Southeast Florida Transportation Council (SEFTC) Peer Review
5.0 Peer Review Panel Recommendations
On the last half day of the meeting, the peer review panel spent about one-and-one-half hours in an executive session, closed to all other participants of the meeting The reason for this closed session was to allow panel members to speak freely and openly among themselves while developing formal recommendations. This section details those panel recommendations
In response to SEFTC's request for guidance on best-practice approaches to assess SERPM 7.0's performance, as outlined in Section 3.2, the panel made the following recommendations:
- Perform the same traditional checks as you would for a trip-based model. Refer to NCHRP Report 716: Travel Demand Forecasting - Parameters and Techniques and NCHRP Report 365: Travel Estimation Techniques for Urban Planning for further guidance.
- Compare SERPM 7.0 constants and coefficients to other Florida models, and compare the overall transfer experience of other regions, such as the experience of transferring DaySim from Sacramento, CA to Jacksonville, FL and then DaySim and parts of FAMOS to Tampa, FL.
- Perform further validation including before-and-after studies, run the model for year 2015 and compare to traffic counts and travel time data, and perform further tests on the travel time and cost elasticity of the model as described in Todd Litman's report titled Understanding Transport Demands and Elasticities How Prices and Other Factors Affect Travel Behavior.
In response to SEFTC's request for guidance on short-term, medium-term, and long-term data collection efforts, as outlined in Section 3.2, the panel made the following recommendations:
- Before creating a data collection plan lay out a long-term vision for SERPM that includes the policies that SERPM needs to address in the near and long-term and the functional capabilities of SERPM to handle those policy analyses.
- The current budget of 1.5 million dollars is not enough to pursue all of the data collection efforts that SEFTC desires. Thus, it is important to prioritize data collection based on the policies SEFTC wants to concentrate on in the medium-term versus long-term.
- The short-term priority for data collection should focus on collecting data to help with the more traditional calibration and validation efforts. This effort would include collecting traffic count and travel time data, and origin-destination data (i.e. purchasing of cell-phone data). This data collection effort could be undertaken with the existing 1.5 million dollar budget.
- The second priority of data collection should be to collect data that aids SEFTC in quantifying and understanding specific regional travel markets such as collecting data to understand the magnitude of visitor and seasonal residents or truck trips. In the short-term cell-phone data could help quantify these markets. Then, SEFTC could undertake a detailed travel pattern survey to further understand the travel patterns of these markets. A targeted survey could focus on understanding different value-of-time groups such as a survey of toll-users. With a little bit more budget, SEFTC could perform these targeted surveys.
- A full regional household travel survey cannot be undertaken without a significant increase in data collection budget. If more budget can be obtained, SEFTC should consider undertaking data collection efforts that focus on collecting data to help re-estimate the location choice models and undertaking a broad establishment survey with more than 20 establishments sampled. Finally, if a full regional travel survey is undertaken it should be a 100% GPS-based survey with prompted recall validation, which has a higher price per survey than traditional household survey efforts.
- Consider moving to ongoing versus an episodic data collection strategy for all data collection efforts.
The peer review panel made several additional recommendations to SEFTC related to improvements of SERPM:
- Develop a task-force for figuring out how to develop employment estimates for both the base and forecast year. Employment should be developed at the regional level rather than by each MPO separately.
- Consider replacing the regional truck model with the statewide freight model for modeling truck travel.
- Put effort into improving model run times and other strategies to ensure efficiency of model application.
- Update the calibration effort by scaling model parameters rather than just modifying model constants as discussed in TMIP's report Guide for Travel Model Transfer.
- Given SERPM currently has a disconnect between disaggregate time-of-day demand model coupled with an aggregate (4 time-period) supply model, improvements should be made to the supply-side assignment model, such as:
- Moving to finer time periods;
- Developing dynamic intersection models; and
- Using more disciplined network coding practices to create a more realistic network.
