United States Air and Radiation EPA420-P-99-004
Environmental Protection M6.SPD.004
Agency February 1999
&EPA Guidance for the
Development of Facility
TypeVMT and
Speed Distributions
DRAFT
> Printed on Recycled Paper
-------
EPA420-P-99-004
February 1999
for the of Type
M6.SPD.004
Assessment and Modeling Division
Office of Mobile Sources
U.S. Environmental Protection Agency
NOTICE
This technical report does not necessarily represent final EPA decisions or positions.
It is intended to present technical analysis of issues using data which are currently available.
The purpose in the release of such reports is to facilitate the exchange of
technical information and to inform the public of technical developments which
may form the basis for a final EPA decision, position, or regulatory action.
-------
SYSTEMSAPPLICA TIONSINTERNA TIONAL
Final Report
GUIDANCE FOR THE DEVELOPMENT OF FACILITY TYPE
VMT AND SPEED DISTRIBUTIONS
SYSAPP-98/32r
September 1998
EPA Contract No. 68-C6-0068
Work Assignment No. 1-03
Prepared for
Acurex Environment Corporation
555 Clyde Avenue
Mountain View, California 94043
Prepared by
Systems Applications International, Inc.
ICF Kaiser Consulting Group
101 Lucas Valley Road
San Rafael, California 94903
98320 ir.wpd Final Report — September 1998
-------
SYSTEMSAPPLICA TIONSINTERNA TIONAL iii
Table of Contents
1.0 PURPOSE, BACKGROUND, AND GENERAL APPROACH 1
1.1 PURPOSE 1
1.2 OVERVIEW OF MOBILE6 CHANGES NECESSITATING THIS
GUIDANCE 3
1.3 SUMMARY OF GUIDANCE 4
2.0 VMT DATA SOURCES 7
3.0 WORKING WITH TRAFFIC COUNT DATA 9
3.1 RECONCILING LOCAL COUNT DATA WITH HPMS DATA . . 9
3.2 ISSUES WITH LOCAL COUNT DATA SOURCES 9
3.3 ESTIMATING VMT BY FUNCTIONAL CLASS USING COUNT DATA
10
3.4 SPEED ESTIMATION PROCEDURES FOR COUNT DATA ... 11
4.0 WORKING WITH REGIONAL TRANSPORTATION MODEL OUTPUTS
17
4.1 PROCESSING TDM OUTPUTS 17
4.2 SPEED ESTIMATION PROCEDURES USING TRAVEL DEMAND
MODEL DATA 17
5.0 APPLICATION OF METHODOLOGY TO DATA FROM FIVE CITIES .... 19
5.1 DATA COLLECTED 19
5.2 APPLICATION OF TRAVEL DEMAND MODEL OUTPUT
METHODOLOGY 23
5.3 APPLICATION OF TRAFFIC COUNT DATA METHODOLOGY 35
GLOSSARY 43
983201r.wpd Final Report — September 1998
-------
iv SYSTEMSAPPLICA TIONSINTERNA TIONAL
List of Tables
1 Summary of speed cycles developed by Sierra Research 4
2 Summary of speed estimation method features 13
3a Charlotte Department of Transportation (CDOT) counts for 1995 20
3b Charlotte DOT and North Carolina DOT count data 21
3c New York traffic count dataset 22
3d Chicago 1996 transportation model outputs 22
3e Ada County 1995 and 2015 transportation model outputs 23
3f Houston 2020 transportation model outputs 23
4a Distribution of 1995 hourly VMT by functional class and speed
during the morning peak (8 a.m.) for Ada County 25
4b 1995 VMT distribution by functional class and speed for Ada County . . 26
4c Overall VMT distributions for Ada County in 1995 (miles/day) 26
4d Fraction of freeway VMT used to calculate ramp VMT by CDOT 26
5a Summary of 1996 light duty vehicle VMT distribution by functional
class and speed during AM peak (8 a.m.) for Chicago (miles/day) 27
5b Summary of 1996 heavy duty vehicle VMT distribution by functional
class and speed during AM (8 a.m.) for Chicago (miles/day) 28
5c Summary of 1996 light duty VMT fraction by functional class
and speed during AM Peak (8 a.m.) for Chicago 29
5d Summary of 1996 heavy duty VMT fraction by functional class
and speed during AM peak (8 a.m.) for Chicago 30
5e Light duty vehicle VMT distributions for Chicago in 1996 (miles/day) . 30
5f Heavy duty vehicle VMT distributions for Chicago in 1996 (miles/day) . 30
5g Total VMT distributions for Chicago in 1996 (miles/day) 31
6a Freeflow speeds (mph) and capacities for Houston 31
6b Summary of 2020 PM Peak (5p.m.) VMT by functional class and
speed for Houston using speed processor (miles/day) 32
6c Summary of 2020 PM Peak (5 p.m.) VMT fractions by functional
class and speed for Houston using speed processor (miles/day) 33
6d Summary of 2020 PM Peak (5 p.m.) VMT by functional class and speed
for Houston using HGAC transportation model speeds (miles/day) .... 34
Final Report — September 1998 98320 Ir.wpd
-------
SYSTEMSAPPLICA TIONSINTERNA TIONAL
6e Summary of 2020 PM Peak (5 p.m.) VMT fractions by functional class and
speed for Houston using HGAC transportation model speed (miles/day) . 35
6f Overall VMT distributions for Houston in 2020 (miles/day) 35
7a Hourly distributions of travel by functional class for Charlotte 37
7b Distribution of link VMT by functional class for Charlotte 37
7c Freeflow speeds for Charlotte (mph) 38
7d Capacities for Charlotte (volume per hour) 38
7e Summary of AM peak (8 a.m.) VMT distributions by functional
class and speed for Charlotte 39
7f Total number of count sites by functional class for Charlotte 39
8a Hourly distributions of travel by functional class for urban New York . . 41
8b Urban Freeflow speeds, capacities, and number of lanes for New York . . 41
8c Distribution of link VMT by functional class for urban New York 42
8d Summary of AM Peak (8 a.m.) VMT distributions by functional class
and speed for urban New York 42
8e Total number of count sites by functional class for urban New York .... 42
983201r.wpd Final Report — September 1998
-------
vi SYSTEMSAPPLICA TIONSINTERNA TIONAL
Final Report — September 1998 98320 Ir.wpd
-------
SYSTEMSAPPLICA TIONSINTERNA TIONAL
1.0 PURPOSE, BACKGROUND, AND GENERAL APPROACH
1.1 Purpose
This guidance is provided to assist users of the MOBILE6 highway vehicle emission factor model
in the preparation of traffic activity inputs. It offers the U.S. Environmental Protection Agency's (EPA)
recommendations on how to develop distributions of vehicle miles traveled (VMT) by time of day, facility
type and average speed. These distributions are required by MOBILE for the development of an areawide
emission factor for light duty vehicles, as well as facility-specific emission factors. Although national
default distributions have been developed, there is sufficient variation in roadway network characteristics
between areas that the use of locally developed distributions is strongly preferred. In almost all urbanized
areas, some local data will be available that can be used to develop improved activity inputs for MOBILE6.
The methods presented here were developed based on a review of locally available data and tools
that could support facility-specific VMT and speed estimates and existing transportation planning tools and
methods. The report documenting this review1 contains both the detailed results obtained for five selected
urban areas on which the methods were tested (summaries of which appear here), and estimated national
time-of-day and speed distributions of urban VMT derived by extrapolation of results for four of the
selected urban areas.
The purpose of this guidance is to give suggestions which will help states prepare region-specific
on-road vehicle area-wide emission factors using the MOBILE6 model. MOBILE6 produces facility-
specific emission factors, unlike previous versions of the model, which produced emission factors
corrected only to user-input average speeds without regard to facility type. This guidance provides
approaches for the determination of the distribution of VMT among the facility types and speeds modeled
in MOBILE6. This distribution of VMT can be used to prepare regional average emission factors, to
disaggregate regional VMT among MOBILE6 facility types, and to more accurately estimate emission
factors for specific areas and/or time periods. Most nonattainment and maintenance areas will develop
emission inventories on a Iink2-by-link basis using emission factors for a specific facility type and average
speed in a post processor. Area-wide emission factors will be used primarily for planning purposes, and
for some smaller areas which submit inventories in this format.
Many areas will already have more sophisticated methods to determine the distributions for VMT
than those described here. Direct measurement or modeling of vehicle activity has traditionally been
conducted in support of highway planning and traffic engineering decisions. State and local transportation
and traffic management agencies collect traffic count data at both permanent and temporary locations. In
addition to their use in the analysis of traffic patterns and trends, these data are commonly used in the
development and calibration of travel demand models3. In addition to counts, a variety of data collection
'SAI (1998). "Development of Methodology for Estimating VMT Weighting by Facility Type," Final Report
SYSAPP-98/llr2, Systems Applications International, Inc., San Rafael, California, September 1998.
2The language of transportation planning and traffic engineering assigns specific meanings to many common words.
For example, roadway segments in transportation networks are commonly referred to as "links." A glossary at the
end of this document presents a number of the more commonly used terms that air quality planners may encounter in
the development of traffic activity inputs for emission modeling.
3 Most commonly used travel demand models follow the "four-step" process in which socioeconomic data are used
to describe the number and type of trips between zones in an urban area, and assign trips to specific paths along the
roadway network. The results are used to evaluate the performance of the roadway network and the effects of
growth, highway construction, etc. on roadway congestion and travel time.
Final Report - September 1998 983210r.wpd
-------
SYSTEMSAPPLICA TIONSINTERNA TIONAL
techniques are used to characterize travel, including household surveys and trip diaries, license plate
surveys, and "floating car" studies. None of these methods provide truly comprehensive data on all travel
activity, and the costs of data collection further limit the availability of sound vehicle activity data.
Nevertheless, both routine monitoring and special studies enhance local knowledge of travel patterns and
traffic characteristics. In response to the EPA transportation conformity rules, many Metropolitan
Planning Organizations (MPOs) are conducting a range of efforts to enhance the quality of vehicle activity
estimates for air quality planning. These efforts focus on both expanded and improved data collection, as
well as studies targeted on model improvement4.
State and local air quality planners are strongly encouraged to coordinate closely with their local
transportation agencies, as well as their EPA regional office, before deciding on the approach to be used to
develop VMT and speed distributions for MOBILE. This coordination is vital to facilitate state
implementation plan (SIP) emission budget development, transportation conformity analyses5, and EPA
approval. In most urban areas, the designated metropolitan planning organizations (MPOs) will be the
primary source of information on regional traffic activity, but data available from MPOs may be
supplemented with information collected by city traffic departments. Because the comprehensiveness of
traffic data collection systems, the sophistication of existing traffic modeling, and the level of involvement
of city, regional and state traffic and environmental agencies varies greatly between areas, an initial
assessment of all potential sources of information is desirable. Any method chosen for providing VMT
inputs for MOBILE6 should be technically sound, based on local data collection and represent a consensus
of transportation and environmental concerns.
It is strongly preferred that local VMT-weighted speed distributions be prepared as inputs to
MOBILE6 so that emission estimates are as representative of local conditions as possible, and so that
emission forecasting procedures can accurately characterize changes in emissions attributable to
anticipated growth and roadway network changes. The emission factors will be much more meaningful,
whether used for planning purposes or for an official submittal to EPA. Extrapolating data from other
similar cities is discouraged. If the needed types of data are not available, EPA strongly recommends that
the area start an appropriate data collection program
In addition to this guidance, states should refer to the EPA's Section 187 VMT Forecasting and
Tracking Guidance6 and EPA's Procedures for Emission Inventory Preparation, Volume IV: Mobile
Sources7 for guidance on the preparation of regional on-road vehicle emission inventories.
4See, for example, papers by Janik ("Enhancing the Highway Performance Monitoring System in Northeastern and
Southeastern Illinois: An Assessment of the State of the VMT Estimating Practice in Illinois"), Stopher and Fu
("Feasible Improvements to Travel-Forecasting Procedures for Air Quality Analysis"), and Suhrbier et al.
("Improved Transportation Air Quality Analysis Methodologies") in Transportation Planning and Air Quality Ill-
Emerging Strategies and Working Solutions, Conference Proceedings, August 17-20, 1997, Lake Tahoe, California,
American Society of Civil Engineers, Reston, Virginia.
540CFR51 Subpart T, and 40CFR93 Subpart A.
6 EPA, 1992. Section 187 VMT Forecasting and Tracking Guidance. Prepared by the U.S. Environmental Protection
Agency. January 1992.
7 EPA, 1992. Procedures for Emission Inventory Preparation, Volume IV: Mobile Sources. Prepared by the U.S.
Environmental Protection Agency. 1992. EPA-450/4-81-026d (Revised).
Final Report - September 1998 983210r.wpd
-------
SYSTEMSAPPLICA TIONSINTERNA TIONAL
1.2 Overview of MOBILE6 Changes Necessitating This Guidance
The Clean Air Act Amendments of 1990 (CAAA) mandated a closer look at "real world driving,"
that is, driving modes not covered by the Federal Test Procedure (FTP) upon which MOBILE historically
was based, and which is used for certifying vehicle compliance with emission standards. Emission factors
developed from FTP data were said to represent standard conditions of average speed, temperature, and
fuels. These factors were adjusted to "non-standard" conditions using relationships encoded in MOBILE.
For example, the emissions effects of average speeds other than the FTP's 19.6 mph were modeled using
emissions versus speed relations developed from vehicle test data collected using several different test
cycles with average speeds that differed from the FTP.
This historical treatment of speed effects did not readily allow emission modelers to address the
fact that the distribution of speeds associated with a particular average speed could be dramatically
different for different types of roadway. For example, an average speed of 30 mph for a freeway will
typically involve relatively constant speeds, but for arterials, would more likely include periods of idle and
acceleration to higher speeds. VMT estimates are available from both direct measurement programs and
from regional transportation models, broken down by the roadway type (also referred to as facility type or
functional class). MOBILE6 capitalizes on the availability of this information by providing different
emission factors for different facility types. These emission factors will better represent driving patterns
than factors from previous versions of MOBILE.
In its review of the FTP driving cycle, EPA collected both chase car data and instrumented vehicle
data in Baltimore, Maryland, and Spokane, Washington, which was supplemented by an instrumented
vehicle study conducted in Atlanta, Georgia and a chase car study conducted by the California Air
Resources Board in Los Angeles, California. The driving patterns in the instrumented vehicle studies
show that some types of facility-specific driving conditions contain more frequent and more extreme
acceleration and deceleration than others, which reach a similar speed but remain at a steady cruise. Based
on the instrumented vehicle and chase car data, EPA has developed a new set of driving cycles that
represent passenger car and light truck operation on a variety of roadway types and under a variety of
congestion levels and average speeds.8 Four types of roadways are treated: freeways, arterials and
collectors, local roads, and freeway on- and off-ramps. The MOBILE6 facility-specific emission factors are
an attempt to quantify the emission differences for facility-specific driving activity.
The disaggregation of emission factors by facility type results in an emission factor most
appropriate for short time periods, as compared to the those provided by earlier versions of MOBILE
which could be applied to daily total VMT for a single average speed. The user-input requirements are
significantly different with these new cycles, in that the user can specify hourly link-by-link traffic volumes
and speeds for detailed spatially and temporally resolved emission inventories.
The guidance presented here describes how vehicle count data or traffic model volumes can be
used to calculate hourly, link-specific speeds based on the link characteristics and level of congestion. By
aggregating the link-specific VMT across all links into fifteen speed "bins" (0-2.5 mph, 2.5-7.5 mph, and
so on up to 67.5-72.5 mph), hourly areawide speed distributions can be developed. To develop area-wide
emission factors using MOBILE6, the user must input a 24-by-4 matrix containing the percent of VMT for
each of the four facility types for each hour. In addition, because emission factors for freeways and
arterials/collectors are speed dependent, the user must provide speed distributions for each hour for these
two facility types for each hour of the day. In summary, the MOBILE6 input requirements addressed here
are: a 24-by-4 matrix containing hourly VMT for each facility type; and a 24-by-15-by-2 matrix containing
8 Sierra Research, 1997. Development of Speed Correction Cycles. Prepared for U.S. Environmental Protection
Agency by Sierra Research, Sacramento, California. April 30, 1997. Report No. SR97-04-01.
Final Report - September 1998 983210r.wpd
-------
SYSTEMSAPPLICA TIONSINTERNA TIONAL
the fraction of two facility types' VMT that occurs in each speed bin for each hour. The MOBILE6 user
does not have the option of specifying speed distributions for local roads or ramps, as emission rates for
these facility types are based on one driving cycle to represent local driving and one driving cycle to
represent ramp driving. The VMT fractions are used to weight the facility-specific emission factors at each
speed to produce an areawide running emission factor which is representative of the urban area.
1.3
Summary of Guidance
As presented above, the MOBILE6 model provides facility-specific emission factors and uses
separate speed correction curves for each facility type to adjust these factors for the user-input speed. This
is quite different from past versions of the model, which used the same speed correction relationships
regardless of facility type. To allow the continued use of MOBILE emissions for estimation of regional
total, as opposed to link-level, emissions, this guidance has been prepared to instruct users on the
allocation of VMT by facility type and speed, which can be weighted together to generate customized area-
wide emissions.
Characteristics of the new facility-specific cycles in MOBILE6 are summarized in Table 1. The
model produces emission factors for four facility classes: freeways, freeway ramps, arterials/collectors, and
local roadways. The factors for freeways and arterials/collectors are based upon cycles developed to reflect
specific levels of service (LOS) on these facilities. Level of service is a measure describing the operating
conditions on a particular roadway as affected by the level of congestion. Although defined in terms of
vehicle density (vehicles per lane-mile), LOS also relates to speed, freedom to maneuver, interruptions, and
safety.
Table 1. Summary of speed cycles developed by Sierra Research.9
Cycle
Freeway, High Speed
Freeway, LOS A-C
Freeway, LOS D
Freeway, LOS E
Freeway, LOS F
Freeway, LOS "G"
Freeway Ramps
Arterials/Collectors, LOS A-B
Arterials/Collectors, LOS C-D
Arterials/Collectors, LOS E-F
Local Roadways
Average
Speed (mph)
63.2
59.7
52.9
30.5
18.6
13.1
34.6
24.8
19.2
11.6
12.9
Maximum
Speed (mph)
74.7
73.1
70.6
63.0
49.9
35.7
60.2
58.9
49.5
39.9
38.3
Maximum
Acceleration Rate
(mph/s)
2.7
3.4
2.3
5.3
6.9
3.8
5.7
5.0
5.7
5.8
3.7
Currently, agencies developing regional emission inventories are not required to disaggregate
vehicle activity for a particular facility type by speed, LOS, or other measures of traffic density, such as the
traffic volume to service capacity (V/C) ratio. EPA guidance on the level of detail required in reporting
} See Id.
Final Report - September 1998
983210r.wpd
-------
SYSTEMSAPPLICA TIONSINTERNA TIONAL
on-road inventories requires, however, that VMT and emissions be disaggregated by vehicle class and
facility type.10 Two methods are typically available for arriving at these VMT estimates. The first derives
VMT estimates from vehicle count data, roadway network information (lane-miles by facility class), and
assumptions regarding the representativeness of the count data. The second uses the link-level traffic
volumes and network information produced by regional transportation models. Both methods can be
adapted to also provide distributions by speed through use of standardized relationships between congested
speed, freeflow speed, and V/C ratio.
If local data on observed speeds by facility type are available, these data should be used to refine
speed estimates produced by the methods discussed in this guidance. There are a variety of methods by
which speed data may have been collected and processed, and it is extremely important that the analyst
understand how to interpret specific speed data bases because of the variety of methods currently in use.
For example, pairs of in-road loop detectors provide instantaneous speeds of individual vehicles at a fixed
location which may not be representative of overall speed distributions along the roadway segment.
Average speeds may be calculated, as either arithmetic means or harmonic means (the inverse of the
average of the inverses of observed speeds). For speed measurements at a specific location, these are
sometimes referred to, respectively, as "time-mean speed" and "space-mean speed"11. The harmonic mean
is preferred because the arithmetic mean provides a positively biased estimate. Arithmetic means from
some other measurement methods (e.g., second-by-second recording in instrumented vehicles) do provide
unbiased space-mean speeds. It is recommended that EPA be consulted when making use of local speed
observation datasets.
The development of VMT distributions from vehicle count data and from network-based
transportation model outputs are described in detail in this guidance document. Examples of their
application are presented for five urban areas in the U.S.
10 EPA, 1992. "Example Documentation Report for 1990 Base Year Ozone and Carbon Monoxide State
Implementation Plan Emission Inventories." Prepared by Radian Corporation for EPA Office of Air Quality
Planning and Standards, Research Triangle Park, N.C. March 1992. EPA-450/4-92-007.
"Bowling et al, 1996."Planning Techniques to Estimate Speeds and Service Volumes," NCHRP 387, Final Report
prepared for National Cooperative Highway Research Program project 3-55(2), Transportation Research Board,
September 1996..
Final Report - September 1998 983210r.wpd
-------
SYSTEMSAPPLICA TIONSINTERNA TIONAL
Final Report - September 1998 983210r.wpd
-------
SYSTEMSAPPLICA TIONSINTERNA TIONAL
2.0 VMT DATA SOURCES
The methods available to a particular urban area for developing VMT fractions by facility class
depend upon the quality and types of local data available for determining the location of vehicle activity
and speeds or traffic density. At a minimum, the Highway Performance Monitoring System (HPMS)
provides data representative of urban areas in each state12. The HPMS data provide state level urban and
rural data on an annual average basis. These data can be obtained directly through the Federal Highway
Administration (FHWA). Summaries of HPMS data are also available on the Internet at
http://www.fhwa. dot. gov/ohim/1994/section5. htm.
HPMS summaries provide a reasonable starting point for estimating urban area VMT, but more
detailed information specific to a particular city can be developed by supplementing HPMS data with
additional local count data. Such data will usually allow the development of VMT distributions by facility
class, time of day, and speed. Vehicle count locations may provide either long-term records or intermittent
data at temporal resolutions from daily to 15-minutes. Within a given area, it is likely that several different
data collection procedures will be in use, necessitating assumptions in the analysis. For example, high
temporal resolution data (e.g., hourly counts) may only be available from a few locations for each facility
class, or for a limited period of time, and the representativeness of these data must be evaluated in
developing a weighting scheme to estimate temporal variation in traffic volumes. Consultation with state,
regional and city traffic engineers regarding the purpose for which data were collected and the
representativeness of different count sites and data sets is necessary if local data are used. Working with
vehicle counts, one can estimate the average traffic count for a specified time period by facility class
which, when combined with total centerline miles of roadway for a facility class, yields total VMT for the
time period. Hourly traffic volumes at count locations (either from direct measurement or the application
of temporal profiles from other sites) can be used to estimate speeds through the use of standard
relationships between freeflow speed, level of congestion (expressed as the ratio of traffic volume to
roadway capacity, or "V/C ratio"), and congested speed.
Similarly, the link-level traffic volumes and link lengths from network-based transportation models
can be used to calculate total VMT by link. Summing VMT by facility class then provides a distribution of
VMT by facility class. As for count data, speed-congestion relationships can be applied to each link to
obtain distributions by speed when working with local transportation models.
12FHWA (1987). "Highway Performance Monitoring System Field Manual," U.S. Department of Transportation,
Federal Highway Administration, FHWA Order M 5600.1 A, 1 December 1987.
Final Report - September 1998 983210r.wpd
-------
SYSTEMSAPPLICA TIONSINTERNA TIONAL
Final Report - September 1998 983210r.wpd
-------
SYSTEMSAPPLICA TIONSINTERNA TIONAL
3.0 WORKING WITH TRAFFIC COUNT DATA
State and local agencies will have available, at a minimum, traffic count data collected from the
HPMS for specific Federal Aid Urbanized Areas (FAUA). HPMS data represent traffic counts taken on a
sample of an area's roadway network and are adjusted for day-of-week and season and expanded to
include the area's entire roadway network. HPMS data may be supplemented by locally collected data.
The local data often provide more widespread, but not necessarily statistically balanced, coverage of local
roadways. Count data are available at different temporal resolutions: typically daily data are available at all
count sites, with a selection of continuous monitors providing data which can be used to arrive at hourly
traffic distributions.
3.1 Reconciling Local Count Data with HPMS Data
HPMS datasets are designed to provide a statistically balanced representation of the traffic
characteristics of the region for which they are collected. Local data can substantially enhance the richness
of this characterization, but may introduce biases due to over- or under-representation of specific facility
types, subregional characteristics, etc. or by the inclusion of data from atypical sites or dates. Therefore,
caution should be used in supplementing HPMS traffic count data with local data, and care taken to ensure
that biases are not incorporated into the dataset which can lead to inaccuracies in the VMT distributions
derived from them. It is recommended that EPA be consulted, as well as data providers and transportation
planners, when supplementing HPMS data with those from local traffic count programs. If uncertainties
are large, or analytical results prove to be sensitive to values derived from limited data sets, the collection
of additional data, potentially including both targeted short term studies and expanded long term
monitoring, is recommended.
3.2 Issues with Local Count Data Sources
Common problems with local count data sets include biases because of the roadway sample or
because of idiosyncrasies of the counting device. For example, areas using road tube counters may have
undercounts on multilane facilities. These result from two cars crossing the tube at the same time. In-road
loop detectors connected to central computer facilities are widely used in some urban areas, but
malfunctions are common, resulting in the need for procedures for identifying and correcting for missing
data. Also, data for a particular area that originates from different sources may be combined without
correcting for underlying differences in the data collection or processing methods. For example, a state
may collect data for a facility and derive average daily traffic (ADT) counts, while a county may collect
average annual weekday traffic (AAWT) counts. Differences in averaging periods would need to be
accounted for if the data were combined.
The traffic count data set may need to be adjusted to remove inherent biases. For example, counts
on freeways in high-density travel areas may be over-represented, leading to an overestimate of the average
freeway volume. If count locations can be matched to area type and distributions of land use by area type
are available, the counts can be proportionately weighted to ensure that they reflect average land use
characteristics. Although many areas have fairly detailed GIS databases that provide area types, at a
minimum, estimates of the proportion of land in different area types should be available from standard
USGS databases. Assigning specific count locations to area types may, in the absence of GIS databases,
require cumbersome review by planning personnel who are familiar with local land use patterns.
In some areas, either the number of counting sites for a particular facility type (or combination of
facility and area type) may be too small, or the duration of count data sets may be too short for the data to
provide a good representation of that facility types volumes and temporal profiles. In these cases, it may be
necessary to either combine similar facility types or to use data from another, similar class. For example, if
Final Report - September 1998 983210r.wpd
-------
7 0 SYSTEMSAPPLICA TIONSINTERNA TIONAL
no hourly traffic volume data are available for local roadways, temporal distributions obtained for minor
arterials or collectors could be assumed to be representative of local roadway VMT. The overall result,
however, is an increase in the associated uncertainty of these estimates.
3.3 Estimating VMT by Functional Class Using Count Data
It is relatively straightforward to estimate total VMT from vehicle count data. Most regions have
used similar methods to that described in this guidance to estimate VMT for regional inventories. The
procedural steps are:
1. Calculate the sum of counts in each facility type (by area type if that information is retained in the
data);
2. Determine the sample size in each facility type (i.e., the number of count sites);
3. Determine the average volume for a facility type by dividing total count by sample size (this will
usually be average daily volume, unless a representative body of hourly data are available to perform
the calculation on an hourly basis);
4. Obtain total centerline miles of each facility type in the modeling domain (these are available from
Departments of Transportation or geographic information system (GIS) databases);
5. Multiply average volume by the number of centerline miles for each facility type to estimate total
VMT for each facility type.
Although there are four facility types explicitly modeled in MOBILE6 (freeways,
arterial/collectors, freeway ramps, and locals), there are a number of facility type definitions in use
throughout the country. For example, HPMS tracks travel activity in urban areas for interstate, other
freeway or expressway, other principal arterial, minor arterial, and collector. It will commonly be
necessary to group functional classes together to obtain VMT totals according to the MOBILE6 classes. In
the examples shown in this document, local terminology has been accepted as the basis for such grouping
(e.g., major arterial, minor arterial, major collector and minor collector would all be assigned to the
arterial/collector facility class). Some caution may be needed in making such assignments, as the principal
criteria for such assignment should include vehicle speeds and the nature of traffic control. As noted later,
cases have been observed in which observed speeds on roadways classified as "local" significantly exceed
the 12.9 mph average of the local roadway driving cycle underlying MOBILE6 emissions. In such cases,
some portion of local roads should be grouped with the arterial/collector class. Ideally, measured speed
distributions for each locally defined class of roadways should be compared with the speed distributions of
the MOBILE6 driving cycles to select the class that matches best.
Many traffic count databases, including HPMS, do not include counts for freeway ramps. It may
be necessary to assume that ramp VMT represents a specific fraction of freeway VMT. Such ramp VMT
fractions may be developed from the data in local transportation models, or they may be estimated by
comparison with ramp VMT data which are available for regions with similar characteristics to the one
modeled. Ramp emissions have the potential for being a significant component of total emissions, and
data and assumptions should be carefully documented and reviewed with transportation planners and EPA.
The result of this procedure is a distribution of VMT by facility type. Often, total centerline miles
for each functional class can also be obtained by area type from geographic information system (GIS)
databases used in conjunction with regional travel demand models. If such data are available, one can use
the data in the above procedure in order to obtain VMT by functional class and area type. At a minimum, it
is expected that most urban areas will have the capability to arrive at these types of estimates using similar
methods to those described above.
Final Report - September 1998 983210r.wpd
-------
SYSTEMSAPPLICA TIONSINTERNA TIONAL 11
Typically, the above procedure is carried out with average annual daily traffic (AADT) counts.
However, one can extrapolate from available hourly count data to develop time of day distributions of
vehicle activity by facility type. Most regions will have hourly count data, although these data may only be
representative of average weekdays and may not fully represent the combination of facility and area types
in the region. It is strongly recommended that areas collect appropriate data, but in the event that local data
are lacking it may be possible to extrapolate from data that have been collected for other cities. One should
look for cities with somewhat similar population and meteorological characteristics. The result of these
calculations is VMT distributions by facility type and time of day. It is recommended that states consult
with EPA before applying data from other areas in order to avoid later problems. Seasonal effects may
also be important, both for VMT totals and distribution by time of day. These effects are particularly
noticeable in areas with seasonal tourist influx. Weekday/weekend differences are large in most areas, and
day of week variation may be important in some areas. Obtaining raw (disaggregated) count data in
electronic form can facilitate the development of appropriate temporal allocation and day of week,
monthly, and seasonal adjustment factors. Care should be taken to consciously select and document the
data used for VMT distribution estimates.
3.4 Speed Estimation Procedures for Count Data
The next level of desired detail is the speed distribution of VMT for the two facility types with
speed-dependent emissions: freeway and arterial/collector. Although many urban areas track level of
service (LOS) for arterials and freeways for peak and off-peak periods, EPA does not believe that agencies
will generally have available robust databases of observed speed. Also, LOS classifications cannot be
directly used as inputs to MOBILE6 in place of speed distributions. Therefore, procedures are needed to
estimate speeds from the traffic count data.
Much of the information presented here on speed estimation procedures is taken from the
previously cited National Cooperative Highway Research Program (NCHRP) study of speed and service
volume estimation procedures13. There are generally two methods available for estimating speeds. The
first uses procedures from the Highway Capacity Manual14 (HCM). The second uses the speed-congestion
relationships based on the ratio of traffic volume to roadway capacity (V/C ratio) of the Bureau of Public
Roads (BPR), known as the "BPR curves" or "modified BPR curves." As traffic volumes approach
roadway capacity (which is determined by factors affecting driving behavior, such as lane width, median
width, roadway curvature, distance between side streets, etc.), speeds can drop and vehicle densities can
increase rapidly. The accuracy of speed predictions under severely congested conditions is limited,
regardless of the methods used. Also, for arterials, the effects of cross traffic and traffic signals on
congested speed are difficult to reduce to a manageable calculation scheme, and the accuracy of both HCM
and BPR procedures for arterials is limited. Table 2 summarizes features of both methods. Each method is
discussed below.
3.4.1 Highway Capacity Manual Procedure
HCM procedures separately address "uninterrupted flow" and "interrupted flow" facilities (i.e.,
freeways and arterials/collectors). For freeways, the basic procedure involves: (1) the calculation of an
adjusted segment volume that addresses lane and shoulder width, driver aggressiveness, fraction of large
vehicles, among other considerations; and (2) the use of a speed lookup chart based on the nominal
13 Bowling et al., 1996."Planning Techniques to Estimate Speeds and Service Volumes," NCHRP 387, Final Report
prepared for National Cooperative Highway Research Program project 3-55(2), Transportation Research Board,
September 1996.
14TRB (1994). "Highway Capacity Manual," Special Report 209, Third Edition, Transportation Research Board,
1994.
Final Report - September 1998 98321 Or.wpd
-------
12 SYSTEMSAPPLICA TIONSINTERNA TIONAL
freeflow speed of the facility. For interrupted flow facilities, the procedure again includes calculating
adjusted volumes, but speed calculation requires separate calculations of the running time between signals
(based on distance and freeflow speed) and the intersection delay (based on signal timing, V/C ratio, and
other factors), which are then combined into an average speed.
For general information, the data requirements for the HCM procedure are as follows.
Uninterrupted Flow Facilities (Freeways):
• Hourly volume
• Number of lane
• Free-flow speed
• Peak hour factor
• Lane and shoulder widths
• Percent trucks
• Percent recreational vehicles
• Terrain type
• Predominant driver type.
Interrupted Flow Facilities Techniques (Arterial/Collectors):
• Hourly volumes
• Number of lanes
• Free-flow speed
• Arterial class
• Density of signals per mile
• Peak hour factor
• Percentage turning traffic from exclusive lanes
• Medians
• Exclusive turn lanes
• Green time per cycle
• Cycle length
• Quality of signal progression
• Signal controller type.
Note that, for both facility types, default look-up tables can be created for many of these variables allowing
one to apply these methods provided one has facility and area type. Of course, use of defaults rather than
facility-specific data reduces the accuracy of the resultant speeds.
The HCM method requires more facility-specific information than is likely to be available, and in
some cases cannot be applied for V/C ratios greater than 1.0. Therefore, because of its lower data
requirements and ease of application, and despite its limitations, the BPR method appears the most
practical for typical urban areas and it is our recommended method for speed estimation. We note,
however, that with the expanding use of GIS for developing integrated data bases with roadway
characteristics for transportation planning, analysis and modeling, urban areas are beginning to develop
richer databases that could allow application of the HCM procedure on a regional basis, or perhaps hybrid
approaches that specifically address signal timing, coordination, and congestion.
Table 2. Summary of Speed Estimation Method Features
Criteria
L Data Reauirements
Volume/Capacity Curves (BPR)
Highway Capacity Manual
Final Report - September 1998 983210r.wpd
-------
SYSTEMSAPPLICA TIONSINTERNA TIONAL
13
Amount
Precision
Feasibility
2. Ease of Use
Complexity
Training Required
Spreadsheet
3. Reliability
Accuracy
Facilities
Area Types
Planning Applications
4. User Confidence and
Acceptance
Overall Use
Planning Applications
Agencies
Geographic Spread
5. Significant Strengths and
Deficiencies
Strengths
Deficiencies
-volume, capacity, free speed
-a 10% error in volume or capacity
translates into a 1 9% change in the
estimated speed at v/c = 1 .00
-all required data are feasible for all
agencies to easily obtain
-single equation
-few minutes to learn
-spreadsheet friendly
-not accurate at high v/c ratios
-all, but not reliable for interrupted
flow facilities
-all
-Good only for RTF models
-used by 22% of all respondents
-predominant technique for RTFs
-most popular with MPOs
-least popular with local traffic
agencies
-used throughout USA
-simple, quick, well-behaved function
1 . Not accurate at V/C's > 1 .00
2. Needs to be refitted to new HCM
data
3. Not sensitive to signal timing
-volume, free speed, plus numerous
additional facility characteristics
-complexity of procedures make it
difficult to determine impacts of data
errors
-40% of MPOs indicated it is not
feasible to obtain some of the required
data items (% heavy vehicles, quality
of coordination were most difficult)
-multiple equations
-one-day training
-adaptable to spreadsheets, but figures
must be translated to look-up tables
-most accurate of available techniques
not in traffic model software
-no planning technique for
uninterrupted flow facilities systems
-interrupted flow technique designed
for only urban application
-rural road procedure limited to 60
mph design speed
-good for all except RTF models
-used by 33% of respondents
-predominant technique for site impact
and congestion management
-most popular with state DOTs
-least popular with MPOs
-most frequently used across the
country but less popular on west coast
-comprehensive, sensitive to many
factors
1 . Extensive data required
2. Complex procedures
3. No procedure for freeway systems
4. Can't do V/C> 1.00
5. Rural roads procedure limited
3.4.2 Bureau of Public Roads Procedure
In contrast to the HCM procedure, the BPR is not data intensive. Default tables of capacity by
functional class are available, although the accuracy of the method is improved if individual facility
capacities are used. The standard BPR equation is:
= s/(l+a(v/c)b)
where:
Final Report - September 1998
983210r.wpd
-------
14 SYSTEMSAPPLICA TIONSINTERNA TIONAL
s = predicted mean speed
sf = free flow speed
v = volume
c = practical capacity
a = 0.05 for signalized facilities (arterials, collector, and local)
a = 0.20 for unsignalized facilities (freeways, highways, and expressways)
b = 10
Different values of the parameters a and b have been developed by some urban areas based upon
speed data sets, resulting in customized BPR curves. Practical capacity is defined as 80% of maximum
capacity. Free-flow speed is defined as the space mean speed15 of traffic when volumes are so light that
they have negligible effect on speed and is estimated to be 1.15 times the speed at capacity. Relationships
for space mean speed have been developed by Bowling et al, as follows:
Uninterrupted facilities with posted speed limits > 50 mph:
Mean speed (mph) = 0.88 * (posted speed limit in mph) + 14
Uninterrupted facilities with posted speed limits < 50 mph:
mean speed (mph) = 0.79 * (posted speed limit in mph) + 12
By entering either coded capacities by facility type or using default look-up tables, along with the
link volumes from traffic count data, link speeds can be predicted with the BPR equation. VMT within
each functional class can be grouped by speed, resulting in distributions of VMT by speed for freeways.
The accuracy with which the BPR curves predicts speeds for both arterials and freeways can
reportedly approach those of the HCM and traffic simulation models, provided that accurate free-flow
speeds and capacities for each facility are known. Predicted speeds are proportional to free-flow speed, but
drop rapidly as v/c approaches 1.0, making it particularly important to use reliable capacity values.
As noted above, the accuracy of these speed relationships is reduced for arterials and locals
because of traffic control effects. However, unless local data on control parameters by facility and area type
are available to at least construct look-up tables, regional planners are probably limited to the BPR curves
for estimating arterial speeds as well. A simple method to better estimate speeds as a function of facility
type is to differentiate the parameter "a" for signalized (a = 0.05) and unsignalized (a = 0.20) facility
classes. As discussed in the following section regarding travel demand models, speed processors have
been developed that make use of look up tables of default signalization data in order to improve speed
estimates for arterials, but the accuracy of these methods is still limited.
3.4.2.1 Time of Day Variations
Time of day variations in speeds can be accounted for by distributing traffic volumes by time of
day, as discussed above, and then applying the BPR equation with the appropriate capacities and volumes.
If available, day-of-week and seasonal effects can also be incorporated by applying appropriate
adjustments to link volumes or using traffic count data specific to the temporal period of interest. The use
of adjustment factors, particularly for time of day, can result in peak hour traffic volumes that exceed
roadway capacity, conditions under which uncertainties in the BPR relationships are greatest. While
significant congestion does occur in many areas, it is possible for adjustments to cause unrealistically low
estimated speeds for some roadways, requiring care and judgment on the part of the analyst.
15Space mean speed is the "true" average speed along a roadway as given by the total distance traveled by all
vehicles divided by the total travel time for the vehicles.
Final Report - September 1998 983210r.wpd
-------
SYSTEMS APPLICATIONS INTERNATIONAL 15
3.4.2.2 Future Year Estimates
Future year estimates can be developed by projecting VMT estimates by functional class and area
type (and speed bin if data permit) to the desired year based either on past trends or travel demand model
predictions. Regional growth and its effects on congestion, travel demand, and spatial distribution of travel
can cause significant shifts in VMT between functional classes and areas. Therefore, agencies should
review and update their VMT distributions when projecting inventories.
3.4.2.3 Distributions by Vehicle Class
The fraction of VMT accumulated by each vehicle class (light versus heavy duty or by FHWA
classifications) can be obtained from HPMS data by functional class for each state. Obtaining similar
distributions for a specific urban area or by speed bin is more difficult. The survey conducted as part of
the NCHRP planning techniques study found that forty percent of respondents would not be able to obtain
data on percent trucks by roadway type.16 In the absence of local data, the state HPMS estimates could be
assumed to be applicable. HPMS estimates do differentiate between urban and rural area types. In some
cases, the particular urban area under study may be the predominate source of data for the HPMS statistics
which makes this less of an extrapolation. EPA is unaware of any other readily available sources of data
that would allow agencies to develop distributions of VMT on freeways and arterials by vehicle class and
speed, although relatively limited data collection efforts could provide information that could either verify
state-level estimates or refine vehicle class information for specific regions..
16 Dowling et al. (1996).
Final Report - September 1998 983210r.wpd
-------
16 SYSTEMSAPPLICA TIONSINTERNA TIONAL
Final Report - September 1998 983210r.wpd
-------
SYSTEMSAPPLICA TIONSINTERNA TIONAL 17
4.0 WORKING WITH REGIONAL TRANSPORTATION MODEL OUTPUTS
Travel demand models (TDMs) provide another source of estimates of vehicle activity by
functional class, time of day, and speed. The modeling process assigns trips (defined by an origin and a
destination within the roadway network) to roadway segments. To the extent that model inputs capture all
trips within a region, TDMs provide comprehensive regional VMT estimates and avoid the uncertainties
associated with extrapolation of traffic volumes from count data at selected locations. They provide less
detail, however, regarding volume fluctuation by time of day, vehicle type, and speeds than can be
obtained from measurements, except to the extent that available data are used to provide such detail in
model output. Agencies with access to TDMs can readily obtain VMT distributions from the link-level
traffic volumes and other outputs of these models. The November 1993 transportation conformity
regulations require network-based modeling for metropolitan planning areas in ozone and carbon
monoxide nonattainment areas classified as serious and higher (40CFR51.452). Modeling improvements
have been made in many areas in response to these regulations, as well as other planning needs.
4.1 Processing TDM Outputs
TDMs produce information for thousands of individual links, depending on the size of the network
being modeled. Several software systems, such as statistical software or inventory software, exist that will
automatically provide summaries similar to those that would be needed to estimate regional VMT
distributions. These software systems process TDM data that in some cases may be pre-processed to
reformat datasets or add specific parameters, such as capacity, to the data.
4.2 Speed Estimation Procedures Using Travel Demand Model Data
The assignment of traffic to the roadway network in travel demand models uses calculated speeds
and route choices to minimize travel time. The effect of speeds on assignments is evaluated primarily in
terms of how well the assigned traffic volumes agree with count data. Historically, lack of agreement
between TDM speeds and observed speeds was of little concern provided that congestion and traffic
volumes were well characterized. TDM inputs (especially the "trip tables" identifying numbers of trips
between each pair of zones) are prepared for specific time periods, and simulation results provide a single
assignment representative of that period. For average daily travel (ADT) assignments, it is not possible
within the model to describe hourly variation in congestion and speeds. Even if modeling is conducted
separately for different times of day (e.g., AM peak, midday, PM peak, and overnight), congestion and
speeds can vary within each period. Consultation with transportation planners and modelers should
include assessment of whether TDM speeds have been calibrated or evaluated against observations. Post-
processing techniques are available that use HCM procedures and the BPR curve to calculate hourly
congested speeds. The general algorithm is:
1. Distribute link-level volumes by hour of day using user input temporal distributions (usually from
count data sets);
2. Calculate hourly VMT by multiplying link distance by hourly volume;
3. Calculate v/c using either link-specific capacities or lookup tables;
4. Apply the BPR curve, using link-specific free flow speeds or lookup tables, to arrive at hourly
congested speeds.
Exogenous volume adjustments can be applied to the loaded networks to account for variations by
day of week or season prior to post-processing speeds. Note that this results in higher inaccuracies in the
assignments, since ideally traffic should be assigned with the actual trip productions and attractions that
Final Report - September 1998 983210r.wpd
-------
18 SYSTEMSAPPLICA TIONSINTERNA TIONAL
correspond to the modeling episode. These types of data are unlikely to be available. However, future year
loaded networks are usually available from local planners which simplifies the development of future year
VMT distributions. The procedure described above is simply repeated, this time with the future year
assignment.
There are several areas in which TDMs may fail to provide comprehensive VMT estimates. These
relate to both the preparation of inputs used in modeling and in the level of detail incorporated in trip and
network inputs. For example, "intrazonal" travel (trips whose origin and destination are within the same
zone of the TDM) and other travel on local roads17 are not directly assigned to the network, and must be
separately addressed. This is typically accomplished through calculations based on assumptions about
intrazonal trip lengths, the sizes of zones, and local roadway speeds. Alternatively, local road travel can be
estimated from count data. Like local roads that are not "coded" into the TDM network, separate freeway
on- and off-ramp links may not be included in network specification. Ramp travel may therefore be either
omitted from TDM VMT estimates, or included as a portion of freeway VMT.
Information on travel by vehicle class is typically not available directly in TDMs. The "trip table"
inputs that identify the number of trips for each purpose (e.g., home-based work trips) between each pair of
spatially defined zones in the model, and this information can be used if data exist on fleet composition for
different trip purposes. However, as TDMs focus primarily on travel by individuals rather than goods
movement, this approach provides little value for identifying medium and heavy truck activity. Goods
movement models are under development, but at present, simple adjustment factors are more commonly
used to estimate incremental freight-related VMT to be added to modeled volumes. Time of day, day of
week, and seasonal variation of freight travel should be evaluated separately, based on local data.
17 TDM networks typically include all freeways and arterials, but may not explicitly include minor streets.
Final Report - September 1998 983210r.wpd
-------
SYSTEMS APPLICATIONS INTERNATIONAL 19
5.0 APPLICATION OF METHODOLOGY TO DATA FROM FIVE CITIES
These two methodologies for developing distributions of VMT by facility type and average speed
are tested in this section using representative datasets for five urban areas. The urban areas selected for
this analysis were:
• Chicago, Illinois;
• Houston, Texas;.
• Charlotte, North Carolina;
• Ada County, Idaho (Boise region); and
• New York, New York.
5.1 Data Collected
Two types of data were collected for this exercise. For some cities, actual hourly or daily traffic
count data were obtained. These data generally included some information about the count locations, such
as functional class and/or area type, as well as temporal information for each count (e.g., time of day, day
of year). For other cities, transportation activity estimates and network information from the regional
transportation model were obtained. The choice of method to be tested in each city was dependent upon
the type of data obtained. Both types of datasets were not obtained for one city, therefore it was not
possible to perform a side-by-side comparison of methods. Tables 3a through 3f present a summary of the
traffic count and transportation model output data that were gathered for each city.
Each of the cities studied used its own set of functional classes, which did not directly correspond
to the four MOBILE6 classes. For example, although estimates of VMT on ramps existed, none of the five
areas had either count data or model link VMT for ramps, and Houston used a total of nine distinct arterial
and collector classes. In applying the speed estimation methodologies to each city, functional classes were
grouped, and in some cases, freeflow speeds and capacities were assigned based on a combination of local
recommendations and judgment. Speed estimation procedures were carried out for all functional classes
possible for purposes of demonstration. As noted in the presentation of results for these cities, actual
space-mean speeds for the different functional classes in a city should be determined so that appropriate
assignment of local functional classes to those of MOBILE6 are made.
As noted, none of the datasets obtained for this project contained direct count or volume estimates
for freeway ramps. A methodology for developing ramp VMT estimates, developed by the Charlotte DOT,
was provided with the Charlotte dataset. This methodology was used to develop rough estimates of ramp
VMT for all five cities as discussed in this section of the guidance.
Final Report - September 1998 983210r.wpd
-------
20
SYSTEMSAPPLICA TIONSINTERNA TIONAL
Table 3a. Charlotte Department of Transportation (CDOT) counts for 1995.
Item
Card Number
Location
LNKNM
CODE
DATE
DAYWEEK
TIME
TWOWAY
COMPASS DIR
AADT
AAWT
SUMCNT
AAWT factor
Card Number
VolOOOO-0015
VolOO 16-0030
Voll 145-1200
Description
= 1 (card 1)
Description of count site
Link number, as used in the transportation model (not
available for all links)
Internal CDOT code
mmddyy
0=Sunday, l=Monday, ...
Time counter placed (counts stored 001 - 2400)
Two-way count? T/F
Direction (NEWS) of A direction counts
Average-annual daily traffic
Average-annual weekday traffic
Sum of counts
Month/Day factor to calculate AAWT
= 2 (card 2)
Volume 0000-0015 - compass direction A
Volume 00 16-0030
Volume 1 145-1200 - compass direction A (total of 48
15-min counts)
Repeat for 1201 - 2400 - compass direction A (total of 48 15-min counts)
Repeat for 001 - 1200 - compass directionB (total of 48 15-min counts)
Repeat for 1201 - 2400 - compass directionB (total of 48 15-min counts)
Final Report - September 1998
983210r.wpd
-------
SYSTEMSAPPLICA TIONSINTERNA TIONAL
21
Table 3b. Charlotte DOT and North Carolina DOT count data (CDOT counts are AAWT,
NCDOT counts are ADT).
Item
FUNCL
AREATP
DOTF
LINKLEN
LNKNM
LOCATION
85VOL
86VOL
87VOL
88VOL
89VOL
90VOL
91 VOL
92VOL
93VOL
94VOL
95VOL
Description
CDOT functional class
CDOT area type (1 = CBD, 2 = CBD fringe, 3 = residential, 4 = commercial area,
5 = rural)
FHWA functional class (First letter: I = interstate, F = oth. Freeway, P = principal
arterial, M = minor arterial, C = collector, L = local; Second letter: U = urban, R =
rural)
Link length (miles)
Travel demand model ID (matches data set above as well)
From demand model - different method than count dataset
Volume - 1985
Volume - 1986
Volume - 1987
Volume - 1988
Volume - 1989
Volume - 1990
Volume - 1991
Volume - 1992
Volume - 1993
Volume - 1994
Volume - 1995
Final Report - September 1998
983210r.wpd
-------
22
SYSTEMSAPPLICA TIONSINTERNA TIONAL
Table 3c. New York traffic count dataset (Data are organized into four subsets, designated
TF1 through TF4).
TF1 Records
Region
County
Route
Milepoint
Station
Card Code ("1")
Section Length
Year
Functional Class
Factor Group
Description
Not Used
Not Used
Not Used
Not Used
Not Used
Not Used
Not Used
TF2 Records
Region
County
Route
Milepoint
Station
Card Code ("2")
Year
Month
Day of Month
Direction
Day of Week
Week of Year
Reference Marker
Hour Counts (4
positions)
Not Used
Not Used
Not Used
Not Used
TF3 Records
Region
County
Route
Milepoint
Station
Card Code ("3")
Year
Month of First Days Count
Day-of-Month of First
Days Count
Direction
Factor Group
Reference Marker
Hourly Counts (24 Times)
Not Used
Not Used
Not Used
Not Used
Not Used
TF4 Records
Region
County
Route
Milepoint
Station
Year
Count Number
Section Length
Beginning Description
Ending Description
AADT
Design Hour
Reference Marker
Direction
Bridge Identification
Number
Functional Class
Factor Group
HPMS Number
Table 3d. Chicago 1996 transportation model outputs.
Description
Anode - link end point
Bnode - link end point
Link distance
Functional class (0 = dummy links; 1 = freeway; 2 = major highway; 3 = area service (arterial); 4
other principal arterials; 5 = minor arterial (urban); 6 = collector (urban); 7 = local; 8 = major
collector (rural); 9 = minor collector)
Daily link volume
Link capacity
Link freeflow speed
Anode coordinates
Bnode coordinates
Final Report - September 1998
983210r.wpd
-------
SYSTEMSAPPLICA TIONSINTERNA TIONAL 23
Table 3e. Ada County 1995 and 2015 transportation model outputs.
Description
Anode - link end point
Bnode - link end point
Link distance
Functional class (1 = freeway; 2 = arterial; 3 = collector; 4 = local)
Daily link volume
Link freeflow speed
Anode coordinates
Bnode coordinates
Table 3f. Houston 2020 transportation model outputs.
Description
Anode - link end point
Bnode - link end point
Link distance
Functional class (0 = locals; 1 = radial freeways w/o frontage road; 2 = radial freeways w/ frontage
road; 3 = crc freeways w/o frontage road; 4 = crc freeways w/ frontage road; 5 = radial tollways w/o
frontage road; 6 = radial tollways w/ frontage road; 7 = crc tollways w/o frontage road; 8 = crc
tollways w/ frontage road; 9 = principal arterial w/ grade separator; 10 = principal arterials divided;
11 = principal arterials undivided; 12 = other arterials divided; 13 = other arterials undivided; 14 =
one-way pairs; 15 = one-way facilities; 16 = collectors divided; 17 = collectors undivided; 18 =
ferries; 19 = saturated arterials; 20 = transitways; 21 = saturated arterials)
Peak and Offpeak period link volumes
Link capacity
Link freeflow speed
Anode coordinates
BNODE COORDINATES
5.2 Application of Travel Demand Model Output Methodology
The first method used to determine vehicle miles traveled (VMT) distributions by functional class
and speeds that is presented uses transportation model outputs. This method was applied for Ada County,
Houston, and Chicago.
5.2.1 Use of Inventory Software
This method was carried out by applying a standard software package, the Direct Travel Impact
Model (DTIM2), which is available from the California Department of Transportation.18 In some
instances, supplemental preprocessor programs were created in FORTRAN to reformat datasets or to add
specific parameters, such as capacity, to the data. The DTIM2 model includes a speed processor that uses
18 Fieber et al. 1994. "DTIM2 User's Guide." Final Report SYSAPP94-94.051, Prepared for the California
Department of Transportation, Systems Applications International, Inc., San Rafael, California, June 30, 1994.
Final Report - September 1998 983210r.wpd
-------
24 SYSTEMSAPPLICA TIONSINTERNA TIONAL
hourly V/C ratios, freeflow speeds, and the standard Bureau of Public Roads (BPR) speed curve to arrive
at hourly link-level speeds.19 Although this speed processor can modify speeds to reflect the impacts of
signalization and queuing, these two functions were turned off for this analysis. Signalization was partially
accounted for by using different parameters in the BPR equations as outlined in Section 3.4.2.
The DTIM speed processor requires that the functional class and capacity be explicitly coded for
each link in the network. These parameters were available for both Houston and Chicago. Therefore, the
DTIM speed processor was used for both cities. For Ada County, functional class was available but
capacity was not available. Therefore, the predicted speeds from the local transportation model, which are
provided by link, were used in developing the VMT distributions for Ada County.
5.2.2 Ada County. Idaho Transportation Model Outputs
The transportation model estimates of traffic volumes and network characteristics that were
obtained for the Ada County region were processed using the DTIM software system. Current (1995)
predictions of vehicle activity were analyzed. Initially, distributions were developed for an entire day of
travel. Subsequently, distributions were developed separately for the AM-, PM-, and Off-Peak travel
periods. Distributions for the AM period are presented here. The Ada County network assigns roadway
links to one of the following four functional classes: 1) freeway; 2) arterial; 3) collector; and 4) local.
In many ways, this was the least detailed database used in this analysis. For example, data for
developing hourly distributions of VMT were not available. As a result, the Ada Planning Association had
used distributions developed by the California Department of Transportation for San Luis Obispo,
California,20 which has a similar population to Boise, Idaho. Extrapolating data from another city is
strongly discouraged for SIP purposes. Each area should start to collect their own data if a program is not
already in place. Individual link capacities were also not available. Therefore, the DTIM speed processor
was not applied for this region. Rather, the link-level speeds provided by the local transportation model
were assumed to be applicable regardless of time-of-day or congestion. As a result, the speed distribution
for each hour of the day was identical, regardless of hourly variation in traffic volume (and congestion).
Table 4a shows the 8 a.m. (morning peak) distribution of VMT by speed and functional class
from the 1995 Ada County transportation files. Table 4b provides similar information expressed as
fraction of total miles. Here, local roadway speed estimates fall mostly in the 15 mph speed bin, with all
other speeds higher. This may indicate either classification of roads as local that would more appropriately
be classed as arterial/collector, or possible errors in assigned freeflow speeds. Users should remain aware
in their development of facility-class-specific VMT distributions that MOBILE6 is not speed-dependent
for local roadways and ramps. Speed distributions are shown here for local roadways based on available
data, and for purposes of demonstrating the results of the BPR speed methodology. Agencies are
encouraged to test and evaluate the VMT and speed estimation procedures presented here, and to critically
assess their results and revise, if appropriate, the facility type classifications or other inputs to the speed
estimation procedure. Consultation with local transportation planning agencies can be extremely valuable,
as many MPOs have made significant investments in developing reliable speed estimates.
Calculating link- and hour-specific speeds with the speed processor could not be done due to the
absence of hourly congestion (v/c) inputs in the 1995 Ada County transportation files. Thus, the speed
19 Bowling, 1994. "Technical Memorandum 8-1 User's Guide for Speed Processor." Prepared for the California
Department of Transportation by Richard Bowling, Bowling Associates, June 14, 1994.
20 Caltrans, 1993. "1991 Statewide Travel Survey." Prepared by the California Bepartment of Transportation, Office
of Traffic Improvement. Becember 1993.
Final Report - September 1998 983210r.wpd
-------
SYSTEMSAPPLICA TIONSINTERNA TIONAL
25
distribution shown in Table 4b applies to all hours of the day. Table 4c summarizes the distribution of
daily VMT between functional class, after merging Ada County's arterial and collector classes to coincide
with those of MOBILE6. Also, because the Ada County model does not track ramp VMT, assumptions
based on analyses by the Charlotte Department of Transportation (CDOT) were used to derive ramp VMT
as a fraction of freeway VMT.21 These fractions are summarized in Table 4d. The ramp VMT varies
greatly among land use type, and for this study, the fraction for commercial land uses was selected as most
applicable on a regional basis. Thus, ramp VMT was estimated as 8.7% of freeway VMT.
Table 4a. Distribution of 1995 hourly VMT by functional class and speed during the
morning peak (8 a.m.) for Ada County.
Speed Range
0.0- 2.5
2.5 - 7.5
7.5-12.5
12.5 - 17.5
17.5-22.5
22.5-27.5
27.5-32.5
32.5-37.5
37.5-42.5
42.5-47.5
47.5-52.5
52.5-57.5
57.5-62.5
62.5-67.5
67.5-72.5
Total
Vehicle Miles
Freeway
0
0
0
147
230
2,318
468
0
0
7,407
42,903
14,612
15,574
0
0
83,659
Arterial
0
0
0
5
1,669
7,720
56,278
67,940
15,866
20,578
0
0
0
0
0
170,056
Collector
0
0
0
0
6,705
6,135
7,241
3,214
3,381
0
0
0
0
0
0
26,676
Local
0
0
0
26,989
2,593
441
493
912
513
0
0
0
0
0
0
31,942
Total
0
0
0
27,142
11,197
16,614
64,481
72,067
19,760
27,985
42,903
14,612
15,574
0
0
312,332
21 CDOT (1997). "Charlotte Travel Demand Models, Vehicle Miles Traveled." Draft Report prepared by the
Charlotte Department of Transportation, Charlotte, NC. July 1997.
Final Report - September 1998
983210r.wpd
-------
26
SYSTEMSAPPLICA TIONSINTERNA TIONAL
Table 4b. 1995 VMT distribution by functional class and speed for Ada County.
Speed Range
0.0- 2.5
2.5 - 7.5
7.5 - 12.5
12.5 - 17.5
17.5-22.5
22.5-27.5
27.5-32.5
32.5-37.5
37.5-42.5
42.5-47.5
47.5-52.5
52.5-57.5
57.5-62.5
62.5-67.5
67.5-72.5
Total
Fraction of Total VMT
Freeway
0
0
0
0.0005
0.0007
0.0074
0.0015
0
0
0.0237
0.1374
0.0468
0.0499
0
0
0.2679
Arterial
0
0
0
0
0.0053
0.0247
0.1802
0.2175
0.0508
0.0659
0
0
0
0
0
0.5445
Collector
0
0
0
0
0.0215
0.0196
0.0232
0.0103
0.0108
0
0
0
0
0
0
0.0854
Local
0
0
0
0.0864
0.0083
0.0014
0.0016
0.0029
0.0016
0
0
0
0
0
0
0.1022
Total
0
0
0
0.0869
0.0359
0.0532
0.2064
0.2307
0.0633
0.0896
0.1374
0.0468
0.0499
0
0
1.0000
Table 4c. Overall VMT distributions for Ada County in 1995
(miles/day).
Functional Class
Freeway
Arterial & Collector
Ramp
Local
1995 VMT
1,486,240
3,495,042
129,303
567,470
1995 VMT
Fraction
0.26
0.62
0.02
0.10
Table 4d. Fraction of freeway VMT used to calculate ramp VMT by
CDOT22.
Area Type
Central Business District
Commercial
Residential
Fraction
0.194
0.087
0.024
5.2.3 Chicago Transportation Model Data Outputs
22 See Id.
Final Report - September 1998
983210r.wpd
-------
SYSTEMSAPPLICA TIONSINTERNA TIONAL
27
Transportation model data for 1996 traffic volumes and network characteristics were processed
using the DTIM system to arrive at VMT distributions by functional class and speed for the Chicago
region. The DTIM speed processor was used to calculate speeds as a function of hourly volume to capacity
and freeflow speeds. Tables 5a and 5b summarize total VMT by functional class and speed for light and
heavy duty traffic, respectively, during the AM peak travel period. Separate results in this form are
obtained from DTIM for each of 24 one-hour periods. The estimates for heavy duty travel activity are
arrived at by applying an overall fraction of VMT assumed to be attributable to heavy duty vehicles to link
volumes, producing heavy duty vehicle volumes by link. This fraction does not vary by area type or
functional class. However, a separate temporal distribution is then applied to the heavy duty portion of the
fleet, which results in different VMT distributions for this portion of the fleet. Tables 5c and 5d summarize
fractional VMT distributions by functional class and speed for AM peak travel. The values shown in these
two tables are fractions of hourly total VMT for all functional classes.
Tables 5e through 5g summarize VMT distributions by MOBILE6 functional classes for light
duty, heavy duty, and total fleet, respectively. Ramps are not tracked as a separate functional class within
the Chicago model, so based on Charlotte NC data, it was assumed that ramp VMT was equal to 8.7% of
freeway VMT. For Chicago, the freeway and highway facility classes were combined, as were the arterial
and collector classes. For this document, no attempt was made to segregate roadways identified as "local"
for which freeflow speeds and congestion allowed speeds significantly higher than the 12.9 mph average
for the local roadway speed cycle. It is likely that many of the roadways classified as local would be better
classified as arterial/collector for this data set.
Table 5a. Summary of 1996 light duty vehicle VMT distribution by functional class and speed
during AM peak (8 a.m.) for Chicago (miles/day).
Speed Range
0.0- 2.5
2.5 - 7.5
7.5 - 12.5
12.5 - 17.5
17.5-22.5
22.5-27.5
27.5-32.5
32.5-37.5
37.5-42.5
42.5-47.5
47.5-52.5
52.5-57.5
57.5-62.5
62.5-67.5
67.5-72.5
Total
Vehicle Miles
Freeway
0
3,362
0
105,660
182,753
181,568
156,724
251,344
198,653
133,224
517,441
309,012
107,232
135,870
0
2,282,844
Highway
0
856
17,769
66,463
201,406
327,280
348,149
240,993
160,016
117,340
57,882
18,407
0
0
0
1,556,560
Arterial
0
4,970
29,132
137,749
530,801
929,526
804,607
538,417
222,657
116,133
22,251
1,153
0
0
0
3,337,395
Collector
0
7,607
25,812
123,324
340,752
409,209
224,273
161,452
152,032
101,917
35,996
881
0
0
0
1,583,256
Local
0
6,416
3,652
10,492
51,658
115,174
72,144
44,870
102,912
61,213
34,334
1,131
0
0
0
503,996
Total
0
23,211
76,365
443,688
1,307,370
1,962757
1,605,897
1,237,076
836,270
529,827
667,904
330,584
107,232
135,870
0
9,264,051
Final Report - September 1998
983210r.wpd
-------
28
SYSTEMSAPPLICA TIONSINTERNA TIONAL
Table 5b. Summary of 1996 heavy duty vehicle VMT distribution by functional class and speed
during AM peak (8 a.m.) for Chicago (miles/day).
Speed Range
0.0- 2.5
2.5- 7.5
7.5-12.5
12.5-17.5
17.5-22.5
22.5-27.5
27.5-32.5
32.5-37.5
37.5-42.5
42.5-47.5
47.5-52.5
52.5-57.5
57.5-62.5
62.5-67.5
67.5-72.5
Total
Vehicle Miles
Freeway
0
981
0
23,528
37,181
35,764
31,576
51,636
42,259
31,789
143,395
100,161
35,520
64,592
0
598,382
Highway
0
86
1,914
8,074
21,387
36,103
36,346
23,862
16,994
12,817
6,516
1,978
0
0
0
166,076
Arterial
0
403
2,658
12,554
51,742
85,488
70,946
47,134
20,102
10,093
3,156
126
0
0
0
304,401
Collector
0
605
2,150
9,896
28,838
33,821
17,884
12,242
11,981
9,361
3,269
59
0
0
0
130,104
Local
0
588
290
771
4,249
8,755
5,243
2,924
8,067
5,895
3,968
45
0
0
0
40,793
Total
0
2,663
7,012
54,823
143,397
199,931
161,995
137,798
99,403
69,955
160,304
102,369
35,520
64,592
0
1,239,756
Final Report - September 1998
983210r.wpd
-------
SYSTEMSAPPLICA TIONSINTERNA TIONAL
29
Table 5c. Summary of 1996 light duty VMT fraction by functional class and speed during AM
peak (8 a.m.) for Chicago.
Speed Range
0.0- 2.5
2.5 - 7.5
7.5 - 12.5
12.5 - 17.5
17.5-22.5
22.5-27.5
27.5-32.5
32.5-37.5
37.5-42.5
42.5-47.5
47.5-52.5
52.5-57.5
57.5-62.5
62.5-67.5
67.5-72.5
Total
Fraction of Total VMT
Freeway
0
0.0004
0
0.0114
0.0197
0.0196
0.0169
0.0271
0.0214
0.0144
0.0559
0.0334
0.0116
0.0147
0
0.2464
Highway
0
0.0001
0.0019
0.0072
0.0217
0.0353
0.0376
0.0260
0.0173
0.0127
0.0062
0.0020
0
0
0
0.1680
Arterial
0
0.0005
0.0031
0.0149
0.0573
0.1003
0.0869
0.0581
0.0240
0.0125
0.0024
0.0001
0
0
0
0.3603
Collector
0
0.0008
0.0028
0.0133
0.0368
0.0442
0.0242
0.0174
0.0164
0.0110
0.0039
0.0001
0
0
0
0.1709
Local
0
0.0007
0.0004
0.0011
0.0056
0.0124
0.0078
0.0048
0.0111
0.0066
0.0037
0.0001
0
0
0
0.0544
Total
0
0.0025
0.0082
0.0479
0.1411
1.1149
0.1733
0.1335
0.0903
0.0572
0.0721
0.0357
0.0116
0.0147
0
1.0000
Final Report - September 1998
983210r.wpd
-------
30
SYSTEMSAPPLICA TIONSINTERNA TIONAL
Table 5d. Summary of 1996 heavy duty VMT fraction by functional class and speed during AM
peak (8 a.m.) for Chicago.
Speed Range
0.0- 2.5
2.5- 7.5
7.5-12.5
12.5-17.5
17.5-22.5
22.5-27.5
27.5-32.5
32.5-37.5
37.5-42.5
42.5-47.5
47.5-52.5
52.5-57.5
57.5-62.5
62.5-67.5
67.5-72.5
Total
Fraction of Total VMT
Freeway
0
0.0008
0
0.0190
0.0300
0.0288
0.0255
0.0417
0.0341
0.0256
0.1157
0.0808
0.0287
0.0521
0
0.4827
Highway
0
0.0001
0.0015
0.0065
0.0173
0.0291
0.0293
0.0192
0.0137
0.0103
0.0053
0.0016
0
0
0
0.1340
Arterial
0
0.0003
0.0021
0.0101
0.0417
0.0690
0.0572
0.0380
0.0162
0.0081
0.0025
0.0001
0
0
0
0.2455
Collector
0
0.0005
0.0017
0.0080
0.0233
0.0273
0.0144
0.0099
0.0097
0.0076
0.0026
0
0
0
0
0.1049
Local
0
0.0005
0.0002
0.0006
0.0034
0.0071
0.0042
0.0024
0.0065
0.0048
0.0032
0
0
0
0
0.0329
Total
0
0.0021
0.0057
0.0442
0.1157
0.1613
0.1307
0.1111
0.0802
0.0564
0.1293
0.0826
0.0287
0.0521
0
1.0000
Table 5e. Light Duty Vehicle VMT distributions for Chicago in 1996
(miles/day).
Functional Class
Freeway
Arterial & Collector
Ramp
Local
1996 LDV VMT
47,071,489
59,593,980
4,095,213
6,098,846
1996 LDV VMT
Fraction
0.40
0.51
0.04
0.05
Table 5f Heavy Duty Vehicle VMT distributions for Chicago in 1996
(miles/day).
Functional Class
Freeway
Arterial & Collector
Ramp
Local
1996 HDV VMT
9,928,019
5,642,914
863,738
529,783
1996 HDV VMT
Fraction
0.59
0.33
0.05
0.03
Final Report - September 1998
983210r.wpd
-------
SYSTEMSAPPLICA TIONSINTERNA TIONAL
31
Table 5g. Total VMT distributions for Chicago in 1996 (miles/day).
Functional Class
Freeway
Arterial & Collector
Ramp
Local
1996 TOTAL VMT
56,999,508
65,236,894
4,958,957
6,628,629
1996 TOTAL VMT
Fraction
0.43
0.49
0.04
0.05
5.2.4 Houston Transportation Model Data
Transportation model outputs for the Houston region were processed using DTIM to arrive at
VMT distributions by functional class and speed. The model's network uses a total of 21 different facility
type designations (e.g., it separately identifies one-way pairs, undivided and divided collectors, transitways,
etc.). These facility types were preserved in the speed analysis, and later grouped based on local
recommendations. Three speed estimation procedures were applied in this analysis. Initially, the model
was run using the speed processor. However, it was determined that the capacities and freeflow speeds
coded on the network were providing unreasonable results. New capacities and freeflow speeds were then
receded using standard assumptions based on data seen for several other cities, FHWA recommendations,
and engineering judgment. Table 6a summarizes these speed and capacity assumptions. This provided
somewhat better results. The model was then rerun using the link-specific speeds originally provided by
the Houston-Galveston Area Council (HGAC) transportation model. This also resulted in reasonable
speeds, but failed to capture any variation in speed by time of day.
The VMT distributions by functional class and speed as calculated by the speed processor (using
the assumed capacities and freeflow speeds in Table 6a) are summarized in Table 6b for the AM peak (7
a.m.) travel period. The fractional VMT distributions for this run are provided in Table 6c. The VMT
distributions by functional class and speed for the run using the speeds output by the HGAC transportation
model are summarized in Table 6d.
The fractional VMT distributions for this run are provided in Table 6e. In Table 6f, the overall
distribution of VMT by functional class has been remapped into the functional classes used in MOBILE6,
with expressways grouped into the freeway class. Again, ramp VMT is estimated as 8.7 percent of
freeway VMT due to the absence of separate ramp volume data.
Table 6a. Freeflow speeds (mph) and capacities for Houston
Speed
Capacity
Freeway
61
1750
Expressway
49
1660
Arterial
41
1400
Collector
41
1400
Local
25
1000023
23In transportation modeling, local roadways are least likely to be comprehensively represented in the model network
of links. A "centroid connector" link is a single artificial link representing local roads within each traffic analysis
zone. The length of this link is the assumed average distance traveled on local roads leading to collectors or
arterials. A high capacity is set for such links to prevent the TDM or the speed processor from assuming that large
traffic volumes (i.e., large numbers of trips originating in a zone) cause local roadway congestion.
Final Report - September 1998
983210r.wpd
-------
32
SYSTEMSAPPLICA TIONSINTERNA TIONAL
Table 6b. Summary of 2020 PM Peak ( 5 p.m.) VMT by functional class and speed for Houston using
speed processor (miles/day).
Speed Range
0.0- 2.5
2.5 - 7.5
7.5 - 12.5
12.5 - 17.5
17.5-22.5
22.5-27.5
27.5-32.5
32.5-37.5
37.5-42.5
42.5-47.5
47.5-52.5
52.5-57.5
57.5-62.5
62.5-67.5
67.5-72.5
Total
Vehicle Miles - PM Peak (5 p.m.)
Freeway
0
40,328
19,328
37,612
32,532
28,383
111,107
31,885
70,849
138,444
263,919
893,947
1,743,590
25,997
16,152
3,454,072
Expressway
0
1,945
0
0
0
0
0
173
0
0
4,565
14,715
0
0
0
21,398
Arterial
0
52,579
28,473
21,826
18,235
33,471
67,263
435,580
992,053
319,085
422,125
156,327
256,274
0
0
2,803,290
Collector
0
6,241
590
1,252
310
0
1,410
25,626
2,061
6,095
58,011
2,391
38,804
0
0
142,789
Local
0
0
0
0
0
866,111
0
1654
151,609
0
0
0
0
0
0
1,019,374
Total
0
101,094
48,391
60,690
51,077
927,965
179,780
494,917
1,216,572
463,623
748,619
1,067,380
2,038,668
25,997
16,152
7,440,921
Final Report - September 1998
983210r.wpd
-------
SYSTEMSAPPLICA TIONSINTERNA TIONAL
33
Table 6c. Summary of 2020 PM Peak (5 p.m.) VMT fractions by functional class and speed for
Houston using speed processor (miles/day).
Speed Range
0.0- 2.5
2.5 - 7.5
7.5 - 12.5
12.5 - 17.5
17.5-22.5
22.5-27.5
27.5-32.5
32.5-37.5
37.5-42.5
42.5-47.5
47.5-52.5
52.5-57.5
57.5-62.5
62.5-67.5
67.5-72.5
Total
Vehicle Miles - PM Peak (5 p.m.)
Freeway
0
0.0054
0.0026
0.0051
0.0044
0.0038
0.0149
0.0043
0.0095
0.0186
0.0355
0.1201
0.2343
0.0035
0.0022
0.4642
Expressway
0
0.0003
0
0
0
0
0
0
0
0
0.0006
0.0020
0
0
0
0.0029
Arterial
0
0.0071
0.0038
0.0029
0.0025
0.0045
0.0090
0.0585
0.1333
0.0429
0.0567
0.0210
0.0344
0
0
0.3767
Collector
0
0.0008
0.0001
0.0002
0
0
0.0002
0.0034
0.0003
0.0008
0.0078
0.0003
0.0052
0
0
0.0192
Local
0
0
0
0
0
0.1164
0
0.0002
0.0204
0
0
0
0
0
0
0.1370
Total
0
0.0136
0.0065
0.0082
0.0069
0.1247
0.0242
0.0665
0.1635
0.0623
0.1006
0.1434
0.2740
0.0035
0.0022
1.0000
Final Report - September 1998
983210r.wpd
-------
34
SYSTEMSAPPLICA TIONSINTERNA TIONAL
Table 6d. Summary of 2020 PM Peak (5 p.m.) VMT by functional class and speed for Houston using
HGAC transportation model speeds (miles/day).
Speed Range
0.0- 2.5
2.5- 7.5
7.5-12.5
12.5-17.5
17.5-22.5
22.5-27.5
27.5-32.5
32.5-37.5
37.5-42.5
42.5-47.5
47.5-52.5
52.5-57.5
57.5-62.5
62.5-67.5
67.5-72.5
Total
Vehicle Miles - PM Peak (5 p.m.)
Freeway
0
0
0
0
0
0
0
3,015
4,435
0
1,938
63,828
3,338,709
5,103
37,045
3,454,072
Expressway
0
0
0
0
0
0
0
173
0
0
6,510
14,715
0
0
0
21,398
Arterial
0
0
0
0
0
2,992
0
370,739
1,139,260
322,949
512,976
140,765
313,608
0
0
2,803,288
Collector
0
0
0
0
0
0
0
28,228
663
11
69,979
1,542
42,366
0
0
142,789
Local
0
0
0
0
0
866,111
0
0
153,263
0
0
0
0
0
0
1,019,374
Total
0
0
0
0
0
869,103
0
402,154
1,297,621
322,959
591,403
220,849
3,694,683
5,103
37,045
7,440,920
Final Report - September 1998
983210r.wpd
-------
SYSTEMSAPPLICA TIONSINTERNA TIONAL
35
Table 6e. Summary of 2020 PM Peak (5 p.m.) VMT fractions by functional class and speed for
Houston using HGAC transportation model speeds (miles/day).
Speed Range
0.0- 2.5
2.5- 7.5
7.5-12.5
12.5-17.5
17.5-22.5
22.5-27.5
27.5-32.5
32.5-37.5
37.5-42.5
42.5-47.5
47.5-52.5
52.5-57.5
57.5-62.5
62.5-67.5
67.5-72.5
Total
Fraction of Total VMT - PM Peak (5 p.m.)
Freeway
0
0
0
0
0
0
0
0.0004
0.0006
0
0.0003
0.0086
0.4487
0.0007
0.0050
0.4642
Expressway
0
0
0
0
0
0
0
0
0
0
0.0009
0.0020
0
0
0
0.0029
Arterial
0
0
0
0
0
0.0004
0
0.0498
0.1531
0.0434
0.0689
0.0189
0.0421
0
0
0.3767
Collector
0
0
0
0
0
0
0
0.0038
0.0001
0
0.0094
0.0002
0.0057
0
0
0.0192
Local
0
0
0
0
0
0.1164
0
0
0.0206
0
0
0
0
0
0
0.1370
Total
0
0
0
0
0
0.1168
0
0.0540
0.1744
0.0434
0.0795
0.0297
0.4965
0.0007
0.0050
1.0000
Table 6f Overall VMT distributions for Houston in 2020
(miles/day).
Functional Class
Freeway
Arterial & Collector
Ramp
Local
2020 VMT
54,357,162
50,073,142
4,729,073
16,366,908
VMT Fraction
0.43
0.40
0.04
0.13
5.3 Application of Traffic Count Data Methodology
The second methodology utilized traffic count data to directly estimate VMT distributions and
speeds. This methodology was applied to traffic count data from Charlotte, North Carolina and New York.
In many cases, FORTRAN programs were developed to manipulate the traffic count data to accomplish
this. Each example is presented below.
Final Report - September 1998
983210r.wpd
-------
36 SYSTEMSAPPLICA TIONSINTERNA TIONAL
5.3.1 Charlotte Traffic Count Data
The traffic count databases obtained for Charlotte, North Carolina were processed using simple
FORTRAN routines. Two sets of count data were available from CDOT, only one of which had functional
class identified. Therefore, the LNKNM variable (see Table 3b) was used to match the functional class for
each count location from one data set with count locations in the other data set. Further, only counts on
freeways, major arterials, and minor arterials were available. VMT for collectors, locals, and ramps are
estimated using procedures developed by CDOT24.
For collectors, VMT is estimated as a fraction of total VMT on major and minor arterials. The
fractions vary by area type and are drawn from the 1990 calibrated transportation model. These fractions
are:
CBD 0.5 percent
Commercial 4.9 percent
Residential/Rural 12.9 percent.
Local street VMT was estimated by CDOT using local GIS data and then distributed by area type
using the following assumptions for fraction total local VMT by area type:
CBD 1.6 percent
Commercial 12.9 percent
Residential/Rural 85.6 percent.
Total 1995 local street VMT was estimated by CDOT as 1,118,051 miles/day.
Ramp VMT was estimated as a fraction of freeway VMT, with area-specific fractions developed
by CDOT:
CBD 19.4 percent
Commercial 8.7 percent
Residential/Rural 2.4 percent.
The CDOT procedures for estimating collector, local, and ramp VMT were combined with the information
derived from the count data to arrive at overall VMT distributions.
Count data were allocated by hour of day (based upon the 15-minute counts supplied by CDOT)
and then used in the BPR formula to calculate speed by hour by count site. Table 7a presents the hourly
distributions of VMT for each functional class (i.e., the fraction of daily functional class VMT occurring in
each hour), and total across functional classes, as derived from the count data. As noted above, the count
sites did not include either local roads or ramps, so no temporal pattern is available for those facilities. In
the absence of such data, it would be necessary to assume temporal distributions for these functional
classes (e.g., using freeway distributions for ramps and minor arterials for locals). Table 7b summarizes
total link VMT calculated by functional class.
Table 7a. Hourly distributions of travel by functional class for Charlotte.
See CDOT (1997).
Final Report - September 1998 983210r.wpd
-------
SYSTEMSAPPLICA TIONSINTERNA TIONAL
37
Hour
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Total
Total
0.0080
0.0046
0.0035
0.0030
0.0040
0.0112
0.0352
0.0708
0.0765
0.0538
0.0492
0.0537
0.0612
0.0623
0.0618
0.0674
0.0754
0.0866
0.0665
0.0460
0.0359
0.0294
0.0203
0.0137
1.0
Freeway
0.0076
0.0057
0.0074
0.0066
0.0126
0.0365
0.0691
0.0688
0.0511
0.0460
0.0474
0.0530
0.0568
0.0571
0.0676
0.0717
0.0805
0.0658
0.0456
0.0368
0.0376
0.0306
0.0241
0.0141
1.0
Major
0.0080
0.0046
0.0035
0.0029
0.0038
0.0106
0.0342
0.0701
0.0772
0.0543
0.0496
0.0540
0.0616
0.0627
0.0618
0.0670
0.0750
0.0868
0.0668
0.0463
0.0359
0.0294
0.0202
0.0137
1.0
Minor
0.0084
0.0051
0.0037
0.0032
0.0052
0.0147
0.0426
0.0797
0.0713
0.0494
0.0455
0.0500
0.0562
0.0586
0.0611
0.0707
0.0794
0.0865
0.0653
0.0446
0.0353
0.0286
0.0208
0.0139
1.0
Table 7b. Distribution of link VMT by functional class for Charlotte.
Functional Class
Freeway
Major Arterial
Minor Arterial
Ramps
Collectors
Local
VMT (miles/day)
1,729,473
5,584,962
1,227,056
150,464
333,789
1,118,051
VMT Fractional Distribution
0.17
0.55
0.12
0.02
0.03
0.11
Final Report - September 1998
983210r.wpd
-------
38
SYSTEMSAPPLICA TIONSINTERNA TIONAL
The hourly count estimates were used to arrive at VMT distributions by speed by functional class.
This procedure is very sensitive to the assumptions made for the number of lanes, freeflow speed and
capacity for each count site. Different assumptions for these parameters were used for each area and
functional class combination. Assumptions for speeds and capacities are summarized in Tables 7c and 7d,
respectively. They are based on a combination of datasets that have been developed for other areas of the
country, FHWA assumptions in the Highway Capacity Manual, and engineering judgment. Freeflow
speeds were determined by using an assumed posted speed limit with the equations for space mean speed
as outlined in Section 3.4. It was assumed that freeways had three lanes in each direction, major arterials
had two lanes in each direction, and minor arterials had one lane in each direction. The data were also run
through the programs assuming that freeways had two lanes in each direction; this produced nearly
identical results to the first run. Table 7e summarizes the resulting VMT distributions for the AM-peak
hour of 8 a.m. for the three functional classes for which count data were available: freeways, major
arterials, and minor arterials. For use in MOBILE6, a VMT-weighted average of the major and minor
arterial speed distributions would be used for the arterial/collector facility type inputs.
Overall, the distributions appear to underestimate congestion and overestimate speeds for
freeways. The results for major arterials appear reasonable. The VMT distributions by speed for minor
arterials appear to overestimate congestion somewhat. The results suggest that this method should be
applied cautiously, and preferably local or link-level data on the number of lanes, freeflow speeds, and
capacities should be used. Table 7f summarizes the number of count sites in the Charlotte database for
each functional class. There were only four freeway sites, which is not enough data for deriving good
speed distributions. This probably explains the lack of congestion seen here.
Table 7c. Freeflow speeds for Charlotte, (mph)
Area Type
Freeway
Major
Minor
CBD
62
32
32
CBD fringe
62
48
40
Residential
62
40
32
Commercial
62
48
40
Rural
62
40
32
Table 7d. Capacities for Charlotte, (volume per hour)
Area Type
Freeway
Major
Minor
CBD
3,500
1,200
600
CBD fringe
3,500
1,600
550
Residential
3,500
1,600
550
Commercial
3,500
1,600
550
Rural
3,500
1,600
550
Final Report - September 1998
983210r.wpd
-------
SYSTEMSAPPLICA TIONSINTERNA TIONAL
39
Table 7e. Summary of AM Peak (8 a.m.) VMT distributions by
functional class and speed for Charlotte.
Speed
0.0- 2.5
2.5 - 7.5
7.5 - 12.5
12.5 - 17.5
17.5-22.5
22.5-27.5
27.5-32.5
32.5-37.5
37.5-42.5
42.5-47.5
47.5-52.5
52.5-57.5
57.5-62.5
62.5-67.5
67.5-72.5
72.5-77.5
Freeway
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
Major Arterial
0
0
0
0
0
0
0.1281
0
0.2838
0.0023
0.5858
0
0
0
0
0
Minor Arterial
0
0.0213
0
0
0.0213
0.0106
0.5319
0
0.4149
0
0
0
0
0
0
0
Table 7f Total number of count sites by functional class for Charlotte.
Counts
Freeway
4
Major
437
Minor
94
5.3.2 New York Traffic Count Data
The New York traffic count databases were processed using similar procedures to those described
for Charlotte. As with Charlotte, FORTRAN programs were used to develop VMT distributions by time of
day, functional class, and speed. Assumptions were made regarding the freeflow speeds and capacities for
each functional class, as these data were not available in the count database. Speed estimates proved to be
quite sensitive to these assumptions, and in practice, specific information regarding number of lanes,
freeflow speeds, and capacities for each count site should be obtained.
The hourly distributions of travel as calculated from count data are summarized in Table 8a for
urban count sites. Freeflow speed and capacity assumptions are summarized in Table 8b. The hourly
congested speed at each count site was calculated using the BPR equations, the hourly counts, and the
corresponding assumed freeflow speeds and capacities in Table 3b. This reliance on assumptions,
particularly with regard to capacity (number of lanes), introduces some risk of introducing bias into speed
calculations. For example, speed estimates for freeway count locations with more than three lanes would
be negatively biased, particularly during high volume periods. Table 8c provides total calculated VMT by
functional class, and the corresponding VMT fractions, on a daily basis. Table 8d summarizes speed
Final Report - September 1998
983210r.wpd
-------
40 SYSTEMSAPPLICA TIONSINTERNA TIONAL
distributions by functional class for urban count sites during the AM peak (8 a.m.) travel period. The total
number of counts by functional class for urban New York is presented in Table 8e.
The grouping of functional classes identified in the data base to correspond to the MOBILE6
facility types should consider the estimated speeds. Speeds are shown for all functional classes in Table 8d
to facilitate this evaluation. A notable feature of the speed distributions in this table is that for all
functional classes at or below principal arterial, more than 90 percent of VMT during the AM peak is
estimated to occur in the speed bin corresponding to the freeflow speed. This indicates that either counts
are quite low relative to capacities, or that the assumed capacities are overestimated. Moreover, the speed
estimates shown for local roadways are probably overstated due to an unreasonably high assumed speed. If
possible, with a count data set as rich as this in terms of number of locations and temporal distribution, it
would be well worth the effort to survey actual speeds at a representative number of count sites for each
functional class, as well as to obtain site-specific information regarding at least the number of lanes at each
count site. This would provide a sound empirical basis for assigning functional classes to the MOBILE6
categories, and would improve the quality of speed distribution estimates for the sites classified as either
freeway or arterial/collector.
Final Report - September 1998 983210r.wpd
-------
SYSTEMSAPPLICA TIONSINTERNA TIONAL
41
Table 8a. Hourly distributions of travel by functional class for urban New York.
Hour
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Interstate
0.0137
0.0093
0.0082
0.0086
0.0116
0.0251
0.0523
0.0742
0.0700
0.0540
0.0490
0.0495
0.0499
0.0505
0.0553
0.0634
0.0690
0.0687
0.0555
0.0445
0.0373
0.0327
0.0273
0.0205
Freeways &
Expressways
0.0152
0.0090
0.0064
0.0062
0.0097
0.0213
0.0487
0.0713
0.0700
0.0558
0.0483
0.0476
0.0481
0.0489
0.0543
0.0624
0.0677
0.0698
0.0597
0.0490
0.0408
0.0361
0.0304
0.0232
Principal
Arterial
0.0120
0.0071
0.0050
0.0043
0.0058
0.0141
0.0360
0.0589
0.0623
0.0542
0.0527
0.0565
0.0602
0.0593
0.0627
0.0687
0.0736
0.0739
0.0607
0.0500
0.0415
0.0347
0.0263
0.0194
Minor
Arterial
0.0086
0.0049
0.0035
0.0030
0.0046
0.0123
0.0335
0.0618
0.0647
0.0546
0.0530
0.0584
0.0636
0.0618
0.0660
0.0734
0.0782
0.0786
0.0610
0.0484
0.0386
0.0313
0.0216
0.0147
Collector
0.0087
0.0050
0.0034
0.0030
0.0041
0.0105
0.0307
0.0610
0.0653
0.0538
0.0530
0.0590
0.0651
0.0618
0.0650
0.0737
0.0795
0.0789
0.0622
0.0506
0.0397
0.0309
0.0207
0.0145
Local
0.0097
0.0051
0.0034
0.0030
0.0042
0.0118
0.0358
0.0666
0.0705
0.0539
0.0488
0.0524
0.0571
0.0554
0.0615
0.0722
0.0802
0.0821
0.0649
0.0503
0.0394
0.0319
0.0231
0.0166
Table 8b. Urban Freeflow speeds, capacities, and number of lanes for New York.
Functional Class
Interstate System
Other Freeways and Expressways
Other Principal Arterial
Minor Arterial
Collector
Local
Freeflow Speed
62
48
40
32
32
32
Capacity
5,250
2,400
1,600
600
800
550
# Lanes
3
3
2
1
1
1
Final Report - September 1998
983210r.wpd
-------
SYSTEMSAPPLICA TIONSINTERNA TIONAL
Table 8c. Distribution of link VMT by functional class for Urban New York.
Functional Class
Interstate
Other Freeways and Expressways
Other Principal Arterial
Minor Arterial
Collector
Local
Ramps
VMT
25,368,696
19,505,362
27,778,980
37,607,872
14,566,768
85,913,472
2,207,077
VMT Fractional
Distribution
0.12
0.09
0.13
0.18
0.07
0.41
0.01
Table 8d. Summary of AM Peak (8 a.m.) VMT distributions by functional class and speed for
urban New York.
Speed
0.0- 2.5
2.5- 7.5
7.5-12.5
12.5-17.5
17.5-22.5
22.5-27.5
27.5-32.5
32.5-37.5
37.5-42.5
42.5-47.5
47.5-52.5
52.5-57.5
57.5-62.5
62.5-67.5
67.5-72.5
72.5-77.5
Interstate
0
0
0
0
0.0192
0.0064
0.0128
0.0192
0.0064
0.0192
0.0321
0.0641
0.8205
0
0
0
Freeways &
Expressways
0.1116
0.0223
0.0134
0.0134
0.0045
0.0045
0.0045
0
0.0268
0.0536
0.7455
0
0
0
0
0
Principal
Arterial
0.0030
0
0.0010
0.0020
0
0.0010
0.0030
0.0050
0.9849
0
0
0
0
0
0
0
Minor
Arterial
0.0213
0.0085
0.0075
0.0085
0.0085
0.0043
0.9414
0
0
0
0
0
0
0
0
0
Collector
0.0029
0
0
0.0029
0
0
0.9942
0
0
0
0
0
0
0
0
0
Local
0.0250
0.0058
0.0038
0
0.0019
0.0019
0.9616
0
0
0
0
0
0
0
0
0
Table 8e. Total number of count sites by functional class for urban New York.
Counts
Interstate
127
Freeways &
Expressways
152
Principal
Arterial
521
Minor
Arterial
574
Collector
44
Local
19
Final Report - September 1998
983210r.wpd
-------
SYSTEMSAPPLICA TIONSINTERNA TIONAL
43
Glossary of Terms and Acronyms
25
Arterial
A class of street serving major traffic movement that is not designated as
a highway. There are principle and minor arterials which are designated
to primarily provide mobility and are a higher class than local or collector
streets which are designed to primarily provide access.
Assignment
Average Daily Traffic
The principal output of a travel demand model, which identifies the
number of vehicles assigned to each link of the highway network
The average number of vehicles passing a fixed point in a 24-hour time
frame. A convention for measuring traffic volume. Annual average daily
traffic (AADT) is determined by using a factor to adjust for the changing
amounts of traffic at different times of the year.
Capacity
Centerline Miles
The maximum number of vehicles that can pass over a given section of a
lane or roadway in one direction during a given time period under
prevailing roadway and traffic conditions.
A measure of the total length of roadways, ignoring the number of lanes
(Cf lane miles)
Central Business District
(CBD)
The most intensely commercial sector of a city. Often referred to as the
downtown.
Collector
A class of street serving neighborhood circulation, and providing a
balance between accessibility to land and through movement of traffic.
Counter
A machine that provides a count of traffic volume on a particular point of
the highway system. The detector is generally embedded into the
pavement in a relatively permanent installation.
Cordon Line
An abstract line encompassing a study area such as a central business
district, a shopping center or a larger planning area. Origin-destination
surveys and traffic counts are typically conducted along points on this
line to determine the characteristics of travel entering and leaving the
study area. It measures the transportation activity generated by the study
area. The line is usually associated with physical barriers, such as rivers,
or major highways with limited crossings.
Corridor
Broad geographical band connecting major sources of trips. Usually
associated with transportation facilities.
25The definitions presented here are in most cases taken from "Talking the Talk-A Pocket Guide to the Language of
Transportation Planning," prepared by the Northwestern Indiana Regional Planning Commission.
Final Report - September 1998
983210r.wpd
-------
44
SYSTEMSAPPLICA TIONSINTERNA TIONAL
Cutline
Expressway, Freeway
External Trip
Forecasting
Functional Classification
Fiighway Capacity Manual
(HCM)
Fiighway Performance
Monitoring system (HPMS)
An arbitrary line strategically drawn across the corridor of a
transportation network to connect all the paths in the corridor. Its
purpose is to check the larger scale comparability of the model with real-
world knowledge of the area.
A divided arterial highway for through traffic with controlled access, the
intersections on which are usually separated from other roadways by
differing grades. It can be a toll road.
A trip with one end inside a study area and the other end outside the study
area.
The processing of estimating the future values of specific variables used
in the transportation modeling process, including population, income and
employment.
The categorization of streets and roadways based on their intended use.
The classifications range from interstate, which is a controlled access
facility that serves through traffic movement and provides no access to
adjacent land, to the local street that primarily serves access to adjacent
land, and provides little movement of through traffic.
A guide for engineers and planners to estimate the capacity of the
elements of the highway stem, including freeways, ramps, arterial streets
and intersections, based on factors that cause the reduction of capacity,
such as parking, curves, topography and other similar factors.
The system used by the FHWA to provide information to Congress, the
States, and the public on the extent and physical condition of the nation's
highway system, its use, performance and needs. For clean air act
conformity analyses, the HPMS provides an official base estimate of
vehicle-miles of travel, which is used to adjust model-derived estimates of
vehicle-miles of travel for base and future years.
Lane Miles
A combined measure of the length and capacity of roadways (Cf.
centerline miles). In estimating VMT from count data, the units of
the count data (vehicles per hour, vehicles per hour per lane) are
known.
Level of Service (LOS)
A set of qualitative descriptions of a transportation system's performance.
The Highway Capacity Manual defines levels of service for intersections
and highway segments, with ratings that range from A (best) to F (worst).
Transportation projects are usually planned and designed to result in a
LOS of C or D, depending on the severity of the congestion problems,
and the ability to make improvements.
Final Report - September 1998
983210r.wpd
-------
SYSTEMSAPPLICA TIONSINTERNA TIONAL
45
Link
A representation of a road segment on transportation model networks.
One part of a chain of trips.
Local Street
Loop Detector
Metropolitan Planning
Organization (MPO)
Mode
Network
Node
A street intended solely for access to properties contiguous to it.
A vehicle detection device imbedded in pavement used to control traffic
signals and count vehicles. Speeds can be estimated from single loop
detectors based on the duration of single vehicle presence over the loop
(with an assumed average vehicle length). More accurate speed data can
be derived from two detectors closely spaced in the same lane.
The organizational entity designated by law with lead responsibility for
developing transportation plans and programs for urbanized areas of
50,000 or more in population. MPOs are established by agreement of the
Governor and units of general purpose local government.
The method used for personal travel or the movement of goods on a
particular trip. Modes include automobile, bus, commuter rail, bicycle,
walking, rail freight and trucking.
A system of links and nodes that represent highway segments and
intersections, and transit services, used in a transportation model to
estimate the utilization of the transportation system.
An element of a transportation model network that represents either an
intersection or the centroid of a traffic analysis zone.
Off-Peak
Peak Hours or Peak Period
Referring to the times and directions of travel not associated with the
major commuting direction; that is, all times and directions other than
toward the central business district or activity center(s) in the morning
and away from it in the late afternoon and early evening.
The period in the morning or evening in which the largest volume of
travel is experienced. Travel peaks are typically the result of trips to and
from work.
Screen Line
An imaginary line bisecting an area. Traffic counts are taken at regular
intervals at all streets intersecting the screen line. The line is associated,
where possible, with physical barriers, such as rivers, or major highways
with limited crossings. Counts taken along the screen line determine the
traffic moving between two areas. These counts are intended to detect
long-range changes in volume and direction of traffic due to significant
changes in land use and travel patterns.
Final Report - September 1998
983210r.wpd
-------
46
SYSTEMSAPPLICA TIONSINTERNA TIONAL
Space-mean Speed
Speed
Time-mean speed
Traffic Analysis Zone
Travel Demand model (TDM)
An average speed derived by dividing total distance traveled by total time
of travel. This definition of average speed is the one used in MOBILE6.
To derive a space-mean speed from VMT distributions across speeds, the
total travel time for the VMT in each speed bin is determined and
summed to divide into the total VMT across all speed bins. This is
equivalent to computing the harmonic mean speed (inverse of the average
of the inverses). (Cf time-mean speed)
Depending on the context, speed in transportation planning or traffic
engineering may refer to an instantaneous vehicle speed, an average
speed of vehicles at a specific location (e.g., from loop detectors), an
average speed for a single vehicle along a particular roadway segment, or
an average speed for all vehicles along a particular roadway segment.
The manner in which data are collected, or speed estimates are derived,
determines which meaning of the term applies (see space-mean speed and
time-mean speed).
An average speed derived as the arithmetic mean of instantaneous vehicle
speeds (e.g., from loop detector data). This definition of average speed
differs from that used in MOBILE6, and should be avoided in emissions
analysis (Cf. space-mean speed).
A subdivision of the metropolitan area used for transportation modeling.
The characteristics of the traffic analysis zone are used to estimate the
number of trips that start and end in the zone, for a base year, and for
specific forecast years.
A process to estimate the utilization of the transportation system under
various scenarios, using specific computer software, combined with
socioeconomic data, forecasts and the transportation system presented by
a network of links and nodes.
Vehicle Hours of Travel
(VHT)
Vehicle Miles of Travel
(VMT)
Volume to Capacity Ratio
(V/C)
A standard measure of total travel time, expressed in vehicle hours.
Regional average speed is determined by dividing VMT by VHT.
A standard areawide measure of travel activity. The most conventional
VMT calculation is to multiply average length of trip by the total number
of trips, or to sum the traffic volumes on links multiplied by link length.
A measure of the performance or utilization of a specific element of the
transportation system, such as a road segment or an intersection. The
capacity of the facility can be calculated using methods described in the
Highway Capacity Manual. The traffic volume is determined through the
traffic counting programs, and adjusted using factors to relate the data
collection date to the annual average for the data collection year. The v/c
is the percentage of the capacity that is being consumed by traffic. A v/c
ratio above 1.0 means that the volume of traffic exceeds capacity and the
road segment or intersection is becoming deficient and congested.
Final Report - September 1998
983210r.wpd
------- |