Environmental impact assessment methodological frameworkfor liquefied natural gas terminal and transport network planning

Maria P. Papadopoulou, Constantinos Antoniou n

National Technical University of Athens, Greece

H I G H L I G H T S

� Determined the optimal location for an LNG terminal in Cyprus.� REGIME multi-criteria analysis used to prioritize alternative LNG terminal locations.� Multiple modes of transportation connections were evaluated and geometric alignments were proposed.� Environmental impact assessment and validation was undertaken based on a structured questionnaire and an expert panel.� Parameters such as safety, existing infrastructure, and access were also considered.

a r t i c l e i n f o

Article history:Received 15 November 2013Received in revised form24 January 2014Accepted 25 January 2014Available online 16 February 2014

Keywords:Environmental planningLiquefied natural gas (LNG)Transport network designEnvironmental impact assessment (EIA)REGIME multi-criteria decision analysis

a b s t r a c t

The recent discovery of significant offshore natural gas reserves in the Aphrodite field, south of the islandof Cyprus in the Mediterranean Sea, changes the energy landscape in the greater Mediterranean-MiddleEast-Caucasian Region. In this paper, different alternative locations for the construction and operation ofa liquefied natural gas (LNG) terminal station in Cyprus were evaluated, explicitly considering also theirconnection to the power generation station of Mari and the country's gateway.

The problem of determining the optimal location for an LNG terminal in Cyprus has been approachedusing multiple methodological components, which consider environmental and transportation issues,both technocratic in nature, as well as more subjective and based on expert opinion. The first step was aREGIME multi-criteria decision analysis used to prioritize alternative LNG terminal locations. Then,multiple modes (railroad and pipeline) of transportation connections were evaluated and geometricalignments were proposed, considering a multitude of restrictions. Finally an environmental impactassessment based on a structured questionnaire and an expert panel was conducted to validate andassess the impact of the alternative options (combination of location and transportation mode androute). During the evaluation process parameters such as safety, existing infrastructure, and access werealso considered.

& 2014 Elsevier Ltd. All rights reserved.

1. Introduction

Energy adequacy is universally recognized as a crucial para-meter for the sustainable development of a nation. The mainobjective of an administration regarding the energy policy of anation is to ensure balance between total energy demand andsupply. This goal will be better achieved if the energy “mix” isobtained involving analysis of existing and potential energysources in two ways by: (a) considering how future supplies can

be guaranteed and (b) identifying future energy needs for varioussectors (Verma, 2007). In order to maintain availability of energysupplies at affordable prices the energy policy of a nation shouldbe based on the diversification rule of source supply. The constantincrease in oil and gas demand requires the exploration, develop-ment and distribution of new energy sources (Tavana et al., 2012).

In late 2011, the Noble Energy Corporation announced thediscovery of 5–7 trillion cubic feet of natural gas in the Aphroditefield (Block 12) of the Exclusive Economic Zone of Cyprus. Thediscovery is estimated to be worth billions and it could cover theelectricity needs of the Republic of Cyprus for the next 210 years.The exploitation of the reserves and the transport to Europe andIsrael could be made by an undersea pipeline, an alternative thathas already been discarded due to economical (high cost),

Contents lists available at ScienceDirect

journal homepage: www.elsevier.com/locate/enpol

Energy Policy

http://dx.doi.org/10.1016/j.enpol.2014.01.0440301-4215 & 2014 Elsevier Ltd. All rights reserved.

n Corresponding author. Tel.: +30 210 7722783.E-mail addresses: mpapadop@mail.ntua.gr (M.P. Papadopoulou),

antoniou@central.ntua.gr (C. Antoniou).

Energy Policy 68 (2014) 306–319

technical (deep seabed) and geopolitical (Middle-East region)issues. The construction of a liquefaction plant and gas transporta-tion by ships to markets offer to Cypriots a particularly importantadvantage of non-dependence from countries where the pipelinemay pass; however, this is also a quite expensive project consider-ing various social-economic and environmental factors.

As of today, high dependency on imported energy sources andstrong dominance of oil in Cyprus' energy balance is mainlyconsidered. Also, the economic development of the country isstrangled between the rapid increase in energy demand anddifficulties in the connection with European energy networksdue to its geographic location, as well as the relatively lowpenetration and utilization of renewable energy sources.

In recent years, Cyprus' energy system has profound capacity asit is undergoing a period of significant change, involving mainlythe liberalization of energy markets as required by EuropeanDirectives. The decision to import natural gas in the energybalance of the country requires the development of moderncogeneration systems and imposes structural interventions toaddress challenges in energy. The energy policy of Cyprus is fullyharmonized with European Union (EU) guidelines ensuringhealthy competition in the market, secure supply of energy andenergy needs coverage of the country with the least possibleeconomic and environmental burden.

The nature of this research is to propose a methodologicalframework for macroscopic infrastructure location decision-making based on environmental impact assessment. The outputof this process would lead to the development of specific spatialplans, which would then be further considered at a lower level, ina more detailed impact assessment study. The aim of the analysispresented in this research is to evaluate various alternativelocations for a liquefied natural gas (LNG) terminal station inCyprus, based on REGIME multi-criteria decision analysis, and toexamine the impact of its construction and operation on thenatural and human environment. The selection of LNG terminallocation will be done considering factors arising from currentEuropean and National legislation for natural gas and moreparticularly for LNG terminals (e.g. area, access, terrain, etc.). Also,the final location will be obtained after examination/evaluation ofother specific parameters such as cost, safety, and existing infra-structure. Also an important issue for Cypriots is the transporta-tion of LNG from the terminal to the major power plant in Mariand the country's gateway in order to be exported. Differentappropriate transport modes (rail and pipelines) will be consid-ered for the transport of LNG between the point it reaches theisland and the energy conversion plant/export gateway. Alterna-tive feasible geometric alignments are proposed and evaluated,considering a number of restrictions and guidelines, using a GIS-based approach. Then an environmental impact assessment (EIA)methodology is proposed to evaluate the alternative transportroutes. The EIA analysis was based on a tableau format question-naire answered by a group of environmental and transportationexperts.

2. Background

2.1. Previous experience

The offshore oil and gas sector has been primarily developed inthe North Sea since the mid 1970s. Land-based developments,such as pipelines, terminals, processing plants, production plat-forms installation, and the large number of construction workerswere the two main issues that were characterized as crucial inearly '80 (Lyddon, 1983). An analysis of the UK and Norwegianenvironmental impact assessment (EIA) adopted procedure

related to North Sea oil development showed that issues relatedto land-use planning, public hearing, discharge standards, coastallegislation and bodies were considered important but for variousreasons, such as bureaucracy, limited public participation andlack of information on biological and chemical effects of oilon marine environment, affected planning to a minor degree(Fischer, 1983).

Offshore oil and gas developments require regional planningand strategic coordination (Salter and Ford, 2001) due to high riskon a global scale to marine environment (Wagner and Armstrong,2010). However, the oil and gas sector has limited experience ofstrategic environmental assessment (SEA) applications legislatedunder EU Directive (2001/42/EC). Fidler and Noble (2012) in theiranalysis about SEA in offshore oil and gas sector applied toNorway, Canada and the UK found that, without a coordinationbetween the different levels of administration, SEA will fail toobtain decisions related to planning and development in a lessrestrictive environmental and socioeconomic context. An evalua-tion of EIA performance within the oil and gas sector in the UKshowed that quality improvements related to the scientific qualityof EIA, enhanced scoping, a greater level of integration and a betterdissemination are necessary to significantly reduce the percentageof unsatisfactory quality environmental statements for decisionmaking purposes (Barker and Jones, 2013).

2.2. Research perspective

The energy dependence of a country is a long-standing pro-blem, which has to be managed by its administration. Thegeopolitical position of countries, such as Turkey and Greece,between the producer countries of Middle East, Caspian andCentral Asia and the major consumer countries of Western Europeallows them to develop a leading role as “energy corridor” (Kilic,2006). Failure to obtain adequate volume, continuous supply andreasonable price could lead into catastrophic impacts for thenational security, economic and political stability of a country(Conant and Gold, 2000). In order to maintain energy adequacyand security, countries must comply with certain rules such as(Verma, 2007):

– diversification of supply from more than one sources;– “security margin” in the energy supply system that will provide

a buffer against shocks and facilities recovery after disruptions;– importance of information; and– recognizing the reality of integration.

The design and implementation of an oil and/or natural gaspipeline is not something trivial considering the impact of energythat will be transferred over great distances. In the planning phaseof a pipeline routing, not only political but also environmentalissues related to topography, land uses, ownership should beconsidered. The accelerated development of remote sensing (RS)in conjunction with the analysis that could be obtained usinggeographical information systems (GIS) contribute to the evalua-tion of alternative pipeline routes. Feldman et al. (1995) developeda prototype to obtain the least cost pathway for pipeline place-ment using remotely sensed data and GIS analysis. Their prototypewas tested in a section of the Caspian pipeline. In their analysis,parameters such as pipeline length, topography, geology, land use,stream wetland road and railroad crossing were considered inorder to determine a least cost pathway. Their results showed thata straight-line path was not always the cost-effective solution.

Jo and Ahn (2002) provided an analysis of various character-istics of pipelines including the potential size of the hazard areaand developed an equation based on release rate, gas jet and heatflux from fire to estimate the hazard area. A quantitative risk

M.P. Papadopoulou, C. Antoniou / Energy Policy 68 (2014) 306–319 307

assessment method to improve the level of safety in transmissionnatural gas pipeline was proposed by Jo and Ahn (2005) bydefining the minimum proximity of the pipeline to buildings tobe approximately proportional to the square root of the operatingpressure of the pipeline. Also Jo and Ahn (2005) introduced twokey factors: (a) the fatal length relating hypothetical accidentsalong the pipeline and (b) the cumulative fatal length of thepipeline section that leads to N or more fatalities. The estimationof these parameters was performed using GIS pipeline geometryand population density information.

Kuwari and Kaiser (2011) used satellite images to monitor landuse changes at Al Khore, a region in Qatar where a natural gas fieldwas discovered. Their analysis focused on the investigation ofrapid urbanization rate and its effects on other land uses at AlKhore city and the impact of Ras Laffan harbor on the coastline. Ahybrid Delphi-SWOT paradigm was developed by Tavana et al.(2012) to evaluate alternative routes of oil and gas pipeline inCaspian Sea basin. In the proposed methodology, the experts'opinion was utilized to segregate various economical, political,legal, environmental, technological, cultural, social and geographi-cal factors into internal strengths and weaknesses and externalopportunities and threats in order to assist Decision Makers (DMs)in evaluating the alternative export routes and enhancing decisionmaking process.

The design of energy pipeline routing could also be evaluatedprior to any implementation based on risk assessment methodol-ogies. Zuniga-Gutierrez et al. (2002) presented an environmentalecological and risk assessment of two alternative gas pipelineroutes in Campeche Mexico. The ecological assessment wasobtained by the modified Yapp Method that determined thedamages along the proposed routes. A comparative risk of accidentindex was also developed to estimate risk assessment associatedwith pipeline length and land use, population density and terraintopography. In their analysis, Zuniga-Gutierrez et al. (2002) wereable to assess alternative routes for natural gas pipelines. Indexingsystem method and analytical hierarchy process were also appliedto assess environmental risks of gas transfer pipelines. Using GIS,investigated risks were classified throughout the Aabpar–Zanjanpipeline route in Iran (Jozi et al., 2012).

2.3. Current situation in Cyprus

The recent discovery of large reserves of natural gas in theExclusive Economic Zone of Cyprus has the potential to signifi-cantly enhance the role of Cyprus in the region, as well as shift andaugment the economy of Cyprus. Besides covering its energyneeds, these reserves could also turn Cyprus into an energyexporter. However, risks such as the peril of myopic, short-termplanning, political tension in the area (Lebanon-Israel, Syria, etc.),the relation with Turkey, the large depth of the reserves (under1700–2500 m of water), and earthquake potential of the region arecrucial. The technical challenges associated with the exploitationof these reserves have been manifested by extremely high cost ofdrilling.

The Republic of Cyprus is a small country, covering 70% of theisland of Cyprus (the remainder is occupied by Turkey since a 1974invasion) with a total population of approximately 839,000, upfrom approximately 615,000 in 1992 and 703,000 in 2001(Statistical Service of Cyprus, 2013). The geomorphology of Cyprusis mountainous, making the location of infrastructure a challen-ging task. In particular, there is the main mountainous complex ofTroodos, located in the central and southwest part of the island,with a summit of 1953 m, and the northern mountain series ofPentadaktilos, with a summit of 1024 m. A central valley is locatedbetween these two mountains, traversed by two rivers, while theisland is surrounded by beaches and seaside valleys.

Careful planning is required in order to ensure that LNG can bequickly and efficiently incorporated into the energy mix of Cyprusand proper long-term planning is performed. The RegulatoryEnergy Authority of Cyprus (REAK) is responsible for a largenumber of related activities, including energy sufficiency, priceregulation, protection of consumers' interests, encouraging the useof renewable energy and advising the government on energyissues. According to Cypriot legislation, the implementation oftransporting natural gas in Cyprus has been assigned to the PublicAuthority for Natural Gas (DEFA) and the Electricity Authority ofCyprus (ΑΗΚ). Negotiations for the selection of strategic partnersfor the exploitation of the natural gas resources are currentlyunderway by the aforementioned parties. According to the strate-gic plan of the Ministry of Commerce of Cyprus, the exploitation ofthe Cypriot natural gas reserves is expected to start around 2016;however, the option of kick-starting the use of natural gas usingimports from neighboring countries is also being considered.

The Republic of Cyprus has revised its legal framework in orderto fully harmonize it with Directive 94/22/EC on conditions forgranting and using authorizations for the prospection, exploration,and production of hydrocarbons (Hydrocarbons Law of 2007 andHydrocarbons Regulations of 2007). Previously, the Republic ofCyprus has also instituted legislation on defining the boundaries ofthe nation's contiguous zone (Law (63(I) of 2004) and ExclusiveEconomic Zone (EEZ Law (64(I) of 2004).

Cyprus also participates as a party on MARPOL and theBarcelona Convention. Under the Barcelona Convention, there isan offshore protocol specifying control measures for hydrocarbonexploration and exploitation; however it has not yet entered intoforce, because it has not yet been ratified by the requisite numberof Parties. Finally a SEA (2008) has been conducted in accordancewith Law (No. 102(I)/2005), which is harmonized with Directive2001/42/EC focusing on those activities most likely to resultfrom the concerning hydrocarbon activities within the EEZ of theRepublic of Cyprus including prospecting, exploration, andexploitation.

The infrastructure associated with the extraction and transportof natural gas comprises several facilities, including LNG terminallocations. Transport of natural gas can take place via multipleroutes, including (high or low pressure) pipelines, distributionnetwork and final infrastructure at the receiver's end. Based onfeasibility and technical-financial analyses, it appears that trans-port via pipeline to Turkey is unlikely (due to the relation betweenCyprus and Turkey), while pipelines to Greece or Israel presentsignificant technical challenges, making the project infeasible. Itseems that liquefying the natural gas and transporting it viacontainer ships would be the preferred solution. Therefore, it isnecessary to determine the optimal locations for the LNG terminallocations (with reference to the selected/candidate ports), dis-cussed in Section 4, as well as the pipeline network for the transferof the natural gas from the extraction point to the terminallocations, discussed in Section 5.

3. Methodological approach

The topic of this research is multidisciplinary in nature, requir-ing a number of different steps and expertise (Fig. 1). A sequentialprocess if followed, which comprises a preliminary data analysisphase (Steps 1 and 2), with rigorous engineering methodologies(Steps 3 through 5) and involvement of experts (Step 6). Inparticular, following the definition of the problem (Step 1), therequired data were collected (maps, locations of infrastructure andfacilities) (Step 2). All data were imported into a GIS platform,where they were reconciled (Step 3) and used together to developcommon influence areas for the determination of candidate

M.P. Papadopoulou, C. Antoniou / Energy Policy 68 (2014) 306–319308

terminal locations (Step 4) and feasible alignment for possiblerailroad/pipeline connections, conditional on the selected terminallocations (Step 5). The findings of this analysis were evaluated andassessed by an expert panel comprising researchers and expertpractitioners specialized in transport and environmental impactassessment (Step 6). Finally, conclusions and recommendationsare drawn, and directions for future research are outlined (Step 7).

4. Evaluation of LNG terminal locations based on multi-criteria analysis

4.1. Identification of criteria

In this paper, the choice of LNG terminal location in Cyprus firstand then the suitable routing of LNG pipeline to transfer the LNGfrom the terminal station to the country's gateway were obtainedbased on the satisfaction not only of geographical and technicalcriteria but also sociopolitical ones. Environmental planners arecommonly using GIS modeling to combine and/or overlay, cate-gories of geographic information to create new model categories,which are often used to represent and evaluate the suitability of aparticular activity that may have various environmental impactsacross the landscape (Towers, 1997). The data collection and apreliminary analysis were performed using the GIS techniques(Tsioupis, 2012; Mamas, 2012).

Particular emphasis was paid in issues related to security andprotection of the environment. The criteria followed in thisanalysis for the LNG terminal station siting were in accordancewith the EU and the Cypriot legislation and were the following:

� Criterion 1: Distance of the terminal station from coastline.Natural gas reserves will be transferred through an underseapipeline onto the shore. The distance between the end point ofthe undersea pipeline to the terminal should not exceed 5 km,for construction and operational feasibility considering thespecial conditions of the island of Cyprus. Based on thiscriterion a 5 km area along the coastline creates a potentialinfluence zone.

� Criterion 2: Safety clearance.An LNG terminal station is defined by EU legislation (Directive96/82/EC) as high dangerous site due to risks incurred in theevent of leakage. Based on previous experience in Revithousaplant in Greece, LNG vapors could be dangerous in a radius of4833 m (Balaouras, 2008). In this analysis, the safe distance ofLNG terminal from cohesive residential areas with population

over 1000 people is specified in 5 km creating an exclusivezone around them.

� Criterion 3: Protected areas.This criterion considers the protection of high importance areasof flora and fauna according to the Natura 2000 Treaty. Anexclusive zone of 5 km away from environmental protectedarea under the Natura 2000 Treaty is created, based on thesame guideline values calculated for the Revithousa plant.

� Criterion 4: Protection of natural water bodies.The fourth criterion concerns the protection of surface waterbodies of the island, namely rivers and lakes as they are definedby 2000/60/EC Directive and the corresponding Water ResourcesManagement Plan for Cyprus (Ministry of Agriculture, NaturalResources and Environment of Cyprus (2011)). A 500 m zonefrom any surface water body is considered as an exclusive zonefor any terminal construction activity that may pose risks ofwater pollution.

� Criterion 5: Access-distance from road network.The road connection of LNG terminal facilities with existingroad infrastructure is a priority for a project of this scale. Theterminal final location should be considered within a max-imum distance of 3 km from primary road network for financialfeasibility purposes.

� Criterion 6: Area.A LNG terminal station consists of liquefaction of natural gas,storage tanks of LNG regasification of LNG and ancillary facil-ities (buildings, utility roads, etc.). For the liquefaction ofnatural gas and LNG regasification facilities and ancillarybuildings an estimated area of 20,000 m2 is considered suffi-cient from an engineering point of view, considering theparameters of the problem.

By imposing criteria 1 and 5 a zone of influence has beencreated in which the potential LNG terminal locations should beplaced after evaluation (Fig. 2a). On the other hand, an exclusionzone has been created by imposing the criteria 2, 3, and 4 repre-senting the areas of the island that are not suitable for the LNGterminal construction (Fig. 2b).

By removing from the zone of influence the exclusion zone fiveinitial candidate locations were obtained that were also satisfiedthe area criterion of minimum 20,000 m2 (Fig. 3). The candidatelocations 3 and 4 are located in the northern coast of the island.Even though both locations satisfied the imposed constraints,nevertheless are considered unsuitable for such a significant scaleinvestment due to its proximity to the Green Line that divides theisland into two regions since 1974.

4.2. Application of the multi-criteria decision analysis methodology

Multi-criteria decision analysis (MCDA) offers a series of widelyused methodological tools for the solution of natural resourcesmanagement such as value measurement models, goal, aspirationor reference level models and outranking models (Mendoza andMartins, 2006). As Belton and Stewart (2002) pointed out in theiranalysis, the outcomes of outranking MCDA methodologies are basedon pairwise evaluation of the various alternatives by assessingpreferences and indifferences. Previous applications of outrankingmethods such as ELECTRE (e.g. Brito et al., 2010), REGIME (e.g. Chungand Lee, 2009), PROMETHEE (e.g. Anagnostopoulos et al., 2003) haveshown that these methods could be suitable to deal with environ-mental energy planning decision problems related to LNG terminalselection location. The comparison analysis performed by Polatidiset al. (2006) between various MCDA methods for energy planningbased on modeling decision-makers' preferences, theoretical and

Fig. 1. Overview of the followed methodology.

M.P. Papadopoulou, C. Antoniou / Energy Policy 68 (2014) 306–319 309

technical features, practical requirements and uncertainty treatmentshowed that no single method has superior performance for allidentified attributes. However, for planning problems, such the onedealt with in this paper, REGIME method seems to have an aboveaverage “desirable level” of acceptance, similar to the others.

Based on the above, all three methods are suitable for theproblem at hand. An evaluation process is performed based onREGIME, in order to evaluate the candidate locations 1, 2 and5 along the western and southern coast of the island. Oneadvantage of this analysis lies on its capacity (along with othermethodologies, e.g. PROMETHEE) to simultaneously deal withquantitative and qualitative data and also considers criteria prio-rities during the evaluation process (Nijkamp et al., 1990). Theapplication of the method is based upon (a) the evaluation criteria

and (b) the set of weights that corresponds the relative importanceof each criterion to the final decision.

The basic information necessary to apply this methodology is:

� problem alternatives (i.e. candidate terminal locations),� quantitative or qualitative criteria to perform the evaluation

process, and� hierarchy of criteria (weight criteria factor).

The methodological structure of the evaluation procedurefollows certain steps:

(a) The definition of the problem: Construction of LNG terminalstation in Cyprus.

Fig. 2. GIS-based analysis of location feasibility (adapted from Tsioupis, 2012). (a) Zone of influence (based on criteria 1 and 5) and (b) Exclusion zone (based on criteria 2, 3 and 4).

M.P. Papadopoulou, C. Antoniou / Energy Policy 68 (2014) 306–319310

(b) The definition of alternatives: Candidate alternative locations 1,2 and 5.

(c) The definition and detailed description of the evaluation criteria:K1: Transportation cost of natural gas – Distance fromAphrodite field.K2: Transportation cost – Distance to existing powerstations.K3: Distance from port facilities.K4: Presence of faults and potential faults in thecandidate location.

(d) Creating the list of consequences: Environmental impact assess-ment of the LNG terminal station in to natural and humanenvironment.

(e) Hierarchy of criteria based on their importance from more to lessimportant: This is usually determined based on engineeringjudgment. In this case, the hierarchy that was used isK44K24K14K3 and the assigned weights are 45% for K4,30% for K2, 20% for K1 and 5% for K3.

(f) Application method – Analysis of Performance – Results: Thealternative with the highest value is considered the best one.

The developed evaluation matrix is composed of elements thatmeasure the effect of each alternative location i, i¼1,..,3 withrespect to each criterion j, j¼1,..,4. The REGIME method is based ona pair wise comparison of the 3 candidate locations, where thecomparison each time is not influenced by the presence andeffects of other alternatives, even though the potential rank orderof a certain alternative is conditioned by the remaining alterna-tives (Hinloopen and Nijkamp, 1986). The basic principle of thismethod is the calculation of the probability for each pair ofalternative locations. One alternative could be ranked higher onthe hierarchy than the other considering all imposed criteriataking into account the assigned criteria weights.

Upon completion of this process, the method compares the3 alternative locations in pairs (Table 1). The best alternative isconsidered the one with the highest average probability value.

Based on the analysis, the alternative location E2 dominateswith probability of 0.68 the other two candidate locations E1 andE5 with average probability values of 0.36 and 0.47 respectively.The analysis shows that the optimum location of the LNG terminal

gas tends to be in region E2 near Pentakomo, Mari and Kalavasosvillages between Limassol and Larnaca.

However, it is noted that due to the recent devastating accidentat the power station in Mari, which left 12 people dead and thefacility severely damaged, even though location E2 appears themost likely by the REGIME analysis, social concerns require furthercareful consideration. While this might be the object of a riskassessment procedure, in this research, the focus is on theenvironmental impact assessment. Therefore, in this context, weconsider the other two candidate locations for further evaluation.

5. Transport approach and network design

With the advent of directives and regulations (such as the“Seveso II” EU Directive), information technologies such as GISbecome increasingly useful in the management and control ofmajor accident risk (Contini et al., 2000). The authors describeARIPAR-GIS, an area risk analysis and control support tool thatuses vector thematic maps to describe road, rail, waterways, andpipeline networks, vulnerability centers (e.g. hospitals, churches,schools, supermarkets), the inhabited areas and the location ofpossible accidents.) Papadakis (2000) discussed the difference oflegislation among EU member states and highlighted the need fora European Regulatory Instrument.

Recent research works were published considering the risk ofpotential accident on gas transmission pipeline failure. Han andWeng (2011) compared qualitative and quantitative risk assess-ment methods for urban natural gas pipeline networks andconcluded that both approaches could be applied to practicalapplications. The choice of the method depends on the actual

Fig. 3. Candidate LNG terminal locations based on GIS analysis (adapted from Tsioupis, 2012).

Table 1Evaluation of candidate terminal location based on REGIME analysis.

Candidate terminallocations

E1 E2 E5 Averageprobability

E1 0.30 0.41 0.36E2 0.70 0.66 0.68E5 0.59 0.35 0.47

M.P. Papadopoulou, C. Antoniou / Energy Policy 68 (2014) 306–319 311

basic data of the gas pipelines and the accuracy requirements ofrisk assessment.

Koopman and Ermak (2007) reported lessons learned from LNGsafety research through large scale tests that were performed inthe US using liquefied gaseous fuel spills (such as those that couldresult from a spill from a LNG pipeline). Simonoff et al. (2010)developed risk measures and scenarios in order to better under-stand the consequences of pipeline failures and find that thecharacteristics of an incident in terms of risk management canvary a lot depending on the type of the pipeline (transmission vs.distribution) and the cost consequence being modeled. Antonioniet al. (2007) presented a methodology for the quantitative riskassessment of major accidents triggered by seismic events andrecognized that large diameter pipelines should be consideredamongst the more critical equipment items in the assessment ofrisk due to seismic events (in process plants).

5.1. Location criteria

The emphasis in this research is to look at the infrequent, high-impact, catastrophic events that can be resulted from a majorfailure or accident of a liquid gas pipeline (and not so much at themore frequent and lower impact events that would e.g. only leadto property damage). The view that is taken is to first look at theconstraints and attempt to find a solution that satisfies them all attheir nominal values. As can be expected, this is often not feasible.Therefore, a process of prioritizing and incrementally relaxingthese constraints is followed, until a feasible solution that satisfiesall (valid at that point) constraints is obtained. Considering thescope of this study, which deals with the preliminary siting of theinfrastructure, detailed parameters such as the pressure andconfiguration of the pipelines are not considered.

Transportation planning is inherently multi-objective in nature(Current and Min, 1986). Determining the optimal location for apipeline is subject to a number of constraints, including legal,administrative, technical, environmental and safety criteria. Theneed for compliance with an increasingly large number of regio-nal, national, and international standards, guidelines and policiesputs additional burden and limitations on these activities. A largenumber of approaches have been developed over the past fewdecades (e.g. Barda et al., 1990). These approaches rely on bothquantitative and qualitative criteria, the relative weighting ofwhich is driven by a number of stakeholders (including engineer-ing, production and planning management). Kim and Moon (2008)developed a generic optimization-based model formulated as amixed integer linear programming problem, which considers notonly cost efficiency, but also safety. Laporte and Pascoal (2012) alsoproposed an exact algorithm for the problem of locating a pipelinebetween two points, using a labeling approach that simulta-neously locates the optimal pipeline and safety valve locations.During the last decades GIS have evolved from a research tool topractical instruments with a multitude of applications, oftentransforming the way that location science works (Church,2002). In the context of this analysis and based on the above, anumber of suitable criteria have been set. These criteria are splitinto three domains: land-use, hydrological and geological criteria:

Land-use (LU) criteria:

Distance from settlements (LU1): LU1 refers to small settlementsscattered across Cypriot territory. Natural gas pipelines transferdaily very large amounts of natural gas, so in the unlikely eventof an accident the peril to nearby settlements can be signifi-cant. Sklavounos and Rigas (2006) indicated that the minimumdistance for the safety of the population is 250 m. Therefore inthis analysis, a distance of 1 km has been selected from thecontiguous part of the settlements.

Distance from municipalities (LU2): LU2 refers to the metropo-litan areas, such as Limassol. As discussed above, the minimaldistance is defined as 250 m; therefore, a safety distance of3 km from the contiguous part of the municipality has beenselected.Distance from cultural and archeological monuments (LU3):Using the same distance of 250 m, the safety distance fromsuch sites of cultural significance is defined as 500 m from thecenter point.

Hydrological (H) criteria:

Distance from riversides (H1): The minimum distance fromriversides is defined as 250 m, to ensure that during the projectconstruction there will be no danger of solid or liquid wasteleakage from the construction site to the water.Distance from lakes (H2): As above, the minimum distance isdefined at 250 m.

Geological (G) criteria:

Distance from seismic faults (G1): The minimum distance fromseismic faults is defined as 1 km.Distance from possible seismic faults (G2): The minimum dis-tance from possible seismic faults is defined as 1 km; this isdefined independently from the distance from the seismicfaults to allow for additional flexibility in relaxing the respec-tive weights, see below.

Environmental (E) criteria:

Distance from protected fauna and flora areas (E1): The mini-mum distance from protected fauna and flora areas is definedas 1 km.

The union of the exclusion zones based on the original criteriaresults in a very limited usable area that does not allow for theconnection of the locations of interest via pipelines or railroad.Therefore, relaxations of these criteria were considered, whichcould result in a balance of protection of the sensitive areas, whileat the same time providing reasonable opportunities for thedevelopment of the infrastructure required for the transport ofthe natural gas. It quickly became evident that achieving acombination of reasonable relaxations of the criteria that wouldallow for acceptable influence zones would not be possible, due tothe rich characteristics of the Cypriot landscape. Therefore, it wasnecessary to violate some of these criteria at special instances,where alternative routes were not afforded. In order to determinethe violations that would minimize the impacts, the criteria wereprioritized using a scale from 1 to 10, where 10 reflects the mostrigid (and less likely to be violated) criterion. The initial andrelaxed values per criterion are shown in Table 2, along with apriority index, which can be used for further relaxation, if there isstill no clear route allowing connection of the various locations.

5.2. Pipeline route design

The details of the alignment and geometric design of thepipeline and railroad are outside the scope of this paper, whichfocuses on a more macroscopic view of the problem. Fig. 4 showsthe pipeline and railroad alignment overview that resulted fromthe analysis for Areas E1 and E5.

M.P. Papadopoulou, C. Antoniou / Energy Policy 68 (2014) 306–319312

5.3. Environmental impact assessment through expert panelvalidation

Following the technical feasibility assessment of the alternativemodes and routes to connect the candidate LNG terminal locationswith the power plant in Mari, an environmental impact assess-ment was undertaken. The impact assessment was based on atableau format questionnaire (appendix) that outlined the poten-tial impacts (columns) on specific parameters of natural andhuman environment (lines). The adopted questionnaire formatwas based on previous environmental impact assessment studieson similar projects (PENSPEN Ltd. – C&M Engineering S.A., 2010).The questionnaire was completed by a number of transportationand environmental experts, who were asked to fill in the intensityand the significance of each specific impact on the correspondingparameter of the environment. In order to get a more balancedand complete picture, we divided the experts from each group(transportation or environment experts) to experts from the

practice and experts from the academia. As a result, four groupsof experts completed the questionnaire (all combinations oftransportation/environment by practice/academia), resulting infour groups of 3 experts each, for a total of 12 completedquestionnaires. This type of analysis is a recognized method ofprimary research and analysis (Patton, 1994).

The methodological approach that was adopted in order to obtainan environmental impact assessment for the proposed four alter-natives routes (2 for pipeline and 2 for railroad, as shown in theprevious section) is similar to that used to capture the opinion of anexpert group about transportation issues (e.g. Papathanasopoulouand Antoniou, 2011; Sarlas et al., 2013). The small size of the experts'group is compensated for by their expertise and their structuredanswers related to both the intensity and the significance of theimpacts (Rye, 1997). The careful selection of the respondents ensuredan in-depth environmental impact assessment of the proposedalternative routes (Gar, 2007). In this analysis, the experts wereasked to assess the impacts of the construction and operation of eachproposed route to the environment based on their perception on theenvironmental characteristics of the areas where the proposed routeswould pass through. It is noted that the methodology performed inthis research is not a full-blown Delphi study. Instead, a single-stepprocess of structured interviews was conducted, which did notinclude a feedback/iteration/consistency analysis.

5.4. Questionnaire

The questionnaire was given to 6 transportation (3 faculty –

Group A and 3 professionals – Group B) and 6 environmentalexperts (3 faculty – Group C and 3 professionals – Group D) inorder to evaluate the intensity and the significance of specificimpacts into specific parameters of the environment during theconstruction and operational phases of LNG transmission network.Two alternatives network routes from each candidate location

Table 2Relaxation and prioritization of the initial criteria.

# Description Initialcriterion(km)

Relaxedcriterion(km)

Priority (higher valuemeans more rigid)

LU1 Settlements 1 0.8 6LU2 Municipalities 3 2 6LU3 Cultural and

archeologicalmonuments

0.5 0.5 5

H1 Riversides 0.25 0.25 7H2 Lakes 0.25 0.25 7G1 Seismic faults 1 1 9G2 Possible seismic faults 1 0.5 7E1 Protected areas 1 0.7 10

Vasiliko Power PlantLNG Transport PipelineCoastlineIsoheightsProposed LNG Plant LocationsInfluence Zone

Legend LegendVasiliko Power PlantLNG Transport PipelineCoastlineIsoheightsProposed LNG Plant LocationsInfluence Zone

Legend

Vasiliko Power PlantCandidate Railroad AlignmentCoastlineIsoheightsAreas with Steep SlopesProposed LNG Plant LocationsInfluence Zone

Vasiliko Power PlantCandidate Railroad AlignmentCoastlineIsoheightsAreas with Steep SlopesProposed LNG Plant LocationsInfluence Zone

Legend

Fig. 4. Optimal pipeline and railroad route alignment from Areas E1 and E5 to power plant in Mari (adapted from Mamas, 2012).

M.P. Papadopoulou, C. Antoniou / Energy Policy 68 (2014) 306–319 313

(Area 1 and Area 5) were evaluated based on the questionnaire,one considering transmission through a high-pressure pipelineand the other considering transmission through a railroad systemnetwork.

Each expert completed the questionnaire considering the pro-vided information related to soil, water and climatic conditions,geomorphologic characteristics, urban and cultural environment,infrastructure and environmental protected areas. The question-naire covered aspects such as the intensity and significance ofspecific impacts like pollution, noise, odors/particle matter (PM),deterioration, construction and congestion of ancillary works,

employment during the construction phase and also impacts duringthe operational phase like noise, leakage, land use change and landoccupation. For each alternative route (two for pipeline and two forrailroad), the expert was asked to complete a value from �10 to 10(�10 being the most negative) to express the intensity that thespecific impact has to the corresponding environmental parameter.Then, a single value from 1 to 10 (10 being most important) ischosen for each pair to express the significance of the specificimpact to the corresponding environmental parameter according tothe expert's opinion. It is stressed that not all cells of this tableauneed to be completed, since all construction and operation

Odo

rs/P

M(d

urin

g co

nstru

ctio

n)Po

llutio

n(d

urin

g co

nstru

ctio

n)

Noi

se (d

urin

g co

nstru

ctio

n)

Noi

se (d

urin

g op

erat

ion)

Det

erio

ratio

n (d

urin

g co

nstru

ctio

n)

Traf

fic fl

ow o

f ac

com

pany

ing

proj

ects

(d

urin

g co

nstru

ctio

n)C

onst

ruct

ion

of

acco

mpa

nyin

g pr

ojec

ts

(bef

ore

and

durin

g co

nstru

ctio

n)

Em

ploy

men

t/GD

P

(dur

ing

cons

truct

ion)

Land

occ

upat

ion

and

land

use

cha

nge

(dur

ing

oper

atio

n)

Leak

age

(dur

ing

oper

atio

n)

Public healthEconomic environmentExisting infrastructureSocial environmentUrban environmentCultural heritageBiodiversity − faunaBiodiversity − floraClimateAtmosphereSea waterSubsurface waterSurface waterLandscape colorLandscape shapeTerrainFaults/EarthquakesGeomorphologySoil

Number of responses (significance) Pipeline / All respondents

2 4 6 8 10

Value

Color Keyand Histogram

Odo

rs/P

M(d

urin

g co

nstru

ctio

n)Po

llutio

n(d

urin

g co

nstru

ctio

n)

Noi

se (d

urin

g co

nstru

ctio

n)

Noi

se (d

urin

g op

erat

ion)

Det

erio

ratio

n (d

urin

g co

nstru

ctio

n)

Traf

fic fl

ow o

f ac

com

pany

ing

proj

ects

(d

urin

g co

nstru

ctio

n)C

onst

ruct

ion

of

acco

mpa

nyin

g pr

ojec

ts

(bef

ore

and

durin

g co

nstru

ctio

n)

Em

ploy

men

t/GD

P

(dur

ing

cons

truct

ion)

Land

occ

upat

ion

and

land

use

cha

nge

(dur

ing

oper

atio

n)

Leak

age

(dur

ing

oper

atio

n)

Public healthEconomic environmentExisting infrastructureSocial environmentUrban environmentCultural heritageBiodiversity − faunaBiodiversity − floraClimateAtmosphereSea waterSubsurface waterSurface waterLandscape colorLandscape shapeTerrainFaults/EarthquakesGeomorphologySoil

Number of responses (significance) Railroad / All respondents

2 4 6 8 10Value

Color Keyand Histogram

0

30

Cou

nt

0

20

50

Cou

nt

Fig. 5. Number of responses (significance) for all respondents. (a) Pipeline and (b) Railroad.

M.P. Papadopoulou, C. Antoniou / Energy Policy 68 (2014) 306–319314

characteristics do not affect each environment aspect. Furthermore,it is expected that the results will have some sort of variability, aseach expert does not consider the same type and magnitude ofeffects of each alternative route to the environment. It is possiblethat groups of experts (as defined above) might have lower within-group variability, than that observed for the entire pool of experts.

5.5. Results analysis

The analysis of the questionnaires from the expert panel groupshowed that the number of responses on environmental impactsrelated to the construction phase such as odors/PM, noise, dete-rioration of the soil and employment opportunities is quite highand got a score whereas for the operation phase only leakage isunanimously considered important and got a score. A specialconcern was indicated by a significant number of respondents inimpacts related to noise during the operation phase of the railroadalternatives and to leakage into the soil for the pipeline alter-natives (Fig. 5).

The results obtained from the analysis of the 12 experts'outcomes were focused on the impacts in the various parametersof the environment and also on the causes. A correlation betweencauses and impacts was obtained by analyzing the results basednot only on the score but also on the significance that wasattributed from each expert. From the analysis it is obvious thatthe major concern for all experts as far the causes is the leakageduring the operation phase (Fig. 6). Although environmental

experts are more sensitive concerning the damage to the environ-ment due to leakage, transportation experts expressed their highconfidence in railroad routes, which ensure low risk of leakageinto the environment. The variation between the two examinedlocations (Area 1 and Area 5) is proportional to the route length(the longer the route, the higher the risk). However, the twoproposed railroad routes cause significant adverse effects inrelation to high-pressure pipelines in terms of additional ancillaryprojects, pollution during construction period, odors and noiseduring the operation phase (Fig. 7).

A small number of outliers is evident in the results presented inthe analysis. However, as these outliers seem to be consistentacross modes and areas, their impact to the overall results isexpected to be minor. Furthermore, the underlying trends anddifferences attributed to the actual factors in the analysis (e.g.mode and area, but, perhaps more importantly, type of expert) areclearly identified. For example, special concern was paid bytransportation experts to impacts related to public health. Onthe other hand, impacts to the geomorphology and soil quality areconsidered significant for the environmental experts (Fig. 8).Impact differentiations for the two different (railroad vs. pipeline)LNG transmission modes are pointed out by the experts onenvironmental parameters such as terrain, surface water andlandscape (Fig. 9). Regarding the design of the proposed routes,the main effects found in environmental variables such as soil,surface and subsurface water, urban environment, biodiversity,geomorphology and public health. The magnitude of the effectvaries slightly with respect to the route length.

Group A Group B Group C Group D

Construction of accompanying projects

(before and during construction)

Deterioration (during construction)

Employment/GDP (during construction)

Land occupation and land use change

(during operation)

Leakage(during operation)

Noise (during construction)

Noise (during operation)

Odors/PM(during

construction)

Pollution(during

construction)

Traffic flow of accompanying projects

(during construction)

Construction of accompanying projects

(before and during construction)

Deterioration (during construction)

Employment/GDP (during construction)

Land occupation and land use change

(during operation)

Leakage(during operation)

Noise (during construction)

Noise (during operation)

Odors/PM(during

construction)

Pollution(during

construction)

Traffic flow of accompanying projects

(during construction)

Railroad

Pipeline

−50 0 50 −50 0 50 −50 0 50 −50 0 50

Score x Significance

Area 1

Area 5

Cause distribution by expert group, mode and position

Fig. 6. Cause distribution by expert group, mode and position.

M.P. Papadopoulou, C. Antoniou / Energy Policy 68 (2014) 306–319 315

It is noted, that a different mix of experts might affect theoutcome and the details of the results. This is why it is importantto put particular emphasis and time in developing a well-structured,balanced design of the impact assessment through the expert panel.As mentioned above, in this study, the resulting panel is comprisedby experts classified across two dimensions (transportation vs.environmental expert, and academic vs. practitioner).

6. Discussion and conclusion

Fossil fuels are still the dominant source of energy for human-kind. Incremental steps towards alternative, renewable or cleanersources of energy signify a slow progress, which may lead to adifferent status-quo after several decades. In the meantime, smallchanges in the relative position of each country in the energy mapresult from new discoveries of natural resources of oil or naturalgas. Such a big discovery was recently made in Cyprus, a countrywith limited energy resources of its own. Cyprus now needs toconsider how to best leverage these resources and turn them intoan opportunity for development and growth, especially amidst the

current precarious financial situation in the region. Furthermore,the tension between the states in the Middle East and the Arabworld is urgently required the discovery of new areas of energyreserves which could be used to safely transport energy todeveloped and developing countries. The discovery of offshoregas deposits in Aphrodite field in Exclusive Economic Zone ofCyprus not only creates new opportunities for the region but alsoreleases the energy dependency of European countries from areaswith a high geopolitical instability (i.e. Iran).

In this research, alternative modes of transporting the LNG froma terminal station on the shores of Cyprus to the energy producingstation (and gateway to foreign markets) are explored. Railroad andhigh-pressure pipelines are considered. The impacts of variousaspects of construction and operation of the infrastructure on theenvironment are explicitly considered. The presented researchresults could be valuable as a decision support tool for stakeholdersand policy-makers. Besides the aspects considered in this research,there are also other considerations, such as cost and politicalobjectives. However, e.g. cost considerations would require a farmore detailed level of study (i.e. feasibility study, preliminary study,and finally detailed study), while political decisions generally fall

Fig. 7. Cause distribution by mode and position for all experts.

M.P. Papadopoulou, C. Antoniou / Energy Policy 68 (2014) 306–319316

outside the control and scope of the engineers. In terms of cost, onecould use unit prices (e.g. €/km of railroad or pipeline), but such alevel of approximation is often not acceptable even as a preliminaryestimate, as it might lead to skewed results.

Ball and Boehmer-Christiansen (2007) discussed the societalconcerns and trade-offs relating risk and recognize that societalconcerns are often viewed very differently by many stakeholders.Therefore, in this research, we have developed and disseminated astructured questionnaire survey to solicit the information ofdomain experts. In particular, a number of questionnaires havebeen completed by experts from four groups: transportationexperts, transportation researchers/academics, environmentalexperts, and environmental researchers/faculty (all with an engi-neering background).

It is noted that natural gas transportation and processingfacilities are sensitive infrastructure that can be the target ofmalicious attacks. One consideration in their development canbe the optimal location of response units in the event of anemergency (Sathe and Miller-Hooks, 2005), something that hasnot been explicitly considered in this research. This is part offuture research, which should also consider related research onthis topic. For example, Sosa and Alvarez-Ramirez (2009) arguethat there are temporal correlations in the occurrence of hazar-dous material pipelines incidents and this might be a usefulconsideration when creating contingency plans for large-scalepipeline projects, such as the one considered in Cyprus.

Sensor and telecommunication technologies may offer signifi-cant support to the improvement of the safety of gas pipelines,

whether natural or man-made. For example, Batzias et al. (2011)present a sensor-based framework for pipeline leakage detectionthat uses rapid, sensitive and selective biosensors in synergy withconventional leakage detection systems.

Also another crucial parameter in the planning of a energytransportation network is the potential impacts to the natural andhuman environment. In their paper Leal and D’Agosto (2011)established priorities based on financial and socio-environmentalconsiderations in order to choose between alternative ways oftransporting bio-ethanol. Their analysis concluded that a multi-modal transport involving pipelines for much of the movement ofbio-ethanol is a better environmental alternative than trucks.

Using multiple methodological approaches, in this research wewere able to address the problem of determining the optimallocation for an LNG terminal in Cyprus. These componentsconsider environmental and transportation impacts both techno-cratic in nature, as well as more subjective issues and based onexpert opinion. Geopolitical and other softer concerns are alsoexplicitly considered. Based on this analysis, the original candidatelocations are progressively pruned from 5 to 3 and then 2. Asthe number of alternatives decreases, the assessment methodol-ogy becomes progressively more detailed. It is noted, that, inthis research, we have not considered feedback from the down-stream steps to the previous steps of the methodology. Futuremethodological extensions of this work could include such feed-back loops.

Based on the findings of this research, it has been demon-strated that the combination of these approaches can be used to

Group A Group B Group C Group D

AtmosphereBiodiversity − faunaBiodiversity − flora

ClimateCultural heritage

Economic environmentExisting infrastructure

Faults/EarthquakesGeomorphologyLandscape color

Landscape shapePublic health

Sea waterSocial environment

SoilSubsurface water

Surface waterTerrain

Urban environment

AtmosphereBiodiversity − faunaBiodiversity − flora

ClimateCultural heritage

Economic environmentExisting infrastructure

Faults/EarthquakesGeomorphologyLandscape color

Landscape shapePublic health

Sea waterSocial environment

SoilSubsurface water

Surface waterTerrain

Urban environment

Railroad

Pipeline

−50 0 50 −50 0 50 −50 0 50 −50 0 50Score x Significance

Area 1

Area 5

Impact distribution by expert group, mode and position

Fig. 8. Impact distribution by expert group, mode and position.

M.P. Papadopoulou, C. Antoniou / Energy Policy 68 (2014) 306–319 317

select and filter candidate infrastructure facility locations androutes by prioritizing them, based on a number of criteria. Theprioritization can then be used as input to the political process ofdecision-making. The characteristics of the proposed methodolo-gical framework make it suitable for a wide range of applications,where one needs to consider multiple heterogeneous criteria froma strategic planning perspective. Such examples include large-scale transport infrastructure, energy policy-making and naturalresources management.

Acknowledgments

The authors would like to thank N. Mammas andG. Tsioupis for assistance with data collection and analysis, andDr. D. Papakonstantinou for his comments during the developmentof the questionnaire. Also, the authors would like to express theirgratitude to the 12 experts for participating in the questionnairesurvey and the two anonymous reviewers whose valuable com-ments considerably improved the manuscript.

References

Anagnostopoulos, K., Giannopoulou, M., Roukounis, Y., 2003. Multicriteria evalua-tion of transportation infrastructure projects: an application of PROMETHEEand GAIA methods. Adv. Transp. 14, 599–608.

Antonioni, G., Spadoni, G., Cozzani, V., 2007. A methodology for the quantitativerisk assessment of major accidents triggered by seismic events. J. Hazard. Mater.147, 48–59.

Balaouras, I., 2008. Development and application of impact model for LNG leakagefrom LNG-carrying vessels (Diploma thesis). School of Naval and MechanicalEngineering, National Technical University of Athens, Greece (in Greek)

Ball, D.J., Boehmer-Christiansen, S., 2007. Societal concerns and risk decisions. J.Hazard. Mater. 144, 556–563.

Barda, O.H., Dupuis, J., Lencioni, P., 1990. Multicriteria location of thermal powerplants. Eur. J. Oper. Res. 45, 332–346.

Barker, A., Jones, C., 2013. A critique of the performance of EIA within the offshoreoil and gas sector. Environ. Impact Assess. 43, 31–39.

Batzias, F.A., Siontorou, C.G., Spanidis, P.-M.P., 2011. Designing a reliable leak bio-detection system for natural gas pipelines. J. Hazard. Mater. 186, 35–58.

Belton, S., Stewart, T.S., 2002. Multiple Criteria Decision Analysis. An IntegratedApproach. Kluwer Academic Publishers, Massachusetts

Brito, A.J., de Almeida, A.T., Mota, C.M.M., 2010. A multicriteria model for risksorting of natural gas pipelines based on ELECTRE TRI integrating utility theory.Eur. J. Oper. Res. 200, 812–821.

Fig. 9. Impact distribution by mode and position.

M.P. Papadopoulou, C. Antoniou / Energy Policy 68 (2014) 306–319318

Chung, E.S., Lee, K.S., 2009. Prioritization of water management for sustainabilityusing hydrologic simulation model and multicriteria decision making techni-ques. J. Environ. Manag. 90, 1502–1511.

Church, R.L., 2002. Geographical information systems and location science. Comput.Oper. Res. 29, 541–562.

Conant, A.M., Gold, F.R., 2000. The Geopolitics of Energy. Westview Press, Boulder,Colorado

Contini, S., Bellezza, F., Christou, M.D., Kirchsteiger, C., 2000. The use of geographicinformation systems in major accident risk assessment and management. J.Hazard. Mater. 78, 223–245.

Current, J., Min, H.K., 1986. Multiobjective design of transportation networks:taxonomy and annotation. Eur. J. Oper. Res. 26, 187–201.

Feldman, S.C., Pelletier, R.E., Walser, E., Smoot, J.C., Ahl, D., 1995. A prototype forpipeline routing using remotely sensed data and geographic informationsystem analysis. Remote Sens. Environ. 53 (2), 123–131.

Fidler, C., Noble, B., 2012. Advancing strategic environmental assessment in theoffshore oil and gas sector: Lessons from Norway. Can. U. K., Environ. ImpactAssess. 34, 12–21.

Fischer, D.W., 1983. A comparison of approaches to assessing the impacts of NorthSea oil in the United Kingdom and Norway. Environ. Impact Assess. Rev. 4 (3–4), 433–456.

Gar, Y.N.g., 2007. Quality of Judicial Organisation and Checks and Balances(Doctoral thesis). Utrecht University, The Netherlands

Han, Z.Y., Weng, W.G., 2011. Comparison study on qualitative and quantitative riskassessment methods for urban natural gas pipeline network. J. Hazard. Mater.189, 509–518.

Hinloopen, E., Nijkamp, P., 1986. Qualitative multiple criteria choice analysis: thedominant REGIME method (Research memorandum). vol. 45. Free University ofAmsterdam, Amsterdam

Jo, Y.-D., Ahn, B.J., 2002. Analysis of hazard areas associated with high-pressurenatural-gas pipelines. J. Loss Prev. Process Ind. 15 (3), 179–188.

Jo, Y.-D., Ahn, B.J., 2005. A method of quantitative risk assessment for transmissionpipeline carrying natural gas. J. Hazard. Mater. A123 (1), 1–12.

Jozi, S.A., Rezaian, S., Shahi, E., 2012. Environmetal risk assessment of gas pipelinesby using of indexing system method. Procedia APCBEE 3, 231–234.

Kilic, A.M., 2006. Turkey's natural gas necessity, consumption and future perspec-tives. Energy Policy 34, 1928–1934.

Kim, J., Moon, I., 2008. Strategic design of hydrogen infrastructure considering cost andsafety using multiobjective optimization. Int. J. Hydrog. Energy 33, 5887–5896.

Koopman, R.P., Ermak, D.L., 2007. Lessons learned from LNG safety research. J.Hazard. Mater. 140, 412–428.

Kuwari, N.Y., Kaiser, M.F., 2011. Impact of North gas field development on landuse/landcover changes at Al Khore North Qatar, using remote sensing and GIS. Appl.Geogr. 31, 1144–1153.

Laporte, G., Pascoal, M.M.B., 2012. The pipeline and valve location problem. Eur. J.Ind. Eng. 6 (3), 301–321.

Leal , I.C., D'Agosto, M., 2011. Modal choice evaluation of transport alternatives forexporting bio-ethanol from Brazil. Transp. Res. Part D 16, 201–207.

Lyddon, W.D.C., 1983. Land use and environmental planning for the development ofNorth Sea gas and oil: the Scottish experience. Environ. Impact Assess. Rev. 4(3–4), 473–492.

Mamas, N., 2012. Siting transmission grid and sustainable development of thesurrounding area (Diploma thesis). National Technical University of Athens,Athens, Greece (in Greek).

Mendoza, G.A., Martins, H., 2006. Multi-criteria decision analysis in naturalresource management: a critical review of methods and new modelingparadigms. For. Ecol. Manag. 230, 1–22.

Ministry of Agriculture, Natural Resources and Environment of Cyprus, 2011.Implementation of Articles 11, 13 and 15 of the 2000/60/EC Directive in Cyprus(in Greek).

Nijkamp, P., Rietveld, P. & Voogd, H. 1990. Decision Support Model for RegionalSustainable Development. Avebury, Aldershot, UK.

Papadakis, G.A., 2000. Assessment of requirements on safety management systemsin EU regulations for the control of major hazard pipelines. J. Hazard. Mater. 78,63–89.

Papathanasopoulou, V. and C. Antoniou 2011. Assessment of congestion pricingprospects for Athens, Greece. In: Proceedings of the 90th Annual Meeting of theTransportation Research Board. January 2011, Washington, D.C.

Patton, J., 1994. Doing Qualitative Research. Sage, Berkeley, CAPENSPEN Ltd. – C&M Engineering S.A. 2010. Environmental impact assessment

study of high pressure gas pipeline from Agious Theodorous to DEI PowerPlantsin Megalopolis, Technical report 80102-RPT-EV-003, pp. 311 (in Greek).

Polatidis, H., Haralambopoulos, D.A., Munda, G., Vreeker, R., 2006. Selecting anappropriate multi-criteria decision analysis technique for renewable energyplanning. Energy Sources, Part B 1, 181–193.

Rye, T., 1997. The implementation of workplace transport demand management inlarge organisations (Unpublished Ph.D. thesis). Nottingham Trent University,Nottingham

Salter, E., Ford, J., 2001. Holistic environmental assessment and offshore oil fieldexploration and production. Mar. Pollut. Bull. 42, 45–58.

Sarlas, G., Papathanasopoulou, V., Antoniou, C., 2013. A simulation-based analysis ofroad pricing prospects for Athens, Greece. J. Urban Plan. Dev. 139 (3), 206–215.

Sathe, A., Miller-Hooks, E., 2005. Optimizing location and relocation of responseunits in guarding critical facilities. Transp. Res. Rec.: J. Transp. Res. Board 1923,127–136.

SEA, 2008. Strategic environmental assessment concerning hydrocarbon activitieswithin the exclusive economic zone of the Republic of Cyprus, EnvironmentalReport (Chapters 1–3), Contract number (MCIT/ES/13/2007), pp. 399.

Simonoff, J.S., Restrepo, C.E., Zimmerman, R., 2010. Risk management of costconsequences in natural gas transmission and distribution infrastructure. J.Loss Prev. Process Ind. 23 (2), 269–279.

Sklavounos, S., Rigas, F., 2006. Estimation of safety distances in the vicinity of fuelgas pipelines. J. Loss Prevent. Process Ind. 19 (1), 24–31.

Sosa, E., Alvarez-Ramirez, J., 2009. Time-correlations in the dynamics of hazardousmaterial pipelines incidents. J. Hazard. Mater. 165, 1204–1209.

Statistical Service of Cyprus, 2013. ⟨http://www.mof.gov.cy/mof/cystat/statistics.nsf/index_en/index_en⟩ (accessed 8.10.13.).

Tavana, M., Pirdashti, M., Kennedy, D.T., Belaud, J.-P., Behzadian, M., 2012. A hybridDelphi-SWOT paradigm for oil and gas pipeline strategic planning in CaspianSea basin. Energy Policy 40, 345–360.

Towers, G., 1997. GIS versus the community siting power in southern West Virginia.Appl. Geogr. 17 (2), 111–125.

Tsioupis, G., 2012. Selecting a location siting LNG terminal to Cyprus and anenvironmental impact assessment (Diploma thesis). National Technical Uni-versity of Athens, Athens, Greece (in Greek).

Verma, S.K., 2007. Energy geopolitics and Iran–Pakistan–India gas pipeline. EnergyPolicy 35, 3280–3301.

Wagner, J., Armstrong, K., 2010. Managing environmental and social risks ininternational oil and gas projects: Perspectives on compliance. J. World EnergyLaw Bus. 3, 140–165.

Zuniga-Gutierrez, G., Arroyo-Cabrales, J., Lechuga, C., Ortega-Rubio, A., 2002.Environmental quantitative assessment of two alternative routes for gas pipe-line in Campeche, Mexico. Landsc. Urban Plan. 59, 181–186.

M.P. Papadopoulou, C. Antoniou / Energy Policy 68 (2014) 306–319 319