URBAN TRAVEL DEMAND FORECASTING PROJECT PHASE I FINAL REPORT SERIES, VOLUME V. DEMAND MODEL ESTIMATION AND VALIDATION

The project attempts to provide transportation engineers and planners with the information necessary to select and use policy-oriented disaggregate behavioral travel demand models, and to assess the applicability and limits of specific alternative models. This volume is devoted to the investigations of demand, forming the core of this project. Data are collected on a sample of individual commuters in the San Francisco Bay Area before the initiation of Bay Area Rapid Transit (BART) service. BART patronage is predicted from demand models fitted to the pre-BART data. The predictions are compared with actual BART patronage, using a second survey taken after BART was in service. Attention is concentrated on work mode-choice. These studies demonstrate disaggregate travel demand forecasting to be a practical policy analysis tool. The limitations of the current generation of these models are spelled out, and suggest that considerable care is needed in their application to new mode forecasting, and in transferring models across populations.