See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/220635544

Airport site selection based on multicriteria analysis: the

case study of the island of samothraki

Article  in  Operational Research · December 2003

DOI: 10.1007/BF02936405 · Source: DBLP

CITATIONS READS

11 3,063

1 author:

Athanasios Ballis
National Technical University of Athens
32 PUBLICATIONS   421 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Systems with Multithematic and Georeferenced Infoemation View project

All content following this page was uploaded by Athanasios Ballis on 09 May 2014.

The user has requested enhancement of the downloaded file.

 Enixetpq61ctrw) "Epeovct/ Operational Research, An InternationalJournal. Vol.3, No.3 (2003), pp.261-279

Airport Site Selection Based on Multicriteria

Analysis:
the Case Study of the Island of Samothraki

Athanasios Ballis
Lecturer, Department of Transportation Planning and Engineering, National Technical University of
Athens, Greece

Abstract
The investigation of alternative sites for the construction of a new airport is a complex task
that involves the cooperation among scientists having complementary disciplines. A typical
investigation comprises a thorough examination of all pertinent factors: existing and foreseen
land use planning, characteristics of wind and weather conditions, operational and safety
aspects, interference with the existing road network, earthworks, construction and
expropriation costs as well as environmental and socio-economic impacts. The preliminary
analysis of the above aspects usually results in more than one alternative solution that satisfies
the above criteria in different ways. The next step is the selection of the most appropriate
solution, where the use of multi-criteria techniques can be of great value. The scope of this
work is to describe such a multicriteria approach, performed within the decision making
process concerning the site selection for the development of a new airport at the island of
Samothraki.

Keywords: Airport site selection, multicriteria decision-making, Analytical Hierarchy Process

1. Introduction
The technological evolution in the field o f aircraft engineering that resulted in the
design of larger, faster and safer aircrafts, in combination with the deregulation of the
air-transport market that has led to a significant reduction of air transport fares and
the supreme appreciation of the modern society for the value of time, have resulted to
the exponential development of air transport sector, during the last decades. The
significant transport demand generated has been converted into requirements for the
262 Operational Research. An Intemational Joumal/ Vol.3, No.3 / September- December 2003

upgrade of airport infrastructure worldwide. For countries, like Greece, where a

significant portion of the population is dispersed in the islands, the air transport
development is of vital importance since it contributes significantly to the abolition of
isolation as well as to the tourist and economic development of the islands. This, in
combination with the tendency of tour -operators intemationally to transport the
majority of tourists by air (80% of the tourists visiting Greece are transported by air)
using preferably direct flights, has led to a continuous demand for the upgrade of
existing airports in the islands as well as for the construction of new ones.
The island of Samothraki is a characteristic example, where the development of an
airport is demanding in order to boost the tourism as well as to offer a vital altemative
of transportation to the inhabitants who are currently depended on sea-transport. The
scope of this work is to present the above airport site selection process focusing on
the multicriteria approach followed for the evaluation of altemative solutions.
In Section 2, the criteria taken into consideration for the selection of an airport site are
described. Section 3, gives an overall description of Samothraki, in respect to specific
characteristics of the island, necessary for a rough identification of possible
restrictions in the selection of the airport site. It also presents the procedures and
actors involved in the decision making process of the airport site selection. The steps
followed for the identification of altemative solutions conceming candidate airport
sites are described in Section 4, while Section 5 accommodates the conclusions.

2. Criteria f o r an airport site selection

Airports are considered significant transportation projects that in most cases have
important impacts on the environment. During the investigation of candidate airport
sites in a given area, a number of environmental, operational criteria as well as land-
use planning and cultural criteria are taken into consideration, [Ashford (1992),
Horonjeff and Mckelvey (1994)] which are as follows:
1. Land-use plan criteria. The existence of residential or industrial areas or even
the existence of constituted (or even officially declared) future residential or
industrial land uses in a candidate area prohibits the development of an
airport on-site.
2. Operational and safety aspects of the airport. This criterion evaluates the
compliance of the proposed alternative solutions, with the international
regulations conceming safety aspectS-for obstructions in the vicinity of the
airport as well as operational aspects due to the impact of wind characteristics
to the operation of the airport, turbulence and microburst effects. Other
operational aspects are imposed due to airport system design (e.g. runways
layout, air traffic control limitations).
A. Ballis / Airport site selection based on multicriteria analysis: The case study of the Island of Samothraki 263

3. Interference with the existing road network. This criterion evaluates the
interference of the foreseen airport construction with the existing road
network, taking into account the degree of restoration (if any) of the existing
road network. In case that far distant alternative airport sites are considered,
the associated access time to and from the airport weighted by the number of
travellers from the various catchment zones of the airport should be taken into
account [Keeny (1973)].
4. Impacts from earthworks. This criterion deals with the earthworks
(embankments and losses) required for the construction of the runway, the
strips, and the runway obstacle-free zones. This criterion is of major
importance, and especially in the case of mountainous landscape, as the
identification of an adequate site with a zone of about 2 to 3 km in length for
the construction of the runway, becomes a rather difficult task.
5. Hydraulic works. This criterion deals with two clusters: hydraulic works
required for the drainage system of the airport and hydraulic works required
for the restoration of the removal of surface runoff due to the alternation of
the existing landscape and the disruption of the water flow through existing
natural watersheds in the area.
6. Structural requirements. This criterion concerns exceptional structural
requirements imposed mainly by the topography at the proposed sites (e.g.
high embankments to align the runway).
7. Impacts on historical or archaeological resources. This criterion evaluates the
impact of the proposed alternative sites to the various archaeological and
historical sites in the surrounding areas.
8. Environmental impacts. This criterion examines the possible interference of
the proposed sites with natural habitats, coasts or protected cultivations in the
area.
9. Air-pollution and noise impacts. The impacts of noise in the airport
surrounding area are measured with the noise patterns according to the NEF
(Noise Exposure Forecast) indicator. This method takes into account aircraft
types, number of flights for a specific time period during the day (including
flights during the night period), etc [Abacoumkin (1990)].
10. Vibration impacts. The vibration from airplane circulation can possibly affect
the structural soundness of antiquities in the vicinity of the candidate airport
site.
11. Impacts on aesthetics that concern the harmonization of the relevant civil
engineering works with the surrounding physical environment.
264 Operational Research. An International Journal / Vol.3, No.3 / September- December 2003

12. Expropriation costs. This criterion evaluates the costs related to the requirec
expropriation of land and existing buildings.
13. Cost of construction. This criterion deals with the estimation of the overall
cost for the construction of the airport. In mountainous landscapes, the cost ot
earthworks for obsessions/obstacles removal and runway-taxiway
construction represents the major cost difference among candidate airport
sites.
14. Socio-economic impacts. This criterion deals with the socio-economic
impacts of the operation of the airport on the community such as relocation of
families and business due to expropriation, changes in employment patterns,
changes in the tax base, requirements for new public services, etc.
The selection of the adequate site for the construction of an airport, at Samothraki,
allows for the demonstration of the customisation/applicability of such criteria in a
real world case study.

3. The case study of SamothraM: Island characteristics, actors

involved and decision making processes
Samothraki is the northern Greek island in the Aegean archipelagos. Its terrain is
mountainous and covers an area of 180 sq.km with a shoreline of 32 km and a
population of about 3.000 inhabitants. Forestland covers one third of the island and is
developed around the mountain of Saos, which has a height of 1,700m. Figure 1
presents the west part of the island where the efforts for the identification of an airport
site were focused, as the sharp mountainous landscape in the rest of the island,
prohibits from an airport site selection investigation. Natural habitats are concentrated
in the northwest part of the island. One of the main watersheds of Samothraki is that
of Rodifili, and is situated it the middle-west part of the island. It is part of a wide
basin that concentrates surface runoffs. During the winter period, when significant
rainfalls occur, these flows are very intense, and fall into the sea. Within this area
numerous oak trees are scattered. Except from its natural beauties and dense greenery,
Samothraki is also well known for its ancient past. The ancient city near the village of
Palaiopolis (at the north part of the island) where the Temple of Great Gods is located
is considered a significant attraction pole for tourists.
The main road network of the island has a length of 65 km. It connects the port of
Kamariotisa with Chora, which is the main town of Samothraki. A branch of this
network forms an interchange of T shape, connecting the Kamariotisa-Chora road
with the south part of the island. A shipping line links the port of Kamariotisa to the
(29 nautical miles distant) mainland Port of Alexandroupolis. In case of stormy
weather/sea condition, this link is cut-off.
 A. Ballis / Airport site selection based on multicriteria analysis: The case study of the Island of Samothraki 265

In order to reduce the isolation and improve the tourist attractiveness of the island, the
construction of an airport is considered. Figure 2 presents the actors and procedures
involved in such a decision making in Greece. The bureaucratic procedures are
initiated by Civil Aviation Authority (CAA) but require the approval of Ministries of
Transport and National Economy. Then CAA assigns the study for the airport site
selection and the feasibility of the project to an Engineering Firm (that should include
scientists from various specialisations). The conclusion/deliverables of the relevant
study should be approved by an authorised committee from the CAA Department of
Planning as well as by the CAA Department of Environment that may accept the
proposed technical solution (airport site selection and relevant justification) or request
for modifications. After its approval from both CAA departments, the study is
forwarded to the Department of Environment of the Ministry of the Environment,
Physical Planning and Public Works (MINENV). This Department forwards copies of
the study to various organisations/Ministries (e.g. Ministry of National Defence,
Ephorates of the Ministry of Culture), asking for their point of view. Based on the
relevant feedback as well as on the evaluation performed by the Department itself, the
proposed airport site selection is approved or returned back to the study team for
modifications. A "hidden" but active actor, in this process is the local community that
through its representatives provides information at technical level as well as
argumentation at political level promoting the local interests.
The above procedure reveals a three stage decision-making process: In the first stage,
scientists from various disciplines have to conclude to one proposed solution, or at
least at a set of hierarchically sorted alternative solutions. In the second phase, the
study team and the CAA staff/engineers should settle in a commonly accepted
proposal. In the third phase, this proposal should be accepted/modified according to
MINENV staff/scientists. This multistage decision-making imposes many
inconveniences as each of the above actor groups (study team, CAA team, MINENV
team) usually weights the criteria in a relatively focused way. The study team usually
focuses on technical/organisational issues; CAA team tries to compensate
technicalities, cost and social impacts while MINENV team focuses solely on
environmental issues. The procedure is usually smoothed thanks to 2 interfaces (see
Figure 2): The project manager of the study team (that "brings" the CAA point of
view in the study team) and the CAA department of environment (that knows the
MINENV group background and "mentality"). All judgments/evaluations performed
in the above "chain of decisions" are usually made according to the "engineering
judgment".
In the case of Samothraki, in addition to the above "conventional" evaluation, the
Analytical Hierarchy Process (AHP) was implemented. The AHP is a multiple
criterion evaluation methodology according to which complexity is structured in a
hierarchy, pairwise relative comparisons are made, and redundancy of judgements is
used to improve accuracy and deal with fuzziness. According to this method, weights
266 Operational Research. An International Journal / Vol.3, No.3 / September- December 2003

or priorities are not arbitrarily "assigned", but are derived from a set of judgements,
either verbal or numerical, using a specific scale. Elements are compared to their
peers at each level of the hierarchy. A comparison between a pair of peers is made
about the relative importance, preference, or likelihood of the elements with respect
to their parent element. Thus, a model should be constructed in such way that peers
are approximately within an order of magnitude of one another [Forman (1990), Saaty
(1994)].

The method requires the formulation of a A matrix that contains the relative weights
of the criteria. Since:

-w, ]
-w~lw,
w2lw,
w, lw2 w~lw3 ..... w, lw.
w2lw~ w~lw~ ..... w~lw. w2 I- nw 1
nw2
w~/w, w~lw~ w~/w3 ..... w~lw.

wo/w, w./w2 w./w3 ..... w./wo

A w = ?/w

and we knew __A,but not % we could solve the above equation for w. The problem of
solving for a non-zero solution to this set of equations is known as an eigenvalue
problem [Forman (1990)].
The same process is then followed for the performance expressions of the alternative
solutions, where the pairwise comparison is made among solutions, per criterion. The
summary of weighted, performances (also called decision alternatives scores) yields
the final score of each alternative solution and the selection of the most appropriate is
possible.
The AHP allows for small errors and inconsistencies in judgment, using the
consistency index to measure such deviations: perfectly consistent judgments produce
a consistency index of 0 while a value of 1 indicates almost random selections. A
value of 0.1 (inconsistency ratio 10%) or less is usually considered "acceptable"
[Forman (1990)].
The AHP has been applied to a wide range of problems [Zahedi (1986), Zanakis et.
al. (1995), Golden et al. (1998)] including the assessment of transport projects [Saaty
 A. Ballis / Airport site selection based on multicriteria analysis: The case study of the Island of Samothraki 267

(1977a), Saaty (1977b), Saaty (1990), Giotis et. al. (1999), Saunders (1994)]. It must
be noted that the case of Samothraki is not a typical site selection case but rather a
very difficult one, thus offers a good test for the applicability of AHP method in
airport site selection projects.

4. Identification of candidate airport sites

As already mentioned, the identification of a new airport site is a difficult task that
involves the consideration of multiple criteria. Taking into account the specific
characteristics of the area that will accommodate the airport project, some of these
criteria become higher in priority in relation to other criteria. This is also the case of
Samothraki, where the decision for the airport site was based on a systematic analysis
of all pertinent criteria performed within a study that aimed to investigate alternative
sites and identify the most suitable [OBK (2003)]. All relevant decisions were based
on the engineering judgment of the members of the study-team and of the evaluators
of the above-mentioned governmental bodies involved. In addition to the above
procedure, AHP was performed, based on the same criteria. The aim was to identify if
AHP leads to the same decisions. The above AHP-based evaluation has not been
included in the project deliverables (no such contractual obligation existed) and is
presented in the current paper.
Both selection approaches (engineering judgment and AHP) were based on a
systematic investigation of all alternative sites which were evaluated according to a
set of criteria. The above process was performed in 3 steps which are described in the
following paragraphs.

4.1 Step 1: Exclude alternatives based on land use and wind

characteristics
In the first step, three criteria (land use, existence of historical or archaeological
resources, wind characteristics) guided the identification of the candidate sites.
The land use criterion excluded all residential areas and town/villages where planned
expansions exist. Nevertheless, sites containing a very limited number of buildings
(given that the land and the existing buildings will be expropriated) were not
excluded. Another aspect of land use was the existence of a wind generator
installation (see Figure 1) that is conflicting with the airport operations.
Many historical resources exist in the island, including the ruins of an ancient city
near the village of Palaiopolis where the Temple of Great Gods is located. All sites of
archaeological interest that are officially characterized as protected areas were
excluded from the airport candidate sites. On the contrary, sites having surface
268 Operational Research. An International Journal / Vol.3, No.3 / September - December 2003

archaeological findings were considered (given that an archaeological research will

be performed to confirm the lack of other antiquities).
Wind characteristics are a major criterion for an airport site selection. Among the
various factors taken into consideration in the determination of the runway direction,
is the direction of the prevailing winds in the area. Aircraft may not manoeuvre safely
on a runway when the wind contains a large component at right angle to the direction
of travel (called crosswind component). Therefore, runways are oriented in the
direction of prevailing winds. Standards of ICAO and FAA agree that runways should
be oriented so that the percentage of time during which the use of the runway system
is not restricted because of an excessive crosswind component (usage factor) is not
less than 95%. A graphical procedure utilizing a wind rose is typically used to
determine the "best" runway orientation as far as prevailing winds are concerned. The
wind analysis required for the identification of the "best" runway direction, includes
wind distribution statistics for at least five years. These wind data are arranged
according to velocity, direction and frequency of occurrence, in order to provide
information for the time percentage that wind velocities within a certain range and
from a given direction can be expected [Ashford (1992)]. No differences exist among
the candidate sites as far as airport capacity is concerned.
In addition, turbulence and microburst effects should also be considered. One of the
most common aviation hazards and sometimes the most damaging is clear-air
turbulence, which can occur even when no rain or other adverse weather conditions
are present. It occurs when a windstorm passes down a steep, rough mountainside
forming a layer of air that often turns suddenly upwards and begins to rotate in
circles. As these "rotors" multiply, they form a series of more violent, spinning air
masses, which can rise up to altitudes of 30,000 feet or more, about the normal
cruising height for most airliners. For this reason the identification of relevant
topography extremities (rough mountainsides, severe slopes) should be taken into
consideration for the selection of the airport site. Thunderstorms and other highly
unstable atmospheric events can cause wind shear, a sudden, rapid change in wind
velocity or direction. The most dangerous form of wind shear is a microburst, which
occurs when the cold air above clouds or thunderstorms falls rapidly to the ground
and fans out in all directions. A plane approaching a microburst experiences
increasing headwinds and a turbulent altered flight path, and as it flies through the
microburst, it may experience increasing tailwinds and loss of lift.
Based on the above 3 criteria, 21 potential airport sites were identified through the
"scanning" of the defined terrain.
 A. Ballis / Airport site selection based on multicriteria analysis: The case study of the Island of Samothraki 269

4.2 Step 2- Exclude alternatives with high construction cost or socio-

economic cost
The above 21 candidate sites were comparatively evaluated in the second step of
investigation according to the following criteria: cost (airport construction and
earthworks to eliminate the obstructions in the vicinity of airport) and socio-economic
impacts due to expropriation of land and buildings.
The elimination of the obstructions in the vicinity of the airport is of major
importance. Airports must be sited in areas where airspace is free from obstruction
that could be hazardous to aircraft turning in the vicinity or on takeoff or approach
paths. It is also necessary to maintain the surrounding airspace free from obstacles,
preventing the development and growth of obstructions to airspace that could cause
the airport to become unusable. Protected airspace around airport is made up of five
principal imaginary surfaces illustrated in the upper part of Figure 3. These surfaces
are namely: the Primary Surface, the Approach Surface, the Horizontal Surface, the
Transition Surface and the Conical Surface [Ashford (1992)]. The selection of a
candidate airport site, should ensure that the above mentioned surfaces are free of
obstructions. The lower part of Figure 3 presents the output of a special software,
used to identify and quantify the required obstacle elimination (through excavation).
By use of special software, the earthworks for the alternative sites were identified and
the results were used for the rating of the relevant criterion.
The "socio-economic impacts" criterion usually refers to the positive and negative
effects of an airport construction to the (local and/or national) economy and society.
All alternative solutions have equal impacts (as the development of an airport is
anticipated to serve the community interests and the economic activity of the island).
On the contrary, the expropriation of land and erections required for the airport
construction have (in addition to the cost aspect) negative social impacts, thus the
required land and erections expropriations were identified for the 21 alternative sites.
Based on the above information the following selection procedure was applied:
o All sites having extremely high construction cost, were excluded.
o All sites requiring extensive land and erections expropriation were excluded.
o If two neighbouring sites had similar cost of construction, the one with the higher
socio-economic impacts was excluded.
The above process led to four candidate sites which are presented in Figure 1.
It must be noted that the above Step 1 and Step 2 site selection play a role of filter that
protects AHP method from erroneous selections: the criteria related to the operational
and safety aspects (expressed through the obstructions in the vicinity of airport and
wind characteristics effects) as well as the criterion for the historical/archaeological
impacts are of paramount importance and cannot be weighted against other criteria.
270 Operational Research. An International Journal / Vol.3, No.3 / September - December 2003

Nevertheless, when an alternative site complies with certain minimum requirements,

the remaining criteria are taken into account (recall the indifferent threshold of
PROMETHEE method or the veto of ELECTRE method).
4.3 Step 3: Selection of the best alternative
As already mentioned, the selection of the airport site was performed by engineering
judgement as well as by AHP. In both cases the same criteria were used. Table 1
presents the criteria and the associated rating of the initial pairwise comparison matrix
(as it is explained later on, additional criteria were introduced leading to new results).
More specifically, Group B contains the 14 criteria taken into account in the initial
phase of the evaluation (i.e. before the study team interacts with the CAA and
MINENV engineers). Such a big matrix necessitated too many pairwise comparisons.
A way to reduce the computational effort could be to cluster some criteria (e.g. wl
and w2, w3 and w4 as well as w13 and w14) creating criteria-subcriteria structures,
that allow to reduce the matrix to 11X11 plus the 2X2 sub-matrixes. Nevertheless, in
the specific case study, the clustering of criteria is not suitable as each criterion has
quite different effects: The wl "effects in n-w area" is related to the future expansion
possibilities of the city. The w2 "wind generators" is related to the limitations
imposed to wing generators operation due to airport construction. The w3 "volumes
transferred" is related to the landscape effects due to earthworks while the w4
"transferring distances" reflects mainly technical inconveniences. The "high value
real-estate" (w13) concerns mainly money for expropriation while the w14 "number
of erections" has also significant social effects. Therefore, the clustering of the above
criteria could reduce the workload but complicates the comparison, thus it was
abandoned. Another way to reduce the computational effort could be to employ
Harker's incomplete AHP algorithm [Carmone et. al. (1977)]. However, this
technique was not implemented in the case of Samothraki, as the trade-off between
the workload reduction and the accuracy reduction were considered worthless.
The criteria of groups C and D were expressed and taken into account in later phases
of the evaluation (see below).
The formulation of this Table was not an easy task. Many scientists of the study team
were not familiar with the whole spectrum of the impacts and avoided the pairwise
comparison among all criteria. Even when they did, they had a tendency to
overestimate the importance of the criteria related to their specialisation. The Table
was finally formulated by two engineers of the study team that had an overall view of
all criteria and took into account the evaluations of the other members expressed for
selected criteria. The rating process was as follows: The first engineer (who was the
project manager) proposed a rating and the second one agreed or disagreed providing
the relevant argumentation. The final rating was decided by compromise.
 A. Ballis / Airport site selection based on multicriteria analysis: The case study of the Island of Samothraki 271

The compliance of the four alternative sites with these criteria is described below:
The first criterion (wl) concerns the compliance of the four alternatives with the
land--use plan. An airport located near the inshore areas of the island, will restrict the
construction of hotels and other poles of tourist attraction. This is quite undesirable,
as the number of beaches in the island of Samothraki is limited. Taking into
consideration the above, the alternative solutions I and II, are most favoured as they
propose the development of the airport far from the inshore areas, thus allowing a
balanced regional development of the island. As already mentioned, the second
criterion (w2: wind generators) is related to the limitations imposed to wing
generators due to airport operations. Wind generators are "obstruction in the vicinity
of the airport". For the airport sites III and IV, the regulations require that wind
generators should be eliminated while for the airport sites I and II advise that wind
generators should preferably be eliminated. This means, that wind generators could be
preserved, under specific safety measures. Concerning the criterion "earthwork
volume transferred" (w3) the rating of the alternative solutions was based on the
output of the relevant earthwork calculation (in m 3) where differences up to 33% were
identified. It must be noted that this criterion reflects the relevant construction costs
as well as the associated environmental impacts. The forth criterion (w4) "transferring
distances" reflects cost but also technical inconveniences during the construction
phase. Impacts on river-beds (w5) occur when the airport's runway, blocks the water
flow to the sea (worst case) or alters the river-bed alignment. Hydraulic works can be
used to overcome these problems but there are negative environmental effects and
risk of failure (e.g. a heavy rain that exceeds the design standards) that are expressed
by the criterion w5 as well as cost and technical problems, expressed by the criterion
w7 "Hydraulic works". Similarly, the criterion w8 "high embankments" reflects cost
but mainly technical difficulties associated to such non-typical constructions.
Another very important criterion considered was the interference of the future airport
with the existing road network (w6). The alternative solutions I, II and III are partly
interfering with the existing road network, while the alternative solution II (in the
central of the area), significantly affects both branches of the road network. Although
a ring road will allow the access to the areas in the periphery of the airport, other
problems arise: the modification of the existing road network, will consequently lead
to alteration in the real estate values (and consequently to reactions from many land-
owners). In fact, this constitutes another criterion (see Table 1, criterion "Real-estate
status-quo") that in the case of Samothraki has played a very significant role in the
airport site selection (see below).
The w9 criterion takes into account the impacts to environment (except from the
impacts due to earthworks, river-beds, noise etc, which are expressed individually).
Due to the absence of studies for the air-pollution and noise impacts (wl0), the distance
from the city of Kamariotissa was used as indicator. Similarly, the wl 1 "vibration" criterion
272 Operational Reseamh. An International Journal / Vol.3, No.3 / September- December 2003

was taken into account through the indicator: distance from the Temple of Great Gods
antiquities.
In relation to the "impacts on aesthetics" (w 12) criterion, the alternative solutions IV
and III are considered less favourable, since they have a significant intervention with
the existing landscape, as a high embankment, will be formed for the construction of
the airport. The alternative sites I and II better comply with the existing topography,
thus impose smoother interventions to the existing landscape.
Two"expropriation" criteria were defined: (a) The w13 addressing the impacts on
high-value real estate (reflecting cost but also impacts to real-estate where demand for
other activities exist), and (b) the expropriation of erections (houses, stores etc) where
-as already mentioned- great socio-economic impacts usual occur.
It must be noted that as the major elements of the construction cost of Samothraki
airport (earthworks, hydraulic works, expropriation cost) are taken into account by
other criteria (the remaining cost for the runway and air-terminal is equal for all
alternative sites) there is no criterion related to the total construction cost, to avoid
double counting.
Table 2 presents the AHP computation that ranks site II as the best altemative
(decision altemative score U2 > U4 > U3 > U1). All consistency indexes are bellow the
0.1 limit (except the one in j 14, but the difference is insignificant). In order to
investigate the sensitivity of the site scores, the relatively higher { Wj x aij } elements
(in other words when a heavily weighted selection criterionj is combined with a high
performance score of the alternative site/) should be altered. For site II, for example,
such strong elements are the earthwork volumes (w3), the high embankments (w8) and
the n-w area impacts (wl). As the difference between U2 and U4 is big enough, the
selection of site II seems quite robust.
The same result (selection of site II) was the outcome of the selection procedure
based on the engineering judgment. Nevertheless, the selection process has not
ended: when the study results were presented to the local authorities, a new criterion
arose: The status quo of the real estate in the island that is affected by the road
modifications. Based on this criterion, site I should be selected. The final outcome of
the study was the ranking of the four altematives as follows: Sites I and II were
ranked equal, higher than III and IV. The study was submitted to the CAA
Department of airport Planning and to the Department of Environment. The
Department of airport Planning was in favour of site I. One main reason for this
decision was that site I was the preference of the local community. The Department of
Environment was in favour of site II due to its relatively low impacts on the
environment. Finally, the site I was selected and the study was submitted to the
Department of Environment of the MINENV for approval. There, focus was given to
the criterion related to river-beds (Table 1, criterion w5) and in order to accept site I,
additional studies were required for the estimation of the environmental impacts, thus
 A. Ballis / Airport site selection based on multicriteria analysis: The case study of the Island of Samothraki 273

the study should be modified and resubmitted. This requirement is interpreted as: the
approval will delay for about two years. That introduced an additional criterion: The
duration until the approval of the project. When the whole "picture" was known, the site
II was selected by all parties (local authorities, CAA, MINENV).

5. Conclusions
The airport site selection process is complex as multiple criteria have to be satisfied
and therefore the implementation of multicriteria methods can be of great value in the
decision process. The case of Samothraki offered a good test for the applicability of
AHP method in airport site selection projects. The relevant study was based on a
systematic three-step analysis of the pertinent criteria. The decisions were based on
the engineering judgment of the study-team members and the evaluators of the
governmental bodies involved. In addition to the above procedure, AHP was
performed, based on the same criteria. Both approaches (engineering judgment of the
study-team and AHP) concluded in the same airport site. Nevertheless, in each of the
following stages of the procedure (comments by the local authorities, approval by the
Civil Aviation Authority, approval by the Ministry for the Environment) the selection
criteria were weighted differently, but also new criteria arose.
The lesson learned from this AHP application is that the site selection procedure can
probably be simplified by introducing -at the early stages of the process- a round table
where all parties involved will have the opportunity to present their selection criteria
as well as the relevant argumentation. In such a meeting, AHP can probably be
implemented, as it can -at least- contribute to a better understanding of the site
selection problem through the pairwise comparison of the criteria.
Acknowledgments
The author would like to acknowledge the anonymous reviewers for their valuable
comments and information provided.

References
Ashford, N. and Wright, H., P. (1992). Airport Engineering, A Wiley-interscience
Publication, John Wiley & Sons, Inc., 3~ Ed.
Abacoumkin, C. (1990). Airports, Symmetria Ed., Athens, Greece.
Carmone, J., F., Kara, A. and Zanakis, H., S. (1977). A Monte Carlo Investigation of
Incomplete Pairwise Comparison Matrices in AHP. European Journal of
Operational Research, Vol. 102, pp. 538-553.
274 Operational Research. An International Journal / Vol.3, No.3 / September - December 2003

Forman, H., E. (1990). Multi Criteria Decision Making and the Analytic Hierarchy
Process, Readings in Multiple Criteria Decision Aid, pp. 295-318, Carlos A. Bana
e Costa (Ed.), Springer-Verlag Berlin, Heidelberg.
Giotis, G., Tsamboulas, D. and Roilos. E. (1999). Applicability of quantitative and
qualitative criteria in the multicriteria evaluation of transport projects, Technica
Chronica, Vol.3, Athens, Greece, (in Greek. Extended summary in English)
Golden, B., L., Wasil, E., A. and Harker, P., T. (1998). The Analytic Hierarchy
process: Applications and Studies. New York, NY: Spinger - Verlag.
Horonjeff, R. and Mckelvey X., F. (1994). Planning and Design of Airports.
McGraw-Hill, Inc., 4th Ed.
Keeny, L., R. (1973). A Decision Analysis with Multiple Objectives: The Mexico City
Airport. The Bell Journal of Economics and Management Science, pp. 101-117.
O' Brien Kreizberg (OBK). (2003). Airport site selection for the island of
Samothraki, Civil Aviation Authority, Athens, Greece.
Saaty, T., L. (1977a). Scenarios and Priorities in Transport Planning: Application to
the Sudan, Transportation Research 11/5.
Saaty, T., L. (1977b). The Sudan Transport Study. Interfaces 20/3, pp. 147-157.
Saaty, T., L. (1990). The Analytic Hierarchy Process: Planning, Priority Setting,
Resource Allocation, RWS Publications, USA.
Saaty, L., T. (1994). How to Make a Decision: The Analytic Hierarchy Process.
Interfaces, Vol. 24(6), pp. 19-43.
Saunders, H., J. (1994). A Comparison of Decision Accuracy in the Analytic
Hierarchy Process and Point Allocation. Third International Symposium on the
AHP, Washington D.C.
Zahedi, F. (1986). The Analysis Hierarchy Process- A Survey of the Method and its
Applications. Interfaces, Vol. 16, pp. 96-108.
Zanakis, H., S., Mandakovic, T., Gupta, K., S., Sahay, S. and Hong, S. (1995). A
Review of Program Evaluation and Fund Allocation Methods Within the Service
and Government Sectors. Social-Economic Planning Science Vol. 29, pp. 59-79.
o~ ~

~ ~~ r~
o

.~- ~

~o 0

o~

-- 0-~ 2
.=.
O

~ <~

<
~

<
 oFHI o
~ddd~

(",I
('i.
II

00

0
II

~ ~ ~i(~ ~

0
0

?1i1:1-I
/ollol
o
II
, ,..,~

i gddd

.IL

II
Figure 1: Candidate area and limitations for the selection of the airport site
 o

E
~
o

o
~J

9~ o

o5

~J
(J

o
~J

,4
;u
 Figure 3: "Free from obstacles" principal imaginary surfaces (upper part) and output of a
software used to identify and quantify the required earthworks

View publication stats