Topics

• The Big Picture
• Model Input
• Model Development
  • Land Use Forecasting
  • Trip Generation
  • Trip Distribution
  • Mode Choice
  • Assignment
  • Time of Day
  • Freight Models
  • Integrated Models
  • Model Validation
  • Uncertainty and Risk
  • Simplified Approaches
  • Other Special Case Models
    • Seasonal Residents
    • Special Events
    • Vehicle Availability
    • Parking Choice
    • Tourism
    • Taxi Trip Estimation
    • Non-Motorized Models
    • Urban Goods and Services
  • Model Transferability
  • Post Processing
• Model Application

Non-Motorized Models

The objective of travel demand forecasting is to predict changes in travel behavior and transportation conditions as a result of proposed transportation projects, policies, and future changes in socioeconomic and land use patterns. For non-motorized forecasting in particular, the objective is generally to predict the change in the number or characteristics of bicycle, pedestrian, or vehicle-trips as a result of facility improvements or policy changes which are designed to make bicycling or walking more attractive. In addition to affecting overall levels of non-motorized travel, changes in non-motorized travel conditions may affect travel behavior in a variety of ways:

• Trip making. A high-quality walking and bicycling environment is likely to increase total person travel, while a poor quality environment may lead some people to choose not to travel.
• Trip location. A high-quality pedestrian and bicycling environment may cause changes in the choice of destinations, e.g., diverting travel from more distant automobile-accessible areas to closer-by pedestrian-oriented locations.
• Mode choice. Changes in the quality of the travel environment may spur changes not only in the number of people who walk and bicycle, but also decrease the propensity to use public transportation, rideshare, or to drive an automobile.
• Route choice. Changes in the quality of the travel environment may spur changes in the use of various routes by pedestrians and bicyclists.
• Trip scheduling. The quality of the travel environment may vary by time-of-day (e.g., with changes in on-street parking regulations or non-peak-period traffic restrictions) and may also affect trip scheduling of motorized travel. For example, bicyclists may choose to make trips when there is less motor vehicle traffic.
• Land use. Changes in the travel environment may spur changes in land use over a period of several years or more, with some locations becoming more or less desirable for certain types of uses. For example, pedestrian-friendly urban environments may be more attractive, thus increasing development in these areas.
• Distribution of effects. Changes in the pedestrian and bicycling environment are likely to have widely varying effects on different segments of the population. For example, some types of improvements will primarily benefit recreational users while others will benefit those for whom bicycling or walking is the primary means of transportation.

Factors Specifically Influencing Bicycling and Walking

Standard travel demand modeling procedures generally predict total trip-making and mode choice based on a limited number of variables, such as household characteristics and the time and cost of competing modes. These factors, however, only partially explain the decision to bicycle or walk. Development of non-motorized travel forecasting methods requires consideration of a range of factors specific to non-motorized modes. From an individual perspective, personal factors, environmental factors, and trip characteristics interact to determine whether a trip is made by bicycle, foot, or other mode. The specific factors which are important vary depending on whether the mode being discussed is bicycling or walking.

If behavior studies are performed from an aggregate-level perspective, factors must be identified which proxy for the personal and environmental factors seen from the individual's perspective. For example, median income of an area may represent household income, or average vehicle travel speeds and parking costs in a city may serve as a proxy for the time and cost of travel by automobile for a particular trip. Figure 1 presents a framework for how a general set of factors, including facility design factors, interact to affect non-motorized travel levels, both overall and for specific facilities (links) in a network. These factors are described in Table 1 below.

Regardless of whether models are developed at the disaggregate or aggregate level, it is important to remember that decision making ultimately occurs at the individual level and that a forecasting procedure should approximate the individual decision-making process as closely as possible.

Figure 1: Relationship of Factors Influencing Non-Motorized Travel

Table 1: Description of Factors Influencing Non-Motorized Travel

Finally, it should be kept in mind that the factors shown in Table 1 may influence an individual's travel behavior decisions at a variety of stages, not just on a trip-by-trip basis. For example, the individual must first decide to even consider bicycling or walking as a viable travel option. Only when this is done does the question of whether to bicycle or walk for a particular trip become relevant.

Differences in Forecasting Bicycle vs. Pedestrian Travel

Bicycle and pedestrian travel are collectively referred to throughout this guidebook as non-motorized travel, and each class of forecasting methods discussed is generically applicable to both. Nevertheless, significant differences exist between the two modes, both in terms of travel characteristics and factors influencing the decision process. These differences are apparent in the specific examples of the methods, most of which were developed for either bicycles or pedestrians, as discussed in the supporting documentation of this guidebook. Some of the most significant differences include:

Pedestrian trips are generally shorter than bicycle trips. This is important because appropriate analysis methods may depend on the spatial scale of analysis. For example, an analysis of pedestrian conditions may consider every block in a small area, while an analysis of bicycle conditions may focus on through bicycle routes.

A large percentage of pedestrian trips are actually trips to access other modes, including the automobile or transit. Bicycle trips, in contrast, are primarily stand-alone trips (although bicycle access to transit is an important type of non-motorized travel). Therefore, local pedestrian travel will largely result from automobile and transit trips rather than replacing these trips, and modeling transit vs. auto mode choice will be relevant to predicting pedestrian travel. Conversely, pedestrian access factors will be important in predicting transit vs. auto mode choice, since the quality of the environment for walking may influence the decision to use transit.

Perhaps most significantly, the decision to ride a bicycle involves a greater conceptual leap than the decision to walk. Everyone is a pedestrian, but not everyone is a bicyclist. Insights from the public health and social marketing fields suggest that the decision to even consider riding a bicycle is a multi-staged process involving a variety of interacting personal, social, and environmental factors. The choice to bicycle for a particular trip depends not only on the specific characteristics of that trip but on the individual's attitude toward and willingness to bicycle. While attitudinal research gives important insights into pedestrian and transit travel choices as well, its implications are perhaps most significant for bicycle travel.[1]

[1] Adapted from Guidebook on Methods to Estimate Non-Motorized Travel: Overview of Methods (July 1999)