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.
87 CHAPTER 6. METHODS WE USED IN THIS PROJECT 6(A) INTRODUCTION AND STRUCTURE INTRODUCTION This ACRP project created several new models to better understand the influence of many causal factors on the decision between the auto and the plane in various aspects of the full long-distance trip. The attitude-based Structural Equations Model was presented in Chapter 5, with other discussions of attitudes, values and preferences. The other new models are described in this chapter. STRUCTURE The first part of Chapter 6 describes the Stated Preference survey undertaken, and the subsequent creation of two new models, designed to emphasize the immediately relevant factors of travel time and costs, in addition to other traditional variables such as travel party size. By design, these models do not emphasize the âsofterâ factors including values, preferences and attitudes. These models are called the Multinomial Logit and Mixed Multinomial Logit Choice Models. The modeling process was then refined with the development of an Advanced Hybrid Choice model. The final part of Chapter 6 describes the Scenario Testing Model that was developed based on an adaptation of a major national model of intercity travel, after the research team made major additions to the treatment of airport routing, airport section, and access/egress options in that model. 6(B) THE PROJECTâS STATED PREFERENCE SURVEY, 2017 For this study, a major national survey was conducted in 2017 in four metropolitan areas to collect long- distance travel behavior and attitudes as well as administer a series of Stated Preference (SP) experiments that deal with the choice between car and air. Sampling plan The objective of this study was to examine the choice to use car or air for long-distance travel in the United States. With this objective in mind, the sample for the study was comprised of travelers who have taken an automobile or air trip of greater than 300 miles within the past year. Quota sampling, a technique that sets a minimum number of respondents for each respective category, was used to establish a diverse sample of respondents. Respondents were recruited through Research Now, a major online sample provider. To be eligible to take the online survey, respondents must have lived in a qualifying region and be over the age of 18. The overall goal for the survey was to obtain 4,000 respondents. Invitations to the online survey were emailed to respondents who resided within the Combined Statistical Area (CSA) centered on each of the studyâs four target regions. Minimum quotas were set for each metropolitan area as well as combinations of reference trip distance, purpose and mode. Soft quotas were used to ensure an equal representation of gender and a range of incomes and ages in,
88 meaning that sample was targeted proportionally along these dimensions and was monitored during the data collection process, but exact minimum quotas were not set. Questionnaire A questionnaire was designed to understand present mode choice behavior and the sociodemographic characteristics of each survey respondent. The survey instrument also collected basic information concerning attitudes affecting the propensity to choose traveling by car or plane for trips of 300 miles or more. The survey questionnaire was drafted and reviewed with the Panel. Once the questionnaire content was finalized, the web-based survey was programmed. The survey first asked a few demographic questions for screening purposes (to ensure the sample was in the geographic areas of interest and determine which quota cells respondents fell into). Then respondents were asked to enter the total number of round-trips they have made within the last year. These totals were collected separately for medium distance trips (300-500-miles from the metropolitan area) and long-distance trips (over 500 miles) and categorized by mode (car or plane) and primary purpose (business or leisure). Each question included an interactive map so that respondents could see which destinations fell into each distance category. The survey also asked if any multi-destination or one-way long-distance trips were taken within the past year. Following this section, respondents were assigned a trip distance, purpose and mode where they had entered at least one round-trip and asked to provide further details about their most recent trip in this category. Trip details included origin/destination, party size, trip duration, trip costs and specific questions relating to air trips or car trips. This information was used to construct a series of eight Stated Preference (SP) trade-off experiments where respondents were asked to choose driving or one of three air options all with varying characteristics for similar trip as they had just described. Attributes that varied for the air options were: departure airport, number of stops, airfare, travel time, airport access mode and access travel time. Airport parking and access gas costs were also included if applicable. Attributes that varied for the car option were door-to-door driving time, gas costs and if the trip was made with an autonomous vehicle. A screenshot of an example Stated Preference experiment is pictured in Figure 6-1. Respondents were then put through a battery of approximately 50 attitudinal questions in 8 sets to get an idea of their general attitudes toward the various aspects of medium- and long-distance air and automobile travel. Respondents were asked to what extent they agreed or disagreed with each statement on a 7-point scale. The survey concluded with an additional set of demographic questions. Survey administration The full field effort occurred in early 2017. Prior to beginning analysis, the records were reviewed to identify potentially bad data. Three criteria were used to identify bad respondents: ï· The respondent had taken the survey in an unreasonably quick time (under 8 minutes) ï· The respondent had indicated that they made an unreasonable number of long- distance trips (over 200 trips in a year).
89 ï· The respondent straight-lined (provided the same response) for at least 4 out of the 8 sets of attitudinal questions. FIGURE 6-1 SCREENSHOT OF AN EXAMPLE OF A STATED PREFERENCE EXPERIMENT
90 Respondents meeting these criteria were removed from the dataset and new respondents were recruited to complete data collection. The tables and charts below provide an information about the sample makeup. Survey results The field effort resulted in a total of 4,223 valid responses, exceeding the goal of 4,000. A description of the surveying process is presented in Chapter 6 of the Technical Appendix. The research team aimed to collect 800 completed surveys in each metro area to provide enough sample for analysis and exceeded that goal in all four metro areas. The final sample is split evenly among the four metro areas. 6(C) DEVELOPING THE MODELS FROM THE PROJECT SURVEY RESULTS THREE KINDS OF KINDS OF MODELS The project developed three separate kinds of mathematical models to aid in the understanding of the choice of mode between the auto and the airplane in the long-distance trip. Attitudes were explored in the first model: the Structural Equations Model (SEM) was presented in the previous chapter on attitudes. The SEM model was designed to emphasize the importance of âsoftâ variables, including values, preferences and attitudes in the selection of modes for the long- distance trip; by design, this model does not emphasize the trip-based times and costs. The second models were designed to emphasize the immediately relevant factors of travel time and costs, in addition to other traditional variables such as travel party size. These models are called the Multinomial Logit and Mixed Multinomial Logit Choice Models. The third model is designed to integrate all relevant factors into the prediction of (in our case) long-distance travel mode. The model takes the form of an Advanced Hybrid Choice model, which is also described as an Integrated Choice Latent Variable model. It was developed to support the development of a nationwide travel demand modeling process, referred to here as the Scenario Testing Model. 6(D) MULTINOMIAL LOGIT AND MIXED MULTINOMIAL LOGIT CHOICE MODELS Early logit models were developed to support the early interpretation of the project survey results, and to be used to support the later development of the Advanced Hybrid Choice model, (described below.) These early results were included in the Interim Report and used as the basis for early discussion of possible conclusions/directions for the study. The early results of the logit models were then refined in the development of the Advanced Hybrid Choice model, which should be seen as the final product of the modeling process undertaken in this study. MODEL ESTIMATION The research team included a stated choice exercise as part of the ACRP survey. During the exercise, each respondent was shown a series of travel scenarios in succession (discussed in greater detail in the Technical Appendix. Flight time, cost, frequency and driving times and costs changed and respondents had to choose a mode for each scenario. The research team simplified the
91 experiments by limiting the choices while still collecting the necessary data to accomplish the projectâs objectives. The estimation work conducted makes use of both multinomial logit (MNL) and mixed multinomial logit (MMNL) models. The research team conducted much of the determination of appropriate model specifications (e.g., in terms of specification of cost) by using MNL models. An MMNL model was then run using the final MNL specification. The MMNL models allow for random variations in sensitivities and preferences across individual travelers. The research team estimated several different models and structures before arriving at the structures presented in this chapter. Throughout the process, the research team examined different cost structures, distance effects, and demographics. LOGIT MODEL RESULTS This section presents results from three multinomial logit models (MNL). The models include coefficients for the level-of-service attributes (tested in the stated preference experiments) and alternative specific constants (a measure of the attractiveness of each option that is not captured in the level-of-service attributes). This model includes shifts on the alternative specific constants for airport access modes, different airports, and demographics, showing how modes are valued differently under different circumstances. In addition, this section presents a MMNL version of the model for all purposes. Air constants in the logit models The air constant includes a base constant and several shifts, including shifts for access modes and different departure airports. For MNL models, the research team shows the estimated parameter value alongside the robust t-ratio against zero, giving a measure of statistical robustness (a one- sided 95% confidence level would be reached with a test value of 1.65). For the MMNL models, the research team shows the mean alongside the standard deviation, giving an indication of the heterogeneity in the sample data. The base constant shows that air is preferred to car overall. This effect is even stronger in the business travel model, which shows that air is even more preferred for business travel. The MMNL models highlight a large amount of heterogeneity in these baseline mode preferences. Airport preference. The model also includes shifts on the air constant for all the large departure airports included in the study. These are best understood regionally, as there are slightly different factors affecting each region. In the Washington, DC, region, the main trade-off is between three large airports: Dulles (IAD), Reagan National (DCA), and Baltimore-Washington (BWI). Respondents in the northeast area of the region also saw experiments with the Philadelphia airport (PHL). Constants for each of these airports are compared to several smaller airports in the region, including Charlottesville, Richmond, and Harrisburg. The model shows that Reagan and BWI are preferred while Dulles is less preferred. In the Boston metro area, the trade-off is between Logan (BOS), a large airport downtown, and several smaller regional airports in Manchester, New Hampshire; Hartford, Connecticut; and Providence, Rhode Island. The model shows that Logan is preferred to these smaller airports. In the Chicago area, OâHare (ORD) and Midway (MDW) are compared to the base of all other airports in this region. The other airports include Milwaukee and
92 several smaller airports on the outskirts of the region. Both large central airports are preferred to others, and OâHare has a higher positive coefficient than Midway. Finally, in the Denver region, the main trade-off is between the centrally located large hub, Denver (DEN), and the small airport at Colorado Springs. A handful of even smaller airports are occasionally included. The model shows that Denver is preferred over Colorado Springs and the other small airports in the region. The MMNL results are consistent with the MNL findings, but again show extensive heterogeneity across travelers. Car constants in the logit models As presented in the Technical Appendix, several shifts are applied to the car constant in the original logit models, primarily for demographic effects. In addition, there is a key difference if a rental car is shown for the long-distance trip. The rental car shift is negative, indicating that using a rental car is much less preferred than a personal vehicle, with substantial heterogeneity in the MMNL models. The model compares the age effects by using individuals between the ages of 45 and 64 as the base. The age shifts on the car constant show a trend where younger people prefer air to car, with car being increasingly preferred as respondents get older. Autos owned. There is a strong negative shift for respondents from households with no cars indicating that air is preferred by these respondents. Households with more than one member are more likely to choose car, and households with two or more unemployed adults are also more likely to choose carâboth logical shifts. Party size. For a trip party size of one, air is preferred versus a base of all other trip party sizes. Air is preferred for round trips made in the same day for MNL but not for MMNL; car is greatly preferred for trips that involve staying away for seven or more nights. Again, these results are logical as making a round trip in the same day for a trip of at least 300 miles would be extremely difficult and even impossible in some cases. It also makes sense that the additional time needed to drive might not be such a deterrent for longer stays at the destination. MIXED MULTINOMIAL LOGIT MODEL APPLICATION AND SIMULATION To demonstrate how the choice models can be used to answer questions about changes in level of service, the research team applied the Mixed Multinomial Logit model to the ACRP survey data using a sample enumeration technique. This technique applies the model to each record of the dataset using the origin and destination of the reference trip as well as other variables associated with that trip, including party size, length of stay, access mode to each airport and demographics. The Base Scenario uses air level of service data based on D1B1 and on-time databases adapted for the national long-distance model. This includes a base cost, flight time and frequency of direct, 1- stop and 2-stop itineraries between each origin and destination pair in the sample. Driving distances for the driving trip and from home to the airport are taken from google directions. Parking and gas costs are input directly to approximate the current conditions. All variable inputs to the simulator are shown in Table 6-1.
93 TABLE 6-1. VARIABLE INPUTS TO THE MMNL MODEL SIMULATOR AIR INPUTS VALUE DESCRIPTION Large Hub airport parking $26.00 per day Medium airport parking $15.00 per day Small airport parking $10.00 per day Airfare multiplier 1 Air travel time multiplier 1 Add frequency of direct flights (per day) 0 Change in direct flights per day Add frequency of 1 stop flights (per day) 0 Change in one-stop flights per day Add frequency of 2 stop flights (per day) 0 Change in two-stop flights per day CAR INPUTS VALUE DESCRIPTION Gas cost per gallon $2.50 per gallon Car travel time multiplier 1 Autonomous vehicle 0 Self-driving vehicles are not available The sample is weighted to the total number of trips by distance (2 categories), purpose (business or leisure) and mode and then the model is calibrated so that the overall mode share is equal to the total number of trips reported by respondents in the survey. While this mode share is perhaps a reasonable estimate for demonstrating the application of the model, it should not be seen as a definitive result for long-distance mode shareâwhich is addressed below in the discussion of the final Scenario Testing model. Early scenarios explored in the testing of the logit models The Base Scenario shows a mode share of 48.9% air and 51.1% car. Of those that fly 84% fly out of a large hub airport while 16% flight out of a smaller or medium airport. These results are based on respondents living within the 4 metropolitan areas in the survey sample: Washington, DC, Boston, Chicago and Denver and would depend on the available departure airports and the geography of the sample. Table 6-2 shows the mode share and airport share (of those that fly) in the Base Scenario as well as percent changes from the absolute values for 6 example scenarios. Parking. The first scenario increases the cost of parking at small and medium airports by 50%. Small airport costs go from $10/day to $15/day while medium airport costs go from $15/day to $22.50/day. This results in, unsurprisingly, a small reduction in overall air trips (0.5%) and a shift in air trips from small and medium airports to large hub airports. In fact, the 50% increase in parking costs leads to a 9.7% decrease in trips at the small and medium airports. TABLE 6-2. BASE SCENARIO AND PERCENT CHANGE FOR ALL EARLY TESTING SCENARIOS EARLY TESTING SCENARIO LARGE HUB AIRPORTS SMALL AND MEDIUM AIRPORTS FLY DRIVE Base Scenario Percent 84.0% (of 16.0% (of 48.9% 51.1%
94 air trips) air trips) Base Scenario Count 1733 331 2064 2159 1. 50% increase in parking cost at small and medium airports 1.2% -9.7% -0.5% 0.5% 2. Gas $4.50/gallon 10.4% 8.8% 10.1% -9.7% 3. Airfare 20% higher -3.1% -5.0% -3.4% 3.3% 4. Add one direct flight to all itineraries -0.2% 10.5% 1.5% -1.4% 5. Subtract one direct flight from all itineraries 0.0% -10.5% -1.7% 1.6% Price. The second and third scenarios involve changing the costs of flying or driving and show the impacts if the gas price were to rise from $2.50/gallon to $4.50/gallon and if the all airfares were to rise by 20%. The $2 increase in gas costs per gallon results in about a 10% increase in air trips. Trips from smaller airports grow at a slightly smaller rate than trips from large airports. This is likely because increased gas costs also increase the cost of driving to the airport and the small airports in this example are more frequently farther away. A 20% increase in airfare results in a 3.4% decrease in air trips; trips from small airports decrease at a higher rate than trips from large airports. This suggests that small airports have less resiliency in dealing with an overall airfare increase. More flights. The final two scenarios involve altering the network of available flights. The example presented here changes the number of direct flights offered on all possible flight itineraries positively and negatively by adding or subtracting one direct flight. The overall increase in frequency leads to a 1.5% increase air trips and a 10.5% increase in air trips from small airports. The five early scenarios used in the model development process demonstrate the types of simulation that can be done by applying the model and can serve as a jumping off point for discussion and for further, more rigorous, applications of the model. 6(E) HYBRID CHOICE MODEL FOR SCENARIO TESTING INTRODUCTION Advanced Hybrid Choice models, also often referred to as Integrated Choice/Latent Variable (ICLV) models were developed by the research team in order to integrate the wide variety of results from processes which examine either (a) effects of times and costs, or (b) effects of preferences, attitudes and values. These models account for the differences across individual respondents in their preferences in terms of their baseline preferences for given modes of transport. We also allowed for further heterogeneity in these modal constants that is linked to attitudinal constructs. These latent attitudes vary both deterministically (e.g., as a function of age) and randomly (i.e., due to unobserved factors) across individuals. At the same time as explaining a share of the heterogeneity in modal preferences across respondents, they are also used to explain the answers that these same respondents give to a set of attitudinal questions. Model estimation jointly maximizes the likelihood of the observed choices and observed answers to the attitudinal questions for a given respondent. This creates the link between the two parts of the data. Separate models were estimated for business and leisure trip purposes.
95 RESULTS: OUTPUTS FROM MEASUREMENT MODEL AND STRUCTURAL MODEL We first looked at the outputs of the model estimation process for the structural equations and measurement model. This helps us to clarify the role and meaning of the five attitudinal constructs which we use to explain the answers respondents provide to 13 separate attitudinal questions. The grouping used relies on the insights of the Structural Equation Model described at the end of Chapter 5. We used the groupings that had the largest effect in the SEM as long as they were based on the attitudinal questions (for example the cost factor in the SEM is not included even though it had the largest effect in the SEM because cost is dealt with in the choice model part of the ICLV). The five latent variables (LV) here are referred to as: Auto Orientation, Values Information Technology, Multiday Trips Unpleasant, Car Stress, and Airport Stress. LV1: Auto Orientation Our first latent construct is used to explain the answers to three separate attitudinal statements. We examined the impact of the latent variable (LV) on the attitudinal statements. This resulted in positive signs, showing that a higher value for this LV means stronger agreement with the attitudinal statement, identifying this LV as pro-car. Four sociodemographic characteristicsâage, number of household vehicles, household size, and household incomeâwere used to explain the value of the LV. Having no household vehicles indicates an anti-car attitude, pro-car attitude increases as household size increases, and pro-car attitude decreases as income increases. LV2: Values Information Technology Our second latent construct is used to explain the answers to two separate attitudinal statements. This LV was only included in the leisure model, due to insignificant effect when tested in the business model. The results show that a higher value for this LV means stronger agreement with the attitudinal statement, identifying this LV indicates a value for access to information and technology. We see, unsurprisingly, that younger age groups and females put a greater value on technology while smaller households put lesser value on technology. Value placed on technology increases with household income. LV3: Multiday Trips Unpleasant Our third latent construct is used to explain the answers to three separate attitudinal statements. The results show that a higher value for this LV means stronger agreement with the attitudinal statement, identifying this LV indicating aversion to multiday car trips. We see that younger people and people traveling alone find longer multiday car trips more unpleasant. LV4: Car Stress Our fourth latent construct is used to explain the answers to two separate attitudinal statements. The results show that a higher value for this LV means stronger agreement with the attitudinal statement, identifying this LV as an anti-car LV. We see that younger respondents feel more stressed with automobile travel than older respondents. People in larger households (4+) find car travel to be more stressful than those in smaller households. Finally, car travel becomes less stressful as income increases.
96 LV5: Airport Stress Our fifth and final latent construct is used to explain the answers to three separate attitudinal statements. The results show that a higher value for this LV means stronger agreement with the attitudinal statement, identifying this LV as an anti-air travel LV. Females and younger age groups are less stressed about air travel while stress with air travel decreases as income increases. Impact of LVs on Utilities in Choice Model We can observe the impact of the five LVs on the mode specific utilities in the choice model. ï· For the Auto Orientation LV, we find that a positive value for this latent attitude increases the utility (and hence probability) for car. ï· For the Values Information/Technology LV, we find a negative impact on car in the leisure model. This LV was not included in the business model. ï· For the Multiday Trips Unpleasant LV and the Car Stress LV, we observe a negative impact car in both the business and leisure models. ï· Finally, for the Airport Stress LV, we observe, as expected, a positive impact on car choice. OTHER DETERMINISTIC AND RANDOM HETEROGENEITY IN MODE SPECIFIC CONSTANTS We next looked at the baseline mode constants which indicate a preference for air or car assuming all other characteristics of the trip are equal. The results show that air is preferred to car overall. This effect is even stronger in the business travel model, which shows that air is even more preferred for business travel. The access mode shift for being dropped off at the airport is positive in the leisure model (although insignificant in the business model) implying that those who are able to be dropped off at the airport have an even greater preference for air. However, the shift for taking another mode, such as a taxi, is negative, which indicates that taxi access is less preferred than driving. This negative shift can be explained because taxi costs were not included in the SP data and were not included in the model. Finally, the model includes a random component for the air constant, indicating that there is significant heterogeneity across respondents that is not explained by the shifts included in the model. Estimates Relating to Explanatory Variables Table 6-3 examines the parameters explaining the sensitivities to the explanatory variables, namely costs, access time, in-vehicle time, frequency as well as direct or indirect flight itineraries. The robust t-ratio is a measure of significance and is defined as coefficient divided by the robust standard error. As a rule of thumb, robust t-ratios with an absolute value of less than two are not significantly different from zero. All cost and time coefficients were negative, indicating that increased cost or time to a travel option would be a deterrent to that option. Flight frequency is split into frequency for direct flights and frequency for indirect flights. Both coefficients are positive, indicating that more flights in a day is better, as expected. Adding frequency for connecting flights is a bit more valuable than
97 adding frequency to direct flights, likely because this might lessen the burden of missing a connection. TABLE 6-3. ESTIMATES RELATING TO EXPLANATORY VARIABLES HYBRID CHOICE MODEL A connecting flight was much less attractive to respondents than a direct flight and a flight itinerary with two stops was less attractive than one stop. Finally, we see a strong income elasticity, showing that for a 10% increase in income, we see a reduction in cost sensitivity of 3.0% in the business model and 1.9% in the leisure model. IMPLIED MONETARY VALUATIONS FROM THE HYBRID CHOICE MODEL While not a core focus of this study, one way to evaluate the sensitivities that are estimated in the model is to calculate the marginal rates of substitution for different attributes of interest. In basic economic theory, the marginal rate of substitution is the amount of one good (e.g., money) that a person would exchange for a second good (e.g., travel time), while maintaining the same level of utility, or satisfaction. Table 6-4 shows the resulting values of time for each of the three models. These values are shown for respondents with the median household income of $87,500. Value of time will change with income based on the income elasticity coefficient. COEFFICIENT BUSINESS LEISURE ESTIMATE T-RATIO ESTIMATE T-RATIO Air fare ($) -0.01 -13.10 -0.01 -16.87 Gas cost for access ($) -0.04 -2.42 -0.06 -4.37 Parking cost for access ($) -0.01 -5.46 -0.02 -7.28 Gas cost for driving ($) -0.01 -3.22 -0.02 -12.44 Rental cost ($) 0.00 -0.79 0.00 -1.09 Flight time (min) 0.00 -8.69 0.00 -13.06 Access time (min) -0.02 -16.81 -0.03 -26.06 Car drive time (min) 0.00 -6.03 0.00 -11.46 Log of frequency of direct flights 0.16 6.82 0.16 7.81 Log of frequency of connecting flights 0.12 5.55 0.15 7.64 Driving trip made by an autonomous vehicle 0.00 -- -0.40 -4.70 One-stop itinerary (vs. direct) -0.74 -10.60 -0.74 -11.66 Two-stop itinerary (vs. one stop) -0.40 -2.98 -0.22 -1.68 Income elasticity -0.30 -3.31 -0.19 -3.72
98 We see that business travelers place a higher value on all attributes than leisure travelers indicating that they are more willing to pay for a faster or more comfortable trip. These results imply that business travelers would be willing to spend $17 to save an hour of driving time but would be willing to spend $36 to save an hour of flying time. Leisure travelers would be willing to spend $10 to save an hour of driving time and $26 to save an hour of flight time. With regards to frequency, the log construction of this variable becomes apparent here. If a direct itinerary currently only has one flight per day, we see that business travelers would be willing to pay $32 more for an additional flight added per day, but if a direct itinerary already has a frequency of 10, adding an additional flight is only worth $3.16 to a business traveler. Finally, in terms of number of stops: the results imply that a business traveler would be willing to pay $143 for a direct flight over a 1-stop flight and that this person would be willing to pay $220 for a direct flight over a 2-stop flight. We see that business travelers are willing to pay nearly twice as much as leisure travelers for a direct itinerary. TABLE 6-4. IMPLIED MONETARY VALUATIONS -HYBRID CHOICE MODEL WILLINGNESS TO PAY (WTP) BUSINESS LEISURE Value of in-vehicle time for car ($/hr) $17.19 $10.41 Value of access time ($/hr) $36.03 $32.77 Value of in-vehicle time for air ($/hr) $35.86 $25.95 WTP for one additional flight per day for direct flights as base freq of 1 $31.63 $17.20 WTP for one additional flight per day for connecting flights as base freq of 1 $22.93 $16.16 WTP for one additional flight per day for direct flights as base freq of 5 $6.33 $3.44 WTP for one additional flight per day for connecting flights as base freq of 5 $4.59 $3.23 WTP for one additional flight per day for direct flights as base freq of 10 $3.16 $1.72 WTP for one additional flight per day for connecting flights as base freq of 10 $2.29 $1.62 WTP for one additional flight per day for direct flights as base freq of 20 $1.58 $0.86 WTP for one additional flight per day for connecting flights as base freq of 20 $1.15 $0.81 WTP for direct vs 1 stop $143.15 $76.85 WTP for direct vs 2 stops $220.54 $99.87 6(F) NATIONAL APPLICATION OF THE SCENARIO TESTING MODEL MODEL DESCRIPTION A new modeling process was created by the research team to support the projectâs national program of scenario testing. The model was subjected to a calibration/validation task to ensure that the models described above could be applied at the national level to reasonably replicate observed car versus air mode shares for trips of various distances, as well as match the number of
99 air passengers on flights within the continental US to and from a representative sample of airports in different regional markets. The model calibration process is described in the Technical Appendix to this report. The FHWA model used as the base for developing the scenario testing model The model application leveraged the output of an existing national modeling framework for long- distance passenger travel, developed by RSG for the Federal Highway Administration (RSG 2015; Bradley et al. 2016). The mode/airport choice models developed for this project incorporate aspects of consumer preferences, attitudes, and values that are not in the current FHWA model. The new models have been applied using an extension of the FHWA model framework to develop scenarios to understand how different market developments are likely to affect demand for air and auto. This process is designed to enhance practitionersâ understanding of the issues associated with auto as an alternative to air. The structure of the FHWA national long-distance model is shown in the upper portions of Figure 6-2. The FHWA model uses the following inputs: ï· A population with over 100 million households for the entire nation (including Alaska and Hawaii), synthesized at the Census-tract level. ï· A national zone system with almost 5,000 zones based the intersection of county and PUMA geography. ï· Zone-to-zone auto networks, with distances, tolls, and estimates of congested travel times. ï· Airport-to-airport air networks, based on the DB1B and on-time databases, with distances, in-flight times, frequencies of direct and indirect connections, on-time performance, and fares paid by class. ï· Access distances sorted by road to/from each airport from each Census tract and zone centroid. ï· Station-to-station rail networks, based on Amtrak schedules and fare tables. ï· Zone-to-zone intercity bus networks, based on schedules and fare tables from various carriers. ï· Zonal land-use data on attractions for long-distance travel, such as households, employment in the lodging and entertainment sector, employment in various other sectors, university enrollment, and percentage of land area in public parklands. For a given household in the synthetic population, the model first predicts auto ownership based on demographic variables and land-use density near the residence. For each residence zone, the model system then applies a joint model of destination and mode choice that evaluates the accessibility by each mode (i.e., car, air, bus, rail) to each possible âlong-distanceâ destination in the country (zones that are 50 miles or more from the residence zone), and the probability of choosing each destination and mode combination. The next step of the modeling process includes a long-distance tour generation and scheduling model that predicts the number of long-distance tours (round-trips) that a household makes during every month of the year for each of five different travel purposes (business, commuting, visit friends or relatives, vacation/leisure, and âotherâ, which includes purposes such as medical,
100 shopping, and college). For each round-trip, the model also predicts the duration of stay at the destination (0 nights, 1â2 nights, 3â6 nights or 7+ nights) and the travel party size (1, 2, 3, or 4+travelers). Finally, based on the household characteristics and the characteristics of the tours, the model uses the precalculated mode/destination choice probabilities to predict the choice of a specific destination zone and mode used for each tour. (For each disaggregate choice, the model software draws a random number and uses it in âMonte Carloâ stochastic simulationâanalogous to spinning a roulette wheelâwhere the number of slots allocated to each possible choice outcome is based on the model probabilities.) The output of the model system is a list of many millions of individual tours, with the characteristics shown at the right of Figure 6-2 along with the predicted airports or stations used for air or rail tours. Model refinement for the new ACRP Scenario Testing Model. The bottom of Figure 6-2 (the shaded portion) shows the new component that the research team has appended to the national long-distance model framework for this project. The predicted car and air tours from the individual tour list are passed to the new ACRP Scenario Testing Model. Most of the tour characteristics will remain fixed (origin, destination, purpose, month, duration of stay, party size). The ACRP model application only re-simulates the choice between the car and air modes, and the choice of the best itinerary for air trips, using the parameters of the new model. In addition to the wider range of variables used in the model, the main advantage of the new model application is that while the FHWA model uses a single âbestâ air route for each zone-to-zone OD pair, the new model is applied using a full set of relevant airport- airport pairs, as well as Census tract-level accuracy for the airport access and egress distances. (There are over 70,000 Census tracts in the United States, compared to approximately 5,000 zonesâthis represents a substantial improvement in spatial accuracy to be applied to questions of airport access.) In summary, the ACRP model is applied to all tours within the continental U.S. of 100 miles or longer in each direction, for which the mode predicted by the FHWA model is either car or air. The synthetic population used in the FHWA model is drawn at the Census tract level, so the Census tract at the home end of the tour is already known. However, the FHWA model only predicts the destination to the zone level of spatial detail. To use the more detailed Census tract-to- airport access and egress data, the ACRP application uses Monte Carlo simulation to select a destination Census tract within the destination zone. The selection probabilities are based on the same destination attraction function as used in the FHWA data, applied to Census tract values for accommodation employment, entertainment employment, retail employment, service employment, other employment, resident households, university student enrollment, and open space parklands, with the relative weights depending on the tour purpose. 6(G) CONCLUSION Chapter 6 has described the development of a set of advanced travel demand models which were used in the creation and evaluation of a base case, and five overarching scenarios for the future.
101 These scenarios were presented and analyzed in Chapter 1 of this report, and the implications of the process are explored in Chapter 7. FIGURE 6-2. COMBINED FHWA MODEL WITH NEW ACRP AIRPORT CHOICE MODULE