Air-Rail Link at Airport in : How competitive is Monorail to /LRT Systems? Prof. Panagiotis Papaioannou1, Dr Alexandros Deloukas2, Prof. Ioannis Politis3, Prof. Christos Pyrgidis4

1 Professor, Aristotle University of Thessaloniki, School of Civil Engineering, Dept of Transport and Construction Management, Corresponding author Tel: +302310 995775, e-mail: [email protected], 2 ATTIKO METRO SA, e-mail: [email protected] 3 Aristotle University of Thessaloniki, School of Civil Engineering, Dept of Transport and Construction Management, e-mail: [email protected]

4 Aristotle University of Thessaloniki, School of Civil Engineering, Dept of Transport and Construction Management, [email protected]

Abstract

The International Airport of Thessaloniki is the third busiest airport in Greece, serving over 6.5 million passengers annually. The airport is located 14 kms from the city center of Thessaloniki and normally it takes 45 mins to travel to/from the airport by . A significant upgrade in terms of capacity is taking place these years according to the airport’s master plan. As of 2017 the airport has been undertaken by Fraport under a 40 years Concession contract. Travelers’ figures have increased significantly in the last 5 years and it is expected that they will continue to grow and may reach double figures by 2030. This increase calls for an improved and more reliable city-airport ground access in the next years. Currently, the airport is served by two routes which account only for a small percentage of all person trips. Most travelers, including people employed at the airport, use private cars and taxis. Creating a fixed track connection between the airport and the city of Thessaloniki has been a hot issue in the last years, given that the Metro line which is under construction terminates 5 kms from the airport site. Many discussions among local politicians, involved authorities and the public, are in progress about the best alternative to link the airport with the city. The do-minimum option will be to schedule an express bus from the Metro terminal to the airport and vice versa so that people can reach their destination with one transfer. However, most stakeholders involved agree that ground access by road suffers from congestion impacts and this problem will continue to deteriorate in the future. For this reason, other solutions relying on a fixed track system are on the table. Three alternative rail technology options were examined in a comparative way in order to identify the least expensive one in social, economic and environmental terms. These options comprise: (a) a further extension of the Metro line which is under construction and it is expected to start operations in 2021; (b) a segregated Tramway/LRT which will start from the Metro terminal station and will follow more or less the alignment of the existing highway to the airport; (c) an elevated Monorail – single track - connecting the Metro terminal station with the airport. All alternatives are not only time reliable but also electrified, so their use is air-pollution free and carbon neutral. The rubber-tired technologies investigated are almost free of noise nuisance. The new fixed track corridor under consideration will consist of 5 stations and will have a total length of 5.1 – 6.3 kms depending on the system. The modal operating capacity selected, covers 10 -15 min policy , a 20- 25% rail transit share of the total trips and a directed loading of 1,300 passengers per hour per direction (pphpd). For determining future demand and modal split, a comprehensive multi modal transport model, developed by Thessaloniki public Transport Authority, was used as a supplementary tool. Model results were consequently used to perform extended cost/benefit comparative analysis. The investigation of cost (e.g. operating cost, user cost, rolling stock and infrastructure) and benefit (e.g. time variability risk, novelty image, employment creation) elements, indicate that Monorail is likely to be the least costly and most beneficial rail alternative in a total cost (benefit) perspective. The Monorail system also seems to create the least effects to land owners and businessmen in the area. It also seems to be associated with the least risks among the examined rail alternatives. However, the final decision depends on both availability of public and private funds that can be secured and on future expansion potential of each alternative.

Keywords: , Rail technologies, Fixed route systems, Monorail, Metro, LRT, Cost Benefit analysis

1

1. Introduction

The international airport of Thessaloniki “” is the third busiest airport in Greece, serving over 6.0 million passengers annually in the last two years. The airport serves local and international flights mainly to European and Middle East countries and operates on a 24-hour basis. Some 50 airlines travelling to/from 40 countries were served during 2018 from which approximately 21 were charter flight airlines. Thessaloniki airport is considered a regional hub and serves the population of Thessaloniki greater area, which is just over 1 million, as well as the area of the Region of Central Macedonia in Northern Greece including the major tourist area of . The airport has two runways and currently there is an ambitious plan under implementation which consists of the extension of one runway to accommodate long haul flights, the construction of a new terminal of 31,000 m2 which will double the airport capacity and of other installations. The passenger demand trend is increasing over the years, and it is expected that the demand may reach 10 million passengers annually within the next ten-year period. Figure 1 presents the total passenger demand for the period 1999-2018

Source: https://www.skg-airport.gr/uploads/sys Figure 1: Evolution of airport passengers at Thessaloniki International Airport “Macedonia” (1999-2018)

It should be noted that as of 2017 the airport has been undertaken by Fraport under a 40 year Concession contract. Many of the airport upgrading projects are part of the investment proposition by the Concessionaire.

The airport is located 14 kms from the city center of Thessaloniki. Available options to reach/leave the airport include private car, taxi and Public Transport bus plus coaches for organized trips. It normally takes 55 to 75 mins to travel from/to Thessaloniki city center to/from the airport by bus, including average waiting time at stops. The respective time by private car or taxi is 50%-60% of the bus time. Extra time of 5-10 mins should be added to the above figure for private car parking and walking time to the terminal. Travel time are considered non satisfactory and it is expected that will be increased significantly in the next years along with passenger demand. Medium to serious congestion problems also appear especially during summer months, given that there is only one ground access link for all possible modes. Figure 2 shows the passenger variation by month for 2018. August is the heaviest month with a share of approximately 12%.

2

Source: https://www.skg-airport.gr/uploads/sys Figure 2: Thessaloniki Airport passenger variation by month (2018)

Creating a fixed track connection between the airport and the city of Thessaloniki has been a hot issue in the last decade, given that the Metro line which is under construction terminates 5 kms from the airport site at station MIKRA. Many discussions among local politicians, involved authorities and the public, are in progress about the best alternative to link the airport with the city. The do-minimum option will be to schedule an express bus from the Metro terminal station MIKRA to the airport and vice versa so that people can reach their destination with one transfer. However, most stakeholders involved agree that ground access by road suffers from congestion impacts and this problem will continue to deteriorate in the future. For this reason, other solutions relying on a fixed track system are on the table.

Figure 3 shows the map of the Thessaloniki Metro system at its final phase, including the potential Metro extension to the airport. This plan foresees that there will be four stations along the Metro line to the airport.

Source: Attiko Metro website (www.ametro.gr) Figure 3: Thessaloniki Metro system map and location of Thessaloniki airport

3

The other fixed track solutions examined at a preliminary level so far are a /LRT surface connection between MIKRA Metro station and the airport and an elevated Monorail connection between the same end points. The three systems, (Metro, Tram/LRT and Monorail), are examined in a comparative way in this paper as future alternative options to connect the terminal station of MIKRA with the airport. The comparison is made against the existing bus connection which will continue to exist no matter if a new system will be implemented. All alternatives, with the exception of bus connection, are not only time reliable but also electrified, so their use is air-pollution free and carbon neutral. The rubber-tired technologies investigated are almost free of noise nuisance. The comparative examination took place in the framework of a CIVITAS CATALIST project completed in 2011 [1], [2] by Thessaloniki Public Transport Authority.

The paper is organized as follows. The available data about the airport ground access as well as about the modal split of trip makers to/from the airport are presented in section 2. Section 3 provides a short description of the three alternative fixed track rail options as well as of the bus option. The methodological approach to estimate/forecast travel behavior of trip makers towards each different option and consequently future demand is presented in section 4. Finally, Section 5 is focused on the results and findings of the analysis including the different cost items that correspond to each future option. It also includes the main conclusions and provides some discussion points. This paper presents the methodology and the results of the comparative examination of these three rail alternatives to the existing bus connection.

2. Airport Ground Access Data

The ground access of Thessaloniki airport is made through a specific road link that connects all the airport facilities (passenger and freight terminals and other airport installations) with the primary road network of the area. The closest intersection, controlled by traffic lights, is approximately 2.4 kms from the airport passenger terminal main entrance. The number of travelers using this link is made up of air passengers, their escorts, the airport employees and the drivers of distribution and other supply vehicles. The latter comprise a very small percentage of all trip makers.

Based on surveys made in the Athens International Airport [3], escorts account for approximately 10% of air travelers. These figures have not been verified lately, but it is believed that they have not been considerably altered. In addition, for those escorted, 1.5 persons in average escort one traveler. In this way it was made possible to estimate the number of escorts on a typical day. Regarding the number of employee trips per day this depends on the people working at the airport and the transport mode used. Employees vary between seasons. In winter time some 1,125 persons are employed whereas in summer time they reach 2,000 persons. Car-pooling/sharing is not yet developed in Thessaloniki and in Greece in general, and therefore each employee trip to/from the airport using private car should be considered as a distinct trip. The daily accumulation of airport employees, as obtained from Thessaloniki Public Transport Authority, is shown in Figure 4.

Traffic to/from the airport was measured using automatic traffic counters which were installed on the road link at both directions for one full week. Vehicles were classified in passenger cars, , light trucks and heavy trucks. Public Transport buses – as a sub-total of all buses- were calculated based on scheduling data, whereas taxi trips were calculated by manual counts in conjunction with automatic traffic counts. In addition, the average vehicle occupancy for private cars and taxis was determined by sampling surveys during different periods of the day. By combining traffic on the airport ground access link with passenger flows – in and out - during the data collection days, it was made possible to estimate the average modal split of all passenger trips to/from the airport. Tables 1 and 2 present the daily modal

4 split figures for all trip makers in absolute and % terms and the % share of all different categories on an annual basis.

Table 1: Overall modal split of trip makers to/from Thessaloniki Airport

Mode Vehicle Occupancy Person trips Modal Split Taxi 1,5 1850 17% Bus 20% 1960 18% Car 1,6 6920 65% Total - 10730 100%

Table 2: Overall modal split of trip makers to and from the airport

Segment Person trips % share Air travellers 6.50 mil 70% Escorts 1.85 mil 20% Employees 0.93 mil 10% Total 9.28 mil 100%

During the year some 2,600 commuter trips/day are made from which 75% by car and 25% by PT bus. Figure 4 presents the daily accumulation of airport employees in a typical day.

Source: Thessaloniki Public Transport Authority Figure 4: Daily accumulation of airport employees at Thessaloniki Airport

3. Alternative Rail Links to Thessaloniki airport 3.1 Overview of alternative fixed track systems

Three main alternative connections between Thessaloniki airport and the city future Public Transport network, relying on different rail technology, were examined: a) a further extension of the Metro line to Municipality which is under construction and will terminate at MIKRA station located at the eastern part of Thessaloniki conurbation, some 5.1 km from the airport; the Metro extension will be a surface line. b) a segregated Tramway/LRT, which will start from MIKRA terminal station and will end at the airport; the ram/LRT route will follow the alignment of the existing highway with minor deviations c) an elevated Monorail connecting the same two terminal points; a single track is foreseen for the next 10 years period, with double track at the stations.

5

The three alternatives comprise Scenarios A, B and C respectively. All scenarios are compared among them and with scenario Do-Minimun which refers to an improved airport bus connection starting from the MIKRA station and offered by the Public Transport Operator of Thessaloniki. The Business as Usual (BAU scenario) is also examined for evaluation purposes.

For passengers travelling from Thessaloniki city area, all but metro alternatives would require a transfer at the terminal metro station of MIKRA, properly designed to reduce the resulting inconvenience. The new fixed track corridor under consideration will consist of 5 stations, including MIKRA and Airport terminal stations, and will have a total length of 5.1 km (Figure 5). Estimations of passenger demand at the intermediate stations of the connection were made by means of a land use inventory, subsequent trip generation calculations and the use of a transport planning model built exclusively for this consideration [4]. The modal operating capacity selected, covers 10 min policy headway during peak periods, a 25% rail transit share and a directed loading of 1,300 pphpd.

Figure 5: Proposed new Rail-Air Link with intermediate stops

A comparative analysis of the above three different rail systems was performed using the two staged approach of the World Bank [5] as refined later in 2001 [6]. According to this methodology in the first stage non cost attributes of the modal options are considered; most adequate systems will come out from this exercise considering also the demand and supply elements in each specific option. In the second stage the choice of the most suitable mass transit technology will be made in terms of the total costs, namely user costs, operator’s cost and community (social) costs. All alternative options were compared against an improved future bus connection starting from the MIKRA terminal station.

Capital costs differ significantly among the three fixed track alternatives as it is the case for the annual operating costs. The Metro alternative bears high capital and expropriation costs as well as operating costs. Buses on the opposite side are linked with high operating costs and low capital costs. On the other hand, Metro bears high benefits to its users, especially because there will be no need for transfer from one PT mode to another. Tram/LRT and monorail lie in between.

6

Investment and O&M cost figures were obtained from Attiko Metro, a state-owned company, responsible for the physical design and construction of Metro and Tram/LRT systems in Greece. Monorail cost figures were obtained from People Mover System of Intamin which was implemented at Bologna airport in .

3.2 System, User and Community Costs

Costs examined within the comparative analysis performed, include three main categories, Operator’s costs, User costs and Community costs.

Operator’s costs can be further broken down into capital costs (Rolling Stock, Land, Infrastructure, Electromechanical, Depots, Overheads) and Operating Cost (Staff, Energy, Materials, Outsourcing, Overheads. Effective life of all systems was taken equal to 30 years. No depreciation cost was taken into account at this preliminary analysis level.

User costs consist of Time cost (linked to value of time) which further distinguish in In-vehicle journey time cost, Transfer penalty costs and Travel variability risk. A mode specific cost (benefit) expressed as Novelty image was also included in the analysis. It is expected that mode specific benefit pertains to Metro, Tram/LRT and Monorail in comparison to buses.

Community costs consist of Employment costs, Land acquisition costs, and Climate related costs (benefits).

Table 3 presents in summary the comparison of the 3 fixed track alternatives and the improved bus connection option.

Table 3: Main characteristics of alternative Airport Connection options

Scenario A B C D Metro Tram /LRT Monorail System (2 trains of (2 trains of (2 trains of Bus 3 cars) 3 cars) 6 cars) Route Length m 5,100 5,100 5,100 6,300 Car Capacity pax 150 200 34 150 Transit Unit Capacity pax 450 200 204 150 Max design capacity pphpd 10,800 4,000 1,300 900 Max speed km/h 80 70 80 80 Commercial speed km/h 32 25 40 21-35 In- Vehicle Journey Time min 10 13 8 15 Transfer Time (Mikra) min 0 1 3 3 Walk Time / Egress Time min 3 3 3 3 Transfer Penalty (Mikra) min 0 1 3 3 Policy Headway min 10 10 10 10 Novelty Image (-) min 1 1 2 0 Minimum Layover Time min 2,5 2,5 1,5 10 Number of transit units p.h. no 4 5 2 6 Capital Cost (2011) K€ 126,777 78,239 95,567 3,600 Operating Cost K€ 4,200 3,300 4,000 2,200 Land Cost K€ 4,000 2,900 525 0

7

4. Travel Demand Forecasting 41. Stated Choice Experiment

To calculate user related costs and benefits per alternative system considered, it is necessary to have a reliable forecast of the passenger demand per day and per year for the analysis time period. Passenger demand is different for each fixed track system based on the time and fare cost associated with each travel option as well as on the image of each travel mode. Given that no such system is in operation in Thessaloniki Greater Area, a combined Revealed and Stated Preference Survey was undertaken for capturing the current and the stated choices of the persons travelling to/from the airport for different purposes as well as their attitudes and preferences.

A special questionnaire form was designed for this purpose and was used at the airport site. Specific questions about the three alternative future fixed track systems were included in the questionnaire form. In total 500 valid questionnaires were collected from five discrete segments, namely Greek domestic travelers, Greek international travelers, foreign travelers, escorts and employees. Trip characteristics of travelers were also gathered. Different criteria with respect to the most attractive mode to the questioned were set. Table 4 presents the responses of all persons in the sample to those criteria.

Table 4: Mode choice criteria per user group category

Criterion Air Travelers Escorts Employees Duration of the trip 92.6% 92.0% 98.0% Cost of the trip 90.0% 91.0% 93.0% Minimization of transits 89.6% 88.0% 91.0% Comfort 82.0% 79.0% 87.0% Reliability 95.6% 93.0% 97.0%

Binary Logit models for determining the travelers’ choices as a function of their personal and trip characteristics were developed from the gathered revealed preference and stated choice data. Travelers were categorized based on their current travel choice (bus, car, taxi). Two binary choice models were set up for each traveler type; a simplified model including only mode specific variables like travel time, travel cost and number of transfers (model 1) and an extended model (model 2) including also socio-economic variables like the age, the gender, the frequency and the trip purpose. In total 18 different models were developed which were calibrated through the freeware package BIOGEME, Bierlaire, [7].

More detailed about the choice modelling can be found in Politis, et al [8]

4.2 Travel Demand Model

A travel demand model using the PTV’s VISUM suite was employed for determining future demand for each different transport mode. The model used was the one developed and calibrated by Thessaloniki Transport Authority in 2011. By using the related behavioral characteristics obtained from the state choice experiment per scenario examined, it was made possible to construct and apply three transport planning models, one per alternative. In addition, two more models were built, one for the existing bus connection, Business As Usual (BAU) and one for an improved bus connection, starting from MIKRA terminal station every 10 min (DO_MIN). The outputs of the model-runs for horizon 2024 in terms of peak hour maximum passenger load per direction for the heaviest sections are shown in Table 5.

Table 5: Projected peak hour Maximum Passenger load per Scenario (2024)

8

Max Passenger Scenario Direction Segment Load ph A From Airport 1,170 ZEDA-MIKRA (Metro)) To Airport 740 MIKRA-ZEDA B From Airport 548 ZEDA-MIKRA (Tram/LRT) To Airport 328 MIKRA-ZEDA C From Airport 794 ZEDA-MIKRA (Monorail) To Airport 456 MIKRA-ZEDA BAU From Airport 76 KRIKELA-25HS MARTIOY (Bus) To Airport 282 FALIRO-SXOLI TIFLON DO_MIN From Airport 788 IKEA-SASTH-VIAMIL (Bus) To Airport 203 EMPORIKO KENDRO-POLYFOTA Fixed Track Route Stations: MIKRA, ZEDA, IKEA, GREEN FIELDS, AIRPORT Bus Route Stops: KRIKELA, FALIRO, SASTH, POLYFOTA

The obtained results clearly indicate that the Metro option attracts the highest passenger load, whilst the Monorail comes second. The improved future bus connection comes very close to the monorail option. These findings are quite expectable given the high cost of the transfer penalty that imposed to all alternatives except the Metro system that does not require a transfer.

5. Comparative Analysis Results and Conclusions The results of the comparative analysis performed among the three fixed track alternatives and the future improved bus connection in terms of total implementation and operating costs are shown in Table 6. It should be stressed that the bus option refers to a link from MIKRA terminal station to the airport and vice versa and not to a direct link from/to Thessaloniki city center.

Table 6: Overview of Total Cost (Benefit) per Alternative Transport Mode

Tram/ Mass Transit Technology Metro Mono-rail DO_MIN (Bus) LRT User Costs In Vehicle Journey Time Costs 8.759 11.180 7.263 10.698 Transfer Penalty Costs ('Mikra' terminal) 0 1.124 3.075 3.075 Cost of Travel Time Variability (Risk) 0 0 0 3.531 approaching the Airport Novelty Image Cost (benefit) -1.149 -1.149 -2.300 0 Sum User Costs 7.609 11.154 8.039 17.304 Operator’s Costs 0 0 0 0 Capital Costs (8%, 30 years) 11.083 6.840 8.354 315 Operating Costs 4.133 3.248 3.937 2.165 Sum Operators' Costs 15.216 10.088 12.291 2.480 Community Costs 0 0 0 0 Employment Costs (benefits) -341 -297 -312 -219 Land Acquisition Costs (4%, 30 years) 227 165 30 0 Climate Cost (benefit) 0 0 -69 0 Sum Community Costs -113 -132 -351 -219

SUM Total Costs (000) 22.712 21.110 19.978 19.565 (annualized 2018 prices)

All costs were annualized in order to allow for a direct comparison. Transit investments assumed to be made by a 30-year loan with an interest rate of 8%. Annual payments cover both principal and interest. Targeted passenger demand is achieved in 2030, whilst in 2024 the lower demand is satisfied at a

9 reduced cost by longer headways. For the Metro option, land expropriations are necessary whilst for the Tram/LRT system the respective needs are smaller. The Monorail does not require any expropriations. In addition, Tram/LRT and Monorail require space for a depot which depends on the “future” number of transit units. There is no such a need for the Metro since a depot will be available from the main Metro line. Regarding operating cost of each system, a detailed cost calculation was made. The same figure applies to energy consumption, spare parts needs and other outsourcing costs. Finally, an average farebox revenue of 2€ for rail alternatives, and 1€ for bus, was assumed as a 100% operating cost recovery.

The summary findings presented in Table 6 indicate that the alternatives examined do not differ significantly in terms of total cost. The bus, as expected is the cheapest one, but it does not secure in the future an adequate level of service. Furthermore, it is more vulnerable to congestion conditions as well as to unexpected events. Monorail seems to be the cheapest option among the fixed track ones and most easy to implement. The Tram/LRT is the least preferred option, given that it is the most expensive after the Metro and at the same time it is associated with many other disadvantages such as need for land space, interaction with traffic and need for a new depot. The Metro on the other hand is associated with many advantages but at the same time is the most expensive and time consuming with a very high capacity reserve (low utilization rate).

An additional aspect that should be taken into account, is the potential for development in the area along the fixed track and the possible gains in land value. Such gains can partially finance implementation and operation of the air-rail link. This is not possible in case of maintaining the existing bus connection or changing to a new shorter one. In any case, a full feasibility study will provide a more detailed determination of costs and benefits and will examine the different risks associated with each alternative rail technology. Risks seem to be higher in the cases of Metro and Tram/LRT given the lengthy procedures required to mature the projects as well as the administration and transition cost. Furthermore, in terms of funding/financing, Monorail requires less advanced funding and it is easier to construct, a dimension that is not fully reflected in the quantifiable analysis presented. Finally, Monorail does not cause traffic impacts during construction and requires, if not at all, very limited expropriation. On the other hand, a constantly developing passenger demand will lead to the need for doubling the Monorail track which will increase the capital and maintenance costs of this system.

In conclusion, the Monorail System seems to be directly competitive to all other alternative rail systems which have been implemented and operate in Athens. A more detailed cost and benefit analysis including determination of wider benefits to the local society and based on an updated and comprehensive traffic forecasting model will be necessary before a final decision is made.

References

1. Thessaloniki Public Transport Authority, Connection of Thessaloniki International “Macedonia” Airport with future fixed track rapid system network of Thessaloniki, Project co-funded by the EU CIVITAS CATALIST Programme, Feasibility Report, Greece, 2011 2. Thessaloniki Public Transport Authority, Connection of Thessaloniki International “Macedonia” Airport with future fixed track rapid system network of Thessaloniki, Project co-funded by the EU CIVITAS CATALIST Programme , Stated Choice Survey Report, Greece, June 2011 3. Athens International Airport, AIA 2003 Passenger and Employee Survey, Greece, 2004 4. Thessaloniki Public Transport Authority, Greater Thessaloniki Area Transport Planning Model for the Metro extension to Kalamaria, Technical Report, 2011

10

5. [Armstrong-Wright Alan, Urban transit systems: Guidelines for examining options World Bank Volume/Edition 52, Washington D.C.1986. 6. Parajuli P., Wirasinghe S., A Line Haul Transit Technology Selection Model, Transportation Planning & Technology, Vol 24, 2001 pp. 271-308 7. Bierlaire, M., “BIOGEME: a free package for the estimation of discrete choice models”, Proceedings of the 3rd Swiss Transport Research Conference, Ascona, , 2003 8. Politis I., Papaioannou P., Georgiadis G., “Airport Ground Access Logit Choice Models for Fixed Track Systems”, Title of the Book: “Transport Infrastructure and Systems”, ISBN: 978-1-1380- 3009-1, Edited by Gianluca Dell’Acqua and Fred Wegman, Published by CRC Press, Taylor and Francis Group, Paper Presented at the AIIT International Congress on Transport Infrastructure and Systems, 10-12 April 2017, , Italy

11