This paper presents data developed for FHWA’s Traffic Analysis Framework: origin-destination trip tables for auto, air, and rail for the United States for 2008 and 2040. The data are intended to enhance USDOTs ability to conduct quantitative analysis for infrastructure investment and operational effectiveness needs and to support the needs of states in their statewide modeling efforts. The tables were developed using existing and, where possible, publicly available data describing travel in 2008 and forecasts of growth to 2040. The paper describes the approach and includes visualizations of the results.
The long distance auto trip tables were developed using 1995 American Travel Survey (ATS) data. The ATS was used as it is the most (and only) comprehensive source of long distance travel in the country. Because the level of cost-sensitivity in mode choice varies considerably by travel market in long-distance passenger travel, it is critical to segment the ATS data by trip purpose; business and non-business segments were used. The ATS trips were regressed against variables such as income, population, employment, and under 18 population. The resulting models and an iterative proportional fitting method were used to generate and allocate trips, with the results validated against data from Ohio and California.
The air trip tables were developed by blending three sources of data: Airline Origin and Destination Survey Data (DB1B) and T-100 data that describe air passenger trips between airports, and a collection of airport ground access surveys that describe access trips from trip origins (e.g. homes, hotels) to airports and egress trips from airports to trip destinations. The combination of trip origin to airport, airport to airport, and airport to trip destination describes a complete air passenger trip from origin to destination.
The rail trip tables were developed using a similar approach to the air trip tables, by blending data on station to station trips with data and models describing station access and egress trip distributions. Unlike the aviation market, there are no publicly available station to station datasets. However, Amtrak, the operator of long distance rail services in the United States, made data available to the research team.