Skip to content U.S. Department of Transportation/Federal Highway AdministrationU.S. Department of Transportation/Federal Highway Administration
Office of Planning, Environment, & Realty (HEP)

Peer Review Process Guide: How to Get the Most Out of Your TMIP Peer Review

Appendix B. Sample Peer Review Application


Request for TMIP Peer Review

Contact: Agency Representative
Phone Number
Contact E-mail address

1. Purpose of TMIP Review

The agency is the Federally designated Metropolitan Planning Organization (MPO) for the metropolitan area. It has served as the MPO for the area since 1973. As the region has grown, the Metropolitan Statistical Area (MSA) expanded to the current five counties. For the agency's 2035 Plan Update that is under development, the modeling area has been expanded to the five counties in the MSA.

To support the agency's decisionmaking process, the modeling department maintains and applies a regional travel model for the long-range plan updates. The current validated model (1997 model) is a traditional four-step model for a 24-hour period. Additionally, the agency has a two-hour AM peak model that was developed based on diurnal factors from the 1997 household survey. We are currently in the process of developing a 2005 base year model that covers the full MSA and will be using survey data collected in 2005 and 2006. These surveys include external station, commercial vehicle, workplace, airport, and household surveys. Additionally, the main transit agency performed an on-board survey in 2005. The agency is improving its model by implementing a speed feedback loop-to-trip distribution, making toll a route choice decision in traffic assignment, and adding a demographic allocation tool for dual track growth scenarios for the 2035 Plan Update, scheduled for adoption in June 2010.

With the advancements being made in the transportation modeling industry, agency staff feel that it is time to have a peer review to assess the processes that have been implemented in the agency model stream and make recommendations for the next model update. Additionally, with the growth of our region and completion of toll roads, it is important to the agency and its member jurisdictions that we move forward with the model in an appropriate direction and make the best use of the limited funds available for updating and running the model for future planning and air quality efforts.

2. The Region

The agency's five county modeling area was estimated to have 1,459,000 people and 698,000 jobs in 2005. The agency board in 2004 approved the population and employment totals to be 3,251,000 and 1,643,000, respectively, for 2035. The five-county regional daily vehicle miles of travel (VMT) reported by the state DOT in 2005 was 37,261,000. This includes on- and off-system roadways. The regional VMT is expected to grow by about 2.7 percent per year. The agency modeling area is comprised of 1,413 internal traffic analysis zones (TAZ) and 49 external stations.

3. The Models

The first regional travel model for the metropolitan area was developed in 1984 by a consultant for the 2020 Plan. The first (1980) model was a four-step model. The city and state departments of highway and public transportation (future state DOT) staff jointly recalibrated the 1980 model to develop the 1985 model in TRANPLAN. The 1985 model contained eight trip purposes for a 24-hour period with a total of 635 TAZs. The mode choice included single occupancy vehicle (SOV), carpool, local and express bus, and fixed-guideway transit.

The 1997 model was jointly developed by and the state DOT in TransCAD. The modeling area expansion resulted in an increase in the number of TAZs to 1,074. A three-way cross-classification model using workers per household, household size, and median income was developed for three home-based work trip purposes: home-based work (HBW) direct, HBW strategic, and HBW complex. The 1997 model had a total of 15 trip purposes and 10 modes of travel. Toll travel was considered a mode in the nested logit model. The 1997 model was applied twice for the agency's 2025 and 2030 Plan Updates. The four-step procedures were performed in a combination of FORTRAN programs and TransCAD functions. As stated earlier, several improvements are being developed or added for implementation in the 2005 model.

The current (2005) model being developed is implemented in TransCAD 5.0. In trip generation and trip distribution, it uses TripCAL5 and ATOM2. These programs are used by the state DOT-TPP, which does the modeling for the majority of the modeling for the MPO in the state. These two programs are FORTRAN based, but the agency has used GISDK code to set up the input files from both static file locations and to call from the TAZ and network geographic files that are in TransCAD format. The main mode choice program is a replication of a former FORTRAN program that the agency's consultant has converted into TransCAD. We have progressed from our previous model in using a barrier file (used to identify areas that were not crossable because, for example, it would have crossed a lake) and straight line transit connectors to using the highway network for transit access. With this implementation, we have also updated to a process of creating ¼ mile and ½-mile radius for transit access. This process is now automated and will be updated with each model run. Previously, this was a manual process. Also being implemented is a feedback loop, as stated earlier, using a successive measure of averages. The setup includes the options of converging either at a certain number of iterations or a predetermined value. Previously, the mode choice model determined the toll-eligible trips. The toll mode has not been removed, but those trips are added together with the non-toll trips; and this decision is now being made during the assignment run. In addition to the daily model run, we will also have an AM and a PM peak model runs. These peak-period runs will largely be developed based on the diurnal factors from the 2005 household survey.

The 2005 model will have the following 15 trip purposes: home-based work complex, home-based work direct, home-based work strategic, home-based non-work retail, home-based non-work other, education-1 (school), education-2 (university), non-home-based work, non-home-based other, non-work airport (ABIA), truck/taxi, non-home-based external local, external-local auto, and external-local truck. In trip generation, the three home-based-work trip purposes use a three-way cross classification using household size, median family income, and workers per household. All other purposes use a two-way cross classification of household size and median family income. All purposes except for education-2 are balanced to productions. In trip distribution, a gravity model is used. In this step, all purposes except education-2, UT, and ABIA are balanced to productions. The highway assignment is set up as a capacity-constrained, multi-class user equilibrium assignment method. A route choice function for toll traffic will be implemented for forecast years.

The agency has undertaken another new effort on developing a replicable and quantitative method for demographic forecasting along with the development of the 2005 model. Previously, a modified Delphi method was used for the allocation of county control totals to TAZs. With assistance from the agency's consultant, a menu-driven program using GISDK code was implemented for allocating both population and employment to TAZs. The tool will facilitate the forecasting work of dual tracks of growth scenarios for the agency board to select a preferred scenario for the 2035 Plan update.

4. Plans for Model Improvement

The agency's overall goal for model improvement and motivation for seeking a TMIP peer review is to continuously maintain and apply a model that is representative of the state of the practice in travel demand forecasting and equips the agency, its policy board, and local jurisdictions the support that is needed for our rapidly growing region. Questions that the agency staff would like to have answered through this peer review are:

  1. Additional improvements to the current feedback loop
  2. How can our model keep up with time-of-day questions?
  3. Do peak spreading and induced travel need to be emphasized more? If so, how?
  4. How can our mode choice model be improved?
  5. Assessment of the reasonableness of toll traffic forecasting
  6. Visualization techniques for better communications to the board and public
  7. Should we consider a micro simulation model and how can our current model assist with that implementation?
  8. Are there better techniques for handling trip generation and trip distribution for the size and diversity of our region?
  9. How can the effects of gas prices or parking cost be better implemented into the agency's model stream?
  10. What are ways that the agency's model can be improved to better answer questions related to environmental justice?
  11. Should we start a freight model? If so, how?
  12. Should we consider transitioning into an activity-based or tour-based model? If so, should we maintain parallel tracks of modeling?
  13. If either of the two model types from the previous question is recommended, how will this affect air quality modeling?
  14. What recommendations are there for improving the current demographic allocation tool or for pursuing a new land-use forecasting model for alternative growth scenario testing?
  15. Are there improvements that could be implemented immediately with limited funds (under $120K) or implemented in a two to three year period for an overall model development?

5. Response to TMIP Selection Criteria

The agency's area is a diverse and rapidly growing region. The members of our local jurisdictions and of our policy board understand the changing dynamics of the region and the need to maintain and update the agency's model to be able to answer the ever-growing questions that are being asked of the model. The agency's staff are in the process of bringing the full model in-house. Currently, the state DOT-TPP maintains control over trip generation and trip distribution and works cooperatively on the traffic assignment phase. At this point, the agency's staff also retain a consultant to assist with the development of the mode choice model. Maintenance and applications of the developed model are performed by the agency's staff. The streamlining of the four-step modeling process at the agency's office was one of the recommendations made in the agency's peer review done in 2001.

Model improvement activities, as well as regular communication, are kept with the local jurisdictions through modeling group meetings. These meetings are held periodically depending on the needs of the modeling group and that of the agency's staff.

The agency has previously demonstrated its commitment to support innovative model techniques with the integration of the speed feedback loop. This was recommended from a previous review of the agency's modeling efforts.

Currently, the agency has funding identified in its FY 2009 Unified Planning Work Program (UPWP) for model improvements. These funds are intended to be used to help implement some of the recommendations from this peer review. Also, the agency staff are very interested in completely updating the mode choice model during the 2010 update. This will be a very beneficial time for this process since, for the first time, we will have local data for calibration of toll roads and a commuter rail line that have or will come on-line in 2006 and 2009, respectively.

6. Proposed Panel and Availability

Due to the agency staff's time constraints with ongoing work, we request the assistance of TMIP staff to put together the members of the peer review panel. The make up of this panel will hopefully consist of the following types of backgrounds: mode choice specialists, toll traffic forecasting, land use modeling integration, and tour-based/activity-based modeling.

7. Schedule and Cost

After researching, it appears that the peer review panel will have sufficient time to accomplish the goals of the committee with a one-day meeting. Agency staff proposes that the Peer Review Committee members meet with agency staff and discuss their findings during the mid-May to early July 2009 time frame.

Agency documentation will be provided to the panel members prior to the meeting, with sufficient time for them to go over the material and ask any questions of agency staff before the panel gathers in the suggested time frame. On the day of the meeting, agency staff will provide a quick introduction to the model and its processes during the morning. Panelists will be encouraged to ask agency staff for specific topics for this presentation. The afternoon will be reserved for the panelist discussion that will form the basis for the final documentation. The final report will provide recommendations for short- and long-term steps with which agency staff should proceed.

The agency requests that our consultant staff attend the peer review to evaluate the process and assist in documenting the review.

The estimated cost for travel, lodging, and per diem for the peer review panel is $4,190 as shown in the following budget.

Task Description

Peer Review Panel Expenses - 5 Members Per Estimate Total
Travel $500 $2,500
Lodging (2 nights) $117 $1,170
Miscellaneous $104 $520
Grand Total $4,190
Updated: 3/25/2014
HEP Home Planning Environment Real Estate
Federal Highway Administration | 1200 New Jersey Avenue, SE | Washington, DC 20590 | 202-366-4000