Distefano N. et al.

Risk Assessment Procedure for Civil Airport

UDC: 656.71.08 DOI: http://dx.doi.org/10.7708/ijtte.2014.4(1).05

RISK ASSESSMENT PROCEDURE FOR CIVIL AIRPORT

Natalia Distefano1, Salvatore Leonardi2
1, 2
Department of Civil and Environmental Engineering, University of Catania, Italy
Received 7 March 2013; accepted 21 August 2013

Abstract: A practical problem in air transport is how to manage risk and safety. In recent
years have been developed special technical and managerial skills to the systematic, forward
looking identification and control of hazards throughout the life cycle of a project, program,
or activity. Safety Management System (SMS) involves identifying, evaluating, and addressing
of hazards or risk. Its sole purpose is to prevent accidents. Safety risk assessment is defined
as the systematic identification and evaluation of the risk posed by the complete spectrum of
possible accident scenarios. Risk assessment is a tool that supports decision making and as
such supports risk management. Risk management comprises the safety optimization of the
system, the verification process and risk acceptance, which support airport operations. This
paper proposed a quantitative methodology for the risk assessment for a civil airport, which
is based on historical data of aircraft accidents, contained in the Aviation Safety Network
database, from 1 January 1980 to 31 December 2010.

Keywords: airport, aircraft accidents, hazard probability, hazard severity.

1. Introduction More than 80% of all aircraft accidents

in commercial air transport operations
During its early years, commercial aviation occur at or near an aerodrome. Statistics
was a loosely regulated activity characterized on accident data show that the greatest
by underdeveloped technology; lack of a number of accidents occurs in flight phases:
proper infrastructure; limited oversight; an ‘approach and landing’ as well as ‘standing
insufficient understanding of the hazards and taxi’ and ‘take-off’. This means that the
u nderly i ng av iat ion operat ions; a nd aerodrome, as well as its surroundings, is the
production demands incommensurate with area which may see the largest proportion
the means and resources actually available of safety events, varying from hazardous
to meet such demands. events (e.g. non-stabilized approaches of
the runway by an aircraft) to fatal accidents.
It is hardly surprising that the early days
of commercial aviation were characterized Technological improvements (due in no
by a high frequency of accidents, therefore small measure to accident investigation),
the overriding priority of the early safety together with the eventual development
process was the prevention of accidents, of an appropriate infrastructure, led to a
and accident investigation was the principal gradual but steady decline in the frequency
means of prevention. of accidents, as well as an ever-increasing
regulatory drive. By the 1950s, aviation was

1
Corresponding author: [email protected]

62
International Journal for Traffic and Transport Engineering, 2014, 4(1): 62 - 75

becoming (in terms of accidents) one of the those hazards. In order to control these
safest industries, but also one of the most risks, risk management techniques must be
heavily regulated. enforced. The first step of managing risks
is to collect data. Once data is collected,
Civil aviation is a complex mosaic of many accident precursors (hazards) are identified
varied, yet interrelated human, technical, and evaluated. Finally, countermeasures are
environmental, and organizational factors developed, communicated throughout the
that affects safety and system performance. organization, and are then implemented in
Aviation accidents result from multiple the system.
cont r i but i ng f ac tor s . L oga n (19 9 9)
mentioned that operational safety data such Hazards can be prioritized according to the
as aircraft reliability, f light data records, probability of an accident occurring, and by
employee safet y reports, enforcement the severity of an accident that may occur due
information, inspector investigations or to the hazard. In order to prioritize hazards,
oversight information were also essential each hazard must be ranked according to the
to aviation risk analysis. most severe or the least severe outcomes and
to its probability of occurring.
ICAO recog n ized t he i mpor ta nce of
Safety Management System (SMS) and Risk management consists of three essential
recommended that all contracting states elements:
implement the requirement for SMS by
November 2005. It relies on an examination • Hazard identification - Identification of
of all operational aviation hazards and risks undesired or adverse events that can lead
which may impact upon safe operation at the to the occurrence of a hazard and the
airport. This includes unwritten risks such as analysis of mechanisms by which these
human factors, cultural and environmental events may occur and cause harm. Both
risks. The potential benefits of SMS are reactive and proactive methods and
generally recognized throughout the world techniques should be used for hazard
to the extent that many ICAO nations identification.
now require commercial operators to have • Risk assessment - Identified hazards are
documented safety management systems. assessed in terms of criticality of their
In this regard ICAO emanates the Safety harmful effect and ranked in order of
Management Manual (Doc 9859, 2009). their risk-bearing potential. They are
assessed often by experienced personnel,
A basic process within the SMS is risk or by utilizing more formal techniques
management; it is composed by description and through analy tical ex per tise.
of the system, identification of the hazards, The severity of consequences and the
assessment of the risk, analysis of the risk, likelihood (frequency) of occurrence
and control of the risk (FAA, 2007). of hazards are determined. If the risk
is considered acceptable, operation
The objective of risk management is to continues without any intervention,
obtain an understanding of how to access if it is not acceptable, risk mitigation
the various levels of hazards and to gain an process is engaged.
insight on logical approaches to deal with • Risk mitigation - If the risk is considered

63
 Distefano N. et al. Risk Assessment Procedure for Civil Airport

to be u nacceptable, t hen cont rol Risk is defined as a measure of the threat to

measures are taken to fortif y and safety posed by the accident scenarios and
increase the level of defenses against their consequences (Preyssl, 1995).
that risk or to avoid or remove the risk,
if this is economically feasible. T he model i ng approach adopted for
the quantitative assessment of the risks
This paper proposes a risk assessment process associated with runway operations involves
that employs quantitative methodologies, several methodological steps which are
which are based on historical data of defined as:
aircraft accidents. In order to determine
the probability of the accident it has been • identification of hazardous conditions
adopted a quantitative technique based on and accident scenarios,
the combination of the likely of occurrence • determination of probabilities of the
of all causes attributed at the accident; accident identified; and
differentiating between typical causes the • definition of consequences of such
site in question (i.e. presence of snow on an accident (fatalities and aircraft
the runway, crosswind, etc.) and those damages).
independent of the particular site (i.e.
outbreak of pneumatic, inadequate crew 2.1. Data Description
competence, etc.). In order to define the
severity of each accident it has been adopted One of the key factors that can affect the
a quantitative model, which is based on the results of a quantified risk assessment is the
study of the consequences of aircraft accident quality of the base data used in the analysis.
from 1 January 1980 to 31 December 2010.
Historical data is the most appropriate source
The results of the current study can be to utilize as it reflects the actual situation
used by a broad range of civil aviation within the industry being assessed. However,
organizations for risk assessment and cost- such data can be extremely difficult and
benefit studies of actions improvements. time consuming to obtain and interpret.
Applied to any specific airport, the analysis One of the principal reasons for such
approach for risk assessment developed in problems relates to the manner in which the
this study will allow users to determine if information is recorded as it rarely matches
the risk is relatively high or low and whether the analyst’s requirements.
there is a need for risk management action.
The approach adopted for this study has been
2. Methodology to utilize the historical data and supplement
any deficiencies with data from generic
R isk assessment is the process which sources. The failure rate data used in the
associates “hazards” with “risks”. When we analysis has been based on the collection
know the various impacts a hazard may have and analysis of reported accidents contained
on our mission and an estimate of how likely in the records of the ANS (Aviation Safety
it is to occur we can now call the hazard a Network) database which is considered
risk. according to the following criteria:

64
International Journal for Traffic and Transport Engineering, 2014, 4(1): 62 - 75

1. Occurrence in the period 1980-2010; The data used in this study were collected
in a database.
2. Serious incident and incident are
excluded;
Using such criteria, 1749 accidents and
3. Accidents to helicopters, militar y serious incidents were selected to compose
aircraft and tourist flights are excluded; the information that was used for developing
the risk models.
4. Accidents only during final approach,
landing, go-around, take-off, initial Fig. 1 shows a screenshot of the database,
climb and parking were analyzed; and from which it is possible to infer the
information collected for each event and
5. Sabotage, terrorism and military actions the Table 1 summarizes the number of events
are excluded. of the database for flight phase.

Fig. 1.
Screenshot of Database

The database includes, for each individual a i rc ra f t d a m age), t he f l ig ht n at u re

event, the f light phase, the location, the and causal or contributing factors and
aircraft type, the runway and environmental parameters required to develop the risk
conditions, the consequences (fatalities, models.

65
 Distefano N. et al. Risk Assessment Procedure for Civil Airport

Table 1
Events of Database for Flight Phase
Parking Taxi Take-off Landing Total
1980-1984 35 16 83 170 304
1985-1989 23 15 104 183 325
1990-1994 28 12 108 220 368
1995-1999 21 14 86 240 361
2000-2004 59 35 114 243 451
2005-2010 72 32 118 368 590
Total 238 124 613 1424 2399

2.2. Identification of Hazardous It was analyzed the historical data contained

Conditions and Accident Scenarios in the database created in order to define
the aircraft accidents most likely. They are
A hazard is any condition, object, activity or classified into seven scenarios:
event with the potential of causing injuries
to personnel, damage to equipment or • Landing overshoot;
structures, loss of material, or reduction • Landing veer-off;
of ability to perform a prescribed function. • Take-off veer-off;
• Landing overrun;
In order to identify the hazards inside airport, • Take-off overrun;
the scenarios were defined in terms of • Ground collision with other aircraft in
consequences, not the causes of loss of control. landing; and
• Ground collision with other aircraft
Looking at absolute values by phase of in take-off.
flights, aerodrome can be seen as the critical
location where efforts have to be constantly The definition of each incident type is
performed to maintain a uniform high level summarized as follows, consistent with
of safety with the involvement of different ACRP (2008):
types of actors on aerodrome platform.
• The overrun accident is an accident
The purpose of this step in the context of during a landing or an aborted take-
this study was to determine relevant causal off, when the pilot is unable to prevent
factors of aircraft accidents and hazards to the aircraft from leaving the paved
aircraft associated with airport operations surface of the runway from its ends. It
(e.g., landing, takeoff roll, and associated is a “longitudinal deviation”, that is the
fault sequences). longitudinal distance traveled beyond
the accelerate/stop distance available (for
The outcome of the hazards identification takeoff events), and beyond the landing
process has the form of a list of hazards; distance available (for landing events).
this hazards logging is useful for subsequent • T he undershoot accident is a
analysis. “longitudinal deviation”, that is the

66
International Journal for Traffic and Transport Engineering, 2014, 4(1): 62 - 75

longitudinal distance the aircraf t and the other in the air close to the
undershoots the intended runway ground, both take-off and landing.
threshold.
• The veer-off accident, both take-off Table 2 shows the number of events of
and landing, is a overruns in which the database for accident scenarios and their
aircraft leaves the side (as opposed to consequences (fatalities/occupants and
the end) of the runway. It is a “lateral aircraft damage). The ranking of aircraft
deviation” that is the lateral distance to damage consist of tree classifications: Minor
the extended runway centerline. (M) if there was no damage to the aircraft;
• The ground collision with other aircraft Substantial (S) if there was substantial
is an event where aircraft collide on the damages to the aircraft; and Total loss (T)
runway or while one is on the ground if the aircraft was totally destroyed.

Table 2
Events of Database for Accident Scenarios and their Consequences
N° Aircraft damage
Accident scenarios Fatalities/occupants
events T S M
Landing veer-off 147 124/5379 35 106 6
Landing overrun 120 208/5368 33 76 11
Landing overshoot 43 15/1127 12 31 0
Landing collision 1 2/3 1 0 0
Take-off veer-off 40 176/ 715 14 25 1
Take-off overrun 40 98/987 12 28 0
Take off collision 5 125/214 4 1 0

2.3. The Causes of the Hazards apron) surface conditions, environmental

conditions, human factors (Janic, 2000.)
The causes of an accident are factors, events,
acts, or unsafe conditions which singly, or In order to determine the main causes of
in combination with other causes, result in each hazard, there were analyzed all events
the damage or injury that occurred and, if of the database.
corrected, would have likely prevented or
reduced the damage or injury. In this study the causes were divided in
four categories: environmental conditions,
Investigating causes of aircraft accidents is runway conditions, aircraft performance
difficult because they generally stem from characteristics, and human factors.
a complex system of mutually dependent,
sequential factors (Owen, 1998). The main Table 3 shows, for each type of accident and
causes of air accidents can conditionally for each potential cause, the percentage of
be classified into: aircraft performance events of the database in which the cause
c h a rac ter i st ic s, r u nw ay (t a x iw ay or was present.

67
 Distefano N. et al. Risk Assessment Procedure for Civil Airport

This data covers the whole range from serious they can be considered to be reasonably
incident to accident of the database, therefore representative for all civil airports.
Table 3
Percentage of Accident in Present of the Cause
Take
Ground Ground
Landing Landing Landing off Take off
collision collision in
short Veer off overrun Veer overrun
in landing take-off
off
Heavy rain 14,89 18,10 19,01 2,78 1,85 2,62 6,67
Environmental conditions

Windshear - 1,72 - - - - -
Tailwind - 1,72 4,23 8,33 1,85 - -
Crosswind 6,38 16,38 3,52 - - 7,21 6,67
Bird strike - 1,72 - - - - -
Fog - 5,17 5,63 - 1,85 9,29 20
Snow 2,13 6,03 0,70 5,56 - - -
Ice - - 2,11 - 3,70 - -
Vortex shedding - 2,59 0,70 - - - -
Inadequate markers 2,13 0,86 - - - 1,31 6,67
Loss of runway lights - 4,31 0,70 - - 7,21 -
Failure of ILS 2,13 - - - - - -
Runway conditions

Absence of VASI 2,13 - - - - - -

High slope of runway - 0,86 - - - - -
FOD - 0,86 - 2,78 - - -
Loss of friction - - - 5,56 - - -
Aquaplaning - 6,03 18,31 - 9,26 - -
Presence of snow on the
2,13 - - - - - -
runway
Presence of ice on the
- 0,86 1,41 - 1,85 - -
runway
Outbreak of pneumatic - 6,03 - 5,56 - - -
Engine failure 4,26 6,90 1,41 27,78 9,26 3,13 6,67
Thrust reverse failure - 3,45 2,11 - 1,85 - -
Aircraft performance

Flap failure 2,13 - 2,82 - - - -

characteristics

Spoiler failure - 1,72 2,11 - - - -

Brake failure - 1,72 4,23 2,78 1,85 9,29 -
Landing gear failure 2,13 21,55 5,63 13,89 3,70 - -
Loss hydraulic power - 1,72 1,41 - - - -
Loss electrical power - - 1,41 - - - -
Anti-skid system failure 2,13 - - - - - -
High speed - 1,72 7,04 2,78 7,41 - -
Inadequate crew
17,02 30,18 30,17 30,56 25,93 35,45 6,67
Human factor

competence
Inadequate ATC service 4,26 0,86 - - - 9,29 13,33
Inadequate maintenance - 2,59 2,11 2,78 1,858 - -
Incorrect loading of the
4,26 1,72 1,41 5,56 18,52 - -
aircraft
Unknown causes 38,30 25 17,61 2,78 24,07 15,20 46,67

68
International Journal for Traffic and Transport Engineering, 2014, 4(1): 62 - 75

For visually the many potential causes backbone of the fish to each accident. These
defined for each accident fishbone diagrams diagrams are shown in the following figures
were constructed; the causes represent the (Figs. 2-8).

Fig. 2.
Landing Overshoot Fishbone Diagram

Fig. 3.
Landing Veer-Off Fishbone Diagram

Fig. 4.
Take-Off Veer-Off Fishbone Diagram

69
 Distefano N. et al. Risk Assessment Procedure for Civil Airport

Fig. 5.
Landing Overrun Fishbone Diagram

Fig. 6.
Take-Off Overrun Fishbone Diagram

Fig. 7.
Ground Collision with other Aircraft in Landing Fishbone Diagram

70
International Journal for Traffic and Transport Engineering, 2014, 4(1): 62 - 75

Fig. 8.
Ground Collision with other Aircraft in Take-Off Fishbone Diagram

2.4. Hazard Probability Assessment of the National data for the causes belonging
to environmental and runway conditions and
Any quantitative risk assessment requires by analyzing the International data for the
an analysis of the probability of each of the causes belonging to the aircraft performance
potential events that may occur. characteristics and human factors.

In the present risk assessment procedure the Using the databases it is possible to assess the
probability of each hazard is proportional frequency at which each cause determines
to the cumulative probability of the causes a hazard. We assume the frequency as the
identified for the hazard, so in order to value of probability (Eq. (2)). Therefore:
calculate the hazard probability it has been
used the Total Probability Theorem (Eq. (2)
(1)). Therefore:
Where:

(1) NE = number of events occurred during take-

off or landing, in a definite period, generated
Where: by the cause i.

P(E|Ci) = probability that, in presence of N MC = number of take-offs or landings, in

the cause i (e.g. heavy rain), the hazard (e.g. a definite period, occurred in presence of
landing overrun) will occur (Conditional the cause i.
probability).
Through the Eq. (2) the probability that, in
P(Ci) = probability that the cause i (e.g. heavy presence of the cause i, the hazard will occur
rain) will occur. it is not calculated for a specific airport,
in fact report P(E|Ci) to the single airport
P(E|Ci) . P(Ci) = probability that the cause i requires a large amount of data, in order to
will produce the hazard. be statistically significant. Since the available
data is quite poor, we refer P(E|Ci) generally
T he aut hors bel ieve t hat t he P(E|C i) to National airports.
assessment must be carried out by analyzing

71
 Distefano N. et al. Risk Assessment Procedure for Civil Airport

The probabilit y P(C i) of the cause i is Eq. (3) for P(Ci) assessment becomes the
assumed equal to the frequency whit which following (Eq. (4)):
the cause occurred in the airport studied
(Eq. (3)).
(4)
(3)
Where:
Where:
P(C i) = probability that the cause i will
P(C i) = probability that the cause i will occur, for causes of the aircraft performance
occur in a specific airport, for causes of characteristics category.
the environmental and runway conditions
categories. N Fj = number of failures associated with
the cause i suffered by the aircrafts of the
N M(a)i = number of take-offs or landings, type j, in a definite period, to refer to total
in a definite period, in the airport studied number of take-offs or landings, which an
occurred in presence of the cause i (e.g. aircraft makes in the same period (e.g. if
number of landings occurred in presence aircraft of type j had one engine failure in
of snow). 800.000 take-offs then N Fj = 1/800.000 =
1,25·10 -6).
NM(a) = total number of take-offs or landings,
in a definite period, occurred in the airport NM(a)j = total number of take-offs (landings),
studied. which an aircraft of type j makes, in a stated
period, in the airport studied.
The total number of take-offs and landings
occurred in the airport and the number of The probability assessment of the causes
flight movements occurred in presence of belonging to the human factors category
the causes belonging to environmental and (e.g. communication misunderstanding,
runway conditions categories are provided inadequate crew competence, airside driver
by the airport management company. competence, etc.) is very difficult.

The probability of the causes belonging to The probable cause of more than 70% of
the aircraft performance characteristics commercial aircraft hull-loss accidents has
category is not dependent on the airport been cited as “human error”. Today, more
where they occur, so data about failures, for accident/incident investigations have been
each type of aircraft, should be provided by focusing on the human factors in each
airlines companies. operation during flight. This includes flight
crew operations, air traffic control, ground
Considering the composition of the traffic operations, and maintenance operations.
flow of an airport which is to be the object
in the study, based on data provided by the This study doesn’t take into account the
airlines companies it must realize a valid category of human factor because of the
process of weighting. So, in this case, the absence of human factors data.

72
International Journal for Traffic and Transport Engineering, 2014, 4(1): 62 - 75

2.5. Hazard Severity Assessment aircraft can have one or two turboprop
driven or jet engines (sometimes three).
In order to determine the severity of each Maximum gross mass is up to 90.000 lb.
hazard identified in step 1, the present • Commuter aircraft (COM A): usually
procedure uses the results of a previous t w i n eng i ne a i rc r a f t w it h a fe w
study by the same authors (Distefano exceptions such as the De Havilland
and Leonardi, 2012). This study provides DHC-/ which has four engines. Their
the estimation of severity of the aircraft maximum gross mass is below 70.000 lb.
accident associated with airport operations • Transport aircraft (TA):
differentiated for class of aircraft, through • Short-range (S-R): their maximum gross
a methodology based on historical data mass usually is below 150.000 lb.
contained in a database similar to that used • Med iu m-ra nge (M-R): t hese a re
in the present procedure. transport aircraft employed to fly routes
of less than 3.000 nm (typical). Their
The different classes of aircraft operations maximum gross mass usually is below
considered are: 350.000 lb.
• Long-range (L-R): these are transport
• General aviation aircraft (GA): typically aircraf t employed to f ly routes of
these aircraft can have one (single engine) less than 3.000 nm (typical). Their
or two engines (twin engine). Their maximum gross mass usually is above
maximum gross weight is below 14.000 lb. 350.000 lb.
• Corporate aircraft (CA): typically these Fig. 9 shows the results of this study.

73
 Distefano N. et al. Risk Assessment Procedure for Civil Airport

Fig. 9.
Hazard Severity for Class of Aircraft

74
International Journal for Traffic and Transport Engineering, 2014, 4(1): 62 - 75

3. Conclusions References
In the past, only accident or fatality data ACRP. 2008. Report 3 - Analysis of Aircraft Overruns
were investigated and used to measure the and Undershoots for Runway Safety Areas, 60 p.
risk or/and safety level of airlines. This is just
a reactive way to manage the aviation risk. Distefano, N.; Leonardi, S. 2012. A qualitative risk
However, commercial aviation is a complex severity assessment for aircraft accident in civil airport.
mosaic of many varied, yet interrelated 5th International Proceeding SIIV “Sustainability of Road
human, technical, environmental, and Infrastructures”, Roma, 29/31 Ottobre 2012 1-11.
organizational factors that affect safety
and system performance. The possible FAA. 2007. Introduction to Safety Management Systems
influencing factors should be included while (SMS) for Airport Operators. Washington, D.C., AC
assessing risk (Shyur, 2008). 150/5200-37, 22 p.

This paper proposed a new quantitative ICAO. 2009. Safety Management Manual (SMM).
methodology for the assessment of risk Second edition. Doc 9859, 263 p.
in civil aviation. In order to estimate the
probability of each accident proportionally Janic, M. 2000. An assessment of risk and safety in
to the cumulative probability of the causes civil aviation, Journal of Air Transport Management. DOI:
identified for the accident, it was used http://dx.doi.org/10.1016/S0969-6997(99)00021-6,
the Total Probability Theorem. In order 6(1): 43-50.
to determine the severity of each hazard
identified in step 1, the present procedure Logan, T.J. 1999. Trend toward wider sharing of safety
uses the results of a previous study by the data is resisted by industry concerns, ICAO Journal,
same authors (Distefano and Leonardi, 54(1): 7-9.
2012). This study provides the estimation
of severity of the aircraft accident associated Owen, D. 1998. Air Accident Investigation: How Science Is
with airport operations differentiated for Making Flying Safer. Patrick Stephens Limited, Yeovil.
class of aircraft.
Preyssl, C. 1995. Safety risk assessment and management
After safety risks have been assessed though - the ESA approach, Reliability Engineering and System
the present procedure, elimination and/or Safety. DOI: ht t p://d x .doi.org /10.1016/0 951-
mitigation must take place. This is known 8320(95)00047-6, 49(3): 303-309.
as safety risk mitigation. Safety risk controls
must be designed and implemented. These Shyur, H.J. 2008. A quantitative model for aviation
may be additional or changed procedures, safety risk assessment, Computer & Industrial Engineering.
new super v isor y controls, changes to DOI: http://dx.doi.org/10.1016/j.cie.2007.06.032,
training, additional or modified equipment, 54(1): 34-44.
or any of a number of other elimination/
mitigation alternatives.

75