Estimating Emission Reductions
from Travel Efficiency Strategies:
Three Sketch Modeling Case Studies
SEPA
United States
Environmental Protection
Agency
Office of Transportation and Air Quality
EPA-420-R-14-003
January 2014
-------
Estimating Emission
from Travel Efficiency Strategies:
Three Sketch Modeling Case Studies
Transportation and Climate Division
Office of Transportation and Air Quality
U.S. Environmental Protection Agency
Prepared for EPA by
ICF International
EPA Contract No. EP-C-12-011
Work Assignment No. 1-08
&EPA
United States
Environmental Protection
Agency
EPA-420-R-14-003
January 2014
-------
Estimating Emission Reductions from Travel Efficiency Strategies: Three Sketch Modeling Case Studies
Acknowledgements
The U.S. EPA Office of Transportation and Air Quality would like to thank the following individuals and organizations
for their partnership and support. Along with their supporting agencies, they provided the critical data, information,
and thoughtful input required for the successful completion of this project.
Amanda Graor - Mid-America Regional Council (Kansas City Metropolitan Area)
Catherine Cagle - Massachusetts Department of Transportation (Boston Metropolitan Area)
Susanne T. Cotty - Pima Association of Governments (Tucson Metropolitan Area)
Sisinnio Concas - Center for Urban Transportation Research, University of South Florida
ICF International U.S. Environmental Protection Agency
10-000
-------
-------
Table of Contents
Executive Summary ES-1
Background ES-1
Analysis and Results ES-2
Results and Lessons Learned ES-4
1. Introduction and Background 1
1.1. Introduction to TEAM 1
1.2. 2010 National Study 1
1.3. Case Study Participants 2
2. Methodology and Key Findings 4
2.1. Scenario Selection and Baselines 4
2.2. Data Collection and Validation 5
2.3. TRIMMS Analysis 6
2.4. MOVES Analysis 11
3. Case Study Results 15
3.1. Pima Association of Governments (PAG) -Tucson, Arizona 16
3.2. Massachusetts Department of Transportation (MassDOT) - Boston Region 26
3.3. Mid-America Regional Council (MARC) - Greater Kansas City 37
4. Conclusions and Recommendations 45
4.1. TEAM Data Requirements 45
4.2. TRIMMS Support of TEAM 45
4.3. MOVES Support of TEAM 48
4.4. Regional Realities and Implications for Using TEAM 50
5. Appendix A: Strategies Identified in the 2010 Potential Changes in Emissions Due to
Improvements in Travel Efficiency Report A-1
6. Appendix B: Emission Change Quantities and Additional Technical Details for the
MOVES Analysis B-1
List of Tables and Figures
Table 1. Regional Scenario Descriptions ES-3
Table 2. Percent VMT and Emissions Changes ES-6
Table 3. TRIMMS Analysis Options 9
Table 4. Data Inputs for MOVES Runs 12
Tables. PAG Scenario Details 20
Table 6. Resulting Emissions Changes for Selected Pollutants 21
Table?. PAG Data Sources 22
Table 8. Emissions Factors for PAG 23
Table 9. PAG Regionwide Percent Emissions Changes 24
Table 10. MassDOT Scenario Details.. ...31
ICF International U.S. Environmental Protection Agency
10-000
-------
Estimating Emission Reductions from Travel Efficiency Strategies: Three Sketch Modeling Case Studies
Table 11. MassDOT Data Sources 32
Table 12. Emission Factors for MassDOT 34
Table 13. MassDOT Regionwide Percent Emissions Changes 35
Table 14. MARC Scenario Details 41
Table 15. MARC Data Sources 42
Table 16. Emission factors for MARC 43
Table 17. MARC Regionwide Percent Emissions Changes 44
Table 18. Travel and Emission Changes by Region and Scenario B-1
Table 19. Full Pollutant List B-3
Table 20. Difference in Activity Values: Hourly-to-Annual Aggregation B-4
Table 21. Difference in Emission Values: Hourly-to-Annual Aggregation B-4
Table 22. MassDOT VMT allocation factors B-7
Figure 1: TRIMMS Employer Demand Management Inputs 7
Figure 2: TRIMMS Financial, Pricing, Access, and Travel Times Inputs 7
FigureS: TRIMMS Land Use Inputs 8
ICF International N U.S. Environmental Protection Agency
10-000
-------
Executive Summary
Executive Summary
Background
The purpose of this report is to document how the "Travel Efficiency Assessment Method"
previously used for a national assessment could be applied to specific regions to estimate the
reductions in vehicle miles traveled (VMT), greenhouse gas (GHG) emissions, and criteria
pollutant emissions. The U.S. Environmental Protection Agency (EPA) collaborated with state
and local government officials in three regions to complete these case studies: Pima County
Association of Governments (PAG) for the Tucson, AZ region, Massachusetts Department of
Transportation (MassDOT) for the Boston, MA region, and Mid-America Regional Council
(MARC), for the Kansas City, MO Region. EPA offered technical assistance and the
collaborating agencies offered their time, expertise, and local data to assess reductions in
greenhouse gas (GHG) and criteria emissions from a set of travel efficiency strategies selected
by and tailored to each particular region.
Travel efficiency strategies represent the broad range of strategies designed to reduce travel
activity, especially single-occupancy travel. The term "travel efficiency strategies" builds on the
traditional Transportation Control Measures (TCMs) listed in Section 108(f)(1)(A) of the Clean
Air Act (CAA) such as provision of transit, high-occupancy vehicle (HOV) lanes, and park and
ride lots, and includes other strategies such as transportation pricing such as parking pricing
and per-mile pricing and smart growth, such as transit-oriented development.
These case studies build on other EPA research to quantify the potential reductions in
transportation-related emissions resulting from travel efficiency strategies that reduce VMT.
EPA developed a methodology to estimate VMT and emissions reduced from travel efficiency
strategies called the Travel Efficiency Assessment Method (TEAM). TEAM uses available travel
data and a sketch model analysis to quantify the change in VMT, combined with the EPA
MOVES2010 (Motor Vehicle Emission Simulator) model's emission factors to estimate the
emission reductions that can reasonably be expected. The method allows evaluation and
comparison of scenarios, and thus provides a useful tool for state and local planners who want
to assess impacts of possible strategies before adopting them. This method was applied to
urban areas nationwide to estimate the potential impact that adopting travel efficiency strategies
could have, and the results were documented in EPA's recent study, Potential Changes in
Emissions Due to Improvements in Travel Efficiency1 (2010 national study).
EPA had also provided a user guidance document that describes the methodology in detail, so
that others interested in such an analysis could follow the method established. In 2011, EPA
issued the guidebook describing the TEAM approach titled, Analyzing Emission Reductions
from Travel Efficiency Strategies: A Guide to the TEAM Approach.2 The guide describes the
TEAM approach to estimating the emission reductions from travel efficiencies at the regional
1 2010 national study, www.epa.gov/otaq/stateresources/policy/420r11003.pdf
2 User Guidance for TEAM, http://www.epa.gov/otaq/stateresources/policy/420r11025.pdf
ICF International ES-1 U.S. Environmental Protection Agency
10-000
-------
Estimating Emission Reductions from Travel Efficiency Strategies: Three Sketch Modeling Case Studies
level using information that is typically available from a travel demand model. The analysis can
also be conducted for smaller geographical areas to the extent that the areas are covered by a
travel demand model and the data for these areas can be extracted from the model. The guide
describes the information and data required for analysis, step-by-step procedures for performing
the analysis, considerations for making assumptions about the strategies of interest, and
considerations for interpreting the results. In addition, it identifies default values, alternative
sources of information, and data that can be used when local data and information is incomplete
or absent.
EPA began these case studies in 2012, by requesting letters from agencies interested in
applying TEAM in their local region to evaluate their own selected travel efficiency strategies.
From those that applied with a letter of interest, three agencies were selected as case studies
for testing the TEAM approach at a regional level: Pima County Association of Governments
(PAG) for the Tucson, AZ region, Massachusetts Department of Transportation (MassDOT) for
the Boston, MA region, and Mid-America Regional Council (MARC), for the Kansas City, MO
Region. The selection of these participants to apply TEAM in a regional analysis provides a
collaborative opportunity to determine its applicability across a range of geographic,
developmental and travel activity contexts. These lead agencies identified stakeholder groups to
support their strategy selection and data collection. For purposes of analysis, each region's
selected strategies were grouped into four scenarios. Both the strategies and their underlying
assumptions represent a broad range of potential scenarios for evaluation of corresponding
emissions reductions.
The results of these regional studies were compared with the results of the 2010 national study
to see if they were of similar magnitude. The aforementioned Guide to the TEAM Approach
recommended this validation approach, and the current study provides an opportunity for
comparison. In most cases, the comparison validated that individual agency results were
reasonable. This 2013 regional study estimates fewer potential emissions reduction in response
to the scenarios selected by the regions, especially in regard to land-use strategies. Because of
this, the 2013 regional study highlights some important lessons learned that can inform future
applications of the TEAM approach.
Analysis and Results
Selected Strategies
The three regions selected travel efficiency strategies for modeling that fall into four categories:
Travel Demand Management (TDM) or Employer Incentives, Transit, Land Use, and Pricing.
PAG defined its scenarios as four separate strategies with no overlap, while the other two case
study regions applied their strategies to achieve a progressive increase in VMT reduction across
scenarios. Table 1 provides a brief description of each region's selected strategies.
ICF International ES-2 U.S. Environmental Protection Agency
10-000
-------
Executive Summary
Scenario
Table 1. Regional Scenario Descriptions
Description
Scenario 1: SunTran,
Pass
I Access
Scenario 2: Expanded Employer-
based Incentives
Provides an unlimited-ride transit pass for almost 90,000 faculty, staff, and students at the
University of Arizona and Pima Community College
Increases subsidies for those currently eligible for commute subsidies
Scenario 3: Bus RapidTransit
(BRT) on 2 Corridors
Scenario 4: Parking Pricing in
Downtown
EPA Scenario: Land Use
changes with all other scenarios
^^^^M
Scenario 1: Expanded Healthy
Modes Program
Scenario 2: Scenario 1 + Land
Use
Scenario 3: Scenario 2 + HOV
Lanes
Scenario 4: Scenario 3 +
Expanded Transit
BRT on two major corridors in the region that bring traffic into and out of downtown
Doubles parking prices in a specific downtown-university area along with expanding the
share of parking that is priced
Concentrates population growth in existing urban centers
MassDOT
Increase the number of employees with access to MassRIDES by 25% and expand access
to employer-paid monetary subsidies to all employees with access to MassRIDES
Adds Smart Growth Land Use to Scenario 1 with an increased emphasis on growth in
existing urban centers
Adds HOV to Scenario 2 with a decrease in rideshare travel time for the entire region based
on a network of HOV lanes
Adds transit network expansion and improvement to Scenario 3 to reduce both transit trip
times and access times (wait times) for the regional population
Scenario 1: Expanded TDM
Scenario 2: Scenario 1 +
Enhanced Transit
Expands the group with access to telework and flexwork programs and adds alternative
mode subsidies for this expanded group
Adds transit improvement and promotion to Scenario 1 by reducing transit trip times,
reducing walking distance to transit, and expanding the University's successful transit pass
program
Scenario 3: Scenario 2 + Land
Use
Scenario 4: Scenario 3 + Pricing
Adds Smart Growth Land Use to Scenario 2 to increase average residential density and
mixed land uses for the entire regional population
Adds transportation pricing to Scenario 3 as an increase in the average cost of auto trips
and increase parking costs in the Downtown Area
An important element of this study was to evaluate how changes in land use, particularly the
implementation of smart growth principles, support emissions reductions. Given the long lead
times needed for cities to implement smart growth development plans, EPA chose a thirty year
time horizon to model the results of the different scenarios.
Both MassDOT and MARC included a land use strategy; however, PAG chose not to test the
effectiveness of a land use strategy. To provide a balanced comparison across regions and a
point of comparison with the previous national study, a reasonable land use strategy was
independently developed by EPA for the PAG region. The results are included as an additional
scenario that uses land use as a foundation for the other strategies requested by PAG.
ICF International
10-000
ES-3
U.S. Environmental Protection Agency
-------
Estimating Emission Reductions from Travel Efficiency Strategies: Three Sketch Modeling Case Studies
After conducting a review of various sketch planning models in EPA's 2010 national study, the
TRIMMS model developed by the Center for Urban Transportation Research (CUTR)3 was
selected to demonstrate the TEAM approach. Several factors recommend using TRIMMS for
this type of analysis, including the type and format of inputs required, the geographic scale of
analysis, and the capability of modeling a variety of TCMs, making it highly useful for this type of
analysis.
MOVES2010b was used to determine appropriate emission factors for all regions. MOVES was
run in inventory mode to produce activity-weighted average emission factors. Four primary
pollutants were considered in this analysis: CO2-Equivalent, NOx, PM2.s, and VOC.
Results and Lessons Learned
While estimates of VMT and emission reductions from each case study region are generally
within the range of results for similar regions in the 2010 national study, they are not directly
comparable to the 2010 national study, or to each other. The differences observed can be
attributed to the underlying assumptions and the affected population for each region and the
scenarios selected to be modeled. For example, in one of the cases, a strategy was applied to
a limited set of corridors rather than to all roads; in a couple of the cases, strategies were
applied to a subset of the population rather than to the entire population. The scenarios were
tailored to each region and based on the agencies' individual goals for the case study. The
regions were diverse in their selection of strategies, as well as availability, type and
completeness of data for the models. This limits their comparability to each other, and to the
2010 national study. However, focusing on specific regions in this way was the one of the
reasons for pursuing case studies. In addition, the differences in the regions provided more
insight to the strengths and challenges of the analytical tools used.
To expand on that point, each of the regions had a different future "business as usual" (BAU)
scenario that was used as the benchmark to compare that region's scenarios of travel efficiency
strategies. The BAU scenario was unique to each region and reflected what is already planned
in that area for the future. Each region made a decision for what to include in their BAU. In
Boston, the MPO developed two land use forecasts, one the continues recent historical
development and one that concentrates future growth in core areas served by transit. The BAU
in the Boston case study included the latter. With this in mind, the differences between the BAU
and the other scenarios modeled in the Boston region is smaller than it would have been if the
BAU scenario reflected the more dispersed land use patterns of the recent past. In the Kansas
City region, the MPO created three land use forecasts: one that continues current trends, one
that focuses growth in core areas, and a compromise between these two that was adopted by
the MPO board. In contrast to Boston, in the Kansas City case study the BAU scenario included
3 Concas, S. and P.L. Winters; Economics of Transportation Demand Management (TDM), 2007, "Estimating Costs and Benefits
of Emission Reduction Strategies for Transit by Extending the TRIMMS model." 2012, and Quantifying the Net Social Benefits
of Vehicle Trip Reductions: Guidance for Customizing the TRIMMS Model, 2009;National Center for Transit Research at the
Center for Urban Transportation Research: Tampa
ICF International ES-4 U.S. Environmental Protection Agency
10-000
-------
Executive Summary
the current trends forecast rather than the focused growth forecast or the adopted compromise.
The fact that each region had a unique BAU scenario is one reason why results from one region
are not directly comparable to another. The results illustrate the flexibility of the method, with
this particular version of the sketch model used.
Agencies wanting to model more aggressive policies, applied to broader populations, showed
larger emission reductions. As in the national-level analysis, pricing strategies outperform the
others when applied to a significant share of the population. The largest reductions result from a
combination of mutually supportive strategies modeled together, compared to the sum of
individual strategies modeled separately. For example, land use, transit and TDM improvements
can be mutually supportive and produce better results than when considered individually. The
scale of implementation of a strategy can have important implications for the results as well.
When the population or geography is reduced to a subset of the region, the benefits may be
large to the affected population, but quite small for the region as a whole.
VMT and emission reductions for each of the future year scenarios, expressed as a percent
change, as compared to the business as usual baseline, are provided in Table 2.
ICF International ES-5 U.S. Environmental Protection Agency
10-000
-------
Estimating Emission Reductions from Travel Efficiency Strategies: Three Sketch Modeling Case Studies
Table 2. Percent VMT and Emissions Changes
Percent Emissions Changes for Selected Pollutants
2040 BAU compared to 2040 Scenario
Scenario
Light-Duty GHGs
VMT (C02
equivalent)
Scenario 1: SunTran All Access Pass
Scenario 2: Expanded Employer-based Incentives
Scenario 3: BRT on 2 Corridors
Scenario 4: Parking Pricing in Downtown
Incentives
n
-0.99%
-0.43%
-0.02%
-0.26%
-0.97%
-0.43%
-0.02%
-0.25%
-0.94%
-0.42%
-0.02%
-0.25%
-0.86%
-0.40%
-0.02%
-0.24%
-0.77%
-0.44%
-0.02%
-0.26%
EPA Scenario: Land Use changes with all other
scenarios
-1.95%
-1.87%
-1.73%
-1.43%
-0.71 <
MassDOT
Scenario 1 : Expanded Healthy Modes Program
Scenario 2: Scenario 1 + Land Use
Scenario 3: Scenario 2 + HOV Lanes
Scenario 4: Scenario 3 + Expanded Transit
-2.80%
-3.88%
-4.06%
-4.40%
-2.80%
-3.89%
-4.06%
-4.41%
-2.80%
-3.88%
-4.06%
-4.40%
-2.79%
-3.88%
-4.05%
-4.39%
-2.77%
-3.84%
-4.02%
-4.36%
Scenario 1 :
Scenario 2:
Scenario 3:
Scenario 4:
Expanded TDM
Scenario 1 + Enhanced Transit
Scenario 2 + Land Use
Scenario 3 + Pricing
-0.93%
-2.35%
-2.49%
-12.05%
-0.93%
-2.35%
-2.49%
-12.05%
-0.93%
-2.35%
-2.49%
-12.05%
-0.92%
-2.35%
-2.48%
-0.92%
-2.34%
-2.48%
-12.03% -12.02%
Data
Regions that undertake a TEAM analysis should allow for a significant amount of preparation
time to identify data requirements, collect or identify substitute data elements and validate the
appropriateness of the data for this type of analysis. TEAM was developed to use regional
planning data without requiring significant additional data collection or extensive re-evaluation of
the information. Data validation to determine the appropriateness for analysis is an essential
step in the TEAM approach. This study found that the reasonableness of data in common use
shouldn't be taken for granted. For instance, the distribution of VMT for transit vehicles among
road types is unlikely to be the same as for passenger cars. Road type has an impact on vehicle
speeds, and thus can have a significant impact on emissions. A critical element for applying
TEAM successfully is the ability to identify questionable data and develop substitutions when
needed.
Analysis Tools
The analysis demonstrated that TRIMMS can be an effective sketch modeling tool to estimate
VMT reductions from a variety of travel efficiency strategies. Many of the model's functions
work by manipulation of travel times and travel costs, common factors used in travel demand
modeling. While TRIMMS can provide a relatively rapid and low cost evaluation of travel
efficiency strategies, more comprehensive land-use and transportation modeling should be used
ICF International
10-000
ES-6
U.S. Environmental Protection Agency
-------
Executive Summary
to confirm the analysis before making final policy decisions. The analysis also raised questions
about the estimated impacts of higher density land use. TRIMMS land use analysis features are
in an early stage of development, and the results appear to underestimate impacts on VMT and
are inconsistent with current literature. The web-based version currently under development is
intended to improve the land use capabilities.
MOVES is EPA's on-road mobile source emissions model, and is required to be used for
regulatory purposes under the Clean Air Act. Specifically, MOVES is required to be used for
estimating on-road emissions in State Implementation Plans (SIP) and transportation conformity
determinations. This project has shown that some regions are becoming increasingly familiar
with its use. Accordingly, many are developing their own local MOVES input data to make the
analyses more sensitive to local transportation characteristics. If local inputs were available,
they were employed for the TEAM analysis and where unavailable, MOVES default data was
used. For this non-regulatory planning purpose, the MOVES default data inputs were sufficient
to compare and contrast different scenarios. TEAM analyses and CAA regulatory analyses may
rely on some common data, but TEAM cannot be used to perform the regulatory analyses
required by the CAA.
For each of the regions, MOVES was used to estimate average emissions per mile and per
start, and these rates were applied to the resulting change in VMT and trips to calculate the
potential benefit of the strategies. These case studies illustrate the flexibility inherent in
MOVES. EPA's MOVES guidance documents describe different methods that can be used to
estimate equivalent emissions in a given area, and in each of the regions, a difference approach
was used.4 In one case, MOVES was run for the one relevant county. In another, MOVES was
run for a custom domain, using inputs from one "representative" county. For the third, several
counties were modeled individually and the results summed. In addition, the three areas had a
variety of locally available data and relied on the default data within MOVES to varying degrees.
The MOVES results in terms of per-mile and per-start rates developed in each of the three
areas can be found in Tables 8, 12, and 16.
Conclusions
The foundation of the TEAM approach is the development of scenarios and translating those
scenarios into reasonable input values that can be used by the TRIMMS and MOVES models.
The scenarios can represent reasonably achievable goals or can be an acknowledged "reach"
for a more significant change from the BAU case. A stakeholder group composed of well-
informed local transportation planners, experienced modelers and land use planners can
develop the scenarios and associated model inputs for a region based on their professional
knowledge and limited additional research. While scenarios can be reasonably achievable in the
4 Using MOVES to Prepare Emission Inventories in State Implementation Plans and Transportation Conformity: Technical
guidance for MOVES2010, 2010a and 2010b, EPA-420-B-12-028, April 2012,
Using MOVES for Estimating State and Local Inventories of On-Road Greenhouse Gas Emissions and Energy Consumption -
Final (PDF) (74 pp. 2.4M, EPA-420-B-12-068, November 2012)
ICF International ES-7 U.S. Environmental Protection Agency
10-000
-------
Estimating Emission Reductions from Travel Efficiency Strategies: Three Sketch Modeling Case Studies
long term, they can also represent aggressive regional goals, unconstrained by shorter term
political, fiscal, or technological challenges.
Although results for each region were compared for reasonableness to the similar
representative regional clusters from the 2010 national study, in some cases it was more
instructive to compare the results to the full range of results of the 2010 national study.
Population size and transit mode share do not capture other regional characteristics that can
affect the results, such as average trip lengths, travel costs, climate and geography. In addition,
strategy assumptions, as translated into model input values, may have a significant effect on the
results. Strategies can be specified in more or less aggressive inputs, and can be restricted to
sub-populations and sub-geographies or a region.
The TEAM approach, utilizing travel sketch modeling, regional travel data, information from the
literature, and emissions estimates from the MOVES model, produces reasonable estimates of
emissions reductions from travel efficiency strategies. These case studies tested the approach
in a variety of regional contexts, and the results, while more conservative, fall within the range
predicted by the 2010 national study. Differences appear more related to the appropriateness
of the data used and the limitations of the travel sketch model than the approach itself. The
regional analysis was conducted with TRIMMS 3.0, which includes some significant changes
since the previous version used for the 2010 national analysis, and may have contributed to
some differing results. Given more time, regions should be able to develop better data and
improve the assumptions in TRIMMS and MOVES, making TEAM a useful approach to support
regional decision-making.
ICF International ES-8 U.S. Environmental Protection Agency
10-000
-------
Introduction and Background
1. Introduction and Background
1.1. Introduction to TEAM
The U.S. Environmental Protection Agency (EPA) has a long history of supporting research to
quantify the potential reductions in transportation-related emissions resulting from travel
efficiency strategies that reduce vehicle miles traveled (VMT). Travel efficiency strategies
represent the broad range of strategies designed to reduce travel activity, especially single-
occupancy travel. The term "travel efficiency strategies" builds on the traditional Transportation
Control Measures (TCMs) listed in Section 108(f)(1)(A) of the Clean Air Act (CAA) by adding
other strategies such as transportation pricing and smart growth.
This report documents EPA's latest effort to contribute to the state of the practice for assessing
the impacts of travel efficiency strategies in reducing transportation emissions. The purpose of
the effort was to develop a collaboration with state or local government officials to demonstrate
a methodology developed by EPA and documented in the recent study, Potential Changes in
Emissions Due to Improvements in Travel Efficiency5 (2010 national study), and to quantify
potential emission reductions from travel efficiency strategies at the regional level. EPA
identifies the methodology as the Travel Efficiency Assessment Method (TEAM), TEAM uses
commonly available regional travel data and sketch modeling analysis to quantify changes in
VMT, combined with emissions estimates from the EPA MOVES2010 (Motor Vehicle Emission
Simulator) model's emission factors to calculate the emission reductions that can reasonably be
expected. In support of this effort, EPA provided two resources for demonstrating the
methodology at the regional scale: (1) a user guidance document that describes the
methodology in detail and (2) technical assistance to three regions interested in assessing
reductions in greenhouse gas (GHG) and criteria emissions from these strategies. The process
and results of this effort are discussed in this case study of the three selected regions.
1.2. 2010 National Study
The 2010 national study was intended to establish a reliable and useful source of information on
the effectiveness of selected travel efficiency strategies and to quantify the potential national
emission reductions that could result from those strategies. The study focused on light-duty
vehicles and as such only considered gas and diesel fueled passenger cars and light duty
trucks. The results represent the reduction in urban VMT and emissions that could be achieved
with selected travel efficiency strategies applied to urban areas nationwide, with rural travel
assumed to remain unaffected.
The study used data from 15 metropolitan regions to determine the potential for national-level
emissions reductions. These regions were grouped into "clusters" based on the size of their
population and the extent to which transit is used. Although the analysis is based on actual
travel data and characteristics of real metropolitan areas, because of the aggregation and
5 2010 National Study, www.epa.gov/otaq/stateresources/policy/420r11003.pdf
ICF International U.S. Environmental Protection Agency
10-000
-------
Estimating Emission Reductions from Travel Efficiency Strategies: Three Sketch Modeling Case Studies
averaging of data, the predicted changes to travel activity and resulting emissions from this
analysis are not intended to represent the effectiveness of the strategies for any particular area.
The strategies analyzed were: travel demand management (TDM), land use policies, transit-
related strategies, and parking and road pricing. The strategies were further combined into
future scenarios building from combinations of the most widely applied strategies to more
aggressive approaches like transportation pricing. A sketch-planning tool developed at the
University of South Florida, named Trip Reduction Impacts for Mobility Management Strategies
(TRIMMS, version 2.0), and the data from representative metropolitan regions were used to
estimate the national potential for reductions in VMT under a variety of scenarios. Emission
factors obtained from EPA's MOVES2010 (Motor Vehicle Emissions Simulator) model were
then used to convert VMT reductions into emissions reductions. Key aspects of the study
included:
A review of recent studies to determine the range of effectiveness of various strategies in
addressing travel demand
Development of an assessment methodology (Travel Efficiency Assessment Method, or TEAM),
with input from a panel of subject matter experts
Defining a set of future scenarios that incorporate various strategies expected to reduce travel
activity and emissions
Sketch-planning analysis of actual metropolitan areas representing a range of populations and
transportation characteristics using available local data from regional planning organizations
MOVES 2010 emissions modeling using results from the sketch-planning analysis of the
surrogate metropolitan areas
MOVES2010 was used to generate national-level, fleet-wide emission factors for this analysis
reflecting emissions from start, refueling, and urban driving activities for years 2010, 2020,
2030, 2040, and 2050. These factors account for all changes incorporated in the model's default
assumptions regarding vehicle technology and fuel characteristics. No additional strategies,
including alternative vehicles and fuels or special use of retrofit technologies, were included.
Although not directly comparable due to differences in approach, the results of this study were
similar to other studies estimating the potential reductions from travel efficiency strategies
conducted in the recent past, such as Moving Cooler (Cambridge Systematics 2009).
1.3. 2013 Case Study Participants
EPA recruited regional partners to collaborate with on this project by soliciting letters of interest
from state and local transportation and air quality planning agencies. The received letters of
interest were considered based on the following key factors:
A demonstrated interest in GHG planning and analysis
Regional population above 200,000 in the MPO or Metropolitan Statistical Area (MSA)
Availability of data and staff resources to support the analysis and collaborate with EPA
ICF International 2 U.S. Environmental Protection Agency
10-000
-------
Introduction and Background
EPA targeted technical support to areas that are interested in exploring tools to conduct an
analysis of transportation-related emissions more quickly and efficiently than the methods they
currently use, if any. TEAM is a sketch planning approach that allows for some strategies to be
analyzed at a regional or sub-regional scale without requiring the use of a sophisticated regional
travel demand model. The approach can therefore be used either in regions with sophisticated
analysis capabilities to prioritize strategies for further analysis, or as the primary analytical
method in regions that do not have any established methodologies. EPA was particularly
interested in partnering with agencies that expressed an intent to collaborate with regional and
local partners such as state agencies, regional air quality boards, and environmental or natural
resource agencies on GHG planning and analysis.
EPA received ten letters of interest from a variety of transportation agencies engaged in
regional planning. Three agencies were selected to participate in the case studies, testing the
TEAM approach at a regional level: Pima County Association of Governments (PAG),
Massachusetts Department of Transportation (MassDOT), and Mid-America Regional Council
(MARC). Each agency brought a unique perspective and interest to the study as identified
below.
Pima Association of Governments - local commitment for GHG reduction through U.S.
Mayor's Climate Protection Agreement; presence of baseline GHG emissions inventory;
interest in wider range of strategies, including transit and TDM with ability to provide detailed
data (rideshare program currently in effect); experience with generalized sketch planning
tools like Climate and Air Pollution Planning Assistant (CAPPA), where the TEAM approach
and use of TRIMMS could add value.
Massachusetts Department of Transportation - strong collaboration across the region
and state; motivation due to state legislation and regional targets for GHG reduction;
capacity to leverage knowledge gained through the case study across MPOs in the state as
well as other Northeastern and Mid-Atlantic states through the Transportation and Climate
Initiative; opportunity to compare results from TRIMMS against their own tools developed to
test transit and non-motorized transportation strategies.
Mid-America Regional Council - availability of CarbonFIT model for land use and TDM
analysis that can be used to compare results from TRIMMS; interest in a wide range of
strategies including transit, land use, TDM; strong collaboration across multiple jurisdictions
and two states and playing a leadership role on GHG reduction efforts in the region.
Each agency selected strategies of interest to their region and developed scenarios with these
strategies based on individual goals for the case study. The regions were diverse in their
selection of strategies, availability, type and completeness of data for the models, and this
provided more insight to the strengths and challenges of the analytical tools used.
ICF International 3 U.S. Environmental Protection Agency
10-000
-------
Estimating Emission Reductions from Travel Efficiency Strategies: Three Sketch Modeling Case Studies
2. Methodology and Key Findings
2.1. Scenario Selection and Baselines
Agencies were asked to consider strategies for their region as identified in the 2010 national
study (see Appendix A for the referenced tables 4 and 6) for inclusion in their scenarios.
Selected strategies must be able to be specified to fit the capabilities of TRIMMS in order to
estimate VMT reductions. This requirement limited the full range of potential strategies under
consideration because some strategies cannot easily be modeled in TRIMMS. For instance,
high-occupancy vehicle (HOV) lanes must be translated into travel time savings for input to
TRIMMS. In each region, thought was given to how the travel efficiency strategies chosen could
be reflected in TRIMMS, as explained further in Section 2.3.
Scenarios were comprised of groups of strategies, and assessed for their overall combined
VMT reduction and emissions benefit. The 2010 national study combined strategies in
increasingly aggressive scenarios by starting with the most basic strategies of travel demand
management (TDM) and land use, and progressing to the more controversial strategies for
many regions to implement, such as road pricing. This combining of strategies (through
sequential application and modeling) helped identify how benefits may incrementally increase
with additional actions over time. Both MassDOT and MARC elected to use a similar approach.
PAG was more interested in specific strategies that were likely to garner political and public
support, and therefore selected single strategies for each scenario. The specific strategies
selected to comprise each scenario, along with the data needed to model them are provided for
each agency in Section 3 of this report.
The TEAM results provide a comparison of potential emissions reductions from selected
strategies with the potential emissions from a business as usual (BAU) scenario. The results are
presented as percent reductions based on this comparison. For this reason the selected BAU
baseline is a critical component of using this approach. The BAU scenario represents likely
emissions based on the existing regional plan for transportation infrastructure and regional
growth. The development of a BAU can be challenging, so participating agencies provided a
range of BAU scenarios for use in this study. In most cases the BAU represents the future year
infrastructure and travel activity without additional travel efficiency strategies. This is typically
the scenario in the adopted long range transportation plan and is well suited for use with TEAM.
It is essential to ensure that the scenarios analyzed for comparison match the geographic
boundaries, fleet characteristics, population, and other parameters of the BAU scenario.
EPA has a strong interest in how land use changes, particularly smart growth principles,
influence emissions reductions. MassDOT and MARC included land use strategies in their
scenarios; however, PAG decided not to explore this strategy. To provide a balanced
comparison across regions, a reasonable land use strategy for PAG was developed by EPA
independent from the region's other strategies. The EPA land use strategy was then combined
with the other strategies selected by the region to provide an additional scenario for PAG. The
results of this analysis are discussed in Section 3 as a part of the PAG case study.
ICF International 4 U.S. Environmental Protection Agency
10-000
-------
Methodology and Key Findings
2.2. Data Collection and Validation
Data collection and validation is the most important task within any analysis approach, and
TEAM is no exception. One of the advantages of TEAM is that it relies primarily on data inputs
and outputs from typical travel demand models used for regional planning, without the need to
actually run the travel demand model for each scenario being considered. This can save
valuable time and other costs associated with performing several travel demand model
analyses. The intent of TEAM is to support early decision making by providing a comparison of
different strategies under consistent future conditions. Strategies that appear most effective and
have the necessary public and political support can then be more rigorously analyzed for
precise impacts using the travel demand model and other tools available to the region. When
the data being used is consistent across these levels of analysis, it is more likely that confidence
in the outcome of the scenario can be maintained.
Data collection for the analysis involved obtaining local data from the agencies to include in
MOVES, processing them as necessary, completing quality assurance reviews of provided
inputs and revising as needed, and filling gaps in the local data with national default data from
MOVES to create a complete dataset. Locally specific data is preferred to default data, and
efforts were made to obtain local data wherever possible. All three regions provided some local
data that generally was used by the agencies for other analytical or planning purposes. Only
PAG performed data collection unique to this analysis.
The data required for the TEAM analysis is primarily from the regional travel demand model, but
not entirely. Responsibility for the analysis that supports transportation planning can vary across
states and regions. This variation in responsibility may require cooperation among local
agencies and stakeholders to support the data collection and scenario development necessary
to undertake the TEAM analysis. The EPA notice for letters of interest explicitly stated that "EPA
would prefer to engage with agencies that collaborate with regional and local partners such as
state agencies, regional air quality boards, and environmental or natural resource agencies on
GHC planning and analysis."
Obtaining the data for this study included stakeholder participation in the PAG and MassDOT
regions. The degree to which stakeholders were involved had a significant impact on the
amount of effort required of the lead agency. MassDOT was the region with the broadest array
of stakeholders, which allowed the agency to spread the data requirements and reduce the
demand on lead agency staff. The region with the most independent process, MARC, was
limited in what information they could provide and relied heavily of the use of default data,
especially with respect to the MOVES inputs. The data input elements required for MOVES are
shown in Table 4. In PAG's case study, stakeholder working group meetings helped identify
detailed data for sub-regional geographies to support their specific sub-regional strategies of
interest.
Initially, some questions arose about whether the TEAM results could be used for transportation
conformity, because some of the data provided for use in the MOVES analysis was drawn from
the data used in the area's most recent regional transportation conformity analysis. Although
both analyses rely on some of the same data, TEAM is not part of the transportation conformity
ICF International 5 U.S. Environmental Protection Agency
10-000
-------
Estimating Emission Reductions from Travel Efficiency Strategies: Three Sketch Modeling Case Studies
process and the TEAM approach does not meet conformity requirements. For example, in the
TEAM approach, changes in VMT estimates are generated with TRIMMS rather than the
region's travel demand model. However, for conformity purposes, use of a travel demand
model is required for certain areas, as well as in any areas where the MPO has a travel demand
model.6 Travel demand models provide vehicle activity estimates beyond the capability of
TRIMMS. This detailed activity, such as volumes and speeds on each facility, is then used with
MOVES to more precisely estimate emissions. The TEAM analysis is a less rigorous analysis
than that required for transportation conformity, as it dispenses with detailed transportation
facility classification and associated speed data as used in transportation conformity analyses
where areas are required to use a travel demand model. While TEAM can inform an area's
overarching conformity discussion about the potential emission benefits of particular strategies,
it does not provide results that can be directly used in a regional transportation conformity
determination.
2.3. TRIMMS Analysis
The use of the TRIMMS model to estimate VMT reduction has been a standard feature of the
TEAM approach since the initial 2010 national study. Several factors recommend using
TRIMMS for this type of analysis, including the type and format of inputs required, the
geographic scale of analysis, and the capability of modeling a variety of TCMs, making it highly
useful for this type of analysis.
The previous study used TRIMMS 2.0; however, the analysis for the 2013 regional study was
conducted with TRIMMS 3.0. The Center for Urban Transportation Research (CUTR) at the
University of South Florida made some significant changes between these versions. The user
interface was updated to reduce the number of steps required to conduct the analysis, and the
outputs were expanded in detail. Most significantly, TRIMMS 3.0 added a new land use module
that includes controls for increasing residential densities, land use mixing, transit station
accessibility, and transit-oriented development (TOD). In the previous version of TRIMMS, land
use strategies could be analyzed only by translating them into changes in travel times and travel
distances. Further investigation into the TRIMMS 3.0 land use module are required in order to
fully understand how this change may account for for the lower response to changes in land use
compared to the 2010 study.
TRIMMS strategies are input into three main tables, as shown in the figures below: Employer
Demand Management strategies (Figure 1); Financial, Pricing, Access, and Travel Times
(Figure 2); and Land Use Controls (Figure 3). The employer demand management inputs work
primarily through the use of "radio buttons" where the user selects either yes or no for the
application of the strategy. Financial, pricing, access, and travel times are input as numerical
values. The land use controls are a mixture of sliding scales, where the user specifies a
percentage change from the baseline, and radio buttons.
6 The conformity rule at 40 CFR 93.122(b) requires the use of travel demand models for serious and above ozone nonattainment
and maintenance areas with a population over 200,000; in addition, the rule at 40 CFR 93.122(d) requires MPOs that have a
travel demand model to use that model for conformity purposes.
ICF International 6 U.S. Environmental Protection Agency
10-000
-------
Methodology and Key Findings
Figure 1: TRIMMS Employer Demand Management Inputs
LJ Program Subsidies
Yes No
iarpool Subsidies
Transit Subsidies
Vanpool Subsidies
Bike Subsidies
Walk Subsidies
Guaranteed Ride Home and Ride Match
Yes No
iarpool matching service offered?
Emergency ride home provided?
Vehicle for non-work trips?
Worksite Characteristics
Accessibility
Bus or train station onsite or within 1/4 mile
Bike lanes onsite or within 1/4 mile
Dedicated sidewalk onsite
Amenities
Shopping onsite or within 1/4 mile
Restaurant onsite or within 1/4 mile
Bank onsite or within 1/4 mile
Childcare onsite or within 1/4 mile
Parking
Pa rki ng cha rge for ca rpool i ng?
Pa rki ng cha rge for va npool i ng?
Number of free onsite parking spaces
Yes No
Program Marketing
Telework and Flexible Work Schedules
Yes No
Yes No
Flexible working hours offered?
Compressed work week offered?
Telework program offered?
Internal snail-mail of promotional material?
Internal promotional email?
Do you hold promotional events
Program management and promotion (hrs./week)
Figure 2: TRIMMS Financial, Pricing, Access, and Travel Times Inputs
t* Financial and Pricing Strategies ($)
Mode
Auto-Drive Alone
Auto-Rideshare
Vanpool
Public Transport
Cycling
Walking
Other
Current
Parking
Cost
New
Parking
Cost
Current
Trip Cost
New Trip
Cost
Access and Travel Time Improvements (minutes)
Mode
Auto-Drive Alone
Auto-Rideshare
Vanpool
Public Transport
Cycling
Walking
Other
Current
Access Time
% Workforce Affected
85.0%
New
Access
Time
Current
Travel
Time
New Travel
Time
ICF International
10-000
U.S. Environmental Protection Agency
-------
Estimating Emission Reductions from Travel Efficiency Strategies: Three Sketch Modeling Case Studies
Figure 3: TRIMMS Land Use Inputs
u Land Use Controls
Encouraging higher densities in residential areas
Gross Population Density (persons/sq. mile)
Increase (%)
Current
JW5
A\ _>,
New
0.0
Encouraging mixed land-use
Retail Establishment Density (number/sq. mile)
Increase (%)
Current
so
New
0.0
Increasing station accessibility
Walking distance to neareststation (miles)
Decrease (%)
Current
i"ico
_lJJ >
New
I
J 0.0
Implementing TOD stations
Presence of TOD stop
% Workforce Affected
75.0%
Yes
No
Strategy Selection
The three regions selected individual strategies of interest that fit within the TRIMMS analytical
capabilities. In general, these fall into the four strategy categories identified in Table 3. Many of
the TRIMMS functions work by manipulation of travel times and travel costs, common factors in
travel demand modeling. To model an expansion of HOV lanes in TRIMMS, for example, the
user must input changes in typical travel times for carpool trips (and possibly single occupancy
vehicle (SOV) trips). In order to use TRIMMS effectively, users must consider how to translate
their strategy interest into the input options within the model. Table 3 provides information about
the data needed and the analysis options within TRIMMS that were used to analyze the
strategies selected.
ICF International
10-000
U.S. Environmental Protection Agency
-------
Methodology and Key Findings
Strategy Categories
TDM or Employer Incentives
Transit
Table 3. TRIMMS Analysis Options
Data Needs
share of regional employees covered
average subsidy offered to employees (by
mode)
are guaranteed ride home, ride match, telework,
and flexible work schedules offered?
share of regional population affected
average decrease in transit trip cost
transit travel time and access time
TRIMMS Ootions
financial and pricing strategy entries:
parking cost and trip cost
program subsidy radio buttons
guaranteed ride home, ride match,
telework, and flexible work schedules
radio buttons
financial and pricing strategy entries:
access time and travel time
Land Use
Pricing
share of regional population in affected areas
increase in weighted average residential density
(persons per square mile)
increase in weighted average retail
establishment density (number per square mile)
average decrease in walking distance to transit
land use controls: residential density,
retail establishment density, and
walking distance to nearest station
share of all parking (public and private) that is
priced
average increase in parking cost per trip
average increase in trip cost
financial and pricing strategy entries:
parking cost and trip cost
Customizing TRIMMS
Elasticities
TRIMMS contains default elasticities that measure the relationship of travel costs and access
times to travel patterns. Direct elasticities describe relationships between travel by one mode
and the cost and time characteristics of that mode. For example, a direct elasticity of drive alone
travel with respect to trip cost of -0.5 means that a 1% increase in drive alone trip cost is
associated with a 0.5% decrease in drive alone travel. Cross elasticities describe relationships
between different modes of travel. A cross elasticity of drive alone travel with respect to transit
access times of 1.0 indicates that a 1% decrease in transit access times is associated with a 1%
decrease in drive alone travel.
TRIMMS does not provide defaults for all elasticities. For example, TRIMMS 3.0 does not have
a default elasticity to represent the shift in drive alone travel relative to the cost of carpool or
vanpool travel. TRIMMS supplies zero values for these elasticities, whereas their true values
are most likely non-zero. TRIMMS allows the substitution of elasticities for its default values.
The national-level research for EPA identified elasticities that are well-supported in the
literature. The research team relied upon these values to supplement and support what was
available in TRIMMS. The TEAM user guidance identifies the desired elasticities for a regional
analysis as well as provides a list of elasticities in its Appendix C.
ICF International
10-000
U.S. Environmental Protection Agency
-------
Estimating Emission Reductions from Travel Efficiency Strategies: Three Sketch Modeling Case Studies
Still, there are several instances in which elasticities not supplied by either TRIMMS or the
literature, and which therefore default to zero, can produce unreasonable analysis results. If the
model is run with these values set to zero, the TRIMMS model may predict increases in trips for
one mode of travel without corresponding decreases for another mode. TRIMMS automatically
rebalances total trip numbers when evaluating employer-based strategies, but not for other
strategy types. The user can correct for this by adjusting the outputs of TRIMMS to rebalance
total trip numbers (the approach taken in this study, as described below) or by supplying a more
complete set of elasticity values.
Alternative Populations and Geographies
In the 2010 national study, TEAM was applied to each region's entire population for all
strategies with the exception of TDM, which considered only the working population. However,
in the 2013 regional study PAG requested modeling strategies that clearly applied to only a
subset of the population and a limited geography within the region. Using TEAM for sub-
geographies and sub-populations is more complicated. Sub-geographies and populations
sometimes require different baseline assumptions (e.g. mode shares and trip lengths for
downtown vs. region), and sub-populations can be difficult to isolate (e.g. traveler population for
a specific corridor). Combining strategies that apply to different sub-population or sub-
geographies requires that the effects of the strategies be summed together outside of TRIMMS
as a post-processing step. In practice this is more likely to overestimate rather than
underestimate impacts. When VMT reduction strategies are applied cumulatively within a single
model run, the second strategy applies to a smaller baseline VMT than the first strategy, and
therefore produces smaller absolute VMT reductions than if applied on its own.
The value of the results using sub-geographies and sub-populations is sometimes limited. The
corresponding emissions reductions are quite small when applied to the entire region and may
be insignificant even for the limited geography. This type of analysis is more suited to use of a
sub-regional model to capture changes in VMT. Regionwide averages and regional scale
strategies are more appropriate for TRIMMS and more consistent with the intent and value of
TEAM.
However, in some cases modeling for sub-populations and sub-geographies is unavoidable,
such as when modeling employer-based strategies. These will naturally apply only to the
employed population. In these cases post-processing is essential. For conducting TRIMMS
model runs, each scenario was divided into the constituent populations and one model run was
conducted for each population. For example, to combine an employer-based strategy with a
pricing strategy we conducted two model runs: one run for the employed population, to which
both TDM strategies and pricing apply; and one run for the rest of the population, to which only
pricing applies. Absolute VMT reductions for each model run were summed together to produce
total VMT reductions for the region.
ICF International 10 U.S. Environmental Protection Agency
10-000
-------
Methodology and Key Findings
Reasonableness Check
For each case study region, a comparison was made to the results for similar representative
regions (based on regional population size and transit mode share) in the 2010 national study.
Results for each strategy type for the case study regions were expected to be similar to the
results for the comparison representative regions. The aforementioned Guide to the TEAM
Approach recommended this approach, and these case studies provide an opportunity to test
the usefulness of this comparison. Where results for case study regions were dissimilar to those
of the comparison region in the 2010 national study, it is noted in the discussion. In some cases,
the comparison helped to identify necessary adjustments to the TRIMMS inputs.
2.4. MOVES Analysis
Modeling and post-processing proceeded in stages. First, the actual MOVES runs were
conducted, including collection and assembly of all input data and runspec control files, quality
assurance and correction of any data issues as necessary and extracting inventory results from
the output database. Next, the results were post-processed into a form useable in this analysis.
This consisted of converting the inventory values of emissions and activity into regional
average, activity-weighted emission factors for each pollutant, vehicle type, and year for the
same composite vehicle types used in the TRIMMS analysis. Each of these steps is described
more specifically in the sections for each region.
MOVES is EPA's most current and capable emissions model, and yet is still somewhat
unfamiliar to some MPOs. Accordingly, there was varying degree of experience from the three
regions on coordinating and collecting the necessary data. The PAG region, in particular,
expressed interest in shadowing our analysis, examining all aspects of data collection and
processing to further their understanding of the model and its use. Of the three regions, the
Boston region MPO showed the most expertise and comfort with the model, and had the most
complete input data set. However, their inputs were based on data from a more geographically
narrow area than that desired for the TEAM analysis. Accordingly, their data required some
modifications for this analysis. MARC was the most straightforward case because little local
data was available, necessitating that the analysis be based largely on model defaults. Each of
these is discussed more fully in Section 3.
ICF International 11 U.S. Environmental Protection Agency
10-000
-------
Estimating Emission Reductions from Travel Efficiency Strategies: Three Sketch Modeling Case Studies
Data Inputs
MOVES allows the user to select the scale and the geographic boundaries for the analysis.
When the county scale is chosen, the user can input data specific to the county of interest, or
can use the default information within MOVES. Since a TEAM analysis is non-regulatory, there
are no restrictions against using default data. However, at the county scale, no default data is
available for vehicle type population or for VMT; the user must enter the appropriate local data.
Default data is available for other data fields. The MOVES default database includes
information that varies by county for fuel and I/M program type, based on survey data at the
time the model was developed. It also contains meteorological data for each county and fuel
supply and formulation information for each county. For the remaining data fields, the default
data are national average data that are then used for the county as-is. For example, using the
default age distribution will apply the U.S. average age distribution to the particular county,
which may over or underestimate the real age distribution of the county.
Table 4. Data Inputs for MOVES Runs
Data Type Description Data Elements
Modeling decision elements that are
selected in the MOVES run
specification file
Fields for which county-specific data
must be entered when using the
county scale
Domain/Scale
Calculation Type
Time Aggregation
Calendar Year
Evaluation Month
Type of Day
Evaluation Hour
Geographic Bounds
Vehicle Type
Road Type
Pollutants and Processes
Strategies
Activity
Emissions Detail
Source (Vehicle) Type Population
Vehicle Type VMT
Fields for which MOVES includes
default county-specific data
Meteorological Data
Fuel Supply/Formulation
Fields for which MOVES includes a
national default that can be used for
TEAM when county data is
unavailable
I/M (Inspection and Maintenance) Program
Age Distribution
Ramp Fraction
Month, Day, Hour VMT Fractions
Average Speed Distribution
Alternative Vehicle and Fuel Technology
ICF International
10-000
12
U.S. Environmental Protection Agency
-------
Methodology and Key Findings
The scale selected for MOVES modeling was the county scale, to make use of more precise
local data and consistent with EPA Guidance,7 but the approach varied across the regions. The
PAG analysis was performed for the entirety of Pima County. The greater Boston area was
modeled as a custom domain encompassing several counties. The seven counties in the MARC
region were modeled individually and the results aggregated outside the model. Data collection
was focused on each element, as identified in Table 4, for analysis as chosen by each region.
Details about the collection, processing, sources, and quality assurance of each of these data
items appear in the regional discussions in Section 3.
Deriving Emissions Factors
MOVES was run in inventory mode for each of the regions, and the resulting totals of emissions
and activities (VMT and number of starts) were ratioed to produce activity-weighted average
emission factors for the region. Four primary pollutants were considered in this analysis: CO2-
Equivalent, NOx, PM2.s, and VOC. Other pollutants were also included in the analysis as done in
the initial study. These are provided in Appendix B.
As noted above, TRIMMS uses composite vehicle types. Emission factors from MOVES were
derived by combining MOVES vehicle types to represent the TRIMMS composite vehicle
category definitions.8 For the TRIMMS auto drive alone and auto rideshare vehicle categories,
composite emission factors representing motorcycles, passenger cars, and passenger trucks
were combined from the MOVES model. For the TRIMMS vanpool vehicle category, composite
emission factors representing MOVES passenger truck and light commercial truck vehicle types
were used.
The MOVES analysis included all road types. MOVES emission process types included Start
Exhaust, Crankcase Running Exhaust, Crankcase Start Exhaust, Running Exhaust, Brakewear,
and Tirewear. All MOVES runs for this project were run without pre-aggregation of the data.
While pre-aggregation saves modeling time, it can reduce precision. As recommended by EPA
for most purposes, an hourly analysis was performed for all hours, days, months of the year.
Emission factors were then calculated as total running emissions (in grams per year) divided by
total running activity (in miles per year); a similar analysis was made for starting emissions.
Emission factors, in grams of pollutant per average mile driven and grams of pollutant per
average start were produced.
From this, a single emission factor was derived for each vehicle type, year, pollutant, and
process type. This represents an overall average for that year and was used for every scenario
in that year. The current year emission factors were paired with baseline activity in the current
year for total baseline emissions, and future year emission factors paired with activity for the
7 Analysis for scale and other parameters adhered to EPA's current guidance for estimating on-road greenhouse gas emissions:
Using MOVES for Estimating State and Local Inventories of On-Road Greenhouse Gas Emissions and Energy Consumption -
Final (EPA-420-B-12-068, November 2012).
8 Other vehicle parameters are also combined within the model. These include fuel type and model year, which are based on the
locally specific inputs collected from the regions.
ICF International 13 U.S. Environmental Protection Agency
10-000
-------
Estimating Emission Reductions from Travel Efficiency Strategies: Three Sketch Modeling Case Studies
BAU and all scenario alternatives to generate total future year emissions from the BAU and
scenarios. Both starting emissions - based on TRIMMS-calculated number of trips - and
running emissions - based on TRIMMS-calculated VMT were included in computing the total
emissions and emission changes. The calculation of total emissions was done in a simple, off-
model spreadsheet calculation that assembled outputs from both the emissions and VMT
modeling results. Regionally specific details and results are presented in the Section 3.
ICF International 14 U.S. Environmental Protection Agency
10-000
-------
Case Study Results
3. Case Study Results
Three agencies were selected as case studies for testing the TEAM approach at a regional
level: Pima County Association of Governments (PAG), Massachusetts Department of
Transportation (MassDOT), and Mid-America Regional Council (MARC). The lead agencies
identified stakeholder groups to support their strategy selection and data collection. The
selection of strategies was grouped into four scenarios for each region based on their individual
interests and data availability. PAG defined its scenarios as four separate strategies with no
overlap, in contrast to the other two case study regions, which incrementally applied strategies
to achieve a progressive increase in VMT reduction across scenarios. The strategies and their
underlying assumptions represent a broad range of potential futures to be evaluated for
corresponding changes in travel activity and emissions reductions. Each of the regional
analyses are examined in detail in the following sections.
ICF International 15 U.S. Environmental Protection Agency
10-000
-------
Estimating Emission Reductions from Travel Efficiency Strategies: Three Sketch Modeling Case Studies
3.1. Pima Association of Governments (PAG) - Tucson, Arizona
Background
PAG is a metropolitan planning organization (MPO) with regional planning responsibilities for
transportation, air quality, and water. The organization is governed by a regional council
composed of representatives of its member jurisdictions: Pima County, Tucson and South
Tucson, the towns of Marana, Sahuarita, Oro Valley, and the Tohono O'odham Nation and
Pascua Yaqui Tribe. PAG is responsible for development of the area's long-range regional
transportation plan.
PAG has an active travel reduction program which includes voluntary computer based ride-
matching for residents and mandatory ride-matching for employers with greater than 100
employees. There is an interest in transit fare and TDM programs. The region's GHG inventory
is done every two years, and both county and city have adopted mandates to reduce GHG
emissions.
Interest in Participation
When submitting its letter of interest, Pima County was in compliance with the EPA's air quality
standards with an ozone level at 90% of the ozone standard, but was concerned that its status
could change with a future, more stringent ozone standard. Gathering more accurate
information on the emission benefits from travel efficiency strategies can potentially help the
region lower VOC and NOx emissions and ultimately ozone concentrations.
Along with the concern about ozone, the region must comply with the requirements in the
carbon monoxide Limited Maintenance Plan (CO LMP) due to CO violations in the late 1970's
and early 1980's. One of these requirements is to implement a Travel Reduction Program (TRP)
for the region to encourage alternate modes of transportation to improve regional air quality.
The TRP is administered by PAG staff and has been in effect since 1988, with over 290
companies and organizations participating. The program includes employers in unincorporated
Pima County, Tucson, South Tucson, Marana, Oro Valley and Sahuarita. The benefits of the
TRP program are quantified by estimated VMT reductions and gasoline savings. PAG is
interested in improving staff ability to more precisely estimate the TRP benefits and tailor future
planning to promote programs with the most significant pollution benefits. In addition to the TRP,
both Tucson and Oro Valley have signed the U.S. Mayor's Climate Protection Agreement
committing to reduce GHG emissions to seven percent below 1990 levels by 2012. Both entities
have initiated aggressive plans to address this reduction goal.
PAG's initial letter of interest indicated that using the TEAM approach to evaluate travel
efficiency strategies could:
Provide assistance in meeting regional policies and actions such as TRP ordinances and the
Mayor's Climate Protection Agreement
Help to refine actions already underway through the TRP program
ICF International 16 U.S. Environmental Protection Agency
10-000
-------
Case Study Results
Refine regional transportation-related emissions using the more complex MOVES model,
resulting in data that are more closely tracking local conditions and GHG emissions
Provide an improvement over the Clean Air and Climate Protection model, which does not
incorporate speed, local meteorological or travel patterns or vehicle registration data
Assist PAG's jurisdictions with resource and program planning to select the travel reduction
components that allow the greatest air quality benefits
PAG had a specific interest in the use of MOVES for this analysis, and requested permission to
"shadow" the analysis effort as a training exercise. This level of interface resulted in more
coordination between PAG and EPA to answer questions and resolve issues; however, it also
produced a stronger understanding of how the MOVES model can be used for this type of
analysis.
Prior Experience with Analysis of GHG Emissions
PAG staff conducts GHG modeling and analyses to develop a Regional GHG Inventory for
eastern Pima County and the City of Tucson communities and their respective government
operations. For the analysis, the most current edition of the inventory was released in 2011 and
covers 1990 through 2008. The preliminary results from the 2011 inventory indicate that
regional transportation contributed 32 percent to the 2010 Pima County's GHG total. From 1990
to 2010, regional, private/commercial VMT increased by 63 percent with an accompanying 34
percent increase in GHG emissions. In 2010, PAG staff conducted GHG modeling and analyses
and developed GHG emission inventories for PAG's outlying jurisdictions (Marana, Oro Valley,
and Sahuarita) spanning the 2000-2008 timeframe. PAG staff continues to provide these
jurisdictions with GHG emissions and transportation data as requested.
PAG's GHG inventories include emissions from private and commercial vehicles, public transit,
government fleet travel and employee commuting, as well as energy use and waste disposal
emissions. Private/commercial vehicle emissions are responsible for 99 percent of the region's
transportation emissions. Private and commercial VMT data used in the inventories was
developed by PAG's transportation planners using their travel demand model. Inventory
commuter data are obtained from PAG's annual employee survey. Public transit VMT by fuel
type is supplied by the various transit providers. To develop GHG inventories and ongoing GHG
analyses, PAG staff uses the Clean Air and Climate Protection (CACP) model and the Climate
and Air Pollution Planning Assistant (CAPPA) software developed by the International Council
for Local Environmental Initiatives (ICLEI) and others.
The City of Tucson formed the Citizen's Climate Change Advisory Committee in response to the
adoption of the U.S. Mayor's Climate Protection Agreement by Tucson's mayor in 2006. The
committee developed a report, Action Plan Tucson, Phase One Climate Mitigation, passed by
the Tucson's Mayor and Council in December 2011.
(http://cms3.tucsonaz.gov/sites/default/files/ocsd/act phasel report final 6dec11.pdf)
PAG staff conducted the GHG emissions modeling and analyses that were the basis for this
report and were instrumental in the development of the technical support document, Community
ICF International 17 U.S. Environmental Protection Agency
10-000
-------
Estimating Emission Reductions from Travel Efficiency Strategies: Three Sketch Modeling Case Studies
Economic Security and Climate Action Analysis, compiled by Westmorland. Both documents
include numerous strategies to reduce emissions through transportation efficiency and land use
strategies. Similarly, when the Pima County Board of Supervisors adopted a resolution to
develop the Five-year Action Plan, implementing wide-ranging strategies aimed at reducing
government GHG emissions to 2007 levels by 2020, PAG's GHG inventory data served as the
baseline for County government emissions. PAG staff continues to supply GHG emission data
and analyses for assessing the plan's progress.
Scenario Development
As the lead agency in the case study, PAG initially expressed an interest in understanding the
impact of a variety of strategies on vehicle miles travelled, GHG emissions, and NOx and VOCs
in PAG's planning area, including parking cost increases (in the downtown University of Arizona
corridor), bike travel increases, bus efficiency improvements, provision of an all-access bus
pass for the populations at two academic institutions, incentives for rideshare, and streetcar
enhancements.
The existing or base year for the PAG analysis is 2010. The future year is 2040 as reflected in
the adopted Regional Transportation Plan (RTP). The BAU scenario provided by PAG included
transportation improvement projects included in the plan; it represents the preferred scenario in
the RTP. Population and employment in 2040 are estimated to nearly double those of 2010;
resulting in substantial increases in travel activity and emissions of some pollutants. Average
travel distances will increase under the plan, but walk share will increase too, with more on
campus housing being built at the University of Arizona.
PAG was unique in its approach to scenario development. Rather than use the analysis in a
cumulative way, building the impacts of one strategy on those of another, the agency wanted to
see the specific impacts of each strategy as a stand-alone policy. The rationale for this
approach was that the region would not likely adopt all of these measures at the same time, and
it would be more meaningful for them to evaluate the benefits of individual strategies.
Much of the PAG data for the TRIMMS analysis was available from the agency's travel demand
model. An important stakeholder that was engaged early in the process was SunTran, the
regional transit agency. Sun Tran's involvement helped PAG provide the data needed for
Scenarios 1 and 3, where transit was a major feature.
Scenario 1 Implement SunTran All Access Pass
This scenario considers how providing an unlimited ride transit pass for almost 90,000 faculty,
staff, and students at the University of Arizona and Pima Community College would affect travel
activity. Purchase of the pass would be mandatory, but would be bundled with student tuition
and employment benefits, with costs possibly shared between the passholders and the
academic institutions. However, the user experience is similar to a case of free transit, since the
purchase is non-discretionary, the cost is largely not perceived by the user, and the marginal
cost of each additional transit trip is zero.
ICF International 18 U.S. Environmental Protection Agency
10-000
-------
Case Study Results
This strategy did not include an increase in transit service, so bus VMT does not change. This
methodology is consistent with the 2010 national study, in which only light-duty vehicle travel is
considered and bus VMT is excluded from the results. The TRIMMS analysis used baseline
mode shares and trip lengths provided by PAG that are specific to the university populations.
The transit pass was modeled as a reduction in transit cost to zero.
Scenario 2 - Expand Employer-Based Incentives for Alternative Commute
About one-third of employees in the region are currently eligible for commute subsidies. This
strategy would increase subsidies as shown in Table 5 with all dollar amounts expressed in
current year dollars throughout the analysis.
The trip subsidy radio buttons in TRIMMS were used, providing a simple apply/do not apply
input. Another option was to use baseline trip costs by mode and adjust these by the subsidy
amounts. The research team initially tested this approach using baseline trip costs estimated by
the research team, since PAG did not have baseline trip costs. The results of that analysis
proved unreasonable, likely due to a combination of incomplete elasticities and underestimated
baseline trip costs. In particular, the TRIMMS model lacks a default cross-elasticity for auto
drive alone travel with respect to carpool/vanpool trip costs (See Section 2.3).
Scenario 3 Implement Bus Rapid Transit (BRT) on Two Corridors
PAG is exploring BRT on two major corridors in the region that bring traffic into and out of
downtown Tucson: Oracle Road, a north-south corridor, and Broadway Boulevard, an east-west
corridor. Although there was no specific data available on the change in transit level of service
associated with the BRT proposals, SunTran estimated that the projects might reduce travel
times and wait times between transit vehicles by 20% on the corridors.
The PAG-supplied daily traffic counts on both corridors were assumed to be roughly equivalent
to traveler population using the assumption of two trips per day (round trips) for this population.
The base year traffic counts were then scaled up using projected regional population growth to
provide the future year BAU estimate of the traveler population using the two corridors.
Scenario 4 - Expand Parking Pricing in the Downtown-University Corridor
This scenario included a doubling of parking prices in a specific downtown-university area along
with expanding the share of parking that is priced in the future year. PAG supplied an estimate
of the share of priced parking in the corridor of interest, along with the average parking price per
trip. The average price of all priced and unpriced parking in the downtown-university area was
estimated as the product of average hourly rates and the proportion of priced parking.
Input parameters are provided in Table 5 for current conditions, a BAU future, and the four
scenarios selected by PAG. Specific input values are provided for the scenarios.
ICF International 19 U.S. Environmental Protection Agency
10-000
-------
Estimating Emission Reductions from Travel Efficiency Strategies: Three Sketch Modeling Case Studies
Table 5. PAG Scenario Details
Scenario
Current Conditions
Business as Usual
Scenario 1 : SunTran
All Access Pass
Scenario 2:
Expanded Employer-
based Incentives
Scenario 3: BRT on 2
Corridors
Scenario 4: Parking
Pricing in Downtown-
University Corridor
Description
Existing conditions across
all strategies in 20 10
2040 conditions with
current levels of transit
pass, employer-based
incentives, BRT coverage,
and parking pricing
Bundle 'free' transit pass
with tuition for faculty,
staff, and students at two
local universities
Increase subsidies by
$10-$50 per mode.
BRT on Oracle Rd and
Broadway Blvd.
Double parking prices and
expand number of priced
spaces.
Geography
Regionwide and
Subarea
Regionwide
Regionwide
Regionwide
Subarea
Subarea
Data Used
mode shares
average vehicle occupancy
average vehicle trip lengths
regional population and employment
TRIMMS default retail establishment density
(0.31 per square mile)
TRIMMS default distance to nearest transit
stations (0.70 miles)
parking pricing
employer-based incentives for alternative
commute modes
mode shares
average vehicle occupancy
2040 regional population and employment
TRIMMS default retail establishment density
(0.31 per square mile)
TRIMMS default distance to nearest transit
stations (0.70 miles)
Parking cost per trip: $2.71/hr. x 80%
population = $2. 17/hr.
current employer-based incentives for
alternative commute modes
average transit trip cost ($1.43)
average transit travel time (72.46 minutes)
average transit wait time (10 minutes)
number of students, faculty, and staff who will
use the pass (86,234)
average transit trip cost ($0)
number of regional employees covered
(236,616)
traveling population affected (62,565)
20% travel time reduction (to 57.97 minutes)
20% headway (wait time) reduction (to 8
minutes)
95% share of all parking (public and private) is
priced
parking cost per trip: $5.41/hr. for priced
parking x 95% of total parking that is priced=
$5.14/hr for all priced and unpriced parking
Additional EPA Land Use Scenario for PAG Region
Unlike MARC and MassDOT, the PAG region did not select a land use strategy as part of its
four scenarios. Given the evidence that land use strategies provide an important base for other
travel efficiency strategies, a land use strategy was independently developed by EPA for the
ICF International
10-000
20
U.S. Environmental Protection Agency
-------
Case Study Results
PAG region. This land use strategy was analyzed in an additional scenario with the four
strategies selected by PAG: a Transit All Access Pass, Expanded Employer-based Strategies,
BRT on Two Corridors, and Downtown-University Corridor Parking Pricing.
The land use scenario was defined based on an existing vision for the region: the Imagine
Greater Tucson (IGT) preferred land use scenario. IGT developed this scenario through a two-
year, community-based visioning process in which residents from across the greater Tucson
region communicated their thoughts and opinions on how the region should develop and grow.
Over 10,000 people participated in the process and the results show a strong regional desire for
smarter growth options.
The regional population growth assumed in PAG's RTP 2040 scenario is roughly equivalent to
that assumed in IGT's preferred scenario, and thus provides a consistent reference point for
comparison. The IGT scenario concentrates population growth in existing urban centers.
PAG provided data on population densities for all traffic analysis zones (TAZs) in the modeling
region for both the IGT scenario and PAG's RTP 2040 scenario. The weighted average
population density9 of the IGT scenario is roughly 50% higher than that of the RTP scenario.
Therefore the land use scenario was modeled as a 50% increase in residential density for the
region. Since higher residential densities would cluster development in central areas with better
transit access, walking distance to the nearest station was reduced by a conservative 10%.
Table 6 provides the results of the analysis of this additional scenario. Adding the land use
option increases total region VMT reductions by only 0.25% on top of the other strategies. This
result is very small compared to the results of previous analyses of land use strategies using
TRIMMS 2.0, and to values estimated in other studies. CUTR has suggested that this result
may be due to the relatively low level of transit service in the PAG region, as land use results in
TRIMMS are driven by transit accessibility. See Section 4.2 for a further discussion of land use
analyses with TRIMMS.
Table 6. Resulting VMT and Emissions Changes for Selected Pollutants
Scenario
Percent Change - 2040 BAD compared to 2040 Scenario
Light-Duty VMT GHGs(C02 PM2.5 NOx VOC
equivalent)
Land use changes plus PAG scenarios 1-4
-1.95%
-1.87%
-1.69°
-1.43% -0.71%
9 Weighted average population density summarizes regional land use densities in a way that is more representative of residents'
daily experience of land use patterns than average density. TAZs with higher populations, which tend to be denser areas, are
weighted more heavily than TAZs with lower populations. In the weighted average density calculation, the density of each TAZ
in the region is weighted by its proportion of total regional population. An increase in weighted average density does not
indicate an increase in total regional population, but rather a shifting of population toward higher density centers.
ICF International
10-000
21
U.S. Environmental Protection Agency
-------
Estimating Emission Reductions from Travel Efficiency Strategies: Three Sketch Modeling Case Studies
Emissions Analysis
MOVES2010b was used to determine appropriate emission factors for the PAG regional
analysis. Based on information from PAG that 98% of the county's population live within the
MPO boundary, the domain was taken as the entirety of Pima County. Input data was provided
by PAG and derived from several local, state, and national sources. Table 7 summarizes the
MOVES inputs PAG provided and the data used in the final modeling. Table 7 also indicates
any modifications to the region-provided data before modeling.
Table 7. PAG Data Sources
Region Provided
Data
Base Future
Year Year
Data Used in Final Modeling
Future
Year
Source Type (vehicle)
Population
Vehicle Type VMT
Road Type Distribution
Meteorological Data
Age Distribution
VMT Hour
Fractions Day
Month
Average Speed
Distribution
Ramp Fraction
Fuel Supply
Formulation
I/M Program
AVFT
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
provided data
provided data
provided data
provided data
provided data
MOVES defaults
provided data
provided data
provided data
MOVES defaults
provided data
provided data
modified data: Changed USEIMYN
field to N, so that I/M data were not
used in modeling.
MOVES defaults
Calculated data. Applied the 2010
population/VMT ratio to the 2040
population to get 2040 HPMS
VMT totals.
provided data
provided data
provided base year data (changed
YEAR to 2040)
provided base year data (changed
YEAR to 2040)
MOVES defaults
provided data
provided data
provided data
MOVES defaults
provided data
provided data
modified data: Changed
USEIMYN field to N, so that I/M
data were not used in modeling.
MOVES defaults
PAG worked closely EPA to determine the locally representative data for the required MOVES
inputs. Site-specific data for PAG were very complete, being derived from local, state, and
federal modeling and datasets. This is a case of geographically specific strategies that require
data specific to that geographic sub-area.
For example, PAG used local data to develop inputs for inspection/maintenance, speed
distribution, fuels, and meteorology. During review, concerns were raised about projections of
these fields to future years and their effect on results. Ultimately inspection/maintenance data in
the analysis was removed due to uncertainty of far-future program design and efficacy on future
vehicles and the concern that uncertain I/M influences on emissions could mask other trends.
This approach was implemented in all three regional analyses. Baseline year meteorological
ICF International
10-000
22
U.S. Environmental Protection Agency
-------
Case Study Results
data provided were also used for the future year, as requested by PAG. Speed distribution
values were extracted from the regional TDM modeling data.
PAG also obtained current information from state agencies to develop model inputs. Current
fuel information was obtained from the Arizona Department of Weights and Measures (model
default fuels were used for future year analyses). Distributions of road type VMT and vehicle
ages came from the Arizona Department of Transportation (ADOT). Vehicle population data
were developed from ADOT vehicle registrations for Pima County with methods based on EPA
guidance10 for future year vehicle populations. In addition, appropriate local vehicle/road type
VMT information and baseline year age distribution data was developed from PAG data, in
close coordination with EPA. PAG also coordinated with EPA in use of EPA's MOVES converter
tools to derive and/or reformat VMT allocation and speed distribution data, including future year
vehicle populations, use of compressed natural gas (CNG) fueled transit buses, and day-type
activity factors to create an accurate regional input dataset.
As discovered during initial model runs, when the analysis year is a leap year and the MOVES
inputs are selected to pre-aggregate the input data to annual resolution at the county scale, the
analyses must be performed with hourly aggregation of inputs.11 Appendix B discusses this
further. No preaggregation was used for the analysis.
Results
The hourly MOVES outputs (for each hour of each day type of each month and year, accounting
for the number of hours in each) were manually post-processed to produce annual total
emissions (g) and activity (starts or miles). The data were then aggregated to the TRIMMS
vehicle types and emission factors were calculated as total emissions divided by total activity,
as discussed previously. Table 8 shows the resulting emission factors.12
Table 8. Emissions Factors for PAG
g/start
Base Year
Future Year Base Year Future Year
Auto (Motorcycles+Passenger Cars+Passenger Trucks)
GHGs (C02-
equivalent)
NOx
PM2.5
VOCs
377.22 309.10 117.93 74.88
0.84
0.02
0.16
0.14
0.01
0.03
1.47
0.03
2.12
0.35
0.01
0.67
10 Using MOVES for Estimating State and Local Inventories of On-Road Greenhouse Gas Emissions and Energy Consumption -
Final (EPA-420-B-12-068, November 2012).
11 Note that this issue would not arise in MOVES runs done for official SIP or conformity purposes, as it is not acceptable to pre-
aggregate over time for these purposes.
12 Note that final results omitted Transit emissions. Accordingly, the transit emission factors are not presented in Table 8.
ICF International
10-000
23
U.S. Environmental Protection Agency
-------
Estimating Emission Reductions from Travel Efficiency Strategies: Three Sketch Modeling Case Studies
Vanpool (Passenger Trucks+Light Duty Trucks)
GHGs (C02-
equivalent)
NOx
PM2.5
VOCs
512.47
1.80
0.04
0.33
383.41
0.36
0.01
0.06
162.37
2.36
0.04
3.70
87.48
0.44
0.01
0.62
PAG Scenario Results
Table 9 provides the results of the analysis in terms of total regional impacts. It is important to
note that regional reductions are often very modest when compared to the impact on the
population targeted by the strategies. Additional explanation is provided on the following page
and numerical results are included in Appendix B.
Table 9. PAG Regionwide Percent VMT and Emissions Changes
Scenario
Percent Change - 2040 BAU compared to 2040 Scenario
Light-Duty VMT GHGs(C02 P
equivalent)
scenario i . ourr i ran MM Access rass
Scenario 2: Expanded Employer-based
Incentives
Scenario 3: BRT on 2 Corridors
-0.99%
-0.43%
-0.02%
-U.3/70
-0.43%
-0.02%
Scenario 4: Parking Pricing in Downtown- n 9RO/ -0.25%
University Corridor -u.^b/o
-u.a^r/o
-0.42%
-0.02%
-0.25%
-U.O070
-0.40%
-0.02%
-0.24%
-U.I 170
-0.44%
-0.02%
-0.26%
The SunTran pass extension has a large impact on the affected population with shifts to transit
from SOV: close to a 16% VMT reduction for 90,000 people. More of the mode shift will happen
at the University of Arizona, which offers higher transit accessibility. There is the potential to
expand to other populations as well, such as residents of downtown and other transit accessible
areas.
Employer-based incentives show a more moderate impact on affected population (3% VMT
reduction for approximately 240,000 people) with shifts from SOV to carpool, vanpool, transit
and cycling.
BRT on two corridors that traverse the region, intersecting in downtown Tucson, shows a small
impact on the affected population with only a 0.5% VMT reduction in shifts to transit. This result
could be affected by more extensive transit improvements and land use changes.
Parking pricing limited to a small downtown area indicates a moderate impact on the affected
population (2% reduction for 400,000 people traveling downtown) with shifts from SOV to
carpool, transit, cycling, and walking.
Overall the PAG scenario analysis represents a pragmatic application of potential strategies in
the region. Stakeholders were highly interested in feasible scenarios instead of making
ICF International
10-000
24
U.S. Environmental Protection Agency
-------
Case Study Results
aggressive assumptions that have a remote chance of being enacted. PAG's selected strategies
were limited to specific sub-geographies and sub-populations, whereas other regions applied
strategies more broadly and used more aggressive assumptions. PAG also chose not to
analyze a land use strategy. Other regions included land use as a foundational strategy in
several scenarios, reflecting an understanding of land use as a key strategy to maximize
alternative mode options and minimize travel distances. As with any region, the aggressiveness
of scenarios is an important decision for lead agencies to make when undertaking a "what if"
analysis of this type. The selection of sub-areas and limited affected populations, as well as the
lack of a land use foundation for the transit and BRT scenarios contributed to the modest
outcome.
PAG has taken a strong interest in the details of this analysis, particularly with respect to the
MOVES analysis. This level of involvement and data availability may encourage a repeat of the
analysis with changes to the assumptions as an internal exercise in the future. This approach
would allow PAG to test impacts before involving a larger group of stakeholders.
ICF International 25 U.S. Environmental Protection Agency
10-000
-------
Estimating Emission Reductions from Travel Efficiency Strategies: Three Sketch Modeling Case Studies
3.2. Massachusetts Department of Transportation (MassDOT) - Boston
Region
Background
The lead agency for the second regional case study is MassDOT. MassDOT's case study
partners include the Metropolitan Area Planning Council (MAPC), the Central Transportation
Planning Staff (CTPS, the Boston MPO (BMPO) staff), and the Georgetown Climate Center
(GCC). This stakeholder group indicated that HOV lanes, carpool/vanpool incentives, rideshare
matching programs, bicycle programs, transit expansion, and smart growth are strategies of
priority for the case study.
The BMPO region is the focus of this modeling study, covering 101 cities and towns and
approximately 1,405 square miles in the greater Boston region. The region covers more than
three million people and includes diverse communities from relatively rural to urban. The base
year for the MassDOT regional analysis is 2009 with a future year of 2035. Population and
employment data is from the community model domain, which is slightly larger and includes 164
towns and cities in eastern Massachusetts.
Interest in Participation
MassDOT expressed an interest in this study based on the state's strong interest in climate
change and GHG emissions reduction. MassDOT has an internal agency GHG emissions target
of a 40% reduction from a 2002 baseline (in the GreenDOT Program).The Commonwealth has
a general target for GHG reductions across the economy of 25% below 1990 levels, established
pursuant to Massachusetts' climate change legislation, the Global Warming Solutions Act.
Under the general target, Massachusetts has also established a target for the transportation
sector of a 7.1% reduction from a 1990 baseline by 2020.
MassDOT is a participant in the Transportation and Climate Initiative (TCI). The TCI is a
regional collaboration of transportation, environment, and energy agencies in 11 Northeast and
Mid-Atlantic states, as well as the District of Columbia, that seeks to develop the clean energy
economy and reduce GHG emissions in the transportation sector. One of TCI's stated goals is
to promote sustainable communities. This project is intended to directly support TCI's work in
that area and facilitate the sharing of lessons learned from this effort with other states and
metropolitan areas in the region.
Prior Experience with Analysis of GHG Emissions
MassDOT worked with other state agencies and consultants to develop the Massachusetts
Clean Energy and Climate Plan 2020, which identified emission reduction potentials from
specific strategies in the transportation sector, including from implementation of Sustainable
Development principles, implementation of MassDOT's GreenDOT program, and
implementation of a Smart Growth policy package. MassDOT has also recently released its
GreenDOT implementation plan, identifying the actions MassDOT plans to take to achieve its
sustainability goals, including GHG emissions reductions.
ICF International 26 U.S. Environmental Protection Agency
10-000
-------
Case Study Results
The Central Transportation Planning Staff (CTPS) has been incorporating climate change into
the planning and programming activities of the MPO through the LRTP and the TIP. The MPO
uses its visions and policies established in the LRTP in the project selection process for both
documents. Project evaluation criteria are used to identify projects that will help slow climate
change through reduction in GHGs and projects that will help adapt to the effects of climate
change by addressing flooding or improving evacuation routes. Several discussion papers
developed by the MPO previous to this project have considered how projects that reduce GHG
emissions can be supported through their programs.
MassDOT is developing performance measures for climate change through the weMove
Massachusetts and GreenDOT processes, and the Boston MPO is developing performance
measures to demonstrate the region's success in managing their transportation network and the
effectiveness of investments in moving towards its visions. The MassDOT and MPO
performance measures will be coordinated as required in federal transportation law.
Scenario Development
MAPC developed two land use forecasts for its most recent regional transportation plan, a
Current Trends forecast and the MetroFuture forecast. The Current Trends would continue
recent historical patterns of land use dispersal in the Boston region, while MetroFuture would
concentrate future growth in core areas well-served by transit. In 2008, MAPC adopted
MetroFuture as the forecast for its 2030 regional land use plan, which includes supportive
implementation strategies. The Boston Region MPO (BRMPO) then adopted the MetroFuture
land use scenario in its most recent RTP (adopted in 2011).
All BAU inputs provided by the MassDOT stakeholder group represent a combination of the
MetroFuture land use scenario and a no-build transportation scenario. Therefore the BAU
scenario already represents a shift from current trends to more smart growth development
patterns. The land use scenario incorporated in the BAU forecast clusters more development
around transit while no new transit service is provided. Thus, the average transit trip length
declines. In addition, mode shares tend to shift from driving to transit and walking given the
assumption about new development. Trip lengths for driving trips increase slightly in the future
due to some continued dispersal of land uses in the outer areas, despite the emphasis on
densification.
The BAU also includes the existing MassRIDES program, the statewide travel options program
that partners with employers to provide information about commuting by carpooling, bicycling,
walking, public transportation, teleworking, and vanpooling. About 700,000 employees in the
Boston region currently have access to MassRIDES programs, which include guaranteed ride
home and ride match, telework and flex work programs, and program marketing. Approximately
100,000 of the 700,000 have access to monetary subsidies for healthy modes. For the BAU
scenario input, the number of covered employees was scaled by the natural rate of regional
employment growth.
ICF International 27 U.S. Environmental Protection Agency
10-000
-------
Estimating Emission Reductions from Travel Efficiency Strategies: Three Sketch Modeling Case Studies
Scenario 1 Expand Programs and Incentives for Healthy Modes
The first MassDOT strategy is based on the existing MassRIDES program, and would increase
the number of employees with access to MassRIDES by 25% - giving them access to
guaranteed ride home and ride match, telework and flex work programs, and program
marketing. The strategy would also expand access to employer-paid monetary subsidies to all
employees with access to MassRIDES, who would each receive a subsidy of $70 a month.
This scenario was modeled as two separate strategies: 1) effect on new MassRides users
(monetary subsidies and non-monetary benefits) and 2) effect on existing MassRides users
(monetary subsidies only).
For monetary subsidies, the baseline trip costs for all modes provided by MassDOT were used,
and reduced according to the subsidy amounts, with some modifications. This is in contrast to
the radio-button method used for a similar strategy for the PAG region. The actual dollar
amounts were able to be used because MassDOT could provide reasonable baseline trip costs
for transit and vehicle trips, which included maintenance, tires, gas, and oil. Bicycling and
walking subsidies were not explicitly modeled through this method; since both the baseline and
final costs of these bicycling and walking trips is assumed to be zero.
For non-monetary benefits, the radio buttons were used for TDM programs and subsidies for
Guaranteed Ride Home and Ride Match and Telework and Flexible Work Schedules.
Scenario 2 - Expand Programs and Incentives for Healthy Modes with Smart Growth Land Use
Scenario 2 adds Smart Growth Land Use to the Scenario 1 strategy. This included an increased
emphasis on growth in existing urban centers, in order to increase weighted average residential
density, and an increase in land use mixing.13 The Smart Growth Land Use Strategy
represents an increase in weighted average density over that of the MetroFuture forecast, which
is incorporated in the BAU. It is therefore not based on any prior detailed analysis or input
process. For the strategy, MassDOT suggested doubling the increase in weighted average
density between the MetroFuture (BAU) and base year. Although this represents a highly
ambitious strategy, it is appropriate for a TEAM analysis of alternative "what if" scenarios.
Population-weighted average density by TAZ was calculated for the entire region for the base
year and the MetroFuture 2035 scenario. There is a projected 11 % increase in density over this
timeframe. Doubling this increased density relative to the base year results in a 22% increase.
Population-weighted average density of retail establishments by TAZ was calculated for the
entire region for the base year and the MetroFuture 2035 scenario, and the results showed
13 Weighted average population density summarizes regional land use densities in a way that is more representative of residents'
daily experience of land use patterns than average density. TAZs with higher populations, which tend to be denser areas, are
weighted more heavily than TAZs with lower populations. In the weighted average density calculation, the density of each TAZ
in the region is weighted by its proportion of total regional population. An increase in weighted average density does not
indicate an increase in total regional population, but rather a shifting of population toward higher density centers.
ICF International 28 U.S. Environmental Protection Agency
10-000
-------
Case Study Results
there is effectively no density increase projected. MassDOT instructed that the scenario should
increase density relative to the base year by 10%.
MassDOT did not have an immediately available method to estimate the change in average
walking distance to transit, so this was not modeled in TRIMMS. However, with the increased
density envisioned in this scenario, this distance would decrease. Had it been modeled, it may
have reduced VMT and emissions further.
The TRIMMS model run for this scenario only includes land use changes, and the entire
regional population is affected. The Healthy modes strategy was added in as a post-processing
step. For residential population density, MassDOT provided the figures 10,205 people per
square mile (BAU) and 11,230 people per square mile (scenario). TRIMMS default density
figures are lower, because they are based on an unweighted density calculation, which gives
equal weight to low and high density TAZs. Adjustments to the default baseline density in
TRIMMS are not easily customizable; however, since the percentage increase drives the
modeling results, the calculated increase can be applied to the TRIMMS default figure.
Therefore an increase of 20% was applied to the TRIMMS default baseline. (The actual
increase calculated was 22%, but TRIMMS only allows increases in 5% increments).
For retail establishment density, MassDOT provided estimates of 698 establishments per
square mile (BAU) and 762 establishments per square mile (scenario). The TRIMMS default
value was increased by 10% for the scenario.
Scenario 3 Expand Programs and Incentives for Healthy Modes with Smart Growth Land Use and
HOV Lanes
The scenario adds HOV lanes to the previously analyzed strategies with a decrease in
rideshare travel time by 10% for entire region through a network of HOV lanes. The HOV lanes
strategy is not based on any specific proposal or plan. Parameters represent a hypothetical
scenario, developed through internal conversations at MassDOT. The assumptions appear to
be reasonable, given that they represent roughly a return to the 2009 base year travel time
value.
The TRIMMS model run for this scenario includes HOV lanes and land use changes (Scenario
2), since these both affect the entire regional population. As in Scenario 2, Healthy modes was
added in as a post-processing step. The resulting reduction in SOV VMT reported was less than
the increase in rideshare VMT; a counter-intuitive result. In order to capture the reduction in
SOV VMT, rideshare VMT was held constant in post-processing. The response of TRIMMS to
this analysis will be discussed more fully in the conclusions and recommendations.
Scenario 4 - Expand Programs and Incentives for Healthy Modes with Smart Growth Land Use, HOV
Lanes and Transit Network Expansion and Improvement
The final MassDOT scenario adds transit network expansion and improvement to the previous
analyses. This strategy is expected to reduce both transit trip times and access times (wait
times) for regional population by 10%. The strategy is not based on any specific regional
proposal or plan. Parameters represent a hypothetical scenario, developed through internal
ICF International 29 U.S. Environmental Protection Agency
10-000
-------
Estimating Emission Reductions from Travel Efficiency Strategies: Three Sketch Modeling Case Studies
conversations at MassDOT. MBTA has approximately 15 key bus routes, some or all of which
could be upgraded to a BRT level of service, as part of the strategy. The model run includes
transit network, land use changes (Scenario 2) and HOV lanes (Scenario 3). As in Scenarios 2
and 3, Healthy modes were added in as a post-processing step.
Input parameters are provided in Table 10 for current conditions, a BAU future, and the four
scenarios selected by MassDOT. Specific input values are provided for the scenarios.
ICF International 30 U.S. Environmental Protection Agency
10-000
-------
Case Study Results
Table 10. MassDOT Scenario Details
Scenario
Current Conditions
Business as Usual
Scenario 1 : Expanded
Healthy Modes
Program
Scenario 2: Scenario 1
+ Land Use
Scenario 3: Scenario 2
+ HOV Lanes
Scenario 4: Scenario 3
+ Expanded Transit
Description
Existing conditions across all
strategies in 2009
2035 conditions with current
levels of employer program,
land use, HOV lanes, and
transit; future growth
assumed to be focused in
areas served by transit (the
"Metro Futures" land use
forecast)
Expand the statewide travel
options program that
partners with employers to
provide information about
commuting by alternate
modes of transportation.
Increase residential density
and mixed use land uses in
selected areas.
Add HOV lanes.
Expand transit network and
improve transit infrastructure.
Geography
Regionwide
Regionwide
Regionwide
Regionwide
Regionwide
Regionwide
Data Needs
mode shares
average vehicle occupancy
average vehicle trip lengths
TRIMMS default vehicle ownership (1.72
vehicles/household)
regional population and employment
employer-based incentives for alternative
commute modes
mode shares
average vehicle occupancy
average vehicle trip lengths
2035 regional population and employment
current employer-based incentives for
alternative commute modes
TRIMMS default vehicle ownership (1 .72
vehicles/household)
current travel times (23.4 minutes for
rideshare and 24.5 minutes for transit)
current transit access times (10.5 minutes)
current trip costs ($2.08 for rideshare, $4.00
for vanpool, and $1 .60 for transit)
residential population density (10,205 persons
per square mile)
retail establishment density (698 per square
mile)
number of regional employees covered
(652,565 existing users, 152,265 new users)
average monthly subsidy offered to
employees ($70 for rideshare, vanpool, and
transit)
are guaranteed ride home, ride match,
telework, and flexible work schedules offered?
22% increase in population density (to 1 1 ,230
persons per square mile)
10% increase in retail establishment density
(to 762 per square mile)
rideshare travel time reduction (10% reduction
to 21.1 minutes)
1 0% transit travel time reduction (22.1
minutes)
10% headway (wait time) reduction (to 9.5
minutes
ICF International
10-000
31
U.S. Environmental Protection Agency
-------
Estimating Emission Reductions from Travel Efficiency Strategies: Three Sketch Modeling Case Studies
Emissions Analysis
As previously described, MOVES2010b was used to determine appropriate emission factors for
the MassDOT scenario analysis. This region was modeled as a single-zone, custom domain to
capture the BMPO region of jurisdiction and best utilize the available input data:14
The BMPO regional boundary does not coincide well with county boundaries, including all or
parts of seven counties. The MOVES analysis experience and capability in the region is well
advanced, including previous modeling for several years and detailed inputs ready for use, but
based on data for a single county, Middlesex. Their MOVES experience has focused on
emission rates for this county for CAA purposes instead of an emission inventory approach as
used in this study.15 With this approach, emission rates from the representative county can be
applied to activity from other counties, as long as they have the same fuels and I/M program as
the representative county. That is the case for this region. Therefore most inputs provided were
based on Middlesex County, except source population and total VMT (the two MOVES inputs
that must be totals), which CTPS provided for the whole MPO rather than split into each of the
seven counties.
Given the available data, it made sense to model the region as a "custom domain" in MOVES
rather than model the counties individually. In the county scale, the custom domain option
allows the user to model a multi-county area using one run. Executing a custom domain
requires that some MOVES inputs be formatted slightly differently than county-scale inputs, but
otherwise the run execution is similar to that of a county-scale run. Table 11 summarizes the
MOVES inputs that MassDOT provided and the data used in the final modeling. Table 11 also
indicates if any modifications to the region-provided data were made before modeling.
Table 11. MassDOT Data Sources
Region Provided
Data
Base Future
Year Year
Data Used in Final Modeling
Base Year
Future Year
Source Type Population
veiiiuie lype vivn
Road Type Distribution
Meteorological Data
A
X
X
A
X
X
*provided data
*provided data with change:
calculated VMT from provided data
by annualizing, summing
auto+truck, and allocating across
source types using factors from
region.
provided data
provided data with change:
ZONEID was changed to 1
*provided data
*provided data with change:
calculated VMTs from provided
data by annualizing, summing
auto+truck, and allocating across
source types using factors from
region.
provided data
provided data with change:
ZONEID was changed to 1
14 http://www.ctps.org/Drupal/mpo. Other options include single county (as for PAG), or series of individual counties (as for
MARC).
15 Either MOVES calculation type, Inventory or Emission Rates, is acceptable for regulatory purposes; the emission factors within
the model are the same regardless of calculation type chosen.
ICF International
10-000
32
U.S. Environmental Protection Agency
-------
Case Study Results
Region Provided
Data
Base Future
Year Year
Data Used in Final Modeling
Age Distribution
Base Year
provided data with change:
Future Year
provided data with change:
VMT
Fractions
Average Sp
Distribution
Ramp Fract
Fuel
Hour
Day
Month
eed
on
Supply
X
x
X
X
X
X
YEARID was changed to 2009
x provided data
x
x
x
x
x
provided data
provided data with change:
ISLEAPYEAR was changed to N
provided data
provided data
"provided data with change:
FUELYEARID was changed to
2009 and COUNTYID was
changed to 99001
Formulation x x provided data
YEARID was changed to 2035
provided data
provided data
provided data
provided data
provided data
"provided data with change:
COUNTYID was changed to
99001
provided data
I Program
AVFT
**Zone
*Zone Road Type
'provided data with change:
YEARID was changed to 2009,
STATEID was changed to 99,
COUNTYID was changed to
99001, and USEIMYN was
changed to N.
MOVES defaults
All factors set to 1 because
modeling used only one zone
"provided data with change:
YEARID was changed to 2009,
STATEID was changed to 99,
COUNTYID was changed to
99001, and USEIMYN was
changed to N.
All factors set to 1 because
modeling used only one zone
MOVES defaults
All factors set to 1 because
modeling used only one zone
All factors set to 1 because
modeling used only one zone
* MassDOT provided inputs for Source Type Population and Vehicle Type VMT for the entire domain in aggregate (not subset by
county). All other provided inputs were specific only to Middlesex County.
** Because the MassDOT MOVES modeling used a custom domain, a zone (with arbitrary ID '1') was used in the modeling,
requiring the ZONEID to be set to 1 in the meteorology data and arbitrary STATEIDs and COUNTYIDs of 99 and 99001,
respectively, in inspection/maintenance and fuel supply inputs. Custom-domain runs also require the Zone and
ZoneRoadType input sheets, which were not provided by MassDOT but contained zone allocation factors that were all set to 1
because the domain included only one zone.
Lessons Learned
The MassDOT case study provides a good example of how a TEAM analysis can be done when
the region of interest covers multiple counties but the available data is not provided for each
county. MassDOT provided most of the other MOVES inputs for a single county in the domain
(Middlesex County), which represents the seven-county region in terms of the MOVES
parameters such as fuels used, fleet age distribution, road type distribution, etc. Although much
of the Middlesex County data was provided for year 2012, it was taken as representative of
2009, so the value of the YEAR field was changed to 2009 where necessary. Age distribution
data were only provided for the base year, so the same data were reused for future year by
changing the YEAR field. Ultimately inspection/maintenance (I/M) was excluded from the model
results by setting the USEIMYN field to 'N'. MOVES default data were used for alternative fuels
ICF International
10-000
33
U.S. Environmental Protection Agency
-------
Estimating Emission Reductions from Travel Efficiency Strategies: Three Sketch Modeling Case Studies
and technology. This was done to remove uncertainty in future program design and efficacy of
I/M for future vehicles, as well as the concern that uncertain I/M influences could mask other
trends.
To capture the entire domain with input data provided for a single county, we ran the MOVES
model, specifying inputs for a custom modeling domain that covers the entire region of the study
(parts or whole of 7 counties: Middlesex, Suffolk, Norfolk, Plymouth, Essex, Worcester, Bristol)
using provided inputs for Middlesex that are applicable to the entire region and provided
regional totals for other fields (population and VMT) as appropriate. The setup for a custom-
domain run is nearly identical to that of a county-level run, and documented in the MOVES
User's Guide,16 except for some minor reformats of some of the input sheets. It also uses a
different Geographical Bounds specification and two additional input sheets (Zone and
ZoneRoadType) with allocation factors that are all set to 1 for a single-zone custom domain
such as the application here.
Summary of Results
The post-processing of hourly MOVES outputs to domain-wide, average emission factors for the
MassDOT case study were calculated as described previously. Total running emissions from
hourly outputs (in grams per year) were divided by total running activity (in miles per year) to
produce regional and annual average, g/mile emission factors. A similar analysis was made for
starting emissions. The resulting emission factors, in grams of pollutant per average mile driven
or grams of pollutant per average start are shown in Table 12.
Table 12. Emission Factors for MassDOT
g/start
Base Year Future Year Base Year Future Year
(2009) (2035) (2009) (2035)
Auto (Motorcycles+Passenger Cars+Passenger Trucks)
GHGs (C02-equivalent)
NOx
PM2.5
VOCs
418.16
0.85
0.02
0.16
316.97
0.17
0.02
0.03
160.18
1.77
0.04
2.68
106.41
0.46
0.02
0.71
Vanpool (Passenger Trucks+Light Duty Trucks)
GHGs (C02-equivalent)
NOx
PM2.5
VOCs
510.61
1.28
0.03
0.23
373.36
0.31
0.02
0.05
191.76
2.46
0.04
3.33
118.59
0.52
0.02
0.61
16 Motor Vehicle Emission Simulator (MOVES): User Guide for MOVES201 Ob, EPA-420-B-12-001b, June 2012.
ICF International
10-000
34
U.S. Environmental Protection Agency
-------
Case Study Results
MassDOT Scenario Results
Table 13 provides the regional results of the analysis. The MassDOT results are similar to the
results of the 2010 national study for regions similar to Boston. Note that regional reductions are
often modest when compared to the impact on the population targeted by the strategies where
the impact is much greater. Additional explanation is provided on the following page and
numerical results are included in Appendix B.
Table 13. MassDOT Regionwide PercentVMT and Emissions Changes
Percent Change for 2035 BAD compared to 2035 Scenario
Scenario
Scenario 1: Expanded Healthy Modes
Light-Duty
VMT
-2.80%
GHGs
(C02
equivalent)
-2.80°
-2.80%
-2.79%
-2.77%
Program
Scenario 2: Scenario 1 + Land Use
Scenario 3: Scenario 2 + HOV Lanes
Scenario 4: Scenario 3 + Expanded Transit
-3.89%
-4.07%
-4.41%
-3.89%
-4.06%
-4.41%
-3.88%
-4.06%
-4.40%
-3.88%
-4.05%
-4.39%
-3.84%
-4.02%
-4.36%
The Healthy Modes scenario shows a large impact on the affected population with about 16%
reduction in VMT for existing MassRides users (650,000 people) and 20% reduction in VMT for
new users (150,000 people). The shifts occur from drive-alone to rideshare, cycling and walking.
When the entire population of the region is considered rather than just the affected population,
the results are more modest, as shown in Table 13.
For the land use strategy the VMT impacts are much greater for MassDOT than they were for
PAG, and more consistent with current literature. Land use impacts may be more reasonable for
MassDOT because Boston is a transit rich region, and the TRIMMS land use algorithms focus on
transit ridership as the impact of land use densification. There is a moderate impact on regional
population (1% VMT reduction) that results from a reduction of drive-alone trips. However, as
noted earlier, the change in average walking distance to transit that resulted from the land use
strategy was not modeled; had it been, greater reductions may have been seen.
HOV Lanes show small impacts with shifts from drive-alone to carpool.
Transit network improvements also show small impacts. While Boston has a robust transit
system, transit still accounts for only 7-8% of trips. Therefore impacts of transit improvements
are applied to a small baseline.
The HOV strategy represents a reduction in rideshare travel times. The TRIMMS analysis showed
a reduction in SOV VMT less than the increase in rideshare VMT; which is unexpected. This is a
result of TRIMMS' lack of controls on trip totals for analysis of pricing strategies (discussed
further in Section 4). This was adjusted in post-processing by removing the increase in VMT due
to rideshare. Making this adjustment produces a reduction in VMT that is consistent with the
literature reviewed for the previous TEAM analysis.
ICF International
10-000
35
U.S. Environmental Protection Agency
-------
Estimating Emission Reductions from Travel Efficiency Strategies: Three Sketch Modeling Case Studies
Non-SOV travel subsidies can be effectively modeled using dollar values rather than the radio
buttons if reasonable baseline trip costs and elasticities are supplied. The dollar value approach
was used for MassDOT. This approach will tend to produce larger VMT reductions in TRIMMS
compared to the radio buttons, especially if combined with non-monetary employer-based
programs. TRIMMS' standard functions for employer-based strategies, which use on/off buttons
for subsidies, use more conservative assumptions to model impacts. However, use of actual
dollar amounts allows the impact to vary with the subsidy amount.
ICF International 36 U.S. Environmental Protection Agency
10-000
-------
Case Study Results
3.3. Mid-America Regional Council (MARC) - Greater Kansas City
Background
MARC is a coordination entity and the metropolitan planning organization for the bistate region
of Kansas City (Kansas and Missouri). The MARC region consists of 9 counties and 120
municipalities, but the transportation planning boundary is slightly smaller and the air quality
boundary is even smaller. The staff works with two different state DOTs and two different air
quality offices.
MARC is responsible for the long range regional transportation plan for the region, as well as
many other planning and coordination initiatives. The LRTP, Transportation Outlook 2040, was
adopted by the MARC Board of Directors and included, for the first time, a GHG goal as a
fundamental element of the policy framework. During the same timeframe, the region's
voluntary Clean Air Action Plan underwent a comprehensive update and included co-benefit
analysis of recommended measures including reduction of GHGs. MARC is in danger of
violating the 2008 75ppb ozone standard after the completion of the 2012 ozone season and
understands the importance of credible assessments of travel efficiency strategies to help attain
the federal standards and estimate the additional benefits of GHG emission reductions.
Interest in Participation
For this study, MARC expressed an interest in investigating a combination of land use controls,
transit-oriented development and smart growth; expansion of transit service; rideshare; and
road pricing. The agency would like to be able to use the results from this study to inform the
LRTP.
The MARC region is already active in operational improvements using strategies such as ramp
metering, ITS infrastructure, signal prioritization and others. Most travel strategies for the region
relate to congestion mitigation rather than VMT reduction. Although their goal in the past has
been to build more lanes to add capacity, this study represents a shift in focus in order to get
more support for VMT reduction. In addition, MARC anticipates the potential need to develop a
SIP, and this study could identify specific measures to include.
Prior Experience with Analysis of GHG Emissions
Transportation Outlook 2040 included two alternative 2040 land-use scenarios that were
subjected to several analyses, including trip travel time, roadway congestion, and cost of new
infrastructure. Parsons Brinkerhoff evaluated the scenarios for their anticipated energy
consumption and GHG emissions using the scenario-analysis tool called CarbonFIT, a model
built on the platform of CommunityViz. Although the region has demonstrated an interest in
GHG and climate change, MARC has not yet set specific goals in this regard.
ICF International 37 U.S. Environmental Protection Agency
10-000
-------
Estimating Emission Reductions from Travel Efficiency Strategies: Three Sketch Modeling Case Studies
Scenario Development
The MARC base year is 2010 with a future year of 2040. MARC was able to provide only a
limited amount of data, so TRIMMS defaults were used heavily. MARC produced three land use
forecasts during the development of the long range plan. These are:
Current Trends: continuing recent historical patterns of sprawling land uses
Adaptive Scenario: focusing growth in the core areas
Adopted Scenario: representing the compromise between Current Trends and Adaptive that
was adopted by the MPO board
MARC provided BAU outputs from its travel demand model for year 2040. For land use, the
BAU outputs use the Current Trends forecast rather than the Adopted Scenario. This approach
is in contrast to the land use scenario Boston included in the BAU, and was chosen so that the
all smart growth improvements could be analyzed. For transportation, the BAU outputs
represent a build scenario, including all the projects identified in the most recent long range
transportation plan, with a horizon year of 2040. The modeled mode shares are almost
unchanged from 2010 to 2040, because the transportation investments in the plan supply
additional roadway and transit facilities in proportion to the population growth, and consistent
with existing travel patterns.
Scenario 1 Expand Ridesharing and TDM programs
About 50,000 employees in the region have access to telework and flexwork programs,
Guaranteed Ride Home and ridematching services. (A regional ridesharing program includes
14,000 participants). The strategy expands the number of people covered by the program from
50,000 to 300,000, an increase of 500%. It also adds alternative mode subsidies for the
300,000 people covered from the current $25 to $50 dollars per month. Population figures were
scaled based on the natural rate of job growth in the region.
This was modeled as two strategies in TRIMMS: 1) telework and flex work plus TDM programs
(non-monetary only) and 2) subsidize work trips using alternate modes plus other programs
(both monetary and non-monetary). For non-monetary strategies, the radio buttons for TDM
programs and subsidies for Guaranteed Ride Home and Ride Match and Telework and Flexible
Work Schedules were used. For monetary subsidies, MassDOT's BAU trip costs were
substituted as reasonable proxies for trip costs in the MARC region, which were not available
from MARC. In reality trip costs are likely to be higher in the MassDOT region, where average
trip lengths are longer and tolling is more prevalent. However, using higher baseline trip costs
errs on the conservative side when modeling VMT reductions, since the subsidy amounts
examined will be smaller in relative terms.
Scenario 2 - Expand Ridesharing and TDM programs along with Transit Improvement and
Promotion
This scenario adds transit improvement and promotion by reducing transit trip times by 20%,
reducing walking distance to transit by 50% and expanding the successful university transit pass
program (U-Pass) to 6% of total regional population. SmartMoves, the regional transit vision
ICF International 38 U.S. Environmental Protection Agency
10-000
-------
Case Study Results
updated in 2008, would support this strategy. Ridesharing and TDM programs were not included
in the model run. Rather the impact of expanded ridesharing and TDM programs was summed
with these TRIMMS model runs as a post-processing step. Because base year values were not
available from MARC, substitute base year values from the 2010 national study for a similar
metropolitan area (San Diego) were used.
As in Scenario 1 this was modeled as two strategies in TRIMMS: 1) reduce transit trip times and
walking distance only and 2) university transit pass program with reduced transit trips times and
walking distance. TRIMMS does not allow specific adjustment of walking distance to transit. It
uses percentage reductions from the provided default walking distance. The default value of
1.26 miles was used and decreased by 50% for the scenario. For the second scenario, a
baseline transit trip cost of $1.50 was used, based on fare information on the KCATA website.17
In the scenario, the transit trip cost is $0, reflecting the marginal cost of each transit trip to the
pass holder. This is not to say that the transit pass would be free, but in the scenario the transit
pass costs are assumed to be bundled with tuition or employment benefits.
Scenario 3 Expand Ridesharing and TDM programs, Transit Improvement and Promotion with
Smart Growth Land Use
Scenario 3 adds smart growth land use to the previous strategies. The smart growth land use
strategy represents a shift from the region's Current Trends scenario (incorporated in the BAU)
to their Adaptive scenario. This increases weighted average residential density18 and mixed use
land uses for the entire region. Thus, the entire regional population is affected.
MARC provided population densities by TAZ, from which population-weighted average density
by TAZ was calculated. The TRIMMS default for retail establishment per square mile for the
Kansas City region is 0.82. MARC provided retail employment/acre by TAZ. Weighted average
density figures were calculated, and the percent change was applied to the TRIMMS default.
For residential population density, MARC provided the figures 2,655 people per square mile
(BAU) and 3,693 people per square mile (scenario). After rounding these figures to units
allowed by TRIMMS, the scenario population density was modeled as a 40% increase above
the BAU population density. For retail establishment density, MARC provided the figures 0.45
establishments per square mile (BAU) and 0.71 establishments per square mile (scenario).
Similarly, after rounding, the BAU retail density was increased by 55% for the scenario. The
VMT impacts for land use are considerably lower than expected.
17 http://www.kcata.org/fares/
18 Weighted average population density summarizes regional land use densities in a way that is more representative of residents'
daily experience of land use patterns than average density. TAZs with higher populations, which tend to be denser areas, are
weighted more heavily than TAZs with lower populations. In the weighted average density calculation, the density of each TAZ
in the region is weighted by its proportion of total regional population. An increase in weighted average density does not
indicate an increase in total regional population, but rather a shifting of population toward higher density centers.
ICF International 39 U.S. Environmental Protection Agency
10-000
-------
Estimating Emission Reductions from Travel Efficiency Strategies: Three Sketch Modeling Case Studies
Scenario 4 - Expand Ridesharing and TDM programs, Transit Improvement and Promotion, Smart
Growth Land Use and Transportation Pricing
The final scenario adds transportation pricing as an increase in the average cost of auto trips by
100% and increase parking costs by 500% in the Downtown Area. Currently about 1% of
regional parking is priced. This was expanded to 6% of total regional parking.
BAU trip costs represents the average fuel cost/mile for a sedan from AAA19 and the average
trip length for Kansas City. Relative to fuel costs only, the scenario represents approximately a
25 cent per mile charge. The rideshare trip cost was adjusted for vehicle occupancy. BAU
parking cost is based on similar metropolitan areas from Colliers Parking Rate Survey.20 The
rideshare parking cost was assumed to be the same as auto-drive alone parking cost.
This was again modeled as two strategies: 1) parking pricing for 6% of the regional population
and mileage pricing and 2) mileage pricing for the remaining 94% of the regional population.
Initially, the entire scenario was applied regionwide by calculating average regionwide parking
price as an intermediate step. However, this produced an unreasonable outcome because it
results in an enormous relative increase in average parking cost. Instead, the parking strategy
was modeled only for the sub-area to which it applies in order to reasonably limit the impact.
Both strategies modeled include land use changes and transit improvement. TDM and university
transit pass (part of transit improvement and promotion) were added in as a post-processing
step
Input parameters are provided in Table 14 for current conditions, a business as usual future,
and the four scenarios selected by MARC. Specific input values are provided for the scenarios.
19 http://fuelgaugereport.aaa.com/?redirectto=http://fuelgaugereport. opisnet.com/index.asp
20 http://www.colliers.eom/~/media/files/marketresearch/unitedstates/colliers 2012 na parking survey.pdf
ICF International 40 U.S. Environmental Protection Agency
10-000
-------
Case Study Results
Table 14. MARC Scenario Details
Scenario Description
Current
Conditions
Business as
Usual
Scenario 1:
Expanded TDM
Geography Data Inputs
Existing conditions across
all strategies in 2010
Regionwide
2040 conditions with
current levels of employer
program, land use, HOV
lanes, and transit
Regionwide
Expand access to telework
and flexwork programs,
Guaranteed Ride Home
and ridematching services.
Regionwide
Scenario 2: Improve transit and expand Regionwide
Scenario 1 + transit pass program.
Enhanced Transit
Scenario 3:
Scenario 2 +
Land Use
Increase residential density
and mixed use land uses
for entire regional
population.
Scenario 4:
Scenario 3 +
Pricing
Implement mileage pricing
and increase and expand
coverage of parking costs.
mode shares
average vehicle occupancy
average vehicle trip lengths
regional population and employment
TRIMMS default vehicle ownership (1.71
vehicles/household)
employer-based incentives for alternative commute
modes
mode shares
average vehicle occupancy
TRIMMS default vehicle ownership (1.69
vehicles/household)
2040 regional population and employment
current employer-based incentives for alternative
commute modes
current travel times (50 minutes for transit)
walking distance to nearest transit station (1.26 miles)
average parking costs ($10 for drive-alone and $4.26
for rideshare)
average trip costs ($2.41 for drive alone, $2.08 for
rideshare, $4 for vanpool, and $1.50 for transit)
residential population density (2,655 persons per
square mile)
retail establishment density (0.45 per square mile)
share of regional employees covered (65,181 for non-
monetary subsidies only and 300,000 for all subsidies)
average monthly subsidy offered to employees ($50
for rideshare, vanpool, and transit)
are guaranteed ride home, ride match, telework, and
flexible work schedules offered?
20% travel time reduction (to 40 minutes)
50% decrease in walking distance to nearest transit
station (to 0.63 miles)
Number of regional employees covered by university
transit pass program (134,834)
average transit trip cost for university transit pass ($0)
Regionwide 39% increase in population density (to 3,693 persons
per square mile)
57% increase in retail establishment density (to 0.71
per square mile)
Regionwide 6% of all parking (public and private) is priced
500% increase in average parking cost per trip (to $60
for drive-alone and to $25.53 for rideshare)
100% increase in average cost per trip (to $4.82 for
drive alone and to $4.16 for rideshare)
ICF International
10-000
41
U.S. Environmental Protection Agency
-------
Estimating Emission Reductions from Travel Efficiency Strategies: Three Sketch Modeling Case Studies
Emissions Analysis
MARC provided a variety MOVES input data at county-level resolution for the seven counties in
the MPO. However, this data only covered a small portion of the required inputs for MOVES, so
a large amount of default data was included in the analysis. MOVES was run at the county scale
for each county and year individually. The final calculated emission factors were then calculated
for the domain as a whole by aggregating the MOVES results from each county. MOVES
outputs for amounts of emissions and activity were summed across the seven counties, creating
domain totals, and the ratio of emissions to activity created domain-wide emission factors. This
approach required greater effort in data collection and manipulation, file management, model
setup and runtime, and post-processing, but incorporated all the provided data to produce the
regional average used in this approach. Table 15 summarizes the data MARC provided and
used in the MOVES simulations, and any modifications made to the data before modeling.
Table 15. MARC Data Sources
Region Provided
Data
Base Future
Year Year
Data Used in Final Modeling
Base Year
Future Year
Source Type Population
Vehicle Type VMT
Road Type Distribution
Meteorological Data
Age Distribution
VMT Hour
Fractions Day
Month
Average Speed
Distribution
Ramp Fraction
Fuel Supply
Formulation
I/M Program
AVFT
X
X
X
X
X
(Kansas
only)
X
X
X
MOVES defaults*
MOVES defaults*
provided data
MOVES defaults
provided data
provided data
MOVES defaults
MOVES defaults
provided data
MOVES defaults
MOVES defaults
MOVES defaults
no data: these counties do not
require emissions inspections
provided data for Kansas counties.
2008 values repeated for 2009 and
2010; MOVES defaults for
Missouri counties
MOVES defaults*
MOVES defaults*
provided data
MOVES defaults
provided baseline year data with
change: YEARID was changed to
2040
provided data
MOVES defaults
MOVES defaults
provided data
MOVES defaults
MOVES defaults
MOVES defaults
no data: these counties do not
require emissions inspections
MOVES defaults
*MOVES defaults for these inputscan be generated from a national scale run
Lessons Learned
The MARC case study is a good example of how emission factors can be estimated for a multi-
county domain when at least some county-level data are available. MARC did not provide
several of the most critical MOVES inputs, including vehicle populations and total VMT by
vehicle type. MARC also did not provide data for meteorology, ramp fraction, monthly and daily
VMT allocations, or fuel supply and formulation. Each of these were modeled with MOVES-
ICF International
10-000
42
U.S. Environmental Protection Agency
-------
Case Study Results
derived defaults. I/M is not required for these counties and was not included in the analysis.
MARC did provide road type VMT and speed distribution data for all counties and years. For the
remaining inputsage distribution and alternative fuels and technologiesMARC provided data
for some years and/or some counties, and otherwise defaults were used or both years shared
the same values, depending on the input.
In some cases, the year of provided data did not agree with the baseline year and had to be
modified. In these cases, any local inputs provided were translated to the baseline year and any
default values were re-extracted for the baseline year. Notably, the AVFT values were provided
for the counties in Kansas but only extending to 2008. To update to the baseline 2010 year, the
last year of provided data was repeated for 2009 and 2010. This approach was selected given
the preference to rely on local data whenever possible, the reasonable doubt that there would
be a significant change across the most recent three years, and the consistency maintained by
using local data for all the years.
Finally, no local traffic, vehicle registration, or other DOT information was available to derive
source populations and VMT totals. Instead, national scale MOVES simulations were performed
for each county of interest to extract defaults for these values, which were then re-imported to
the county-scale runs. This is less ideal than using local data. The data for age distribution and
alternative fuels and technologies could also be more complete, instead of a mix of local and
default data. However, the case does serve as a demonstration of this method with very limited
local input values.
Summary of Results
The post-processing of hourly MOVES outputs to domain-average emission factors for the
MARC case used the same methods as discussed for PAG, except that the emissions and
activity were aggregated across all seven counties before domain-wide, activity-weighted,
emission factors were calculated. The resulting emission factors are shown in Table 16.
Table 16. Emission factors for MARC
g/start
Base Year Future Year Base Year Future Year
(2010) (2040) (2010) (2040)
Auto (Motorcycles+Passenger Cars+Passenger Trucks)
GHGs (C02-equivalent)
NOx
PM2.5
VOCs
398.53
0.90
0.03
0.18
291.07
0.14
0.01
0.03
153.99
1.76
0.04
2.51
97.94
0.40
0.02
0.68
Vanpool (Passenger Trucks+Light Duty Trucks)
GHGs (C02-equivalent)
NOx
PM2.5
VOCs
509.56
1.50
0.05
0.28
361.64
0.30
0.02
0.05
181.22
2.34
0.05
3.06
110.55
0.49
0.02
0.59
ICF International
10-000
43
U.S. Environmental Protection Agency
-------
Estimating Emission Reductions from Travel Efficiency Strategies: Three Sketch Modeling Case Studies
MARC Scenario Results
Table 17 provides the regional results of the analysis. As noted earlier, regional reductions are
often modest when compared to the impact on the population targeted by the strategies.
Additional explanation is provided below and numerical results are included in Appendix B.
Table 17. MARC Regionwide Percent VMT and Emissions Changes
Percent Change for 2040 BAD compared to 2040 Scenario
Scenario
PM2.5 NOx VOC
equivalent)
Scenario 1 :
Scenario 2:
Scenario 3:
Scenario 4:
Expanded TDM
Scenario 1 + Enhanced Transit
Scenario 2 + Land Use
Scenario 3 + Pricing
MARC
-0.93%
-2.35%
-2.49%
-12.06%
-0.93%
-2.35%
-2.49%
-12.05%
-0.93%
-2.35%
-2.49%
-12.05%
-0.92%
-2.35%
-2.48%
-12.03%
-0.92%
-2.34%
-2.48%
-12.02%
Ridesharing and TDM have a moderate impact on affected population with a 3.6% VMT
reduction and shifts from drive-alone to rideshare, cycling, walking and transit.
Transit improvements have a small impact on affected population (0.5% VMT reduction) for
transit trip times and walking distance. However, there is a large impact on affected population
for university transit pass (18.3% VMT reduction) with shifts from drive-alone and rideshare to
transit.
Land use also shows a small impact on affected population (0.14% VMT reduction) with shifts
from drive-alone to transit. This is an unreasonably small impact, similar to that observed for
PAG. CUTR has suggested that this result may be due to the relatively low level of transit
service in the MARC region, as land use results in TRIMMS are driven by transit accessibility.
See Section 4.2 for a further discussion of land use analyses with TRIMMS.
As expected, transportation pricing has a large impact on the affected population (19.5% VMT
reduction) for parking and mileage pricing. There is also a large impact on the affected
population for mileage pricing only (9% VMT reduction) with shifts from drive-alone and
rideshare to transit, vanpool, cycling and walking.
ICF International
10-000
U.S. Environmental Protection Agency
-------
Conclusions and Recommendations
4. Conclusions and Recommendations
Although each region participated independently in the selection of scenarios and with respect
to the available data, there were some common themes and overall comparisons that may be
useful to other regions interested in applying TEAM. The options selected in TRIMMS and
MOVES for the analysis of scenarios were similar across regions, except for adjustments made
based on data availability. This standard analytical approach provides a basis of comparison
that is independent of individual strategy performance in each regional context. The lead
agencies along with their stakeholders can apply the appropriate regional context for evaluation
of the scenario outcomes. Their context and experience drawn from participating in the case
studies will inform their view of how best to use the information gathered from this study.
4.1. TEAM Data Requirements
Regional data collection and validation was the most challenging aspect of the analysis. Many
factors contributed to the extensive time and interaction that this task involved. MPOs have
standard data elements that are used for various routine planning functions, and TEAM was
developed to interface with this data without requiring additional data collection or extensive re-
evaluation of the information. Every MPO is unique in this regard. The availability of data
depends on regional priorities and what data can be collected or borrowed. A strong
understanding of both TRIMMS and MOVES allows adjustments to account for these regional
data differences; however, some regions may find this challenging.
Data validation is an essential step in the TEAM approach. It is not enough to ensure that the
right type of data is used, but also necessary to consider the reasonableness of this data. For
instance, the distribution of VMT for transit vehicles among road types is unlikely to be the same
as for passenger cars. This study found that the reasonableness of data that may already be in
use cannot be taken for granted. A critical element for applying TEAM successfully is the ability
to identify questionable data and develop substitutions when needed. In some cases this meant
several revised data sets. In other cases it resulted in the extensive use of defaults. Knowledge
about the underlying data and previous experience in their use is an advantage.
Use of local data is the best way to ensure that the strategy effectiveness identified through
TEAM is applicable to the region. Default data is available in both TRIMMS and MOVES to use
when required, but care must be taken to ensure the default data is applicable to the region and
to the strategies being evaluated. Regions that undertake a TEAM analysis should allow
significant amount of preparation time to identify data requirements, collect or substitute data
elements and validate the appropriateness of the data for this type of analysis.
4.2. TRIMMS Support of TEAM
TRIMMS 2.0, which was used for modeling the 2010 national study, was in the process of being
updated at the time. TRIMMS 3.0 is the current version and was used for these case studies., A
web-based version is now under development. To the extent possible, the new features and
functionality in TRIMMS 3.0 were used in an effort to meet the strategy interests of the region.
By putting TRIMMS to the test in this way, some potential shortcomings were identified. Many
ICF International 45 U.S. Environmental Protection Agency
10-000
-------
Estimating Emission Reductions from Travel Efficiency Strategies: Three Sketch Modeling Case Studies
of these issues have been discussed with the TRIMMS developers and in some cases, ways to
work within the tool and through post-processing were identified. In a very small number of
instances, TRIMMS just might not be the right tool for a selected strategy at this time. The
information below is provides some understanding of the more significant issues.
Users of TEAM should be prepared to translate their travel efficiency strategies into TRIMMS-
ready inputs. Many of TRIMMS' functions work by manipulation of travel times and travel costs,
common factors in travel demand modeling. To model an expansion of HOV lanes in TRIMMS,
for example, the user must input changes in typical travel times for carpool trips (and possibly
SOV trips). Precise changes would be difficult, if not impossible, to estimate without a detailed
transportation network analysis. TRIMMS users can instead rely on assumptions derived from a
literature review or expert opinions. For example, in the case studies conducted here, a travel
demand modeler at the Boston MPO suggested that a 10% reduction in average carpool times
could be reasonably achieved by an expansion of the HOV network in the region. Providing this
kind of assumption makes the analysis a goal-based one. Determining what investments must
be made to achieve that travel time reduction is not necessary to conduct an exploratory
analysis of alternative scenarios.
Land Use
Land use is one of the most critical elements in the evaluation of travel efficiency strategies
because it is so central to transportation planning and represents an almost universally
identified approach to reducing travel. By shortening distances among travel destinations,
increasing mixed use zoning, and concentrating growth around transit nodes, smart growth
strategies make walking, cycling, and transit more viable modes of transportation and can even
reduce the distances of some car trips. TRIMMS land use features are in an early stage of
development, and the predicted impact of land use in these case studies is less than in other
current literature. TRIMMS 3.0 includes a new function for estimating the impact of land use
strategies, including increasing residential densities, land use mixing, transit station
accessibility, and transit-oriented development (TOD). Although this capability was initially
viewed as a positive attribute, the TRIMMS land use function raised concerns for two main
reasons. First, it is based on a limited data sample. Second, it only considers the propensity of
people living in denser areas to increase their use of transit. It does not consider the well-
documented effects of increased biking and walking, or shortened trips in private vehicles. As a
result, the land use analyses conducted for this study produced smaller VMT reductions in all
regions than for the clusters in the 2010 national study used for comparison, where land use
impacts were assessed outside of TRIMMS. Land use strategies produced slightly lower
impacts than expected in the MassDOT region, and dramatically lower impacts than expected in
the PAG and MARC regions. While the underestimation of land use impacts is believed to be a
result of the model's response to the availability of transit, (transit availability is much greater in
the MassDOT region than in the PAG and MARC regions) more testing is required to determine
the cause.
To use the land use functions in TRIMMS 3.0 for a regional analysis, using weighted average
densities was determined to be most appropriate. Weighted average population density
ICF International 46 U.S. Environmental Protection Agency
10-000
-------
Conclusions and Recommendations
summarizes regional land use densities in a way that is more representative of residents' daily
experience of land use patterns than average density. TAZs with higher populations and which
tend to be denser areas, are weighted more heavily than TAZs with lower populations. This
approach was used as a pre-processing step in scenario evaluation
There are other approaches to analyzing land use strategies that we recommend for future
efforts. The approach used in the 2010 national study with TRIMMS 2.0 (which did not have an
explicit land use component) is still valid. In the 2010 study, the TRIMMS model was used to
calculate the change in VMT using elasticity values for travel time. These values for expected
changes in travel time (access time and in-vehicle time) and trip lengths resulting from land use
measures were based on a review of the literature and values included in EPA's Smart Growth
Index (SGI) model. It is also possible to use the widely accepted density variable elasticities put
forward by Ewing and Cervero.21 The web-based version of TRIMMS is expected to improve
and expand the land use capabilities.
Mode Shifts and Trip Lengths
The 2010 national study identified the importance of trip lengths in determining strategy
effectiveness. TEAM is focused on the shift from automobile travel to other modes and to
shorter trip lengths. TRIMMS 3.0 does not appear to ensure that trips of equal length are
substituted for one another when travel mode shifts. For example, TRIMMS does not account
for the fact that a 40 mile vanpool trip replaces a 40 mile car trip. It always applies the average
trip length for each mode, which is typically shorter for a car trip than for a vanpool trip. Thus if
car trips are replaced with a vanpool trip, the VMT benefit may be affected by the assumption of
longer trips.
CUTR has also confirmed that there are no internal controls in TRIMMS to ensure that the
number of trips remains constant (or near constant) while trips are shifted between various
modes, with the exception of the algorithms used to analyze employer-based TDM strategies.
TRIMMS sometimes shows an increase in rideshare VMT without a commensurate decrease in
drive-alone VMT when trip cost values are entered. For modeling HOV lanes (reduction in
rideshare travel times), the reduction in SOV VMT is less than the increase in rideshare VMT.
This result can be corrected within the TRIMMS model by adjusting elasticities or outside of the
model by adjusting the changes in SOV and carpool trips to be more comparable.
In general, the total VMT results are more reliable than VMT by mode. This study presents
results for light-duty vehicles only, and therefore does not include bus VMT. TRIMMS tends to
over-estimate increases in transit VMT, because it does not allow for increases in transit vehicle
occupancy.
There is some interest in strategy types that cannot be evaluated by TRIMMS at the regional
scale in its current form. Bike strategies are one consistent example noted. Other interests
21 Ewing, Reid and Cervero, Robert(2010) Travel and the Built Environment', Journal of the American Planning Association, First
published on: 11 May 2010 (first)
ICF International 47 U.S. Environmental Protection Agency
10-000
-------
Estimating Emission Reductions from Travel Efficiency Strategies: Three Sketch Modeling Case Studies
expressed were park-and-ride, transit fare integration, peak/off-peak fares, transit marketing,
public education, and operational improvements such as intelligent transportation system (ITS).
TRIMMS remains the most viable sketch planning tool to support TEAM, and these
observations are not intended to detract from its usefulness. The developers at CUTR suggest
that improvements to TRIMMS can be supported by users reporting their results and experience
using the model.
4.3. MOVES Support of TEAM
MOVES is the best model for estimating emissions for TEAM. It is EPA's current mobile source
emission inventory tool, and is required to be used for state and local regulatory analyses, such
as for State Implementation Plans (SIP) and transportation conformity determinations.22 These
uses of the model are addressed in other EPA guidance and, although addressing model use
for these purposes is beyond the scope of the present analysis, this project has shown that
regions are becoming increasingly familiar with its use. Accordingly, many are developing their
own, custom MOVES inputs.
Although the inputs developed for those purposes may be used for TEAM, the use of the
MOVES model for TEAM differs.23 Primarily, in a sketch analysis, detailed emission factors are
not needed and the additional complexity in producing and using them is not warranted. Instead,
as noted previously, overall regional, average emission factors representing the activity-
weighted mean of all starting and running activities in the region is produced and coupled with
the TRIMMS outputs. These emission factors are calculated as total running emissions from
hourly outputs (in grams per year) divided by total running activity (in miles per year) across the
modeled region; a similar analysis is made for start emissions. Resulting emission factors are in
units of grams of pollutant per average mile driven or grams of pollutant per average start. This
is explained further in the TEAM User's Guide24 and in Section 2.4, above.
A range of technical capability related to the use of MOVES was evident among the regions.
There were also some policy questions about the analytical procedures for this study and the
regulatory analysis conducted for CAA purposes. Although the same data may often be used, a
TEAM analysis cannot be used for emission inventories, air quality demonstrations and
transportation conformity determinations required by the CAA. Key topics on the use of MOVES
for the case studies are discussed below.
Selection and Use of Base and Future Years
Base and future years in MOVES should be selected to agree with those of the strategies being
analyzed. A common issue encountered was that the baseline year selected for the scenario
22 E.g..Using MOVES to Prepare Emission Inventories in State Implementation Plans and Transportation Conformity: Technical
guidance for MOVES2010, 2010a and 2010b, EPA-420-B-12-028, April 2012
23 More explanation in the use of this approach is given in: Analyzing Emission Reductions from Travel Efficiency Strategies: A
Guide to the TEAM Approach, EPA-420-R-11-025, September 2011.
24 http://www.epa.gov/otaq/stateresources/policy/420r11025.pdf
ICF International 48 U.S. Environmental Protection Agency
10-000
-------
Conclusions and Recommendations
analysis was different than the year for which data had previously been prepared in the region,
e.g., for a regional emissions analysis. In these cases, the emission inputs for the baseline year
were required to be modified (or default values extracted) to agree with the scenario baseline
year. The effort needed to adjust for differences in the baseline year should be reduced as more
regional data collection and preparation takes place with TEAM analysis in mind.
Data Collection and Validation
Of the three regions, MARC provided the least developed MOVES data. The data MARC
provided had been prepared by a consultant for a previous analysis and MARC was not able to
collect data beyond this. In some cases, MARC had data, or access to data, but it had not been
formatted for or used in any MOVES applications. MARC's data collection also was affected by
its regional span across state boundaries, which complicated the efforts.
The Boston region showed the most expertise with MOVES, but a small number of differences
between their uses of the model, such as for transportation conformity analyses, and the
methods of the TEAM approach had to be considered in this analysis. This is principally related
to MassDOT's use of "Emission Rates" calculation type instead of "Inventory" for a single,
representative county, and how to translate that parameterization to the TEAM approach
documented in Section 3.2. No unique data collection was needed for this TEAM analysis.
Accordingly, CTPS indicated an interest in performing a comparison of the emission factors
generated by the TEAM approach to those generated through their emission rate-based
approach, since few regions have chosen this approach. Their preliminary comparison showed
that the two sets of factors agreed well except for baseline year (2009) NOx running emissions,
for which the TEAM values were greater than the curves of values generated by their previous
analyses. Subsequent review showed no obvious errors in either analysis and this was noted as
an interesting result that should be documented for potential future exploration. It is possible
that this was related to removing I/M in the TEAM approach, however this has not been
demonstrated conclusively.
The PAG region showed great interest in increasing staff understanding and mastering of
MOVES. PAG collected extensive data for this effort and consistently updated their existing data
and methods per discussions with EPA. This led to several revisions of input values, but
produced one of the most locally specific analyses. In addition, PAG also independently
produced their own emission factors. Comparisons of those with the factors generated in this
study showed essentially identical results.
It is not clear that any data element was generally less available than others. Instead, the case
studies illustrated that the experience level of each MPO and the availability of resources to
collect and process data if it was not already in house are factors in the amount of time a TEAM
analysis will take. Both of these factors should improve with increasing familiarity with the
models and better coordination among various analytical planning efforts by the MPOs.
ICF International 49 U.S. Environmental Protection Agency
10-000
-------
Estimating Emission Reductions from Travel Efficiency Strategies: Three Sketch Modeling Case Studies
4.4. Regional Realities and Implications for Using TEAM
Making Assumptions
Using TRIMMS requires agencies to make assumptions about future conditions in terms of
inputs that TRIMMS uses, such as trip costs and travel times. These assumptions, especially
about a distant future, are not easily made and may represent sensitive issues in the region.
Agencies are faced with the decision of how aggressively strategies should be framed simply in
the choice of inputs. Stakeholders are not always comfortable coming up with appropriate
assumptions or dealing with the types of inputs that TRIMMS uses. For example, transit
improvements need assumptions about impacts on access, headways and travel times. Dollar
values must be applied to subsidies and pricing. Participating agencies dealt with this
requirement differently. Some used it as a way to identify the value of much stronger policies.
Others evaluated strategies that would be reasonably feasible in the region. This represents a
broad range of approaches and will impact the results of the analysis.
The TEAM approach works best when the considered hypothetical scenario analysis is based
on reasonable goals. A stakeholder group composed of well-informed local transportation
planners, modelers and land use planners can draft a set of goals for a region based on their
professional knowledge and limited additional research. While goals need to be reasonably
achievable in the long term, they do not need to be constrained by shorter term political, fiscal,
or engineering challenges.
Comparisons and Validations
As discussed in Section 2.3, results for each region were validated against a comparison cluster
from the 2010 national study. Clusters were defined by their population size and transit mode
shares. These characteristics were used to select a comparison cluster for each of the three
case study regions. Results for each strategy were expected to be similar to those from the
comparison cluster.
This comparison was used to identify differences in model results that required further
examination. Where the case study results were similar to those for the comparison cluster, the
case study results were considered to be validated. In some cases it became clear that a
selecting a single comparison cluster was not enough to validate results. For example, results
for pricing strategies and employer TDM strategies varied broadly from the comparison cluster
selected for the MARC region. This difference highlighted that a single cluster may be too
limited a comparison. In these cases it was more valuable to compare case study results to the
full range of results across all urban clusters. Population size and transit mode share are useful
factors to characterize urban regions, but they do not capture other nuances that can affect the
results of strategies, such as average trip lengths, travel costs, and geography.
It is also important to consider the affect that the strategy assumptions have on results.
Strategies can be specified in more or less aggressive terms, and can be restricted to sub-
populations and sub-geographies. These types of variations must be considered when
comparing results to those for the 2010 national study.
ICF International 50 U.S. Environmental Protection Agency
10-000
-------
Appendix A: Strategies Identified in the 2010 Potential Changes in Emissions Due to Improvements in Travel Efficiency
Report
5. Appendix A: Strategies Identified in the 2010 Potential Changes
in Emissions Due to Improvements in Travel Efficiency Report
The table below provides the strategy categories for reducing vehicle travel demand selected for
analysis:
Table 4. TCM Strategies Analyzed
Strategy Categories
Travel Demand Management (TDM)
Land Use / Smart Growth
Transit
Pricing
TCMs Included in the Analysis
Rideshare Programs
Employer-based Programs
Public Outreach/ Education
TOD: Improved Transit Access
Mixed Land Use
Promotion of Higher Density
Increased Transit Frequency
Lower Fares or Transit Subsidies
Parking Pricing
Mileage Fees
The table below provides detailed descriptions of strategies and the assumptions used in the
analysis of scenarios.
Table 6. Scenario Assumptions and Modeling Approach for TCM Strategies
TCM Strategy
Employer-
based TDM
strategies
Specific strategy
Flexible work hours
Incentives for carpooling
Guaranteed ride home
programs
Ride sharing/ ride matching
TDM outreach/public
outreach programs
Subsidies/discounts for
transit, pedestrian and bike
modes
Telecommuting
Strategy information
Whether or not
employer offers
(TRIMMSasksfora
yes/no answer) to take
these programs into
consideration
2010 - 2030
30% of employers
Region-wide offer these
programs; includes all
TDM strategies except
walk and bike subsidies
2030 - 2050
50% of employers
Region-wide offer these
programs; includes all
TDM strategies
ICF International
10-000
A-1
U.S. Environmental Protection Agency
-------
Estimating Emission Reductions from Travel Efficiency Strategies: Three Sketch Modeling Case Studies
Table 6. Scenario Assumptions and Modeling Approach for TCM Strategies
TCM Strategy
Land use
policies
Transit
projects and
policies
Pricing
policies
Specific strategy
TOD, smart growth, increase
in density, mixed use
developments
Transit service
expansion/increase in
frequency, improved access
Fare discounts, reduction,
subsidies, or free transfers
Parking charges
VMT fees or congestion
pricing
Strategy information
Change in travel times
for all modes, change
in average trip lengths
Improvement in transit
travel time and access
time
Change in transit fares
Increase in auto
parking costs
Increase in peak hour
driving costs
2010 - 2030
3% reduction in all
access times, 5%
reduction in transit travel
time and walk/bike times;
5% increase in auto travel
time due to density/
congestion effects
Note: Access time taker
5% reduction in transit
travel time
10% reduction in transit
fares
$2 increase per day
$0.10 increase per mile
2030 - 2050
6% reduction in all
access times, 10%
reduction in transit travel
time and walk/bike times;
10% increase in auto
travel time due to density/
congestion effects
as proxy for trip length.
10% reduction in transit
travel time
20% reduction in transit
fares
$5 increase per day
$0.25 increase per mile
The table below represents the range of impacts of each TCM on travel activity from the
literature. Where elasticity values were available and could be compared, the travel time and
travel cost elasticities for each mode used in the 2010 national study fall within the reported
ranges shown in Table A-3.
The ranges provided show estimates of the change in automobile travel or transit ridership for a
given change in user travel time or travel cost.
Where specific elasticities are not available, Table A-3 lists impacts in terms of percentage
reductions in travel demand (trips or VMT).
The elasticities are travel demand elasticities, defined as the percentage change in travel (VMT
or trips) caused by a one-percent change in user travel costs or travel time. In this study, travel
costs are equivalent to out-of-pocket operating costs for auto drivers and passengers, and
transit fares per trip for transit riders.
Table A-3. Quantitative Estimates of Travel Activity Impacts of TCMs from Literature
Examples of Measures
Elasticity/ VMT Reduction %
Ridesharing Programs and Investments
Park-and-ride facilities
High-Occupancy Vehicle (HOV) lanes
Rideshare matching programs
Regional implementation: 0.1 to 0.5% reduction in
Long run (LR) travel time elasticity, regional: -1.0,
0.2 to 1.4% VMT reduction
VMT
urban: -0.6, rural: -1.3
0.1 to 2.0% VMT reduction
ICF International
10-000
A-2
U.S. Environmental Protection Agency
-------
Appendix A: Strategies Identified in the 2010 Potential Changes in Emissions Due to Improvements in Travel Efficiency
Report
Table A-3. Quantitative Estimates of Travel Activity Impacts of TCMs from Literature
Examples of Measures
CarpoolA/anpool incentives
Car-sharing
Elasticity/ VMT Reduction %
0.2 to 3.3% VMT reduction
Limited quantitative data
Bicycle and Pedestrian Facilities and Programs
Bike paths / lanes / routes
Bike/ped facilities to support transit
<0.1% VMT reduction
Limited quantitative data
Transit Projects and Policies
Transit service expansion /increase in
frequency
Improved transit travel times and operations
(busways, BRT, signal prioritization for transit
vehicles, heavy and light rail, managed lanes)
Improved transit access through shuttle and
feeder bus services, paratransit
Transit service integration and intermodal
transfer centers
Fare integration for easy transfers
Improved transit marketing, information,
amenities
Commuter discounts/fare reductions
Peak/off-peak transit fares
Transit improvement policies, overall
-0.6 to -1.0; for buses
-0.5 (time between buses) for service frequency alone
-0.4 (travel time elasticity with respect to ridership)
Relates to improving travel time above, not measured separately
Relates to improving travel time above
Relates to improving travel time above
Limited quantitative data
-0.3 to -0.4 (fare elasticity with respect to ridership)
-0.1 to -0.3 (peak fares) and -0.1 to -0.7 (off-peak fares, depending on trip
purpose; lower for work trips)
Studies estimate 0 to 2.6% VMT reduction
Parking Management and Incentives
Parking cash-out
Preferential parking for carpools and vanpools
Parking duration restrictions
Elasticities are not available; although some quantitative data on percentage
reduction in regional VMT are available from specific projects and studies.
Employer-based Programs (effects depend on level of adoption)
Flexible work schedules
Telecommuting
Compressed work weeks
Employer-provided transit passes
Guaranteed ride home programs
Elasticities are not available; although some quantitative data on percentage
reduction in regional VMT are available from specific projects and studies.
ICF International
10-000
A-3
U.S. Environmental Protection Agency
-------
Estimating Emission Reductions from Travel Efficiency Strategies: Three Sketch Modeling Case Studies
Table A-3. Quantitative Estimates of Travel Activity Impacts of TCMs from Literature
Examples of Measures
Elasticity/ VMT Reduction %
Pricing Policies
Area-wide road pricing/congestion pricing
Distance-based pricing or mileage fees
Peak period pricing/ variably priced lanes
Parking pricing/fees
High Occupancy Toll (HOT) lanes/toll
increases
Pay-as-you-drive Insurance
Fuel taxes
Freight vehicle pricing
-0.1 to -0.4 (urban road pricing); 10-25% reduction in central city VMT with
cordon pricing; 0.2 to 5.7% regional VMT reduction
LR: -0.1 to -0.8 (price elasticity). Conservative LR estimate for the U.S. would
be -0.3
-0.03 to -0.4 (depending on time of day)
Overall LR elasticity: -0.1 to -0.5
LR regional: -0.3; at sites: -0.1 to -0.2
LR (non-commute): -0.2 to -0.4
Studies show 0.5-4% reduction in work-related VMT; 3.1 to 4.2% reduction in
non-work VMT
-0.1 to -0.4; data from specific projects are available
-0.3
LR: -0.1 to -0.3, tending towards the lower end
-0.25 to -0.35 (price elasticity); -0.3 to -0.7 (travel time elasticity)
Integrated Land Use and Transportation Strategies
Transit-oriented development and incentives
(Design and transit access)
Smart growth and mixed use development
(Diversity)
Land use controls for compact, dense urban
development (Density)
Improved regional accessibility due to
combined measures
Land use measures, overall
-0.05 (vehicle trips) and -0.03 to -0.08 (VMT)
-0.03 (vehicle trips) and -0.05 (VMT)
-0.05 (vehicle trips) and -0.05 to -0.12 (VMT)
-0.18 to -0.22 (VMT); studies estimate regional VMT reduction by 2-20% in 20
years with doubling of results in 40 years.
Regional VMT reduction of 0 to 5.2%
Vehicle Restrictions by Geographic Area or in Peak Periods
Freight vehicle controls
No-drive days
Urban non-motorized zones
Elasticities are not available; although some quantitative data on percentage
reduction in regional VMT are available from specific projects and studies.
Public Education and Outreach Programs
TDM outreach programs by employers
Episodic programs (e.g. ozone action days)
Public communication about the impacts of
travel decisions
These measures are typically implemented as part of other measures. Difficult
to estimate impacts separately as it could lead to double-counting.
ICF International
10-000
A-4
U.S. Environmental Protection Agency
-------
Appendix B: Emission Change Quantities and Additional Technical Details for the MOVES Analysis
6. Appendix B: Emission Change Quantities and Additional
Technical Details for the MOVES Analysis
As discussed previously, MOVES was run in inventory mode for each of the regions to produce
activity-weighted, average emission factors each the region. These emission factors were
coupled with the TRIMMS-predicted changes in activity to produce net emissions changes by
scenario for each region. The corresponding relative reductions are presented for each region in
the body of the report. Table 18 shows the calculated total emission and travel changes for each
of the main pollutants.
Table 18. VMT and Emission Changes by Region and Scenario
Resulting VMT and Emissions Changes for Selected Pollutants (kg), relative to Baseline or BAD Level, by Scenario
VMT and Emissions Changes - 2040 BAU to 2040 VMT and Emissions Changes - 2010 Baseline to 2040
Scenario Scenario
oueiidiiu
Scenario 1 :
Light-Duty VMT
Implement .40,970,861
SunTran All
Access Pass
Scenario 2:
Expand
Employer-
Based
Incentives for
Alternative
Commute
Modes
Scenario 3:
-40,744,298
GHGs
(C02
equivalent)
-127,117
-55,843
PM2.5
-5
-2
BRTonTwo -40,575,312 -2,148 0
Corridors
Scenario 4:
Expand
Parking
Pricing in the
Downtown-
University
-40,672,387
-33,130
-1
NOx
-71
-33
-1
-20
VOC
-38
-21
Light-Duty
VMT
21,645,838
21,872,402
GHGs
(C02
equivalent)
PM2.5
NOx
VOC
5,542,039 81 -12,456 -3,821
5,613,313
83
-12,418
-3,804
-1 22,041,387 5,667,007 85 -12,387 -3,784
-13
21,944,312
5,636,026
84
-12,405
-3,796
Scenario
VMT and Emissions Changes - 2035 BAU to 2040 VMT and Emissions Changes - 2009 Baseline to 2035
Scenario Scenario
GHGs
Light-Duty VMT (C02 PM2.5
equivalent)
Scenario 1:
Expand
Programs and -103,148,176
Incentives for
Healthy Modes
NOx VOC
PM2.5 NOx
VOC
equivalent)
-920,323 -48 -598 -283 4,465,026 -8,459,568 -573 -74,147 -29,627
ICF International
10-000
B-1
U.S. Environmental Protection Agency
-------
Estimating Emission Reductions from Travel Efficiency Strategies: Three Sketch Modeling Case Studies
Scenario 2:
Expand
Programs and
Health'^odes -104,237,520 -1,276,754 -66 -830
+ Smart
Growth Land
Use
Scenario 3:
Expand
Programs and
-393 3,375,683
-8,815,999
-592
-74,378
-29,737
Incentives for
Healthy Modes
+ Smart
Growtl
Use +
Lanes
Scena
Expand
Prograr
Incentiv
Healthy
+ Smart
Growth
Use + Hi
Lanes +
Transit
Net\
Exp
and
-104,415,406 -1,334,958 -70 -867 -411 3,197,796 -8,874,203 -595 -74,416 -29,755
th Land
HOV
>
jrio 4:
id
ams and
tives for
ny Modes
art
th Land
HOV
> +
it
)rk
ision
upmpnt
-104,757,958
-1,447,040
-75
-940
-445
2,855,245
-8,986,285
-600
-74,489
-29,789
VMT and Emissions Changes - 2040 BAU to 2040 VMT and Emissions Changes - 2010 Baseline to 2040
Scenario Scenario
oueiidiiu
Scenario 1 :
Expand
Ridesharing
and TDM
programs
Scenario 2:
Expand
Ridesharing
and TDM
programs +
Transit
Improvement
and Promotion
Scenario 3:
Expand
Ridesharing
and TDM
programs +
Transit
Light-Duty VMT
GHGs
(C02 PM2.5
equivalent)
-51,104,975
-51,824,957
-144,155
-364,530
-8
-20
NOx
-92
-234
-51,895,054 -385,975 -21 -248
Improvement
and Promotion
+ Smart
Growth Land
Use
voc
Light-Duty
\/MT
GHGs
(C02
PM2.5
NOx
equivalent)
-63 12,821,202
-159
12,101,220
VOC
-409,306 -410 -33,971 -14,180
-629,681
-169 12,031,123 -651,126
-422
-34,113 -14,276
-423 -34,126 -14,286
ICF International
10-000
B-2
U.S. Environmental Protection Agency
-------
Appendix B: Emission Change Quantities and Additional Technical Details for the MOVES Analysis
Scenario 4:
Expand
Ridesharing
and TDM
programs +
Transit
Improvement -56,738,678
and Promotion
+ Smart
Growth Land
Use +
Transportation
Pricing
-1,868,248
-100
-1,200
-820 7,187,499 -2,133,399 -503 -35,079 -14,937
In addition to the four primary pollutants presented here, numerous other pollutants were also
included in the analysis as done in the initial study. Table 19 lists the full set of pollutants
modeled.
Ammonia (NH3)
Nitrous Acid (MONO)
Table 19. Full Pollutant List
Pollutants
Primary PM10 - Brakewear
Particulate
Primary PM2.5 - Organic Carbon
Atmospheric C02
Nitrous Oxide (N20)
Primary PM10 - Elemental Carbon
C02 Equivalent Non-Methane Primary PM10- Organic Carbon
Hydrocarbons
Carbon Monoxide
(CO)
Methane (CH4)
Nitrogen Dioxide
(N02)
Nitrogen Oxide
(NO)
Non-Methane Organic
Gases
Oxides of Nitrogen (NOx)
Primary Exhaust PM10 -
Total
Primary Exhaust PM2.5 -
Total
Volatile Organic Compounds
Primary PM10 - Sulfate Particulate
Primary PM10-Tirewear
Particulate
Primary PM2.5 - Brakewear
Particulate
Primary PM2.5 - Elemental Carbon
Primary PM2.5 - Sulfate
Particulate
Primary PM2.5 - Tirewear
Particulate
Sulfur Dioxide (S02)
Total Energy Consumption
Total Gaseous Hydrocarbons
Total Organic Gases
A variety of MOVES vehicle and fuel types were included in the analysis to characterize the
vehicle types used in the TRIMMS model. Those were:
Diesel Passenger Car
Diesel Passenger Truck
Diesel Transit Bus
Gasoline Light Commercial Truck
Gasoline Motorcycle
Gasoline Passenger Car
Gasoline Passenger Truck
Gasoline Transit Bus
ICF International
10-000
B-3
U.S. Environmental Protection Agency
-------
Estimating Emission Reductions from Travel Efficiency Strategies: Three Sketch Modeling Case Studies
For the majority of strategies considered here, additional VMT from increased transit ridership is
minimal, so transit emissions were excluded. For the PAG region, the influence of pre-
aggregation of input data on predictions was briefly considered. Our original approach was
designed to use annual aggregation for speed, but EPA cautioned against this because it would
compromise accuracy. The comparisons made in Table 20 and Table 21 indicate that the
differences in emissions can be significant, although activity differences are very small. Cases
where the data uses annual aggregation tend to underestimate annual emissions somewhat
relative to hourly inputs (without aggregation). Aggregation has minimal influence on activity.
The hourly activity results are within 5% of the yearly results - usually <1% difference.
Emissions vary more widely. Hourly results are usually greater than annual results - by an
average of 7%, although the range reached 75-85% for buses, especially for off-network
activities (starting). In a small number of cases hourly results were seen to be smaller than
annual results, by up to 5%.
Table 20. Difference in Activity Values: Hourly-to-Annual Aggregation
Distance Traveled Starts
Rural Restricted Rural Unrestricted Urban Restricted Urban Unrestricted Off-
Access Access Access Access Network
Light Commercial
Truck
Motorcycle
Passenger Car
Passenger Truck
Transit Bus
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
-3%
0%
1%
-4%
Table 21. Difference in Emission Values: Hourly-to-Annual Aggregation
Vehicle
Light Commercial
Truck
Pollutant
Atmos C02
CO
C02 Equivalent
Methane (CH4)
N20
NMHC
NOx
PM10 Brakewear
PMIOEIemC
PM 10 Organic C
PM10 Sulfate
PM10 Tirewear
PM2.5 Brakewear
Off-Network
-10%
-11%
-7%
-12%
0%
-8%
-4%
-24%
-24%
-11%
Rural
Restricted
Access
-1%
-4%
-1%
-1%
0%
-1%
-1%
0%
-1%
-17%
-1%
0%
0%
Rural
Unrestricted
Access
-1%
-4%
-1%
-1%
0%
-1%
-2%
0%
-1%
-13%
-1%
0%
0%
Urban
Restricted
Access
-1%
-4%
-1%
-1%
0%
-1%
-1%
-1%
-1%
-16%
-1%
0%
-1%
Urban
Unrestricted
Access
-4%
-6%
-4%
-5%
-6%
-5%
-7%
-6%
-3%
-13%
-4%
-2%
-6%
ICF International
10-000
B-4
U.S. Environmental Protection Agency
-------
Appendix B: Emission Change Quantities and Additional Technical Details for the MOVES Analysis
Vehicle
Rural Rural Urban Urban
Pollutant Off-Network Restricted Unrestricted Restricted Unrestricted
Access Access Access Access
Motor-cycle
Passenger Car
nviz.o ciem w
PM2.5 Organic C
PM2.5 Sulfate
PM2.5 Tirewear
PM2.5 Total Exh
Total Energy
Total Gas HC
VOC
Atmos C02
CO
C02 Equivalent
Methane (CH4)
N20
NMHC
NOx
PM10 Brakewear
PMIOEIemC
PM 10 Organic C
PM10 Sulfate
PM 10 Tirewear
PM2.5 Brakewear
PM2.5 Elem C
PM2.5 Organic C
PM2.5 Sulfate
PM2.5 Tirewear
PM2.5 Total Exh
Total Energy
Total Gas HC
VOC
Atmos C02
CO
C02 Equivalent
Methane (CH4)
N20
NMHC
NOx
PM10 Brakewear
PMIOEIemC
-ZT/O
-24%
-11%
-24%
-10%
-8%
-8%
-6%
-48%
-6%
-24%
3%
-15%
-30%
-29%
-29%
-9%
-29%
-29%
-9%
-29%
-6%
-17%
-15%
-10%
-29%
-8%
-15%
0%
-13%
-4%
-31%
- I/O
-17%
-1%
0%
-7%
-1%
-1%
-1%
0%
0%
0%
0%
0%
0%
0%
0%
-21%
-21%
0%
0%
0%
-21%
-21%
0%
0%
-21%
0%
0%
0%
-1%
-5%
-1%
-1%
0%
-1%
-2%
0%
-19%
- I/O
-12%
-1%
0%
-5%
-1%
-1%
-1%
0%
0%
0%
0%
0%
0%
0%
0%
-21%
-21%
0%
0%
0%
-21%
-21%
0%
0%
-21%
0%
0%
0%
-1%
-4%
-1%
-1%
0%
-1%
-2%
0%
-20%
- I/O
-16%
-1%
0%
-7%
-1%
-1%
-1%
0%
0%
0%
0%
0%
0%
0%
-1%
-21%
-21%
0%
0%
-1%
-21%
-21%
0%
0%
-21%
0%
0%
0%
-1%
-4%
-1%
-1%
0%
-1%
-2%
-1%
-20%
-O70
-13%
-4%
-2%
-8%
-4%
-5%
-5%
-1%
-1%
-1%
-4%
-6%
-4%
1%
-7%
-22%
-22%
-1%
-2%
-7%
-22%
-22%
-1%
-2%
-22%
-1%
-4%
-4%
-4%
-6%
-4%
-4%
-6%
-5%
-6%
-6%
-22%
ICF International
10-000
B-5
U.S. Environmental Protection Agency
-------
Estimating Emission Reductions from Travel Efficiency Strategies: Three Sketch Modeling Case Studies
Vehicle
Pollutant
Rural Rural Urban Urban
Off-Network Restricted Unrestricted Restricted Unrestricted
Access Access Access Access
Passenger Truck
Transit Bus
PM 10 Organic C
PM10 Sulfate
PMIOTirewear
PM2.5 Brakewear
PM2.5 Elem C
PM2.5 Organic C
PM2.5 Sulfate
PM2.5 Tirewear
PM2.5 Total Exh
Total Energy
Total Gas HC
VOC
Atmos C02
CO
C02 Equivalent
Methane (CH4)
N20
NMHC
NOx
PM10 Brakewear
PMIOEIemC
PM 10 Organic C
PM10 Sulfate
PMIOTirewear
PM2.5 Brakewear
PM2.5 Elem C
PM2.5 Organic C
PM2.5 Sulfate
PM2.5 Tirewear
PM2.5 Total Exh
Total Energy
Total Gas HC
VOC
Atmos C02
CO
C02 Equivalent
Methane (CH4)
-31%
-13%
-31%
-31%
-13%
-31%
-10%
-13%
-13%
-11%
-13%
-8%
-12%
-1%
-8%
-3%
-32%
-32%
-13%
-32%
-32%
-13%
-32%
-11%
-8%
-8%
-7%
3%
-7%
-85%
-21%
-1%
0%
0%
-19%
-21%
-1%
0%
-20%
-1%
-1%
-1%
-1%
-4%
-1%
-1%
0%
-1%
-1%
0%
-6%
-20%
-1%
0%
0%
-5%
-20%
-1%
0%
-17%
-1%
-1%
-1%
-1%
0%
-1%
0%
-21%
-1%
0%
0%
-19%
-21%
-1%
0%
-20%
-1%
-1%
-1%
-1%
-4%
-1%
-1%
0%
-1%
-2%
0%
-4%
-19%
-1%
0%
0%
-4%
-19%
-1%
0%
-15%
-1%
-1%
-1%
-1%
0%
-1%
0%
-21%
-1%
0%
-1%
-20%
-21%
-1%
0%
-21%
-1%
-1%
-1%
-1%
-4%
-1%
-1%
0%
-1%
-1%
-1%
-6%
-20%
-1%
0%
-1%
-6%
-20%
-1%
0%
-17%
-1%
-1%
-1%
-1%
0%
-1%
0%
-23%
-4%
-2%
-6%
-22%
-23%
-4%
-2%
-23%
-4%
-5%
-5%
-4%
-6%
-4%
-5%
-6%
-5%
-5%
-6%
-7%
-20%
-3%
-2%
-6%
-6%
-20%
-3%
-2%
-17%
-4%
-5%
-5%
-3%
-3%
-3%
-4%
ICF International
10-000
B-6
U.S. Environmental Protection Agency
-------
Appendix B: Emission Change Quantities and Additional Technical Details for the MOVES Analysis
Vehicle
Pollutant
Rural Rural Urban Urban
Off-Network Restricted Unrestricted Restricted Unrestricted
Access Access Access Access
N20
5%
0%
0%
0%
-6%
NMHC
NOx
PM10 Brakewear
PMIOEIemC
PM 10 Organic C
PM10 Sulfate
PM10 Tirewear
PM2.5 Brakewear
PM2.5 Elem C
PM2.5 Organic C
PM2.5 Sulfate
PM2.5 Tirewear
PM2.5 Total Exh
Total Energy
Total Gas HC
VOC
-74%
-85%
0%
0%
-7%
1%
0%
-7%
0%
-7%
-75%
-74%
0%
0%
0%
0%
0%
-1%
0%
0%
0%
0%
-1%
0%
0%
-1%
0%
0%
0%
0%
0%
0%
0%
-1%
0%
0%
0%
0%
-1%
0%
0%
-1%
0%
0%
0%
0%
-1%
0%
-1%
-1%
0%
-1%
0%
-1%
-1%
0%
0%
-1%
0%
0%
-5%
-3%
-6%
-1%
-5%
-3%
-2%
-6%
-1%
-5%
-3%
-2%
-2%
-3%
-5%
-5%
Based on these issues, all MOVES runs for this project were run without pre-aggregation, and
hourly outputs from MOVES were manually aggregated to annual.
Another special case involved disaggregating MassDOT provided data. The source populations
and total VMTs from MassDOT were cumulative across the seven-county domain. The total
VMTs were not allocated to the HPMS vehicle types required by MOVES, so VMT by vehicle
type was estimated from MassDOT-provided factors by multiplying the total VMT by the
MassDOT-provided allocation factors shown in Table 22.
Table 22. MassDOT VMT allocation factors
HPMS Vehicle Type
Motorcycle
Passenger Car
Passenger Truck/Light Commercial Truck
Intercity Bus, Transit Bus, School Bus
Refuse Truck, Short-Haul Single Unit, Long-Haul Single Unit, Motorhomes
Short-Haul Combination, Long-Haul Combination
Allocation
Factor
0.50%
51.40%
45.20%
0.10%
0.70%
2.00%
ICF International
10-000
B-7
U.S. Environmental Protection Agency
------- |