The Sacramento Activity-based Travel Demand Model: Estimation and Validation Results

The Sacramento Activity-based Travel Demand Model: Estimation and Validation Results


John L. Bowman, Mark A. Bradley, Independent Researchers and Consultants, US; J Gibb, DKS Associates, US


We present the latest information about the integrated econometric activity-based travel demand microsimulation model being implemented for Sacramento, with results of estimation, sensitivity tests, calibration and forecast validation.


This paper presents the latest information about the activity-based regional travel forecasting model system being implemented for the Sacramento, California, Council of Governments. The system includes an integrated econometric activity-based demand microsimulation model with a highly disaggregate treatment of the purpose, time of day and location dimensions of the modeled outcomes.

Recent planning efforts in the region focused attention on the importance of development patterns at the neighborhood scale. Planning tools currently in use for scenario analysis generate detailed descriptions of development scenarios, providing detailed attributes for each parcel in the region. It is desired to accurately predict in the travel forecasting models the travel impacts of alternative neighborhood scale development patterns, including the effects of things such as increased development density, mixed use development, improved walkability, and convenient transit access, as captured by the current scenario planning tools.

To help achieve this objective, the new model system represents travel in the context of an integrated disaggregate econometric model of each resident's full-day activity and travel schedule. Sensitivity to neighborhood scale is enhanced through disaggregation of the modeled outcomes in three key dimensions: purpose, time, and space. Each activity episode is associated with one of eight specific purposes, and with a particular property parcel on which it occurs. The beginning and ending times of all activity and travel episodes are identified within a specific 30-minute time period.

At ETC 2005 we presented the design of the model system, emphasizing the disaggregate treatment of time, space and purpose. Some estimation results were presented, but not all the models in the system had yet been estimated. As of January, 2006, the demand models have all been estimated and are operational in an application program. Integration with the traffic assignment models, and model system validation have begun. By the time of the ETC 2006 conference, the model system should have been calibrated for a base year of 2000, and validated for forecasts to the year 2005. Accordingly, the paper will provide new estimation results, explain the integration with the traffic assignment model, describe calibration and validation procedures, present empirical results from model validation and sensitivity tests, and discuss the major issues encountered in designing, developing and implementing such a model system.


Association for European Transport