A REVIEW OF THE INTERNATIONAL MODELING
LITERATURE: TRANSIT, LAND USE, AND AUTO PRICING
STRATEGIES TO REDUCE VEHICLE MILES TRAVELED
AND GREENHOUSE GAS EMISSIONS
A Report for the California Air Resources Board and
the California Department of Transportation
Caroline Rodier, Ph. D. Senior Researcher
Transportation Sustainability Research Center
Institute of Transportation Studies
University of California, Berkeley
1301 S. 46th Street, Richmond Field Station ( RFS), Bldg. 190, Richmond, CA 94804
( 510) 665- 3524 ( O) ( 510) 665- 2183 ( F)
caroline@ tsrc. berkeley. edu
October 1, 2009
ii
ACKNOWLEDGEMENTS
The author would like to thank the California Air Resources Board ( ARB), California
Department of Transportation ( Caltrans), California Energy Commission ( CEC), and the Energy
Efficiency Center at the University of California, Davis for their generous contributions to this
research. In particular, we would like to acknowledge Jeff Weir, Lezlie Kimura, Kurt Karperos,
and Lynn Terry of ARB; Reza Navai, Nancy Chinlund, and Larry Orcutt of Caltrans; and
Panama Bartholomy of CEC. I would also like to thank Susan Shaheen, Bob Johnston, Gordon
Garry, Dan Sperling, and for their support and advice. A special thanks to undergraduate
researcher, Martin Brown, for his dedicated assistance to this project. The contents of this report
reflect the views of the author, who is responsible for the facts and the accuracy of the data
presented herein.
iii
ABSTRACT
California led the nation by passing the first global warming legislation in the U. S. California is
tasked with reducing green house gas ( GHG) emissions to 1990 levels by 2020 and 80% below
1990 levels by 2050. The California Air Resources Board estimates that significant GHG
reductions from passenger vehicles can be achieved through improvements in vehicle technology
and the low carbon fuel standard; however, these reductions will not be enough to achieve 1990
levels if current trends in vehicle miles traveled ( VMT) continue. Currently, most operational
regional models in California have limited ability to represent the effects of transit, land use, and
auto pricing strategies; efforts are now underway to develop more advanced modeling tools,
including activity- based travel and land use models. In the interim, this report reviews the
international modeling literature on land use, transit, and auto pricing policies to suggest a range
of VMT and GHG reduction that regions might achieve if such policies were implemented. The
synthesis of the literature categorizes studies, by geographic area, policy strength, and model
type, to provide insight into order of magnitude estimates for 10-, 20-, 30-, and 40- years time
horizons. The analysis also highlights the effects of modeling tools of differing quality, policy
implementation timeframes, and variations in urban form on the relative effectiveness of policy
scenarios.
Key Words: Travel modeling; land use modeling; land use and transit measures; auto pricing;
green house gas reductions
iv
TABLE OF CONTENTS
Executive Summary........................................................................................................... vi
Chapter 1: Introduction ........................................................................................................ 1
Chapter 2: Methods.............................................................................................................. 2
Chapter 3: Summary of Studies Reviewed .......................................................................... 4
Chapter 4: Synthesis ........................................................................................................... 7
Chapter 5: Conclusions ..................................................................................................... 20
References..................................................................................................................... .... 22
v
LIST OF TABLES AND FIGURES
TABLE 1 Model Type and Quality Categories
TABLE 2 Policy Strength and Type Categories
TABLE 3 Summary of Studies Reviewed by Source, Location, Model, and Number of Scenario
Types
FIGURE 1 Box Plots of Single Policy VMT Reductions by Time Horizon.
FIGURE 2 Distributions of Single Policy VMT Reductions for 20- Year Time Horizon.
FIGURE 3 Box Plots of Combined Policy VMT Reductions by Time Horizon.
FIGURE 4 Distributions of Combined Policy VMT Reductions for 20- Year Time Horizon.
vi
EXECUTIVE SUMMARY
California led the nation by passing the first global warming legislation in the U. S. California is
tasked with reducing green house gas ( GHG) emissions to 1990 levels by 2020 and 80% below
1990 levels by 2050. The California Air Resources Board estimates that significant GHG
reductions from passenger vehicles can be achieved through improvements in vehicle technology
and the low carbon fuel standard; however, these reductions will not be enough to achieve 1990
levels if current trends in vehicle miles traveled ( VMT) continue. Currently, most operational
regional models in California have limited ability to represent the effects of transit, land use, and
auto pricing strategies; efforts are now underway to develop more advanced modeling tools,
including activity- based travel and land use models. In the interim, this report reviews the
international modeling literature on land use, transit, and auto pricing policies to suggest a range
of VMT and GHG reduction that regions might achieve if such policies were implemented. The
synthesis of the literature categorizes studies, by geographic area, policy strength, and model
type, to provide insight into order of magnitude estimates for 10-, 20-, 30-, and 40- year time
horizons. The analysis also highlights the effects of modeling tools of differing quality, policy
implementation timeframes, and variations in urban form on the relative effectiveness of policy
scenarios.
The results of this report provide some order- of- magnitude estimates for policies that appear to
have some promise of near term implementation. Employee parking pricing may result in
approximately a 1% reduction in VMT over the 10- year time horizons. Pay- as- you- drive
insurance policy may produce reductions ranging from 4% to 5% reduction over all time
horizons. Moderate cordon pricing schemes are likely to reduce VMT by 2% to 3% over time.
Increased transit investment may reduce VMT by 0.1% to 1% during a 10- year time horizon, and
in future 10- year increments, this may increase by 1 percentage point at the higher reduction
level. Land- use- only scenarios may reduce VMT by up to 2% in the 10- year time horizon, which
may increase by approximately 2 to 3 percentage points at the higher reduction level at 10 year
increments. Land use and transit scenarios may reduce VMT by 2% to 6% during a 10- year time
horizon, and these figures may increase by approximately 2 to 5 percentage points at each future
10- year increment. Combined land use, transit, and pricing policy measures would bring
significantly greater reductions both in the shorter and longer term time horizons.
In general, the results confirm that even improved calibrated travel models are likely to
underestimate VMT reductions from land use, transit, and pricing policies. These models simply
are not suited for the policy analysis demands in the era of global climate change. For example,
when similar transit scenarios were simulated with the improved calibrated travel model and the
integrated land use and transport model, the latter produced significantly larger results ( 6.0%
versus 0.3%). Despite the very aggressive pricing measures simulated by the improved travel
model in the San Francisco region, the results are significantly lower than weaker pricing
policies simulated in the same region using an advanced travel model.
However, even the advanced models used in the reviewed studies exhibit limitations. Scenarios
simulated with integrated land use and travel models of relatively moderate policy strength in
regions with high quality transit tended to show very small reductions in VMT distributed widely
above the median. These integrated models use relatively large zones and thus have coarse
vii
geographic resolutions, which may overestimate the share of vehicle trips relative to walk and
bike trips from transit- oriented development policies. On the other hand, the advanced travel
model used in the pricing studies may fail to identify possible consequences arising from land
use and transport interactions. For example, pricing policies simulated with integrated land use
and travel models showed that in some cities these policies may actually increase VMT by
shifting housing and employment to outer areas of the regions and increasing average shopping
trip lengths. Theoretically advanced land use and travel models are needed that have fine- grained
geographic resolutions and represent greater variation in the socio- economic attributes of
travelers.
The results of the extrapolation analysis in this study also illustrate the challenge of
implementing land use and transit strategies in a regulatory framework that emphasizes near-term
compliance. For example, the Sacramento Area Council of Government’s blueprint land use
and transport plan was simulated over a 50- year time horizon; the extrapolated results, which
evenly distribute VMT reduction over time, show a 4.2% reduction in VMT in the 10- year time
horizon. However, a much more aggressive scenario, simulated with the improved travel model
in the region over a 10- year time horizon, only showed a 0.4% reduction in VMT.
The analysis of consistent policies across different regions also provides insight into how VMT
reduction may vary given existing land use densities and transit infrastructure. For example,
analyses of land- use only policies suggest that these policies may be less effective in various
European regions and in Washington, D. C. relative to the more sprawling and rapidly growing
regions ( e. g., Sacramento) where trend land use patterns do not take full advantage of existing
transit capacity. The results of the auto pricing policies tend to show greater reductions in VMT
in European cities because of higher quality modal options to the auto. As a result, care should
be taken in generalizing such results to U. S. cities without high quality alternatives.
1
CHAPTER 1: INTRODUCTION
California led the nation by passing the first global warming legislation in the U. S. The
Global Warming Solutions Act ( AB 32) requires California’s green house gas ( GHG)
emissions be reduced to 1990 levels by 2020, and the Governor’s Executive Order targets an
additional 80% reduction below 1990 levels by 2050. Transportation accounts for 36% of
total GHG emissions in California and 27% in the U. S. The California Air Resources Board
( CARB) estimates that significant GHG reductions from passenger vehicles can be achieved
through improvements in vehicle technology and the low carbon fuel standard; however,
these reductions will not be enough to achieve 1990 levels if current trends in vehicle miles
traveled ( VMT) continue. As a result, land use and transport strategies to reduce growth in
VMT are an important part of achieving California’s GHG goals.
Currently most operational models used by state, regional, and local governments in
California have limited ability to represent the effects of transit, land use, and auto pricing
strategies. The major metropolitan planning organizations ( MPOs) in California are in the
process of developing more advanced modeling tools ( activity- based travel and land use
models); however, it is likely to be at least three years before such models are fully
operational. In the interim, this report reviews the international modeling literature to suggest
a range of VMT and GHG reductions that regions might achieve if such policies were
implemented over 10-, 20-, 30-, and 40- year time horizons. The analysis also provides
insights into the effects of varying modeling tools, policy types, regulatory timeframes, and
urban form on the relative effectiveness of discrete and combined policy alternatives.
The report begins with a description of the methods used in the evaluation of the scenarios
including the categorization of models, area type, and policy strength. Next, a general
overview of the studies reviewed is provided, including the location, models, and scenarios
by policy type. This is followed by a literature synthesis, which presents results separately for
single- and combined- policy scenarios. Finally, key conclusions are drawn.
2
CHAPTER 2: METHODS
The literature reviewed in this study consists of studies conducted by regional or state
government agencies, academic researchers, and community groups. To be included in this
review, the study must report VMT and/ or GHG effects of a policy alternative relative to a
base case ( typically a trend or business- as- usual) in the same horizon year. The results are
presented as per capita percentage change in VMT and include both personal and commercial
vehicle travel. GHG results from reduced vehicle travel are used from one study ( Lautso et
al., 2004) because VMT results were not available. Most studies provide simulation results
for only one or two time horizons ( most typically 20 or 30 years); however, the AB 32
legislation has an initial 10- year time horizon, and the Governor’s Executive Order has a 40
year time horizon. Incremental progress toward GHG reduction goals will have to be
monitored. As a result, compound annual growth rates were calculated using the current base
case ( e. g., year 2005) for each future policy scenario time horizon or horizons. The growth
rates were then applied to estimate results for the time horizons of 10, 20, 30, and 40 years.
However, if a pricing study included only one time horizon, then future overestimates were
addressed by applying average extrapolation changes from studies of the same policies in
similar regions ( i. e., size and transit infrastructure). It is important to note that the timing of
implementation could change the estimates for these time horizons and, in general, near term
effects may be overestimated and outer- year effects may be underestimated. Study intervals
( SI), free from distribution assumptions, are identified for a 68% and 95% range of study
scenario results.
EVALUATION
In the evaluation of these studies, the type and quality of the model are categorized as
described in Table 1. The model types include ( 1) travel and/ or land use models of varying
quality, calibrated to specific regions and used for regulatory compliance and planning; ( 2)
experimental or research models typically of high quality but lacking more rigorous
calibration of official models; and ( 3) sketch planning or visioning tools used by community-based
groups to explore different development futures, but not to make official forecasts.
3
TABLE 1 Model Type and Quality Categories
Model Type Quality
Poor Calibrated Travel Limited sensitivity to changes in travel time and cost ( zone- based
without feedback to trip distribution) ( 4- step without feedback)
Typical Calibrated Travel Some sensitivity to changes in travel time and cost ( zone based with
feedback to trip distribution) ( 4- step with feedback of uncertain quality)
Improved Calibrated Travel Better sensitivity to changes in travel time and cost ( smaller zones with
feedback to trip distribution) and higher geographic resolutions ( 4- step
with feedback and greater sensitivity to transit, walk, and bike
variables)
Advanced Calibrated Models More advanced representation of travel behavior, land use, and
economic theories; good sensitivity to modal changes in travel time
and costs; land use effects; and high geographic resolutions ( Travel
and land use models; activity- based models)
Experimental/ research models Similar to advanced models but without the rigorous calibration of
official models
Visioning tools Sketch planning for quick scenario analysis; exploratory analysis of
alternative policies ( unofficial 4- step model; UPLAN; PLACES; INDEX)
To address generalizability, study results are categorized by area type, defined by population
size and transit commute mode share ( in approximately 2000). A region with a population of
seven million or more is categorized as large, between seven and one million is medium, and
less than one million is small. Regions with transit commute mode share greater than or equal
to 10% are categorized as having high quality transit, and those with mode share less than
10% have moderate to low quality transit.
Policy type and strength are also identified in this analysis in Table 2. Land use and auto
pricing policies are widely considered to be effective policies to reduce VMT; however,
historically, in California and the U. S., the adoption and implementation of these policies
have been difficult for a variety of political and institutional reasons. Some of the literature
included in this study attempts to “ bookend” or represent extreme ends of the policy-implementation
spectrum. For example, some assume all new development over a 20- year
period would be accomplished through infill and redevelopment in areas near transit. Others
include congestion pricing policies on all congested roadways or combine multiple auto
pricing policies in one scenario ( e. g., fuel pricing, VMT pricing, and parking pricing). In the
near term, such aggressive implementation of land use and pricing policies seems unlikely.
TABLE 2 Policy Strength and Type Categories
Policy Strength Policies Typically Included
Moderate Improve transit service; reduce transit fares.
Aggressive Land use and transport strategies in official planning documents and/ or
that represent moderate changes relative to historical development
patterns; cordon pricing; pay- as- you- drive insurance; parking pricing in the
urban core; widespread carsharing and telecommuting; traffic calming.
Very Aggressive Land use and transport strategies that depart significantly from historical
patterns and are not included in official planning documents; VMT pricing;
congestion pricing on all roadways; fuel pricing; and region- wide parking
pricing.
4
CHAPTER 3: SUMMARY OF STUDIES REVIEWED
In Table 3, the studies reviewed in this report are summarized by source, location, model,
and number of scenarios by type.
TABLE 3 Summary of Studies Reviewed by Source, Location, Model, and Number of Scenario Types
Size/ Transit Region Studies Models Scenario #
TR LU L
UT
R
P
R
Chicago Chicago Metropolis,
2003
LU ( CRIEM/ GIS)+ TDM 4
Yorkshire Simmonds et al., 2006 LU ( DELTA)+ TDM 7 5
Washington DC Safirova, et al., 2007 LU ( LUSTRE) 4 6
Nelson et al., 2003 START TDM 1
Philadelphia DVRPC, 2003 DVPCP TDM 1
Deakin et al., 1996 STEP TDM 10
Large/ High
San Francisco
MTC, 2007 MTC TDM 1 1 1 2
Deakin et al., 1996 STEP TDM 10
SANDAG, 1998 3
San Diego
SANDAG, 2007
SANDAG TDM
1 1
Deakin et al., 1996 STEP TDM 10
SCAG, 2004 1
Large/
Moderate
Los Angeles
SCAG, 2008
SCAG TDM
1 1
Brussels, BEL Lautso et al., 2004 LU/ TDM ( TRANUS) 1 1 9
Naples, ITA Lautso et al., 2004 LU/ TDM ( MEPLAN) 1 9
Dortmund, Lautso et al., 2004 LU/ TDM ( IRPUD) 1 1 13
GER BCI et al., 2006 LU/ TDM ( Dortmund) 3
Medium/ High
Bilbao, ESP Lautso et al., 2004 LU/ TDM ( MEPLAN) 1 1 9
Austin ENVISION TX, 2003 NA 3
Salt Lake City Envision Utah, 1998 NA 2
Governor's Office, 2000 LU ( UrbanSim)+ TDM 1
Deakin et al., 1996 STEP TDM 10
Johnston et al., 1998 SACMET TDM 2 1
Johnston et al., 2000 1
Rodier, 2002 2 2 4
Johnston et al., 2005
LU/ TDM ( MEPLAN)
1 1 1 2
SACOG, 2004 LU( MEPLAN)+
SACMET TDM
1
Sacramento
SACOG, 2008 SACSIM TDM 1
Twin Cities CEE et al., 1999 3
Barnes, 2003
GIS + TDM
1 3
Portland CSI, 1996 2
METRO, 1998
METRO TDM
1
Seattle PSCOG, 1990 PSCOG TDM 2
Baltimore BMC, 2002 BMC TDM 2
Medium/
Moderate
Orlando HDR, 2003 LU ( ULAM)+ FSU TDM 1
Small/ High Helsinki, FIN Lautso et al., 2004 LU/ TDM ( MEPLAN 1 1 11
Edinburgh, UK BCI et al., 2006 LU ( LUTI)+ TDM 4
Small/
Moderate
Vicenza, ITA Lautso et al., 2004 LU/ TDM ( MEPLAN) 1 1 10
5
San Joaquin Bai et al., 2007 LU ( Small/ Poor UPLAN)+ TDM 1
Pee Dee Pee Dee COG, 2003 TDM 1
Scenarios: TR is transit; LU is pricing; and PR is auto pricing.
Models: TDM is travel model; LU is land use model; and LU/ TDM is integrated.
CALIFORNIA
Special attention is paid to recent transport, land use, and/ or pricing studies conducted by the
four major MPOs in California because of their relevance to the GHG goals of AB 32 and the
subsequent executive order. The Sacramento Area Council of Governments ( SACOG) has
pioneered “ Blueprint” planning in California: an MPO- sponsored participatory planning
process used to develop a common land use and transport vision for the region, which is
ideally accompanied by high- quality modeling of travel, environmental, and economic
impacts. The San Francisco Bay Area Metropolitan Transportation Commission ( MTC), the
San Diego Association of Governments ( SANDAG), and the Los Angeles South Coast
Association of Governments ( SCAG) have also conducted blueprint planning processes that
are more or less similar to SACOG’s approach. The San Joaquin Valley region is currently
conducting its blueprint planning process. In a dramatic departure from the past, four major
MPOs have included their land use strategy in official regional transportation planning
documents ( SACOG, 2008; SANDAG, 2007; SCAG, 2008; MTC, 2007). SACOG was
allowed by the U. S. Environmental Protection Agency to use its land use plan in its official
regional transportation plan alternative as part of its air quality conformity process. The
results of earlier visioning studies of land use and transport scenarios in these regions are also
presented in this study ( SACOG, 2004; SCAG, 2004; SANDAG, 1998). These studies
typically simulate scenarios for a 30- year time horizon. However, the earlier SACOG
Blueprint study ( SACOG, 2004) simulated a 50- year time horizon.
Deakin et al. ( Deakin, et al., 1996) use an advanced calibrated travel model ( the STEP
model) to conduct analyses of a common set of pricing policies across the San Francisco, Los
Angeles, Sacramento, and San Diego regions. The STEP model ( separately calibrated to the
four regions) is particularly well suited to evaluate pricing policies because of its
disaggregate representation of the costs experienced by travelers. Policies are simulated for a
current base year as well as a 20- year future time horizon.
Rodier and Johnston conduct a series of simulation studies using the Sacramento region’s
improved travel demand model ( SACMET) ( Johnston, et al., 2000) as well as an
experimental land use and transport model ( the Sacramento MEPLAN model) ( Rodier, 2002;
Johnston, et al., 2006) to explore transit, land use, and pricing policies for time horizons of
10, 20, and 50 years.
More recently, Bai et al. ( Bai, et al., 2007) use an experimental modeling framework that
includes the UPLAN land use model and a TP+/ Viper travel demand model to examine
transit and land use scenarios in the San Joaquin Valley region for a 25- year time horizon.
6
OTHER STATES
Outside of California in the U. S., simulations have been conducted in three large regions in
the U. S. with high quality transit. Safirova et al. ( Safirova, et al., 2007) and Nelson et al.
( Nelson, et al., 2003) use the experimental LUSTRE land use model and/ or START travel
model to simulate transit, pricing, and land use scenarios in the Washington, D. C., region for
a 20- year time horizon. Thirty- year visioning studies of land use and transit scenarios are
conducted for the Chicago region using an advanced land use and travel model. In the
Philadelphia region, which is part of the states of Delaware, Pennsylvania, and New Jersey, a
travel model of uncertain quality is used to evaluate alternative land use and transit scenarios
for a 20- year time horizon.
Numerous studies have been conducted in medium- sized city regions with moderate quality
transit. In Portland, Oregon, an improved travel demand model is used to simulate land use,
transit, and pricing scenarios in the famous LUTRAQ study ( 20- year time horizon) ( CSI, et
al., 1996). Later, in an official planning study, the improved travel model is used to simulate
future land use scenarios for a 50- year time horizon ( Metro, 1998). In Salt Lake City,
Envision Utah explores land use and transit scenarios as part of a regional visioning planning
process for a 20- year time horizon ( Envision Utah, 1998). Later, an official regional planning
document includes the results of a modified land use and transport plan, with roots in the
Envision Utah process, and simulated with an advanced land use model ( UrbanSim) and an
improved calibrated travel model for a 20 year time horizon ( Governor’s Office of Planning
and Budget, 2000). Visioning studies are also conducted in the Twin Cities ( Barnes, 2003;
CEE, 1999), Austin ( ENVISION Central Texas, 2003), Baltimore ( BMC, 2002), Seattle
( PSCOG, 1990), and Orlando ( HDR, 2003).
INTERNATIONAL
Several studies simulate consistent sets of policy scenarios across European regions. In the
PROPOLIS study, advanced calibrated land use and travel models ( MEPLAN, TRANUS,
and/ or IRPUD) are used to simulate the effects of common transit, land use, and auto pricing
policies for 10- and 20- year time horizons in six European regions ( Lautso, et al., 2004).
Dortmund, Naples, and Bilbao are medium- sized regions with high quality transit. Helsinki is
small sized with high quality transit, and Vicenza is small with moderate transit quality.
In Europe, the STEPS study, also uses advanced land use and travel models to simulate the
effect of common policies in Dortmund and Edinburgh for 20 year time horizons ( BCI, et al.,
2006). The Dortmund and the Edinburgh SPM models are advanced calibrated land use and
travel models. Edinburgh is categorized as a small sized city with relatively high quality
transit.
Simmonds et al. ( Simmonds, et al., 2006) use an advanced land use and travel model
calibrated to the Yorkshire region ( SWYSM which includes the DELTA, START, and DTM
sub- models) to evaluate a range of transit and pricing policies in an official planning
document for a 25- year time horizons. The Yorkshire region is large with high quality transit.
7
8
CHAPTER 4: SYNTHESIS
SINGLE POLICY SCENARIOS
Transit
In the four major regions of California, scenarios are simulated that represent transit service
improvements ranging from 2.9% to 475% ( SANDAG, 2007; MTC, 2007; Johnston, et al.,
2000; Rodier, 2002; Johnston, et al., 2006; SCAG, 2008). Scenarios simulated in six
European cities ( Lautso, et al., 2004) reduce transit travel time by 10%. In Yorkshire,
( Simmonds and Parkman, 2006) transit service is expanded incrementally over subareas with
a 30% reduction in fares and a 20% increase in frequency. Percentage change in VMT for the
four time horizons for these transit scenarios ( N= 9) is as follows ( as illustrated in Figure 1).
• 10 years: median - 0.3%; 68% SI - 1.1% to - 0.1%; 95% SI - 3.7% to - 0.0%
• 20 years: median - 0.7%; 68% SI - 2.1% to - 0.2%; 95% SI - 6.0% to - 0.0%
• 30 years: median - 0.9%; 68% SI - 3.1% to - 0.2%; 95% SI - 8.9% to - 0.0%
• 40 years: median - 1.0%; 68% SI - 3.5% to - 0.3%; 95% SI - 10.4% to - 0.0%
Figure 2 illustrates the distribution of transit results for the most frequent time horizon
represented in these studies, the 20- year horizon. Most scenarios were simulated with land
use and travel models. Those simulated with travel models only, in San Diego, San
Francisco, and Sacramento, tend to fall around the median within the 68% SI ( SANDAG,
2007; MTC, 2007; Johnston, et al., 2000). Scenarios with similar transit investment are
simulated in both the Sacramento MEPLAN model ( Rodier, 2002) and the official calibrated
travel model ( Johnston, et al., 2000) but produce very different VMT reductions: 6.0%
versus 0.3%. The extreme ends of the distribution are represented by a very aggressive transit
investment scenario simulated with the Sacramento MEPLAN model ( Johnston, et al., 2006)
and a transit and highway scenario simulated with a calibrated travel model in the Los
Angeles region, which indicated a 0.5% increase in VMT ( Johnston, et al., 2000). The transit
scenarios simulated with a land use and travel model in Yorkshire tend to rank with the level
of transit service improvement, and most fall above the median within the 95% SI. Yorkshire
is a large region with high quality transit, and thus the relative level of transit service
improvement may be small compared to existing services ( Simmonds and Parkman, 2006).
9
FIGURE 1 Box Plots of Single Policy VMT Reductions by Time Horizon.
10
FIGURE 2 Distributions of Single Policy VMT Reductions for 20- Year Time Horizon.
11
Land Use
Aggressive to very aggressive land- use- only scenarios are simulated in regions of varying
size and quality of transit. In Washington, D. C., Safirova ( Safirova, et al., 2007) simulates
land use scenarios that include high preference for living inside the beltway ( 25% more
attractive); increased residential housing density ( 20% more dense inside the beltway); live
near your work program ( closing cost assistance of $ 8,000 for first- time home buyers); and
an inclusionary zoning program ( increased stock of affordable housing). Elsewhere,
simulations include a land use plan developed as part of the blueprint process in the San
Francisco region ( MTC, 2007); a recentralized land use scenario in an official Philadelphia
region report ( DVRPC, 2003); transit- oriented development policies in six European regions
( Lautso, et al., 2004); visioning scenarios in the Twin Cities ( Barnes, 2003; CEE, 1999); and
finally a very aggressive urban growth boundary policy in the Sacramento region ( Johnston,
et al., 2006). Percentage change in VMT for these scenarios ( N= 19) is as follows:
• 10 years: median - 0.5%; 68% SI - 2.0% to - 0.1%; 95% SI - 3.1% to - 0.0%
• 20 years: median - 1.1%; 68% SI - 4.0% to - 0.0%; 95% SI - 6% to 0.1%
• 30 years: median - 1.4%; 68% SI - 5.9% to - 0.1%; 95% SI - 7.5% to 0.1%
• 40 years: median - 1.7%; 68% SI - 7.7% to - 0.1%; 95% SI - 9.8% to 0.2%
Some interesting patterns develop in the ordering of scenarios around the median. See Figure
2. Scenarios simulated with integrated land use and travel models of relatively moderate
policy strength in regions with high quality transit ( Washington, D. C., Helsinki, Brussels,
Vicenza, and Naples) tend to show very small reductions in VMT distributed above the
median ( Safirova, et al., 2007; Lautso, et al., 2004). VMT is actually increased in two
scenarios, one in Washington D. C. and the other in Helsinki ( Safirova, et al., 2007; Lautso, et
al., 2004). These integrated models use relatively large zones and thus have coarse
geographic resolutions, which may overestimate the share of vehicle trips relative to walk
and bike trips from transit oriented development policies. The exception to this trend,
however, is the very aggressive land use scenario simulated with the experimental land use
and travel model in the Sacramento region, which has the greatest level of VMT reduction
falling outside the 95% SI. This may be explained by the relative densities and transit quality
of the regions: dense European and Washington D. C. regions with high quality transit may
limit the relative effectiveness of the additional densification policies compared to the more
sprawling and rapidly growing Sacramento region where trend land use patterns do not take
full advantage of existing transit capacity. Results for Twin Cities, a region similar to
Sacramento, also fall below median between the 68% SI and the 95% SI ( Barnes, 2003; CEE,
1999). Scenarios simulated with travel models only tend to fall around the median in
Philadelphia ( DVRPC, 2003), Pee Dee ( Pee Dee COG, 2003), San Francisco ( MTC, 2007),
and Orlando ( HDR, 2003).
12
Cordon Pricing
Studies of a range of cordon pricing policies are conducted in Washington D. C. as well as in
Yorkshire and in six other European cities. In Washington, D. C., Safirova el al. ( Safirova, et
al., 2007) evaluate three cordon pricing scenarios: downtown cordon ($ 4.70); downtown
cordon ($ 2.18) and a beltway cordon around the urban core ($ 3.43); and a broader beltway
cordon ($ 2.84). Simmonds et al. ( Simmonds and Parkman, 2006) simulate cordon charges
around the towns and cities of the Yorkshire region. In the PROPOLIS study, cordon pricing
is set at 20% and 60% of the value of commuters’ travel time ( Lautso, et al., 2004).
Percentage change in VMT for scenarios ( N= 16) is as follows:
• 10 years: median - 2.8%; 68% SI - 5.8% to - 1.3%; 95% SI - 14.5% to - 1.1%
• 20 years: median - 2.1%; 68% SI - 6.1% to - 1.3%; 95% SI - 11.0% to - 0.9%
• 30 years: median - 1.8%; 68% SI - 6.4% to - 0.7%; 95% SI - 7.4% to - 0.6%
• 40 years: median - 1.7%; 68% SI - 4.0% to - 0.5%; 95% SI - 6.9% to - 0.4%
All of the cordon pricing policy scenarios are simulated with integrated land use and
transport models, which allow for land uses to reallocate in response to the cordon charge
and thus the effect of a static policy may be reduced over time. Generally, policies rank with
the magnitude of the cordon charge by region. See Figure 2. Below the median at the tail end
of the distribution, the Helsinki scenario includes two cordons that appear to affect a
significantly larger share of trips than in the other regional cordon pricing scenarios. This
result is unlikely to be transferable to regions with multiple employment centers.
Parking Pricing
Parking pricing studies are available for the major California regions and six European cities.
Deakin et al. ( Deakin, et al., 1996) simulate two employee parking pricing charges,
representing a minimum daily price of $ 1.00 and another of $ 3.00 for drive alone work trips.
In the PROPOLIS study, parking pricing is set at 20% and 60% of the value of commuters’
travel time ( Lautso, et al., 2004). Percentage change in VMT for these parking pricing
scenarios ( N= 16) is as follows:
• 10 years: median - 2.2%; 68% SI - 3.2% to - 0.8%; 95% SI - 6.9% to 0.1%
• 20 years: median - 2.2%; 68% SI - 2.9% to - 0.8%; 95% SI - 7.1% to 0.0%
• 30 years: median - 2.2%; 68% SI - 2.8% to - 0.6%; 95% SI - 7.0% to - 0.2%
• 40 years: median - 2.0%; 68% SI - 2.6% to - 0.7%; 95% SI - 6.1% to - 0.0%
The high parking pricing scenarios simulated with an advanced travel model in the California
regions fall below the median within the 68% SI, and the low parking pricing scenarios fall
13
above the median within the 68% SI ( Deakin, et al., 1996) with approximately 1% reductions
across all time horizons. See Figure 2. In the PROPOLIS study, the scenarios simulated with
the integrated land use and travel models tend to rank by policy strength for regions. The
regions of Helsinki and Naples tend to be most responsive to the pricing policies, and
Dortmund and Brussels tend to be least responsive. The small change in Dortmund is
explained by the policy tendency to reduce the auto mode share and to increase average
shopping trips lengths ( Lautso, et al., 2004). In Brussels, per capita VMT is increased by
0.02% in one scenario because of housing and employment shifts from the city center and
inner urban areas to outer areas of the regions ( Lautso, et al., 2004). As households and
employers are able to adjust to the parking pricing policies in scenarios simulated by the land
use and transport models, some results are slightly dampened, and some are increased over-time.
Congestion Pricing
Congestion pricing charges are imposed on all regional roadways to reduce volume of
capacity ratios to the 0.9 level in the major California regions ( Deakin, et al., 1996). In
Washington, D. C., different congestion tolling schemes are simulated, including a variable
comprehensive toll ( similar to Deakin, et al., 1996) and a variable freeway toll ( a more
limited application) ( Safirova, et al., 2007). In Yorkshire, the marginal external cost of
pricing is imposed on roadways. Percentage change in VMT for these scenarios ( N= 9) is as
follows:
• 10 years: median - 2.3%; 68% SI - 6.6% to - 1.6%; 95% SI - 6.8% to - 1.0%
• 20 years: median - 2.8%; 68% SI - 7.1% to - 2.1%; 95% SI - 7.3% to - 1.4%
• 30 years: median - 3.3%; 68% SI - 7.6% to - 2.6%; 95% SI - 7.8% to - 1.7%
• 40 years: median - 3.8%; 68% SI - 8.1% to - 3.1%; 95% SI - 8.3% to - 2.1%
As population grows over time, so does congestion, and thus these policies are more
effective. In general, the stronger congestion pricing policies simulated in the California
regions fall at or above the median, and congestion pricing of similar strength in Yorkshire
and Washington, D. C., fall below. See Figure 2. This result is likely explained by relative
congestion levels in these studies. The California region scenarios were simulated with 1990
and 2010 time horizons and thus tend to have lower relative congestion than latter studies
with a 2020 time horizon. However, it is also possible that the interaction between land use
and transport and greater modal alternatives to the auto contribute to the larger effects.
VMT Pricing
VMT pricing scenarios are evaluated in the California regions ( Deakin, et al., 1996; Rodier,
2002), Washington, D. C., ( Safirova, et al., 2007), and six European regions ( Lautso, et al.,
2004). Deakin et al. ( Deakin, et al., 1996) simulate a VMT fee ( two cents per mile/ 1.6
kilometer increase in auto operating costs) in the four major California regions, which may
represent an aggressive but feasible policy strategy in the form of pay- as- you- drive
14
insurance. Rodier ( Rodier, 2002) simulates a higher VMT pricing fee ( five cents per mile
increase in auto operating costs) in the Sacramento region. Safirova et al. ( Safirova, et al.,
2007) simulate an even more aggressive VMT fee ( a 10 cent per mile increase in auto
operating costs) in the Washington, D. C., area. The PROPOLIS study includes VMT pricing
scenarios that increase per- mile auto operating cost by 25%, 50%, and 100% over existing
levels ( Lautso, et al., 2004). Percentage change in VMT for these scenarios ( N= 27) is as
follows:
• 10 years: median - 9.9%; 68% SI - 14.2% to - 4.4%; 95% SI - 22.7% to - 2.2%
• 20 years: median - 10.4%; 68% SI - 18.4% to - 4.6%; 95% SI - 29.5% to - 3.6%
• 30 years: median - 11.2%; 68% SI - 22.4% to - 5.0%; 95% SI - 43.2% to - 3.9%
• 40 years: median - 11.1%; 68% SI - 24.4% to - 5.0%; 95% SI - 54.2% to - 3.8%
Moderate VMT pricing falls above the median within the 68% SI, and higher VMT pricing in
Sacramento and Washington D. C. falls below the median within the 68% SI. See Figure 2. In
the PROPOLIS study, the scenarios simulated with the integrated land use and travel model
tend to rank by region by policy strength. The regions of Vicenza and Naples tend to be most
responsive to the pricing policies, and Dortmund and Bilbao tend to be least responsive. In
the PROPOLIS study, over time, as the regional urban economies adjust to the policy, there
is a slight dampening of the VMT reductions at the lower VMT price levels and a
heightening of the reductions at the highest VMT price levels. The low VMT scenarios in
Deakin et al. ( Deakin, et al., 1996) scenarios could represent a pay- as- you- drive insurance
scenario in California, and these results suggest a 4% to 5% reduction over the four time
horizons.
Fuel Tax
Fuel tax studies are examined in California ( Deakin, et al., 1996) and in Washington, D. C.,
( Nelson, et al., 2003) for the 20- year time horizon. In California, the following scenarios are
simulated: $ 0.50 per gallon/ 3.8 liters (- 0.13 fuel elasticity); $ 2.00 per gallon/ 3.8 liters (- 0.13
fuel elasticity); $ 2.00 per gallon/ 3.8 liters (- 0.05 fuel elasticity); and $ 2.00 per gallon/ 3.8
liters (- 0.22 fuel elasticity). In Washington, D. C., Nelson et al. ( Nelson, et al., 2003) simulate
a lower fuel tax ($ 0.25 cents per gallon/ 3.8 liters). The results of these fuel tax studies show
that policies rank above and below the median by policy strength. See Figure 2. Percentage
change in VMT for these scenarios ( N= 17) is as follows:
• 10 years: median - 8.4%; 68% SI - 16.6% to - 4.1%; 95% SI - 17.6% to - 3.9%
• 20 years: median - 8.2%; 68% SI - 16.1% to - 4.2%; 95% SI - 17.4% to - 3.8%
• 30 years: median - 8.2%; 68% SI - 15.5% to - 4.1%; 95% SI - 17.1% to - 3.6%
• 40 years: median - 12.9%; 68% SI - 14.9% to - 4.0%; 95% SI - 16.9% to - 3.5%
15
COMBINED SCENARIOS
Land Use and Transit
Analyses of the VMT effects of land use and transit scenarios are available from a series of
official planning and visioning studies in the U. S. Aggressive but feasible land use plans are
included in official planning documents for the following regions: San Francisco ( MTC,
2007), San Diego ( SANDAG, 2007), Los Angeles ( SCAG, 2004; SCAG, 2008), Sacramento
( SANDAG, 2007; SACOG, 2004), Baltimore ( BMC, 2002), Seattle ( PSCOG, 1990),
Portland ( Metro, 1998), and Salt Lake City ( Governor’s Office of Planning and Budget,
2000). More aggressive visioning studies are conducted in Chicago ( Chicago Metropolis
2020, 2003), Salt Lake City ( Envision Utah, 1998), Portland ( CSI, et al., 1996), Austin
( ENVISION Central Texas, 2003), San Diego ( SANDAG, 1998), and the Twin Cities
( Barnes, 2003). More aggressive studies are also included in experimental studies in
Sacramento ( Johnston, et al., 2000; Rodier, 2002, Johnston, et al., 2006) and the San Joaquin
Valley ( Bai, et al., 2007). Percentage change in VMT for these scenarios ( N= 34) is as
follows:
• 10 years: median - 3.9%; 68% SI - 5.7% to - 1.5%; 95% SI - 7.7% to - 0.4%
• 20 years: median - 8.1%; 68% SI - 11.4% to - 3.4%; 95% SI - 14.9% to - 1.4%
• 30 years: median - 11.9%; 68% SI - 16.5% to - 5.1%; 95% SI - 21.4% to - 2.0%
• 40 years: median - 15.8%; 68% SI - 20.7% to - 6.7%; 95% SI - 27.5% to - 2.7%
In general, the results of the very aggressive visioning studies ( SANDAG, 1998; Envision
Utah, 1998; ENVISION Central Texas, 2003) and the experimental academic studies
( Rodier, 2002; Johnston, et al., 2006; Bai, et al., 2007) fall below the median. See Figure 4.
These studies tend to rank by the relative aggressiveness of plan, and those that employ land
use and travel models ( i. e., Chicago, San Joaquin Valley, and Sacramento) are more likely to
fall below the median at the tail end of the distribution. Most of the studies above the median
are official planning documents or more conservative plans in visioning studies. The studies
above the median and at the tail end of the distribution tend to be less aggressive and use
weaker travel models ( SCAG, 2004; CSI, et al., 1996; PSCOG, 1990; SCAG, 2008).
16
FIGURE 3 Box Plots of Combined Policy VMT Reductions by Time Horizon.
17
FIGURE 4 Distributions of Combined Policy VMT Reductions for 20- Year Time Horizon.
18
Combined Pricing
Combined pricing scenarios are available for the four major regions in California. A
comprehensive auto pricing policy scenario is simulated by MTC ( MTC, 2007) in the San
Francisco region that includes a 100% increase in per- mile/ 1.6 kilometer auto operating
costs, 4.9% increase in the average parking cost for work trips, and a congestion pricing
charge of $ 0.25- per mile on all roads when volume to capacity ratios exceed 0.9. Deakin et
al. ( Deakin, et al., 1996) also explore combined pricing policies, which include a region- wide
congestion pricing policy with an average cost of $ 0.13 per mile; a region- wide employee
parking pricing policy with a minimum charge of $ 1.00 per day; a fuel tax of $ 0.05 per
gallon; and VMT/ emissions- based fees of approximately $ 0.01 per mile. Despite the
aggressive pricing measures included in the MTC scenario, the results are the lowest of all
scenarios and low compared to the results of the single pricing policies, described above,
which illustrates improved travel models lack of sensitivity to pricing policies relative
advance models ( i. e., STEP model). Percentage change in VMT for these scenarios ( N= 5) is
as follows ( SI is high to low because of sample size) ( see Figures 3).
• 10 years: median - 4.5%; 68% SI - 4.6% to - 4.3%
• 20 years: median - 8.7%; 68% SI - 8.9% to - 8.5%
• 30 years: median - 12.8%; 68% SI - 13.1% to - 12.5%
• 40 years: median - 16.6%; SI - 17.0% to - 16.3%
Transit and Pricing
In California, the comprehensive auto pricing policy scenario ( described above) is added to
the transit scenario for the San Francisco region ( MTC, 2007). Deakin et al. ( Deakin, et al.,
1996) also add expanded transit to more aggressive pricing policies, including region- wide
congestion pricing ( mean $ 0.13 per mile); region- wide employee parking pricing ( minimum
$ 3.00 per day); fuel tax ($ 2.00 per gallon); and VMT/ emissions based fees ($ 0.01 per mile).
In Sacramento, experimental studies examine a $ 0.05 VMT pricing policy with an aggressive
transit scenario ( Rodier, 2002) and an even more aggressive transit scenario with a gas tax
($ 1.00 per gallon) and parking pricing ($ 6.00 downtown and $ 1.00 elsewhere) ( Johnston, et
al., 2006).
Outside the U. S. in Yorkshire, the congestion pricing policy ( described above) is combined
with increased transit frequencies and reduced transit fares ( Simmonds, et al., 2006). In
Dortmund and Edinburgh ( BCI, et al., 2006), the combined pricing policy ( fuel tax, VMT
pricing, and congestion pricing), transit enhancements ( increased speeds and reduced fares),
and traffic auto calming are simulated with low, high, and/ or extreme fuel price levels. In the
PROPOLIS study, 75% increase in per mile/ 1.6 kilometers auto operating costs is added to a
5% reduction in transit travel times.
19
Percentage change in VMT for these scenarios ( N= 15) is as follows:
• 10 years: median - 10.3%; 68% SI - 16.6% to - 1.6%; 95% SI - 20.0% to - 1.0%
• 20 years: median - 14.4%; 68% SI - 20.3% to - 3.2%; 95% SI - 22.2% to - 1.5%
• 30 years: median - 16.8%; 68% SI - 28.3% to - 4.7%; 95% SI - 31.4% to - 1.5%
• 40 years: median - 17.1%; 68% SI - 35.8% to - 6.3%; 95% SI - 39.5% to - 2.0%
All the PROPOLIS and the Deakin et al. ( Deakin, et al., 1996) results fall below the median
within the 95% SI. See Figure 4. Again, in Deakin et al. ( Deakin, et al., 1996) the regions
with relatively fewer modal alternatives to the auto are more strongly affected by the auto
pricing policies. The Sacramento scenarios simulated by Rodier ( Rodier, 2002) and Johnston
et al. ( Johnston, et al., 2006) tend to be less aggressive than the Deakin et al. ( Deakin, et al.,
1996) scenarios and fall just above the median. In the STEPS study ( BCI, et al., 2006), the
extremely high ( Dortmund) and low ( Edinburgh) fuel price scenarios fall at the ends of the
distribution.
Land Use, Transit, and Pricing
Pricing, expanded transit, and land use studies are available from studies in Sacramento as
well as European regions ( Johnston, et al., 2000; Rodier, 2002; Johnston, et al., 2006;
Lautso, et al., 2004; BCI, et al., 2006). Scenarios in Sacramento include very aggressive land
use, transit, and pricing policies ( VMT tax and parking) ( Johnston, et al., 2000); VMT
pricing policy with an urban reserve, subsidy for infill development, and transit expansion
( Rodier, 2002); a VMT pricing policy with an urban growth boundary and transit expansion
scenario ( Rodier, 2002); and a combined pricing and transit scenario ( described above) with
an urban growth boundary ( Johnston, et al., 2006). In the PROPOLIS study, the transit-oriented
development policy is combined with a 75% increase in auto operating costs, a 50%
reduction in transit fares, and a 5% increase in transit travel speeds. In Helsinki, the transit-oriented
development scenario is also added to a 20% reduction in transit fares, a 5%
increase in transit travel speeds, and a distance based congestion pricing charge ( Lautso, et
al., 2004). In Dortmund and Edinburgh ( BCI, et al., 2006), the combined land use,
carsharing, telecommuting, fuel tax, congestion pricing, and traffic calming policies scenario
is simulated at the low, high, and/ or very extreme fuel price levels. Percentage change in
VMT for these scenarios ( N= 15) is as follows:
• 10 years: median - 14.5%; 68% SI - 22.5% to - 7.1%; 95% SI - 33.1% to - 4.9%
• 20 years: median - 18.0%; 68% SI - 21.9% to - 13.7%; 95% SI - 55.2% to - 8.8%
• 30 years: median - 21.4%; 68% SI - 25.8% to - 14.6%; 95% SI - 70.0% to - 12.9%
• 40 years: median - 24.1%; 68% SI - 32.8% to - 16.8%; 95% SI - 79.9% to - 12.7%
20
The results below the median at the tail end of the distribution include very extreme fuel
price levels and a broader range of travel demand management measures ( e. g., carsharing,
telecommuting, and traffic calming). See Figure 4. These policies may be considered very
aggressive in the U. S. context. In general, policies rank by strength given their geographic
context.
21
CHAPTER 5: CONCLUSIONS
The results of this report provide some order- of- magnitude estimates for policies that appear
to have some promise of near term implementation. Employee parking pricing may result in
approximately a 1% reduction in VMT over the 10- year time horizons. Pay- as- you- drive
insurance policy may produce reductions ranging from 4% to 5% reduction over all time
horizons. Moderate cordon pricing schemes are likely to reduce VMT by 2% to 3% over
time. Increased transit investment may reduce VMT by 0.1% to 1% during a 10- year time
horizon, and in future 10- year increments, this may increase by 1 percentage point at the
higher reduction level. Land- use- only scenarios may reduce VMT by up to 2% in the 10- year
time horizon, which may increase by approximately 2 to 3 percentage points at the higher
reduction level at 10 year increments. Land use and transit scenarios may reduce VMT by
2% to 6% during a 10- year time horizon, and these figures may increase by approximately 2
to 5 percentage points at each future 10- year increments. Combined land use, transit, and
pricing policy measures would bring significantly greater reductions both in the shorter and
longer term time horizons.
In general, the results confirm that even improved calibrated travel models are likely to
underestimate VMT reductions from land use, transit, and pricing policies. These models
simply are not suited for the policy analysis demands in the era of global climate change. For
example, when similar transit scenarios were simulated with the improved calibrated travel
model and the integrated land use and transport model, the latter produced significantly
larger results ( 6.0% versus 0.3%). Despite the very aggressive pricing measures simulated by
the improved travel model in the San Francisco region, the results are significantly lower
than weaker pricing policies simulated in the same region using an advanced travel model.
However, even the advanced models used in the reviewed studies exhibit limitations.
Scenarios simulated with integrated land use and travel models of relatively moderate policy
strength in regions with high quality transit tended to show very small reductions in VMT
distributed widely above the median. These integrated models use relatively large zones and
thus have coarse geographic resolutions, which may overestimate the share of vehicle trips
relative to walk and bike trips from transit- oriented development policies. On the other hand,
the advanced travel model used in the pricing studies may fail to identify possible
consequences arising from land use and transport interactions. For example, pricing policies
simulated with integrated land use and travel models showed that in some cities these
policies might actually increase VMT by shifting housing and employment to outer areas of
the regions and increasing average shopping trip lengths. Theoretically advanced land use
and travel models are needed that have fine- grained geographic resolutions and represent
greater variation in the socio- economic attributes of travelers.
The results of the extrapolation analysis in this study also illustrate the challenge of
implementing land use and transit strategies in regulatory framework that emphasizes near-term
compliance demonstration. For example, SACOG’s blueprint land use and transport
plan was simulated over a 50- year time horizon; the extrapolated results, which evenly
distribute VMT reduction over time, show a 4.2% reduction in VMT in the 10- year time
22
horizon. However, a much more aggressive scenario, simulated with the improved travel
model in the region over a 10- year time horizon, only showed a 0.4% reduction in VMT.
The analysis of consistent policies across different regions also provides insight into how
VMT reduction may vary given existing land use densities and transit infrastructure. For
example, the analysis of land- use- only policies suggest that these policies may be less
effective in various European regions and Washington, D. C. relative to the more sprawling
and rapidly growing regions ( e. g., Sacramento) where trend land use patterns do not take full
advantage existing transit capacity. The results of the auto pricing policies tended to show
greater reductions in VMT in European cities because of higher quality modal options to the
auto. As a result, care should be taken in generalizing such results to U. S. cities without high
quality alternatives.
23
REFERENCES
Bai, S., Niemeier, D., Handy, S., Gao, S., Lund J., and Sullivan D. Integrated Impacts of
Regional Development, Land Use Strategies and Transportation Planning on Future Air
Pollution Emissions. Transportation Land- use Planning & Air Quality Conference, 2007.
Baltimore Metropolitan Council ( BMC). Baltimore Vision 2030. Transportation Indicators.
Baltimore Regional Partnership. Smart Mobility, Inc. et al., Baltimore, ML, November,
2002.
Barnes, G. Transportation- Related Impacts of Different Regional Land- Use Scenarios. Final
Report. Humphrey Institute of Public Affairs, University of Minnesota. Minnesota
Department of Transportation, September, 2003.
BCI, AUEB, ITS, JRC IPTS, KUL- SADL, LT, Novem, Speikermann und Wegener, Stratec,
TIS, TRL, TRT, TTR, UPM. Scenarios for the Transport system and Energy Supply and
Their Potential Effects: STEPs. Work Package 4 Scenario Impacts: Deliverable 4.2, 2006.
Cambridge Systematics Inc. ( CSI) et al. Analysis of Alternatives. Making the Land Use
Transportation Air Quality Connections. 1000 Friends of Oregon, Portland, Oregon, May,
1996.
Center for Energy and Environment ( CEE), Minnesotans for an Energy- Efficient Economy,
1000 Friends of Minnesota. Two Roads Diverge: Analyzing Growth Scenarios for the Twin
Cities Region. MN, June, 1999.
Chicago Metropolis 2020. The Metropolis Plan: Choices for the Chicago Region. Chicago,
IL, 2003.
Deakin, Elizabeth, Greig Harvey, Randall Pozdenza, and Geoffrey Yarema. Transportation
Pricing Strategies for California: An Assessment of Congestion, Emissions, Energy, and
Equity Impacts. California Air Resources Board, November, 1996.
Delaware Valley Regional Planning Commission ( DVRPC). Regional Analysis of What- If
Transportation Scenarios: Final Report. Philadelphia, PA, 2003.
ENVISION Central Texas. Envision Central Texas Briefing Packet. Austin, TX, 2003.
Envision Utah. Scenario Analysis: Executive Summary. Salt Lake City, UT, 1998.
Governor’s Office of Planning and Budget. Strategy Analysis: Quality Growth Efficiency
Tools. Salt Lake City, UT, 2000.
HDR. Community Connections A Transportation Vision for the Next 25 Years: Land Use
Experiment Technical Memorandum No. 8. Orlando, FL, July, 2003.
24
Johnston, R. A., C. J. Rodier, M. Choy, and J. E. Abraham. Air Quality Impacts of Transit
and Supporting Land Use and Pricing Policies. A Report for the U. S. Environmental
Protection Agency, February, 2000.
Johnston, R. A., S. Gao, and M. J. Clay. Modeling Long- Range Transportation and Land Use
Scenarios for the Sacramento Region, Using Citizen- Generated Policies. In Transportation
Research Record: Journal of the Transportation Research Board, No. 1902, Transportation
Research Board of the National Academies, Washington D. C., 2005, pp. 99- 106.
Lautso, Kari, Klaus Spiekermann, Michael Wegener, Ian Sheppard, Philip Steadman, Angelo
Marino, Roberto Domingo, Sylvie Gayda. PROPOLIS: Planning and Research of Policies
for Land Use and Transport for Increasing Urban Sustainability: Final Report, European
Commission, Energy, Environment, and Sustainable Development Thematic Programme.
February, 2004.
Metro. Region 2040: Concepts for Growth. Portland, OR, 1998.
Metropolitan Planning Commission ( MTC). Transportation 2035 in Motion Forecasts for the
San Francisco Bay Area 2009 Regional Transportation Plan. Oakland, CA, November,
2007.
Nelson, Peter, Kenneth Gillingham, and Elena Safirova. Revving up the Tax Engine: Gas
Taxes and the DC Metro Area’s Transportation Dilemma. Resources for the Future.
Washington, DC., 2003.
Pee Dee Regional Council of Governments ( Pee Dee COG). Planning Implications of
Alternate Development Patterns on Infrastructure and Existing Planning Policies. Florence,
SC, 2003.
Puget Sound Council of Governments ( PSCOG). Summary and Comparison between
Alternatives. Vision 2020. Growth Strategy and Transportation Plan for the Central Puget
Sound Region. Final Environmental Impact Statement. Seattle, WA, September, 1990.
Rodier, C. J. Air Emissions Impacts of Regional Transportation and Land Development
Policies. A Report for the U. S. Environmental Protection Agency, June, 2002.
Sacramento Area Council of Governments ( SACOG). Tall Order Forum 2004: Regional
Choices for Our Future. Sacramento, CA, 2004.
Sacramento Area Council of Governments ( SACOG). Metropolitan Transportation Plan
2025. Sacramento, CA, 2008.
Safirova, Elena, Sébastien Houde, and Winston Harrington. Spatial Development and
Energy Consumption. Resources for the Future, Washington, DC., 2007.
25
San Diego Association of Governments ( SANDAG). 2020 Cities/ County Forecast Land Use
Alternatives Summary. San Diego, CA, 1998.
San Diego Association of Governments ( SANDAG). 2030 Regional Transportation Plan:
Pathways for the Future. San Diego, CA, 2007.
Simmonds, David, Andy Skinner, Olga Feldman, James Nicoll, and Caroline Sinclair. Wider
Economic Impacts of Transport Interventions: Final Report. London, England, 2006.
Simmonds, David and Mouchel Parkman, MVA Consultancies. Wider Economic Impacts of
Transport Interventions: Final Report. London, England: Authors, 2006.
Southern California Association of Governments ( SCAG). Southern California Compass.
Growth Vision Report. Los Angeles, CA, 2004.
Southern California Association of Governments ( SCAG). Draft Regional Transportation
Plan. Los Angeles, CA, 2008.
Southern California Association of Governments ( SCAG). Regional Transportation Plan:
Transportation Conformity Report. Los Angeles, CA, 2008.