Dynamic Validation at a New Level: Establishing the New Twin Cities Regional Model for Transit Forecasting
Corresponding Author: Patrick Coleman, AECOM
Presented By: Patrick Coleman, Metropolitan Council
Abstract
The Metropolitan Council (Met Council) of Minneapolis-St. Paul has introduced a new state-of-the-practice activity-based regional travel demand forecasting model for use in regional planning and project-level forecasting. The introduction of any new model is inevitably accompanied by a learning process and some surprises. When the last fully new model was introduced in 2004, two years of analysis and re-validation were required before it was accepted for FTA New Starts forecasts developed from that model.
Given those lessons learned the Met Council took a different path with the new model by “dynamically validating” it outside of the project forecasting environment with the following objectives in mind:
• Build confidence in the new model among Council staff and Federal and local partner agencies.
• Understand the differences in expected project-level forecasts under the new model compared to the old.
• Recommend any changes to model parameters, structure, and coding, if necessary, to improve model performance in corridor transit ridership forecasting.
The new model was applied to four test corridors which have previously been forecasted using the previous regional model. Forecast results between the two models are compared with the level of analytical rigor to these forecasts typical to an “FTA New Starts” project forecasting effort to ensure that the new model reasonably forecasts major transit and highway corridor ridership and traffic. A number of refinements were also implemented as part of this dynamic validation exercise resulting from open and positive communication between the project team, the model developers, and the Project Advisory Panel.