PROPOSAL TO
THE MINETA INSTITUTE FOR SURFACE TRANSPORTATION POLICY STUDIES
APPLYING AN INTEGRATED URBAN
MODEL TO THE EVALUATION OF TRAVEL DEMAND MANAGEMENT POLICY IN THE
SACRAMENTO REGION
Principal Investigator:
Robert A.
Johnston, Professor
University
of California, Davis
Co-Investigator:
Tomas
de la Barra, Modelistica, Caracas, Venezuela
TABLE
OF CONTENTS
(Headings
are links)
Abstract:
We
have applied a test version of an integrated urban model (travel and land
use) to the Sacramento, California region over the last four years. In
order to make the model more sensitive to Travel Demand Management (TDM)
policies, such as transit and compact land use, we wish to make several
improvements to the model. We will then run a range of policy scenarios
for interest groups in the region. These findings will then be brought
into the regional transportation planning process by these groups.
Nontraditional interest groups are beginning to participate in
transportation planning in this region and the time is ripe to model TDM
policies with an urban model.
We
will widely disseminate our experience, in order to spur similar
improvements to travel models and to other urban models. As many regions
are improving their travel models and are beginning to use urban models,
this work will positively influence these efforts. More complete models
will better represent the effects of TDM policies and so increase their
acceptance throughout the U.S. TDM policies are unlikely to get adopted
until models are improved, because current travel models do not simulate
several critical travel behaviors. The EPA national office and several
regional offices are pushing MPOs to improve their travel models and so
this research is timely.
| TOP |
Background:
The
Sacramento Metropolitan Planning Organization (MPO), SACOG, like most in
the U.S., is proposing to build more freeway capacity, while at the same
time improving transit service. In many corridors, the roadway expansions
will compete with the rail expansions. The major increases in
accessibility are on roads, due to new carpool lanes, a new bridge, bridge
widenings, and new expressways. There are no plans for increasing capacity
further in existing freeway corridors, when there is no longer room to add
lanes in those rights-of-way. That is, there is no long-range plan for
good transit coverage. The overall investment "strategy" is
geographic equity, spreading the funds around. Compact land use
policies were explicitly rejected, as a regional policy. Recent
studies, leading up to the 1999 regional transportation plan, seem to be
laying the groundwork for new concentric freeways, to connect suburban
employment centers. Two such outer beltways were rejected in the 1980s,
but the MPO seems now to be advocating them as needed, by emphasizing the
need to connect suburbs directly, while at the same time finding that
transit is unlikely to work for these trips. This
is traditional thinking, where the travel demand patterns of the past are
projected into the future and then the MPO tries to "meet" these
"needs."
SACOG,
like all MPOs, runs their travel models only out 20 years, which does not
permit one fully to see the land use effects of expanding road capacity in
outer areas in the region. They also use only one land use projection for
modeling all policies and so the travel-reducing land use effects of
transit policies are not modeled.
The
Principal Investigator has considerable experience in the Sacramento
region doing research on land use policy (Johnston, Schwartz, and Tracy
1984; Johnston, Shabazian, and Schwartz, 1999) and transportation planning
(Johnston and Ceerla 1995; Johnston and Rodier 1998; Johnston, Rodier, and
Choy 1998). He is currently the Chair of the UC Davis Sustainable
Communities Consortium, a group of faculty and outreach staff that has
completed a series of research projects and stakeholder meetings in the
region directed at the analysis of sustainable development.
The
most serious problem in this region, in our opinion, is the falling
incomes of the lower half or three-quarters of households by income, a
long-term trend statewide.
Furthermore, the region has lost employment in low-skilled jobs and also
has lost employment in the inner areas, greatly disadvantaging these
households. These problems were also indentified by a Sacramento Housing
and Redevelopment Agency study on economic growth in the region (SHRA
1997). We have adapted a traveler economic benefits model that we run
using calculations from the SACOG travel model. It projects changes in net
benefits to three household income classes and so is useful for equity
analysis. We have done about 30 scenarios for the year 2015 using this
travel model set and have found that freeway widenings do not favor
lower-income households, whereas transit improvements do, if accompanied
by density increases around the light rail stations and good feeder bus
service to the rail stations (Johnston, Rodier, and Choy 1998). We are
currently performing a welfare-to-work transportation study in the region. SACOG does not analyze equity effects of its
transportation plans or projects.
The
SHRA study interviewed a large number of local decisionmakers and leaders,
both public and private, and
found that the most important problems in the region were regional
planning, transportation, education, and quality of life. Fiscal
disparity among localities and geographic segregation of households by
income were found to contribute to all these problems. SACOG does not
address these two underlying problems. With our urban model, we can
simulate open (deregulated) housing markets, regional tax sharing, and
improved schools and lower crime rates in inner city areas, to see how
these policies may help increase transportation equity. The SHRA survey
indicated a fair degree of consensus from respondents across organization
types on these underlying problems and a desire to advance toward regional
coordination of land use planning and the consideration of regional tax
sharing. The business groups now carrying on from the initial SHRA survey
and recommendations, however, have narrowly construed their charge and are
not examining land use planning or tax sharing or fiscal disparity. The
Regional Mobility Task Force, for example, is advocating new freeways and
explicitly does not want to examine land use interactions or effects on
the poor or the environment.
Two
nontraditional interest groups are now becoming involved in transportation
planning in this region,
ECOS, an environmental umbrella group, and SAC-TE, an umbrella group of
neighborhood and social equity groups. We at UC Davis have begun
discussions with leaders of these groups and with the business groups
about improving our existing urban model and running long-range scenarios,
to the year 2040. Whereas we operate the most advanced model in the
region, one that represents the land use effects of transportation
policies, we need to improve our urban model, so that it represents all
relevant travel and locational behaviors.
A
major shift in thinking about travel demand has been occuring over the
last ten years.
The realization that auto travel can be induced by adding roadway capacity
was firmly established by the United Kingdom SACTRA report in 1994 and
acknowledged in the U.S. by the Transportation Research Board (TRB)
Special Rept. 245, Expanding Metropolitan Highways, in 1995. The idea
found further acceptance in TRB Transportation Research Circular No. 481
in 1998. Much of the grudging acceptance of the fact of induced travel has
been spurred by the empirical papers by Hansen and Huang (1997) and by the
recent paper by Noland (1999) at EPA. Further, a whole issue of the
journal Transportation was devoted to induced travel in 1996. Phil
Goodwin, at University College, London, is leading the research effort in
Europe and has just published a book with dozens of examples of
"evaporated traffic," where bridges, tunnels, and roadways were
closed temporarily or permanently. Evaporated, or suppressed, travel shows
that travel demand can go up or down, as road capacity is changed. In
Goodwin's inaugural lecture on accepting the Professorship of Transport
Policy at University College London, (Oct., 1997) he emphasized the
need to see travel demand as affected by policy. He recommended models
that capture all of the behaviors that result from transportation system
changes.
Demand
for auto travel can be induced by adding roadway capacity and it can be
suppressed by letting roadway congestion worsen, by road tolls, by
developing transit service, and by compact and transit-oriented land use
policies. Regional transportation modeling has not caught up with this
new way of looking at demand. Almost all MPOs in the U.S. use outdated
travel demand models that do not represent changes in number of trips,
trip lengths, auto ownership, time of travel, and land development due to
changes in transportation policies. The EPA air quality conformity
analysis regulations now require serious and worse ozone and carbon
monoxide nonattainment regions to account for changes in land use
patterns, number of trips, and trip lengths in regional transportation
planning, but none of the affected MPOs is meeting these requirements (as
far as we know). A recent U.S. District Court case in the Chicago region
held that NEPA requires the consideration of land use changes when a new
freeway segment is proposed. Regional EPA offices pressured Baltimore into
developing a land use model and are strongly critical of travel modeling
in Salt Lake City and in other regions. So, the pressure is on MPOs to
improve their models.
| TOP |
Proposed
Research:
We
operate an integrated urban model, TRANUS, on the Sacramento, California
region. This model was calibrated on datasets for this region and run on
four scenarios for the year 2015, under previous funding. The model
represents route shifting, mode shifting, changes in trip generation and
trip lengths, and land use changes. Under EPA funding, we are currently
improving the model to represent demolition/redevelopment and
refurbishment of buildings in the land supply submodel. It is also being
run out to the year 2040 with a Trend scenario. When doing 50-year
projections, redevelopment is essential to represent in an urban model.
To
be fully sensitive to the complete range of Travel Demand Management (TDM)
policies available, regional models must represent all household behaviors
affected by such policies. We wish to add an Auto Ownership submodel
to the urban model, so that the effects of TDM policies on household auto
holdings can be represented. We also wish to add a Departure Time
Choice model, so peak spreading can be represented. The first will
draw on the Auto Ownership submodel in the SACOG travel model, which we
operate, and the underlying datasets. The second will involve drawing from
applied work in other regions, such as Edmonton, and on academic models.
In
addition, we will make several other improvement to TRANUS
that will make it more useful for evaluating TDM and other transportation
policies. 1. We will recalibrate it on selected 1980 datasets, in addition
to the 1990 database it is currently calibrated on. We will also validate
the model on selected 1995 data. This exercise will make the model more
accurate, overall. 2. A floorspace submodel will be developed inside of
the current land consumption model. We have already collected floorspace
consumption data and lease rate data for the region. The floorspace model
will permit a more accurate depiction of rents by zone and, therefore, of
the effects of policies on housing affordability, a major concern in
regions examining transit and compact growth policies. 3. Goods movement
by heavy truck will be added, based on the goods movement (distribution)
submodel recently added to the SACOG travel model and on the underlying
datasets. 4. A third residential density category will be added, so that
rural residential development can be more accurately represented. 5. We
may go to smaller zones (from 53 to 127), to better represent compact
development near to rail stations. This work will be accomplished by de la
Barra, at Modelistica in
Caracas. The UC Davis team will develop, clean, and transmit datasets to
the Modelistica team. With these improvements, TRANUS will be the most
comprehensive urban model in the world, in terms of representing all
behaviors affected by transportation and land use policies. It will have
made most of the improvements recommended at the Travel Model Improvement
Program conference on land use models in 1995 (TMIP 1995) and in the
recent report from Transit Cooperative Research Program project H-12 on
transit and land use (Miller, Kriger, and Hunt 1998).
After
testing the improved TRANUS, we will rerun the four scenarios we have
already run and evaluate the results, for 2015 and 2040 (No Build, New
High Occupancy Vehicle Lanes,
HOV Plus Beltways, and Light Rail, Buses, and Tolls). The land consumption
outputs will be disaggregated with our Geographic Information System land
allocation model, to produce very detailed land use maps. We have
developed this method already. The results from this previous work with
TRANUS and the GIS models were generally reasonable (Johnston and de la
Barra, 1999). This work interrelates transportation, land use, and the
economy. The TRANUS vehicle activity data are fed into the DTIM2
California emissions model, to get mobile emissions. TRANUS also includes
data on land consumption and numbers of households and employees per zone,
as well as land prices. It also gives a locator (consumer) surplus measure
for three household income classes, which allows us to perform economic
efficiency and equity analyses.
We
will then take these results out to several community groups,
including the Business Mobility Task Force, ECOS, and SAC-TE. We will
determine if they consider the results reasonable and the model
trustworthy. We will also identify other scenarios they want us to run.
Our work will also be shared with two other ongoing private groups, the
Indicators Task Force and the Green Valley (open space planning) group.
Both are part of the Regional Action Partnership business group. We are
known to the leaders and many members
of these groups, as well as to the planning directors in Sacramento City
and County. The majority of our
effort will be expended on these meetings and on the requisite practical
modeling work.
From
these meetings we will outline additional improvements to make to TRANUS
and the GIS model and make those that can be done rapidly. Also, we
will run the requested scenarios. The new version of TRANUS runs in
Windows and ArcView and network editing is easy, so scenarios with
different road or transit networks can be programmed
done in our lab at UC Davis. The GIS model is easy to change, due
to the way we designed it. All maps and output data tables will be posted
on our Web site.
We
will then meet again with the groups
and evaluate the scenarios with them. This will lead to a detailed
writeup, recommendations for the region (really a discussion of which
groups favor which scenarios for what reasons), and recommendations for
modeling at SACOG. The SACOG staff tentatively favors TRANUS and would
like to implement it, but believes it is not acceptable to their Board and
Director. So, this research project may lead to SACOG being asked by some
or all of the outside groups to use TRANUS. The region is in ferment now,
with ECOS and SAC-TE attempting to affect the criteria used by SACOG and
the member cities and counties for selecting projects under TEA-21. These
groups are also pressuring the local governing bodies to open up the
regional transportation planning process to more TDM policies and to
engage in planning for a longer period than the conventional 20 years.
| TOP |
The
Role of Models:
We
do not approach this modeling in a technocratic fashion.
Analysis and prediction can help to identify common ground among user
groups, as they come to understand and accept current problems and future
likely problems. Regarding prescription, we believe our role to not be one
of advocacy of particular viewpoints, but rather to be one of clarifying
value choices, especially the values that are traded off when one selects
any particular policy set. In order for models to help us to understand
these tradeoffs, the models must be complex enough to represent a great
variety of social, economic, and environmental phenomena.
In
addition to being "truth-telling" devices, models also help us
ask the right questions. In this heuristic role, models allow us to learn
together about the urban region and to test prescriptive concepts. Models
can greatly facilitate bargaining, by bringing all interest groups into
the planning process and allowing the testing of the ideas of all
participants.
Models
can be useful in all these roles, because they are systematic assemblages
of our assumptions about how the world works. They provide a consistent
framework for our discussions and analyses. Models do not provide answers,
they just illustrate various points of view. Since they employ graphical
outputs, models can greatly help get interest groups to meet together and
bargain, because of the evocative methods of analysis and portrayal used.
| TOP |
Dissemination:
Last,
we will write up our practical findings on the performance of the
scenarios, our methods findings regarding TRANUS and the improvements
made, and the process of working with the various groups. These results
will find their way directly into policy circles, as the Principal
Investigator is a member of the TRB Committee on Transportation and Land
Development, advises a member of the Travel Model Improvement Program
advisory committee, advises EPA on the conformity regulations, is an
expert affiant for the Chicago lawyers on the freeway lawsuit, and is
currently working with national environmental groups in developing a
manual on land use modeling for small, medium-sized, and large urban
regions. Johnston participated in the 1995 TMIP conference, the TRB
conference on land use models held at UC Berkeley in 1998, and will attend
the NSF conference on land use models and sustainable development at
Berkeley in March, 1999. Papers will be presented at TRB and other
meetings and published in widely read journals. It is possible that the
new director of Caltrans will authorize the development of guidelines for
MPO travel modeling in the state.
| TOP |
Practical
Value of This Work:
Travel
demand management policies, such as transit and compact land use, or
tolls, will tend to not get adopted if they are modeled with the current
generation of inadequate
travel models.
These models do not represent most of the behaviors that occur as a result
of TDM policies and so underproject the travel reductions and economic
benefits of such policies. Several regions have used the Dram/Empal land
use model, but it has produced poor results, because it does not represent
land markets with prices and it uses the region's travel model, instead of
having its own. Since virtually all regional travel models lack most of
the submodels that are needed, TDM policies evaluated with Dram/Empal do
not look very good.
The
TRANUS model is now being applied to the Baltimore region and to the State
of Oregon. In addition, URBANSIM, a quite advanced model, is being
developed for Honolulu, Eugene-Springfield, Oregon, and for Salt Lake
City. Also, METROSIM, another new and advanced urban model, is being
applied to New York City. None of these models, however, has an Auto
Ownership submodel and none has a Departure Time Choice submodel. The
time is ripe for demonstrating an urban model with all behaviors
represented in it. Many other regions in the U.S. will adopt urban
models in the next few years. These models should be improved to include
the capabilities that we will build into TRANUS and demonstrate in this
project. Such models will be much more sensitive to TDM policies and so
will greatly assist in accurately evaluating transit, supportive land use
policies, parking charges, and tolls.
The
strong movement in EC countries toward TDM policies has occured partly
because of the critiques of standard travel models and the discrediting of
their results (for example, in the SACTRA report). The Principal
Investigator has published many papers using an advanced travel model,
which showed that new HOV lanes (for example) increase emissions (contrary
to conventional analysis by MPOs). The suit in Chicago and the findings in
the TRB report, Expanding Metropolitan Highways, were a result of
critiques of the Chicago travel model and of U.S. travel models in
general. Accurate models are essential to gaining a fair evaluation of
TDM policies. Sacramento is a good region in which to demonstrate the
application of such an urban model on TDM policies, as it is a serious
nonattainment region experiencing rapid growth. Furthermore, the most
populous local jurisdicitions, Sacramento City and County, both support
light rail and have land use policies for compact growth centers around
the light rail stations.
| TOP |
Tasks
and Dates:
1.
Task One:
Collect, clean up and assemble all datasets needed for model
improvements.
May 1, 1999 – June 30, 1999 (before this project starts).
2.
Task Two:
Make improvements to TRANUS.
July 1, 1999 – August 31, 1999
3.
Task Three:
Rerun scenarios and evaluate results. Recalibrate submodels, if necessary.
September 1, 1999 – September 30, 1999
4.
Task Four:
First round of meetings with community groups.
October 1, 1999 – October 31, 1999
5.
Task Five:
Make improvements to TRANUS, if necessary. Run additional scenarios.
November 1, 1999 – February 28, 2000
6.
Task Six:
Second meetings with community groups.
March 1, 2000 – March 31, 2000
7.
Task Seven:
Draft final report to IISTPS.
April 1, 2000 – April 31, 2000
8.
Task Eight:
Draft journal articles. Finalize report to IISTPS.
May 1, 2000 – May 31, 2000
9.
Task Nine:
Get approval of journal articles from IISTPS.
June 1, 2000 – June 30, 2000
10.
Task Ten:
Present papers and submit to journals.
July 1, 2000 – on (past this project time period).
Budget:
July
1, 1999 – June 30, 2000
A. Salaries
R.
Johnston, Prof. III @ $6,807/mo.
$3,404
50%
for 1 summer month
Students
1.
Post Grad Res 4 @ $2,652
50%
for 9 months, academic
$11,934
100%
for 3 months summer
$7,956
2.
Post Grad Res 3 @ $2,200
50%
for 9 months, academic
$9,900
100%
for 3 months summer
$6,600
3.
Post Grad Res 3 @ $2,200
50%
for 9 months, academic
$9,900
100%
for 3 months summer
$6,600
Total Salaries: $56,294
B.
Employee Benefits
R.
Johnston @ 8.9%
$302
Student
PGR @ 1.1% academic
$349
Student
PGR @ 2.7% summer
$571
C.
Tuition and Fees
Resident
fees (3) @ $1,331/qtr
11,979
Grad
Student Health Insurance
Resident
fees @ $168.50/qtr
$1,517
D.
Other Expenses
Phones,
mailing, copying, research supplies
$6,000
Local
travel
$500
E.
Subcontract to Modelistica
$20,000
Total
This Proposal: $97,512
| TOP |
References:
Hansen,
Mark and Yuanlin Huang 1997. Road Supply and Traffic in California Urban
Areas. Transp. Res.:A 31:3, pp. 205-218.
Johnston,
Robert A. and Tomas de la Barra 1998. Comprehensive Regional Modeling for
Long-Range
Planning: Integrated Urban Models and Geographic Information Systems.
Transp.
Res.: A.
In press.
Johnston,
Robert A., Seymour I. Schwartz, and Steve Tracy 1984. Growth phasing and
resistance
to infill development: The Sacramento county experience." J.
of the Am.
Planning
Assoc.,
Autumn (Vol. 54, No. 4), pp. 434‑446.
Johnston,
Robert A. and Caroline J. Rodier 1998. Regional Simulations of Highway and
Transit
ITS: Travel, Emissions, and Economic Welfare Effects. Invited paper for J.
of
Mathematical
and Computer Modeling.
Vol. 27, nos. 9-11.
Johnston,
Robert A., Caroline J. Rodier, and Melanie Choy 1998. Transportation,
Land
Use, and Air Quality Modeling, pp. 306-315 in Transportation,
Land Use, and
Air
Quality: Making the Connection,
ed. by Said Easa and Donald Samdahl. American
Society
of Civil Engineers.
Johnston,
Robert A. And Raju Ceerla 1995. The Effects of New High-Occupancy
Vehicle
Lanes on Travel and Emissions. Transp. Res.: A 30:1, pp. 35-50.
Johnston,
Robert A., David R. Shabazian, and Seymour I. Schwartz 1999. Twenty Years
of Growth Management: The Implementation of Sacramento County's 1973
General Plan. Submitted to the J. of the Am. Planning Assoc. Being
revised.
Miller,
Eric, David S. Kriger, and John Douglas Hunt, 1998. TCRP Project H-12:
Integrated Urban Models for Simulation of Transit and Land-Use Policies.
Final Report. University of Toronto. Sept.
Noland,
Robert B. 1999. Relationships Between Highway Capacity and Induced Vehicle
Travel. TRB Annual Meeting. Washington, D.C. preprint.
SACTRA
1994. Trunk Roads and the Generation of Traffic. U.K. Dept. of Transport.
London.
SHRA.
1997. Sacramento Housing and Redevelopment Agency. Phase I Vision Project:
Urban Area Economic Development Plan for Metropolitan Sacramento,
California Region. Sacramento, CA.
TMIP
1998. Travel Model Improvement Program. Land Use Modeling Conference
Proceedings. USDOT. Washington, D.C.
TRB
1995. Transportation Research Board, Expanding Metropolitan Highways.
National Academy of Sciences. Washington, D.C.
|