EVALUATION OF THREE INEXPENSIVE TRAVEL DEMAND MODELS FOR SMALL URBAN AREAS

Conventional urban travel demand models, which are data-hungry, costly, and mainly meant for use in large cities and metropolitan areas, are not suitable for small urban areas with a population of 500,000 or less. These small urban areas generally lack the staff, expertise, and budget to operate the conventional models. Three simplified travel demand models are evaluated that are suitable for small urban areas and make use of routinely collected ground counts. These three models are applied in a common setting to the City of Pullman (1980 population 23,579) in the State of Washington. Socioeconomic data and routinely collected ground counts for 1970 were used as inputs to run the models, and the outputs (travel forecasts) were compared with 1980 ground counts, to determine their forecasting capability. All three models tested performed very well. The RMS error ranges between 9 and 15%, and the link volume forecasting capability for most of the links ranges between 10 to 15% of the observed volumes. Contacts with selected planning organizations in the State of Washington reveal that such methods will be useful in small urban areas, considering their staff, expertise, time and budget limitations. Currently, these small urban areas use unproven heuristic methods. The models described in this paper will considerably help small urban areas to forecast travel demands, using routinely collected traffic ground counts and socioeconomic data, with confidence.