Flagstaff Metropolitan Planning Organization (FMPO) Peer Review
3.0 Flagstaff Metropolitan Planning Organization (FMPO) Model Overview
To conduct quantitative assessments of the regional transportation system, FMPO employs a traditional 4-step trip-based modeling framework in TransCAD. The modeling area consists of 340 TAZ over 525 square miles with 8 external stations. FMPO's current TransCAD setup was first developed in 1997 to aid the drafting of Arizona's first regional land use and transportation plan. The update in 2007 introduced a &ldquot;3-D&rdquot; process to estimate demand for pedestrian, bicycle and transit trips based on the trip density of each traffic analysis zone (TAZ), the balance of home-based-work productions versus attractions or diversity of each TAZ, and the design of each TAZ based on the quality of the respective modal facilities. The current TransCAD model was updated in 2013.
This is a traditional 4-step trip-based model producing 24-hour estimates that calibrate to a percent root mean square error (RMSE) of 18%. Though transit trips (estimated at 3% of total personal trips) are not currently assigned, demand between districts of aggregated TAZs showed strong correlation to an on-board origin-destination survey conducted by the transit agency. Overall mode share for pedestrian, bicycles and transit compare well to data gathered in two trip diary surveys (2006 and 2012).
There are five primary trip purposes: home-based work, home-based university, home-based shopping, home-based other, and non-home-based. There are no truck trip purposes. Trip production is based on dwelling unit type (5 types) and trip attraction is based on land use type (60 types) and is not stratified. External trip data are usually obtained from the Arizona DOT Statewide Traffic Model. PM peak period trips are estimated but an extensive calibration effort has not been made.
All steps, provided here in order, are activated with a GISDK scripted interface within the TransCAD environment. Separate model runs may be made for development phasing, development build out, and horizon-year, to name a few.
The model process is summarized below:
- Land Use: TAZ-level future land use data are created by a hands-on land use allocation model that considers regional population control total checked with historical land use growth patterns, local planning knowledge of upcoming developments, housing start data from local jurisdictions, and square feet of non-residential data from the assessor's office.
- Build Networks: in this stage, the link capacities and travel times are set and the highway skims are generated.
- Trip Generation: trip generation is based on rates from the Institute of Transportation Engineers Trip Generation Manual (ITE rates). Some of the ITE rates are modified by FMPO. Housing types (single family detached, single family attached, multifamily, manufactured homes) serve as proxy for income and have distinct trip rates. Trips are balanced to productions with the exception of university trips which are balanced to attractions based on university enrollment. The ITE trip rates produce vehicle trips. The model has been run with a vacancy rate applied to schools, hotels, 2nd homes, and areas known to have high levels of seasonal business to generate summer and fall scenarios.
- Trip Distribution: trip distribution is done using a purpose-specific gravity model and generates a trip length frequency report.
- Mode Split: The vehicle trips are first expanded to represent person trips. Occupancy factors by trip type from the Regional Trip Diary Survey are used together with a general expansion factor. The share of non-auto trips is derived by a process that considers the density, diversity, distance (3-D) as well as the non-auto levels of service and proximity to NAU. The mode split process first splits transit trips from the set of person trips and next splits the bike trips from the remaining person trips. The non-auto trips are removed from the trip table before assignment.
- Traffic Assignment - Auto Only: The remaining person trips are factored to vehicle trips and assigned to the network using stochastic user equilibrium (SUE).
The next version of the model currently under development will have a planning horizon to year 2040 as interpolated from a build out model. Planned improvements include more quantitative methods for assigning pedestrian, bicycle and transit level of service to each TAZ and bicycle trips assigned based on a &ldquot;bicycle comfort index.&rdquot;
The FMPO model is used primarily by FMPO to prepare the LRTP and all financially constrained projects are included in the model network. The model is used by consultants for specific traffic impact analyses (TIA) and has been included to some extent in the new statewide model.
The FMPO model outputs are used in ranking projects in the Transportation Improvement Program (TIP). Forty percent of the project score depends on the model output.
Prior to the peer review meeting, FMPO identified several areas for which they wanted the peer review panel to comment and make recommendations. The areas and specific questions are detailed below:
- An objective assessment of the FMPO's existing travel demand forecasting framework;
- Industry standard procedures and best practices for conducting a TIA and a comparison to the City of Flagstaff TIA procedures;
- Within the context of an acceptable TIA procedure the best practices or means to determine the following and the appropriate use of a regional model for each (with feedback specific to the FMPO regional model, desired):
- Trip generation - Compare and contrast trip rates within the model versus trips rates used at the site development scale (typically ITE Trip Generation Rates).
- Trip distribution - Compare and contrast model generated project trip distributions based on constrained and unconstrained networks versus engineering judgment distributions. Discuss the methodology for estimating background traffic within a model versus the use of one or more growth factors.
- Background traffic - Discuss and define the purpose of background traffic, its estimation and its role in determining the impacts assigned to the proposed project.
- Mode share - Compare and contrast the effectiveness of a regional transportation model versus other methods such as ITE for determining mode splits for a proposed development.
- Pass-by and Internal Capture trips - Compare and contrast how pass-by and internal capture trips are represented, if at all, within a regional model to standard ITE practice.
- Proportional share - what are effective and equitable means of calculating proportional share for improvements related to all modes? What is the role a regional model can play in identifying these?
- Relevant to each stage in the modeling process the agency would benefit from the panel addressing the following questions:
- What are the best ways to apply the different stages of the modeling process to the different expectations out of the TIA process?
- Where are there limitations in the use of the model and how are these best explained to and mitigated by the local jurisdiction and development applicant?
- When there are known errors in the model (i.e., under or over estimates) how are these best explained and mitigated?
- How do other MPO's officially approve of their model performance and how do their member agencies adopt it for their use?
- What are the range of model runs (i.e., with and without the project, current and future case, phases of the project, build out, etc.) needed to most effectively answer questions of trip generation, distribution, assignment and proportional share?
- Do agencies use a different process for regional plan amendments vs. zoning cases?
- Are there agencies that have pre-application processes in place and how well do those work?
- Examples of concept planning using the regional model to explore capacity issues, constraints and network opportunities.
- Examples of detailed capacity analysis of roadways and intersections by the agencies on a regular basis whereby that information is a &ldquot;given&rdquot; for the applicant.
- Examples of detailed master plans for roadways and intersections where the level and types of improvements are pre-determined.
- What are reasonable expectations of the development community to mitigate impacts across a broad geographic area when costs or agency practice prohibit or strongly inhibit addition of capacity in the immediate proximity of the development?
- How can the model assist in identifying potential improvements and their relative effectiveness?
- Are there any agencies that actively test the accuracy of their TIA post-development?
