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.

142 C H A P T E R 1 0 Introduction and Structure Introduction: Good Practices and Data Gaps In the process of undertaking the case studies in the earlier phases of the project, the Research Team explored many data sources and interviewed several practitioners associated with the integration of air and rail planning in the United States and Europe. This Chapter presents a summary review of the status, quality, and availability of essential data; it summa- rizes key positions of those interviewed about data and tools, and presents a description of the key gaps in that data—some of which will be dealt with in the concluding chapters of this report, and some of which will remain significantly beyond the scope of this research. Structure First, Chapter 10 outlines a series of general observations about the attitudes and reactions of practitioners inter- viewed. Second, a brief review of selected sources of data in Europe is presented, emphasizing the range of their use for multimodal and intermodal analysis in the planning pro- cess. Third, similar sources in the United States are briefly reviewed, revealing the extent to which data types are and are not made available to the American analyst/planner in a manner similar to that observed in the European context. Fourth, this section reviews the substantive gaps that seem to exist, despite the considerable efforts summarized in the earlier sections of the report. This is followed by a summary table, which takes the form of a “checklist” of available data and methods, and their present status in existing and pro- posed research programs. This final section is a prelude to the presentation in Chapter 11 of the new modeling activity undertaken in the project to address some of the largest gaps in the tools available to practitioners examining combina- tions of air and rail. Differing Views of the Urgency of the Problem of Lack of Intermodal Tools In undertaking the case studies for this report, multiple experiences were found in identifying the quality of the data, the availability of tools, and the gaps revealed in the process. In summary, there was a major difference between the ori- entations of day-to-day practitioners in the field, and those of professionals vested with the responsibility to understand the public policy implications of decisions and policies. In general, those practitioners interviewed who were involved in operational management decisions had found “workarounds” to deal with the reality that multimodal data is usually not available or not shared across jurisdictional boundaries. At the Amsterdam Airport, the Research Team was told, “They run their operations, we run ours.” In Chicago it was found that, in absence of knowledge of how and if HSR would come to Chicago O’Hare International Airport, the manage- ment had developed a workable location for a new rail station that might, or might not, be developed. In short, airport plan- ners there had figured out how to manage the contingency that rail policies might change, without endangering their need to continue with the planning of the airport. In interviews in San Diego, there was an acknowledgment that more data would be produced in the ongoing regional studies about the potential for rail to alter plans for the airport reconfiguration, but the parties agreed that the scale of inter-relationship was not strong enough to impede progress on making key capital investment decisions. Importantly, in all of the interviews, no practitioners stated that the lack of optimally integrated data was keep- ing them from making the immediate decisions they need to make. Managers involved in market research for Amtrak have undertaken market research activities which support the immediate need for corporate management decisions, with or without optimal integration with data stemming from Analytical Tools and Data Sources for Policy Planning

143 aviation planning. The managers of the private airlines have access to propriety programs that help guide their investment and service planning functions—without the detailed infor- mation about rail available to Amtrak managers. All of these are examples of how the managers in each mode have carried out their fiduciary responsibilities, without waiting for a next generation of improved intermodal data. At the same time, those interviewed agreed that better data and methods—if they were available—would improve decisions. However, this must be sharply contrasted with the views of a smaller number of researchers and planners who were tasked with preparing truly multimodal and intermodal analyses. By far, the strongest feeling of frustration encountered was expressed by the authors of the path-breaking multimodal and intermodal study, “Upgrading to World Class: The Future of the New York Region’s Airports” for the Port Authority of New York and New Jersey (Zupan et al. 2011). In the interview, the authors expressed concern about their ability to undertake good market analyses about the 100 to 400 mile trip with no knowledge of the number of vehicle trips in the corridor, and no publicly available data on the corridor’s long-distance rail volumes. The report that was created, however, must be seen as a model of how solid analysis may be undertaken in a cli- mate of less-than-perfect data. A major concern expressed to us by the authors of the “Upgrad- ing to World Class” study of the Regional Plan Association was the lack of a publicly available, citable data source of long-distance rail in the corridor; in addition, concern was raised over the absence of documentable vehicle flow data. Similarly, those involved in regional and system-wide multi- modal analyses felt a far greater sense of urgency about the need to improve the quality of both the data and the tools available for integrated multimodal and intermodal analyses. Thus, those involved in the formulation of good public policy in northern California, for example, explained to us the urgency of improving the quality of basic data and the need to make that data available for all stakeholders in the political process. Senior agency officials at the public agency that managed the creation of the model were concerned that the major demand analysis resource available to support policy decisions about air and rail in California requires days of computation time to analyze a single scenario. Importantly, within the modally based organizations, there was a core of transportation professionals who understood that new demands were about to be made on the transporta- tion planning process, and that those demands would ulti- mately require the improvement of the tools and data sources needed for true multimodal and intermodal analyses involv- ing both air and rail together. The exceptionally broad multi- modal emphasis of the “Upgrading to World Class” report was developed in the aviation department of a public agency. The need to explore the possible role of rail in a feeder mode has been advocated (among other innovative strategies) at the highest management level at the San Francisco Inter- national Airport; the lack of evident consideration by rail officials of an on-airport HSR station at SFO is correlated with generally weak methods to analyze the potential for the role of rail as a feeder mode to air. The commencement of major environmental impact analyses on both east and west coastal areas has made rail decision makers aware of the need for commonly shared data, available to all stakeholders in the processes mandated for several decades by the National Environmental Policy Act (NEPA) of 1969. In short, practi- tioners who place high value on the creation of better tools for intermodal analysis were found; also, there needs to be more public availability of key data to support the application of analysis tools. Other Concerns from the Project Interviews Most of the interviewees did not focus on the quality of transportation modeling as a major concern. In one of the case studies, the strongest statement about the quality of ridership forecasts was that two proposing organizations had created two separate demand forecasts, and that this was a political negative for the advocates of HSR. Anecdotally, the quality of either of the two forecasts was not cited as a concern. But, while most of those interviewed did not focus on the quality of the models per se (interviews in northern California and New York being the clear exception), there are still anecdotal reports about problems with the results of the process in general. One of the most important of these cases was in San Diego, where the question concerned the variation in HSR ridership in response to variation in the location of the southernmost terminal in the project. This was a rare case where the output of the planning process varied between studies undertaken contemporaneously. For those working with the SANDAG- based planning processes, terminating the rail line at a major intermodal transfer center adjacent to the airport was deter- mined to be the preferred alternative. The airport-based RASP, on the other hand, suggested that more rail riders would be generated by a more central downtown location, by the pres- ent Santa Fe terminal—it also suggested that diversion from air to rail would be more pronounced with the station located away from the airport. This level of difference implies there is a need to improve the quality of analysis tools applied at this geographic micro-level. This all suggests an underlying skepticism about the qual- ity of the rail demand forecasting process, whether or not it was phrased in these terms during the interviews. On the one hand, it could simply be posited that a statewide demand modeling process might not be the best method to resolve a matter of station location in a scale of under 10,000 feet. The

144 lishing prototypical applications that could be applied on a multi-national basis. But European decision makers have no plans to undertake any continent-wide survey on the scale of the earlier American Travel Survey (Bureau of Transporta- tion Statistics 1995). In short, this section of the report will provide some anecdotal examples of multimodal data collec- tion and data sharing; it does not conclude that fundamental issues concerning the cost of collecting highway vehicle flow data have been solved in Europe. Multimodal Demand Forecasts by the UIC (International Union of Railways) A brief review of the available literature on multimodal demand in Europe shows that much of the case-by-case (corridor-by-corridor) analysis was largely based on one seminal study. As discussed in Chapter 2, the study was com- missioned by the UIC (International Union of Railways) and provides a corridor-by-corridor prediction of market share between air, rail, and auto. The document, titled “Passenger Traffic Study 2010/2020” (Intraplan et al. 2003) provides base data that set a multimodal context for understanding the role of rail and the role of aviation in Europe (Figure 10-1). The data can be purchased from the UIC in DVD format at a modest cost (under $200). There is currently no comparable document available about intermodal demand in the United States; recent work by the FHWA is noted herein. A major multi-disciplinary effort like this helps to establish the base case against which a wide variety of policy scenarios can be tested, and it provides a sense of scale to the question of the impact of HSR on other modes. credibility of the demand modeling process is challenged by a lack of understanding of the importance of key variables (e.g., access time by car, access time by bus, access time by taxi, etc.) in both the decision to forgo air for rail in a total trip (e.g., competitive mode) and the decision to forgo air for rail as a first segment of a trip that later includes long- distance air (complementary mode). The reader is referred to a more general discussion of the adequacy of the air- port access modeling process in ACRP Synthesis 5: Airport Ground Access Mode Choice Models. That report includes a discussion of the application of various modeling processes (airport specific vs. general) in the regional planning pro- cess, but the scope of the Synthesis study did not address the issue of substitution of long-distance rail travel for air travel, either for an entire trip or in place of a short-haul feeder segment. Multimodal Planning Documents and Data Sources in Europe In the ACRP project research in Europe, several good examples of data analysis that were truly multi-modal in nature were found, which result in publicly available data to support public policy debates. However, in almost all cases, documents created in Europe suffered from the same lack of auto mobile flow data as in the case of the United States. In Europe, there is no single strategy for attaining multi-coun- try highway vehicle trip table data. The innovative project “KITE” (Knowledge Base for Intermodal Transportation in Europe) supported early studies in Switzerland, Portugal, and the Czech Republic (Axhausen 2010) with the hope of estab- Figure 10-1. Multi-modal demand for long-distance travel in Europe is documented in a major study for the International Union of Railways (UIC), and is readily available to the public. Source: Intraplan Consult GmbH et al. 2003.

145 sioning its innovative study, Air and Rail Competition and Complementarity (SDG 2006). A key aspect of this EU study was the creation of a model of air/rail mode share that was purposefully simple and transparent; its authors note that “the ultimate objective of the model is to be able to test sce- narios for the development of short-haul transport over the next ten years” (SDG 2006). Figure 10-2 illustrates how the results of this innovative study were shared with the public, taken directly from a PowerPoint presentation by the project manager (Rizzo 2007). As discussed in Chapter 4, the model has four elements: time-related factors; price; access time and cost; and a set of more subjective factors. The model allows the forecasting of market share for eight corridors in 2011 and 2016, and the calibration year of 2005. • Reviewing the content of Chapter 4, it can be observed that the creation of a location-specific model (i.e., calibrated with the data from eight European corridors) allows the analysis of a wide variety of public policy scenarios. This was accomplished by overcoming the fact that several major railways did not vol- unteer to share their base data, and the analysts were forced to work around this condition. The result is a publicly available data source based on an understandable and largely transpar- ent methodology, which can be used to support a public dialog and debate. French Studies of Intermodal Connections Of particular merit here are the series of studies undertaken by the “Executive Air Transport” of the French Directorate General of Civil Aviation (DGAC), which has conducted a survey every 3 years to better understand the relationship of HSR to longer distance aviation at Charles de Gaulle and Lyon The effort to build one model of long-distance travel behavior by all modes was accomplished by merging fore- casting techniques of several major research organizations. The demand modeling process based trip generation on such factors as “GDP, population, employment development, car ownership, market regulations, user cost, transport policies” (Intraplan et al. 2003) and changes in the nature of supply by modes. They note that the macro process “relies on projecting past pattern on the country level to the future by time series analyses, whereas the micro models calculate the demand effect on each origin-destination link, taking into account the development of structure in traffic zones and supply factors particular to each mode of transport.” The demand response to new rail services were forecast by combining the results of two different micro models, the German Intraplan model, and the French M.A.T.I.S.S.E model developed at INRETS. • An observation from Chapter 2 is that having a common vision of overall trip making, whether on a corridor-by-corridor basis, or for a continent as a whole, expressed in a multimodal con- text can serve as anchor from which more specific analysis can be based. While forecasts of aviation demand growth are read- ily available in the United States, publicly available forecasts of growth in long-distance travel by highway and rail have up until now been lacking. The EU Study of Air and Rail Competition and Complementarity The European Union’s Directorate of Transportation and Mobility made a major contribution to the quality of the public dialog about the relationship between rail and air by commis- Figure 10-2. The EU has developed a model which is simple and transparent, supporting a public dialog about policy options. Source: Rizzo presentation, 2007.

146 travel in its complexity” (KITE Project) and apply that method in three countries. The project proposed “uniform design of intermodal passenger travel surveys” for use in later efforts. The logic is that, as individual European nations voluntarily chose to collect long-distance travel data, data collection would happen in a manner that could later contribute to the creation of a continent-wide database. In short, accurate sources of data about the origins and destinations of long-distance, multi-national travel by pri- vate modes (e.g., car, van, truck, and motorcycle) do not seem to exist either in Europe or in North America. Simulated Multimodal Demand in Europe This report contains the first major United States application of recent academic and consulting teams’ research in the simu- lation of major flows of long-distance travel, including access to and from airports. To develop the descriptions of a “natural” rail marketshed for each airport, the Research Team found that many airports did not have detailed descriptions of their success or failure in competing for wider geographic markets, or simply did not wish to share it with the public; thus, a newly develop- ing process of simulating passenger flows affecting these air- ports was examined. Using the tool “x-via web” developed by MKMetric and jointly marketed with STRATA Consulting, a simulated answer to the question of airport ground origins was provided. As described by its developers, “x-via web currently covers all of Europe and focuses on air passenger transport for more than 2350 regions plus 1150 airports worldwide with land-based feeders and the air links connecting them” (Travel- matrix 2011). This information was used to create map-based summaries of the geographic scale of origins of airport users in seven major airports. That information reveals, for example, Airports in France, including a survey undertaken in the year 2011. These studies provide publicly available data about such details as the wait time between train arrivals and air departures at CDG, which between 2005 and 2008 actually worsened to an average wait of 3 hours and 40 minutes. The extensive mar- ket research about the intermodal connection has been made available to the public, with summaries like Figure 10-3 on the DGAC website (DGAC 2009). This screenshot from the website presents the main reasons that air travelers selected the train over air feeder service; after cost, the second-highest reason was the lack of an air alternative. Had such a research program existed in the United States, a government agency would have been able to document the number of long-distance rail riders at Newark or BWI, pro- vided market research about how they purchased their ticket, and analyzed what factors would cause them to increase the use of the integrated service. Knowledge Base for Intermodal Transportation in Europe (KITE) The question of how to collect accurate descriptions of long-distance travel, particularly for modes where the trav- eler does not have to purchase a ticket is equally problematic in Europe. While there is a general belief that the EU would not support (or even encourage) a top-down approach, where data collection is forced on the member states, various research efforts have been undertaken to create what might evolve into a decentralized, or distributed, approach to data collection. The KITE study first reviewed the results of various national efforts to estimate the amount and location of longer distance travel in Europe. An additional element of the project devel- oped “a survey methodology suited to capture long-distance Figure 10-3. The French Civil Aviation Agency (DGAC) has published data on the use of HSR as a feeder mode. Source: DGAC.

147 The DB1B Origin-to-Destination (O-D) data available from the US DOT takes the form of a massive dataset, and can only be managed through tools capable of managing large amounts of data simultaneously. However, a good introduction to the O-D data can be found in the set of tables included in ACRP Report 31, which presented this information for all the major airports in the Northeast Corridor and in California. These tables present both O-D data from the DB1B system, and segment volumes from the T-100 system. Both the DB1B and the T-100 data are made available through the Bureau of Transportation Sta- tistics, and their use is discussed further in the discussion of the ACRP Project 03-23 Air and Rail Simulation Model, presented in Chapter 11. Improvement in Forecasting Aviation Data: Future Air Traffic Estimator (FATE) In the longer term, the ability to merge the planning pro- cesses for air with those for auto and those for rail will require a standardization of procedures so that common assumptions can be made as input to the travel demand calculations. The “four-step” metropolitan transportation planning process is multimodal in nature, and its methods have largely been applied on a statewide level. Such methods include common assumptions about how travel demand is associated with demographic factors, including income and employment levels. However, the process of predicting aviation demand has to date been accomplished through a separate—and incompatible—mode-specific process. Then, in the seminal document titled “Capacity Needs in the National Airspace System—2007–2025” (referred to as the “FACT 2” report) the authors revealed a transition to an additional system for longer term forecasting (MITRE 2007). The method, known as the FATE, was different from previous forecasting methods used by FAA because it: “[E]stimates the amount of passenger traffic between metro- politan areas rather than estimating demand at individual air- ports. Population, income and market structure all influence passenger demand, as does a host of other factors. Inputs to the model include socioeconomic forecasts from the consul- tancy Global Insight, as well as historical data on O&D traffic from the Department of Transportation” (MITRE 2007, p. 29, Appendix D). This evolution in the methodology of demand forecasting will facilitate the kind of integration of multiple data sources discussed in Chapter 12. Amtrak has shared key ridership information with the public on its website (http://amtrak.com). Figure 10-5 shows an example of a station summary that includes a certain amount of station-to-station demand information just how different the marketshed for Zurich is from Frankfurt, and how different that of Frankfurt Airport is from all other airports in Europe. • It would be highly desirable to have estimated airport origins (built up from actual ground access surveys when possible) available to American practitioners on a systematic basis. Rea- sonably, these would be an amalgam of the best available data, the quality of which would vary by location. Use of Single Mode Data in Multimodal Analysis: The Civil Aviation Authority, London (CAA) An ongoing source of support for the researcher in the United Kingdom is the continuous process of surveying that occurs at the major London airports and all others in the United Kingdom; the results are made available to the public via the CAA. The process is highly unusual in that the survey- ing at UK airports is continuous, with reports issued yearly by the CAA. A massive amount of data is made available through prepared tables via their website, with individual queries answered for a small cost, starting at under $200 (Civil Avia- tion Authority 2011). Table 10-1 is presented to demonstrate the level of detail of airport market data that is routinely made available to the public by the CAA in London. By way of example, Chapter 2 noted that Manchester Airport has developed a long-distance (beyond metropolitan) rail system directly serving the airport in the feeder mode. Any analyst can access the characteristics of the marketshed supporting Manchester Airport by down- loading a pre-made table. Table 10-1 shows that about 34% of those using the airport are from greater Manchester, with 66% from outside the metro area. Leisure travelers, and not busi- ness travelers, would dominate the market for longer distance rail, as the airport pulls its leisure travelers from a wider geo- graphic area than its business travelers. Only 29% of United Kingdom leisure travelers are coming from the metro area; of those non-residents coming to the area for business, 41% have destinations in the metro area. Analysts who seek more infor- mation (e.g., the same categories with access broken down by mode) can purchase specific tabulations for a modest cost. American Sources of Multimodal Data American Aviation Data The United States Department of Transportation (US DOT) has an active program of sharing with the public its available data about air travel. Available in many formats, the quick airport summary is shown in Figure 10-4, describing flows at Newark; the volumes of air travelers on the airport’s ten most important airport pairs are shown on the bottom right on the figure (Transtats 2011).

148 been incorporated in the FHWA’s publicly available data program. Amtrak’s decision was a major breakthrough for inclusion in both the ACRP Project 03-23 modeling pro- cess, and that of the FHWA’s research program, as discussed in Chapter 11. Long-Distance Highway Vehicle Flow Data from the U.S. DOT As noted herein, the overwhelming gap in the analysis of long-distance flows does not involve air or rail, but the that was provided in 2011. At a higher level of aggregation, yearly ridership by line is available on the website, which allows the analyst to derive trend data over time. Until the recent decision to participate directly in the ACRP Project 03-23 modeling effort (see Chapter 11), Amtrak did not have a formal method to make rail demand data available to other practitioners; this meant that while key aviation data had been restructured to be compatible with the tra- ditional transportation planning process, rail had not. Although detailed station-to-station demand data will not be publicly available, data on county to county flows have Source: CAA. Scheduled origin/destination patterns of terminating passengers at Manchester Airport in 2000. UK Foreign Grand Region County Business Leisure Business Leisure Total 000's % 000's % 000's % 000's % 000's % East Anglia Cambridgeshire 0.7 0.0 0.6 0.0 0.3 0.0 0.2 0.0 1.8 0.0 Norfolk 0.2 0.0 1.4 0.0 0.0 0.0 0.0 0.0 1.6 0.0 Suffolk 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 East Midlands Derbyshire 46.6 1.9 71.5 2.0 15.0 1.6 15.5 1.3 148.5 1.8 Leicestershire 3.7 0.1 19.3 0.5 4.1 0.4 1.5 0.1 28.6 0.4 Linconshire 2.6 0.1 9.3 0.3 2.3 0.2 1.4 0.1 15.6 0.2 Northamptonshire 1.3 0.1 3.6 0.1 0.1 0.0 0.0 0.0 5.0 0.1 Nottinghamshire 8.9 0.4 28.4 0.8 3.6 0.4 7.3 0.6 48.2 0.6 North West Cheshire 488.7 19.7 386.5 10.7 148.9 16.1 118.4 10.3 1142.6 14.0 Greater Manchester 921.2 37.2 1038.4 28.7 381.1 41.3 416.1 36.2 2756.8 33.8 Lancashire 195.8 7.9 378.8 10.5 80.6 8.7 110.4 9.6 765.5 9.4 Merseyside 228.1 9.2 349.9 9.7 83.7 9.1 105.3 9.2 767.0 9.4 Northern Cleveland 2.2 0.1 18.2 0.5 0.2 0.0 5.0 0.4 25.6 0.3 Cumbria 49.3 2.0 87.1 2.4 26.2 2.8 34.2 3.0 196.8 2.4 Durham 3.3 0.1 19.0 0.5 1.3 0.1 2.3 0.2 25.9 0.3 Northumberland 0.0 0.0 4.1 0.1 0.0 0.0 0.1 0.0 4.2 0.1 Tyne and Wear 4.1 0.2 25.6 0.7 2.0 0.2 9.3 0.8 41.0 0.5 West Midlands Hereford & Worcs 1.9 0.1 7.9 0.2 0.5 0.1 0.8 0.1 11.2 0.1 Shropshire 22.7 0.9 40.5 1.1 7.1 0.8 6.4 0.6 76.8 0.9 Staffordshire 65.3 2.6 110.2 3.0 19.6 2.1 22.4 1.9 217.5 2.7 Warwickshire 1.1 0.0 3.0 0.1 0.0 0.0 0.2 0.0 4.2 0.1 West Midlands 8.5 0.3 76.1 2.1 6.6 0.7 9.7 0.8 100.8 1.2 Yorkshire Humberside 23.9 1.0 70.7 2.0 7.0 0.8 15.8 1.4 117.4 1.4 North Yorkshire 39.7 1.6 112.0 3.1 20.5 2.2 41.4 3.6 213.6 2.6 South Yorkshire 91.2 3.7 173.4 4.8 27.7 3.0 47.1 4.1 339.4 4.2 West Yorkshire 164.8 6.7 389.8 10.8 55.3 6.0 97.2 8.5 707.1 8.7 Wales Clwyd 67.8 2.7 86.6 2.4 18.5 2.0 37.7 3.3 210.6 2.6 Dyfed 1.0 0.0 3.0 0.1 0.0 0.0 0.5 0.0 4.5 0.1 Gwent 0.9 0.0 3.2 0.1 0.4 0.0 0.0 0.0 4.5 0.1 Gwynedd 15.5 0.6 21.9 0.6 3.6 0.4 16.0 1.4 57.0 0.7 Mid Glamorgan 1.1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.1 0.0 Powys 2.7 0.1 5.9 0.2 0.4 0.0 1.6 0.1 10.7 0.1 South Glamorgan 0.0 0.0 0.6 0.0 0.7 0.1 0.2 0.0 1.6 0.0 West Glamorgan 0.2 0.0 0.8 0.0 0.1 0.0 0.0 0.0 1.2 0.0 Scotland 2.6 0.1 43.3 1.2 1.5 0.2 17.6 1.5 65.0 0.8 Grand Total 2474.8 100.0 3617.2 100.0 923.0 100.0 1149.8 100.0 8164.7 100.0 Table 10-1. Example of origin data by trip category, for use in defining marketshed of the airport.

149 showing how lower-income travelers have a higher propen- sity to stay with the private auto as the trip lengths become longer. As discussed in more detail herein, the FHWA is currently developing a nationwide long-distance trip table, whose vehicle flows are generated by a variety of variables, and whose trip distribution patterns are based on the best available data. This trip table will allow a common set of assumptions to be utilized in the analysis of longer distance highway travel, although it is not based on a recent original survey effort. • Based on the research undertaken in the project the propo- nents of air and rail modes could support the further devel- opment of research efforts, possibly using newly developing technologies in which one’s travel behavior is monitored on a voluntary basis, to create more accurate descriptions description of origins and destinations of vehicles using the highway system. Given the focus of this project on air and rail, the vexing lack of good long-distance highway flow data is beyond the scope of this report. In review of the pres- ent utility of either the 1995 American Travel Survey, or the 2001 National Household Travel Survey for the analysis of long-distance travel, proponents have suggested that both can supply valuable data about trip generation rates—and various characteristics of long-distance trip making. But there is general consensus that their direct use in determin- ing the distribution of trip ends for trip tables is more prob- lematic. Figure 10-6 (FHWA 2006) is presented here to give an example of how this kind of data can be used to under- stand fundamental characteristics of longer distance travel. In this case, data are organized at a high level of aggregation, Figure 10-4. High-quality aviation data is shared with the public on the Transtats website from the Bureau of Transportation Statistics. Source BTS

150 of long-distance, multi-state private vehicle travel. But, improved data resulting from methods will not be available for incorporation into the conclusions of this report. Ground Access Data Surveys by Airport In order to gain an understanding of the actual origins of passengers at an airport, available airport ground access surveys may be used in combination with T-100 data; this was demonstrated in Chapter 6, which examined the poten- tial for new rail services to act in a feeder mode capacity at O’Hare. A good example of how ground access surveys can be merged with other data sources can be found in the Regional Plan Association’s “Upgrading to World Class” study for the PANYNJ. As summarized in Table 3-1 of this report, for each major New York City airport, the reader could easily break out the number of people who access the airport from beyond the metropolitan area compared with the more traditional metro- politan markets. Although this data can be used to establish observations at a high level of aggregation (i.e., metro vs. beyond-metro market), issues of sample size constrain its use for further stratification. Strong traditions of consistent sur- veying exist at the PANYNJ, at the Metropolitan Washington Council of Governments, and in major MPOs in California. While most data collection happens within a metropolitan area (sometimes posing difficulties for merging of data), two major, multi-state exceptions to this pattern were the New England Regional Airport Systems Plan, and New York Region Plan of the FAA. When examined on an airport-by-airport basis, good data often does exist to support decision making. In the pro to- typical study of Chicago O’Hare presented in Chapter 6, the number of air passengers going between the airport and Figure 10-5. Amtrak provided ridership information in their station summaries. Figure 10-6. The Household Travel Survey was applied to the question of mode share as a function of distance in this analysis from FHWA. Source: FHWA Website.

151 By contrast, the work program undertaken for Chapter 11 includes the use of newly structured aviation demand data, newly released and restructured rail data, and new national vehicle flow data organized to be consistent with the other two data formats. The decision by Amtrak to provide the Research Team with base data that can be expressed in tradi- tional origin-to-destination format represents a major break- through in this area. The FHWA’s ongoing Travel Analysis Framework project, designed to create an integrated package of trip tables with a highway component, a rail component, and an aviation component is similarly a major breakthrough (Jenkins and Vary 2010). The Need for a “Quick Turn-around” Model of Air/Rail Mode Share Earlier sections of this Chapter described a major difference between the European Union and the United States in the qual- ity of publicly available information to be used in the public dialog about future investments in air and rail. Through their early investment in both the UIC study of long-distance travel demand, and the EU study of competition and complemen- tarity in eight corridors served by air and HSR, policy makers have the benefit of a common set of terms and assumptions on which to frame the public debate. Alternative scenarios of vari- ous forms have already been analyzed, and the public debate continues. Based on interviews in California, it is clear that models designed to produce extraordinary detail about such issues as sub-mode of access to and from stations were not designed to support quick turn-around planning-level analysis. Clearly, there is a trade-off between the level of detail of the analysis and the time required to develop and run the associated mod- els, and for some purposes a simpler analysis that can be run more quickly has some attraction, even if it does not produce the same level of detailed results. Aviation Response to Change in Competing Supply The report illustrates that more research needs to be done concerning the market reaction from the airlines in response to a decrease in demand for air services in a cor- ridor where high-quality HSR services have been intro- duced. In the corridors the Research Team examined, there were certain similarities in this pattern of response; with the exception of Frankfurt-Cologne (and portions of the Paris-Brussels market), most airlines responded by reducing flights to reflect the reduction in O-D traffic; however, these airlines still retained a market presence with flights needed to feed connecting traffic to their hub airport. Thus, while destinations up to 300 miles away was defined. The percent of the airport’s originating passengers that had destinations near the proposed rail lines was also defined, along with the proportion of those transferring at O’Hare who were using connecting flights. What was less clear in the case study was the ability to “predict” how many air passengers would choose rail as a feeder mode when competing air services were offered. This issue seems to require more data on which to calibrate any reasonable prediction model. • A conclusion from the analysis of market characteristics from one airport (as opposed to a corridor or a system) is that data collected describing ground access patterns can be combined with national aviation flow data to create valuable summaries of the potential markets that might be served by rail. American Sources of Information on Airport Access Issues In addition to the raw data available from the survey pro- cesses, the literature in the United States has several sources of guidance on the question of the adequacy of the ground access modeling process in a more general context, including the characteristics of the models themselves. ACRP Synthesis 5: Airport Ground Access Mode Choice Models provides a valuable review of applied demand modeling, much of which is directly relevant to questions addressed in the case study interviews. Also relevant is ACRP Report 26: Guidebook for Conducting Airport User Surveys. As noted before, that Synthesis did cover regional planning applications of airport ground access pat- terns, but it did not explicitly deal with the interaction between aviation and long-distance rail. Other sources of value to the analyst (but, again not specifically aimed at this study issue) include ACRP Report 4: Ground Access to Major Airports by Public Transportation and TCRP Report 83: Strategies for Improving Public Transportation Access to Large Airports and TCRP Report 62: Improving Public Transportation Access to Large Airports. Major Gaps Revealed in the Research Analysts and scholars have addressed the issue of gaps in American data. Two experts in long-distance travel at Oak Ridge National Labs have noted, “By maintaining mode-specific datasets that are rarely combined, it is also a difficult and resource intensive activ- ity to compare modal travel options with existing data sources. Rather, this information must be pieced together from a variety of sources with little or no consistency or coordination in the data collection methods being used” (Hu and South- worth 2010).

152 • A major conclusion to be drawn from Chapter 4 is that an analysis process is needed that specifically incorporates various scenarios of service responses by the aviation sector. The work program described in Chapter 11 was designed to address this important issue. The Need to Find New Methods to Better Understand Long-Distance Highway Flows This Chapter has emphasized ways in which policy makers in the European Union have made investments in research methods to better understand the relationship between air and HSR; this pattern is not true for the subject of improving the understanding of long-distance auto vehicle flow on the highway system. At present, analysts of long-distance travel flows in Europe are not assuming any massive investment in comprehensive origin-destination surveying for users of private vehicles. They, too, are seeking ways to estimate these flows through methods not yet developed. Given the dominance of the automobile in short and mid- distance travel there is clearly a larger market to be diverted from auto than from air. It is a concern that the scale of pos- sible diversion from auto cannot be well understood in a set- ting in which long-distance highway demand is essentially undocumented. While a narrow definition of the role of air and rail would technically not need solid data about markets currently on the highway system, a more holistic approach would acknowledge this very significant gap in understand- ing the totality of the mid and longer distance market. In this discussion of gaps in understanding, the need to solve the issue of better describing longer distance auto mobile travel Paris-Lyon was one of the original archetypes for a corridor in which the O-D market is virtually entirely served by rail, there are still seven flights a day between the two cities. Cer- tain differences were also revealed; in general airlines in the Madrid-Barcelona corridor have responded with less service cuts that those in the Paris-Marseille corridor. Within the pattern in which the airlines lower the number of flights, there remains significant variation in the extent to which the airlines ceded the O-D market to the competing rail (Figure 10-7). In particular, Chapter 4 shows that even after the opening of the HSR service between Madrid and Barcelona, the “air-bridge” shuttle service is still assertively marketed. Between Madrid and Seville, about eight planes a day are operated, and the rail share of air plus rail is approxi- mately 80%. Between Madrid and Barcelona (a much big- ger and more important corridor for the airlines) the air operators have reacted in a fundamentally different way: the shuttle operations continue throughout the day, with a service frequency of three flights per hour in peak hours. Given this different reaction from the air industry, the rail share is markedly lower than experienced in other Euro- pean corridors with rail service faster than three hours from terminal-to-terminal. The market reaction to the HSR service from Madrid to Barcelona is similar to the market reaction to the Boston–NYC services; both attained just over a 50% share of O-D riders. This is important for this study, as the report documents that between Boston and NYC, the aviation industry responded by keeping the headways on critical Logan–LaGuardia ser- vices largely unchanged, while flying smaller aircraft. Figure 10-7. Market reactions of the aviation sector. Madrid to Seville Air Competition is Weak (85%) Madrid to Barcelona Air Competition is Strong (53%) Boston to NYC Air Competition is Strong (55%)

153 and Vary 2010). That project will also create long-distance trip tables for air and rail as part of this unified effort to improve the quality of long-distance travel data. Chapter 10 has sought to identify best practices and gaps. A lack of proper understanding of long-distance auto trip making must be reported as a major gap, a problem without any clear solution in sight. The information presented in this Chapter is now summarized in the form of a “Checklist” of available data and methods and their status in terms of further research from various programs (Table 10-2). patterns (with emphasis on origins and destinations) needs to be flagged. Solutions would range from a full-scale replica- tion of the 1995 American Travel Survey at one extreme, to incorporation of data originally collected for calibration of in-vehicle navigational devices at the other end of the technol- ogy spectrum—or some merging of various elements of sev- eral strategies at once. Importantly, the FHWA has addressed the problem of creating a standard reference for a national long-distance trip table, with the inauguration of its innova- tive Travel Analysis Framework (Part IIA) project (Jenkins A “Checklist” of Available Tools and Methods/Needs for Further Research Need for Data, Tool or Method Examples from the Case Studies of this Project Concern Implications for Present and Future Research Difficulty of application Told by the MTC that Policy analysts need to New modeling tool has been of demand model for rail model literally takes test different scenarios in (Chapter 11) applied in the competition with air to several days to be a transparent, and testing of alternative scenarios and use in policy analysis. prepared and run. understandable way. hypotheses. Difficulty in exploration of the role of the airline industry in response to competition. Issue has been a concern for some time; what would the dominant airline in that corridor do? Pricing and schedule characteristics are treated as exogenous in most modeling processes—not impacted by the change in competitive environment. The relationship between level of rail service and amount of air service is addressed in Chapter 11, in a preliminary manner. Lack of consistent method to predict the role of rail in a longer term scheme to have rail serve in a feeder mode to longer distance flights. Chapters 2 and 3 present the most complete review yet of use of rail-as-feeder mode in Europe, and failure to do so in the United States, even with Continental/ Amtrak agreement in place in Newark. Participants in Northern California were concerned that rail alignment decisions were being made under the assumption that rail could not work as feeder to air. Further research examining European experience would be appropriate, lasting beyond the course of this ACRP project. In areas with existing patterns of rail reliance, use of long-distance rail to gain access to airports is high; efforts to “replace” feeder flights have mixed results. Issue needs to be understood in both California and NEC. A nationwide forecast for rail volumes by corridor. Work of the UIC, based on research methods from Germany and France. There is very little agreement about the scale of ridership between corridors in the USA. Since the FRA’s publication of HSR in the United States Report (Volpe Center) there is probably a need to update national visions of ridership based on studies currently underway. A nationwide forecast for highway-based long-distance trip making volumes expressed as zone to zone trip tables. One of the major concerns expressed by NYC RPA research team—lack of long- distance trip making descriptions. Lack of success with direct use of 1995 ATS and 2001 NHTS data has led researchers to downplay long-distance trip making by highway modes. FHWA’s presently ongoing project, Traffic Analysis Framework Part IIA, will build a county to county national trip table based on existing data resources. This will allow a common framework over which policy scenarios can be reviewed and analyzed. Table 10-2. Checklist of Available Tools and Methods. (continued on next page)

154 154 A “Checklist” of Available Tools and Methods/Needs for Further Research Need for Data, Tool or Method Examples from the Case Studies of this Project Concern Implications for Present and Future Research A nationwide forecast of Given there is little With separate projects The FHWA’s Traffic Analysis A full description of bus travel O-D patterns in Europe. long-distance travel demand by highway, air and rail. consensus about present long-distance, participants in the case study interviews had very little optimism about longer term forecasts. around the nation making their own visions about future growth, it is extremely difficult to compare and contrast longer term visions. Framework project is undertaking the expansion of the base year 2008 trip tables for highway, air, and rail to the forecast year 2040. This is being done in a method consistent with that of the FAA’s FACT 2/FATE project and work undertaken for Chapter 11. The Research Team found no examples where the private sector bus industry shares its “proprietary” ridership data with the governments. It is difficult to understand how competition between air and rail is impacted by bus ridership, even where it is clearly occurring in the United Kingdom. The Research Team is not aware of any effort in the UK to standardize, and make available to the public long- distance bus ridership data. (Swiss survey methods would capture short distance bus, but long-distance bus is not a major force in Switzerland.) A full description of bus travel O-D patterns in the United States. No one interviewed in the case study process believes that bus ridership is properly documented in the United States. Anecdotal data suggests that in some sub-corridors in the Northeast, more people ride the long-distance bus than take the train. Decision makers are concerned about the lack of systematic data about bus ridership. Episodic surveying may be undertaken in the Northeast Corridor, but these plans are not final. Issue remains unsolved. A full description of air travel, by O-D pattern and by segment. The US DOT’s Bureau of Transportation Statistics has an unparalleled program for sharing highly detailed aviation data with the public Consistently available information supports the ability of the public to engage in dialog and debate. Data from US DOT’s BTS is readily available for use by researchers, and by the public in general. Publicly available descriptions of airport- specific ground access markets. The Civil Aviation Authority (CAA) of the United Kingdom has a continuous program of data collection that is unmatched in its availability to the public. Consistently available information supports the ability of the public to engage in dialog and debate. No program in the United States compares with the kind of ground access market data commonly distributed by the UK’s CAA. Some coordinated data collection has occurred in New England and in the New York region. A standardized analysis tool to be applied over a wider variety of markets. The “Competition and Complementarity” studies performed for the EU by SDG provide a consistent framework for scenario evaluation. Strong emphasis on locally managed project development makes comparison across projects difficult. FRA is aware of the difficulty of making comparisons across projects, and might continue tradition started with Volpe’s HSR in the United States summary of corridor data. A nationwide forecast for aviation volumes expressed as county to county trip tables. Most aviation forecasts have been done on a uni-modal basis, in a manner that makes integration with other modal data difficult. The FAA’s FACT 2 report introduced the FATE forecasts by the MITRE Corporation, which present data in county to county trip table format. Aviation data prepared by/for FAA can now be integrated with new data development effort by FHWA Traffic Analysis Framework. A nationwide description of present trip making on Amtrak, expressed as county to county trip tables. Several case study participants noted that Amtrak considers most of its ridership data to be proprietary in nature, and used in its own business analysis Need for openness and transparency in ongoing analysis and public debate. The Research Team has formatted the Amtrak ridership data into a county to county trip table format, for use in the Chapter 11 work program, in a manner consistent with later integration with other modal data by FHWA. Table 10-2. (Continued).

155 Intraplan Consult GmbH, INRETS, and ImTrans. 2003 (February). “Passenger Traffic Study 2010/2020 Executive Summary.” On Behalf of the International Union of Railways (UIC). Jenkins, D., and D. Vary. 2010 (November 5). “FHWA Traveler Analysis Framework, Part II, Status Report.” AMPO Travel Modeling Meet- ing, Washington, D.C. KITE Project. Accessed at http://www.linkforum.eu/docs/604/Cyprus_ conf_-KITE_project_Last-Sender.pdf. (As of Spring 2011). Leigh Fisher Associates, M. A. Coogan, and MarketSense. 2000. “TCRP Report 62: Improving Public Transportation Access to Large Air- ports.” Transportation Research Board, Washington, D.C. Leigh Fisher Associates, M. A. Coogan, and MarketSense. 2002. “TCRP Report 83: Strategies for Improving Public Transportation Access to Large Airports.” Transportation Research Board, Washington, D.C. MITRE Corporation, The. 2007 (May). “Capacity Needs in the National Airspace System—2007–2025.” Federal Aviation Administration, Washington, DC. Rizzo, G. 2007. “Competition Between Air and High Speed Rail: Conclu- sions of the Study Prepared for the European Commission by Steer Davies Gleave.” A Presentation to the International Air/Rail Associa- tion, Vienna, Austria. Steer Davies Gleave (SDG). 2006. “Air and Rail Competition and Com- plementarity.” Prepared for the European Commission DG TREN. London, U.K. Transtats. 2011, accessed at http://www.transtats.bts.gov. (As of Spring 2011). Travelmatrix, 2011 http://www.travelmatrix.eu. (As of Spring 2011). U.S. Department of Transportation, Washington, DC. Civil Aviation Authority. 2011. http://www.caa.co.uk/docs/5/Catchment%20 area%20analysis%20working%20paper%20-%20FINAL.pdf. Zupan, J. M., R. E. Barone, and M. H. Lee. 2011 (January). “Upgrading to World Class: The Future of the New York Region’s Airports.” Bibliography Axhausen, K. W. 2010. “KITE—A Knowledge Base for Intermodal Passenger Travel in Europe.” Travel Survey Metadata Series, 31, Institute for Transport Planning and Systems (IVT), ETH Zürich, Zürich, Switzerland. Bureau of Transportation Statistics. 1995. “American Travel Survey.” Civil Aviation Authority, http://www.caa.co.uk (As of Spring 2011). Coogan, M. A., MarketSense Consulting LLC, and Jacobs Consultancy. 2008 (July). “ACRP Report 4: Ground Access to Major Airports by Public Transportation.” Transportation Research Board, Wash- ington, D.C. DGAC. 2009 (February). “Enquête sur la complémentarité modale TGV/ Avion Synthèse des résultats.” Federal Aviation Administration. (No date). “The New England Regional Airport System Plan.” Sponsored by the New England Airport Coalition. Accessed at http://www.faa.gov/airports/ new_england/planning_capacity/airport_system_plan/media/ nerasp_complete.pdf. Federal Aviation Administration. 2007 (May). “FAA Regional Air Ser- vice Demand Study, Summary Report.” Port Authority of New York & New Jersey, New York State Department of Transportation, and Delaware Valley Regional Planning Commission. Federal Highway Administration. 2006 (January). “NPTS Brief.” US Department of Transportation, Washington, DC. http://www. amtrak.com. (As of Spring 2011). http://www.caa.co.uk. (As of Spring 2011). http://www.transtats.bts.gov. (As of Spring 2011). Hu and Southworth. 2010 (January). Paper presented at the Workshop on Multimodal Inter-Regional Passenger Travel Demand, Univer- sity of Maryland.