This report documents and presents the results of an evaluation and review of techniques used to forecast socioeconomic variables at sub-county geographic levels of aggregation. This project extends the work documented in NCHRP Report 266. A sensitivity analysis indicated that the final Urban Transportation Planning Process output, link traffic volumes, is sensitive to errors in district level population and employment forecasting. Corridor level studies should thus be aware of the reliability and accuracy of district level forecasts. Local transportation facilities are most sensitive to the errors in the subdistrict (or zone) level forecast of socioeconomic variables. Site-specific studies concerning local road design may be severely affected by data errors at the traffic zone level. Greater attention should be paid to suburban areas where the potential for introducing large input errors is high. Travel demand models must be applied using traffic zone level data, not district level data. The practice of expanding a district-to-district trip table to a zone-to-zone trip table may produce high estimates of vehicle miles travelled and trip volumes on major facilities after the assignment of the expanded trip table. A combination of an intuitive knowledge-based technique, such as the Delphi method, and formal analytical method(s) seems to be the most desirable approach to forecasting sub-regional level socioeconomic variables. Considering the resource requirements and complexity, sophisticated mathematical land use forecasting models seem suitable for large urban areas. Small and medium sized urban area agencies will find simple analytical methods and the Delphi method to be the most preferred techniques. (A)
Abstract