost of the replacement. The cost of the replacement is assumed to happen
every seven years.


Page 60




6.8.4 Sensitivity Analysis
This section analyses the influence of the individual parameters in the final results
(NPV).
The value of the following parameters is considered in a range from the best case
(optimistic view) to the worst case (pessimistic view). For instance it is considered that in
the best case the infrastructure costs would be 10% less than the base case considered;
likewise, in the worst case a 10% increase over the base case is considered.

Parameter Best case Base case Worst case
Infrastructure
Costs
Total value
(installation,
Equipments,
elements..)
-10% 5 000 K +10%
Maintenance,
salaries..
100 K 140 K 180 K
Operational Costs Replacement
(each seven
years)
500 K 750 K 1000 K
CBA, Safety
Studies...
Not included Not Included Not included Non Recurring
Costs
Training -15% 480 K +15%
Benefits from delay reduction in
queue
-10% 0.051 K/min +10%
Environmental Benefits (CO2
Emissions costs)

-10% 0.013K/ton +10%
Table 15: Parameter ranges for sensitivity analysis

The sensitivity analysis, that is, the analysis of the NPV with respect to the parameters
ranges listed in the table above, are represented in the following Tornado Diagram:



Page 61




Figure 24: Sensitivity Analysis. Tornado diagram
The data shown in the previous figure reflects the existence of three variables whose
weight is, in different measure, relevant for the calculation of the NPV:
Demand variation percentage: It is the variable with the greatest weight in the
analysis, with a total weight of 63.7% on the NPV. It indicates that the NPV is
very sensitive to changes in demand. As a result, with a 13% increase in the
demand, the NPV can reach values of around 47M. We should also remember
that the demand will determine both the benefit due to decrease in delays (the
more aircraft there are, the greater the delay and therefore the greater the
possibility to reduce it), and the one due to the greater number of aircraft in the
system, as the total savings coming from the decrease in the delays and in the
CO
2
production will be greater.
Delay costs per minute: This variable has a weight of 38.6% on the NPV. If the
delays cost is 51 euros per minute, the NPV will have a value of 42 M. If we
shift the cost of the delay 10% the NPV will reach almost 44M in the positive
case, and 37M in the negative. Therefore, the NPV is very sensible to the
estimation of the delay cost per aircraft.
Implementation costs: Costs in infrastructure, although relevant, do not have a
very significant weight in the NPV. With a weight of 3%, the cost of infrastructure
which involves introduction, verification and approval of all the elements
necessary to launch CREDOS can make the NPV shift between 41.5 and 42.5
M, considering a 10% range in the cost.

6.8.5 Risk Analysis
The risk analysis will be performed with the aid of a probability curve which is shown in
the figure below and which enables us to know the probability of the NPV obtained from
this exercise being exactly that or smaller.


Page 62





Figure 25: Probability Curve. Risk analysis
As we can see from the figure, with the economical information used for the model, there
is a 50% probability that the NPV will be exactly the value obtained (42M) or smaller,
and a 90% probability that the NPV will be smaller than 49M.
The probability curve also indicates that the NPV cannot be smaller than 36M, because
there is a 0% probability of this happening.
6.9 CBA conclusions and recommendations
The CBA described in this document relative to the introduction and operation of the
CREDOS Project at the Madrid-Barajas airport, has been carried out taking into account
two types of benefits due to the introduction of this system (the reduction of delay and
the reduction of CO
2
emissions).
After having developed the economical model following the EMOSIA methodology and
taking into account all the assumptions mentioned in this document, the analysis of the
results enables to state that the necessary investment for the introduction of CREDOS in
Madrid-Barajas airport is an investment project which offers positive results in the period
considered, as the NPV brought up by the study is 42M. On the other hand, carrying
out the investment implies a very small risk, as there is a 0% probability of the NPV
being less than 36M.
For these reasons, the introduction and operation of CREDOS at Madrid-Barajas would
be considered advisable from an economic point of view as, compared to the base-line
described before (do nothing), the introduction of the project offers low risk positive
results.
The CBA carried out also identifies the variables whose relative weight with respect to
the NPV is relevant: traffic demand, delay decrease and infrastructure costs. Faced with


Page 63



big variations in any of these parameters, a considerable variation in the result obtained
will occur. We must highlight that the variable with the biggest weight is the demand
variation (63.7% out of the NPV value), so there is a need to focus more attention on the
analysis of this variable, and if possible, perform a detailed study of it. We should take
into account that demand is given by external factors (GDP, transport development
parallel to aviation such as trains, roads, etc, or simply the passengers time or security
perception), which can modify its value without taking into account the project we are
studying.
It is important to highlight as part of these conclusions that the analysis was performed
using a global model, that is considering costs and benefits as part of the whole system,
without making any distinction between airlines and ANSPs. It is clear that in this first
analysis the ANSPs would bear the costs while the airline will enjoy the benefit, but
assuming the current Full Recovery System of the ANSP, the cost would eventually be
transferred to the airlines, so it is reasonable to consider a global model.
The realisation of a more precise analysis of the costs included under infrastructure
costs is also advisable, as their influence in NPV is not negligible, and we could re-
adjust these values or study the sensitivity which each one of them brings up.
We can conclude that although the introduction and operation of CREDOS is
recommended by the positive results offered by the CBA, due to the fact that its results
depend to a large extent on a variable impossible to control as the demand, a more
detailed study of this variable is recommended. We should remind ourselves that we
have obtained the demand data from the data offered by AENA in 2007 (last values
published by the Service Provider), which is, without doubt, more positive than the data
offered by more recent studies.


Page 64



7 Methodology used and guidelines
The following graph shows in a logical sequence the activities carried out in the present
study to serve as a guideline other case studies aimed at analysing the applicability of
CREDOS in other airports. The detailed description of the steps has been already
explained in the previous sections.

Figure 26: Case study methodology
After the initial analysis, where we could get a feeling if the use of CREDOS could be
justified (for instance if there is no crosswind or if the mix has few Heavy-Medium pairs),
the path followed by INECO takes in parallel the detailed Operational analysis and the
Human Factors analysis, which would feed into the Cost Benefit analysis at the end.
Once the system components and architecture of CREDOS is further defined, another
analysis should be undertaken (we only did it briefly in the present case study), and that
is the Interoperability analysis to check the necessary additions or modifications in the
current airport systems. These findings will also feed the Economic analysis.
The present guidelines propose a case study at the present level of development of
CREDOS, which would give an impression of the applicability and some initial figures.
For the actual operational implementation of CREDOS at an airport, before the actual
investment is done, a more detailed study would need to be undertaken; the steps
should not be very different from those proposed in the present study, but they should be
performed with a greater level of detail and definition.


Page 65



8 References
[1] CREDOS D4.2 Initial CREDOS Concept of operation version B, Mar 2008
[2] EUROCONTROL Medium-Term forecast. IFR flight movements 2004-2014, Feb
2008
[3] EUROCONTROL Long-Term forecast. Flight movements 2008-2030, Nov 2008
[4] CREDOS. D4.7 CREDOS Preliminary System Safety Assessment, Sept 2008
[5] Report prepared by Air Nav, J un, 2009 (www.airnav.com/fuel/report.html)
[6] Standard Inputs for EUROCONTROL Cost Benefit Analysis, 2007 Edition
[7] Aena Traffic Statistics (2008)







END OF DOCUMENT

You might also like