Transport Model Improvements
Improving Methods for Evaluating The Effects and Value of Transportation System Changes
~~~~~~~~~~~~~~
Victoria Transport Policy Institute
~~~~~~~~~~~~~~~~~~~~
Updated
22 July 2008
This chapter describes various ways to reform transport models to increase their accuracy when comparing modes and evaluating TDM strategies. Current models tend to be biased in various ways that exaggerate the benefits of roadway capacity expansion and undervalue the impacts and benefits of TDM strategies that encourage use of alternative modes and reduce motor vehicle travel.
Models are simplified descriptions of a system used to predict and evaluate the results of system changes. Various models are used to predict impacts and evaluate options for Transport Planning and Evaluation. The assumptions and analysis methods used in these models can affect planning decisions. Commonly used models tend to undervalue alternative modes and TDM solutions in various ways, so model improvements can help support TDM implementation.
Several types of models are used for transport planning:
Traffic models are designed to predict roadway traffic volumes and traffic problems such as congestion and pollution emissions. Most are four-step models, meaning that they follow these steps:
1.
Trip generation. Predict total trips that start and end in a particular area (called Traffic Analysis Zones or TAZs), based
on factors such as each zone’s land use patterns, number of residents and jobs,
demographics, transportation system features (number of roads, quality of
transit service, etc.), and distance between two zones.
2.
Trip distribution. Trips are distributed between pairs of zones, based on the distance
between those zones.
3.
Mode split. Trips
are allocated among the available travel modes (usually auto and transit).
4.
Route assignment. Trips are assigned to specific facilities included in the highway and
transit transportation networks.
These models use travel survey and census data to establish baseline conditions and identify trends. Trips are often predicted separately by purpose (i.e., work, shopping, other) and then aggregated into total trips on the network. Because they are designed primarily to identify congestion problems they mainly measure peak-period motor vehicle trips on major roadways. The generally report roadway Level-of-Service (LOS), which is a letter great from A (best) to F (worst) that indicates vehicle traffic speed and delay.
These models often incorporate several types of bias favoring automobile transport over other modes and undervaluing TDM strategies (TRB, 2007). The travel surveys they are based on tend to ignore or undercount Nonmotorized travel and so undervalue nonmotorized transportation improvements for achieving transportation planning objectives (Stopher and Greaves, 2007). Most do not accurately account for the tendency of traffic to maintain equilibrium (congestion causes travelers to shift time, route, mode and destination) and the effects of generated traffic that results from roadway capacity expansion, and so tend to exaggerate future congestion problems and the benefits that result if roadway capacity is expanded. They are not sensitive to the impacts many types of TDM strategies have on trip generation and traffic problems, and so undervalue TDM benefits.
Economic models are used to evaluate and compare the value of particular transportation improvements, such as the benefits of widening a roadway, improvement public transit service or implementing a TDM strategy. They compare various categories of benefits and costs. They tend to consider a relatively limited set of benefits, since most of these models were originally developed to evaluate roadway improvement options and generally assume that total vehicle mileage is constant, and so are not well designed to evaluate the full benefits of TDM strategies that reduce automobile trips. For example, these models often ignore parking and vehicle ownership cost savings that result when travelers shift from automobile travel to alternative modes, and they generally ignore the safety benefits that result from reductions in total vehicle mileage.
These models are designed to predict how transportation improvements will affect land use patterns, for example, the location and type of development that will occur if a highway or transit service is improved. They are often integrated with traffic models. These are considered the best tools for evaluating transportation policies and programs because they can measure Accessibility rather than just mobility, but are costly to develop and complex, and so may be difficult to apply, particularly for evaluating individual, small-scale projects (Dong, et al, 2006). Some models predict how particular land use factors, such as density and mix, affect travel behavior, and their impacts on congestion and pollution emissions (Donoso, Martinez and Zegras, 2006).
This newer approach models the behavior and needs of individual transport users (called agents), rather than aggregate groups, which improves consideration of modes such as walking and cycling, the travel demands of non-drivers, cyclists and the disabled, and the effects of factors such as parking supply and price, transit service quality, and local land use accessibility factors. Simulation models can provide a bridge between other types of models, since they can incorporate elements from the conventional traffic, economic and land use models. Simulation models have been used for many years on individual projects, and are increasingly used for area-wide analysis.
Some special models have been developed to help evaluate particular types of TDM programs, such as the Commuter Model (USEPA, 2005), the TRIMMS model (USF, 2006) and which can predict the effects of Commute Trip Reduction programs on commute travel behavior.
Price
elasticities are defined as the
percentage change in consumption of a good caused by a one-percent change in
its price or other characteristics (such as traffic speed or road capacity).
For example, an elasticity of -0.5 for vehicle use with respect to vehicle
operating expenses means that each 1% increase in these expenses results in a
0.5% reduction in vehicle mileage or trips. Economists have collected
information on transportation price elasticities, including how changes in
transit service quality and fares affect transit ridership (Litman, 2004; DfT,
2006), the effects of changes in parking fees, fuel price and road tolls on
vehicle travel (Litman, 2006), and information on how various Land Use Factors Affect Travel Behavior.
Because conventional models primarily measure motor vehicle travel rather than Accessibility they tend to undervalue alternative modes and alternative ways of improving accessibility, such as Smart Growth land use (Loudon and Parker, 2008). This is both inefficient (since it undervalues what actually may be the most cost effective way to improve transportation) and Inequitable, since it tends to overlook travel activities and needs of non-drivers.
Travel models tend to focus on quantitative factors (travel speed, operating costs and crash rates) and undervalue qualitative factors such as travel convenience, comfort and security (Litman, 2007a). Conventional traffic models often use simplified travel time cost functions which assumes that any shift from driving to an alternative mode increases travel time costs. This is wrong for two reasons. First, alternative modes are sometimes as fast as driving. Cycling is often as fast as driving for short trips, door-to-door. Ridesharing and transit are sometimes faster than driving with grade separated systems or HOV Priority. Second, travelers sometimes prefer using alternative modes even if they are slower than driving, because they are less stressful or enjoyable (particularly walking and cycling). This tends to favor higher speed modes, such as automobile travel, and undervalues improvements to alternative modes.
Most large Metropolitan Planning Organizations (MPOs) run their travel models with Full Feedback, meaning all model steps are run until the model equilibrated (results in each step no longer change with more iterations). This allows models to indicate how congestion affects trip lengths. This is an important and simple model improvement that helps predict the induced travel effects of adding or expanding roadways. This satisfies the National Environmental Policy Act (NEPA) requirements. This is demonstrated by showing very little change in some output, between model runs N and N+1, called the Convergence Criterion. Some official Federal software, including STEAM (which must be used for new rail starts analyses) have induced travel factoring in them. The U.S. Clean Air Act Air Quality Conformity Rule requires this in regions with Severe and worse air quality ratings. Most model shells (software packages) can now do this, and this type of modeling is common practice in good MPOs, but many MPOs perform too few model iterations to achieve full equilibrium.
To show effects on trip generation (number of trips per day per household) models require an Auto Ownership step at the front end, that is, the model must be upgraded from four to five steps. Many MPOs have done this. Reduced automobile congestion tends to induce slightly higher vehicle ownership rates which slightly increases Trip Generation. This effect is generally small.
The third model improvement is to put land use variables in the Mode Choice step, and increase land use density and mix, which results in more walk and bike trips and fewer car trips. Some large MPOs have done this, but few middle-sized and small ones. This is relatively easy and inexpensive to incorporate. Portland Metro did all of these in 1991 and the Sacramento MPO (SACOG), did them in 1994.
Another major area of model improvement is adding a land use
model to evaluate how land use factors affect travel behavior and how
transportation planning decisions affect land use development patterns (Ewing,
et al., 2007). About half of the induced travel effect is actually caused by
more sprawl, in situations where sprawl land use scenarios are being studied. The
simplest way to model these impacts is to use an Expert Panel. The panel
marks up maps indicating where growth is likely to increase if a transportation
facility is improved. The Conformity rule sort of requires this for regions
with Severe and worse air quality ratings. However, panel members can be
handpicked to be biased, for example, to understate the amount of land use
development that is likely to be induced by a highway improvement, which will
understate induced travel effects. Simple GIS-based land use models are
available, such as UPlan, which is available free and can be set up and run in
days if a MPO has a GIS section and appropriate data (
The biases in current models tend to exaggerate the benefits
of roadway capacity expansion and understate the value of alternative modes and
TDM solutions (
Current models can be improved in various ways summarized in Table 1.
Table 1 Improving Transport Models
|
Factor |
Problems With Current Models |
Appropriate Corrections |
|
Accessibility |
Most transportation models primarily evaluate mobility (movement), and fail to reflect accessibility (people’s ability to obtain desired goods and activities). |
Develop multi-modal models which indicate the quality of nonmotorized and transit travel, and integrated transportation/land use models which indicate accessibility. |
|
Modes considered |
Most current models only consider automobile and public transit. |
Expand models to evaluate other modes, including walking and cycling. |
|
Travel data |
Travel surveys often undercount short trips, non-motorized travel, off-peak travel, etc. |
Improve travel surveys to provide more comprehensive information on travel activity. |
|
Consumer Impacts |
Most economic evaluation models apply relatively crude analysis of consumer impacts. For example, they assume that shifts from driving to slower modes increase costs. |
Use consumer surplus analysis to measure the value to users of transport system changes. Recognize that shift to slower modes in response to positive incentives provide overall benefits. |
|
Travel time |
Most models apply the same travel time value to all travel, regardless of conditions. |
Vary travel time cost values to reflect travel conditions, such as discomfort and delay. |
|
Generated traffic and induced travel |
Traffic models fail to account for the tendency of roadway expansion to generate additional peak-period traffic, and the additional costs from induced travel. |
Incorporate various types of feedback into the traffic model. Develop more comprehensive economic analysis models which account for the economic impacts of induced travel. |
|
Qualitative impacts |
Focus on quantitative factors such as speed and user fees, and undervalues qualitative factors such as convenience and comfort. Level-of-service ratings are provided for roadway conditions but not other modes. |
Develop Multi-Modal Level-of-Service rating systems to help evaluate walking, cycling and public transit travel conditions, in order to identify problems and trade-offs between automobile traffic and other modes. |
|
Nonmotorized travel |
Most travel models do not accurately account for nonmotorized travel and so undervalue nonmotorized improvements. |
Modify existing models or develop special models for evaluating nonmotorized transportation improvements. |
|
Impacts Considered |
Current models only measure a few impacts (travel time and vehicle operating costs). |
Use more comprehensive impact analysis, including crash risk, pollution emissions, pedestrian delays and land use impacts, etc. |
|
Transit elasticities |
Transit elasticity values are largely based on short- and medium-run studies, and so understate long-term impacts. |
Use more appropriate values for evaluating long-term impacts of transit fares and service quality. |
|
Self-fulfilling prophesies |
Modeled traffic projections are often reported as if they are unavoidable. This creates self-fulfilling prophecies of increased roadway capacity, generated traffic, increased traffic problems and sprawl. |
Report travel demand as a variable (“traffic will grow 20% if current policies continue, 10% if parking fees average $1 per day, and 0% if parking fees average $3 per day.”) rather than a fixed value (“traffic will grow 20%”). |
|
Construction impacts |
Economic models often fail to account for construction activity external costs such as congestion and pollution. |
Take congestion delays into account when evaluating projects and comparing capacity expansion with TDM solutions. |
|
Transportation diversity |
Models often underestimate the benefits of improved travel options, particularly those used by for disadvantaged people. |
Recognize the various benefits that result from improving accessibility options. |
|
Impacts on land use |
Models often fail to identify how transport decisions will affect land use patterns, how this affect accessibility and strategic planning objectives. |
Develop integrated transportation and land use planning models which predict how transport decisions affect land use patterns and how land use decisions affect accessibility. |
This table summarizes common problems with current transportation models, and ways to correct those problems. These improvements are particularly important for evaluating alternative modes and mobility management strategies.
Transport Model Improvements are generally implemented by local or regional transportation agencies, often with the support of higher levels of government, professional organizations and academic institutions. Professional standards for transportation models has improved over time, so improvements in a particular community may simply involve bringing local models up to best current practices. New, more comprehensive models are being developed, including generic simulation and integrated land use models, suitable for application in more situations.
Because models affect many transportation planning decisions, Transport Model Improvements can have many travel impacts.
Table 2 Travel Impact Summary
|
Objective |
Rating |
Comments |
|
Reduces total traffic. |
2 |
|
|
Reduces peak period
traffic. |
2 |
|
|
Shifts peak to off-peak
periods. |
2 |
|
|
Shifts automobile travel to
alternative modes. |
2 |
|
|
Improves access, reduces
the need for travel. |
2 |
|
|
Increased ridesharing. |
2 |
|
|
Increased public transit. |
3 |
|
|
Increased cycling. |
3 |
|
|
Increased walking. |
3 |
|
|
Increased Telework. |
2 |
|
|
Reduced freight traffic. |
2 |
|
Rating from 3 (very
beneficial) to –3 (very harmful). A 0 indicates no impact or mixed impacts.
Transport Model Improvements lead to more cost effective planning, particularly implementation of TDM strategies and more accessible land use.
Table 3 Benefit Summary
|
Objective |
Rating |
Comments |
|
Congestion Reduction |
2 |
Supports development of
more efficient transport system. |
|
Road & Parking Savings |
2 |
" |
|
Consumer Savings |
2 |
" |
|
Transport Choice |
2 |
" |
|
Road Safety |
2 |
" |
|
Environmental Protection |
2 |
" |
|
Efficient Land Use |
3 |
" |
|
Community Livability |
2 |
" |
Rating from 3 (very beneficial)
to –3 (very harmful). A 0 indicates no impact or mixed impacts.
Transport Model Improvements tend to better identify the full impacts of transportation decisions, including external impacts such as traffic congestion, parking costs, accident risks and pollution emissions, and so can help reduce these impacts. It also tends to support development of more balanced and efficient transportation systems, which tends to benefit disadvantaged people and improve basic mobility. Multi-Modal Level-of-Service analysis is particularly helpful for identifying problems facing non-drivers and trade-offs between automobile traffic and other modes.
Table 4 Equity Summary
|
Criteria |
Rating |
Comments |
|
Treats everybody equally. |
1 |
|
|
Individuals bear the costs
they impose. |
3 |
Better identifies external
costs of planning decisions. |
|
Progressive with respect to
income. |
2 |
Supports development of
more balanced transport system. |
|
Benefits transportation
disadvantaged. |
2 |
" |
|
Improves basic mobility. |
2 |
" |
Rating from 3 (very
beneficial) to –3 (very harmful). A 0 indicates no impact or mixed impacts.
Transport Model Improvements can be applied in many situations, particularly rapidly-growing urban areas. It is usually implemented by state, regional and local governments.
Table 5 Application Summary
|
Geographic |
Rating |
Organization |
Rating |
|
Large urban region. |
3 |
Federal government. |
2 |
|
High-density, urban. |
3 |
State/provincial
government. |
3 |
|
Medium-density,
urban/suburban. |
3 |
Regional government. |
3 |
|
Town. |
2 |
Municipal/local government. |
3 |
|
Low-density, rural. |
1 |
Business Associations/TMA. |
1 |
|
Commercial center. |
2 |
Individual business. |
1 |
|
Residential neighborhood. |
2 |
Developer. |
1 |
|
Resort/recreation area. |
2 |
Neighborhood association. |
1 |
|
College/university
communities. |
2 |