U.S. Department of Transportation
Federal Highway Administration
1200 New Jersey Avenue, SE
Washington, DC 20590
Federal Highway Administration Research and Technology
Coordinating, Developing, and Delivering Highway Transportation Innovations
This magazine is an archived publication and may contain dated technical, contact, and link information.
|Publication Number: FHWA-HRT-05-005 Date: May/June 2005|
Publication Number: FHWA-HRT-05-005
Issue No: Vol. 68 No. 6
Date: May/June 2005
These FHWA panels are helping States and metropolitan areas refine their travel demand models to enhance transportation decisionmaking.
|Interstate 25, known locally as the Valley Highway, passes under a viaduct as it heads south toward downtown Denver, a city that uses modeling information when decisions are made on which corridors to develop or expand.|
"Travel models are key tools for making the decisions that shape our transportation system," says Federal Highway Administration (FHWA) Associate Administrator for Planning, Environment, and Realty Cynthia Burbank. "Every year the United States invests billions of dollars in highways and transit, relying on travel models to enable transportation officials to make the highest payoff on that investment. Through peer reviews, FHWA brings State and local planners and modelers together to learn from each other and sharpen these decisionmaking tools."
Today, as community leaders debate growth policies, traffic congestion, urban design, and air quality issues, they need improved information about the potential impacts and tradeoffs of the various transportation alternatives. A partnership called the Travel Model Improvement Program (TMIP) is addressing that need by conducting research and delivering technical assistance and training to transportation planning professionals. The Federal Transit Administration (FTA), Office of the Secretary of Transportation, the U.S. Environmental Protection Agency, and FHWA cooperatively support TMIP. The partnership is helping metropolitan planning organizations (MPOs) and State departments of transportation (DOTs) improve the techniques they use to inform decisionmakers on how growth in population and employment, development patterns, and investments in transportation infrastructure are likely to affect travel, congestion, air quality, and quality of life.
Most MPOs and DOTs use a traditional four-step model of trip generation, trip distribution, mode choice, and trip assignment. These "trip-based" models were originally developed in the 1950s and 1960s to help plan major highway and transit investments. Trip-based models represent each trip—such as an employee's trip from home to work, from work to home, or a side trip, perhaps to a shopping center—as separate trips.
Although these models have been and still are adequate tools for many areas, increasing complexity of travel patterns and severe congestion problems in many urban areas demand more robust and sensitive tools for planning and analysis. Larger metropolitan areas, such as Los Angeles and Denver, are heading toward tour or activity-based models, which model travel differently than trip-based models. Tour-based models, for instance, keep track of travel activity throughout the day and can assemble multiple trip legs (chained trips) into tours. For example, a parent may leave work, pick up the children at day care, and stop at the grocery store on the way home. These separate trips would be linked together into a tour and, when taken as a whole, the modeled travel behavior of this parent would likely be different than if all of these trips were considered separately.
Travel demand models enable planning agencies to ask and answer more complex questions. Keith Lawton, former director of technical services for the Metro Planning Department, an MPO that plans for 1.7 million people in three counties in the Portland, OR, region and adjacent Clarke County in Washington State, notes that, for one thing, these tools can model pedestrian travel as a mode that is responsive to the built environment. "Walking is a significant mode as more trips are made by walking than by transit," says Lawton, "and that's true anywhere in the United States if it's measured." He adds that walking as a mode varies from 3 or 4 percent in the suburbs to up to 25 percent in an urban environment. Metro also has developed and used tour-based models to consider travel over the whole day for studies on road pricing options (that is, toll roads), which are increasingly being considered as congestion management tools for metropolitan areas. Additionally, Metro's ability to link its land-use and travel demand models makes the forecast responsive to the transportation alternatives being considered.
"Federal Highway is concerned with the quality of travel modeling to the point of supporting local agencies and bringing in experts to help each agency evaluate and improve its modeling," says Frederick Ducca, FHWA travel model leader. "In the rapidly changing world of travel modeling, peer review has emerged as the primary tool for improvement."
Between June 1, 2003, and May 30, 2004, TMIP sponsored nine peer review panels held at the following locations:
The nine peer review panels consisted of representatives of MPOs, DOTs, private consulting firms, educational institutions, and environmental interest groups. Frank Spielberg, principal of BMI-SG of Vienna, VA, participated as a peer in the DRCOG peer review panel in Denver. He describes the peer reviewer's role as performing a "honeybee" function. "We travel from agency to agency bringing ideas, concepts, and experiences that are valuable to planning agencies."
The Denver, CO, region, with an area of 12,950 square kilometers (5,000 square miles), has a population of 2.6 million and is expected to grow by another 1 million over the next 20 years. Although DRCOG has used a traditional four-step trip-based model for the past decade and a half, the MPO is moving to a tour or activity-based model with a target completion date of 2006.
|Erik Sabina of DRCOG points out the modeling area in greater metro Denver.|
"The trip-based model and the tour-based model are the great divide right now," says Erik Sabina, travel forecasting group leader, DRCOG.
DRCOG conducted two peer review panels through TMIP. The local technical and policy panels developed a prioritized list of issues they wanted to address through an improved model. The list included such issues as the effects of development patterns on travel behavior and improved sensitivity to price and behavioral changes. Existing models are relatively insensitive to the different travel behaviors people exhibit who live in mixed-use (residential and commercial) versus single-use neighborhoods (residential only). An activity- or tour-based model will better show the extent to which mixed-use neighborhood residents tend to reduce their automobile use by taking transit, walking or bicycling, or accomplishing several activities in one automobile trip in cases where mixed-use development places retail, entertainment, and office locations close together.
|Concrete barriers separate traffic on the south sections of Interstate 25, part of the T-REX megaproject in Denver. Modeling information from DRCOG is used in the environmental impact studies that precede construction decisions.|
Similarly, an improved model will better permit planners to predict trends in travel behavior due to shifts in population composition. Current models do not separate the elderly from younger population groups, though it is well known that the elderly make very different travel choices than do younger people, and it is widely projected that the percentage of elderly people in the population will increase significantly in the coming decades. A disaggregate tour model will do a better job of showing the extent to which the elderly make different choices than younger people, both in where they live and the extent to which they take transit instead of driving, as they become less comfortable with operating an automobile.
In the late 1990s, DRCOG completed a travel behavior inventory, a $1.5 million initiative that included a demographic survey of 5,000 households. Conducted by a consulting firm, the survey recorded trips of all family members during an assigned fall weekday. The consultants recruited participants randomly; sent out diaries for each household member, including children, to record trips on the assigned day; and called participants to review the diaries and correct any discrepancies.
"How you gather data is a major issue in modeling," Sabina stresses. "The survey needs to randomly survey, for example, both the rich and the poor, so the sample is representative of everyone traveling, both in vehicles and on public transit. One of the basic fears, if you get a low response rate, is whether it's a self-selected sample, such as only those people who are interested in transportation."
After building a trip-based model based on the survey, DRCOG staff validated the model with data received from the Regional Transportation District and from traffic counts by municipalities and the Colorado Department of Transportation (CDOT).
|In Colorado, CDOT planners used modeling information from DRCOG in decisions on the T-REX corridor alignment and expansion and in the environmental impact studies that preceded construction. This photo shows a lightrail station being built at Louisiana Street as part of the T-REX project.|
|The barricades and dirt in the middle of Colorado Boulevard, a major commercial thoroughfare, allows for construction of the light-rail tunnel. Here, the light rail splits from the highway and dives under Colorado Boulevard to a belowground station. The contractor has shifted traffic of the three lanes either direction to cross this boulevard with the open trench method.|
|Looking south from Steele Street Bridge toward Colorado Boulevard. Note the light-rail tracks along the highway.|
With a population of 8.2 million people, North Carolina covers approximately 126,172 square kilometers (48,715 square miles) and has 17 MPOs and 20 rural planning organizations (RPOs). Nine of the 17 MPOs are included in three large regional travel demand models. Within the last 3 years, NCDOT has created a Model Research and Development Unit to support these models and other modeling activities across the State.
NCDOT is also a leader in travel demand modeling for MPOs in small urban areas. Engineers have long used travel demand models as an integral part of the development of long-range comprehensive transportation plans. Recently, however, the questions that need to be addressed by travel models have become more sophisticated, increasing the complexity of the models.
|NCDOT used modeling software to analyze the recent widening of Interstate 40 (shown here) through Durham and Wake counties in the Raleigh, NC, area.|
"A new North Carolina law requires long-range plans to be multimodal," says Rhett Fussell, leader of NCDOT's Model Research and Development Unit. "Even for areas of 5,000 in population or less, we're now developing a comprehensive transportation plan that considers transportation by automobile, walking, bicycles, and public transportation. This new requirement has significantly expanded our role within the department, necessitating that we take all types of transportation into consideration when we develop travel models and transportation plans."
The presence of numerous rural areas throughout the State makes it challenging for many North Carolina communities to evaluate multimodal options. Therefore, the TMIP panel recommended implementing other types of multimodal analysis for areas that do not have travel models, and NCDOT is now in the process of developing a multimodal toolbox that will handle these issues without having to rely solely on travel demand models.
According to Fussell, given the recent advances in travel demand modeling, the department is reassessing its practice of employing generalists to work with the travel models. NCDOT prefers to assign specialists to provide quality control for developing and running the new models. Fussell's current staff of six specialists is working to establish best practices guidelines and standard procedures to support the modeling activities of NCDOT and MPOs throughout the State. To provide assistance for this relatively small staff, NCDOT is also following the peer review panel's recommendations to use consultants to support critical functions and is developing research projects with local universities.
To help address modeling issues and develop modeling tools that meet the needs of the entire State, a statewide model users group composed of representatives from NCDOT and MPOs, as well as consultants, has been formed to exchange ideas, generate new concepts, and develop a strong partnership across North Carolina. "We remain confident that NCDOT's current efforts, along with the implementation of the other recommendations from the TMIP peer panel, will continue to improve travel demand modeling within North Carolina and help the department better meet the comprehensive transportation planning needs of its citizens," says Fussell.
The Anchorage, AK, region has a population of 270,000 and covers an area of approximately 4,400 square kilometers (1,700 square miles). AMATS uses a traditional four-step trip-based model. The primary focus of the May 2004 peer review panel was to review the current status of the AMATS travel model improvement process and provide guidance on near-term and future model development.
|This van is on the paratransit service (AnchorRIDES) operated by the municipality of Anchorage, AK.|
Among other achievements, AMATS performed a survey of trip data records to better understand major trip patterns and link home-based trips through intermediate stops to their primary destination. In addition, AMATS developed a new set of household disaggregation models for trip generation and mode choice. Unlike the original model, mode choice is performed separately for different residential market segments. Further, management and handling of highway and transit networks was completely revamped. For highway networks, separate travel time data are maintained for peak and offpeak periods. For transit, provision was made for submodes such as local and express buses. AMATS also restructured mode choice models to provide more sensitivity to transit and nonmotorized options by adding measures of walk accessibility, land use composition, and density.
AMATS is implementing a major recommendation from the peer panel to conduct further sensitivity tests. The additional sensitivity analysis will use more than one land use scenario to test how well the land use variables used in the four-step model reflect changes in land development patterns.
"Like most MPOs our size," says Jon Spring, senior transportation planner, AMATS, "we depend to a large extent on consultants to develop our models so the peer review gave us an independent review of our consultants' work and also helped assure our public that the model design is appropriate for our area."
ARC serves as the MPO for an 18-county region as of 2003. The core 10-country region of more than 7,720 square kilometers (2,981 square miles) in the Atlanta area has a population of approximately 3.5 million. Atlanta's rapid growth continues, with population and employment expected to nearly double by 2030. "That's equivalent to adding a Portland, OR, or two cities the size of Jacksonville, FL, to our area," says Guy Rousseau, senior principal, ARC Transportation Planning Division.
ARC implemented its current transportation model with input from peer review panels in 1995, 2000, and 2004. As people move farther from the city yet continue to commute into Atlanta for work and activities, the MPO is finding that it must expand the boundaries of the modeling area to include 20 or more counties. The MPO has set an 18-month target date of mid-2006 to ready, calibrate, and code the expanded model.
|The increasing need for the Atlanta region’s travel demand model to address complicated policy choices such as vehicle emissions and air quality is a major impetus for the Atlanta Regional Commission to use the TMIP peer review to help improve its model. Shown here is the Atlanta skyline.|
Model improvements over the past decade generally have been incremental approaches designed to produce a travel demand model that successfully addresses all Federal planning and air quality requirements and sufficiently represents all transportation modes in the Atlanta region. Travel demand modeling is a constantly evolving field; as data collection efforts improve and research within the field expands, travel models and the transportation plans they support also must evolve. Improvements currently being implemented by ARC include better representation of travel by the time in which it occurs and improvements in predicting vehicle-related emissions.
Iowa has nine MPOs, each with its own model, and has a population of 2.9 million and about 140,000 square kilometers (57,853 square miles). In March 2004 the peer review panel shared expertise on best practices for travel model calibration and validation with IaDOT.
"We're headed in the right direction, but also know we have a lot more to accomplish," says Phil Mescher, IaDOT traffic forecasting and modeling team leader.
IaDOT's modeling team has established effective partnerships with the peers of eight MPOs and looks forward to working with the new Ames MPO. One of the next steps is to develop a framework that all the MPOs can follow to give consistency to model development, maintenance, and use. To address the problem of staff turnover in MPOs, the State agency is assembling a team of in-house modelers to assist and train new MPO staff in modeling. This team will help the MPOs maintain regular modeling operations if a key modeling person leaves the MPO, according to Mescher.
|This 42nd Street interchange reconstruction is one of a group of projects to revamp Interstate 235 through the Des Moines, IA, metropolitan area. The 5-year undertaking is estimated at $429 million with completion set for 2007.|
The peer panel recommended that IaDOT develop a statewide program framework for best practices and that the State draw on similar programs underway in other States such as Florida, Michigan, North Carolina, Ohio, and Texas. In addition to statewide standards, IaDOT is also developing a statewide travel demand model and eventually will incorporate freight into the equation.
OKI, the MPO for an eight county region spanning the three States of Ohio, Kentucky, and Indiana, held a peer review panel in June 2003 in Cincinnati, OH. The OKI region has a population of approximately 1.9 million people and covers 6,713 square kilometers (2,592 square miles). The region conducted peer review panels on OKI's travel demand model in 1994, and OKI implemented a panel recommendation by expanding the coverage of its travel demand model to include an adjacent region.
The major purpose of the 2003 TMIP peer review was to assess the ability of the models to address the planning issues of importance to the region, the reliability of the forecasts produced by the models, the consistency of the model set with the current state-of-the-practice, and further necessary enhancements.
The peer panel noted that there has been significant progress on the model procedures to meet Ohio DOT's standards for traffic assignments. In addition to improvements to some model elements, the panel recommended that the MPO develop a regional forecasting capability for socioeconomic data. Finally, the panel recommended using tour models to better simulate trip-making behavior.
SCAG serves as the MPO for six southern California counties with a population exceeding 16 million people and an area of more than 98,420 square kilometers (38,000 square miles). During the 1990s, SCAG embarked on the first comprehensive overhaul of its travel simulation capabilities since the mid-1970s. As part of its past efforts to update its travel demand model, SCAG held a peer review panel in January 2002 to utilize new data to update and recalibrate its travel simulation modeling process.
SCAG is updating its models with recent data and adding capabilities to the models for improved accuracy and policy sensitivity. For instance, the vehicle availability model has been expanded to account for all vehicles beyond automobiles owned by households. The trip generation and mode choice models have been expanded to include greater details on the purposes of trips, including the identification of intermediate stop locations between work and home, and the addition of Metrolink, high speed rail, and toll roads as separate modes. In general, the models also have been enhanced to better capture characteristics of households and workers, including the age of the head of the household. The increased refinement in the level of details within the models allow for improved ability to account for variability in factors that may comprise a given set of policy options.
A variety of data sources have been and will be integrated into the model update. The trip generation models have incorporated data collected through a global position system survey of households. In the near future, the models will be updated using the following additional data sources: a regional household travel survey, transit on-board origin-destination surveys, a regional cordon survey, a street/highway inventory survey, and an arterial speed study.
Individual reports on the issues and recommendations discussed at each peer review panel are posted on the TMIP Web site at http://tmip.fhwa.dot.gov/services/peer_review_program/status.stm. In addition, this Web site contains a synthesis report providing an overview of the key themes of the TMIP peer reviews held from June 2003 to May 2004. Proceedings for peer reviews held after June 2003 are also on the Web site. These include the Baltimore Metropolitan Council; Memphis Metropolitan Planning Organization; Metropolitan Transportation Commission, Oakland, CA; and Southeast Michigan Council of
Governments in Detroit.
Although the discussion at each panel review was unique to the individual region, many common themes emerged. Some of the highlights of the recommendations follow.
Managing the modeling process and results:
TMIP shares many of its objectives with other stakeholder groups and national organizations. TMIP works cooperatively with the U.S. Department of Transportation's (USDOT) Transportation Planning Capacity Building Program, which is focused on enhancing the capabilities of State and local transportation staffs to meet planning requirements and needs. TMIP also works with the Association of Metropolitan Planning Organizations and the American Association of State Highway and Transportation Officials.
Looking to the future, TMIP has funded the development of the Transportation Analysis and Simulation System (TRANSIMS) at the Los Alamos National Laboratory. The design for TRANSIMS is based on requirements in the Intermodal Surface Transportation Efficiency Act of 1991 (ISTEA), the Transportation Equity Act for the 21st Century (TEA-21), and Clean Air Act Amendments (CAAA) of 1990. TRANSIMS is designed to meet the needs of State DOTs and MPOs for more accurate and more sensitive travel forecasts for transportation planning and air quality analysis.
Michael Culp, a transportation engineer with FHWA's Office of Environment and Planning (HEP), manages training, technical assistance and outreach for the TMIP Peer Review Program. He has focused on training in his 11-year tenure with FHWA and has served on the Transportation Research Board's Committee on Transportation and Land Development since 2000. He has a bachelor's degree in civil engineering from the University of Arkansas and is currently pursuing a master's in transportation policy, operations, and logistics from George Mason University.
Esther J. Lee, a program and policy analyst for the Volpe Center, USDOT, Research and Special Programs Administration, manages the coordination and documentation of peer reviews for the TMIP Peer Review Program. In her 4-year tenure at Volpe, she also has coordinated peer exchanges and authored reports on metropolitan and rural transportation planning best practices for the Transportation Planning Capacity Building Program (FHWA and FTA). She earned her master's degree in city planning from Massachusetts Institute of Technology and a bachelor's in government from Dartmouth College.
For more information, contact Esther Lee, 617–494–3130, firstname.lastname@example.org.