United States
Environmental Protection
Agency
EPA/600/R-07/096
July 2007
Development Work for Improved
Heavy-Duty Vehicle Modeling
Capability Data Mining - FHWA
Datasets
-------
EPA/600/R-07/096
July 2007
DEVELOPMENT WORK FOR IMPROVED HEAVY-DUTY VEHICLE MODELING
CAPABILITY DATA MINING - FHWA DATASETS
Prepared
By
Chris E. Lindhjem
Stella Shepard
ENVIRON International Corporation
101 Rowland Way, Suite 220
Novato, CA 94945-5010
EPA Contract No. 68-C-01-164
Work Assignment No. 4-9
EPA Project Officer
Sue Kimbrough
Air Pollution Prevention and Control Division
National Risk Management Research Laboratory
Research Triangle Park, NC 27711
National Risk Management Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, OH 45268
-------
This page left blank deliberately.
11
-------
ABSTRACT
Heavy-duty vehicles have been seen as contributing a large fraction of emissions from
on-road vehicles and are coming under more intense scrutiny because light-duty emissions have
been controlled to a greater extent than heavy-duty vehicle emissions. A heavy-duty vehicle can
produce 10 to 100 times the emissions (of NOX and PM emissions especially) of a light-duty ve-
hicle. Thus, heavy-duty vehicle activity needs to be better characterized. Key uncertainties with
the use of MOBILE6 regarding heavy-duty vehicle emissions include the fraction of heavy-duty
vehicles on all types of roadways at all times of day. In addition, there may be regional variabil-
ity in both the fraction of different vehicle classes and the vehicle weights within each class.
With the Motor Vehicle Emissions Simulator (MOVES) model, greater emphasis is given
to physical parameters affecting the engine loads and therefore the emissions from individual
vehicles. One primary factor affecting the engine load is the vehicle weight; the weight of the ve-
hicle on the road is needed to estimate its in-use emissions. Because the effect of vehicle weight
may be nonlinear for certain types of driving, it is important to incorporate the weight distribu-
tion of vehicles into emission estimates.
Databases collected by the Federal Highway Administration (FHWA) include vehicle
count and classification from the Highway Performance Monitoring System (HPMS) using au-
tomated traffic recorders (ATR) used to produce the Travel Volume Trends (TVT) reports. Other
data sets compile the results of data collection from weigh in motion (WIM) sensors, and other
data sources (visual observation, weigh stations, and other special projects) maintained by the
FHWA and compiled in the Vehicle Travel Information System (VTRIS). A discussion of these
data sources including original sources, representativeness, and quality and data reduction proce-
dures used in this work are provided in Appendix A.
This work consisted of an investigation and evaluation of these databases for the purpose
of assisting in the development of improved emissions estimates of heavy-duty vehicles. The
goal of the project was therefore to produce estimates of the fraction of heavy-duty vehicles of
all vehicle traffic, and weight distributions for those vehicles according to the time of day, day of
week, and other temporal variables, and an investigation of regional differences.
in
-------
FOREWORD
The U.S. Environmental Protection Agency (EPA) is charged by Congress with protect-
ing the Nation's land, air, and water resources. Under a mandate of national environmental laws,
the agency strives to formulate and implement actions leading to a compatible balance between
human activities and the ability of natural systems to support and nurture life. To meet this man-
date, EPA's research program is providing data and technical support for solving environmental
problems today and building a science knowledge base necessary to manage our ecological re-
sources wisely, understand how pollutants affect our health, and prevent or reduce environmental
risks in the future.
The National Risk Management Research Laboratory (NRMRL) is the agency's center
for investigation of technological and management approaches for preventing and reducing risks
from pollution that threaten human health and the environment. The focus of the laboratory's
research program is on methods and their cost-effectiveness for prevention and control of pol-
lution to air, land, water, and subsurface resources; protection of water quality in public water
systems; remediation of contaminated sites, sediments, and ground water; prevention and control
of indoor air pollution; and restoration of ecosystems. NRMRL collaborates with both public and
private sector partners to foster technologies that reduce the cost of compliance and to antici-
pate emerging problems. NRMRL's research provides solutions to environmental problems by:
developing and promoting technologies that protect and improve the environment; advancing
scientific and engineering information to support regulatory and policy decisions; and providing
the technical support and information transfer to ensure implementation of environmental regula-
tions and strategies at the national, state, and community levels.
This publication has been produced as part of the laboratory's strategic long-term re-
search plan. It is published and made available by EPA's Office of Research and Development to
assist the user community and to link researchers with their clients.
Sally Gutierrez, Director
National Risk Management Research Laboratory
IV
-------
EPA REVIEW NOTICE
This report has been peer and administratively reviewed by the U.S. Environmental Pro-
tection Agency and approved for publication. Mention of trade names or commercial products
does not constitute endorsement or recommendation for use. This document is available to the
public through the National Technical Information Service, Springfield, Virginia 22161.
-------
TABLE OF CONTENTS
ABSTRACT iii
FOREWORD iv
EPA REVIEW NOTICE v
TABLE OF FIGURES viii
TABLE OF TABLES xii
TABLE OF ACRONYMS xiv
1. INTRODUCTION 1
1.1 Average Vehicle Weight and Weight Distributions 3
1.2VMTMix 7
2. INVESTIGATION OF REGIONAL AGGREGATION 9
2.1 Average Weight by State 9
2.2 Vehicle Mix 13
3. REGIONAL VARIATIONS IN WEIGHT BIN DISTRIBUTIONS BY ROAD
TYPE FOR VEHICLE CLASSES 5-13 22
4. TEMPORAL VARIATIONS IN WEIGHT BIN DISTRIBUTIONS FOR
VEHICLE CLASSES 5-13 23
4.1 Monthly Variation 23
4.2 Hourly Variation Over an Average Week 25
5. ANALYZE WEIGHT DISTRIBUTIONS FOR VEHICLE CLASSES 1-4 35
5.1 Monthly Variation 35
5.2 Daily and Hourly Variation Over an Average Week 37
6. DEVELOP NATIONAL AVERAGE WEIGHT DISTRIBUTIONS BY
FHWA VEHICLE CLASSES 1-13 40
7. DEVELOP NATIONAL AVERAGE WEIGHT BIN DISTRIBUTIONS
BY FHWA VEHICLE CLASSES 1-13 42
8. VEHICLE CLASS FRACTIONS 46
8.1 Monthly Variation 46
8.2 Hourly Variation Over an Average Week 48
9. TRAFFIC VOLUME TRENDS 50
10. CONCLUSIONS AND RECOMMENDATIONS 52
REFERENCES 54
APPENDIX A - QUALITY ASSURANCE/QUALITY CONTROL
EVALUATION AND REVIEW A-l
VI
-------
A-l. PROJECT OBJECTIVES, ORGANIZATION, AND RESPONSIBILITIES A-2
A-l.l Purpose of Study A-2
A-1.2 Project Objectives A-3
A-1.3 Secondary Data Required by the Project A-4
A-1.4 Approach for Evaluating Project Objectives A-4
A-2. SOURCES OF SECONDARY DATA A-6
A-2.1 Sources of Secondary Data Used A-6
A-2.2 Rationale for Selecting Data Sources A-7
A-3. QUALITY OF SECONDARY DATA A-8
A-3.1 Quality Requirements of Secondary Data A-8
A-3.2 QAProcedures A-9
A-3.2.1 VTRISData A-9
A-3.2.2 Travel Volume Trends (TVT) Data A-12
A-3.3 Data Representativeness A-12
A-3.4 Data Sample Sizes A-19
A-4. DATA REPORTING, DATA REDUCTION, AND DATA VALIDATION A-22
A-4.1 Data Reduction Procedures A-22
A-4.2 Data Validation Procedures A-22
APPENDIX B - CROSS REFERENCE METHOD TO CONVERT FHWA
VEHICLE CLASSES TO MOBILE VEHICLE TYPES B-l
B-l. INTRODUCTION B-2
B-2. CROSS REFERENCE FROM FHWA TO MOBILE VEHICLE CLASSIFICATION...^-4
B-2.1 ALTERNATIVE METHODS FOR CROSS REFERENCE B-8
B-2.2 GEORGIA TECH INSTITUTE METHOD B-8
B-2.3TTI Method of VMT Mix B-9
APPENDIX C - NATIONAL AVERAGE TEMPORAL PROFILES FOR
FOUR ROAD TYPES C-l
C-l. INTRODUCTION C-2
APPENDIX D - ANALYSIS OF 2000 TRAVEL VOLUME TRENDS (TVT) DATA
AND 2000 VEHICLE TRAVEL INFORMATION SYSTEM (VTRIS) DATA D-1
D-l. INTRODUCTION D-2
D-2. DATA HANDLING PROCEDURES D-3
D-3. RESULTS OF REGIONAL VARIABILITY AND TVT/VTRIS NATIONAL AVERAGE
COMPARISONS D-9
vn
-------
TABLE OF FIGURES
Figure 1. Weight bin distribution by day of week for vehicle class 2 (passenger cars)
on road type 11 in 2000 5
Figure 2. Weight bin distribution by day of week for vehicle class 3 (7000-9000 Ib
two-axle, four-tire single-unit trucks) on road type 11 in 2000 5
Figure 3. Weight bin distribution by day of week for vehicle class 5 (12,000-14,000 Ib
six-tire, two-axle single-unit vehicles) on road type 11 in 2000 6
Figure 4. Weight bin distribution by day of week for vehicle class 6 and 7
(24,000-58,000 Ib three or more axle single-unit vehicles) on road type 11
in 2000 6
Figure 5. Weight bin distribution by day of week for vehicle type 8-13 (26,000-92,000
single- and multi-trailer vehicles) on road type 11 in 2000 7
Figure 6. Relative average weight by state for vehicles 8 - 13 in 1999 and 2000 on road
type 1, rural interstates. (Number of observations used for each state average
reported along axis. 1999 data is solid blue and 2000 is red on white hash.) 10
Figure 7. Relative average weight by state for vehicles 8 - 13 in 1999 and 2000 on road
type 11, urban interstates. (Number of observations used for each state average
reported along axis. 1999 data is solid blue and 2000 is red on white hash.) 10
Figure 8. Relative average weight by state for vehicle 5 in 1999 and 2000 on road type 1,
rural interstates. (Number of observations used for each state average reported
along axis. 1999 data is solid blue and 2000 is red on white hash.) 11
Figure 9. Relative average weight by state for vehicle 5 in 1999 and 2000 on road type 11,
urban interstates. (Number of observations used for each state average reported
along axis. 1999 data is solid blue and 2000 is red on white hash.) 12
Figure 10. State by state average weight for vehicle classes 8 - 13 on rural interstates in
1999 and 2000 12
Figure 11. Vehicle mix for weekdays on rural interstates for 1999 and 2000 data. (Number
of site-days of observations used in the state average reported along axis.) 13
Figure 12. Vehicle mix for weekdays on rural principle arterials for 1999 and 2000.
(Number of site-days of observations used in the state average reported along
axis.) 14
Vlll
-------
Figure 13. Vehicle mix for weekdays on urban interstates for 1999 and 2000. (Number
of site-days of observations used in the state average reported along axis.) 15
Figure 14. Regions denned for rural interstate vehicle mix 19
Figure 15. Average vehicle weight for vehicle classes 5 - 13 on urban interstates, road 11.
(Number of site-days of observations used in the state average reported along
axis.) 22
Figure 16. Average weight by month for lighter heavy vehicle classes in 1999 on road
type 1. (Uncertainty ranges were based on 90% confidence levels of the
sample.) 23
Figure 17. Average weight by month for heavier vehicle classes in 1999 on road type 1 24
Figure 18a. Average weight for vehicle 5 by hour over atypical week in 1999 25
Figure 18b. Average weight for vehicle 5 by hour over a typical week in 1999
(Without Indiana Data) 26
Figure 19. Average weight for vehicle group (8 - 13) by hour over a typical week in 1999. ..26
Figure 20. Day of week weight bin distribution for vehicle class 5 on road type 1 in 1999. ...27
Figure 2 la. Day of week weight bin distribution for vehicle class 5 on road type 11 in
1999 (including the Indiana data) 28
Figure 21b. Day of week weight bin distribution for vehicle class 5 on road type 11 in
1999 (excluding the Indiana data) 28
Figure 23. Fraction of class 5 vehicles on urban interstates in the 1999 Indiana
weight data 30
Figure 24. Day of week weight bin distribution for vehicle group (8-13) on road type
lin 1999 31
Figure 25. Day of week weight bin distribution for vehicle group (8-13) on road type
11 in 1999 31
Figure 26. Day of week weight bin distribution for vehicle group (8-13) on road
type 1 in 2000 32
Figure 27. Day of week weight bin distribution for vehicle group (8-13) on road
type 11 in 2000 32
Figure 28. Vehicle 5 weight bin distribution over an average Wednesday 33
Figure 29. Vehicle group (8 - 13) weight bin distribution over an average Wednesday 34
Figure 30. Average weight by month for lighter vehicle classes in 1999 on road type 1 36
Figure 31. Light vehicles weight over atypical week in 1999 37
Figure 32. Light vehicles weight over atypical week in 2000 38
Figure 33. Bus weight over atypical week in 1999 39
Figure 34. Bus weight bin distribution over atypical week in 1999 39
IX
-------
Figure 35. Vehicle 2 weight bin distribution in 1999 by functional class 42
Figure 36. Vehicle 3 weight bin distribution in 1999 by functional class 43
Figure 37. Vehicle 4 weight bin distribution in 1999 by functional class 43
Figure 38. Vehicle 5 weight bin distribution in 1999 by functional class 44
Figure 39. Vehicle 6 and 7 weight bin distribution in 1999 by functional class 44
Figure 40. Vehicle 8-13 weight bin distribution in 1999 by functional class 45
Figure 41. Month to month variability for road 1 in 1999 46
Figure 42. 1999 Hourly average-week vehicle fractions on road type 1, rural interstates 49
Figure 43. 1999 Hourly average-week vehicle fractions on road type 11, urban interstates. ..49
Figure 44. Typical national average weekly total traffic volume for rural roads 50
Figure 45. Typical national average weekly total traffic volume for urban roads 51
Figure B-1. Vehicle fractional mix over a week for urban interstates in Wisconsin B-6
Figure C-l. National average daily volumes by month with 90% confidence intervals C-4
Figure C-2. National average daily volumes by day of week with 90% confidence
intervals C-5
Figure C-3. National average hourly volumes with 90% confidence intervals C-5
Figure C-4. National average monthly vehicle counts per site for rural limited access
roads with 90% confidence intervals on vehicle types 2, 3, 5, and 9 C-7
Figure C-5. National average monthly vehicle fleet fractions for rural limited
access roads C-7
Figure C-6. National average monthly vehicle counts per site for rural other roads
with 90% confidence intervals on vehicle types 2, 3, 5, and 9 C-8
Figure C-7. National average monthly vehicle counts per site for urban limited access
roads with 90% confidence intervals on vehicle types 2, 3, 5, and 9 C-8
Figure C-8. National average monthly vehicle counts per site for urban other roads
with 90% confidence intervals on vehicle types 2, 3, 5, and 9 C-9
Figure C-9. National average day of week vehicle counts per site for rural limited
access roads with 90% confidence intervals on vehicle types 2, 3, 5, and 9 C-10
Figure C-10. National average day of week vehicle counts per site for rural other roads
with 90% confidence intervals on vehicle types 2, 3, 5, and 9 C-10
Figure C-ll. National average day of week vehicle counts per site for urban limited
access roads with 90% confidence intervals on vehicle types 2, 3, 5, and 9 C-ll
Figure C-12. National average day of week vehicle counts per site for urban other
roads with 90% confidence intervals on vehicle types 2, 3, 5, and 9 C-ll
Figure C-13. National average time of day vehicle counts per site for rural limited
access roads with 90% confidence intervals on vehicle types 2, 3, 5, and 9 C-12
-------
Figure C-14. National average time of day vehicle counts per site for rural other roads
with 90% confidence intervals on vehicle types 2, 3, 5, and 9 C-12
Figure C-15. National average time of day vehicle counts per site for urban limited
access roads with 90% confidence intervals on vehicle types 2, 3, 5, and 9 C-13
Figure C-16. National average time of day vehicle counts per site for urban other roads
with 90% confidence intervals on vehicle types 2, 3, 5, and 9 C-13
Figure D-1. Day of week variability in hourly volumes from Sunday to Saturday D-3
Figure D-4. National VTRIS monthly profile calculated from averages of state volumes D-7
Figure D-5. National VTRIS monthly profile calculated from averages of state profiles D-8
Figure D-6. State and national average TVT hourly profiles for rural interstates D-9
Figure D-7. National average TVT hourly profile plus or minus 1 standard deviation D-10
Figure D-8. National average TVT and VTRIS hourly temporal profile with uncertainty D-l 1
Figure D-9. State and national average TVT day of week profiles for rural interstates D-12
Figure D-10. National average TVT and VTRIS day of week temporal profile and
uncertainty on rural interstates D-13
Figure D-l 1. National average and state temporal profiles by day of week D-l 4
Figure D-12. Illustration of a different day of week temporal profile for Oklahoma on
rural principal arterials D-14
Figure D-13. Illustration of a different day of week temporal profile for Oklahoma for
rural minor arterials D-l5
Figure D-16. State and nationa average TVT monthly profiles for rural interstates D-17
Figure D-17. National average TVT and VTRIS monthly temporal profile
and uncertainty D-l8
XI
-------
TABLE OF TABLES
Table 1. FHWAroadway functional classification (types) in VTRIS 2
Table 2. FHWA vehicle classifications 2
Table 3. Vehicle identifiers and typical average vehicle weight range 3
Table 4. Vehicle weight bin descriptions 4
Table 5. Availability of VTRIS vehicle counts for weekday rural interstates and
suggested regional categories 16
Table 6. Vehicle class fractions for weekdays on road type 1, rural interstates 18
Table 7. Vehicle class fractions for weekdays on road type 2, rural principal arterials 20
TableS. Vehicle class fractions for weekdays on road type 11, urban interstates 21
Table 9. Number of observations for Figures 16 and 17 24
Table 10. Range in the number of observations for Figures 18 and 19 27
Table 11. Number of observations for weight distributions, vehicle class 5, road types 1
and 11 29
Table 12. Number of observations for weight distributions, vehicle classes 8-13, road
types 1 and 11 33
Table 13. Number of observations for Figure 30 36
Table 14. Average (of the monthly averages) and 90% confidence levels (±) using the
month-month variability of vehicle weight in 1999 40
Table 15. Average (of the monthly averages) and 90% Confidence Levels (±) using the
month-month variability of vehicle weight in 2000 41
Table 16. Monthly-average annual relative vehicle class counts and monthly uncertainly 47
Table A-l. TVT data coverage by state and important facility types A-14
Table A-2. VTRIS data coverage for vehicle mix by state and important facility types A-16
Table A-3. VTRIS data coverage for vehicle weight by state and important facility types A-17
Table A-4. Number of VTRIS weigh-in-motion observations by state A-19
Table A-5. Number of VTRIS vehicle classification observation days by site-direction in
each state A-20
Table A-6. Number of 2000 TVT observation days (site-directions counted separately) in
each state A-21
xn
-------
Table B-4. Default 2002 VMT mix by the MOBILE6 16 vehicle classes and crosswalk
calculation method from FHWA vehicle classes B-5
Table B-5. Converting MOBILE6 vehicle types to MOBILES vehicle types - diesel
fractions B-6
Table B-6. Raw average annual vehicle mix estimates B-7
Table B-9. EPA Vehicle Types - 28 Categories B-12
Table C-l. FHWA roadway functional classification (types) in TVT and VTRIS C-2
Table C-2. FHWA Vehicle classifications C-3
Table C-3. FHWA roadway functional classification (types) in TVT C-3
Table D-l. Number of States used in the calculation of the monthly temporal profiles D-2
Table D-2. Chi Square statistical tests for difference from national average for rural
interstates (Low p-values and large number of hours different than the mean
are statistical indicators of difference) D-ll
Xlll
-------
TABLE OF ACRONYMS
AM ante meridiem
ATR automated traffic recorders
EPA U. S. Environmental Protection Agency
FHWA Federal Highway Administration
FIPs Federal Information Processing Standards
HPMS Highway Performance Monitoring System
Ibs. pounds
MEASURE Mobile Emissions Assessment System for Urban and Regional Evaluation
MOBILE EPA's mobile source emission factor model
MOVES Motor Vehicle Emissions Simulator
NOX oxides of nitrogen
OMB Office of Management and Budget
PM particulate matter
PM post meridiem
QA quality assurance
QA/QC quality assurance/quality control
QAPP quality assurance project plan
SIP state implementation plan
TCEQ Texas Commission on Environmental Quality
TTI Texas Transportation Institute
TVT Travel Volume Trends
VMT vehicle miles of travel
VTRIS Vehicle Travel Information System
WIM weigh-in-motion
xiv
-------
1. INTRODUCTION
Heavy-duty vehicles have been seen as contributing a large fraction of emissions from
on-road vehicles and are coming under more intense scrutiny because light-duty emissions have
been controlled to a greater extent than heavy-duty vehicle emissions. A heavy-duty vehicle can
produce 10 to 100 times the emissions (of NOX and PM emissions especially) of a light-duty ve-
hicle. Thus, heavy-duty vehicle activity needs to be better characterized. Key uncertainties with
the use of MOBILE6 regarding heavy-duty vehicle emissions include the fraction of heavy-duty
vehicles on all types of roadways at all times of day. In addition, there may be regional variabil-
ity in both the fraction of different vehicle classes and the vehicle weights within each class.
With the MEASURE1 model and the developing MOVES2 model (the eventual replace-
ment for MOBILE3), greater emphasis is given to physical parameters affecting the engine loads
and therefore the emissions from individual vehicles. One primary factor affecting the engine
load is the vehicle weight; the weight of the vehicle on the road is needed to estimate the in-use
emissions of given vehicles. Because the effect of vehicle weight may be nonlinear for certain
types of driving, it is important to incorporate the weight distribution of vehicles into emission
estimates.
Databases collected by the Federal Highway Administration (FHWA) include vehicle
count and classification from the Highway Performance Monitoring System (HPMS) using au-
tomated traffic recorders (ATR) used to produce the Travel Volume Trends (TVT) reports. Other
data sets compile the results of data collection from weigh in motion (WIM) sensors, and other
data sources (visual observation, weigh stations, and other special projects) maintained by the
FHWA and compiled in the Vehicle Travel Information System (VTRIS). A discussion of these
'MEASURE = Mobile Emissions Assessment System for Urban and Regional Evaluation. Model. This model is a
prototype GIS-based modal emissions model.
2MOVES = Mobile Vehicle Emissions Estimator, next generation mobile source emissions model. The model will be
used for State Implementation Plan emission inventories and will replace the current MOBILE model.
3MOBILE = Current mobile source emissions model used for State Implementation Plan emission inventories.
-------
data sources including original sources, representativeness, and quality and data reduction proce-
dures used in this work are provided in Appendix A.
The primary goals of this work were to investigate the vehicle weights and mix of ve-
hicle classes depending upon a number of regional and temporal factors by vehicle and roadway
types. ENVIRON reviewed and in this report suggests how the TVT data can be used to estimate
temporal variability (by month, day of week, time of day) of total traffic volumes for all vehicles
types combined. Using the VTRIS data the results of this work are provided as summary data
in a series of files that combine and average weight, weight distributions, and vehicle mix de-
pending upon the state where the measurement was taken or as a national average, time period
(month, day of week, or hour of day), roadway type as described in Table 1, and vehicle classifi-
cation as described in Table 2.
Table 1. FHWA roadway functional classification (types) in VTRIS.
Rural Urban
Code
1
2
6
7
8
9
Classification Description
Principal Arterial - Interstate
Principal Arterial - Other
Minor Arterial
Major Collector
Minor Collector
Local System
Code
11
12
14
16
17
19
Classification Description
Principal Arterial - Interstate
Principal Arterial - Other Freeways or Expressways
Principal Arterial - Other
Minor Arterial
Collector
Local System
Table 2. FHWA vehicle classifications.
FHWA VTRIS Vehicle Type
1
2
O
4
5
6
7
8
9
10
11
12
13
14
15
Motorcycle
Passenger cars
Other 2-axle, 4-tire single unit vehicles
Buses
2-axle, 6-tire single-unit vehicles
3 -axle, 6-tire single-unit vehicles
4+ axle single-unit vehicles
4 or less axle combination vehicles
5-axle combination vehicles
6+ axle combination vehicles
5-axle multitrailer vehicles
6-axle multitrailer vehicles
7+ axle multi-trailer vehicles
Unclassified
Unclassifiable
-------
1.1 Average Vehicle Weight and Weight Distributions
The vehicle weight observations were not grouped by any method prior to averaging by
the categories described in this report, namely roadway class, vehicle class grouping, month, day
of week, or hour. Each observation was treated with equal weight in the calculation of the sum-
mary statistics.
The vehicle weight can be presented as both an average and as a distribution of the ve-
hicles across a weight bin spectrum. Table 3 provides the average weight range over the states
by vehicle class across various roadway types and using the 1999 and 2000 VTRIS data. The
smaller vehicle classes have vehicle weights that are higher than one might expect and have more
relatively variability than heavier vehicles, especially vehicle classes 1-3. The variability could
be a function of the error in the measurement itself where the error is constant without regard
to the vehicle weight, but a detailed evaluation of the measurement error is beyond the scope of
the current project. The average weight likely demonstrates which vehicles are most like other
vehicle classes, for instance vehicle class 7 is more like vehicle classes 9-13 while vehicle class 8
is more like vehicle class 6.
Table 3. Vehicle identifiers and typical average vehicle weight range.
FHWA
„. Description Average Vehicle Weight (Ibs.)
L^iass
1
2
O
4
5
6
7
8
9
10
11
12
13
Motorcycles*
Passenger vehicles
Two-axle, four-tire single-unit trucks
Buses
Six-tire, two-axle single-unit vehicles
Three-axle single-unit vehicles
Four or more axle single-unit vehicles
Three or four axle single-trailer vehicles
Five-axle single-trailer vehicles
Six-axle multi-trailer vehicles
Five or less axle multi-trailer vehicles
Six-axle multi-trailer vehicles
Seven or more axle multi-trailer vehicles
8,000-25,000
4,500-9,000
7,000 - 9,000
25,000-29,000
12,000 - 14,000
24,000 - 30,000
41,000-58,000
26,000-31,000
48,000 - 58,000
60,000 - 65,000
50,000-61,000
56,000-63,000
72,000 - 92,000
*Motorcycle data highly variable and not used in this analysis.
When investigating vehicle weight distribution, the weight bin distribution listed in Table
4 was used to demonstrate the range of vehicle weights.
-------
Table 4. Vehicle weight bin descriptions.
Weight Bin
Number
Weight weight Range (Ibs)
Olgoo
, Low Weight (<) High Weight (<)
Mid-point
*
Weight (Ibs)
0
1 -B20
2-B25
3-B30
4-B35
5-B40
6-B45
7-B50
8-B60
9-B70
10-B80
11 -B90
12-B100
13-B140
14-B160
15-B195
16-B260
17-B330
18-B400
19-B500
20 - B600
21 -B800
22-B1000
23-B1300
24 - B9999
0
20
25
30
35
40
45
50
60
70
80
90
100
140
160
195
260
330
400
500
600
800
1000
1300
9999
NA
0
2,000
-------
0.6
0.5 -
0.4 -
0.3 -
0.2 -
0.1 -
_J1_ _m_ _jt.
rS>
<
^Sunday BMonday ^Tuesday ^Wednesday BThursday HFriday ^Saturday
Figure 1. Weight bin distribution by day of week for vehicle class 2 (passenger cars) on road type 11 in
2000.
0.6
0.5 -
0.4 -
0.3 -
0.2 -
0.1 -
<> <>
dSunday BMonday ^Tuesday ^Wednesday BThursday HFriday ^Saturday
Figure 2. Weight bin distribution by day of week for vehicle class 3 (7000-9000 Ib two-axle, four-tire
single-unit trucks) on road type 11 in 2000.
-------
0.6 1
0.5 -
0.4 -
0.3 -
0.2 -
dSunday BMonday dTuesday dWednesday BThursday HFriday ^Saturday
Figure 3. Weight bin distribution by day of week for vehicle class 5 (12,000-14,000 Ib six-tire, two-axle
single-unit vehicles) on road type 11 in 2000.
0.6
0.5 -
0.4 -
0.3 -
0.2 -
0.1 -
DSunday • Monday ^Tuesday ^Wednesday ^Thursday ^Friday ^Saturday
Figure 4. Weight bin distribution by day of week for vehicle class 6 and 7 (24,000-58,000 Ib three or
more axle single-unit vehicles) on road type 11 in 2000.
-------
0.6
0.5 -
0.4 -
0.3 -
0.2 -
0.1 -
^Sunday • Monday ^Tuesday ^Wednesday • Thursday HFrday ^Saturday
Figure 5. Weight bin distribution by day of week for vehicle type 8-13 (26,000-92,000 single- and
multi-trailer vehicles) on road type 11 in 2000.
1.2 VMT Mix
The vehicle mix data were provided in FHWA vehicle classes. Appendix B provides a
suggested method to cross-reference the FHWA vehicle class into MOBILE vehicle types. The
FHWA vehicle mix categories do not necessarily correspond to the MOBILE vehicle types, so
some estimates and governing assumptions about the vehicle fleet make-up must be made to
cross-reference the FHWA classes into vehicle classes useful for emission estimation. Many
states including Texas, Wisconsin, Illinois, Michigan, and Minnesota are using the FHWA vehi-
cle classification data to better estimate the vehicle mix for their emissions modeling. The vehicle
mix results presented in this work also provide an understanding of the more important vehicle
classes when investigating the vehicle weight.
In order to ensure that the vehicle classes count data was not more heavily weighted by
sites with longer periods of observation than others, but rather weighted by sites with heavier
traffic volume, the class counts were averaged at individual sites before being averaged across
sites. The steps followed in processing the class count data are as follows:
1. All counts across lanes in the same roadway direction were totaled. Different direc-
tions at site were treated separately.
7
-------
2. All counts (either total volume or count for each vehicle class) were averaged for each
site-direction pair by hour, day of week (i.e., Sunday through Saturday), month, and
roadway classification. This means that at most five values were averaged together,
corresponding to the total number of days in a week during one month. In other
words, all Monday counts during January for hour 10 were averaged together at each
site-direction pair.
3. The hourly class counts were averaged across the sites by roadway function class,
vehicle class, month, day of week, and hour of the day.
-------
2. INVESTIGATION OF REGIONAL AGGREGATION
An investigation was conducted to determine if are were any regional differences in the
vehicle weight and vehicle mix. In this study, the average vehicle weight was used rather than
the weight bin distribution because it is more difficult to understand state-to-state differences
in the distributions and any differences in the weight distribution are nearly always reflected in
the average weight. The vehicle mix was also grouped into primarily light-duty and heavy-duty
vehicle classes to avoid confusion that might result from a large number of vehicle classes. The
term regional may describe groups of adjacent states or states that exhibit similar travel patterns
perhaps for similar reasons.
2.1 Average Weight by State
The VTRIS site information (where vehicle class counts are made and vehicle weights
are measured) contains the state and county FIPs codes. Using this information, it is possible
to aggregate vehicle class count and vehicle weight distributions by designated state and county
groupings, where the groupings could extend from one state into another. ENVIRON extracted
the data corresponding to interstates and freeways by county to look for possible regional effects.
The vehicle weight information was compared using average weight ranges for vehicle
classes 8-13, the larger combination vehicles, and vehicle classes 5-7, single-unit trucks. The
larger combination vehicles represent a nearly homogenous grouping of class 8 trucks, while
vehicle classes 5-7 include all types of trucks. Therefore the vehicle classes 8-13 show a more
uniform vehicle weight range than other truck types.
Figures 6 and 7 show the relative average weight by state for vehicle classes 8 - 13 on
rural and urban interstate (average weight 50,000 Ibs.) roadway types, which were those with the
greatest number of vehicle measurements. The 1999 data is represented by the solid blue bars,
and the 2000 data is represented by the red hashed bars. The number of observations is reported
underneath each bar. No individual state had an average vehicle weight in excess or less than
30 percent of the national average. On rural interstates, only Indiana, Michigan, and New Jersey
had consistent (two year) averages with an average weight less than 10% of the national average,
while only South Dakota had consistently higher average weight readings. On urban interstates,
the data are more variable from state to state with some states showing extraordinary averages,
especially Georgia where the estimate was based on only 597 observations and South Carolina
based on less than 2,000 observations for each year. However, Connecticut and Wyoming both
show consistently higher average vehicle weights.
-------
100% 1 -
60%
40%
20%
0)
AR
5
|
ODmcoaico'S-r^ococoinmco';
i-incOCTiCOOCOCOtfrtOtMCViCOU
o co co in -^ T
•fl" i- ,CN) T
of CM" JT-"
\RCACACTCTF
566991
: | 18 8 J 5 |
L ID ID IN IN IA IA KSK
r
-LnSS>R™sss™S!js§ffs;ra?s°
-c\icoT-^-ooioc\iin^i-^-tNtQin^CN*-eO
Sco" to o" r--" i*-" CD" CM m" eo CD" CD" 10" : oi T- co" f--"
T-a>r-i^iriLO
SKYKYMI MIMSMSMCMOMTNENVNVNJNJNMNMNCNCOHPAPARI Rl SCSCSDSDTXTXUTVAWAWAWI WIWY
2 16 16 18|18 19, 19|20J20 21,22,26 26 28 28 29 29 30 31 32132 34,34,35,35,37,37,39 ,42,42i44i44,45 45,46,46 48 48,49,51 53,53 55 55 56
CD
3
i/VY
56
Figure 6. Relative average weight by state for vehicles 8 - 13 in 1999 and 2000 on road type 1, rural in-
terstates. (Number of observations used for each state average reported along axis. 1999 data is solid blue
and 2000 is red on white hash.)
120%
100% - n
80% -
60% -
40%
20%
o r*- co CM
° S t-- Sj
tO O5 !•*- f^
T- CM CO CO
CD O) CO 1^
OJ CO OJ
AR AR CA C/
5566
n f
•^•r--coLncNcococ7>iocoi^i-o
CM
-------
Figures 8 and 9 show the relative average weight by state for vehicle class 5 on rural (av-
erage weight 13,000 Ibs.) and urban interstates (12,000 Ibs.) roadway types. The average weight
for this vehicle class varies more widely from state to state and year to year, perhaps because
the vehicle represented by this vehicle class could be one of several gross vehicle weight ranges
from light-duty up to Class 8 trucks.
200%
180%
160%
140%
120%
100%
80%
60%
40%
20%
0%
co m m in m in
.CACTCTFLID ID IN
IAKSKSKYU\MIMIMNMSMSMCMOMTNENVNVNJNJNIVWMNCNCOHPA!PARI RIJSCSCSDSDTXJTXUTJVAW,
'AWAWI WIWWVY
9 9 12 16 16 18 18 19 19 20 20 21 22 26 26 27 28 28 29 29 30 31 32 32 34 34 35 35 37 37 39 42 42 44 44 45 45 46 46 48 48 49 51 53 53 55 55 56 56
YWY
Figure 8. Relative average weight by state for vehicle 5 in 1999 and 2000 on road type 1, rural interstates.
(Number of observations used for each state average reported along axis. 1999 data is solid blue and 2000
is red on white hash.)
11
-------
140%
120%
100%
80%
60%
40%
20%
0%
AR AR CA
5 6
CACO
6 8
CT
•^ CM CO
cvi co g
CO CD CO
GA IN IN KS KS KY Ml MNMSMOMONE NV NJ NJ OH Rl Rl SC SC VA WAWAWV Wl WYWY
13 18 18 20 20 21 26 27 28 29 29 31 32 34 34 39 44 44 45 45 51 53 53 54 55 56 56
CO *^r "^ N- CO CO CO
CD UO CO ^f I*- m CD
O) T- O5 CO CD O O
of co" *-"
Figure 9. Relative average weight by state for vehicle 5 in 1999 and 2000 on road type 11, urban inter-
states. (Number of observations used for each state average reported along axis. 1999 data is solid blue
and 2000 is red on white hash.)
No discernable regional pattern in vehicle weights could be determined from their data,
as shown in Figure 10 for rural interstates. Some states (e.g., Wyoming, Indiana) show consis-
tently higher or lower weights than the national average, but neighboring states do not show a
similar pattern. Therefore a clear determination of state-to-state regions that affect vehicle weight
could not be found.
1999 Average Vehicle Weight Classes 8-13: Rural Interstates 2000 Average Vehicle Weight Classes 8-13: Rural Interstates
Figure 10. State by state average weight for vehicle classes 8 - 13 on rural interstates in 1999 and 2000.
12
-------
2.2 Vehicle Mix
The vehicle mix could depend upon a number of factors including the road type, month
and day of week as well as regional definitions. It was discovered that by and large the weekdays
could be combined, although the day of week does have some subtle effects on the class frac-
tions. This is discussed in more detail later.
The vehicle mix information shows some potential regional variability, especially on
rural interstates during the week. Figure 11 shows the distribution of vehicle mix by state for
1999 and 2000, where the state and year is indicated at the base of the bar. States that show a
low heavy-duty (vehicle classes 4-13) mix were California, Florida, New Jersey, and Rhode
Island. Those states with high heavy-duty mixes were Arkansas, Georgia, Iowa, Missouri, New
Mexico, Ohio, Oklahoma, and Pennsylvania. Therefore, ocean coastal states tended to have low
truck activity relative to that of light-duty vehicles while interior states had higher truck activity
on rural interstates.
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
O Class B-13
LLLL.LJ LJ U LI.LJ
n ja £> .Q
a s
111
cn O 6i
O) O CT)
(D O
o to
C' >:-j
s *
60 to 00
3 S
O Ot f
*• A g
— (N
5 I
o ™
Figure 11. Vehicle mix for weekdays on rural interstates for 1999 and 2000 data. (Number of site-days of
observations used in the state average reported along axis.)
13
-------
The vehicle mix pattern was less discernable when other rural and urban roadway types
were considered. Figures 12 and 13 show the rural principal arterial and urban interstate vehicle
mix results. Regional patterns were not as clearly defined for these road types, although one may
find similarities among the coastal states, which tend to have higher fractions of vehicle classes 1-3.
inn6'
90% |
80%
70%
60%
50% |
40% 1
30% |
20% I
10%
0% I
1 - Jj]
1
n Class 8-1 3
• Class 4-7
Q Class 1-3
MJUUnUUL
Illlllllllllllllllllllllllllllllllllllllllllll
f-- (Q OO if} *— .. . . f^- . . •*- CO
r- ^— .. . . ^- O) Q »™ O ^~ Q
< < O ~
*— i— if) . . i— ^ , . O O if! »— f*J »— 5) ..CMO^P
. ,,01 .. _.0i^i£)ilttliOr-- -- ..QtMU") -. .
ifli^i^iii|ii^||s;
>..O>OCT>OC3>O. OOi^OlO
"X O QQ<< 5>^>
O to 03 g g S* *
Figure 12. Vehicle mix for weekdays on rural principal arterials for 1999 and 2000. (Number of site-days
of observations used in the state average reported along axis.)
14
-------
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
aa
oo
•g -8
-8
5
O) *- *-
§
5
*
g i
o
Figure 13. Vehicle mix for weekdays on urban interstates for 1999 and 2000. (Number of site-days of
observations used in the state average reported along axis.)
One of the difficulties with the 1999 and 2000 VTRIS data is the inconsistency in data
availability by state as shown in Table 5. Many adjacent states are missing when defining region-
al aggregations. Because of the lack in the geographical coverage of the VTRIS data, it may not
be possible to establish specific state-to-state travel patterns using the VTRIS data. In order to
begin to discern regions of like rural interstate vehicle class mixes, five regional categories were
defined as shown in Table 5 and Figure 14 primarily based on states with like vehicle mix, which
maximized the chi square statistical significance of each region/state combination. The choice of
regional aggregation was therefore made on the basis of the empirical observations rather than an
assumption of typical travel behavior.
15
-------
Table 5. Availability of VTRIS vehicle counts for weekday rural interstates and suggested regional categories.
Vehicle Class Counts
State Regional Group
1999 2000
Alabama
Alaska
Arizona
Arkansas
California
Colorado
Connecticut
Delaware
Florida
Georgia
Hawaii
Idaho
Illinois
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maine
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
Missouri
Montana
Nebraska
Nevada
New Hampshire
New Jersey
New Mexico
New York
North Carolina
North Dakota
Ohio
Oklahoma
Oregon
Pennsylvania
Rhode Island
South Carolina
South Dakota
Tennessee
Texas
Utah
Vermont
Virginia
Washington
West Virginia
Wisconsin
Wyoming
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
4
1
1
5
4
3
3
3
2
2
4
2
1
5
2
4
4
4
1
2
3
2
x = data is present.
16
-------
The average vehicle mix for each region is shown in Table 6 for weekdays. An average
weekday was used to combine data to demonstrate the state-to-state differences, though it is also
demonstrated later in this report that each weekday can have a distinct average. The vehicle
categories were combined in either two (vehicle classes 1-3 or 4-13) or three (shown in Table 6)
different categories, though the regions were denned with the two-category groupings. In gen-
eral, group 1 consisted of coastal states (east and west), while groups 2-5 consisted of interior
states. Among the interior states rural interstates might be expected to have higher fractions of
heavy-duty vehicles engaged in interstate commerce. Five regional groups were needed to show
that the vehicle mix distribution was similar for the states within a group and significantly differ-
ent between groups. The chi square probability, comparing the individual state distribution to the
regional average, is reported and shows a high probability (>0.05) for most states and years that
they are reasonably explained by the average for that region. The probability is usually higher
but not always so when using a two-category (group 1-3 and 4-13) test compared with the three-
category test.
17
-------
Table 6. Vehicle class fractions for weekdays on road type 1, rural interstates.
Chi Square, p-
w 0* A. i-ir. 0* A. r, • Class Class Class Number of ,,oi,,oc
Year State F Ps State Region „ „ . , _ „„ „ .. values
a 1-3 4-7 8-13 Observations
2 Cat. 3 Cat.
1999
2000
1999
2000
1999
2000
1999
2000
1999
2000
1999
2000
1999
2000
1999
2000
2000
2000
1999
2000
2000
1999
1999
2000
1999
2000
1999
2000
1999
1999
2000
1999
2000
5
5
6
6
12
13
19
19
20
20
21
22
26
26
28
28
29
30
34
34
35
37
39
40
42
42
44
44
53
55
55
56
56
All
Region 1
Region 2
Region 3
Region 4
Region 5
Arkansas
Arkansas
California
California
Florida
Georgia
Iowa
Iowa
Kansas
Kansas
Kentucky
Louisiana
Michigan
Michigan
Mississippi
Mississippi
Missouri
Montana
New Jersey
New Jersey
New Mexico
North Carolina
Ohio
Oklahoma
Pennsylvania
Pennsylvania
Rhode Island
Rhode Island
Washington
Wisconsin
Wisconsin
Wyoming
Wyoming
Average
Average
Average
Average
Average
Average
4
4
1
1
1
5
4
4
3
3
3
3
2
2
2
2
4
2
1
1
5
2
4
4
4
4
1
1
2
3
3
2
2
All
1
2
3
4
5
61%
65%
87%
89%
83%
54%
61%
67%
67%
75%
69%
69%
78%
81%
81%
82%
67%
79%
80%
85%
52%
80%
65%
63%
69%
58%
84%
82%
81%
78%
73%
81%
79%
77%
85%
80%
72%
64%
53%
5%
7%
4%
4%
4%
25%
4%
4%
4%
4%
4%
5%
9%
6%
4%
3%
4%
5%
7%
4%
11%
4%
4%
12%
4%
5%
4%
5%
5%
4%
6%
2%
3%
5%
5%
5%
5%
5%
20%
28%
9%
7%
13%
20%
34%
29%
29%
20%
27%
26%
13%
13%
16%
15%
29%
16%
13%
10%
37%
16%
31%
24%
27%
37%
12%
13%
14%
18%
21%
17%
17%
18%
10%
15%
23%
31%
26%
9432
84096
1896
1344
1104
96
2760
47184
72
144
696
192
96
768
1680
2160
3648
30168
33408
22344
912
240
10872
264
528
1608
864
624
1872
6864
10464
960
912
0.16
0.56
0.08
0.00
0.07
0.70
0.27
0.12
0.13
0.16
0.04
0.22
0.16
0.61
0.73
0.35
0.30
0.70
0.00
0.94
0.61
0.81
0.58
0.84
0.06
0.01
0.36
0.01
0.73
0.00
0.44
0.78
0.84
0.32
0.11
0.21
0.00
0.02
0.01
0.29
0.27
0.23
0.34
0.05
0.47
0.00
0.59
0.20
0.10
0.48
0.90
0.00
0.92
0.00
0.42
0.27
0.00
0.17
0.01
0.28
0.02
0.91
0.01
0.18
0.20
0.45
18
-------
State-Region Code
region 1
region 2
region 3
^| region 4
^| regions
Figure 14. Regions denned for rural interstate vehicle mix.
While there are similar regional differences for other road types, the same regional pat-
terns for rural interstates do not completely match those for other road types. For instance, as
shown in Table 7 for road type 2 (rural principal arterials), the coastal states (including Cali-
fornia, Connecticut, Florida, New Jersey, and Rhode Island) have higher fractions of light-duty
vehicles (group 1-3) similar to road type 1, but other states (such as Michigan and Montana) also
have high fractions for this vehicle group (1-3). Likewise as shown in Table 8 for road type 11
(urban interstates), the coastal states have higher fractions of light-duty vehicle group (1-3), but
so does the interior state of Kansas.
19
-------
Table 7. Vehicle class fractions for weekdays on road type 2, rural principal arterials.
Year State Class 1-3 Class 4-7 Class 8-1 3 Number of Observations
1999
2000
1999
2000
2000
1999
2000
1999
2000
1999
2000
1999
2000
1999
1999
2000
1999
1999
2000
1999
2000
2000
2000
1999
1999
2000
1999
2000
1999
2000
2000
1999
2000
1999
2000
1999
2000
1999
2000
1999
2000
2000
1999
2000
1999
2000
Arkansas
Arkansas
California
California
Colorado
Connecticut
Connecticut
Florida
Georgia
Iowa
Iowa
Kansas
Kansas
Kentucky
Louisiana
Louisiana
Maine
Michigan
Michigan
Mississippi
Mississippi
Missouri
Montana
Nebraska
Nevada
Nevada
New Jersey
New Jersey
North Carolina
North Carolina
Ohio
Oklahoma
Oklahoma
Pennsylvania
Pennsylvania
Rhode Island
Rhode Island
South Dakota
South Dakota
Washington
Washington
West Virginia
Wisconsin
Wisconsin
Wyoming
Wyoming
82%
79%
86%
87%
62%
92%
94%
89%
90%
84%
83%
83%
84%
61%
87%
81%
91%
84%
87%
80%
80%
84%
87%
83%
84%
89%
86%
88%
91%
93%
58%
93%
78%
85%
77%
95%
95%
85%
84%
83%
82%
91%
84%
83%
94%
81%
4%
5%
4%
5%
5%
4%
3%
7%
9%
5%
4%
5%
5%
15%
6%
7%
3%
9%
5%
3%
3%
5%
5%
3%
5%
4%
8%
7%
4%
5%
4%
3%
8%
3%
5%
3%
3%
5%
6%
7%
8%
6%
5%
6%
4%
6%
14%
15%
9%
8%
34%
4%
3%
4%
2%
11%
13%
13%
11%
24%
7%
12%
6%
8%
8%
16%
17%
11%
8%
15%
11%
7%
6%
5%
5%
2%
37%
3%
14%
11%
18%
2%
2%
10%
10%
10%
10%
3%
11%
12%
2%
13%
31,704
71,688
864
576
21,168
48
96
1,704
24
1,104
20,832
168
192
528
72
120
480
48
1,056
6,072
4,560
5,160
67,248
192
960
48
72,216
75,936
408
240
11,472
48
2,904
360
408
960
960
16,560
24,720
64,848
70,056
96
7,680
18,600
528
4,176
20
-------
Table 8. Vehicle class fractions for weekdays on road type 11, urban interstates.
Year State Class 1-3 Class 4-7 Class 8-1 3 Number of Observations
1999
2000
1999
2000
2000
1999
1999
2000
1999
1999
1999
2000
1999
2000
2000
1999
2000
1999
2000
1999
2000
1999
Arkansas
Arkansas
California
California
Connecticut
Florida
Kansas
Kansas
Kentucky
Michigan
Mississippi
Missouri
New Jersey
New Jersey
Oklahoma
Rhode Island
Rhode Island
Washington
Washington
Wisconsin
Wisconsin
Wyoming
83%
85%
92%
92%
86%
84%
92%
90%
72%
83%
74%
88%
88%
88%
88%
95%
94%
92%
92%
84%
80%
63%
10%
5%
3%
3%
3%
10%
3%
4%
12%
8%
4%
5%
5%
5%
5%
2%
3%
4%
4%
4%
4%
6%
7%
10%
5%
5%
10%
5%
5%
6%
16%
9%
22%
7%
7%
7%
6%
2%
4%
4%
4%
11%
16%
31%
6,216
30,288
5,184
2,328
24
816
72
96
216
72
1,440
1,872
51,720
46,416
696
144
288
48
1,032
1,296
4,944
144
In summary, the regional vehicle mix does not generally vary by region or state. While
vehicle mix differences between states identified for rural interstates were found to be significant,
no discernable patterns could be identified for other road facility types. Regional groupings made
on the basis of the empirical data identified similar states that were not contiguous. One might
speculate that the different regions identified for rural interstates may be explained in part by
cross-country interstate freight movements, where core interior states experience higher fractions
of heavy truck activity than states outside of this core. Therefore regions geographically dis-
persed, such as the east and west coasts of lower 48 states, may be more similar in the nature of
their traffic than would regional groupings based on proximity.
21
-------
3. REGIONAL VARIATIONS IN WEIGHT BIN DISTRIBUTIONS BY ROAD TYPE
FOR VEHICLE CLASSES 5-13
One of goals of this effort was to determine the regional variability in weight distribu-
tions by road type, if any. Weight bin distributions for each state by road type were prepared
for review. During the analysis, it became apparent no regional truck traffic groupings could be
clearly defined and the weight distributions could not be well defined either. Figure 15 (combines
the data in Figures 7 and 9 for vehicle classes 5 and 8-13 with the more rare vehicle classes 6 and
7 converted to actual measured weights) shows the state-by-state variability for vehicles 5-13,
but these averages were based on very little data for the states of Georgia (lowest average weight
ratings for vehicles 5-13), Kansas, Missouri, Nebraska, Nevada, South Carolina, West Virginia,
and Wyoming.
80.000
70,000
60.000
50.000
40.000
30.000
20.000
10,000
O Vehicle 5
• Vehicle 6
D Vehicle 7
• Vehicle 8-13
£ X S
10 co r^ 01 d m
ARARCACACOCTCT FLGA IN IN KS KS KY Ml WNMSMGWONENVNJ NJOH Rl Rl SCSCVAW/WA/VVWIWWVY
o> to to o o
Figure 15. Average vehicle weight for vehicle classes 5 - 13 on urban interstates, road 11. (Number of
site-days of observations used in the state average reported along axis.)
22
-------
4. TEMPORAL VARIATIONS IN WEIGHT BIN DISTRIBUTIONS FOR VEHICLE
CLASSES 5-13
The temporal weight distributions are typically difficult to visualize, so average vehicle
weights are also reported here to better understand the temporal distributions in vehicle weight.
Changes in the weight bin distributions are normally reflected in the average weight. Overall the
variability in the average monthly weights does not reflect a consistent pattern, but the variability
in the daily and hourly average weights do reflect patterns.
4.1 Monthly Variation
Vehicle weight does not appear to vary much or consistently by season. Both the average
vehicle weights and weight bin distributions show little change from one season to the next. Figures
16 and 17 demonstrate that the month-to-month variability in the average vehicle weight does not
depend upon the season. Based on the vehicle mix and the sample sizes, the vehicle classes with
the highest fraction of the fleet are in order, vehicles 9, 5, and 8. The variability between months in
Figures 16 and 17 are well within the standard deviation, but the very large sample sizes provided
in Table 9 reduce the 90% confidence level ranges to those shown in Figures 16 and 17. However
the uncertainty ranges shown in Figures 16 and 17 do not include sampling variability by site.
35,000 1
30,000
25,000
~ 20,000
15,000
10,000
5,000
10 11
12
Figure 16. Average weight by month for lighter heavy vehicle classes in 1999 on road type 1. (Uncertainty
ranges were based on 90% confidence levels of the sample.)
23
-------
100,000
90,000
80,000
70,000
60,000
50,000
40,000
30,000
20,000
10,000
10 11 12
Figure 17. Average weight by month for heavier vehicle classes in 1999 on road type 1.
Table 9. Number of observations for Figures 16 and 17.
Month
1
2
3
4
5
6
7
8
9
10
11
12
V
5R 7 R
D / O
198,451
110,475
137,986
179,017
235,213
116,013
177,481
143,708
95,184
187,238
98,449
58,532
49,242
24,247
102,717
81,381
88,309
40,642
31,622
31,968
25,002
46,237
33,368
36,540
3,381
3,433
9,220
2,627
15,012
5,332
1,050
6,028
4,404
2,952
2,603
2,922
97,131
48,656
158,003
129,504
143,477
57,473
60,793
56,318
56,406
79,567
70,070
45,128
'ehicle class
9 10 11 12 13
1,323,544
664,163
1,225,367
1,278,718
1,505,906
640,491
889,347
655,697
742,091
805,333
767,994
471,771
30,501
6,324
22,381
22,394
30,995
19,282
11,578
6,916
18,476
15,474
33,015
20,871
107,188
33,788
63,106
104,815
72,635
30,660
79,424
32,430
30,451
87,282
33,589
21,389
19,345
3,075
9,392
18,480
11,828
5,829
11,760
5,112
5,948
13,709
9,200
7,061
22,733
3,413
4,850
3,980
21,048
17,490
5,205
4,200
19,382
6,815
32,130
23,352
24
-------
4.2 Hourly Variation Over an Average Week
Using national averages, the average weight of vehicles is shown to vary by day of week
and hour of day in Figures 18 and 19 for vehicle class 5 and vehicle group (8-13). (Hour 1 in
Figures 18 and 19 is 12-1 a.m. on Sunday.) The average weight for vehicle class 5 is clearly
lower on weekends and had a distinctive hourly profile during weekdays that was also apparent
in the 2000 data as in the 1999 data shown in Figure 18. For the vehicle group (8-13), the vehicle
weight increases on weekends and overnight during the week, with a similar pattern in 2000 as
that shown in Figure 19 for 1999.
The average weight by hour for all seven days in the week shown in Figures 18a and 19 reflect
the distribution of vehicle weights for various vehicle classes. (Outlier data from Indiana, as presented in
Figure 23, was identified that greatly affected the average weights for vehicle class 5. While there was
no obvious reason to eliminate this data, alternative versions of Figures 18b and Figure 21b are provided
without the Indiana data.) Figures 20 to 28 show the effect from the day of week for vehicle class 5
and vehicle class group (8-13) on the population by weight bin. For vehicle class 5, the distribution of
weight shifts to lower weight bins on weekend days for road types 1 (rural interstates) and 11 (urban
interstates). This effect is demonstrated dramatically in Figure 21a of the 1999 data including the state
of Indiana, but it is not as great in Figure 21b of the 1999 data excluding Indiana. The opposite effect is
demonstrated for vehicle group (8-13) in Figure 23, where vehicle weights on weekends are higher than
during the week. This effect is most apparent in the weight bin B800 vehicle fractions.
18,000
16,000
14,000
12,000
« 10,000
g 8,000
6,000
4,000
2,000
121 145
Figure 18a. Average weight for vehicle 5 by hour over a typical week in 1999.
25
-------
18000
16000
14000
12000
10000
g 8000
6000
4000
2000
145
Figure 18b. Average weight for vehicle 5 by hour over a typical week in 1999. (Without Indiana Data)
70,000
60,000
50,000
•5- 40,000
3 30,000
20,000
10,000
145
Figure 19. Average weight for vehicle group (8 - 13) by hour over a typical week in 1999.
26
-------
The hourly averages in Figures 18 and 19 represent data for each hour. Typically there
were more observations during daytime hours and during weekdays than at night or on the week-
end days. The range in the number of observations for each hour is shown in Table 10.
Table 10. Range in the number of observations for Figures 18 and 19.
\/6 h I C 1 6
Road Type Minimum Observations Maximum Observations
CI3SS
5
5
5
5
8-13
8-13
8-13
8-13
1
11
2
12
1
11
2
12
1,495 at Sam Sunday
2,242 at 4am Sunday
674 at 4am Sunday
711 at 4am Sunday
25,056 at Sam Sunday
7,331 at 4am Sunday
3,039 at Sam Sunday
1,674 at 2am Sunday
20,228 at 3pm Wednesday
41,021 at 3pm Tuesday
14,888 at 3pm Friday
22,292 at 2pm Friday
126,553 at 1pm Wednesday
88,510 at 11am Tuesday
43,055 at 11 am Tuesday
27,829 at 11am Thursday
0.3 -i
0.25 -
0.2 -
0.15 -
0.1 -
0.05 -
Day of Week Weight Bin Fractions
Vehicle Class 5, Roadway Class 1
ISun "Mon ClTue QWed "Thu DFri "Sat
Figure 20. Day of week weight bin distribution for vehicle class 5 on road type 1 in 1999.
27
-------
Day of Week Weight Bin Fractions
Vehicle Classes 5, Roadway Class 11
0.3
0.25 -
0.2 -
0.15 -
0.05 -
hmU
jnm.
• Sun HMon OTue Owed HThu OFri BSat
Figure 2 la. Day of week weight bin distribution for vehicle class 5 on road type 11 in 1999 (including the
Indiana data).
Day of Week Weight Bin Fractions Without Indiana
Vehicle Classes 5, Roadway Class 11
0.25 -i
0.2
0.15
0.1
0.05
HHl
rS>
.np
ISun HMon DTue DWed HThu DFri BSat
Figure 21b. Day of week weight bin distribution for vehicle class 5 on road type 11 in 1999 (excluding
the Indiana data).
28
-------
0.3
0.25 -
0.2 -
0.15 -
0.1 -
0.05 -
0*
<$-
DSunday ^Monday DTuesday DWednesday ^Thursday DFriday ^Saturday
Figure 22. Day of week weight bin distribution for vehicle class 5 on road type 11 in
2000.
The number of weigh-in-motion observations used to generate the 1999 and 2000 weight-bin
distributions for vehicle class 5 on road types 1 and 11 in Figures 20 through 22 are shown in Table 11.
Table 11. Number of observations for weight distributions, vehicle class 5, road types 1 and 11.
No 1QQQ
Road Vehicle _. „.,. . No. 1999 «J *- No. 2000
Type class Day Of Week Observatjons Obsen/at.ons observations
*^ without Indiana
01
01
01
01
01
01
01
11
11
11
11
11
11
11
05
05
05
05
05
05
05
05
05
05
05
05
05
05
Sun
Mon
Tue
Wed
Thu
Fri
Sat
Sun
Mon
Tue
Wed
Thu
Fri
Sat
175,808
251,611
270,264
281,260
279,778
279,961
199,065
261,751
521,583
549,043
541,184
544,315
561,867
349,511
-
-
-
-
-
177,075
428,490
467,148
455,228
470,080
466,715
254,994
332,949
463,760
490,135
494,138
506,475
528,547
379,390
249,728
550,843
583,937
592,730
623,218
661,306
352,577
29
-------
The distribution of weights by bin for vehicle class 5 on roadway type 1 (rural interstates)
is very similar between 1999 and 2000. The large fraction of weights in bin 25 in the 1999 data
for roadway type 11 (urban interstates) was largely influenced by the 1999 Indiana weight data.
Figure 23 displays the fraction of class 5 vehicles from the 1999 Indiana data. Further inquiry
into the 1999 Indiana weight data would be advisable before using the 1999 national weight
distribution such as that shown in Figures 18a and 18b. This situation demonstrates also how a
single data set added to VTRIS can affect national averages. The weight data is composed of a
vehicle classification, the number of axles, and the weight on each axle. Each of these variables
can potentially add erroneous readings. So one might conclude that in the Indiana data, smaller
light-duty vehicles were often misidentified or mislabeled as vehicle class 5, where both class 5
vehicles and light-duty vehicles have two axles.
Indiana 1999 Weight Bin Distribution by Day of Week for Vehicle Type 5 on Road Type 11
0.8
0.7
0.6
0.5
0.4
0.3
0.2
Ikd
DSun
Mon Djue DWed
Thu
Sat
Figure 23. Fraction of class 5 vehicles on urban interstates in the 1999 Indiana weight data.
The weight bin distributions for vehicle classes 8-13 on roadway types 1 and 11 are dis-
played in figures 24 through 27 below. The distributions are nearly identical between 1999 and
2000. For classes 8 through 13 combined, there are often more than one million observations
used in the generation of the histograms below as demonstrated in Table 12.
30
-------
Day of Week Weight Bin Fractions
Vehicle Classes 8-13, Roadway Class 1
0.45 -i
0.4 -
0.35 -
0.3 -
0.25 -
0.2 -
0.15 -
0.1 -
0.05 -
[Ml
• Sun "Mon ClTue Owed "Thu ClFri "Sat
Figure 24. Day of week weight bin distribution for vehicle group (8 - 13) on road type 1 in 1999.
Day of Week Weight Bin Fractions
Vehicle Classes 8-13, Roadway Class 11
0.45 i
0.4 -
0.35 -
0.3 -
0.25 -
0.2 -
0.15 -
0.1 -
0.05 -
rS>
<
• Sun HMon OTue Owed BThu dFri BSat
Figure 25. Day of week weight bin distribution for vehicle group (8 - 13) on road type 11 in 1999.
-------
Day of Week Weight Bin Fractions
Vehicle Classes 8-13, Roadway Class 1
0.45
0.4 -
0.35 -
0.3 -
0.25 -
0.2 -
0.15 -
0.1 -
0.05 -
ClSun HMon DTue dWed HThu DFri BSat
Figure 26. Day of week weight bin distribution for vehicle group (8 - 13) on road type 1 in 2000.
Day of Week Weight Bin Fractions
Vehicle Classes 8-13, Roadway Class 11
0.45
0.4 -
0.35 -
0.3 -
0.25 -
0.2 -
0.15 -
0.1 -
0.05 -
Sun HMon DTue DWed HThu BFri BSat
Figure 27. Day of week weight bin distribution for vehicle group (8 - 13) on road type 11 in 2000.
32
-------
Table 12. Number of observations for weight distributions, vehicle classes 8-13, road types 1 and 11.
_ ._ ., . , , _ -.... . No. 1999 No. 2000
Road Type Vehicle class Day Of Week _. .. _. ..
' Observations Observations
01
01
01
01
01
01
01
11
11
11
11
11
11
11
05
05
05
05
05
05
05
05
05
05
05
05
05
05
Sun
Mon
Tue
Wed
Thu
Fri
Sat
Sun
Mon
Tue
Wed
Thu
Fri
Sat
1,347,601
1,722,437
2,193,374
2,365,475
2,216,493
1,894,417
1,453,452
425,530
1,234,483
1,455,748
1,435,950
1,410,331
1,235,550
569,823
2,598,508
3,530,039
4,338,483
4,522,336
4,339,262
3,795,638
2,820,523
547,073
1,358,486
1,570,902
1,623,461
1,631,364
1,499,633
680,314
Besides the day of week profiles just described, the time of day clearly affects the hourly
weight bin distributions in the manner demonstrated with the average vehicle weight. Figures
28 and 29 show the effect of the time of day during the week by showing the change in distribu-
tion. The hourly weight distribution is reflected in the more easily demonstrated average weight
profiles in Figures 18 and 19.
100%
90%
B0%
70%
60%
50%
40%
30%
20%
10%
0%
089999
D8130D
OB1000
neaoo
DB600
DB500
DB400
BB330
• B260
• B195
DB160
• B140
DB1000
DB90
• B80
• B70
DB60
• BSD
DB45
• 840
DB35
0830
• 826
DB20
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Hour of Day
Figure 28. Vehicle 5 weight bin distribution over an average Wednesday.
33
-------
100% -
90%
80%
70%
60%
50%
40%
I-
30%
20%
10%
0%
-
I
~
_
—
;
~
~~
i
-
!
"~
_
I
_
—
_
-
—
_
P
—
_
-I 1
u u •
mil
_
-
—
-
—
_
_
_
"'
—
—
-
• • • • U L
HlkBIB
_
S
-
^
—
:
-
_
_
=
^
~
-
.
-
—
:
-
—
1
DB9999
DB'300
DB1000
QB800
• B600
DB500
OB400
• B330
• B260
• B195
• B160
• B140
QB100
DB90
• B80
• B70
QB60
• B50
QB45
• B40
DB35
DB30
• B25
DB20
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Hour of Day
Figure 29. Vehicle group (8 - 13) weight bin distribution over an average Wednesday.
34
-------
5. ANALYZE WEIGHT DISTRIBUTIONS FOR VEHICLE CLASSES 1-4
The lighter vehicle classes including buses (class 4) show a decreased potential for dif-
ferences in vehicle weight primarily because the average vehicle weight is lower limiting the
variability in the overall weight that individual vehicles could attain in use. The weight of these
vehicles is typically higher than one might expect to find with light-duty vehicles. Vehicle class
2, so called passenger vehicles, have average weights ranging from 3,000 pounds on rural roads
to 8,000 pounds on urban interstates. Vehicle class 3 has weight ranges more consistent with
those of vehicle class 2 with an average weight range more typically between 6,000 and 7,000
pounds. The average weight for vehicle class 1, so called motorcycles, varies in a wide range and
is usually in great excess of that typically considered reasonable for motorcycles, so the weight
of vehicle class 1 was ignored in this analysis. It is possible that the FHWA method for classify-
ing motorcycles is incorrect, at least when WTM measurements are conducted.
Regional categories could not be defined for vehicle classes 1 through 4. Not many states
measured vehicle weights for vehicle classes 2 and 3, especially on urban road types. There were
no consistent regional trends; states throughout the country could either be typically higher or
lower without any regional grouping.
5.1 Monthly Variation
As with the heavier vehicles, there was no consistent or significant month of year trend in
the average weight of vehicle classes 1-4 as shown in Figure 30. The sample sizes are provided
in Table 13 for Figure 30. Vehicle class 1 had very little data in the months where data existed,
no data at all for some months, and the average results are inconsistent for motorcycles (vehicle
class 1). Vehicle classes 2 and 3 are considered to be primarily light-duty vehicles and have low
average vehicle weights. The monthly averages were higher than generally considered for light-
duty vehicles. These higher than expected vehicle weights suggest that the measurement accu-
racy of the weigh-in-motion stations should be investigated for lower vehicle classes.
35
-------
30,000
25,000
20,000
15,000
10,000
5,000 -
\
6 7
Months
10 11 12
Figure 30. Average weight by month for lighter vehicle classes in 1999 on road type 1.
Table 13. Number of observations for Figure 30.
Month Vehicle 1 Vehicle 2 Vehicle 3 Vehicle 4
1
2
3
4
5
6
7
8
9
10
11
12
931
0
0
0
1,832
1,170
2
0
2,486
1
937
1,314
197,520
17,149
18,135
38,065
479,653
277,274
7,564
6,035
517,459
7,112
224,672
208,587
197,610
12,637
19,140
57,302
249,326
240,001
51,563
33,240
270,325
51,681
185,229
189,706
33,702
18,346
30,843
30,690
41,867
22,754
22,147
22,900
23,264
15,525
29,016
21,174
36
-------
5.2 Daily and Hourly Variation Over an Average Week
The day of week and hour of day trends in the average weight for vehicle class 2 and 3
show no consistent trend by time of day or day of week as shown in Figures 31 and 32. On urban
road types (road types 11 and higher), the average weight of vehicle class 2 is higher and more
variable during the week than on weekends or on rural roads (road type 1); a trend reflected in
the 2000 data as well as in the 1999 data shown in Figures 31 and 32. The high and variable
weekday average values for vehicle class 2 appears in the both the 1999 and 2000 data, and are
largely due to data from Connecticut. Vehicle weight data is generally sparse for the smaller
vehicle classes, compared with vehicle classification or total volume counts, so often only a
few States (five States contributed to Figure 31) provided weight data to VTRIS. When parsing
this data by hour of each day of the week, individual state data entries can significantly affect
the mean. Because the average vehicle weight for the smaller vehicles (classes 2 and 3) is low,
misidentified vehicles or measurement errors can have a larger affect on the estimated mean.
Results, such as that shown in Figures 31 and 32, suggest that vehicle weight data for vehicle
classes 2 and 3 should be ignored or at least considered carefully given the level of uncertainty.
10,000
Vehicle 2 Road 1
2 Road 11
-Vehicle 3 Road 1
'Vehicles Road 11
49 73 97 121
Hour (Sunday - Saturday)
145
Figure 31. Light vehicles weight over a typical week in 1999.
37
-------
10,000 -f
8,000
6,000
g>
1
< 4,000
2,000
Vehicle 2 Road 1
Vehicle 2 Road 11
1 Vehicle 3 Road 1
Vehicles Road 11
25
49 73 97
Hour (Sunday - Saturday)
121
145
169
Figure 32. Light vehicles weight over atypical week in 2000.
The bus (vehicle class 4) data show a consistent trend toward higher average vehicle
weights on weekends than during the week as shown for 1999 in Figure 33; a trend also observed
in the 2000 data. Figure 34 shows how the day of week affects the weight distribution: on week-
end days the weight distribution was shifted to higher weight bins. There is no consistent hourly
trend in the vehicle weight for buses.
38
-------
35,000
30,000
25,000
•- 20,000
15,000
10,000
5,000
25
49
73 97
Hour (Sunday - Saturday)
121
145
Figure 33. Bus weight over a typical week in 1999.
0.3
0.25
0.2
0.15
0.1
0.05
0
rjlla ijff" ajr" mjim !•- rflVfl fllll Illll Illll Illll
I
I
ntlflnrn™
DSunday
• Monday
OTuesday
^Wednesday
• Thursday
O Friday
• Saturday
1 2 3 4 5 6 7 8 9 10 1 1 12 13 14 15 16 17 18 19 20 21 22 23 24
Figure 34. Bus weight bin distribution over a typical week in 1999.
39
-------
6. DEVELOP NATIONAL AVERAGE WEIGHT DISTRIBUTIONS BY FHWA VEHICLE
CLASSES 1-13
National average weight distribution profiles were calculated for 1999 and 2000 and
were provided to EPA for further investigation. However for purposes of this report, it is too
resource intensive to demonstrate and discuss average weight profiles for each vehicle class and
road type. Average vehicle weights are provided in Table 14 and 15 and indicate some aspects of
the national average profile in that the vehicle weight varies by road type and in-use year.
One interesting comparison is between the average weight by vehicle class on rural
road types (1-9) and urban road types (11-16) roads. Vehicle classes 2 and 3 had higher average
weight on urban road types than rural roads. Vehicle class 9, the most populated vehicle class
within the group of vehicle classes 5 through 13, had higher average weight on rural roads com-
pared to urban roads. Vehicle class 5, the second most populated vehicle within vehicle classes
5-13 showed no significant difference in average weight between rural and urban roads.
Table 14. Average (of the monthly averages) and 90% confidence levels (+/-) using the month-
month variability of vehicle weight in 1999.
Road Vehicle Class
1 2 3 4 5 6 7 8 9 10 11 12 13
1
90% Cl
2
90% Cl
6
90% Cl
7
90% Cl
8
90% Cl
9
90% Cl
11
90% Cl
12
90% Cl
14
90% Cl
16
90% Cl
10,624
3,416
24,284
3,144
15,804
21,340
4,790
831
4,926
652
5,961
974
6,009
481
8,010
1,468
7,798
1,077
6,625
693
3,950
2,180
6,504
371
6,456
541
6,966
377
11,811
7,515
7,156
76
8,932
4,004
9,233
4,080
13,167
16,637
24,328
1,081
24,928
829
26,460
2,678
25,427
3,347
41 ,998
68,748
24,152
2,079
29,373
1,019
23,300
3,069
27,208
3,122
13,080
445
11,981
486
11,376
1,115
11,938
2,030
27,450
20,729
9,174
2,348
11,044
1,543
12,351
1,198
13,442
2,430
12,722
1,755
25,124
2,060
26,836
1,202
30,672
1,983
27,312
2,294
27,865
13,591
26,841
37,735
26,652
1,331
28,497
3,055
30,460
3,519
30,236
3,517
47,797
7,047
58,248
2,330
58,498
6,333
50,909
7,845
60,169
3,129
60,787
5,287
63,438
6,772
55,950
10,981
27,011
1,361
27,257
1,313
30,093
2,345
28,756
3,690
45,562
5,092
27,172
15,124
29,620
1,798
29,907
2,299
32,323
4,767
31 ,582
3,642
57,053
1,458
55,720
1,354
51,983
3,859
50,059
6,996
39,019
27,532
36,597
17,722
51 ,546
1,308
48,431
3,140
53,196
3,315
51 ,450
8,565
63,825
2,142
63,496
2,342
63,848
3,530
58,655
7,962
62,689
2,346
60,691
5,403
62,183
3,849
71,069
6,494
57,141
1,747
54,815
2,041
55,055
4,470
47,499
3,994
53,000
1,609
54,344
3,268
54,647
2,132
56,219
10,125
59,244
1,763
55,527
1,897
62,794
6,650
56,922
21 ,259
58,641
1,412
56,915
2,655
62,347
5,331
74,287
13,802
79,102
3,388
75,327
3,143
74,405
2,569
67,109
9,380
77,693
3,324
72,611
6,779
86,976
22,572
100,753
36,512
40
-------
Table 15. Average (of the monthly averages) and 90% Confidence Levels (+/-) using the month-
month variability of vehicle weight in 2000.
Road Vehicle Class
mmt
1
90% Cl
2
90% Cl
6
90% Cl
7
90% Cl
8
90% Cl
9
90% Cl
11
90% Cl
12
90% Cl
14
90% Cl
16
90% Cl
mm
27,418
3,033
12,557
4,830
19,433
14,349
mm
3,178
236
4,927
709
7,037
208
8,310
1,832
9,191
2,281
7,705
5,043
mm
5,612
246
6,661
342
7,297
347
21,936
37,582
8,547
3,229
7,090
99
6,649
702
6,940
2,762
WMMK WMMM
26,144
2,320
24,406
713
24,374
973
24,099
2,419
20,209
886
27,045
1,667
28,985
434
26,378
2,384
24,148
2,862
mm
12,984
434
13,411
344
11,453
624
12,669
2,095
14,865
6,562
12,353
852
14,498
680
12,879
748
15,731
1,312
^^••^^H
29,535
2,912
29,562
778
29,489
845
28,161
4,705
35,917
5,386
25,460
1,000
32,030
892
31,055
3,081
29,829
3,804
mm
45,066
2,972
56,031
1,905
62,388
2,464
39,924
11,325
54,073
3,793
50,005
3,891
67,957
11,548
60,454
9,384
Bifl
29,010
1,728
29,653
1,799
32,849
2,712
28,891
6,092
11,649
21 ,290
29,246
1,750
32,903
661
33,328
4,413
31 ,662
3,433
^^Ks^^H
59,721
1,140
55,835
1,224
51 ,944
3,106
49,720
6,949
44,552
23,883
52,041
3,697
51 ,244
1,233
53,266
6,556
46,298
4,278
• [••
64,749
5,841
63,230
1,257
64,829
3,552
55,550
5,260
64,805
3,144
62,767
2,264
72,883
11,426
56,544
6,799
mm
63,789
2,046
55,763
2,384
60,004
5,898
56,285
18,050
54,119
5,070
50,454
2,524
51,000
6,839
51,487
11,275
•a
65,295
5,652
57,943
1,490
59,317
4,224
57,087
11,437
60,466
4,834
56,752
1,889
58,811
7,215
91,701
46,139
K£l
88,194
8,597
78,702
2,232
76,185
2,986
77,611
7,233
90,037
9,886
79,858
7,309
95,877
21 ,852
54,555
22,550
41
-------
7. DEVELOP NATIONAL AVERAGE WEIGHT BIN DISTRIBUTIONS BY FHWA VE-
HICLE CLASSES 1-13
As has been discussed in this report, the vehicle weight profiles vary by roadway func-
tional class. The weight profiles can also vary by time period, especially by day of week. Howev-
er, to show the typical weight profiles, the annual average weight bin fractions for vehicle classes
2-13 are shown in Figures 35-40 for the major functional classes. The smaller functional
classes (6-9 for rural and 14-16 for urban) show much more variable weight bin distributions
because there were fewer sites and fewer vehicles weighed on these roadway types.
09
08
0.7
06
0.5
04
0.3
0.2
0.1
ILLiuj
jriljftJl _JL_ .
• 01
a 02
• 11
012
• 14
Figure 35. Vehicle 2 weight bin distribution in 1999 by functional class.
42
-------
0.7
0.6
0.5
0.4
03
0.2
0 1
-HEL.
• 01
a 02
an
a 12
• 14
Figure 36. Vehicle 3 weight bin distribution in 1999 by functional class.
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
Mkj
• 01
D02
• 11
D12
• 14
Figure 37. Vehicle 4 weight bin distribution in 1999 by functional class.
43
-------
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
JJ
• 01
Q02
• 11
D12
• :
Figure 38. Vehicle 5 weight bin distribution in 1999 by functional class.
Figure 39. Vehicle 6 and 7 weight bin distribution in 1999 by functional class.
44
-------
0,9
0.8
0.6
0,5
0.4
0.3
02
0.1
01
o2
11
i2
14
• --
Figure 40. Vehicle 8-13 weight bin distribution in 1999 by functional class.
45
-------
8. VEHICLE CLASS FRACTIONS
In developing national vehicle counts/VMT/vehicle mix fraction estimates, it was under-
stood that some regional differences described in this report would be explicitly included in the
average. As noted in Section 2, the data for the vehicle mix by type showed a clear regional trend
on rural interstates but no clear trends for other roadway types. This regional trend may influence
the national average calculated here depending upon which states submitted data and the number
of sites in each state that were reported to VTRIS.
8.1 Monthly Variation
A sample of month-to-month variability and uncertainty in the vehicle mix is shown in
Figure 41 for rural interstates in 1999 and Table 16 for all road types in 1999 and 2000. Shown
by example in Figure 41, and with other road types and years, there was no clear seasonal effect
on the vehicle mix. Table 16 demonstrates that the most important vehicle classes are 2 and 3
(typically associated with light-duty vehicles) and 5 (light heavy-duty), 9 (heavy heavy-duty) and
8 (heavy heavy-duty) for heavier vehicle classes.
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
U H H
Figure 41. Month to month variability for road 1 in 1999.
46
-------
Table 16. Monthly-average annual relative vehicle class counts and monthly uncertainty.
Road \ Result Vehicle
Year
1 2 3 4 5 6 7 8 9 10 11 12 13
1 - 1 999
1 - 2000
2 - 1 999
2 - 2000
6 - 1 999
6 - 2000
7 - 1 999
7 - 2000
11 -1999
11 -2000
12-1999
12-2000
14-1999
14-2000
16-1999
16-2000
Mean
90% Cl
Mean
90% Cl
Mean
90% Cl
Mean
90% Cl
Mean
90% Cl
Mean
90% Cl
Mean
90% Cl
Mean
90% Cl
Mean
90% Cl
Mean
90% Cl
Mean
90% Cl
Mean
90% Cl
Mean
90% Cl
Mean
90% Cl
Mean
90% Cl
Mean
90% Cl
0.008
0.002
0.005
0.001
0.002
0.000
0.004
0.001
0.008
0.001
0.007
0.001
0.001
0.001
0.002
0.001
0.006
0.001
0.007
0.002
0.007
0.002
0.007
0.002
0.014
0.003
0.006
0.002
0.003
0.001
0.007
0.002
0.623
0.025
0.592
0.020
0.634
0.012
0.620
0.006
0.672
0.018
0.677
0.011
0.640
0.022
0.552
0.021
0.724
0.019
0.716
0.016
0.744
0.014
0.769
0.013
0.702
0.027
0.702
0.014
0.725
0.051
0.728
0.032
0.162
0.011
0.158
0.006
0.230
0.006
0.228
0.004
0.225
0.011
0.231
0.009
0.269
0.022
0.330
0.012
0.161
0.003
0.167
0.010
0.168
0.007
0.162
0.011
0.211
0.016
0.209
0.011
0.176
0.019
0.172
0.024
0.005
0.000
0.007
0.001
0.002
0.000
0.003
0.000
0.001
0.000
0.001
0.000
0.001
0.000
0.002
0.001
0.003
0.001
0.003
0.000
0.002
0.000
0.004
0.001
0.003
0.001
0.003
0.000
0.002
0.001
0.003
0.001
0.035
0.003
0.026
0.002
0.042
0.004
0.036
0.003
0.052
0.012
0.037
0.004
0.040
0.007
0.044
0.003
0.032
0.003
0.029
0.002
0.038
0.003
0.023
0.003
0.038
0.008
0.053
0.009
0.055
0.040
0.052
0.021
0.009
0.002
0.014
0.002
0.009
0.001
0.010
0.000
0.007
0.001
0.008
0.000
0.007
0.001
0.011
0.002
0.014
0.014
0.008
0.001
0.005
0.000
0.005
0.000
0.008
0.002
0.007
0.000
0.010
0.006
0.008
0.001
0.002
0.000
0.001
0.000
0.002
0.000
0.002
0.000
0.001
0.000
0.001
0.000
0.001
0.000
0.001
0.000
0.001
0.000
0.001
0.000
0.002
0.000
0.001
0.000
0.001
0.000
0.001
0.000
0.002
0.002
0.000
0.000
0.015
0.002
0.022
0.002
0.011
0.002
0.014
0.001
0.005
0.001
0.007
0.001
0.007
0.004
0.021
0.006
0.008
0.001
0.011
0.001
0.005
0.000
0.005
0.000
0.005
0.001
0.005
0.001
0.009
0.004
0.013
0.002
0.129
0.014
0.156
0.011
0.058
0.008
0.073
0.004
0.022
0.005
0.027
0.003
0.024
0.007
0.031
0.006
0.046
0.006
0.054
0.005
0.027
0.003
0.023
0.003
0.017
0.004
0.014
0.001
0.013
0.003
0.015
0.003
0.003
0.001
0.004
0.001
0.003
0.001
0.003
0.000
0.002
0.000
0.002
0.000
0.003
0.001
0.002
0.001
0.001
0.000
0.001
0.000
0.001
0.000
0.000
0.000
0.000
0.000
0.001
0.000
0.002
0.001
0.000
0.000
0.007
0.001
0.007
0.000
0.003
0.001
0.002
0.000
0.001
0.000
0.001
0.000
0.001
0.000
0.001
0.000
0.002
0.000
0.002
0.000
0.001
0.000
0.001
0.000
0.000
0.000
0.000
0.000
0.001
0.000
0.001
0.000
0.001
0.000
0.002
0.000
0.001
0.000
0.001
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.001
0.000
0.003
0.001
0.002
0.001
0.004
0.002
0.002
0.001
0.002
0.000
0.005
0.002
0.002
0.001
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.002
0.001
0.000
0.000
47
-------
8.2 Hourly Variation Over an Average Week
The vehicle class counts show clear differences by hour of day and day of week, espe-
cially distinguishing between weekdays and weekend days. Figures 42 and 43 show the hourly
change in the vehicle mix throughout the average weekly activity.
In these figures, it can be seen that Sunday (first day of week), Saturday, and weekdays
are clearly different from one another by comparing the fraction of vehicle 9 (in bold green) and
Vehicle 8 (bold blue). Differences between each weekday are less clear, but indicate that each
weekday could also be considered a unique day.
The hourly change in vehicle mix is more dramatic with an overnight and secondary mid-
day peak in the mix of heavier heavy-duty vehicles (primarily vehicle classes 8, 9, and 11) higher
than the average daily fraction.
48
-------
Vehicle 1
Vehicle 2
'Vehicles
— Vehicle 4
Vehicle 5
Vehicle 6
Vehicle 7
Vehicle 8
Vehicle 9
Vehicle 10
Vehicle 11
Vehicle 12
Vehicle 13
25
49
73
97
121
145
Figure 42. 1999 Hourly average-week vehicle fractions on road type 1, rural interstates.
Vehicle 1
Vehicle 2
Vehicles
Vehicle 4
Vehicles
Vehicles
Vehicle 7
Vehicle
Vehicle 9
Vehicle 10
Vehicle 11
Vehicle 12
73
97
121
145
Figure 43. 1999 Hourly average-week vehicle fractions on road type 11, urban interstates.
49
-------
9. TRAFFIC VOLUME TRENDS
The Traffic Volume Trends data for 2000 were ported into MySQL and summary data
were provided to EPA in Access® files. The data for 1999 were not available. A sample of the
2000 results for rural and urban road types is shown in Figures 44 and 45. The TVT data results
shown in these graphs provide a consistent understanding of the typical hourly traffic profiles.
As the road types move to lower traffic volumes, the hourly profile maintains a similar shape but
lower in magnitude.
49
73
97
121
145
Hour
Figure 44. Typical national average weekly total traffic volume for rural roads.
50
-------
4,000
3,500
145
Figure 45. Typical national average weekly total traffic volume for urban roads.
As with the vehicle class count data, the total volume counts were averaged at individual
sites before being averaged across sites. The steps followed in processing the class count data
were the same as the VTRIS vehicle class counts data:
1. All counts across lanes in the same roadway direction were totaled. Different directions at
site were treated separately.
2. All counts (either total volume or count for each vehicle class) were averaged for each site-
direction pair by hour, day of week (i.e., Sunday through Saturday), month, and roadway
classification. This means that at most five values were averaged together, corresponding
to the total number of days in a week during one month. In other words, all Monday counts
during January for hour 10 were averaged together at each site-direction pair.
3. The hourly class counts were averaged across the sites. These averages were calculated by
roadway function class, month, day of week, and hour of the day.
EPA requested that temporal profiles be provided for four aggregate road types in addition to
the more numerous types shown in Figures 44 and 45. These aggregate temporal profiles are pro-
vided in Appendix C.
In addition, Appendix D provides an analysis of the regional variability of the TVT total
volume temporal profiles and a comparison of the temporal trends between the TVT and VTRIS
data. A regional pattern of the temporal profiles could not be found in general, however individ-
ual States had significant differences for some road types and temporal profiles. The VTRIS total
volume counts exhibited similar trends to the TVT profiles indicating that VTRIS vehicle mix
could be used in concert with TVT total volume estimates.
51
-------
10. CONCLUSIONS AND RECOMMENDATIONS
This work demonstrates that the VTRIS and TVT data can be imported into standard da-
tabase programming tools that can be used to generate averages and typical temporal or regional
profiles useful for emissions modeling. The summary results presented in this report can inform
EPA of vehicle characteristics, weight and class fractions of the in-use fleet.
The results of this work were provided with this report as a series of database and Excel
summary files, as it was impossible to present all the summary results in a reasonable length re-
port. The summaries presented in this report were selected to provide EPA a flavor of the results.
The more important conclusions of this work are:
• There were no clear regions that could be defined with similar vehicle weight profiles, but
there were indications of a regional effect on the vehicle fleet mix. state specific summaries
were produced to further investigate regional effects on both vehicle weight distributions
and vehicle mix.
• Temporal profiles of weight or weight distribution indicate that month of year had little
effect, while the day of week and in some cases the hour of day had a noticeable effect
on the average vehicle weight and weight distribution. The temporal profile of vehicle
mix was more dramatic showing clear diurnal and weekly profiles especially of the larger
heavy-duty vehicle fractions. The temporal profiles of the vehicle mix will have an effect
on modeled emissions because heavy-duty vehicles typically emitNOx and PM emissions
at much higher rates than light-duty vehicles, so overall emission estimates will be sensi-
tive to these temporal profiles of the vehicle mix.
• The road type, especially urban or rural, has an effect on all elements described in this
report. The road type where vehicle weight was measured can affect the average weight
for some vehicle classes, but the overall and temporal profiles of the vehicle mix and total
traffic volume were more clearly affected by the road type measured. Heavy-duty vehicle
mix tends to be highest with rural road types and higher traffic volume roadways. The total
traffic volume profiles were more sensitive to time of day on higher traffic volume road-
ways.
• Vehicle weights for the smaller vehicle classes, 2 and 3, seem unreasonably high and may
need to be ignored. It is unclear whether the vehicle weight measurements for the lighter
vehicles were affected by the detection limits of the measurement method or had been
calibrated only for the higher weights of heavy-duty vehicles. In either case the lighter
vehicles had average weight readings of up to twice the gross vehicle weight rating for the
vehicles supposedly measured.
• The results in this work suggest that vehicle grouping be reinvestigated or that no group-
ings be made maintaining as much specificity as is provided in the FHWA vehicle classifi-
52
-------
cations. Vehicle classes 6 and 8 had similar average vehicle weights while vehicle class 7
was more typical of vehicle classes 9-13. Therefore the typically used vehicle groupings
of (8 - 13) and (5 - 7) were not consistent with the vehicle classes found.
There are several potential areas for future work. While for the most part the data from
1999 and 2000 provided a consistent understanding of the regional and temporal profiles, addi-
tional years of VTRIS data might be evaluated to provide a more robust understanding of typical
weight and vehicle mix profiles. From analysis of traffic data that ENVIRON has performed for
several states, we know that states do not always submit all their data to VTRIS, so additional
data can be gathered directly from the state agencies, especially for states not included in VTRIS
to fill in missing regions.
VTRIS and TVT are part of the Heavy Vehicle Travel Information System (HVTIS),
which is a data collection system authorized by the Office of Management and Budget (OMB).
States must have traffic monitoring systems, and FHWA requests a copy of some of their data
reducing the added burden when submitting data to providing it in the requested format. The
traffic data received is only from state DOTs and not county or municipal transportation depart-
ments. As has been demonstrated and explained in this report, missing data is the primary cause
for variability and discontinuous trends, and there are various reasons why a state may have
missing data including equipment malfunction, budget limits, or unwillingness to submit it to
VTRIS. The Weigh-in-Motion (WIM) systems have the most problems and states often struggle
to keep them operating, especially when the systems have endured the harsh environment of the
roadway after a few years. FHWA encourages the states to edit data before submitting to VTRIS
or TVT, so that they are comfortable with publishing the data. Both TVT and VTRIS entry sys-
tems perform basic data edits. However, individual data entries by station or other delineation
may be peculiar compared with national, state, or metropolitan aggregates. These outliers may
be unique situations and influence the average while being valid data. Outlier identification and
review were performed to the extent possible within the scope of the current work, although ad-
ditional work on such data anomalies is recommended.
Another area relevant for current modeling is to better cross-reference the FHWA and
MOBILE vehicle classes. The vehicle classification data presented in this report provide field
verification of national averages and better delineation of vehicle travel patterns by road type and
region. These results have potential importance in current emissions work both for overall road
type mix and temporal profiles. The results of this study point out many of the failings of current
cross-referencing methods described in Appendix B.
53
-------
REFERENCES
EPA. (2004), personal communication with Dave Brzezinski, U.S. Environmental Protection
Agency, September, 2004, modeling tool to estimate national average VMT mix by in-
use month and year.
EPA. (2003), personal communication from Megan Beardsley, U.S. Environmental Protection
Agency, September 2003.
Gillman, Ralph. (2005), Acting Chief, Travel Monitoring and Surveys Division Federal Highway
Administration, personal communication, September 29, 2005.
TCEQ. (2004), Houston-Galveston State Implementation Plan. [TTI (2003), "2000 On-Road
Mobile Source Episode-Specific Emission Inventories for the Houston/Galveston
Ozone Nonattainment Area, prepared for the Texas Council on Environmental Quality
by the Texas Transportation Institute, August, 2003.]
Yoon, S., Zhang, P., Pearson, J.R., Guensler, R.L., and Rodgers, M.O. (2004). "AHeavy-Duty
Vehicle Visual Classification Scheme: Heavy-Duty Vehicle Reel as sifi cation Method for
Mobile Source Emissions Inventory Development," Presentation at the 2004 AWMA.
54
-------
APPENDIX A - QUALITY ASSURANCE/QUALITY
CONTROL EVALUATION AND REVIEW
A-l
-------
A-l. PROJECT OBJECTIVES, ORGANIZATION, AND RESPONSIBILITIES
A-l.l Purpose of Study
Heavy-duty vehicles have been seen as contributing a large fraction of emissions from
on-road vehicles and are coming under more intense scrutiny because light-duty emissions have
been controlled to a greater extent than heavy-duty vehicle emissions. A heavy-duty vehicle can
produce 10 to 100 times the emissions (of NOX and PM emissions especially) of a light-duty ve-
hicle. Thus, heavy-duty vehicle activity needs to be better characterized. Key uncertainties with
the use of MOBILE6 regarding heavy-duty vehicle emissions include the fraction of heavy-duty
vehicles on all types of roadways at all times of day. In addition, there may regional variability in
both the fraction of different vehicle classes and the vehicle weights within each class.
With the MEASURE1 model and the developing MOVES2 model (eventual replacement
to MOBILE3), greater emphasis is given to physical parameters affecting the engine loads and
therefore the emissions from individual vehicles. One primary factor affecting the engine load is
the vehicle weight, so the weight of the vehicle on the road is needed to estimate the in-use emis-
sions of given vehicles. Because the effect of vehicle weight may be nonlinear for certain types
of driving, the weight distribution of vehicles is useful knowledge to incorporate into emission
estimates.
Databases collected by the Federal Highway Administration (FHWA) include vehicle
count and classification from the Highway Performance Monitoring System (HPMS) which was
automated traffic recorders (ATR) used to produce the Travel Volume Trends (TVT) reports.
Other data sets hold the results of data collection from weigh-in-motion (WEVI) sensors, and
other data sources (visual observation, weigh stations, and other special projects) maintained by
the FHWA and compiled in the Vehicle Travel Information System (VTRIS).
This report investigates and evaluates these databases to assist in the development of im-
proved emissions estimates of heavy-duty vehicles. The goal of the project was therefore to pro-
duce estimates of the ratio of heavy-duty vehicles to all vehicle traffic, and weight distributions
'MEASURE = Mobile Emissions Assessment System for Urban and Regional Evaluation. Model. This model is a
prototype GIS-based modal emissions model.
2MOVES = Mobile Vehicle Emissions Estimator, next generation mobile source emissions model. The model will be
used for State Implementation Plan emission inventories and will replace the current MOBILE model.
3MOBILE = Current mobile source emissions model used for State Implementation Plan emission inventories.
A-2
-------
for heavy-duty vehicles according to the time of day, day of week, and other temporal variables
as well as investigating regional differences.
A-1.2 Project Objectives
The objectives of this project were to:
A. Generate detailed weight distributions of FHWA classes 1-13 individually
B. Develop regional aggregations of data.
C. Analyze the regional variability in weight distributions by road type.
D. Analyze the temporal variability in the weight distributions by road type.
E. Develop national average weight distributions for FHWA classes 1-13 individually for
each road type.
F. Develop national average temporal distributions for FHWA classes 1-13 individually
for each road type.
G. Develop VMT fractions for the FHWA classes, especially groups 1 thru 4 individually,
5-7 as a group, and 8-13 as a group. These fractions will be by month, day of week,
and hourly.
H. Report the results of uncertainty analysis on A through G above.
I. Evaluate the Traffic Volume Trends dataset for the most appropriate platform for its
incorporation into the analysis for task G above.
National or state total vehicle miles traveled (VMT) estimates have historically been pro-
vided by the Federal Highway Administration (FHWA) as annual or average daily totals or other
similar general estimates for all vehicle classes together. In order to properly use these estimates
to estimate emissions for air quality planning, temporal adjustments and the vehicle class frac-
tional mix must be determined. The TVT data were used to provide the temporal adjustments of
the total (all vehicle classes summed together) VMT because TVT data has been used by FHWA
to provide the VMT estimates. The vehicle mix was determined using the VTRIS data. In addi-
tion, VTRIS also provides vehicle weight data useful as input data for future estimation tools that
EPA is developing.
The TVT and VTRIS databases include vehicle counts (by class and weight in VTRIS)
for a number of sites across the country defined by roadway type and provided by month, day,
and hour. These two databases do not consist of identical sites, so the temporal distribution of the
total VMT was determined separately from the temporal variability in the vehicle mix or vehicle
weight distribution. Roadway functional classes (type of roadway such as interstate, arterial, col-
A-3
-------
lector, etc. and either rural or urban) are standard FHWA road type designations and were used to
associate the temporal trends of total VMT and the vehicle mix and weight estimates.
Several statistical procedures were used to estimate uncertainties in the final aggregate
national estimates of vehicle weight distributions and vehicle class fractions (Objective H.). Sta-
tistical hypothesis testing procedures are also available for evaluating the statistical significance
of regional and temporal differences in these distributions. Uncertainties in individual class or
weight fraction estimates obtained from aggregated data (for example, national estimates of the
Class 5 vehicle travel fraction) were obtained by treating the data as binomial and computing the
standard error of the sample estimate of the binomial probability. This process was repeated for
each vehicle class of interest. In addition, significance tests were applied to determine if vehicle
class or weight distributions differ by facility type or by time of day (or day of week). These
significance tests were based on the chi-square statistic computed from contingency tables of
vehicle counts such as tables of counts by vehicle class and hour of day.
The results of this study were provided along with the accompanying uncertainty analy-
sis in files along with the final report. The databases used in the study are large and have been
provided on a hard drive to which the final report, results, and uncertainty files can be added. The
summary files primarily provided in Access® database tables were also large, but provided aggre-
gate results by the spatial and temporal delineations requested in the work assignments.
A-1.3 Secondary Data Required by the Project
The secondary data used in this project was all the available data contained in the Fed-
eral Highway Administration's (FHWA) Vehicle Travel Information System (VTRIS) dataset for
the years 1999 and 2000. This data was previously evaluated for its usefulness and analyzed in
Phase I of this work. This work assignment extended the analysis of the 1999 and 2000 VTRIS
data and provided national and state average activity with the combined datasets.
The other dataset was the Traffic Volume Trends (TVT), also maintained by FHWA. The
TVT data was used to generate estimates of VMT fractions under item (G) in section 1.2 above.
A-1.4 Approach for Evaluating Project Objectives
In Phase I of this work, the raw 1999 and 2000 VTRIS was incorporated into two Micro-
soft SQL Server databases (one containing the 1999 VTRIS data, the other for the 2000 VTRIS
Data). Algorithms (discussed at length under "QA Procedures" below) were developed for re-
ducing the dataset to include only those data deemed appropriate for this analysis. Starting from
A-4
-------
these "cleaned" versions of the datasets, the analyses of elements (A) through (H) of the project
objective were conducted.
The Traffic Volume Trends (TVT) data has been examined. Only monthly summaries
of the 1999 TVT data was available, so only the 2000 TVT data was used. For the 2000 data, a
QA procedure that FHWA applies to this type of data and detailed in Section 3.1 was followed.
There were additional restrictions required that were placed on the 2000 TVT data clearly docu-
mented in this appendix. Because only one year's worth of data was available for TVT, specific
emphasis was placed on evaluating whether sufficient data was available to provide aggregate
results that could be widely applied.
A-5
-------
A-2. SOURCES OF SECONDARY DATA
A-2.1 Sources of Secondary Data Used
There are two datasets used in this project, namely the VTRIS and TVT datasets, both
compiled and maintained by FHWA.
VTRIS and TVT are part of the Heavy Vehicle Travel Information System (HVTIS),
which is a data collection system authorized by OMB. States must have traffic monitoring sys-
tems, and FHWA asks for a copy of some of their data in the Traffic Monitoring Guide (TMG)
format. The traffic data is only from state DOTs.
The automatic traffic recorders (ATR) that generate data for TVT and the automatic
vehicle classifiers (AVC) and weigh-in-motion (WEVI) systems that generate data for VTRIS are
usually at different locations. Since traffic volume is also an output of AVC, AVC sites provide
traffic volume to the TVT database as well. FHWA is beginning the process of merging TVT and
VTRIS software into the new TMAS (Travel Monitoring Analysis System) and more AVC sites
and data are expected in the future, so there will be greater overlap between VTRIS and TVT
databases (Ralph Gillman, FHWA, 2005).
States often have missing entries, and there are various reasons for this. Individual detec-
tors may not be working, communications with detectors may have broken down, their data pro-
cessing software may not be working correctly, they may be in the process of changing data pro-
cessing software or operating systems, they may have difficulty providing data in TMG format,
they may be short on staff, or they may be late, etc. The WEVI systems have the most problems
and states often struggle to keep them operating, especially when the systems have endured the
harsh environment of the roadway a few years.
FHWA encourages the states to edit their data before submitting it to VTRIS or TVT, so
that they are comfortable with publishing the data. Both TVT and VTRIS entry systems perform
basic data edits. However, individual data entries by station or other delineation may be peculiar
compared with national, state, or metropolitan aggregates. These outliers may be unique situa-
A-6
-------
tions and influence the average while be valid data. Such outlier identification and review were
beyond the scope of the current work, however the data aggregations performed in this study
identified suspect data.
A-2.2 Rationale for Selecting Data Sources
The VTRIS dataset was selected for this analysis because of its size and content. There
were roughly 30 states with data in VTRIS for 1999 and 2000 with hourly vehicle classifica-
tion data from numerous sites within each state, so it is a large dataset. The VTRIS dataset also
contains weigh-in-motion data, which consists of the vehicle weights of all vehicles passing over
a roadway for a period of time. This weight data was used to help determine the distributions of
vehicle weights, in particular, the distributions of vehicle weights for heavy-duty classes. In ad-
dition to the weigh-in-motion data, VTRIS contains class count data by hour of the day (usually
collected using automatic traffic counters). The class count data was used to generate temporal
profiles for all the FHWA classes.
The TVT dataset was selected to provide estimates of vehicle miles traveled. When com-
bined with the class count data, it is possible to estimate VMT by vehicle class, hour of day, and
roadway type.
A-7
-------
A-3. QUALITY OF SECONDARY DATA
A-3.1 Quality Requirements of Secondary Data
FHWAhas published guidance that describes requirements for data collection for the two
sets of secondary data used in this work: Vehicle Travel Information System (VTRIS) and the
Travel Volume Trends (TVT). For VTRIS, FHWA requires that State Departments of Transpor-
tation follow the Traffic Monitoring Guide (FHWA-PL-01-021, http://www.fhwa.dot.gov/ohim/
tmguide/index.htm) for which requirements for collecting class count information is detailed in
Section 4 and vehicle weight information in Section 5. This guide includes a lengthy description
of the data collection requirements including data collection equipment, site selection, sampling
periods, and other data handling procedures used in the compilation of this data set. As the data
is input into the VTRIS system, FHWA also describes in a manual (http://www.fhwa.dot.gov/
ohim/ohimvtis.htm) the requirements of data and how the data is handled by the VTRIS sys-
tem. Likewise the TVT data is a compilation of the Highway Performance Monitoring System
(HPMS) and must follow the HPMS Field Manual (http://www.fhwa.dot.gov/ohim/hpmsmanl/
hpms.htm). This guidance has detailed descriptions of the site selection, sampling procedures,
data collection and verification, reporting and data handling.
Under this work, additional quality assurance checks were applied to the data to find and
eliminate spurious data. The additional quality assurance requirements of the VTRIS data were
determined during Phase I of this work. They are listed briefly below:
• All site identification fields were required to have a match in the detailed site informa-
tion table so that the observation could be properly placed.
• For the vehicle class count data, a record was not used if the percent of unknown
vehicles (classes 14 and 15) contained more than 2% of the observed counts for that
hour.
• For the vehicle class counts, only days with all 24 hours measured were included.
• For the vehicle class counts data, data was only used if all lanes in a direction were
measured to reduce bias by heavier vehicles tending to travel in the right most lanes.
• For the weight data, if the sum of the axle weights differed from the total vehicle
weight by more than 5%, the data was thrown out.
A-8
-------
As part of this work, for the TVT data, the QA requirements were further refined during
the course of investigating the database structure and performing the uncertainty analysis. These
requirements include those implemented by FHWA, which are as follows:
• For each month, a station must have at least one valid day of observations for each day
of the week. If there is a day of the week with no valid observations for that month,
that station's observations are dropped for that month.
• Records with more than seven consecutive hours of zero traffic volumes are dropped.
• All days must have 24 hours of valid volume counts to be considered valid.
• An hourly volume count of zero is considered invalid if an adj acent hour has a count
greater than 50.
A-3.2 QA Procedures
The quality assurance requirements and procedures of the VTRIS and TVT data were
developed under Phases I and II of this work. Only the "clean" VTRIS and TVT data were used
in subsequent phases of the work. The methods for "cleaning" the VTRIS data are described
below.
A-3.2.1 VTRIS Data
The Vehicle Travel Information System (VTRIS) is a database management system writ-
ten in FoxPro 2.6a for Windows. It is maintained by the FHWA to house vehicle travel charac-
teristic data. It is designed to import, edit, and summarize data.
All the VTRIS data analyzed in this project are those that were already imported into the
VTRIS system. As part of this work effort, we exported all the VTRIS data out of VTRIS and
into two Microsoft SQL Server databases (one for the year 1999 and one for 2000).
The VTRIS data we analyzed had already undergone some quality assurance checks upon
data import into VTRIS. The ASCII files that were loaded in to VTRIS were validated by the
VTRIS program as follows :
1. Determination of the record's type according to FHWA formats set for coding of STA-
TION, CLASSIFICATION and WEIGHT data. The data type, record length, and other
record parameters were also checked.
2. Validation of single fields within a record to ensure that they hold a valid field value
or are within a specific range of values. In addition, cross-validation within a single
A-9
-------
record between two or more fields was done to ensure that data fields were not contra-
dicting each other.
3. Checks for duplicates and consistency between the records that are being loaded into
the same table. This mostly concerns STATION and CLASSIFICATION data since
there should not be duplicate records with the same key value. For WEIGHT data it
is not a validation issue, but rather a matter of data maintenance since the specified
key may identify an unlimited number of records. This corresponds to the fact that the
table contains one record per truck measured.
4. Cross-validation between the fields of the new records and the records from the
VTRIS table to prevent duplicates and support referential integrity between different
VTRIS tables as well as consistency within a single table. The integrity requires that
the CLASSIFICATION and WEIGHT data checks against station data to make sure
that the key is valid (e.g. Station-Direction-Lane exist).
The VTRIS User's Manual further lists error levels upon import of the data:
1. Junk - those records that are detected at the earliest stage of validation and result in
the record being put into the JUNK file. No further validation is possible for these
records until some manual editing is done.
2. Fatal - those records that cannot be admitted "as is" even if User would like them to.
For those errors, an appropriate correction through the ERROR table Browse/Edit fa-
cility is required. Those are typically errors in the key fields and other very significant
fields that would violate consistency and referential integrity.
3. Caution - those errors that can be fixed or can be flagged by User as acceptable and
put into the VTRIS tables "as is". If User accepts and flags them, an appropriate Flag
Code will be placed into a VTRIS table along with the record.
Records that are classified by VTRIS as "junk" or "fatal" are automatically rejected by
VTRIS. This means that there were no "junk" or "fatal" records in the data that we exported for
this project.
The data flags were not exported by VTRIS. This means that data assigned a "caution"
flag were included in our analysis. In order to capture some of these records that VTRIS would
have flagged, ENVIRON duplicated some of the error checking in the VTRIS program using the
default data ranges and error margins in VTRIS.
What follows below is a discussion of the quality assurance methods, including the
removal of data due to criteria developed during this work. The calculations performed and the
QA/QC decisions were programmed in Microsoft SQL, and the text for these scripts was pro-
vided at the end of phase I and phase II with the ".sql" extension.
Vehicle Class Counts
A-10
-------
The first step in processing the vehicle class counts data was to remove bad data. A bad
record was one for which one of the unknown vehicle classes (14 or 15) contained more than 2%
of the observed counts for that hour. This criterion was set based on the default VTRIS program
configuration.
Records were eliminated where the site identification in the class counts data had no
match in the site table making it impossible to place the data into a state and county or identify
the roadway type. The site table field labeled "Method of Vehicle Classification" included codes
to identify two types of automatic identification and one type labeled "Human Observation." The
records taken by human observation were eliminated to avoid the inclusion of subjective data
into the final data set.
The next step was to find all stations and days for which all lanes in a direction were mea-
sured, and all 24 hours were observed for each of the lanes. This QA/QC criterion was to elimi-
nate records where only partial data was available for a road type. For road types with multiple
lanes, there may be significant differences in vehicle types between these lanes such as where
heavy-duty vehicles may be required to preferentially use the right most lanes on freeways. Also,
daytime or other partial diurnal measurements were taken possibly biasing the hour-of-day esti-
mates and making it unable to be scaled for relative day-of-week activity.
A check of the vehicle class counts against VTRIS default maximums and minimums was
explicitly not performed because EPA did not want to eliminate values that would have qualified
as outliers, but that might be entirely accurate.
Initially we proposed to use only data where all 7 days of the week were measured. But
this criterion would have eliminated most of the data available. The intention of the vehicle class
counts data was to provide estimates of the vehicle class activity by one vehicle class relative to
another rather than the total vehicle counts between days.
The number of class count records in the analysis was reduced from 3,130,642 to 824,112
(or 34,338 complete days) in 1999, from 4,070,127 to 1,468,200 (or 61,175 complete days) in
2000. For all the remaining data, the class counts were summed across all lanes in a given direc-
tion, and then they were averaged for different combinations of state FIPs, roadway function
class, county FIPs, month, hour, and day of week. For the day of week, 1 = Sunday, 7 = Satur-
day.
Vehicle Weights
A-ll
-------
Upon importing the weight data, bad data was removed and stored in tables with the
suffix "_bad." These records were ones for which the sum of the axle weights differed by more
than 5% from the total vehicle weight. This criteria was not part of the VTRIS program, but was
implemented on our part to ensure data integrity. As with the vehicle class counts tables, records
were eliminated where the site identification in the weight data had no match in the site table,
making it impossible to place the data into the state and county or identify the roadway type.
There were 50,559,506 weight records for the 1999 dataset, and 1,197,410 were dropped. There
were 69,910,356 weight records for the 2000 dataset, and 3,029,156 were dropped.
A-3.2.2 Travel Volume Trends (TVT) Data
The data we received was 2000 TVT data in the standard format prescribed by the "Traf-
fic Monitoring Guide" published by the FHWA at:
http: //www. fhwa. dot, gov/ohim/tm gui de/tm g6. htm
The quality assurance we implemented was that recommended by FHWA staff as follows:
• For each month, a station must have at least one valid day of observations for each day
of the week. If there is a day of the week with no valid observations for that month,
that station's observations are dropped for that month.
• Records with more than seven consecutive hours of zero traffic volumes are dropped.
• All days must have 24 hours of valid volume counts to be considered valid.
• An hourly volume count of zero is considered invalid if an adjacent hour has a count
greater than 50.
Out of 1,926,976 records, less than 1% of records were dropped. The remaining number
of records analyzed was 1,922,822 (where one record contained all 24 hours of observations).
After summing the volumes across all lanes in one direction, there were 1,872,708 days of obser-
vations.
A-3.3 Data Representativeness
Data representativeness was defined for this work to determine the data coverage, in
terms of the number of states represented by the road types. Five major road types were respon-
sible for 99% of the VMT (according to the FHWA Highway Statistics publication, http://www.
fhwa.dot.gov/policy/ohpi/hss/index.htm. Though the proportion of VMT by road type varies by
A-12
-------
local metropolitan and state analysis methods where other road types may responsible for more
VMT), so the analysis of the data coverage was described for these major road types to simplify
the presentation. For both vehicle mix (the fraction of vehicles by class) and weight (for the
heavy-duty vehicle types), the data was sampled from states and road types covering approxi-
mately 50% of the national VMT.
A-13
-------
Table A-l. TVT data coverage by state and important facility types.
Rural Rural Urban Urban Other Urban
State Interstates Arterial Interstates Freeways Arterial
2000 2000 2000 2000 2000
Alaska
Alabama
Arkansas
Arizona
California
Colorado
Connecticut
District Of Columbia
Delaware
Florida
Georgia
Hawaii
Iowa
Idaho
Illinois
Indiana
Kansas
Kentucky
Louisiana
Massachusetts
Maryland
Maine
Michigan
Minnesota
Missouri
Mississippi
Montana
North Carolina
North Dakota
Nebraska
New Hampshire
New Jersey
New Mexico
Nevada
New York
Ohio
Oklahoma
Oregon
Pennsylvania
Rhode Island
South Carolina
South Dakota
1
1
1
0
1
0
1
0
0
1
1
0
1
1
1
1
1
1
1
1
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
0
1
0
1
0
0
1
1
1
1
1
1
1
1
1
1
1
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
0
1
0
1
0
0
1
1
1
1
1
1
1
1
1
1
1
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
0
0
1
0
1
0
1
0
0
1
1
1
0
0
0
1
1
1
0
1
0
0
1
1
1
1
0
1
0
0
1
1
0
1
1
1
0
1
1
1
0
0
1
1
1
0
1
0
1
0
0
1
1
1
1
1
1
1
1
1
1
1
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
0
1
1
1
1
1
A-14
-------
Rural Rural Urban Urban Other Urban
State Interstates Arterial Interstates Freeways Arterial
2000 2000 2000 2000 2000
Tennessee
Texas
Utah
Virginia
Vermont
Washington
Wisconsin
West Virginia
Wyoming
VTRIS VMT Data Coverage by Facility*
National VMT by Facility*
1
1
1
1
1
1
1
1
1
93%
22%
1
1
1
1
1
1
1
1
1
94%
18%
1
1
1
1
1
1
1
1
1
94%
32%
0
1
0
1
1
1
1
1
0
89%
13%
1
1
1
1
1
1
1
1
1
93%
13%
The VMT by state and facility were provided by FHWA Highway Statistics (http://www.fhwa.
dot.gov/policy/ohpi/hss/index.htm) for 2000.
1 = Data present.
0 = Data absent.
The data coverage with the VTRIS data included only some states as shown in Tables
A-2 and A-3. Little regional variability beyond that for rural interstates was described in the
main report, so the lack of data from the states not represented would not be expected to bias
the national average results much. For instance, the eastern seaboard states could be considered
a region with lower truck traffic for rural interstate roads as described in Section 2 of the report
because the states are located outside of the main freight corridors. The mid-Atlantic states of
Maryland, Delaware, and Virginia and the north Atlantic states of New York, Massachusettes,
Maine and New Hampshire were entirely missing from the database, but these states comprise
13% of the national VMT and only 10% of the rural interstate VMT. So while there is a potential
that the missing data in the VTRIS 1999 and 2000 database may bias the national averages, any
bias would be relatively minor.
A-15
-------
Table A-2. VTRIS data coverage for vehicle mix by state and important facility types.
Rural _ ... . . Urban Urban Other Urban
. . . . Rural Arterial ,,.,.,. r- „_,. - ,
State Name Interstate Interstates Freeways Arterial
1999 2000 1999 2000 1999 2000 1999 2000 1999 2000
Alaska
Alabama
Arkansas
Arizona
California
Colorado
Connecticut
District Of Columbia
Delaware
Florida
Georgia
Hawaii
Iowa
Idaho
Illinois
Indiana
Kansas
Kentucky
Louisiana
Massachusetts
Maryland
Maine
Michigan
Minnesota
Missouri
Mississippi
Montana
North Carolina
North Dakota
Nebraska
New Hampshire
New Jersey
New Mexico
Nevada
New York
Ohio
Oklahoma
Oregon
Pennsylvania
Rhode Island
South Carolina
South Dakota
0
0
1
0
1
0
0
0
0
1
0
0
1
0
0
0
1
1
0
0
0
0
1
0
0
1
0
1
0
0
0
1
0
0
0
1
0
0
1
1
0
0
0
0
1
0
1
0
0
0
0
0
1
0
1
0
0
0
1
0
1
0
0
0
1
0
1
1
1
0
0
0
0
1
1
0
0
0
1
0
1
1
0
0
0
0
1
0
1
0
1
0
0
1
0
0
1
0
0
0
1
1
1
0
0
1
1
0
0
1
0
1
0
1
0
1
0
1
0
0
1
0
1
1
0
1
0
0
1
0
1
1
1
0
0
0
1
0
1
0
0
0
1
0
1
0
0
0
1
0
1
1
1
1
0
0
0
1
0
1
0
1
1
0
1
1
0
1
0
0
1
0
1
0
0
0
0
1
0
0
0
0
0
0
1
1
0
0
0
0
1
0
0
1
0
0
0
0
0
1
0
0
0
0
0
0
0
1
0
0
0
0
1
0
1
0
1
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
1
0
0
0
0
0
0
1
0
0
0
1
1
0
0
1
0
0
0
0
1
0
1
0
0
0
0
1
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
1
0
0
0
1
0
0
0
0
1
0
1
1
0
0
0
0
1
0
0
0
0
0
1
0
0
0
0
0
0
0
1
0
0
0
0
0
0
1
0
0
0
1
1
0
0
1
0
0
0
0
1
0
0
0
0
0
0
1
0
0
0
0
0
0
1
1
1
0
0
0
1
0
0
1
0
0
0
1
0
1
0
1
0
0
0
0
1
0
0
0
0
0
1
0
0
0
1
0
0
0
0
0
0
0
0
0
1
0
1
0
0
0
0
0
1
1
0
0
0
0
0
1
0
0
0
0
1
0
1
0
0
0
A-16
-------
Rural Rural Arterial Urban Urban Other Urban
State Name Interstate Interstates Freeways Arterial
1999 2000 1999 2000 1999 2000 1999 2000 1999 2000
Tennessee
Texas
Utah
Virginia
Vermont
Washington
Wisconsin
West Virginia
Wyoming
VTRIS VMT Data Cover-
age by Facility
National VMT by Facility*
0
0
0
0
0
1
1
0
1
0
0
0
0
0
0
1
0
1
52%
22%
0
0
0
0
0
1
1
0
1
0
0
0
0
0
1
1
1
1
62%
18%
0
0
0
0
0
1
1
0
1
0
0
0
0
0
1
1
0
0
46%
32%
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
1
0
50%
13%
0
0
0
0
0
0
1
0
0
0
0
0
0
0
1
1
0
0
34%
13%
The VMT by state and facility were provided by FHWA Highway Statistics (http://www.fhwa.dot.gov/policy/ohpi/hss/
index.htm) for 2000.
1 = Data present.
0 = Data absent.
Table A-3. VTRIS data coverage for vehicle weight by state and important facility types.
Rural Rural Urban Urban Other Urban
State Name Interstate Arterial Interstates Freeways Arterial
1999 2000 1999 2000 1999 2000 1999 2000 1999 2000
Alaska
Alabama
Arkansas
Arizona
California
Colorado
Connecticut
District Of Columbia
Delaware
Florida
Georgia
Hawaii
Iowa
Idaho
Illinois
Indiana
Kansas
Kentucky
Louisiana
Massachusetts
Maryland
0
0
1
0
1
0
1
0
0
1
0
0
1
1
0
1
1
1
0
0
0
0
0
1
0
1
0
1
0
0
0
0
0
1
1
0
1
1
0
1
0
0
0
0
1
0
1
0
1
0
0
1
0
0
1
1
0
1
1
1
1
0
0
0
0
1
0
1
1
1
0
0
0
0
0
1
1
0
1
1
0
1
0
0
0
0
1
0
1
0
1
0
0
1
0
0
0
0
0
1
1
1
0
0
0
0
0
1
0
1
1
1
0
0
0
0
0
0
0
0
1
1
0
0
0
0
0
0
1
0
1
0
1
0
0
1
0
0
0
0
0
0
1
1
0
0
0
0
0
1
0
1
1
1
0
0
0
1
0
0
0
0
0
1
0
0
0
0
0
0
1
0
0
0
1
0
0
1
0
0
0
0
0
1
1
1
1
0
0
0
0
1
0
0
0
1
0
0
0
0
0
0
0
0
0
1
0
0
0
0
A-17
-------
Rural Rural Urban Urban Other Urban
State Name Interstate Arterial Interstates Freeways Arterial
1999 2000 1999 2000 1999 2000 1999 2000 1999 2000
Maine
Michigan
Minnesota
Missouri
Mississippi
Montana
North Carolina
North Dakota
Nebraska
New Hampshire
New Jersey
New Mexico
Nevada
New York
Ohio
Oklahoma
Oregon
Pennsylvania
Rhode Island
South Carolina
South Dakota
Tennessee
Texas
Utah
Virginia
Vermont
Washington
Wisconsin
West Virginia
Wyoming
VTRIS VMT Data
Coverage by Facility*
National VMT by
Facility*
0
1
1
1
1
0
1
0
1
0
1
1
1
0
1
0
0
1
1
1
1
0
1
0
1
0
1
1
0
1
0
1
0
1
1
1
1
0
0
0
1
1
1
0
0
0
0
1
1
1
1
0
1
1
0
0
1
1
0
1
69%
22%
1
1
1
1
1
0
1
0
1
0
1
1
1
0
0
0
0
1
1
1
1
0
1
0
1
0
1
1
0
1
0
1
0
0
1
1
1
0
0
0
1
0
1
0
0
0
0
1
1
0
1
0
1
0
0
0
1
1
1
1
73%
18%
0
1
1
1
1
0
0
0
1
0
1
0
1
0
1
0
0
0
1
0
0
0
0
0
1
0
1
0
0
1
0
0
0
1
0
0
0
0
0
0
1
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
1
1
0
1
53%
32%
0
0
0
0
0
0
0
0
0
0
1
0
1
0
0
0
0
0
1
1
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
1
1
0
0
0
0
0
0
1
1
1
0
48%
13%
0
1
0
0
1
0
0
0
1
0
1
0
1
0
0
0
0
1
0
1
0
0
0
0
0
0
0
1
0
0
0
0
0
0
1
0
1
0
0
0
1
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
1
1
0
0
37%
13%
The VMT by state and facility were provided by FHWA Highway Statistics (http://www.fhwa.dot.gov/policy/ohpi/hss/
index.htm) for 2000.
1 = Data present.
0 = Data absent.
A-18
-------
A-3.4 Data Sample Sizes
The following tables display the number of complete sampling days by site for each of
the three datasets; the VTRIS vehicle classification, VTRIS weight-in-motion, and TVT total vol-
ume data. Note that for the VTRIS and TVT volume data, all lanes in each direction are summed
together and each direction is counted separately in the TVT and VTRIS number of observations
below.
Table A-4. Number of VTRIS weigh-in-motion observations by state.
State 1999 Observations 2000 Observations
AR
CA
CO
CT
FL
GA
IA
ID
IN
KS
KY
LA
MD
ME
Ml
MN
MO
MS
MT
NC
NE
NJ
NM
NV
OH
PA
Rl
SC
SD
TX
UT
VA
WA
5,911,103
10,519,290
0
240,640
281,598
0
1,508,979
10,290,043
3,404,073
21,746
1,054,358
24,236
3,738
5,997
96,392
652,758
25,879
1,120,899
0
902,050
59,922
2,267,252
2,760,347
129,418
3,603,002
115,180
257,360
84,216
435,198
229,985
0
165,814
2,434,354
17,421,478
7,320,284
2,986,607
207,286
0
1,753
12,987,453
4,793,371
5,745,172
32,927
0
53,391
10,543
0
543,081
0
508,349
639,944
3,012,214
446,567
0
2,094,257
161,306
5,327
0
347,318
261,191
41,696
878,554
134,674
72,611
0
2,124,689
A-19
-------
State 1999 Observations 2000 Observations
Wl
wv
WY
Total
504,084
0
252,185
49,362,096
3,934,765
4,945
109,447
66,881,200
Table A-5. Number of VTRIS vehicle classification observation days by site-direction in each state.
State 1999 Observation Days 2000 Observation Days
AR
CA
CO
CT
FL
GA
IA
KS
KY
LA
ME
Ml
MO
MS
MT
NC
NE
NJ
NM
NV
OH
OK
PA
Rl
SD
WA
Wl
WV
WY
Total
4,179
888
0
6
356
0
259
39
104
16
32
15
0
752
0
97
17
16,879
0
532
776
2
148
259
966
6,545
1,165
0
306
34,338
16,390
467
2,253
9
0
6
4,523
62
0
26
0
137
892
486
7,353
28
0
15,674
50
2
690
417
209
231
1,436
7,225
2,148
6
455
61,175
A-20
-------
Table A-6. Number of 2000 TVT observation days (site-directions counted separately) in each state.
State 2000 Observation Days
AK
AL
AR
CA
CT
FL
GA
HI
IA
ID
IL
IN
KS
KY
LA
MA
Ml
MN
MO
MS
MT
NC
ND
NE
NH
NJ
NM
NV
NY
OH
OK
OR
PA
Rl
SC
SD
TN
TX
UT
VA
VT
WA
Wl
WV
WY
Total
48,676
54,687
29,044
20,868
23,458
95,177
36,877
5,990
91,494
98,820
22,310
18,918
61,708
39,081
1,422
18,466
77,359
102,401
14,161
29,386
20,797
44,438
25,936
30,337
35,656
48,247
34,804
27,640
48,041
58,060
5,708
65,253
43,661
8,085
13,671
26,803
4,841
94,645
44,085
120,784
23,268
25,667
67,472
20,011
44,495
1,872,708
A-21
-------
A-4. DATA REPORTING, DATA REDUCTION, AND DATA VALIDATION
A-4.1 Data Reduction Procedures
The data reduction procedures used for the VTRIS data are described in detail in section
(3) above. The SQL scripts used to reduce the data were delivered to EPA. The programming
scripts written for the purpose of eliminating and processing data were also delivered to EPA.
A-4.2 Data Validation Procedures
The data validation used by FHWA in preparing the database is the primary validation of
the raw data. When reducing this data to useful summary data, additional quality checks were
used and documented in Section 3 of this appendix. Final comparisons across temporal and geo-
graphic definitions provided additional understanding of the data and provided a validation of the
summaries produced. The comparison summaries demonstrated outliers that deserve additional
analysis and validation.
A-22
-------
APPENDIX B - CROSS REFERENCE METHOD TO CONVERT FHWA VEHICLE
CLASSES TO MOBILE VEHICLE TYPES
B-l
-------
B-l. INTRODUCTION
Heavy-duty vehicles are believed to generate a large fraction of emissions from on-road vehicles
and are coming under more intense scrutiny because light-duty emissions have been controlled to a greater
extent than heavy-duty vehicle emissions. A heavy-duty vehicle can produce 10 to 100 times the emissions
(of NOX and PM emissions especially) because heavy-duty engines emit at a higher rate per unit of power
than light-duty engines, and the vehicles themselves weigh more requiring greater engine loads. Key uncer-
tainties with the use of MOBILE6 regarding heavy-duty vehicle emissions include the fraction of heavy-
duty vehicles on all types of roadways at all times of day. In addition, there may be regional variability in
both the fraction of different vehicle classes and the vehicle weights within each class.
Heavy-duty vehicle activity needs to be better characterized in terms of the fraction of vehicles
on the road. One key uncertainty with the use of the current MOBILE6 model and future versions of
on-road emission estimates is the unknown fraction of heavy-duty vehicles on all types of roadways at all
times of day.
Traffic count data can be collected using a number of electronic devices. These can be road
tubes, loops, or weigh-in-motion (WTM) technology. Traffic counting devices can be either portable
or permanent. Some of the devices can measure time of day, vehicle speed, axle weight, total weight,
distance between axles, and total length, and then determine a fairly reliable vehicle classification. Some
devices are only able to collect an estimated total vehicle count (where the vehicle count is estimated to
be the number of axle hits divided by two). For the purposes of determining temporal distributions by
vehicle class, an estimate of vehicle classification is necessary. The vehicle mix by class is critical to un-
derstanding emissions, especially for NOx and particulate emission because the emission rates for these
pollutants from heavy-duty vehicles are orders of magnitude higher than those from light-duty vehicles.
The site characteristics of the data are also required for this analysis. The roadway type, num-
ber of lanes measured, and the total number of lanes in that direction must be indicated. In particular, it
is important that all lanes in a direction are measured. This is necessary to avoid any bias that could be
introduced from the fact that heavy-duty trucks tend to travel in the right lanes.
The vehicle count data consists of loop counter and pneumatic (tube counters). Typically there
are approximately 20 to 50 counters per state, primarily for multi-lane interstate and highway links.
For each site, the site characteristics required for the analysis include roadway functional classification,
county, number of lanes, and number of lanes measured.
The data provide vehicle classifications in FHWA standard class format, which are different from
those in MOBILE6. These classifications are listed in Table B-l and clearly distinguish light-duty pas-
senger vehicles from other vehicles. However, the vehicle classifications do not exactly match the MO-
BILE vehicle groupings. Historically, FHWA vehicle classifications are by the number and configuration
B-2
-------
of axels for a given vehicle. EPA vehicle classifications are by engine size. Thererfore, a cross-walk is
necessary between the FHWA vehicle classifications and the MOBILE6 classifications.
Table B-l. FHWA vehicle classifications.
FHWA Class Description
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Motorcycle
Passenger cars
Other 2-axle, 4-tire single unit vehicles
Buses
2-axle, 6-tire single-unit vehicles
3-axle, 6-tire single-unit vehicles
4+ axle single-unit vehicles
4 or less axle combination vehicles
5-axle combination vehicles
6+ axle combination vehicles
5-axle multi-trailer vehicles
6-axle multi-trailer vehicles
7+ axle multi-trailer vehicles
Unclassified
Unclassifiable
The vehicle mix is provided by the FHWA roadway functional class as listed in Table B-2,
though vehicle classification counters are usually sited on busy roadways so many of the road
types less traveled do not have data.
Table B-2. FHWA roadway functional classifications.
_ Classification Description
C°de RURAL
1
2
6
7
8
9
Principal Arterial
Principal Arterial
- Interstate
- Other
Minor Arterial
Major Collector
Minor Collector
Local System
URBAN
11
12
14
16
17
19
Principal Arterial
Principal Arterial
Principal Arterial
- Interstate
- Other Freeways or Expressways
- Other
Minor Arterial
Collector
Local System
-------
B-2. CROSS REFERENCE FROM FHWA TO MOBILE VEHICLE CLASSIFICATION
There are three methods available to cross-reference the vehicle counts by FHWA clas-
sification to the EPA classification scheme. The first was a joint effort by EPA and FHWA to
produce estimates for the EPA Trends report, and the second is a recent research effort. The EPA
method was used for this work because it has been vetted, but the other method is described here
for future reference.
EPA (2003) provided ENVIRON estimates of the crosswalk between the FHWA truck
classifications and the MOBILE6 vehicle types used in the NEI emission inventory development;
these are shown in Table B-3. The crosswalk for FHWA vehicle classes #2 and #3 was assumed
in this work to be the default light-duty mix as shown in Table 3 rather than an explicit result of
an EPA analysis. The vehicle counts can be aggregated to MOBILES or MOBILE6 groupings.
The reported vehicle class estimates by FHWA class was converted using both the EPA default
for light-duty and EPA crosswalk for heavy-duty vehicles to produce estimates by specific MO-
BILE6 vehicle classes. If MOBILES formats are needed, then the MOBILE6 vehicle classifica-
tions can be aggregated into the MOBILES groupings. It is not possible to determine the diesel
and gasoline fraction from the road counters, so either state registration or national averages
(such as provided in the MOBILE6 model) are used to apportion the vehicles by fuel type.
Table B-3. FHWA and MOBILE6 crosswalk estimates for heavier vehicles. (EPA, 2003).
MOBILE Weight
Ratings\FHWA
Types
Other 2-
Passenger Car axle 4-tire,
FHWA #2l FHWA #3
Single-Unit
Trucks,
FHWA #5-7
Combination
Trucks,
FHWA #8-13
LDV
6,000 Ibs or less
LDT1 & LDT2
6001 -10.0002
LDT3, LDT4, Class 2b
10,001 -14, 000 Class 3
14,001 -16,000 Class 4
19,500 Class 5
26,000 Class 6
33,000 Class 7
60,000 Class 8a
> 60, 000 Class 8b
52.3%
35.4%
12.3%
(1% Class 2b)
0
0
0
0
0
0
0
98.3%
(0.524% Class
2b)
0.44
0.14
0.13
0.24
0.12
0.05
0.006
0%
24%
21%
12
5.0
4.8
12
6.8
11
2.5
0%
0%
0.77%
0.61
0.65
0.64
3.3
3.7
28
62
1 - Default 2002 light-duty vehicle VMT distribution (EPA, 2004).
2 - 8% were estimated to be Class 2b, GVWR (8,500 - 10,000 Ibs) heavy-duty vehicles and of
those 24% diesel.
The definition in Table B-3, however, is not sufficient to map the vehicle identification
to vehicle class in either MOBILE6 or MOBILES. In order to map the vehicle classification
B-4
-------
into MOBILE6 groups, the default vehicle mix can be used to apportion between LDGT1 and
LDGT2 or between LDGT3 and LDGT4. Another problem with the method described in Table 3
is that it uses the default vehicle mix for FHWA Class 2, but the better denned method for FHWA
Class 3 would over allocate the vehicle counts to LDT and under allocate to LDV. The suggested
remedy is that light-duty portion of FHWA Class 3 be combined with FHWA Class 2 prior to
redistributing using the default light-duty allocation shown in Table B-3. The crosswalk for con-
verting FHWA vehicle classes into MOBILE6 vehicle classes is described in Table B-4.
Table B-4. Default 2002 VMT mix by the MOBILE6 16 vehicle classes and crosswalk calcula-
tion method from FHWA vehicle classes.
MOBILES Diesel Vehicle /*,,*• .. *u .,
._.. . . , _. _ .. „ ... Calculation Method
16 Vehicle Classes Fraction* Mix
LDV
LDT1
LDT2
LDT3
LDT4
HDV2B
HDV3
HDV4
HDV5
HDV6
HDV7
HDV8A
HDV8B
MDBS
HDBT
MC
0.0016
0.0007
0.0007
0.0138
0.0138
0.2414
0.7264
0.8307
0.4906
0.7075
0.8882
0.9996
1.0000
0.7500
1.0000
0.0000
0.459
0.072
0.238
0.074
0.034
0.038
0.004
0.003
0.002
0.008
0.010
0.011
0.038
0.002
0.001
0.006
0.523 x FHWA Vehicle Class 2 + 0.983 x 0.523
FHWA Vehicle Class 3
0.082 x FHWA Vehicle Class 2 + 0.983 x 0.082
FHWA Vehicle Class 3
0.272 x FHWA Vehicle Class 2 + 0.983 x 0.272
FHWA Vehicle Class 3
0.078 x FHWA Vehicle Class 2 + 0.983 x 0.084
FHWA Vehicle Class 3
0.036 x FHWA Vehicle Class 2 + 0.983 x 0.039
FHWA Vehicle Class 3
0.0099 x FHWA Vehicle Class 2 + 0.0052 x FHWA
Class 3 + See Table B-3 for other FHWA Classes
See Table B-3
See Table B-3
See Table B-3
See Table B-3
See Table B-3
See Table B-3
See Table B-3
FHWA Vehicle Class 4 & Fraction of Vehicle Mix of
HDBS and HDBT
FHWA Vehicle Class 4 & Fraction of Vehicle Mix of
HDBS and HDBT
FHWA Vehicle Class 1
*Default registration distribution x Default diesel fraction summed over all model years.
Diesel fraction from MOBILE6 defaults.
The diesel fraction is used to convert the 16 vehicle categories to the 32 vehicle catego-
ries used in MOBILE6 before combining categories to group the MOBILES vehicle categories as
shown in Table B-5. This conversion is performed if the user needs to run MOBILES instead of
or in addition to MOBILE6.
B-5
-------
Table B-5. Converting MOBILE6 vehicle types to MOBILES vehicle types - diesel fractions.
w uM?B^E5 Calculated from MOBILES
Vehicle Classes
LDGV
LDGT1
LDGT2
HDGV
LDDV
LDDT
HDDV
MC
LDV-LDDV
(1-Diesel fraction) x LDT1 + (1-Diesel Fraction) x
(1-Diesel fraction) x LDT3 + (1-Diesel Fraction) x
LDT2
LDT4
(1-Diesel fraction) x HDV2b + (1-Diesel Fraction) x HDV3 + (1-Diesel fraction) x HDT4 +
(1-Diesel Fraction) x HDV5 + (1-Diesel fraction) x HDV6 + (1-Diesel Fraction) x HDV7 +(1-
Diesel fraction) x HDV8A+ (1-Diesel Fraction) x HDV8B + (1-Diesel fraction) x HDBS
Diesel Fraction x LDV
SUM (LDT1, LDT2, LDT3, LDT4) - LDGT1 - LDGT2
SUM (HDV All, Buses All) - HDGV
MC
A sample of the results is shown in Figure B-l and demonstrates the higher fractions of
heavy-duty vehicle traffic overnight as well as distinguishing the day of week activity.
Wisconsin Function Class 11, August
Hourly Class Fractions Sunday through Saturday
0.6 1
0.1
1 5 9 13 17 21 1 5 9 13 17 21 1 5 9 131721 1 5 9 13 17 21 1 5 9 13 17 21 1 5 9 131721 1 5 9 131721
LDDT LDDV -*~LDGT1 -»~LDGT2 -+-|_DGV —MC
Figure B-l. Vehicle fractional mix over a week for urban interstates in Wisconsin.
Overall the vehicle mix results are consistent with the national average for light-duty/heavy-duty as
shown in Table B-6. Without a complete understanding of the total VMT for each roadway type within a given
region, it is difficult to determine if the regional average is similar to the national average. The fraction of heavy-
duty vehicles is highest on interstates and freeways, and it is typically lower on roads less traveled. Rural inter-
B-6
-------
states have higher heavy-duty vehicle fractions than urban interstates. One concern about the cross-reference
method used in this work is that heavy-duty diesel vehicles are a larger portion of the heavy-duty fleet than the
national average, at the expense of heavy-duty gasoline vehicles. In essence, the method may be biased towards
heavy-duty diesel vehicles and thus produce values that underrepresent heavy-duty gasoline vehicles.
Table B-6. Raw average annual vehicle mix estimates
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^H
RoadType Data HDDV HDGV LDDT LDDV LDGT1 LDGT2 LDGV MC
1
2
6
7
11
12
14
16
17
Wl
IL
Ml
MN
Wl
IL
Ml
MN
Wl
IL
Ml
MN
Wl
IL
Ml
MN
Wl
IL
Ml
MN
Wl
IL
Ml
MN
Wl
IL
Ml
MN
Wl
IL
Ml
MN
Wl
IL
Ml
MN
EPA Average
0.167
0.210
0.162
0.103
0.100
0.055
0.085
0.063
0.077
0.032
0.029
0.037
0.098
0.069
0.112
0.074
0.059
0.059
0.047
0.031
0.024
0.032
0.046
0.018
0.082
0.013
0.014
0.013
0.013
0.013
0.011
0.012
0.012
0.013
0.009
0.011
0.012
0.016
0.012
0.012
0.012
0.011
0.011
0.011
0.010
0.010
0.010
0.012
0.010
0.036
0.002
0.001
0.002
0.002
0.002
0.002
0.002
0.002
0.002
0.002
0.002
0.002
0.002
0.002
0.002
0.002
0.002
0.002
0.002
0.002
0.002
0.002
0.002
0.002
0.002
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.293
0.276
0.296
0.316
0.318
0.334
0.324
0.330
0.326
0.341
0.343
0.341
0.319
0.329
0.313
0.328
0.332
0.333
0.337
0.342
0.345
0.341
0.337
0.347
0.310
0.098
0.093
0.099
0.107
0.107
0.110
0.109
0.111
0.110
0.111
0.114
0.114
0.111
0.109
0.104
0.109
0.110
0.110
0.111
0.112
0.114
0.112
0.111
0.112
0.107
0.424
0.399
0.427
0.455
0.458
0.485
0.467
0.477
0.469
0.499
0.499
0.492
0.452
0.478
0.454
0.475
0.483
0.484
0.489
0.498
0.502
0.498
0.489
0.509
0.458
0.003
0.006
0.001
0.004
0.002
0.002
0.001
0.004
0.002
0.005
0.001
0.002
0.002
0.002
0.003
0.000
0.002
0.000
0.003
0.004
0.002
0.004
0.004
0.001
0.006
B-7
-------
The method described in this work to cross reference the FHWA classification into MO-
BILE vehicle categories is a relatively novel technique, though the state of Texas is currently us-
ing a similar method developed by TTI (2003) (and the description of their method is described
here) to adjust their emission inventories in the Houston-Galveston SIP (TCEQ, 2004). EN-
VIRON worked with EPA to develop this cross-reference method, however EPA may consider
further development of this method to reconcile the field measurement results with those derived
from registration, diaries, and surveys.
B-2.1 ALTERNATIVE METHODS FOR CROSS REFERENCE
At least two alternative methods have been suggested to cross reference FHWA catego-
ries with MOBILE classes of vehicle types. Georgia Institute of Technology and Texas Transpor-
tation Institute (TTI) have each proposed a method to cross reference vehicle count information
from automatic traffic recorder data to MOBILE vehicle classifications.
B-2.2 GEORGIA TECH INSTITUTE METHOD
One alternative method has been forwarded by researchers in the Atlanta area (Yoon, et
al., 2004) and is shown in Table B-7. There are two reasons why this method was not used for
this work: it has not been vetted, and the FHWA class #8 is distinguished into 3 and 4 axle cat-
egories, which is not provided in the data set. In addition, the EPA method would still be required
to map the FHWA class #3 results in the MOBILE classes. The FHWA #5 group needs to be
divided into several GVWR classifications presumably from registration or historic manufactur-
ers sales data making the cross-reference difficult without additional information.
Table B-7. FHWA and MOBILE6 crosswalk estimates for heavier vehicles. (Yoon et al., 2004).
FHWA Types \ MOBILE MOBILE MOBILE rciass 8W
MOBILE Weight (Class 2b- 7) (Class 8a) GVWR rx/WR SRO nnn IHQ\
Ratings GVWR (8,500 - 33,000) Ibs* (33,000-60,000) ^VVVK 1 bu.uuu IDS-'
#5
#6 & #8 (3-axle)
#7, #8(4-axle), #9-13
100%
0%
0%
0%
100%
0%
0%
0%
100%
* Uncertain fraction of FHWA #3 into GVWR 8,500 - 10,000 Ibs, Class 2b heavy-duty vehicles.
-------
B-2.3 TTI Method of VMT Mix
Another alternative method has been used by TTI (TCEQ, 2004) for constructing alterna-
tive VMT mix profiles for Texas on-road emission inventories. What follows is the TTI docu-
mentation describing the method quoted from the Texas SIP (TCEQ, 2004). The method relies
on regionally specific registration data for Texas and Houston-Galveston so cannot be used for
national average cross-reference method.
"For the 2000 estimate, 1997 - 2000 TxDOTvehicle classification data were used. The
eight-county area data were aggregated. TxDOT classification counts classify vehicles into the
standard FHWA vehicle classifications (based on vehicle length/number of axles) using best
practice vehicle classification count methods. "
[ENVIRON Note: The TTI classification follows the FHWA definition with the elimination of
unknown and unclassified vehicles and presumably the merging of motorcycles with passenger
vehicles. The tables have been renumbered to follow the document format. Table B-8 is a table of
the TTI definition and the inferred FHWA class definition.]
Table B-8. TTI vehicle identifiers.
FHWA Class TTI DEFINITION Description
1
2
3
4
5
6
7
8
9
10
11
12
13
C
P
B
SU2
SU3
SU4
SE4
SE5
SE6
SD5
SD6
SD7
Motorcycles
Passenger vehicles
Two-axle, four-tire single-unit trucks
Buses
Six-tire, two-axle single-unit vehicles
Three-axle single-unit vehicles
Four or more axle single-unit vehicles
Three or four axle single-trailer vehicles
Five-axle single-trailer vehicles
Six-axle multi-trailer vehicles
Five or less axle multi-trailer vehicles
Six-axle multi-trailer vehicles
Seven or more axle multi-trailer vehicles
EPA and MOBILE use a different vehicle classification scheme than the FHWA cat-
egories. The 28 EPA vehicle categories are defined as a function of gross vehicle weight rating
(GVWR) and fuel type (see Table B-9). The FHWA axle/vehicle length-based classification
categories must be converted into 28 MOBILE GVWR/fuel type based categories.
B-9
-------
The FHWA vehicle classification counts were first aggregated into three intermediate groups:
Passenger Vehicles (PV) C + P;
Heavy-Duty Vehicles (HDV) SU2 + SU3 + SU4 + SE4; and
HDDVSb (HDX) SE5 + SE6 + SD5 + SD6 + SD7.
This is followed by a second intermediate allocation that separates light-duty vehicles (LDV)
into PVs and light-duty trucks (LDT) based on TxDOT registration data:
LDV 0.695 x PV (by county, 2002 Harris registration data shown); and
LDT 0.305 x PV (by county, 2002 Harris registration data shown).
A third intermediate allocation further separates LDTs into LDT1 and HLDT (note that
LDT1 is itself intermediate and is further divided into LDGT1 and LDDT):
LDT1 0.813 x LDT (by county, 2002 Harris registration data shown); and
HLDT 0.187 x LDT (by county, 2002 Harris registration data shown).
Next, the remaining FHWA categories are disaggregated into EPA vehicle groups, as shown.
Note that TxDOT vehicle classification count procedures do not distinguish between gasoline
and diesel LDTs. Consequently, MOBILE defaults for the year of interest are used. As before, ac-
tual TxDOT vehicle registration data are used to separate gasoline from diesel heavy-duty trucks.
Note also that motorcycles are not counted separately and are included as a default (subtracted
from LDGV):
LDGV 0.9989987 x LDV (MOBILE6 default for 2007 shown);
LDDV 0.0010013 x LDV (MOBILE6 default for 2007 shown);
LLDT 0.9947975 x LDT1 (MOBILE6 default for 2007 shown);
LDDT 0.0052025 x LDT1 (MOBILE6 default for 2007 shown);
HDGV0.358 x HDV (by county, 2002 Harris County registration data shown);
HDDV0.642 x HDV (by county, 2002 Harris County registration data shown);
MC 0.001 of total (subtracted from LDGV).
This converts the FHWA axle count-based categories into GVWR categories. This part of the
conversion procedure is summarized schematically in Table 10. Starting with the TxDOT vehicle
classification data, these data themselves provide sufficient information to complete the first step
in the conversion process, the allocation of vehicles into PVs, HDVs, HDDVSbs, and buses (B).
Steps 2 and 3 further allocate these categories using TxDOT registration data. Finally, Step 4 al-
locates light-duty vehicles by fuel type using EPA MOBILE diesel fractions and motorcycles are
separated from light-duty gasoline vehicles using a nominal constant.
The MOBILE6 28-category typology is a subset of this typology. A combination of EPAMO-
BILE6 defaults and area vehicle registration data are used to expand these intermediate catego-
ries.
For the 28-category EPA scheme, HDVs — HDGV and HDDV — are separated into eight and
B-10
-------
seven categories respectively. HDDVSb vehicles are counted directly. The 15 HDV categories
are separated from total HDV, which have been separated by fuel type using TxDOT registra-
tion data by county. Each HDV category (HDGV and HDDV) is then divided into sub-categories
based on TxDOT area vehicle registration data. Buses are treated separately.
The 28-category EPA scheme also further divides the two LDT categories based in part on as-
sumed loading. The previous LDGT1 and LDGT2 categories (previously denned as GVWR <
6,000 and GVWR > 6,000 to 8,500, respectively) are separated into subcategories in terms of ad-
justed loaded vehicle weight (ALVW). ALVW is the average of vehicle curb weight and GVWR.
Thus, two new intermediate categories are introduced. These are light light-duty trucks (LLDT)
and heavy light-duty trucks (HLDT), which are denned as:
• LLDT - any light-duty truck rated through 6,000 pounds GVWR, and
• HLDT - any light-duty truck rated greater than 6,000 pounds GVWR.
These two new intermediate categories are then used to define the four LDT categories using
EPAMOBILE6 defaults for the year of interest. The four LDT categories are:
• LDGT1 -light light-duty trucks through 3,750 pounds loaded vehicle weight (LVW);
• LDGT2 - light light-duty trucks greater than 3,750 pounds LVW;
• LDGT3 - heavy light-duty trucks to 5,750 pounds ALVW; and
• LDGT4 - heavy light-duty trucks greater than 5,750 pounds ALVW.
Similarly, the LDDT category is sub-divided into two categories based on GVWR (less than or
equal to 6,000 GVWR and 6,000 to 8,500 GVWR). This is accomplished using EPAMOBILE6
default values for the year of interest.
Finally the three bus categories are separated from the TxDOT classification counts bus category
using EPAMOBILE6 default values. (Under MOBILE6 the HDV category does not include
buses.)
For historical VMT mix estimates, the MOBILE6 default values consistent with the historical
year are used. No other adjustments are made to alter the count data and conversion procedure to
accommodate historical years. Table 11 shows the VMT mix estimation procedure summary fol-
lowed by explanatory notes. For this analysis, VMT mix estimates were developed for applica-
tion with three functional classification groups (see Table 31 in Emissions Calculations section)
[not shown here] and four time-of-day periods (See Table 5 [not shown here]).
This procedure is performed as described for weekdays. TxDOT vehicle classification data are only
collected for weekdays (Monday through Thursday), consequently other data is used to estimate VMT mix
for Fridays, Saturdays, and Sundays. The procedure used to estimate Friday, Saturday, and Sunday VMT
mix relies on vehicle classification data collected over several years in urban areas. The ratio of weekday
VMT mix to Friday, Saturday, and Sunday VMT mix is applied to the weekday VMT mix to produce
region specific Friday, Saturday and Sunday VMT mix. (No seasonal changes are assumed).
B-ll
-------
Table B-9. EPA Vehicle Types - 28 Categories.
Category Description GVWR
LDGV
LDGT1
LDGT2
LDGT3
LDGT4
HDGV2b
HDGV3
HDGV4
HDGV5
HDGV6
HDGV7
HDGVSa
HDGVSb
HDGB
LDDV
LDDT12
LDDT34
HDDV2b
HDDV3
HDDV4
HDDV5
HDDV6
HDDV7
HDDVSa
HDDVSb
HDDBS
HDDBT
MC
Light-duty gasoline vehicle
Light-duty gasoline truck
Light-duty gasoline truck
Light-duty gasoline truck
Light-duty gasoline truck
Heavy-duty gasoline vehicle
Heavy-duty gasoline vehicle
Heavy-duty gasoline vehicle
Heavy-duty gasoline vehicle
Heavy-duty gasoline vehicle
Heavy-duty gasoline vehicle
Heavy-duty gasoline vehicle
Heavy-duty gasoline vehicle
Heavy-duty gasoline bus
Light-duty diesel vehicle
Light-duty diesel truck
Light-duty diesel truck
Heavy-duty diesel vehicle
Heavy-duty diesel vehicle
Heavy-duty diesel vehicle
Heavy-duty diesel vehicle
Heavy-duty diesel vehicle
Heavy-duty diesel vehicle
Heavy-duty diesel vehicle
Heavy-duty diesel vehicle
Heavy-duty diesel school bus
Heavy-duty diesel transit bus
Motorcycle
< 6,000
< 6,000
< 6,000
6,001 -8,500
6,001 -8,500
8,501 -10,000
10,001 -14,000
14,001 -16,000
16,001 -19,500
19,501 -26,000
26,001 -33,000
33,001 -60,000
> 60,000
all
< 6,000
< 6,000
6,001 -8,500
8,501 -10,000
10,001 -14,000
14,001 -16,000
16,001 -19,500
19,501 -26,000
26,001 -33,000
33,001 -60,000
> 60,000
all
all
all
Table B-10. Initial Vehicle Classification Conversion Procedure.
Start Step 1 Step 2 Step 3 Step 4
Total
Vehicles
PV
HDV
LDV
LOT
LDGV
MC
LDGV
LDDV
LDDT
LDGT
LLDT
LDDT
LDGT 1 & 2
LDGT 3 & 4
HLDT
HDGV
HDDV
HDDVSb
Buses
B-12
-------
Table B-l 1. VMT Mix Estimation Procedure Summary.
EPA-8 EPA-28 Conversion
LDGV
LDGT1
LDGT2
HDGV
LDDV
LDDT
HDDV
MC
LDGV
LDGT1
LDGT2
LDGT3
LDGT4
HDGV2b
HDGV3
HDGV4
HDGV5
HDGV6
HDGV7
HDGVSa
HDGVSb
HDGB
LDDV
LDDT12
LDDT34
HDDV2b
HDDV3
HDDV4
HDDV5
HDDV6
HDDV7
HDDVSa
HDDVSb
HDDBT
HDDBS
MC
0.9990 x LDV
0.231 Ox LLDT
0.7690 x LLDT
0.6850 x HLDT
0.3150 x HLDT
0.519 x HDGV
0.194x HDGV
0.094 x HDGV
0.034 x HDGV
0.091 x HDGV
0.032 x HDGV
0.032 x HDGV
0.004 x HDGV
0.0931 x B
0.0010 x LDV
0.0337 x LDDT
0.9663 x LDDT
0.278 x HDDV
0.134x HDDV
0.081 x HDDV
0.053 x HDDV
0.168 x HDDV
0.102x HDDV
0.184x HDDV
HDX
0.3239 x B
0.5830 x B
MC
B-13
-------
Notes to VMT Mix Estimation Procedure Summary
Intermediate category factors and sources:
LDV 0.695 x PV (by county, 2002 Harris County registration data shown)
LOT 0.305 x PV (by county, 2002 Harris County registration data shown)
LDT1 0.813 x LOT (by county, 2002 Harris County registration data shown)
HLDT 0.187 x LOT (by county, 2002 Harris County registration data shown)
LLDT 0.9948 x LDT1 (EPAMOBILE6 default, 2007 shown)
LDDT 0.0052 x LDT1 (EPAMOBILE6 default, 2007 shown)
HDV SU2+SU3+SU4+SE3+SE4
HDX SE5+SE6+SD5+SD6+SD7
HDGV 0.358 x HDV (by county, 2002 Harris County registration data shown)
HDDV 0.642 x HDV (by county, 2002 Harris County registration data shown)
Category conversion factors and sources:
LDGV 0.9990 x LDV (EPA MOBILE6 default, 2007 shown)
LDGT1 0.2310 x LLDT (EPA MOBILE6 default, 2007 shown)
LDGT2 0.7690 x LLDT (EPAMOBILE6 default, 2007 shown)
LDGT3 0.6850 x HLDT (EPAMOBILE6 default, 2007 shown)
LDGT4 0.3150 x HLDT (EPAMOBILE6 default, 2007 shown)
HDGV2a 0.519 x HDGV (HGAC area registration data)
HDGV3 0.194 x HDGV (HGAC area registration data)
HDGV4 0.094 x HDGV (HGAC area registration data)
HDGV5 0.034 x HDGV (HGAC area registration data)
HDGV6 0.091 x HDGV (HGAC area registration data)
HDGV7 0.032 x HDGV (HGAC area registration data)
HDGVSa 0.032 x HDGV (HGAC area registration data)
HDGVSb 0.004 x HDGV (HGAC area registration data)
HDGB 0.0931 x B (EPA MOBILE6 default, 2007 shown)
LDDV 0.0010 x LDV (EPA MOBILE6 default, 2007 shown)
LDDT12 0.0337 x LDDT (EPAMOBILE6 default, 2007 shown)
LDDT34 0.9663 x LDDT (EPAMOBILE6 default, 2007 shown)
HDDV2b 0.278 x HDDV (HGAC area registration data)
HDDV3 0.134 x HDDV (HGAC area registration data)
HDDV4 0.081 x HDDV (HGAC area registration data)
HDDV5 0.053 x HDDV (HGAC area registration data)
HDDV6 0.168 x HDDV (HGAC area registration data)
HDDV7 0.102 x HDDV (HGAC area registration data)
HDDVSa 0.184 x HDDV (HGAC area registration data)
HDDVSb HDX (TxDOT classification counts)
HDDBT 0.3239 x B (EPAMOBILE6 default, 2007 shown)
HDDBS 0.5830 x B (EPAMOBILE6 default, 2007 shown)
MC MC (default subtracted from LDGV, no conversion)"
B-14
-------
This page left blank deliberately.
B-15
-------
APPENDIX C - NATIONAL AVERAGE TEMPORAL
PROFILES FOR FOUR ROAD TYPES
C-l
-------
C-l. INTRODUCTION
This appendix was prepared to satisfy a request by EPA to provide summary temporal
profiles grouped into four road types. The groupings combine several different road types that
each have unique profiles and different sample sizes, so the groupings could include sampling
bias by the selection of the sites by the local and state transportation departments.
The results here provide state and national temporal aggregations requested by road type
or also called roadway functional classification. The EPA work assignment requested national
and state temporal profiles by month, day of week, and time of average day of total vehicle travel
from the Traffic Volume Trends (TVT) data and by vehicle type using the Vehicle Travel Infor-
mation System (VTRIS). The temporal profiles requested were to be reclassified into the four
road types described in Table C-l.
Table C-l. FHWAroadway functional classification (types) in TVT and VTRIS.
A . Classification Description
C°de RURAL
1
2
6
7
8
9
Principal Arterial
Principal Arterial
- Interstate
- Other
Minor Arterial
Major Collector
Minor Collector
Local System
URBAN
11
12
14
16
17
19
Principal Arterial
Principal Arterial
Principal Arterial
- Interstate
- Other Freeways or Expressways
- Other
Minor Arterial
Collector
Local System
The vehicle types that are reported in VTRIS are shown in Table C-2. The most preva-
lent vehicle types in the database are vehicle 2 and 3, which are primarily light-duty vehicles.
Other significant vehicle types are vehicle type 9 (18 wheel line-haul trucks), and vehicle type 5
(typical of local delivery trucks). The remainder of the vehicles types represent typically 3% or
less of the traffic volume except on rural interstates where other truck types are found in higher
numbers.
C-2
-------
Table C-2. FHWA Vehicle classifications.
FHWA Class VTRIS Vehicle Type
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Motorcycle
Passenger cars
Other 2-axle, 4-tire single unit vehicles
Buses
2-axle, 6-tire single-unit vehicles
3-axle, 6-tire single-unit vehicles
4+ axle single-unit vehicles
4 or less axle combination vehicles
5-axle combination vehicles
6+ axle combination vehicles
5-axle multi-trailer vehicles
6-axle multi-trailer vehicles
7+ axle multi-trailer vehicles
Unclassified
Unclassifiable
ENVIRON analyzed Traffic Volume Trends (TVT) data to provide national average and
individual state temporal profiles by both the FHWA and the reclassified road types described in
Table C-l. The 2000 TVT data represented several states and a number of sample sites as shown
in Table C-3. For national and state EPA-requested road type groupings (described in Table C-l
and outlined in alternating white and light gray background in Table C-3), the vehicle counts and
number of sample sites were summed to determine average vehicle counts per site (where op-
posite directions at a site were treated as separate sites). So the importance of each road type in a
group was determined by the number of sampling sites regardless of the relative miles of road-
way or vehicle miles traveled (VMT) on each roadway type. To avoid a possible day of week
bias, data for a site month was dropped if not all seven days were measured at that site during
that month. The national average volumes were then calculated as the average of the state aver-
age volumes.
Table C-3. FHWA roadway functional classification (types) in TVT.
FHWA .......... „. ... .. _, ... EPA MOVES _. . _.
_ . FHWA Classification Description ^ States Represented Sites
1
2
6
7
8
9
11
12
14
16
17
19
Principal Arterial - Interstate
Principal Arterial - Other
Minor Arterial
Major Collector
Minor Collector
Local System
Principal Arterial - Interstate
Principal Arterial - Other Freeways or Expressways
Principal Arterial - Other
Minor Arterial
Collector
Local System
RLA
RO
RO
RO
RO
RO
ULA
ULA
UO
UO
UO
UO
44
45
44
42
13
11
45
29
44
39
24
8
964
1646
927
564
57
45
913
354
901
390
132
44
c-:
-------
National temporal allocations by road type grouping were prepared for this work as-
signment. Figures 1, 2, and 3 show the monthly, day of week, and time of day average temporal
allocations. The trends for the monthly vehicle counts shown in Figure 1 incorporate to some
extent the year-to-year growth in VMT and therefore increase from the beginning of the year to
the end perturbed by a seasonal trend. The trend for day of week activity shows higher activity
during the week peaking on Friday, with Saturday and Sunday traffic much lower. The time of
day activity show the typical diurnal mid-day traffic increase with the urban commuting period.
The national average volumes generated from the TVT data in figures 1 through 3 are calculated
as the average of the state averages, following the same procedures as outlined in Appendix D.
For these averages, the sample size corresponds to the number of states represented.
45,000
40,000
35,000
a, 30,000
E
3
co
Q
25,000
g, 20,000
01
<
15,000
10,000
5,000
. . -)f
.-I-I-I-i-
i i i i'
I
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
- Rural Limited Access • • a • • Rural Other
-Urban Limited Access • X- Urban Other
Figure C-l. National average daily volumes by month with 90% confidence intervals.
C-4
-------
50,000
5,000
Sunday Monday Tuesday Wednesday Thursday Friday Saturday
• Rural Limited Access
•Rural Other
-Urban Limited Access - * • Urban Other
Figure C-2. National average daily volumes by day of week with 90% confidence intervals.
3,500
•Rural Limited Access --Q-- Rural Other —*—Urban Limited Access - X- Urban Other
Figure C-3. National average hourly volumes with 90% confidence intervals.
C-5
-------
For the VTRIS data analysis of vehicle classifications, the 1999 and 2000 databases were
merged to generate the aggregate average profiles.
Monthly temporal profiles for the four EPA groups of road types were developed to
provide an understanding of the likely results. The results shown in Figure 4 demonstrate the
month-to-month variability in the results for rural interstates. Vehicle type 2 (so called passenger
cars) shows high travel fractions for the month of July on both rural road types and urban limited
access, and high variability from month to month. Vehicle type 3 also shows variability from
month to month with July results higher for rural and urban limited access road types. Other
vehicle types show less variability in count magnitude, however similar relative (percentage)
variability. The average volumes calculated from the VTRIS data presented in Figures 4 through
16 were calculated directly from the site-values instead of from the average of state averages.
Thus, the confidence intervals are small because of the large sample size though state to state
variability remains high.
The number of vehicle counts varies more than the relative vehicle fractions of the total
because of the site selection and especially the relative number of sites at higher or lower volume
roadways. So the vehicle mix fractions shown in Figure C-5 do not vary as much as the total
counts of each vehicle types as shown in Figure C-4. Even the dramatic July peak in Vehicle 2
and 3 traffic counts is much less apparent when observing the vehicle fleet fractions.
Similarly variable vehicle counts for other roadway types are provided in Figures C-6
- C-8, and also show more variability than the fleet fractions as the total volume counts vary
from month to month. This suggests that the fleet fractions are a valid result, and can be used to
represent the temporally averaged vehicle mix.
C-6
-------
600
500
j2 400
c
3
o
O
o 300
I
0
O>
5
0
< 200
100
-X .
—•—Vehicle 1
—*—Vehicle 2
- X- Vehicles
—*— Vehicle 4
- O- Vehicles
—i—Vehicle 6
Vehicle 7
—f> Vehicles
• Vehicle 9
Vehicle 10
Vehicle 11
—Vehicle 12
—*- Vehicle 13
Figure C-4. National average monthly vehicle counts per site for rural limited access roads with 90%
confidence intervals on vehicle types 2, 3, 5, and 9.
1
o
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
Figure C-5. National average monthly vehicle fleet fractions for rural limited access roads.
C-7
-------
•ion
IZU T
100
« 80
3
0
•z1
o 60
I
a)
s
oi
< 40
20
o -1
^
x -
S -
jX
^^4
jj,
• -A- -
p^_
• x* *
^x
. X..
— -* —
-e -
• "Fl^™
*^
*x-
1 — » —
- -e- -
- -en — !
X'
-x -
— * —
- •©- -
- R — •
/^
• -x- •
•—-» —
- e -
• Fl ^
^\
• *• ,
— « —
_ ra
V
" 'X- '
— • —
- -e- •
. _ ra
I
^
. X
— 4 —
- ©- •
m.
-\
-x.
»— ,
-© -
\
' *
— n ,
— • — Vehicle 1
— * — Vehicle 2
- X- Vehicles
— * — Vehicle 4
- €>- Vehicles
— i — Vehicle 6
Vehicle 7
— D- Vehicles
* Vehicle 9
Vehicle 10
Vehicle 11
Vehicle 12
Vehicle 13
1 2 3 4 5 6 7 8 9 10 11 12
Month
Figure C-6. National average monthly vehicle counts per site for rural other roads with 90% confidence
intervals on vehicle types 2, 3, 5, and 9.
1400
1200
1000
01
'c
3
O
^ 800
3
O
I
S) 600
(0
oi
<
400
200
0
S-.
x -
1
---i —
x
- -X- -
2
-V
- x- •
3
1^
'
-X -
• .A .
-^~
4
-"*-
• -x- -
5
- X* '
6
A
A
— R—
7
1 -x -
8
V
• -x- -
^4^
9
Jr
/
. x- •
- i .
-fU
10
T
-^\
-X .
. A.
^q
N
• -x
— ft
— fl_
12
—•—Vehicle 1
— A — Vehicle 2
- X- Vehicles
-Hie— Vehicle 4
- 0- Vehicles
— i — Vehicle 6
Vehicle 7
— f> Vehicle 8
» Vehicle 9
Vehicle 10
Vehicle 1 1
Vehicle 12
Vehicle 13
Month
Figure C-7. National average monthly vehicle counts per site for urban limited access roads with 90%
confidence intervals on vehicle types 2, 3, 5, and 9.
-------
350
300
1 250
8
•g1
o 200
I
0
S1
(o
«> 150
<
100
50
o -I
ft
X
e •
A
.X- -
- •©- -
~-i"
- e-
-r
x •
• -••
I
A
*
- *-•
\.
^\
• •*• .
- e -
V
• x* '
• -e- -
A
,* .
- e- .
. s
^^
I
%
*"
•*?-
A
. *• *
\
V
* M m
- e-
^
. -X
-e
—•—Vehicle 1
— * — Vehicle 2
- * - Vehicle 3
— * — Vehicle 4
- 0- Vehicles
— i — Vehicle 6
Vehicle 7
— E> Vehicle 8
• Vehicle 9
Vehicle 10
Vehicle 1 1
Vehicle 12
Vehicle 13
1 2 3 4 5 6 7 8 9 10 11 12
Month
Figure C-8. National average monthly vehicle counts per site for urban other roads with 90% confidence
intervals on vehicle types 2, 3, 5, and 9.
The day of week and time of day summary estimates are shown in the Figures C-9 to C-
16. The general trends follow those of the monthly averages where the trucks (primarily vehicle
types 5 and 9) show higher activity on rural limited access and lower activity on urban other
road types. The trucks also show less diurnal and weekday variability compared with the light-
duty vehicles (vehicle types 2 and 3) where Friday travel and morning and afternoon peak travel
times are evident. Trucks show little deviation from one weekday to another and regular diurnal
profiles without rush hour peaks.
C-9
-------
450.00
400.00 -
350.00 -
| 300.00 -
3
o
>. 250.00 -
3
o
co
< 150.00 -
100.00 -
50.00 -
0.00
X- .
—•—Vehicle 1
—A— Vehicle 2
- X- Vehicles
—*—Vehicle 4
• O- Vehicles
—i—Vehicle 6
Vehicle 7
—D- Vehicles
—*—Vehicle 9
Vehicle 10
Vehicle 11
Vehicle 12
—*- Vehicle 13
Figure C-9. National average day of week vehicle counts per site for rural limited access roads with 90%
confidence intervals on vehicle types 2, 3, 5, and 9.
120.00
100.00 -
| 80.00
o
O
o 60.00
I
01
O>
£
01
< 40.00
20.00 -
0.00
. - *• ,
..... e ..... e ..... e ..... e
—•—Vehicle 1
—*— Vehicle 2
- * - Vehicle 3
—*—Vehicle 4
- O- Vehicles
—i—Vehicle 6
Vehicle 7
Vehicle 8
Vehicle 9
Vehicle 10
Vehicle 11
Vehicle 12
Vehicle 13
Figure C-10. National average day of week vehicle counts per site for rural other roads with 90% confi-
dence intervals on vehicle types 2, 3, 5, and 9.
C-10
-------
1200.00
1000.00 -
800.00 -
o 600.00
I
01
01
<
400.00 -
200.00 -
0.00
M *
4 1
Sunday Monday Tuesday Wednesday Thursday Friday Saturday
—•—Vehicle 1
—±— Vehicle 2
- *- Vehicles
—*—Vehicle 4
- O- Vehicles
—i—Vehicle 6
——Vehicle 7
—Q- Vehicles
» Vehicle 9
Vehicle 10
Vehicle 11
Vehicle 12
—*- Vehicle 13
Figure C-ll. National average day of week vehicle counts per site for urban limited access roads with
90% confidence intervals on vehicle types 2, 3, 5, and 9.
JOU.UU -
300.00 -
250.00 -
Hourly Counts
b
o
| 150.00 -
01
<
100.00 -
50.00 -
o.oo -
__ — - — t-- ' — — ^ ^"" ^S\
v*-~
*. — * * * x * - . .
. - * * * x
X*
_ . . . 9 e © e o o
Sunday Monday Tuesday Wednesday Thursday Friday Saturday
— • — Vehicle 1
— *— Vehicle 2
- X- Vehicles
— * — Vehicle 4
- e- Vehicles
— i — Vehicle 6
Vehicle 7
— D- Vehicle 8
* Vehicle 9
Vehicle 10
Vehicle 1 1
Vehicle 12
Vehicle 13
Figure C-12. National average day of week vehicle counts per site for urban other roads with 90% confi-
dence intervals on vehicle types 2, 3, 5, and 9.
C-ll
-------
800
—•—Vehicle 1
—*— Vehicle 2
- *- Vehicles
Vehicle 4
- €>- Vehicles
Vehicle 6
Vehicle 7
I- Vehicles
—Vehicle 9
Vehicle 10
Vehicle 11
Vehicle 12
— Vehicle 13
100 -
0^a^^?^M
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Figure C-13. National average time of day vehicle counts per site for rural limited access roads with 90%
confidence intervals on vehicle types 2, 3, 5, and 9.
ZOU 1
200
o 150
O
3
O
I
HI
O)
™ -i nn -
5 1UU
>
50
0 -J
^
*-
-rP
^
-4-
-1T1
-*-
3»r
:TF
y
^
/
/
,*
-r>
/
X*
A
-rr
/
^
r«r
i— 1_
^^
"iv
*(-
r*-
_rn_
--
K-
_m_
^
X'
4-
£-|_
X
X
•4-
m
,*-
X
.
—ITL
nX
-*
•X
"ej
=TT^ ii j L ' ' J > — ' |-j
y
-X
s~
JSl,
^^i~^
X
.x
rt-
J-|
T^r- •
i-
/^
-X-
-*•
M^-
^
A
"X
^*»
^CT
•r>-
\
\
3E
•-TT-
\
vx.
•ff
X
x.
X
X.
-±-
^rl4
X
X.
•*-
-n-
\
^
^u-
— •— Vehicle 1
—*— Vehicle 2
• X- Vehicles
— * — Vehicle 4
- 0- Vehicles
— i — Vehicle 6
Vehicle 7
— f> Vehicles
—•—Vehicle 9
Vehicle 10
Vehicle 1 1
Vehicle 12
Vehicle 13
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Figure C-14. National average time of day vehicle counts per site for rural other roads with 90% confi-
dence intervals on vehicle types 2, 3, 5, and 9.
C-12
-------
1800
1600
1400
w 1200
'c
3
o
^ 1000
3
o
0) 80°
s
0
< 600
400
200
0 \
L£
V
^*-
J
1
m
4
i
/
f
^X
I^Pfp^V^M-VW-
T
f-
/"-
/
.X
-I
\
•x
V
•X-
af
•K-
»
^
K-
1
*
X
f
-r
X
• X
*
/
-X
A
"X
\
'x.
\
x.
V
X.
X
X.
X
X,
^
—•—Vehicle 1
— *— Vehicle 2
- K- Vehicle 3
— * — Vehicle 4
- 0- Vehicles
— i — Vehicle 6
Vehicle 7
— E> Vehicle 8
» Vehicle 9
Vehicle 10
Vehicle 1 1
—Vehicle 12
Vehicle 13
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Figure C-15. National average time of day vehicle counts per site for urban limited access roads with
90% confidence intervals on vehicle types 2, 3, 5, and 9.
600 -
500 -
« 400 -
£
o
0
o 300 -
I
0
O)
s
"* 200 -
100 -
o -I
*
M
V
fc
'Sfc,
•*.
T*"
^
M
/
|
/T
y
H>
/
/
X
f-
,*
-e-
^
A
-X-
•e-
3*M
V
*-
*-
e-
^^_
=^P
V
-*
*•
JL
TT
jr
>4
Z '
-9
^^_
^•p
*-
f
X
—Ai-
=rp
-i-
_L
-*
— A_
^*P
^
-XT
«.
__^t_
^*P
^
JL
'*
e<
/t-
•**
«
I
A
•-.
•»
\
X
-e-
\
'x,
•».
V
X ,
X
X ,
-^
X
x,
\
X
— • — Vehicle 1
—*— Vehicle 2
• * - Vehicle 3
— * — Vehicle 4
- O- Vehicles
— i — Vehicle 6
—a- Vehicles
—*— Vehicle 9
Vehicle 10
Vehicle 1 1
Vehicle 12
Vehicle 13
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Figure C-16. National average time of day vehicle counts per site for urban other roads with 90% confi-
dence intervals on vehicle types 2, 3, 5, and 9.
C-13
-------
This page left blank deliberately.
C-14
-------
APPENDIX D - ANALYSIS OF 2000 TRAVEL VOLUME TRENDS (TVT) DATA AND
2000 VEHICLE TRAVEL INFORMATION SYSTEM (VTRIS) DATA
D-l
-------
D-l. INTRODUCTION
The goal of the analysis in this Appendix is two-fold: (1) the first goal is to examine the
state-to-state variability in the TVT temporal profiles, and (2) to compare the temporal patterns of
the TVT data with those of the VTRIS data. The first is to determine if regions should consider
using a specific profile for that region. The second goal is to determine if the temporal variation
of the VTRIS data is sufficiently compatible to those of the TVT data to use TVT for total traffic
volume and VTRIS for vehicle mix fractions by roadway. The TVT is a more robust source of
total traffic volume, and so has historically been considered the best source for those estimates.
For this analysis, only the TVT and VTRIS data sets collected in the year 2000 were used
because only the 2000 TVT data was available to us in a database format. Note that the number
of states for which there were VTRIS data is substantially smaller than for the TVT data. To
illustrate, Table D-l below displays the number of states by facility class that were used in the
calculation of the monthly profiles (after dropping states for which there were fewer than twelve
months of data). One should also bear in mind that data submittal to VTRIS is voluntary on the
part of the states. Thus, the VTRIS data for a given state may only be a subset of vehicle clas-
sification data that is truly available for that state. However, for the purpose of this analysis, it is
worthwhile to see if the national average profiles of each dataset have a semblance to one an-
other.
Table D-l. Number of States used in the calculation of the monthly temporal profiles.
Facility Class
Number of Number of
TVT States VTRIS States
Rural
01 - Principal Arterial - Interstate
02 - Principal Arterial - Other
06- Minor Arterial
07- Major Collector
08 -Minor Collector
09 - Local System
32
33
30
28
10
8
3
5
2
1
0
0
Urban
11 - Principal Arterial - Interstate
12 - Principal Arterial - Other Freeways or Expressways
1 4 - Principal Arterial - Other
16- Minor Arterial
17 -Collector
19 - Local System
32
22
29
24
15
7
2
3
1
1
0
0
The vehicle classification data in the VTRIS data measures the hourly volumes of ve-
hicles on various roadway types by vehicle class. The vehicle class volumes for the VTRIS data
were summed together to make a total hourly volume that would be comparable to the hourly
volumes in the TVT data.
D-2
-------
D-2. DATA HANDLING PROCEDURES
For the two datasets to be compared, it was necessary to use the same data formatting and
aggregation routines. The more important quality assurance routines are described along with an
illustrative example of the reason for that routine.
In general, the hourly profiles were calculated by first dropping a day's worth of observa-
tions at a site where there were more than eight hours of zero volumes. This quality assurance
step is similar to one employed by the Department of Transportation to eliminate records with
more than seven consecutive hours with zero volumes in order to drop records where equipment
failure was likely. The next quality assurance step was to drop any site data in a given month
for which there were not all seven days of the week represented for that month. This step was to
eliminate any bias in the hourly profiles due to day of week variability.
Figure D-l below illustrates the day of week variability in the hourly data for Urban
Interstates (Class 11). Figure D-l shows typical morning and evening peak periods on Monday
through Friday, but only single peak periods on Sunday and Saturday.
National and State Average TVT Hourly Profiles By Day Of Week for
Urban Interstates
1 5 9 13 17 21 1 5 9 13 17 21 1 5 9 13 17 21 1 5 9 13 17 21 1 5 9 13 17 21 1 5 9 13 17 21 1 5 9 13 17 21
Hour
AK
IA
Ml
NH
— PA
— VT
— AL
— ID
— MN
NJ
Rl
WA
AR
— IL
— MO
NM
SC
— wi
CA
IN
MS
NV
— SD
— wv
CT
KS
— MT
NY
— TN
WY
FL
— KY
— NC
OH
TX
^~Nat Avg
GA
— LA
ND
OK
UT
— HI
— MA
NE
OR
VA
Figure D-l. Day of week variability in hourly volumes from Sunday to Saturday.
Figure D-2 below displays the average daily total volumes at two VTRIS Washington
sites for urban interstates where the average daily volume at site P3N was an order of magnitude
greater than that at site 111. Averaging together the Monday hourly P3N volume with the other
D-3
-------
days of the week at site 111 resulted in an hourly temporal profile with an unusually large peak
on Monday . To remove this potential source of bias, site-months were required to have observa-
tions for all seven days of the week.
Average Daily VTRIS Volumes at Washington Sites for Urban Interstates (Class 11)
E
3
•
•t • - - • 1 .
1 234567
Day of Week
-*-111
-«-P3N
Figure D-2. Average daily volumes at two VTRIS Washington State urban interstate
sites.
Finally, the November TVT volumes for the state of Alaska appeared to be shifted by one
column. Every hourly volume ended in the digit zero, and they were typically ten times greater
in magnitude than the other months. Figure D-3 displays the monthly temporal profile with
the month of November included in the Washington profile on rural interstates. November for
Alaska was dropped in the analysis.
D-4
-------
State and National Average Monthly TVT Profiles for Rural Interstates (Class 01)
0.45
0.40
0.35
0.30 -
0.25 -
0.20 -
0.15
0.10 -
0.05
0.00
Jan
Feb
Mar
Apr May
Aug Sep
Oct
Nov
Dec
— »-AK
KY
-A-NJ
-K-VT
-•-
-*-
-*-
AL
MA
NY
WV
AR
*-MI
-*-OH
WY
CA
-*-MN
•-OR
^^Nat
-*-
-•-
— 1—
—*—
CT
MS
PA
VTRIS
-•-GA
MT
Rl
— 1— IA
NC
SC
ID
— ND
— »-SD
— IL
— »-NE
-•-TX
KS
NH
-A-VA
Figure D-3. Monthly temporal profile including the month of November for Alaska.
Calculation of Temporal Profiles
The total volume counts were averaged at individual sites before being averaged across
sites (VTRIS hourly total volume counts were calculated as the sum of the hourly class counts).
The reason for doing this was to ensure that sites with longer periods of observation were not
more heavily weighted than others, but rather sites with heavier traffic volumes would be more
influential in averages across sites. The site averages were calculated by first totaling the counts
across lanes in the same roadway direction. Different directions at a site were treated separately.
Then the counts were averaged for each site-direction pair by hour, day of week (i.e., Sunday
through Saturday), and month. This means that at most five values (because there are fewer than
5 weeks in any month) were averaged together, corresponding to the total number of days of a
week during one month. For example, all Monday counts during January for hour 10 were aver-
aged together at each site-direction pair. The site averages were used as the starting point for all
further analyses.
The hourly profiles were calculated as follows:
1. The hourly volumes for each site-direction were averaged together by day of week for
each state.
D-5
-------
2. The state fractional profiles for each day of the week were calculated as the hourly
fraction of the daily total volume for each day of the week at each site.
3. The national average hourly profile by day of week was calculated as the average of
the state fractional averages. As discussed below, the national fractional profiles were
calculated as the average of the state fractional profiles.
4. The overall hourly state volumes by day of week were averaged across the days of the
week to create a single hourly profile for each state.
5. The overall hourly state fractional profiles were calculated as the average hourly vol-
ume over the average daily volume for each state.
6. Finally, the national average hourly profile was calculated as the average of the state
fractional profiles1'2.
In generating the day of week profiles, we started from the hourly data that was already
checked for completeness at each site for each day of the week by month.
1. The average hourly volumes by day of week and state were totaled to generate an
average day of week volume by state.
2. The state fractional day of week profiles were calculated as the fraction of the daily
volume to the weekly total volume.
3. The national average day of week fractional profile was calculated as the average of
the state day of week profiles.
The monthly profiles were generated as follows:
1. The monthly daily total volume averages by state was calculated from the day-of-
week complete dataset generated above.
2. All states without twelve months of daily averages were dropped.
3. The state fractional profiles were calculated as the fraction of the average daily vol-
ume to the sum of the average daily volumes over all twelve months.
4. The national fractional profile was calculated as the average over the state fractional
profiles.
:It would be possible at this point to generate a weighted average national profile where different weighting factors
could be assigned by state. For example, it would be possible to generate a population-weighted or VMT-weighted
national profile instead of weighting each state equally.
2The national fractional profiles were calculated from national average volumes initially, but upon finding that, in
the VTRIS data, the state of New Jersey had volumes that were typically ten times higher than the other states. This
forced the national VTRIS profile to be overly influenced by the trends in the New Jersey data. By averaging the
state fractional profiles, the national profile is more influenced by the trends in the state-level data, rather than the
magnitude of the volumes from one state to another. In order to be consistent in the analysis of both the VTRIS
and TVT data, the TVT national profiles were recalculated from the averages of the state fractional profiles. A more
detailed discussion is provided under the "Calculation of National Average Profiles" section.
D-6
-------
Calculation of National Average Profiles
The national temporal profiles were initially calculated from the average of the state vol-
umes. However, the VTRIS volumes from the state of New Jersey were substantially greater in
magnitude than the other states, so the national profile that was generated was almost exactly that
for the state of New Jersey. Figure D-4 below is an example the over-influence of the large vol-
umes on the national profile. In particular, the point for July shows a spike in the national profile
that is representative of only the state of New Jersey. The other two states, Arizona and Iowa, do
not show the same July spike.
State and National Monthly VTRIS Total Volume Profiles
for Rural Minor ArteriaIs (Class 06)
Feb Mar Apr May Jun Jul Aug Sep Oct Nov
-AR -»-|A —*-NJ -"—Nat
Figure D-4. National VTRIS monthly profile calculated from averages of state volumes.
To capture the temporal changes from hour to hour, from one day in the week to the next,
or from one month to the next, the national temporal profiles presented below were calculated as
averages of the normalized state temporal profiles. This procedure was performed to eliminate
the influence of single states, such as that of New Jersey in Figure D-4. Figure 5 illustrates the
revised national average VTRIS monthly profile calculated as the average of the state profiles to
compare with Figure 4. In this figure, the July spike in the national average profile was dimin-
ished, and so considered more representative of the larger sample average.
D-7
-------
State and National Monthly VTRIS Total Volume Profiles for Rural Minor Arterlals
(Class 06)
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
-AR -•—IA -A-NJ -"-Nat
Figure D-5. National VTRIS monthly profile calculated from averages of state profiles.
D-8
-------
D-3. RESULTS OF REGIONAL VARIABILITY AND TVT/VTRIS NATIONAL
AVERAGE COMPARISONS
While it can be demonstrated that some states were statistically significantly different
from the national average by road type and for some hourly, day of week, and monthly profiles,
no consistent regional grouping could be identified to define regional profiles.
The state-to-state variability was examined by conducting a chi-square goodness-of-fit
test, and we also determined if a state's volumes were often more than one standard deviation
away from the national average. Both methods reveal that there are typically a number of states
that would be considered to be significantly different than the national average.
Hourly Profiles
Following is a plot of the TVT hourly state profiles for rural interstates along with the
TVT national average in bold black and the VTRIS average in bold red. This demonstrates that
while many states demonstrate apparent differences from the national average, these differences
are not large.
State and National Average Hourly TVT Profiles for Rural Interstates (Class 01)
0.01
0.00
1234567
10 11 12 13
Hour
14 15 16 17 18 19 20 21 22 23 24
— »— AK
IN
-A- NO
-*-OR
-*-WA
-•-AL
KS
-X-ND
-*-PA
-•-wi
AR
*-KY
-*-NE
Rl
— 1— WV
CA
*-LA
-•-NH
— 1— SC
WY
-*-CT
-•-MA
— 1— NJ
— SD
^^Nat
-•-FL
Ml
NM
TN
^^VTRIS
— 1— GA
MN
NV
»— TX
IA
— MO
— »-NY
-•-UT
— ID
— »-MS
-•-OH
-A-VA
IL
MT
-A- OK
-X-VT
Figure D-6. State and national average TVT hourly profiles for rural interstates.
In Figure D-6, there is a clear diurnal pattern that all the states follow, with the greatest
variance in the early morning commute hours between 6 a.m. and 10 a.m. With the exception of
D-9
-------
New Jersey, the state profiles follow the national profile rather closely during the mid-day hours,
followed by a slightly greater variance around the evening commute hours between 5 p.m. and 6
p.m. Figure 7 below is a plot of the national profile with one-standard deviation error bars, and
Figure 8 shows the national profile generated from the TVT data along with the VTRIS national
average, both with 90% confidence intervals. The 90% confidence intervals on the TVT na-
tional average are quite small, given that the national profile is calculated as the average of the
state profiles, which is not a substantially large sample size. The confidence intervals around the
VTRIS data are larger, reflecting primarily the smaller sample of states used in the average.
National Average Hourly TVT Profile Plus or Minus 1 Standard Deviation
for Rural Interstates (Class 01)
0.08
0.07 -
0.06
0.05
°-04 "
0.03
0.02
0.01
o.oo
Figure D-7. National average TVT hourly profile plus or minus 1 standard deviation.
D-10
-------
National Average Hourly TVT Profile and 90% Confidence Intervals
for Rural Interstates (Class 01)
0.08
0.07
0.06
0.05
o 0.04
0.03
0.02
0.01
0.00
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
VTRISAvg
-TVTAvg
Figure D-8. National average TVT and VTRIS hourly temporal profile with uncertainty.
Despite the seemingly close agreement between the state hourly profiles and that of the
national average, there were only 10 out of 44 states that were not statistically significantly dif-
ferent from the national average. This is primarily because the categorical chi square tests for
even small differences in profiles, so only a few hours need to be different to see a statistical dif-
ference. Following in Table D-2 are the significance levels as well as the number of hours a state
value was more than 1 standard deviation away from the national average for rural interstates.
Table D-2. Chi Square statistical tests for difference from national average for rural interstates
(Low p-values and large number of hours different than the mean are statistical indicators of dif-
ference).
Number of Hours Greater than 1 Number of Hours Greater than 1
State p-value Standard Deviation State p-value Standard Deviation
Away from the National Mean Away from the National Mean
AK
AL
AR
CA
CT
FL
GA
IA
ID
IL
0.028
0.044
0.000
0.000
0.000
0.171
0.000
0.992
0.191
0.000
18
0
10
8
6
2
4
0
1
5
NE
NH
NJ
NM
NV
NY
OH
OK
OR
PA
0.002
0.000
0.000
0.171
0.000
0.000
0.073
0.000
0.000
0.000
3
14
11
2
10
10
0
7
13
5
D-ll
-------
Number of Hours Greater than 1 Number of Hours Greater than 1
State p-value Standard Deviation State p-value Standard Deviation
Away from the National Mean Away from the National Mean
IN
KS
KY
LA
MA
Ml
MN
MO
MS
MT
NC
ND
0.000
0.916
0.000
0.000
0.000
0.001
0.000
0.000
0.620
0.000
0.522
0.819
11
2
8
13
19
4
6
10
0
17
0
5
Rl
SC
SD
TN
TX
UT
VA
VT
WA
Wl
WV
WY
0.041
0.000
0.000
0.000
0.043
0.001
0.000
0.000
0.000
0.002
0.000
0.469
0
0
11
7
2
8
1
16
12
0
8
4
Day of Week Profiles
The day of week profiles show a similar pattern in that individual States may differ from
the national average for a day of the week, but overall one cannot discern a clear regional pattern
from the data. Figure D-9 shows a comparison of the TVT national and individual state profiles
by the day of week on rural interstates. Figure 10 shows the national average TVT day of week
profile along with the VTRIS average profile, both with 90% confidence intervals.
State and National Average TVT Day Of Week Profiles for Rural Interstates (Class 01)
— »-AK
•-IN
-A- NO
-*-OR
-*-WA
-•-AL
KS
-X-ND
-*-PA
-•-Wl
AR
K-KY
-*-NE
-•-Rl
— 1— WV
CA
-*-LA
-•-NH
— 1— SC
WY
-*-CT
-•-MA
— 1— NJ
— SD
^^Nat
-•-FL
Ml
NM
TN
^^VTRI
— 1— GA
MN
NV
TX
3
IA
— MO
— »— NY
-•-UT
— ID
— »-MS
-•-OH
-A-VA
IL
MT
-A- OK
-*-VT
Figure D-9. State and national average TVT day of week profiles for rural interstates.
D-12
-------
National Average Day Of Week TVT Profile and 90% Confidence Intervals
for Rural Interstates (Class 01)
0.18
0.16
0.14
0.12
0.10
0.08
0.06
0.04
0.02
0.00
Sun
Mon
Tue
Wed
Thu
Sat
VTRIS Avg
-TVT Avg
Figure D-10. National average TVT and VTRIS day of week temporal profile and uncertainty on rural
interstates.
In general, the day of week temporal profiles are fairly consistent from state to state. Fig-
ure 11 displays the day of week temporal profiles for urban interstates by state with the national
average in bold black and the VTRIS average in bold red. Again there is quite close agreement,
both from state to state and with the VTRIS average.
D-13
-------
State and National Average Day Of Week TVT Profiles for Urban Interstates (Class 11)
0.18
0.16
0.14
0.12
0.08
0.06
0.04
0.02
Sun
Mon
Tue
Wed
Sat
— »— AK
•-IL
-A-MT
-X-OK
-SK-VT
-•-AL
IN
-X-NC
-«- OR
-•-WA
AR
K-KS
-*-ND
-•-PA
— 1— Wl
-X-CA
-*-KY
-•-NE
— 1— Rl
WV
-*-CT
-•-LA
— 1— NH
— SO
— *-WY
-•-FL
t— MA
NJ
SD
^^Nat
— 1— GA
Ml
NM
TN
^^VTRk
HI
— MN
— »-NV
-•-TX
3
— IA
— »-MO
-•-NY
-A-UT
ID
MS
-A- OH
-X-VA
Figure D-l 1. National average and state temporal profiles by day of week.
There were two instances when the data from the state of Oklahoma was substantially
different from all others in the daily profiles, presented in Figures D-l2 and D-l3 below.
State and National Average Day Of Week TVT Profiles for Rural Principal Arterials (Class 02)
0.3
0.25 -
0.2 -
u 0.15 -
0.05 -
Sun
Mon
Tue
Wed
Thu
Fri
Sat
— •— AK
•-IL
-A-MT
-X-OK
-*-VT
-f-
-*-
-*-
-•-
AL
IN
NO
OR
WA
AR
K-KS
-*-ND
-•-PA
— 1— Wl
CA
-*-KY
-•-NE
— 1— Rl
WV
-*-CT
-•-LA
— 1— NH
^SC
-A-WY
-•-FL
t— MA
NJ
SD
^^Nat
-1-
-^
GA
Ml
NM
TN
VTRIS
HI
— MN
— »-NV
-•-TX
— IA
— »-MO
-•-NY
-A-UT
ID
MS
-A- OH
-X-VA
Figure D-12. Illustration of a different day of week temporal profile for Oklahoma on rural principal arterials.
D-14
-------
State and National Average Day Of Week TVT Profiles for Rural Minor Arterials (Class 06)
— »-AK
•-IL
-A-NC
-H-OR
-*-WA
-•-AL
IN
-*-ND
-SK-PA
-•-wi
AR
*-KS
-*-NE
-•-Rl
— 1— WV
K-CA
-*-KY
-•-NH
— 1— SO
WY
^ie-CT
-•-LA
— 1— NJ
— SD
^^Nat
-•-FL
Ml
NM
TN
^^~VTRI
— 1— GA
MN
NV
TX
3
HI
— MO
— »— NY
-•-UT
— IA
— »-MS
-•-OH
-A-VA
ID
MT
-A- OK
-X-VT
Figure D-13. Illustration of a different day of week temporal profile for Oklahoma for rural minor arterials.
Upon further investigation, we found that there were two sites, one at facility class 02
(urban principal arterials) and one at facility class 06 (rural minor arterials) that were both sub-
stantially different than the other sites. These sites also had substantially greater volumes than
at the other sites. Figures D-14 and D-15 below illustrate the average daily volumes by site for
these roadway types for Oklahoma. Note that these sites fulfilled the completeness criterion that
there be a complete set of seven daily averages, but they do not exhibit typical weekly patterns.
The greater volumes at these sites overly influence the state average profile. A further investiga-
tion into these two sites may reveal special circumstances that may justify removing them from
the analysis.
D-15
-------
Oklahoma Average Daily Volumes by Site for Rural Principal Arterials
20000
18000
16000
14000
12000
= 10000
Day of Week
-OOC016 -"-OOC042 -A-OOC044
Figure D-14. Daily average volumes by site for Oklahoma, roadway class 02.
Oklahoma Average Daily Volumes by Site for Rural Minor Arterials
45000
40000
35000
30000
25000
20000
15000
10000
5000
A
Day of Week
-OOC025
-OOC045
Figure D-15. Daily average volumes by site for Oklahoma, roadway class 06.
D-16
-------
Monthly Profiles
The monthly temporal profiles also show the same general trend among the states with a
slightly higher volume during summer months. Figure D-16 below illustrates the monthly pro-
files for rural interstates.
State and National Average Monthly TVT Profiles for Rural Interstates (Class 01)
0.12
0.00
May
Aug
Sep
— »- AL-
MA
-A- NY
-X-WV
-•-AR
Ml
-X-OH
-3K-WY
-A-CA
K-MN
-*-OR
^^Nat
-•-
^~
CT
MS
PA
VTRIS
IK GA
-•-MT
— 1— Rl
-•-LA
*— NC
SC
— 1— ID
ND
SD
— »-
IL
NE
TX
— KS
— »-NH
-•-VA
KY
NJ
-A-VT
Figure D-16. State and national average TVT monthly profiles for rural interstates.
The comparison between the VTRIS and TVT summary profiles indicates small differ-
ences in the overall profile for this road type. The uncertainty ranges in Figure D-17 were cal-
culated based on the state-to-state variability, and so may not include all of the uncertainty in the
data ranging from site selection bias and other site-to-site variability.
D-17
-------
National Average Monthly TVT Profile and 90% Confidence Intervals
for Rural Interstates (Class 01)
0.12
0.10
Jan Feb Mar Apr May Jun
Aug Sep Oct Nov Dec
0.02
0.00
Figure D-17. National average TVT and VTRIS monthly temporal profile and uncertainty.
Summary results for other road types demonstrate that similar conclusions could be made
for all road types. Smaller road types have fewer states providing data and therefore higher
uncertainty levels. State-to-state variability in the TVT temporal profiles exists, but no consistent
regional pattern could be discerned in the temporal profiles. Likewise, the VTRIS total volume
temporal profiles were very similar to the TVT profiles for all road types indicating that VTRIS
could be used for the vehicle mix fractions applied to the TVT total volume trends.
D-18
------- |