Integrating Aviation and Passenger Rail Planning (2015)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

35 C H A P T E R 3 Introduction and Structure Chapter 3 now examines the North American experience with accessing airports by long-distance rail, presented in three parts. Part One presents a brief introduction to the five United States airports where rail stations have been built on rights-of-way served by Amtrak, and reviews what is known about their markets in general and the extent of their use by intermodal passengers using rail from beyond the traditional metropolitan area of the host airport. Part Two presents an original analysis of the roles of various parties in determining whether or not to have a long-distance rail stop at a candi- date airport; a detailed analysis is presented concerning the business case for stopping and not stopping longer distance trains to serve an airport. Part Three uses the North American experience to update the Site Planning Typology presented in Chapter 2, which had been based on European observations, and notes the similarities and dissimilarities between the two sets of site planning characteristics. The Chapter concludes by noting that, in many cases, the manner in which rail services participate in a fully integrated multimodal transportation system may have little to do with whether or not service is provided directly to the physical site of an airport. Part One: United States Airports Served by Long-Distance Rail A key aspect of analyzing the potential for improving rail access to airports is to provide an assessment of the effective- ness of regional and intercity rail service in operation today at selected airports in the United States. To this end, five air- ports where long-distance rail access is currently in place were examined with regard to their physical planning/design char- acteristics, current ridership/modal split, and their degree of success in attracting air travelers to the rail mode for access/ egress within their regions. The five airports are: Baltimore/ Washington International Thurgood Marshall Airport (BWI), Bob Hope/Burbank Airport (BUR), Milwaukee’s General Mitchell International Airport (MKE), T. F. Green Airport, serving Providence (PVD), and Newark Liberty International Airport (EWR). In Canada, examples of connections between long-distance rail and airports include a shuttle car service (upon a telephoned request) between Via Rail’s Dorval sta- tion and the passenger terminal of Montreal-Trudeau Inter- national Airport, a distance of about 4,500 feet. This Chapter will focus on the five airports in the United States. Examination of these airports will provide a framework that can be utilized to interpret the efforts undertaken to date in the planning, design, and implementation of rail as a feeder mode to O’Hare (ORD), San Francisco (SFO), and San Diego (SAN) in Chapters 6, 7, and 8. Each of the five airports has been examined and illustrations have been created to describe the rail alignment’s proximity to the airport terminal com- plex, as well as the modes of transport providing connections to the terminals from the rail station. The results of this analysis can serve as the framework for intermodal terminal development in other study airports and offers “lessons learned” for airports where intermodal termi- nals are being considered as part of an expanded regional rail network and HSR. Baltimore/Washington International Thurgood Marshall (BWI) Airport Rail Station The Setting: The Airport and Long-Distance Rail The Baltimore/Washington International Thurgood Mar- shall Airport (Figure 3-1) has had rail service for over 25 years. The State DOT, the Airport Department, and Amtrak have sponsored numerous planning studies to investigate ways to enhance and expand the service. Amtrak developed a small sta- tion west of the airport terminal along the Northeast Corridor Connecting Airports with Long-Distance Rail in the United States

36 (NEC) in the 1980s. Between Amtrak and Maryland Area Regional Commuter (MARC) train service, there is frequent service from Baltimore’s Penn Station and Washington’s Union Station to the BWI train station, with travel times of 15 minutes and 30 minutes, respectively. Passengers transfer onto a bus at the BWI station to access the airport terminal, approximately 2 miles away. Estimates of Intermodal Ridership In the 2009 survey conducted by the Metropolitan Wash- ington Council of Governments (MWCOG), there were 8.7 million originating passengers at BWI (MWCOG 2010). According to that multi-state survey effort, 140,000 pas- sengers (annually) traveled by Amtrak and MARC com- bined. If Amtrak is getting half of that, that would be 70,000 air passengers, or about 200 per day. About 880 passengers board Amtrak services every day (with a similar amount dis embarking from trains), which suggests that airline pas- sengers are not the largest segment of users of this station. Intercity Amtrak riders who board at this station are going primarily to Philadelphia, Newark, and New York City, as shown in Figure 3-2. The station is located in a prosperous suburban area south of Baltimore; it would be a good candidate for both park- and-ride and kiss-and-ride patrons making the trip north. A major feature within the station area is a large parking garage that is well utilized by Washington-bound, MARC commuters living in the airport area. Although the current station is outdated in terms of basic amenities, it was recently updated by Amtrak to assure longevity—new high level plat- forms, lighting, and elevator enhancements. Light rail serving Amtrak BWI Rail Station Air Terminal Area 9,000 Feet by bus Figure 3.1. The BWI Marshall Airport Rail Station is located 9,900 feet from the passenger terminal. Imagery © 2014 Commonwealth of Virginia, DigitalGlobe, U.S. Geological Survey, USDA Farm Service Agency, Map Data © Google, 2014. Figure 3-2. Amtrak station summary. Source: Amtrak 2011.

37 No tangible steps were taken to implement this concept to date; however, the level of improvements at the airport since this timeframe would not preclude future consideration of this concept. The concept allows for a re-routing of the cur- rent airport light rail line to make way for major runway construction on the eastern portion of the airport. However, the concept would take a direct connection between the light rail and the terminal complex and replace it with an indirect one, relying on a transfer to a people-mover to gain access to the terminal complex. Importantly for the themes of this case study, the existence of the “intermodal terminal” did nothing to overcome the problem that the long-distance rail station is located a considerable distance from the airport passenger terminal, dealing instead with a highly unlikely maglev line. Bob Hope/Burbank (BUR) Airport Rail Station The Setting: The Airport and Long-Distance Rail The local commuter rail system (Metrolink) and Amtrak’s Pacific Surfliner route (10 trains per day) service the Burbank Airport Rail Station (Figure 3-3) daily. The Surfliner is the second busiest service in Amtrak’s national network. With a boarding platform only, it is not supported by a separate station structure, service amenities, or parking lots, although rail users can park at airport rates on the airport lots. Estimates of Intermodal Ridership The majority (about 60%) of Amtrak users at BUR are traveling to destinations within 100 miles, including Santa Barbara and Oceanside, while about one third are going to San Diego. Figure 3-4 shows that very few are going to Los Angeles, as there are other, cheaper modes in the region, including Metrolink from the station. At present, about 60 passengers per day board one of the five Amtrak Pacific Surfliner trains (in each direction) serving BUR. Because this is the only Amtrak stop in the city, it must be assumed that some of the users are gaining local access, and not air- port access. The two highest origins for Amtrak users are San Diego and Santa Barbara. The Research Team estimates that, if there are any airport users on the train, most do not come from either San Diego or Los Angeles, because both airports offer better air service, including service by Southwest Airlines, the main operator in Burbank. Thus, it is estimated that only a very small number of the persons boarding the long-distance train daily are coming from markets that would logically feed the airport. downtown Baltimore also has its southern terminus at BWI at the extreme west end of the terminal complex. The Research Team concludes from the continuous sur- veying process of the local metropolitan planning organiza- tion that the BWI catchment area has shrunk over the last decade. The Research Team estimates that in 2000, 28% of the BWI-originating passengers came from outside the suburban ring of communities surrounding Baltimore and Washington DC. By 2007, that portion of originating riders had shrunk to 17%. MWCOG noted this change and a decrease in relative importance of the District of Columbia as an origination source over that period; they write, “the percentage of pas- sengers originating in the District of Columbia declined by 22 percent and by 43 percent from the outlying areas between 2000 and 2007” (MWCOG 2010). This direct evidence suggests that BWI is gaining strength from its close-in markets, well served by modes such as taxis and shared ride vans, rather than by fast trains in either direction. Coincidentally, the period in question represents a time when Southwest Airlines altered its strategy of serving the Philadelphia area from BWI, and started serving other more NEC airports. According to the interviews, air passen- gers’ use at the BWI Amtrak station was highest during the period when Southwest did fly to BWI, but did not fly to PHL. Today, the Texas-based airline flies to PHL, EWR, and LGA, and will expand its regional presence considerably with the ongoing consolidation of services of its AirTran airlines subsidiary. Background and Other Activities Shortly after September 11, a major feasibility study was conducted to develop a remote intermodal terminal com- plex (ITC) to consolidate some rail access modes, future maglev service, and consolidation/centralization of air pas- senger processing functions, including ticketing, security, check-in, and bag claim. An automated “people-mover” sys- tem would provide the landside distribution system to the airline gate boarding piers. The motivation for this study focused on: • Enhancing BWI’s competitive position in the region; • Alleviating ground access congestion and parking shortages; • Providing more airline gate positions; • Creating joint development opportunities, such as conference/ meeting centers, hotels, airport-related offices, etc.; and • Improving intermodal connections and convenience to attract a higher rail mode share (at least in theory, if the people-mover were later extended to the rail station).

38 Amtrak Burbank Airport Rail Station Distance to Terminal 1,500 Feet Figure 3-3. The Amtrak rail station is 1,500 feet from the Bob Hope/Burbank Airport. Figure 3-4. Amtrak station summary. Source: Amtrak 2011. Burbank Airport Station is an example (similar to Lyon Airport, Chapter 2) of an airport well situated from a physi- cal perspective, but with little market for intermodal services. Background and Other Activities In the fall of 2012, the Burbank Airport commenced the construction of the Regional Intermodal Transportation Center, across Empire Avenue from the Amtrak rail station. According to the airport management, “The RITC will be a three-level structure housing a consolidated rental car facility and rental car customer service building and will include a bus transit station on the ground level. An elevated moving walkway will convey rental car customers and rail and bus passengers between the airport passenger terminal and the RITC, making Bob Hope Airport uniquely convenient and accessible via multiple transportation modes” (Bob Hope Airport 2012). Relationship with HSR According to the Research Team’s interviews California High-Speed Rail Authority (CHSRA) is very skeptical about adding additional stops to its main line HSR service, even when the station would be close to an airport. At the time of this writing, the CHSRA plan calls for a single combined Burbank/Glendale station that is not in downtown Burbank and not at the BUR location. This station plan could, of course, change over time, as the plans for rail investment are still evolving. Thus, the rider would be encouraged to take some other form of rail, possibly a future Orange Line, to get to a transfer point with the HSR corridor. Reportedly, BUR was looking to the Orangeline Development Authority (OLDA),

39 In light of this capacity problem, it would make little sense to use HSR to feed the development of this airport. General Mitchell International Airport (MKE) Rail Station The Setting: The Airport and Long-Distance Rail The MKE rail station was intended to serve as an airport rail link for General Mitchell International Airport, which serves the Milwaukee area. Figure 3-6 shows the location of the Amtrak Airport Rail Station, which is connected by shut- tle bus to the primary passenger terminal check-in area. The station serves about 156,000 passengers per year (Figure 3-7); by way of comparison, the downtown Milwaukee terminal of the same line attracts about 576,000 passengers per year. now renamed as Eco-Rapid Transit, to develop a regional high-speed transit connection to cities between Santa Ana and Santa Clarita, with stops in Glendale, downtown Burbank, and BUR (AirRail News 2011, Orangeline Development Authority 2011). In examining the possible future role of a Burbank Rail station as part of a larger strategy to maximize the quality of access to key airports, it is important to note that Burbank Airport is highly constrained in its growth potential. A regional analysis done in Southern California suggests that BUR will experience a significant capacity constraint by the year 2020. As part of the San Diego Regional Airport Strategic Plan, the firm of Leigh|Fisher (operating as Jacobs Consultancy at the time) concluded that the existence of HSR between Los Angeles and northern California would delay the impact of the capacity constraint at Burbank Airport by several years; however, it would still occur by 2025, as shown in Figure 3-5. BUR enplanement growth with no HSR BUR enplanement growth with HSR in place Capacity cap at BUR airport Figure 3-5. Capacity constraint at Burbank Airport. Source: Jacobs Consultancy, Presentation #2 Preliminary Findings–Remaining Scenarios, Slide 39, Dec 9, 2010. Annotations added. Connection by bus Figure 3-6. Amtrak station is 7,500 feet from General Mitchell International Airport. Imagery © 2011 DigitalGlobe. USDA Farm Service Administration, US Geological Survey, Map Data © Google.

40 itself as Chicago’s third airport, attracting a significant number of airport users from the north side of the city (Trains 2006). Before its demise, Midwest Airlines (whose hub was MKE), supported the airport station because it identified North Shore customers as a potential market. Midwest attracted more than 6% of its customers from northern Illinois (Destination: Freedom 2004). Estimates of Intermodal Ridership There is no evidence that an appreciable number of air passengers are using Amtrak to gain long-distance access to this airport. A recent study by the Texas Transportation Institute concluded that “[u]se of the airport rail station to transfer to airline service was minimal, reported by 6 percent of current Hiawatha Service passengers” (Sperry and Morgan 2011). Their calculations found 9,600 passengers per year used the station to transfer to the airport. This is about 26 passengers per day, or 13 passengers boarding the train at the airport station. Background and Other Activities Station History The idea for an Amtrak station at the airport came from executives at Milwaukee-based Midwest Airlines in 1991. Midwest executives hoped to generate revenue partly by negotiating a deal with Amtrak that would allow custom- ers to use one ticket for their rail and air travel. The airline allowed its customers to use their Amtrak travel for frequent flier miles (Allison 2004). The project began with a 2000 tour of the site by then- Wisconsin Governor Tommy Thompson, who had served as a director of the national railroad from 1990 to 1994. In 2000, it was thought the station would become part of a statewide HSR network that had been proposed. The project hoped to attract airport passengers and commuters living in Milwaukee’s southern areas who did not want to cope with traffic and park- ing downtown. Under the then-proposed $4 billion Midwest Regional Rail Initiative, the existing Hiawatha Service would be expanded, and perhaps extended, to serve Green Bay and Madison (Allison 2004). In 2001, Midwest Airlines and Amtrak reached an agree- ment to allow for code-sharing, construct a station, and increase the number of trains serving the station by 10. The plan envisioned a net positive cash flow as a result and collaboration with the freight railroad (Martin 2011). The September 11 terrorist attacks forced the suspension of plan- ning to provide direct baggage handling to MKE via Amtrak. Funding the construction of the station was an issue; neither the airport nor Milwaukee County officials were willing to commit funds. The station includes a Quik-Trak ticket vending machine, restrooms, a seating area, covered walkways to both the drive-up/drop-off area, and a parking lot. As the station is unstaffed, all tickets from the station must be purchased from an onboard conductor or from Quik-Trak in advance. The station parking lot contains 300 spaces and costs $5 per day per vehicle. All revenue generated from parking fees is used to finance the station’s operating costs. MKE is served by the Hiawatha Service, an intercity pas- senger rail service operated by Amtrak that comprises an 86-mile train route on the western shore of Lake Michigan. At present, 14 trains (seven round-trips, six on Sunday) run daily between Chicago, Illinois and Milwaukee, Wisconsin; these trains make intermediate stops in Glenview, Illinois; Sturtevant, Wisconsin; and General Mitchell International Airport. The line is partially supported by funds from the state governments of Wisconsin and Illinois. The corridor rail service carried 783,060 riders in FY 2010, a 6.1% increase from FY 2009’s total of 738,231 passengers (Amtrak MPR). It is Amtrak’s ninth-busiest route, and the railroad’s busiest line in the Midwest. A one-way trip between Milwaukee and Chicago takes about 90 minutes. MKE views Figure 3-7. Amtrak summary. Source: Amtrak 2011.

41 Providence/T. F. Green Airport (PVD) Intermodal Center The Setting: Airport and Long-Distance Rail In the winter of 2010–2011, a new air/rail station was opened at PVD (Figure 3-8), which serves Providence, Rhode Island. Amtrak would not serve the station, but there was a hope that MBTA commuter service would provide service appealing to some air passengers. The rail service is viewed as a park-and-ride for commuters to Boston whose origins are actually south of Providence. The parking at the intermodal station is designed for daily commuters, with overnight parking discouraged. The intermodal facility can be seen as having six compo- nent elements, including: • Commuter parking garage ($28.1M); • Rental car garage ($46.9M); • Rental car desk and service areas ($40.2M); • Commuter rail platform ($22.9M); • Skywalk, 1250 feet ($43.5M); and • Connection from airport terminal to skywalk ($14.1M). Estimates of Intermodal Ridership Since the summer of 2012, there are eight trains per day departing T. F. Green station with service to Boston. There are seven trains per day departing from Boston going to T. F. Green. The train departure times are not designed to serve air passengers, as the service is primarily designed to take commuters from the area of the airport to Boston. By way The station opened for service in January 2005. At the time, it was only the fourth Amtrak station to have direct service to an airport; the others included Baltimore, Newark, and Burbank. Amtrak’s vice-president of sales and marketing noted at the opening ceremony, “Amtrak looks forward to building upon its growing Hiawatha Service by introducing another rail to air connection. The successful experience in the East with the Baltimore/Washington International Airport (BWI) station demonstrates the great potential for the Milwaukee airport station” (Destination: Freedom 2005). Midwest Airlines chair- man and chief executive officer at Midwest summarized their hopes: “It will provide passengers traveling to and from north- ern Illinois and Chicago faster and more convenient access to Mitchell International Airport and Milwaukee’s hometown airline” (Mitchell Airport 2009). On June 23, 2009, Republic Airways Holdings acquired Midwest Airlines and discontinued the code-sharing arrangement between the airline and Amtrak (Daykin 2009). Relationship with HSR The Milwaukee Airport Rail Station is included in every long-term HSR plan now under consideration, as described in Chapter 6. It is seen as a major potential air/rail transfer point in both the near-term plans for 110 mph service in the region, and in the longer term vision of 220 mph rail in the region. When the state decided to cease planning for services between Milwaukee and Madison, plans for invest- ing in services between Milwaukee and Chicago were not abandoned. Combined Rail and Rental Car Facility Moving Walkway Connector to Terminal Figure 3-8. The airport rail station is 1,300 feet from the air terminal area. Imagery © 2011, DigitalGlobe, dGIS, GeoEye, MassGIS, Commonwealth of Massachusetts, Map data © Google.

42 include new stops at Philadelphia and Westchester Airports, but not at the existing facility at T. F. Green. Newark Liberty Airport (EWR) Rail Station The Setting: The Airport and Long-Distance Rail The Port Authority of New York and New Jersey (PANYNJ), in partnership with New Jersey Transit and Amtrak, devel- oped a new ITC on the NEC (see Figure 3-9) immediately adjacent to Newark Liberty that became operational in 2001. This very significant development included an airline bag- gage check-in facility and a direct connection to an extended landside automated people-mover system—AirTrain—that had been serving the airport complex’s parking, rental cars, and three terminals since the mid-1980s. Currently, Amtrak stops only 10 trains in each direction at the Airport Rail station. By contrast, New Jersey Transit (NJT) operates more than 80 trains in each direction on weekdays, resulting in a good amount of service, but sometimes with gaps inappropriate for an airport rail service. The primary features of this intermodal terminal include very frequent rail service from Manhattan and Newark by NJT, the regional commuter rail provider. While the 10 Amtrak regional trains do stop, there is no Acela service. Taking boardings and alight- of example, after 4:00 p.m., the only train from the airport to Boston departs at 10:30 p.m., with four trains in total between 7:00 a.m. and 4:00 p.m. Similarly, there are only four trains to take passengers from Boston to the airport before 4:00 p.m. The number of air passengers on the service is expected to be minimal given the difficulty of pairing rail schedules with air schedules. Background and Other Activities The Intermodal Center at PVD was conceptualized as a major application of the principles of passenger intermodal- ism, with the hope of uniting air and rail services at a rapidly growing “low-cost” airport. Most relevant for this case study is the fact that Amtrak, unlike service decisions at BWI and at Newark, decided not to stop any trains at the new facility. In order to understand the dynamics of these decisions, this report includes a thorough review of the business case for providing a new stop at this airport, which is presented in Part Two of this Chapter. Relationship with HSR At present, there is no planned activity to link the airport serving Providence, Rhode Island into the HSR system of the NEC. Amtrak’s original NextGen proposal, for example, did Newark Airport NEC Rail Station Location Connection to the terminal area is via an older people- mover Figure 3-9. The center of the terminal area is about 6,500 feet from the rail station at EWR. Imagery © Bluesky, DigitalGlobe, Landsat, Sanborn, USDA Farm Service Agency, Map data © 2011 Google.

43 States, which is documented in ACRP Report 4. Although the number of tickets sold through this program is considered proprietary, the reader can gain a sense of the scale of the program from the more general discussion of long-distance ridership that follows. Additional studies of the intermodal operations at EWR were undertaken by the I-95 Corridor Coalition; these studies are available through the I-95 web site (I-95 2004 and I-95 2004a). Estimates of Intermodal Ridership Understanding the Long-Distance Market to Newark Given Newark’s prime location along the main intercity rail line of the nation, it is reasonable to expect that the mode share to rail for the long-distance traveler would be good, and it is. Table 3-1 shows that 9.6% of those EWR air pas- sengers who are coming from beyond the metropolitan area are coming by rail and transferring to the Newark “AirTrain” people-mover (Zupan, et al. 2011). This percentage is in stark contrast to findings at MKE, BUR, and PVD; it is also seem- ingly stronger than at BWI. The Research Team has compared the mode share of the long- distance ground access market in EWR to that of JFK in order to place this relatively high American mode share to rail in context. JFK is not located on a major interstate rail corridor (although the volumes from Long Island are important). For those who do not live on Long Island, gaining access from a long-distance ground origin requires making a transfer in Manhattan to either the subway or the Long Island Railroad. From there, the air passenger from beyond the metro region gains access to the AirTrain at either Jamaica or Howard Beach. Table 3-1 shows that the mode share to rail/AirTrain at JFK is virtually the same as at EWR. Approximately 1.4% of all non-transferring airline passengers boarding a plane at EWR were long-distance rail users, transferring onto the AirTrain. For comparison, of all non- transferring airline passengers boarding a plane at JFK, about 1.0% of them were from a long-distance origin and arrived by rail/AirTrain (calculated from Zupan et al., page 130). Background and Other Activities In light of the fact that EWR offers competitive air ser- vice for mid-haul, long-haul, and overseas flights, PANYNJ is studying the potential of extending their existing rail system (PATH) from downtown Newark at Penn Station to EWR. This service would reduce transfers for PATH travelers com- ing from lower Manhattan, Hoboken, Jersey City, and other major employment and emerging residential centers en route to the airport. The initial service would tie into the ITC and require a transfer to AirTrain. ments together, Amtrak serves about 115,000 riders per year at the station (Figure 3-10). Overwhelmingly, these are airline passengers rather than employees, as lower cost alternatives are available for most employee destinations. Since the opening of the station, annual ridership on Amtrak has risen from 55,000 to 115,000 in 2010. However, as shown in Figure 3-11, the Newark Airport Rail station is dominated by NJT rail services and not Amtrak. The Research Team con- cluded that persons using the stations were from the airport, a majority being airline passengers; this is because local park-and- ride or kiss-and-ride users are barred from using the station. In addition, the rail station at EWR benefits from the only surviving air/rail joint ticketing agreement in the United Figure 3-10. Amtrak Summary. Source: Amtrak 2011. 0 500,000 1,000,000 1,500,000 2,000,000 2005 2007 2008 2010 NJT Riders Amtrak Riders Figure 3-11. Relationship between local and long-distance markets. Source: NJT.

44 Future Options for Long-Distance Rail to Connect to Newark Liberty Airport PANYNJ recently began investigating the possibility of extending PATH directly into the terminal complex itself, providing seamless access to the three unit terminals at their “front doors.” This was a result of increased interest in improving rail access to EWR and recognizing the cur- rent limitations of the automated people-mover in terms of capacity. The current people-mover system is located near the airside of the terminals and travelers must backtrack to the ticketing areas for check-in unless pre-ticketed, with board- ing passes and carry-on luggage only. PANYNJ is giving this concept further consideration as enhanced HSR is being studied for the NEC. Figure 3-12 shows four options defined in the Regional Plan Association (RPA) study undertaken for the Port Authority: 1. New AirTrain to Newark Penn Station. The EWR AirTrain, built in 1996 currently has insufficient peak hour capacity and will have to be replaced with a higher capac- ity, more technologically advanced and reliable system. This provides an opportunity to extend the replacement service from the NEC station northward into Newark Penn Station. 2. Extension of PATH to the NEC Station Combined with AirTrain Upgrade. With this option, the PATH ser- vice now terminating at Newark Penn Station would be extended about 2 miles to the NEC station, creating a two seat ride (PATH and AirTrain) for Lower Manhattan and Jersey City riders. Table 3-1. Mode share for long- and short-distance access to two NYC airports. EWR JFK Mode Mode Share of Short- distance Group Mode Share of Long- distance Group Mode Share of Short- distance Group Mode Share of Long- distance Group Personal Car 48.5% 54.5% 45.1% 42.4% Hired Car 30.4% 21.9% 32.2% 27.8% Rail to “AirTrain” 14.2% 9.6% 12.4% 9.6% Bus 4.4% 3.3% 3.6% 3.5% Local Shule 2.6% 10.7% 6.6% 16.8% Total 100.0% 100.0% 100.0% 100.0% Source: Zupan, et al. 2011. Figure 3-12. Options for the future. Source: Zupan et al. 2011.

45 The implications of providing better travel times for a minority of corridor riders, at the cost of worsening travel times for the vast majority of train riders on the corridor, needs some additional analysis and documentation. Chapter 3 pro- vides that analysis in Part Two. Looking at the Newark design options issue only from the vantage point of the major themes of this study (i.e., as if other concerns did not exist), Amtrak would be a major beneficiary in the option of abandoning the existing NEC station. One-transfer airport connections would be offered for every train operating through Newark Penn Station. However, Amtrak would not be the only operator that could provide faster running times for its patrons; NJT could improve the travel times and speeds on more than 160 trains per day. If people-mover-like schedules (such as those in opera- tion at JFK) could be provided to Newark Penn Station, the intermodal connecting options would be exceptional. From that location, there are over 375 train movements by NJT and Amtrak (not including the PATH services, nor local light rail) presented as Table 3-2 from data reported in Wikipedia. The PANYNJ announced in February of 2014 that it plans to fund a $1.5 billion extension of the PATH rail system to Newark Liberty International Airport (Wall Street Journal 2014). This is highly interrelated with the agency’s study of possible reconfiguration of the airport passenger terminal area, associated with possible construction of new runways. In effect, this funding decision focuses planning attention on RPA’s Options 2 and 3. In Option 2, PATH service would meet an airport people-mover at the existing NEC rail station; in Option 3, the PATH service would be further extended to some number of on-airport terminals. Thus, questions of how future rail facilities would interact with the future terminal facilities are being addressed on an ongoing basis by the PANYNJ. Part Two: The Business Case for and Against an Airport Rail Station Chapter 2, in its summary of the European experience in access to airports by long-distance rail, reported the conclu- sion of one of the architects of the Frankfurt air/rail system, 3. Extension of PATH onto EWR. This option takes the previous option a step further by extending PATH onto the airport to one or more terminal stations. This would eliminate the transfer for Lower Manhattan, Jersey City, and Downtown Newark originating passengers. 4. Amtrak Service Added at NEC Station. Theoretically, Amtrak could stop more than the nine trains each way each day that stops today. However, the trade-off would be greater intercity travel time for the passengers not des- tined for the airport (Zupan, et al. 2011). Part Two of this Chapter will review the logic of adding more Amtrak trains, where option four is explored from an operations perspective. The remaining three options are combinations of design options. Option 2 has the advantage of administrative simplic- ity, at the cost of direct service for any group. Under this option, automated, driverless service could continue to be supplied for whatever schedule/frequency required for airport operations or integration with other market needs, such a supplying a con- tinuous level of service for the users of the rental car facilities. A remarkable design opportunity exists because of the fact that the existing EWR AirTrain needs to be re-conceptualized, redesigned, and rebuilt in any event; at the same time, new run- way configuration needs are causing the re-examination of the present terminal locations. From a design point of view, “one or more” terminal stations could be rebuilt with clearances appro- priate for PATH vehicle technology. This makes possible the consideration of the RPA’s option three. It also makes possible a hybrid solution in which the PATH vehicle serves as the airport people-mover only when it is needed for its regional schedule requirements. The same track built for the PATH train could support an “extension” of the redesigned airport people-mover to Newark Penn Station. This would allow the PATH vehicle to operate in all three passenger terminals when its regional sched- ule justifies it; this would also ensure the ability of the airport managers to provide internal circulation at a frequency appro- priate for the other requirements of the airport. As noted in the RPA study, all of this allows for the possibil- ity of running the appropriate frequency of service from the airport to the regional system, perhaps by keeping the exist- ing station, and perhaps operating an adequate amount of service to Newark Penn Station and abandoning the existing Newark Airport Rail station altogether. The RPA notes some of the complications in its report: “A sub-option would eliminate the NEC stop altogether, shift- ing the transfer point for NEC riders to Newark Penn Station . . . Those coming from the south on the NEC would have to back- track from Newark Penn Station if the NEC station were dropped. Eliminating this station may also create complications, since the Passenger Facility Charge levied on passengers at EWR was used to build the station” (Zupan, et al. 2011). Corridor Number of Trains (weekday) Westward from Newark 188 NJ Transit from New York Penn 106 NJ Transit from Hoboken 5 NJ Transit to the CNJ main line 26 Amtrak 51 Total Daily Train Movements 376 Source: Wikipedia 2011. Table 3-2. Potential rail connections at Newark Penn Station.

46 hub without rail service. The airport would then calculate whether the cost of the airport rail station is justified by added traffic. The Airlines The airline calculation is more complex. In general, airlines would not favor a HSR link at an airport if it led to increased competition for a given profitable air market. If the market in question is short and/or unprofitable, the airline might look at the question differently, especially if the airline also oper- ated the railway service and could do so profitably. Airlines might also favor HSR connections to some of their “spokes” if shifting some of the “spoke” traffic to rail would permit better (or increased) longer-haul air service at the hub. In a more general way, airlines might also favor HSR connections if the rail service permitted better and more frequent service to a number of spokes than the airline can offer, thus (in principle) generating more demand for the longer-haul air links. It deserves emphasis that the institutional relationship among airlines and railways is important. First, if the passen- ger perceives the link as seamless (single ticket and baggage, integrated schedules, improved service), all might benefit. If airline and railway are competitors, they are unlikely to provide such service (and, indeed, many air/rail connections suffer from this problem). If the airline and railway are sepa- rately owned and managed, it may be more difficult for them to devise a common ticketing service if one of the stakeholders derives more benefit than the other. The Railway Company In order to better understand the business case for and against the addition of intercity rail stations on existing (or proposed) rail lines, this project commissioned a study of the detailed economics faced by the rail managers (Thompson 2011). This section of Chapter 3 summarizes key aspects of that study of concerns experienced by the railroad managers concerning joint air/rail projects. For a local rail connection, the calculation for the railway is simpler: do the added passenger revenues (and related public subsidy, if any) cover the investment and operating cost involved. In some cases (the WMATA link to IAD), the calculation will include the fact that the airport link will also develop a major new, local market (Reston and Tyson’s Corner) that is of interest to the rail operator. In the lon- ger run, most local rail connections tend to generate related economic activity that adds to the financial and economic performance of the local link. Adding an intercity rail link (including HSR) to an airport would raise a different calculation (unless the link is simply a stub connection to an airport having the effect of better local that for such a scheme to work, there must be buy-in from all parties—something he doubted would happen in many attempts to replicate the Frankfurt experience. As noted ear- lier, for an airport railway connection to work, it must be in the interests of a wide group of stakeholders—including the public, the airport, the airline, and the railroad company. This section of Chapter 3 reviews the role of various players, and then focuses primarily on the role of the long-distance rail company in examining the business case for stopping its long-distance trains at the airport station. The Public It is unclear exactly who represents “the public” in their dealing with the airports, the airlines, and the rail compa- nies (each of whom have business cases to justify). However, it is clear that public, civic concerns do exist; in some cases, these issues are clearly defined (e.g., European Union public policy) and in some cases they are not. Connecting railways to airports could satisfy a number of social objectives, includ- ing reduction of pollution and CO2 emissions, reduction of highway and airway congestion, improvement of safety, and focusing use of airport capacity on longer-haul services rather than making short, intercity hops, among others. In principle, the European Commission has decided that the public investment needed to improve rail/airport connec- tions will be justified by these benefits. The Commission’s clear emphasis on HSR airport connections indicates that the objective will include shifting air traffic to rail for short-haul trips and improving local connections. In practice, this suggests that much of the existing air traffic between major airports in France, Belgium, the United Kingdom, the Netherlands, and Germany might eventually shift to center-city rail—a transition that was reported in Chapter 2. The Airports Reliable, local rail connections could be in the interest of airports, subject to there being enough volume to justify the airport’s share (if any) of the cost of the airport railway station, and subject to the impact of loss of parking rev- enues, if any. It may not be in the interest of that airport to see improved rail service to other airports if the airports are competing for the same traffic. On a local level, for example, DCA might lose passengers to IAD when the new Metro line is completed (though this might well be in the overall public interest). Connecting HSR to an airport extends the airport’s catchment zone. As a result, a traveler through a major hub (international or domestic) can choose an air connection to his or her eventual destination or take rail instead. A major hub with good HSR service might benefit from a larger ser- vice area and an improved competitive position vis-à-vis a

47 Modal split models are then used to allocate demand among the available modes, based on the relative cost (line haul plus access/egress costs), trip time (line haul, average waiting times that are determined by service frequency, schedule reliability, and access/egress), and a number of other factors (perceived safety and comfort). Modal splits are also commonly segre- gated by trip purpose—business versus pleasure. The effect of the modal split variables can be significant. For example, business travelers attach a much higher weight to the value of their time (and comfort) and less to cost than do nonbusiness passengers. As with the total demand estimates, modal split projections are specific to particular markets and, if experi- ence is any guide, subject to a great deal more uncertainty than developers would like to acknowledge. As a consequence, the rail planner considering the addi- tion of an intercity stop (HSR or otherwise) faces a decision in which the ratio of qualitative to quantitative measurement can vary widely. Generally, the measurement of costs (invest- ment and operating) is somewhat more precise (though still subject to error) than the ability to calculate benefits, so con- servatism will be natural. This is especially true for rail carriers that will not receive any explicit public support for the new service. These carriers must predict some level of financial (as opposed to economic) surplus from the service added; this must be done while the effect on demand is question- able and while the rail carrier’s costing information is aligned more to accounting for public support than it is to measur- ing financial profitability. Where no demand forecasting tools exist, the resulting decision is based on judgment and tends towards the negative if financial performance matters—and towards the political where it does not. In a few cases, such as the new connection at the PVD airport station, Amtrak would have the option of an experi- ment. Since the station is already built, Amtrak could add an Acela or Regional stop at PVD and measure the effect on demand and cost. This would add a data point that does not now exist, and it might permit a more confident assessment of effect on demand in other cases. Unfortunately, addition of one stop might not make much difference, because there are so few services offered. In addition, the PVD airport station does not have electrified tracks connecting to the high-speed main line, so adding an effective high-speed connection would require significant investment from either Amtrak or the airport authority before the service could be added (cur- rent service by MBTA is provided by diesel traction). In some cases on the NEC, the rail planner would have both approximate existing traffic data and a demand model cali- brated on the available dataset. The Research Team cannot rep- licate their exact information since Amtrak has not historically released detailed market flows by O-D and by passenger class. An approximate set of O-D flows was, however, constructed connectivity). In this case, the rail operator would be adding a new market (air travelers that will complete their trip by rail rather than by air), but doing so at the cost of increasing the trip time for existing rail passengers. This could then pose a trade-off for the rail operator between the passengers gained by the airport link versus those potentially lost because of longer trip times for existing passengers and the added costs of establishing the station, along with the added operating costs of maintaining the station and of stopping trains at it. There are no completely clean examples of this issue because all connections are a blend of local and intercity con- nectivity. One example might be the new airport connection at T.F. Green Airport in Providence, RI (PVD). If Amtrak’s Acela trains stopped at the PVD airport station, they would clearly increase the trip times between other markets on the line [including New York Penn Station (NYP), New Haven (NHV), and New London (NLC) to Route 128 (RTE), Back Bay (BBY) and South Station (BOS)], thus exposing those pas- sengers to increased competition from air between New York and Boston and to diversion to road. On the other hand, a direct stop at PVD could at least theoretically increase rail traffic by diverting some longer-haul air traffic from the con- gested Logan Airport to PVD, with an excellent rail link into Boston, or from LaGuardia (LGA) with an excellent rail link into NHV. Other examples might involve travelers going through BWI rather than PHL to Wilmington, which would expose the rail market south of NYP to the added stopping time at BWI. Travelers could go through EWR rather than LGA or JFK on the way to PHL or NHV (or, indeed, to many areas in the New York metropolitan area that are poorly served by congested access to LGA and JFK), but this would again expose some of the travel market south of NYP to stopping delays. In more general terms, total travel demand between two markets is a function of the populations of the markets, the incomes (and values of time) of the population, and the trip time between the markets. Each of these factors can only approximate reality. For example, the total population of an area can be spatially distributed in a number of ways that will affect the trip generation of the market. Also, income distri- butions can be skewed in ways that will affect total demand and modal preference; trip times as perceived by the pas- senger include not only the line haul time of the trip, but also access/egress times, waiting time (trip frequency), and the probability and severity of delays. Moreover, acquiring actual data on travel can be difficult because some of the most important data (auto travel) are often not collected, while other data (intercity air and rail passengers) are subject to various levels of confidentiality. The net result is a model of total demand that must be calibrated to specific O-D pairs and can give only approximate results subject to a significant range of variation.

48 Similarly, the 460,000 Regional passengers would repre- sent revenues of about $48 million. This figure assumes that Business Class passengers are about 10% of the total and would yield average revenue of about $0.46/passenger-mile. This is compared with AC of $0.16/passenger-mile and FAC of $0.32/passenger-mile, thus losing Amtrak a contribution of $67.50/passenger over AC and $31.50/passenger over FAC for every passenger lost if the additional stop causes an adverse shift in market share. The current trip time from NYC to BOS is 3 hours and 40 minutes (220 minutes) for Acela. For Regional service, it is about 4 hours and 30 minutes (270 minutes). An extra stop at the Providence airport would add about 3 minutes, or about 1.4% to the Acela schedule and 1.1% to the Regional schedule. The time elasticities for these services are not available, but assuming that the loss is the same as the percentage increase in schedule time, this would mean 8,000 Acela passengers/ year and 5,000 regional passengers/year. In total dollar terms, this would be around $500,000/year over FAC and nearly $1,000,000/year over AC. These losses do not include losses in smaller markets also affected (NHV to BOS, or NLC to BOS). A more accurate calculation would look at these markets and might add 10–20% to the above numbers. The increase in passengers arriving at PVD would then be examined, with the primary question being where they might go and how the added traffic would affect the inter- city traffic. Table 3-3 summarizes the revenues and related costs for both the added traffic that the connection might generate and the traffic losses that might result from the lengthened schedule from NYC to BOS. It is clear that an added passenger from the airport stop will not balance a lost passenger from the NYC/BOS market, even given that the costs in the Providence to Boston traffic are almost certainly understated. In fact, the outcome would be dependent on the cost to Amtrak of using the station and on whether the diverted passengers used Acela service or Regional trains. Since the trip time of Regional service to BOS is essentially the same as Acela, it seems likely that most diversions, if they occurred, would be to Regional trains rather than Acela. With this said, the Providence airport has around 2.1 million boardings annually: if only 1% of the passengers were going to/from Boston and could be diverted to rail (from road), there could be a rough balance of losses and gains, assuming that the existing MBTA service would not continue to carry a significant share of the traffic. Only an experiment would answer this question. Intercity Service to the Newark Airport (EWR) Calculations of the effect of adding station stops at EWR are more complex because there are more markets potentially by class based on data that Amtrak has released over the years in its Monthly Performance Reports (MPR from Amtrak web- site) and a judgment as to the relative populations involved—a “stylized” set of O-D flows. The base data table is meant only to illustrate the analysis in a reasonable way, and was not based on recent, actual information. Actual Amtrak tariffs and sched- uled trip times taken from the Amtrak website have been added to this information, along with approximate Acela and NEC Regional cost information taken from the MPRs. Airline data are likewise taken from a combination of the T-100 dataset along with scheduled flight times from airline websites and from the Kayak.com website. Again, these data only reflect a single point in time, both for demand and for schedules, and say nothing about individual airline flows. Given the complexity and volatility of yield management pricing and the multiplicity of aircraft types in use depending on volumes and length of haul, it is not possible to character- ize airline fares in detail. Two examples represent the planner’s challenges: addition of an intercity stop at Providence airport and addition of more trains stopping at EWR. Intercity Service to the Providence Airport (PVD) For simplicity, one can focus on the potentially adverse impact on the NYC to BOS market (other markets would be handled similarly) where there are approximately 570,000 Acela passengers and 460,000 Regional passengers annu- ally as of 2007 (when the air data were available). First Class Acela fares north of NYC are $108/passenger plus $0.37/mile (based on a regression of data from the Amtrak website) and Acela Business Class fares north of NYC are $67/passenger plus $0.22/mile. Regional Business Class fares north of NYC are $33/passenger plus $0.44/mile and Regional Coach Fares are $24/passenger plus $0.34/mile (Thompson 2011). The ratio of First Class seats to total seats on Acela trains is about 15 percent, so 570,000 passengers would be around 85,000 First Class and 485,000 Business Class (assuming a passen- ger balance close to that of the seating). This would produce Acela revenue of about $73 million and an average revenue/ passenger-mile of about $0.57. In past years, Amtrak has published an estimate of avoidable costs (AC) and Fully Allo- cated Costs (FAC) for each route (e.g., the Amtrak MPR for September 2008, pg. C-1) that calculated that, on average, the AC/passenger-mile for all Acela passengers is about $0.22 and the FAC/passenger-mile is about $0.42. This means that each Acela passenger in the NYC to BOS market is yielding around $0.15/passenger-mile over FAC, and $0.35/passenger-mile over AC: at an average distance of 225 miles, this would be between $33.75 and $78.75 of contribution lost for every passenger that is lost if the additional stop causes any shift in market share.

49 market would be reduced by as much as the percentage trip time increase, which could cost Amtrak a contribution of $6 million over FAC, and $9 million over AC. This loss could go up if the ridership loss is greater than the percentage increase in trip time, which could happen on the longer trips affected, especially Washington to NYP. The gains to Amtrak are equally hard to estimate. Poten- tial gains can be approximated by assuming that the traffic gained would have the same distribution of O-Ds as existing traffic south of NYP, and that the distribution of Acela ver- sus Regional passengers would also remain the same. Under these assumptions, added demand at EWR would have to be around 110,000 passengers annually to make up for the $9 million AC and $6 million FAC contributions estimated above (see Table 3-4). This would make the Amtrak EWR sta- tion one of Amtrak’s larger stations, but would only consti- tute about 0.6% of the roughly 17 million airline passengers boarding annually at EWR. As was the case with Providence, this does not seem to be an implausibly large percentage, especially given the ease of the air/rail connection at New- ark, but it would in effect put Amtrak in competition with NJT, which might split the total rail traffic to the detriment of both. in play. A station stop at EWR would add 3 minutes to the schedule of essentially all passengers to and from all points south of NYP to NY Penn, the most important single Amtrak market. In addition, Amtrak tariffs are structured differently south of NYP than north, with southern fares cheaper for short trips and more expensive for long trips, while northern fares are more expensive for shorter trips and less expensive for longer trips (Thompson 2011). Table 3-4 shows the gains per passenger that Amtrak might realize if the added station stop at EWR were to add a pas- senger to or from WAS, BAL, and PHL. Table 3-5 shows the Amtrak passengers, revenue, and contribution that would be at risk by lengthening the trip times between NYP and WAS and BAL and PHL, along with the percentage impact on trip times of an added 3 minute delay added by a stop. It is difficult to say with any confidence whether the gains would balance the losses. From one point of view, since all of the markets are well within the 3.5 hour “barrier,” there might not be any significant demand losses, so that any pas- sengers gained would be worthwhile (of course, this would be true of any added stop within the 3.5 hour limit, which would be a reduction ad absurdum that few planners would accept). At the other limit, it is possible that demand in each Gains to/from Providence airport and BOS ($/passenger) Average Revenue Avoidable Costs (AC) Total Attributed Costs (TAC) Margin over AC Margin over TAC Acela 1st Class 124.28 9.68 18.48 114.60 105.80 Acela Business 76.68 9.68 18.48 67.00 58.20 Average 83.82 9.68 18.48 74.14 65.34 Regional Business 52.36 7.04 14.08 45.32 38.28 Regional Coach 38.96 7.04 14.08 31.92 24.88 Average 40.30 7.04 14.08 33.26 26.22 Losses to/from BOS to NYC Average Revenue Avoidable Costs (AC) Total Attributed Costs (TAC) Margin over AC Margin over TAC Acela 1st Class 191.25 49.50 94.50 141.75 96.75 Acela Business 116.50 49.50 94.50 67.00 22.00 Average 127.71 49.50 94.50 78.21 33.21 Regional Business 132.00 36.00 72.00 96.00 60.00 Regional Coach 100.50 36.00 72.00 64.50 28.50 Average 103.65 7.04 14.08 96.61 89.57 Table 3-3. Gains and losses from an additional stop at T. F. Green Airport.

50 WAS ACELA WAS Regional Distance 212 Distance 212 Revenue ($) 197.5 Revenue ($) 110.2 Avoidable Cost ($) 46.6 Avoidable Cost ($) 33.9 Total Attributed Costs ($) 89.0 Total Attributed Costs ($) 67.8 Margin over AC ($) 150.9 Margin over AC ($) 76.2 Margin over TAC ($) 108.5 Margin over TAC ($) 42.3 BAL ACELA BAL Regional Distance 182 Distance 182 Revenue ($) 173.5 Revenue ($) 95.6 Avoidable Cost ($) 40.0 Avoidable Cost ($) 29.1 Total Attributed Costs ($) 76.4 Total Attributed Costs ($) 58.2 Margin over AC ($) 133.4 Margin over AC ($) 66.5 Margin over TAC ($) 97.0 Margin over TAC ($) 37.4 WIL ACELA WIL Regional Distance 103 Distance 103 Revenue ($) 110.0 Revenue ($) 57.4 Avoidable Cost ($) 22.7 Avoidable Cost ($) 16.5 Total Attributed Costs ($) 43.3 Total Attributed Costs ($) 33.0 Margin over AC ($) 87.4 Margin over AC ($) 40.9 Margin over TAC ($) 66.8 Margin over TAC ($) 24.5 PHL ACELA PHL Regional Distance 78 Distance 78 Revenue ($) 90.0 Revenue ($) 45.3 Avoidable Cost ($) 17.2 Avoidable Cost ($) 12.5 Total Attributed Costs ($) 32.8 Total Attributed Costs ($) 25.0 Margin over AC ($) 72.8 Margin over AC ($) 32.8 Margin over TAC ($) 57.2 Margin over TAC ($) 20.4 AVERAGE ACELA AVERAGE Regional Distance 167 Distance 167 Revenue ($) 161.4 Revenue ($) 88.4 Avoidable Cost ($) 36.7 Avoidable Cost ($) 26.7 Total Attributed Costs ($) 70.1 Total Attributed Costs ($) 53.4 Margin over AC ($) 124.7 Margin over AC ($) 61.7 Margin over TAC ($) 91.3 Margin over TAC ($) 35.0 Average for both Acela (32.3%) and Regional (67.7%) Margin over AC 82.0 Margin over TAC 53.1 Extra passengers to cover AC losses 109,980 Extra passengers to cover TAC losses 113,078 Table 3-4. Potential gains per passenger diverted to rail at Newark Airport.

51 distance rail riders to the airport. The parallels with France’s Lyon Airport rail station are striking: while the architecture of the connection is unique in its grandeur, most of the trains passing through the station simply do not stop, consistent with the experience in Providence. While the question of the logic of the rail operator was explored in Part Two, the characteristics of the physical site planning considerations are reviewed here. Three Options for Connections to North American Airports Planners in the North American experience have examined many approaches to the question of how to best get rail access to airports. This Chapter focuses on three major approaches to this question, which include: (1) full integration; (2) con- nection by shuttle to nearby air terminals; and (3) connections Part Three: Typology of Options to Connect with Long-Distance Rail Planning Considerations from the Research Consistent with the European experience, market analysis of the five airports reviewed in this Chapter reinforces the obser- vation made in Chapter 2: the relationship between the airport and its catchment area primarily explains the market share gained, rather than the details of physical inter connection. Clearly, sheer proximity is not the answer. Chapter 3 shows that the two United States rail stations closest to the major air passenger check-in terminals are Providence airport and Burbank. In Providence airport, a high-quality moving side- walk equipped “skybridge” makes the 1,300-foot connec- tion. In Burbank, the same distance can be traversed on foot. Neither is currently serving a meaningful number of long- Table 3-5. Passengers, revenue, and contribution potentially lost by stopping at Newark Airport. ACELA WAS BAL WIL PHL Total Passengers (000) 650 265 175 475 1,575 Distance 225 195 116 91 Percent time increase 1.8 2.3 3.9 4.1 Potential Passengers Lost 11.7 6.1 6.8 19.5 44.1 Revenue ($000) 2,433 1,121 822 1,956 6,332.1 Avoidable Cost ($000) 579 261 174 390 1,404.7 Total Attributed Costs ($000) 1,106 499 333 744 2,681.7 Margin over AC ($000) 1,854 859 648 1,566 4,927.4 Margin over TAC ($000) 1,328 622 490 1,211 3,650.4 Regional WAS BAL WIL PHL Total Passengers (000) 1,260 580 275 1,165 3,280 Distance 225 195 116 91 Percent time increase 1.5 1.8 3.3 3.3 Potential Passengers Lost 18.9 10.4 9.1 38.4 76.9 Revenue ($000) 2,201 1,064 578 1,984 5,827.8 Avoidable Cost ($000) 680 326 168 560 1,734.3 Total Attributed Costs ($000) 1,361 651 337 1,120 3,468.6 Margin over AC ($000) 1,521 739 410 1,425 4,093.5 Margin over TAC ($000) 840 413 241 865 2,359.2 Total TOTAL GAIN OVER AC 3,375 1,598 1,058 2,990 9,021 TOTAL GAIN OVER TAC 2,168 1,034 731 2,076 6,010

52 reached in Newark and BWI, in which some, but not all, long- distance services use the station. This site planning option is now under consideration for a long-distance rail station along the NEC alignment proxi- mate to Philadelphia International Airport, as shown in the Amtrak NextGen HSR proposal. Questions of just how many Amtrak trains would option to use the additional stop were not resolved in that conceptual plan. In virtually any rail operations plan, the traveler who chose to transfer at this sta- tion, as opposed to the existing 30th Street station, would be offered far fewer trains, and thus longer waiting times. Site Planning Concept #3: Connect to Network at a Central Place: Connect with the Best Consolidated Rail Transfer Point Possible A review of the possible options for the future of rail ser- vice to Newark Liberty International Airport highlights the possibility of not having an airport rail stop at all. Instead, a connection with a major, existing intercity rail terminal could be created. As will be documented later in this report, serious consideration to schemes that do not include on-airport or near-airport long-distance rail stations is being given in San Diego and Chicago, in addition to Newark. In this manner, the number of trains that will in fact stop at the designated airport connector is maximized, with potential interface with other intermodal connections existing at the main rail station. In Europe, cities in which long-distance access schemes do not include an airport rail station include London, Madrid, Barcelona, Munich, Vienna, Hamburg, and many others with metropolitan rail connections only. Conclusion Air/rail connectivity is often presented as a design issue. The empirical observations from Chapters 2 and 3 suggest that a successful air/rail transfer could be first examined from a point of the view of the markets for the services offered. When a given airport offers some kind of unique service, whether differentiated by directness of service (hubbing) or by price (the “Southwest Effect”), passengers can be expected to arrive from originating distances beyond the airport’s logi- cal catchment area. When a less unique airport offers essen- tially the same set of airline connections as those airports surrounding it, the logic for taking a long train ride to access a similar set of aviation options is weak. At face value, this observation might be seen as a positive for rail feeder links at airports in Newark, Philadelphia, Chicago, and JFK. For example, each of those airports offers some kind of direct air service to long-distance destinations not offered by competing airports. A trip optimized from the viewpoint of the passenger might include the selection of a check-in with regional systems, which may not be close to the airport. Thus, the three categories of connection type developed in Chapter 2 for European airports are reviewed here for their relevance. Site Planning Concept #1: Full Integration at Airport: Reroute the Long-distance Rail Line to Go to a Point from Which the Air Traveler Can Walk a Short Distance from the Train to the Check-in Terminal To date, no North American airport has routed its longer distance rail tracks into the terminal complex itself. This is the design concept first accomplished in Zurich, which was the first long-distance trunk line to be placed in a subterranean level of the airport terminal, followed by Amsterdam Schipol, Frankfurt, and Copenhagen. In the optimal case, all distribu- tion between the rail station and the passenger processing areas is by foot, as was the case in the single terminal configu- ration at London Gatwick. Perhaps importantly, a review of the North American experience reveals no airport that has succeeded in providing long-distance, intercity rail directly into the air passenger terminal complex. Design visions of such a solution could be gleaned by looking at success- ful examples of local, metropolitan on-airport rail stations, including Reagan–National, Atlanta, SFO, Portland, St. Louis, Chicago, and Cleveland, among others. Planning considerations for Site Planning Concept #1 are dominated by costs and the horizontal and vertical design requirements of the rail alignment. Indeed, when examining the option of full architectural integration, the more com- mon approach is “bring the airport to the rail” by moving the air passenger terminal, rather than re-routing the rail. This is clearly under discussion in San Diego (see Chapter 6), and is one of a family of options that must be considered at Newark Airport if the plans for a third parallel runway are pursued. Another variation on the “bring the airport to the rail” theme would occur if the City of Chicago revived the Terminal 7 concept as discussed in Chapter 6 of this report. Site Planning Concept #2: Shuttle to Nearby Rail Alignment: Build a Separate Airport Rail Station on the Rail Alignment at a Point As Close As Possible to the Airport Passenger Terminal Complex The present rail station at Newark Liberty International Airport is a good example of placing a new station on an existing rail alignment, from which some form of connec- tion is built to connect with the airline terminal complex. The long-distance railways did agree with this solution in Bur- bank and Milwaukee, but the long-distance railway did not agree with this solution in Providence. Compromises were

53 Daykin, T. 2009 (June 23). “Republic Airways to Buy Midwest Airlines.” JSOnline. Destination: Freedom. 2004 (January 19). “Midwest Airlines Likes Amtrak Connection.” Vol. 5, No 3. Destination: Freedom. 2005 (January 24). “Amtrak Opens Newest Air- port Station.” Vol. 6, No. 4. European Commission. 2011 (March 28). “White Paper—Roadmap to a Single European Transport Area—Towards a Competitive and Resource Efficient Transport System.” COM 144 Final, Brussels. Howell, J. 2008 (January 31). “Warwick Station Hinges on Amtrak Deal.” http://www.warwickonline.com/. Great American Stations. 2011. http://www.greatamericanstations. com/Stations/MKE. (As of Spring 2011). Mitchell Airport. 2005. http://www.mitchellairport.com/news.html. “Governor Doyle, Senator Kohl Dedicate New Passenger Rail Sta- tion At GMIA.” (As of Spring 2011). Mitchell Airport. 2009. Accessed at http://www.mitchellairport.com/ blog/archives/2009/11 (As of Spring 2011). I-95. 2004. Accessed at http://www.i95coalition.org/i95/Portals/0/ Public_Files/pm/reports/summary269.pdf. (As of Spring 2011). I-95. 2004a. Accessed at http://www.i95coalition.org/i95/Portals/0/ Public_Files/pm/reports/full295.pdf. (As of Spring 2011). Jacobs Consultancy. 2010 (December 9). “San Diego Regional Airport Systems Plan, Presentation #2 Preliminary Findings—Remaining Scenarios.” MWCOG and National Capital Region Transportation Planning Board. 2010. “2009 Washington-Baltimore Regional Passenger Survey.” Washington, DC. Martin, P. 2011 (March 22). Interview by Research Team, Chicago, Illinois. Orangeline Development Authority. 2011. Accessed at http://www.eco. rapid.org/Project/description.shtml. SNCF. 2009 (September 14). “Study of California Corridor.” Submitted to US Department of Transportation, Federal Railroad Adminis- tration, Washington, DC. Sperry, B. and Morgan, C. (2011). “Intercity Passenger Rail: Implications for Urban, Regional, and National Mobility” University Transpor- tation Center for Mobility™ Texas Transportation Institute, The Texas A&M University System, College Station, TX. T-100 Data. 2007. See http://www.surveycafe.com/acrp310/Select Airport.asp. Thompson, L. S. 2006 (January). “Options for Federal Ownership of the Northeast Corridor (NEC) Infrastructure.” Federal Railroad Administration, Washington, DC. Thompson, L. S. 2008 (December). “Railway Access Charges in the EU: Current Status and Developments Since 2004.” International Transport Forum. Thompson, L. S. Thompson Galenson and Associates, 2011. “High- Speed Rail (HSR) Interactions with Air Travel” for ACRP Project 3-23, available from ACRP. Trains. 2006 (June 1). “Chicago ‘Jet-setters’ Prefer Milwaukee and Its Amtrak Connection.” Wall Street Journal. 2014. “Port Authority Funds PATH Link to Newark Airport.” Accessed at http://online.wsj.com/news/articles/SB10001 424052702304626804579363013633022416. Wikipedia. 2011. http://en.wikipedia.org/wiki/Pennsylvania_Station_ (Newark)#Current_operations. Accessed December 2012. Zupan, J., R. Barone and M. Lee, Regional Plan Association. 2011. “Upgrading to World Class: The Future of the New York Region’s Airports,” New York. airport outside of his/her metropolitan area, implying that the market for rail access might be seen as a set of longer- than-usual ground access links, rather than the replacement of existing feeder flights from more distant origins. Thus, the market package needs to make sense to a wide vari- ety of stakeholders in the process; the case study of Midwest Airlines in Milwaukee was, for a time, a good example of where good rail service might improve the business case for all the parties. When a new airline management was less focused on the Milwaukee hub, efforts to create better coordination with Amtrak came to a close. And with this, the ability to deal with the serious structural challenges of serving an airport with six trains a day was weakened. Arguably, the need is to get the air traveler directly to that point on the rail network where the connecting options are best. By way of example, it may make more sense to carry Heathrow connecting rail passengers to a point where all HS2 trains are going to stop, than to reroute a portion of rail line haul services into a transfer point more proximate to the airport, at which some form of transfer would still be needed to get to the decentralized terminals. In a parallel manner, if the funds could be found to connect all Newark Liberty air passenger terminals to Newark Penn Station, it might make sense for Amtrak to abandon the “airport station” stop entirely, encouraging a transfer at the major point of system connectivity. The case has been made in this Chapter that it is not in the interest of operators of HSR services to incur the added travel time involved in stopping at airport stations that generate relatively few additional passengers. The traveler will base his/her choice of service on a ser- vice’s door-to-door times and costs. Unless the impediment of transfer time (reflecting poor schedule coordination) is minimized, the empirical evidence suggests that efforts to integrate ticketing, or even baggage handling, will be of little comparative effect. Bibliography AirRailNews. 2011. http://www.airrailnews.com/index.php/component/ simplelists/item/673/component/simplelists/. Allison, M. 2004 (January 13). “Midwest Airlines Hails Connection to Amtrak.” Chicago Tribune. Amtrak. MPR. “Monthly Performance Report.” This report is issued monthly, with the September issue including the full fiscal year data. www.amtrak.com. Amtrak. 2010 (September). “A Vision for High-Speed Rail in the Northeast Corridor.” Washington, DC. Amtrak. 2011. www.amtrak.com. Accessed Spring 2011. At present, station summary information is provided at http://www.great americanstations.com/Stations/. Bob Hope Airport. 2012. http://www.bobhopeairport.com/home/ news/initiatives.html.