Updating Existing Travel Simulation Models with Small-Sample Survey Data Using Parameter Scaling Methods

The travel simulation models for many metropolitan areas were originally developed and calibrated with older large-sample travel surveys that can no longer be undertaken given today’s funding constraints. Small-sample travel surveys have been collected as part of model update activities required by the Intermodal Surface Transportation Efficiency Act and the Clean Air Act Amendments. Although providing useful information, these surveys are inadequate for calibrating elaborate simulation models by traditional techniques. Parameter transfer scaling based on small-sample surveys and other secondary source data can be a cost-effective alternative to large-sample surveys when existing models are being updated, particularly when the models tend to be robust and the required changes are relatively small. The use of parameter scaling methods to update the Delaware Valley Planning Commission’s existing travel simulation models is demonstrated. All available sources of data are incorporated into the update process including current survey data, census work trips from the Census Transportation Planning Package (CTPP), transit ridership checks, highway screenline counts, and Highway Performance Monitoring System travel estimates. A synopsis of experience with parameter scaling techniques including the model changes and resulting accuracy is provided. Overall, small-sample-based parameter scaling techniques were judged to be effective. The census CTPP data were evaluated versus the home interview and were found to be useful in the model recalibration effort as a source of small-area employment data by place of work and as a supplement to home interview data for model validation. However, a home interview survey is required as the primary source of travel data for both work and nonwork trips.