9 TRANSPORTATION ANALYSIS PERFORMANCE

MEASURES

9.1 Introduction

Transportation analysis performance measures, sometimes referred to as measures of

effectiveness (MOEs), are quantitative estimates on the performance of a transportation facility,
service, program, system, scenario or project with respect to policies, goals and objectives. Some
common performance measures used in traffic engineering include v/c ratio, level of service
(LOS), crashes, vehicle delay, travel time, mode share and capacity. Performance measures in
this chapter focus on the objectives of mobility and safety.

Performance measures can be based on empirical observations/data measurements of existing

conditions or may be outputs of models that estimate or predict the performance of potential
future scenarios, programs or alternative strategies.

Performance measures typically have some type of established threshold or target value or rating
which defines the acceptable conditions for a facility. Any case where conditions do not meet
that level is defined as a deficiency or need that should be reviewed. The term ‘need’ as used by
transportation professionals has generally been defined as any case where the current or planned
facility conditions fall below an established threshold.”

The greater the deviation of the measured value from the performance threshold, the greater the
need. Thresholds provide a critical element of the decision-making framework for assessing
deficiencies and improvement alternatives since they are developed to maximize overall system
performance while limiting liability to the agency responsible for construction, operations and
maintenance. Thresholds may be known as goals, targets, or benchmarks. Thresholds may be
adopted by a jurisdiction as part of a plan or policy.

Most road authorities (state, county or city) maintain adopted performance standards for
operational efficiency that identify specific performance thresholds. It is important to identify all
applicable performance standards and corresponding performance measures and thresholds for
study roadways to provide a basis for evaluating the results of transportation analysis and to
determine if project goals and objectives are being achieved. Methods of calculation or tools may
also be prescribed.

9.1.1 Selecting Performance Measures

Performance measures to be used in a system plan, corridor study, development review, or

project alternatives analysis are driven by the goals and objectives of the project and are
identified during scoping and methodology development. Appendix 9A is provided as a general
guide to aid in the consideration of potential transportation analysis performance measures by
plan or project type, including RTPs, TSPs, MMAs, Facility Plans, Development Review, and
NEPA/project development. For each type of study, potentially applicable performance measures
are noted and identified as best practice/recommended, supplemental, or for screening purposes.

Analysis Procedure Manual Version 2 9-1 Last Updated 12/2019

Performance measures may require varying levels of efforts depending on factors such as project
type and tools used. Refer to Chapter 2 for level of effort information.

Some analysis performance measures are required. For example, state highway project v/c ratios
are needed in order to compare the performance of alternatives with ODOT Highway Design
Manual (HDM) mobility thresholds. Other analysis performance measures are often necessary,
depending on the needs of the project. These selected performance measures become project
evaluation criteria by defining them specifically to the project. In addition, project-specific
thresholds and desired confidence or significance levels may be defined.

Performance measures should be SMART: Specific, Measurable, Agreed upon, Realistic, and
Time-bound. See Chapter 10 for guidance on the process of developing project evaluation
criteria and performance measures. Performance measures need to be sensitive enough to
differentiate between analysis years and alternatives, scenarios or options. For example, a highly
congested area with v/c’s in excess of 1.0 and LOS F will not be sensitive to an increase in
volume. In this case, different/additional measures would be needed such as travel time, safety,
and reliability.

There is no one-size-fits-all performance measure that can address all the policies or objectives
of a plan or project. Many performance measures address only one dimension of a problem while
ignoring other important considerations. For example, a ratio or percentage based performance
measure such as v/c ratio by itself does not indicate the number of users affected. Two roadways,
one with a high volume and one with a low volume, may both have the same v/c ratio, but the
high volume roadway affects more users than the low volume roadway. Multiple performance
measures are typically needed.

The applicability or priority of performance measures depends on the purpose, need, goals and
objectives of the project or plan, as well as on the facility and area type. In some cases the same
performance measure can address multiple objectives. For example travel time can be used to
assess emergency vehicle trips, or freight, or other modes. The number of performance measures
chosen for any particular aspect for a project should be minimized. Too many performance
measures for a given area may create conflicts, confusion, unnecessary work, and may result in
measures not being used for that decision process. Some measures may not be clearly
understandable to the desired audience or practically creatable based on data and tools available.

A matrix of typical analysis performance measures including definitions, purpose, modes, level
of resolution, data and tool requirements is available in SPR Report 716 1.

1
Development and Sensitivity Testing of Alternative Mobility Metrics, SPR 716, ODOT, John P. Gliebe and James
G. Strathman, March, 2012, Table 3.1.

Analysis Procedure Manual Version 2 9-2 Last Updated 12/2019

9.1.2 Purpose of this Chapter

Performance measures in this chapter are grouped into categories. Transportation can be
measured in terms of its primary functions such as safety, accessibility and mobility. It can also
be measured in terms of its impact or consequence, such as on the environment, and
socioeconomics. It should be noted that while the performance measures identified below are
assigned to a single primary category, some measures relate to multiple objectives and
categories.

This chapter is limited to performance measures commonly reported out using APM methods
and tools. These measures are used in plans and projects to identify needs, compare scenarios
and alternatives, and identify benefits and impacts. The chapter focuses on facility level
performance measures. System level performance measures generated by APM tools are
discussed at a higher level. The performance measures covered are:

• Mobility
• Reliability
• Level of Service (LOS)
• Accessibility
• Safety
• Other Multimodal Performance Measures
• Infrastructure

The performances measures contained in this chapter are not an exhaustive list, but focus on
those that are the most widely used and practical. Measures which have a good potential for
application for a range of studies are also discussed. TPAU can provide assistance in selecting
appropriate analytical performance measures for a specific project. A given project will use only
a small subset of all possible measures. This chapter provides measure definitions, calculations,
strengths and weaknesses, but leaves the application of performance measures to other
referenced chapters. The use of performance measures to evaluate alternatives is discussed
further in Chapter 10. For a broader discussion on mobility performance measures, see FHWA
Traffic Analysis Tools Volume VI (1).

Not addressed in this chapter

• Factors that contribute to or are components of performance but are not typically reported
out as stand-alone performance measures. Although not performance measures per se, in
many instances these can provide additional useful information on the causes behind
performance, which helps to understand or interpret the performance measure result. This
includes analysis outputs used as inputs into performance measure calculations performed
by other methods. For example, forecasted traffic volumes and speeds which are used to
report out air quality and noise performance, or predicted crashes or delays which are
used in economic analysis to report out reduced travel time or crash reduction cost
performance.
• Preliminary screening criteria or flags are more intermediate in nature, such as those that
are used as inputs into following steps and are not typically reported out as performance

Analysis Procedure Manual Version 2 9-3 Last Updated 12/2019

 measures, for example preliminary signal warrants or turn lane criteria. Refer to Chapter
10.
• Performance measures or evaluation criteria produced using methods or tools outside of
the APM. For example, measures produced for managing the ODOT TSMO program
such as average time to clear an incident. This also includes right of way, construction
cost, funding, economics, design criteria, and environmental impacts.
• Performance measures relating to broad high level policy areas that are not specifically
transportation-related, such as economic vitality; land use; environmental stewardship;
quality of life, livability and health; equity; and funding and finance.
• Agency key performance measures (KPMs)/benchmarks/goals/targets for agency-wide or
policy/strategic planning/investment strategies or monitoring purposes such as
performance measure reporting required by the FAST Act; for example, ODOT Key
Performance Measures. Refer to Chapter 18 for more information on the performance
management planning and programming.

The broader KPMs are not comparable to analysis performance measures because they are
different in purpose and resolution, and are likely to be based on different measurements, tools,
level of aggregation, networks, assumptions, definitions, variables, data sources, formulas,
and/or time periods. In contrast, analysis project alternatives are typically smaller in scale with
greater resolution, focusing on study area roadway sections and intersections. Large scale
performance measures would not be as useful on smaller projects and plans such as small city
TSPs because they would not be likely to show a significantly measurable change in order to
make comparisons useful.

9.2 Mobility

Mobility refers to the movement of both people and goods regardless of mode. Mobility
performance relates to both supply and demand, as affected by land use and other policies.
Supply could include the road network, transit routes, bicycle lanes, or any other modal
infrastructure. Demand is the rate of flow which could include total persons, motorized vehicles,
transit vehicles, etc., desiring to be able to traverse a point or section over a period of time such
as an hour or a day. For additional detailed information on multimodal mobility related
performance measures refer to HCM 6.

9.2.1 Volume to Capacity Ratio

The principal performance measure ODOT uses when evaluating motor vehicle operating
characteristics on the state highway system is the volume to capacity (v/c) ratio, which is a
measure of how close to capacity a roadway is operating. It reflects the ability of a facility to
serve motorized vehicle traffic volume over a given time period under ideal conditions such as
good weather, no incidents, no heavy vehicles, no geometric deficiencies. The volume to
capacity ratio is the degree of utilization of the capacity of a segment, intersection or approach.
The v/c ratio is not defined over 1.0. Under those conditions it is considered to be a demand to
capacity ratio. A lower ratio indicates smooth operations and minimal delays. As the ratio

Analysis Procedure Manual Version 2 9-4 Last Updated 12/2019

approaches 1.0, congestion increases and performance is reduced. At 1.0 the capacity is fully
utilized.

For example, when v/c equals 0.85, 85 percent of a highway’s capacity is being used ; 15 percent
of the capacity is still theoretically available. However, as the v/c ratio approaches 1.0, flow
becomes unstable, speeds decrease, and bottlenecks can easily occur.

Performance measures
• Critical Intersection v/c ratio Xc (signalized intersections)
• Intersection Approach v/c ratio (unsignalized intersections)
• Segment v/c ratio (freeways, uninterrupted flow multilane highways and two-lane
highways)
• Weave, merge, and diverge v/c ratio (freeways and uninterrupted flow multilane
highways)

Example evaluation criteria

• Number of locations exceeding applicable v/c ratio standards
• Percent of intersections operating at V/C > 1.0

The generic formula for v/c ratio is as follows.

𝑣𝑣
= 𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣/𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐
𝑐𝑐
There are other variations on calculating the v/c ratio. For example, for signalized intersections,
Xc is used for total intersection v/c ratio as shown below. Refer to the HCM for fully detailed
procedures.

Critical Intersection Volume to Capacity Ratio (for signalized intersections)

𝐶𝐶
𝑋𝑋𝐶𝐶 = � � � 𝑦𝑦𝑐𝑐,𝑖𝑖
𝐶𝐶 − 𝐿𝐿
𝑖𝑖∈𝑐𝑐𝑐𝑐

With
𝐿𝐿 = � 𝑙𝑙𝑡𝑡,𝑖𝑖
𝑖𝑖∈𝑐𝑐𝑐𝑐

Where
𝑋𝑋𝐶𝐶 = critical intersection volume to capacity ratio
𝐶𝐶 = cycle length (sec)
𝑣𝑣𝑖𝑖
𝑦𝑦𝑐𝑐,𝑖𝑖 = critical flow ratio for phase 𝑖𝑖 =
(𝑁𝑁𝑠𝑠𝑖𝑖 )
𝐿𝐿𝑡𝑡,𝑖𝑖 = phase i lost time = 𝑙𝑙1,𝑖𝑖 + 𝑙𝑙2,𝑖𝑖 (sec)
𝑐𝑐𝑐𝑐 = set of critical phases on the critical path
𝐿𝐿 = cycle lost time (sec)

Analysis Procedure Manual Version 2 9-5 Last Updated 12/2019

 𝑣𝑣𝑖𝑖 = lane group flow rate for phase 𝑖𝑖
𝑁𝑁 = number of lanes for lane group 𝑖𝑖
𝑠𝑠𝑖𝑖 = lane group saturation flow rate for phase 𝑖𝑖

The v/c ratio can account for changes in either volume or supply (capacity). Volume is a measure
of the rate of flow of traffic expressed as the number of vehicles passing a given point on a
roadway over a specified time period, such as vehicles per hour or day. Volume is most
commonly reduced at a location or facility by adding alternative routes or connections which
may shift traffic to other routes. Other means of shifting volume include TDM or TSMO
measures such as ramp metering, traveler information, tolling or congestion pricing. Procedures
for developing traffic volumes are found in APM Chapter 5 for Existing and Chapter 6 for Future
Year.

Capacity is the supply side measure of the ability of a facility to carry traffic. It is the maximum
number of motorized vehicles per hour that can travel on a particular stretch of roadway under
relatively ideal conditions such as proper lane widths, no parking, no bus blockages, etc.
Capacity is a function of a number of variables including number of lanes, lane width, shoulder
width, presence and type of control devices, free flow speed, and other features. Capacity may be
calculated by HCM methods or measured in the field in locations and conditions where demand
exceeds capacity. Procedures for calculating capacity and v/c ratios are found in the APM
chapters on segments and intersections and are primarily based on methodologies in the
Highway Capacity Manual and implemented by various software tools.

ODOT uses v/c-based measures for reasons of application consistency and flexibility,
manageable data requirements, forecasting accuracy, and the ability to aggregate into area-wide
targets that are fairly easy to understand and specify. In addition, since v/c is responsive to
changes in volume as well as in capacity, it reflects the results of demand management, land use
and multimodal policies. Other advantages of v/c ratio include:

• Standardized calculation methodologies and tools

• Easily applied and forecasted
• Planning level methods are available to estimate segment v/c ratios. Volumes are
estimated using AADTs along with K30 factors and directional factors. Capacity estimates
can include the use of default values in estimating v/c ratios with the results reported out
as below, near, or at capacity, as example, HERS-ST performs this level of v/c ratio
analysis (refer to Chapter 7). For urban signalized arterials, segment capacity can be
estimated using approximate green time to cycle time (g/c) ratio assumptions.
• Can be calculated for segments, intersections, approaches, and turn movements
• Travel demand models calculate a link-based demand to capacity ratio (d/c). Refer to
Section 9.2.5 on model based demand to capacity ratios.

Requirements/Limitations
• Does not directly apply to or address safety, non-motorized vehicle modes, operational
improvements, and other policy objectives often under consideration because these

Analysis Procedure Manual Version 2 9-6 Last Updated 12/2019

 aspects of the transportation system cannot be directly measured in terms of vehicle
demand and vehicle capacity.
• Identifies when capacity is exceeded, but does not address the extent or duration of
congestion or queue spill-back effects. By definition, the volume of traffic using a
roadway cannot exceed the roadway’s capacity. When demand exceeds capacity, a
demand-to-capacity (d/c) ratio may be used (see section on Demand to Capacity Ratio).
A d/c ratio that exceeds 1.00 indicates that more vehicles would use a roadway in a given
time period if capacity constraints were not present.
• Is focused on a recurring peak period and does not address non-recurring congestion such
as from incidents, weather, or special events.

9.2.2 Oregon Highway Plan Mobility Targets

ODOT has adopted specific v/c ratio thresholds for identifying current and future needs in the
Oregon Highway Plan (OHP) which are used for identifying needs in planning. These are
different from the performance thresholds for project design in the Highway Design Manual
(HDM), which are accepted by FHWA for design and need to be lower than the planning need
threshold in order to allow for the project to have a design life.

Volume to capacity ratio was selected as the performance standard for motor vehicle mobility on
state highways in the Oregon Highway Plan (OHP) after an extensive analysis of candidate
highway performance measures. The review included the effectiveness of the measure to achieve
other policies (particularly OHP Policy 1B, Land Use and Transportation), implications for
growth patterns, how specifically ODOT should integrate transportation policy with land use,
flexibility for modifying targets, and the effects of Portland metro area targets on the major state
highways in the region.

Targets for state highway motorized vehicle mobility needs are established in the current OHP
Policy 1F. Tables 6 and 7 within Policy 1F contain the v/c ratio targets for various combinations
of highway classifications and surrounding land uses, with Table 7 applying to the Portland
metropolitan area and Table 6 applying to the remainder of the state.

The targets vary the priority for mobility according to facility, area and designation type;
mobility is a high priority on freeways, expressways and freight routes, but is a lower priority on
District highways or local interest roads in Special Transportation Areas (STA) and Metropolitan
Planning Organizations (MPO). It should be noted that the text within Policy 1F contains
exceptions to the targets listed in these tables and, therefore, must be consulted as well.
Furthermore, the OHP Registry of Amendments webpage should be checked for amendments to
the OHP mobility policy where alternative mobility targets have been adopted; for an example
refer to the report US 101 Seaside Alternate Mobility Standards.

The analyst should refer to OHP Policy 1F for appropriate application of the OHP
mobility targets in specific contexts. For plan amendment applications also refer to
TPR 0060.

Analysis Procedure Manual Version 2 9-7 Last Updated 12/2019

9.2.3 Oregon Highway Plan Alternative Mobility Targets

The v/c ratio targets were generally designed to provide continued operation in an under capacity
condition. Increasingly in urban areas, there are roadways that are projected to be over-capacity,
or that are currently operating in an over-capacity mode. Circumstances exist where v/c targets
cannot reasonably be met due to financial, environmental or land use constraints. In these
circumstances, where it is not feasible or desirable to make infrastructure investments to fully
accommodate the existing and projected vehicular demand, it is possible to explore alternative
mobility targets.

If meeting OHP v/c ratio targets is not practical or feasible due to financial, environmental or
land use constraints or impacts, OHP Action 1F.3 contains provisions for creating alternative
mobility measures and targets through a planning process and adoption by the Oregon
Transportation Commission (OTC). Adjustments to the OHP targets may include changing the
v/c ratio target (increase or decrease), changing the analysis methodology (e.g., from 30th highest
hour to average annual traffic volumes or adjusting peak hour factors), and/or acknowledging
that a facility will likely operate at capacity for more than just a single peak hour. Alternative
(non v/c-based) performance measures may involve other analysis methods that address safety
performance, travel time reliability and delay.

The process for consideration of alternative mobility targets is detailed in the Planning Business
Leadership Team PBLT Operational Notice PB-02. This process involves the participation,
commitment and mutual agreement of local and regional jurisdictions and includes exploring a
variety of transportation-related solutions, including a number of system and demand
management activities to maximize the efficiency of transportation movements and to identify
solutions that are realistic to implement and have the potential to be effective. Under most
circumstances, local jurisdictions must adopt appropriate local policies, codes and ordinances
that are necessary to help support and implement the alternative mobility target and achieve other
policy and performance objectives.

In some cases such as a rural interchange area management plan, more restrictive alternative v/c
ratio targets may be adopted as part of OHP Action 1F.4. More restrictive targets may help to
maintain mobility in an identified area. This can be an effective tool where it is desirable to
further preserve a significant investment, such as in the vicinity of an interchange.

Developing OHP Alternative Performance Targets or Measures

The following v/c-based methodology is recommended as a first option when developing

alternative mobility targets for state highways outside the Portland Metro area. OHP policy and
current analysis practices use a v/c-based methodology as the initial measure to standardize and
simplify implementation through a quantifiable, consistent and reproducible measure. Where v/c-
based approaches may not meet all needs and objectives, developing alternative mobility targets
using non v/c-based measures may also be pursued. Any alternative mobility target per the OHP,
including new methodology, will not be final until adopted by the OTC.

Analysis Procedure Manual Version 2 9-8 Last Updated 12/2019

 1. In cases where v/c is forecasted to be greater than the OHP mobility target but less than
capacity (v/c = 1.0) during the design hour using standard analysis procedures, establish
the proposed alternative target consistent with the v/c values used in the OHP (0.75, 0.80,
0.85, 0.90, etc.).

2. In cases where v/c is forecasted to be greater than or equal to capacity during the
design hour using the standard analysis procedures evaluate the actual peak hour traffic
volume for future year design hour projections rather than expanding the peak 15 minutes
to be the design hour traffic volume (e.g. peak hour factor) for projection purposes. If v/c
is less than 1.0, establish the proposed alternative target.

3. In cases where v/c is forecasted to be greater than or equal to capacity during the
design hour using the actual peak hour projection of traffic and in areas where design
hours are affected by high seasonal traffic volumes, evaluate the Annual Average
Weekday PM Peak as the future year design hour rather than the 30th highest hour. If v/c
is less than 1.0, establish the proposed alternative target.

4. In cases where v/c is forecasted to be ≥1.0 using the Annual Average Weekday PM
Peak as the future design hour, determine the duration of the period during which the
future Annual Average Weekday PM Peak hour will have a v/c ≥1.0. Establish the
proposed alternative target by increasing the number of hours that v/c can be ≥1.0 (i.e.,
v/c ≥1.0 for not more than 1 hour, or not more than 2 hours, etc.).

If a v/c-based mobility measure does not by itself meet the needs of the jurisdiction, the state or
the particular facility under consideration, then it is reasonable to explore non v/c-based
measures for defining mobility on the state highway system. At a minimum, all non v/c-based
measures must:

1. Be consistent with OHP Policy 1F, with particular attention to Actions 1F.1 and 1F.3;

2. Follow the PBLT Operational Notice PB-02 Attachment A Checklist; and

3. Develop a measurable and defensible target value, with defined geographic limits and
a defined analysis methodology that can be compared between alternatives, recognizes
data needs, availability and quality, and considers requirements for implementation
including the availability of analysis tools, staff responsibilities and associated costs.

Recognize that, even when exploring non v/c-based measures, there may still be advantages to
keeping v/c measures as well. The v/c ratio along with other measures provides a complete
picture of operations.

9.2.4 Highway Design Manual Mobility Guidelines

Motor vehicle mobility thresholds for design of modernization projects are identified in Exhibit
10-1 of ODOT’s HDM. These v/c ratios (the functional equivalents of the LOS standards in the
American Association of State Highway and Transportation Officials [AASHTO] Green Book)
represent the level of operation for which state facilities are expected to be designed and are

Analysis Procedure Manual Version 2 9-9 Last Updated 12/2019

intended to be applied to an analysis year occurring 20 years beyond the year of completion.
These thresholds are applicable to future build alternatives on state highways associated with all
project types except Traffic Impact Studies associated with development, unless an interchange
or interstate freeway is involved. It should be noted that for ramp terminals, the HDM mainline
maximum v/c ratio is the standard that applies. There is no equivalent ramp terminal v/c ratio in
the OHP as there is in the HDM.

Exhibit 9-1 illustrates the appropriate sources of adopted mobility performance measure
standards for different project types.

Exhibit 9-1 Sources of Adopted Mobility Targets/Standards for State Highways by Study
Type
Corridor and
TIS/TIA Projects TSPs Refinement Plans

Existing Conditions OHP OHP OHP OHP

Future No-Build OHP OHP OHP OHP

HDM
Future Modernization
OHP HDM (OHP in HDM
Build(s) Portland Metro
Area1)
1
In the Portland metropolitan area, future modernization build alternatives on state highways are scoped and
analyzed in corridor plans, refinement plans or projects rather than as part of TSPs.

HDM mobility thresholds are generally more restrictive than the OHP mobility targets; however,
there is a design exception process that allows variation from the HDM when appropriate.
Transportation System Plans (TSPs) generally identify needs and the function, mode, location,
and parameters (e.g. number of lanes) of solutions. The precise location, alignment, and
preliminary design of solutions is typically deferred to refinement studies or project
development.

HDM FHWA-ODOT MOU

In order to be used as baseline standards for future project design, alternative mobility targets
being considered as an amendment to the OHP must be established in coordination with FHWA.
This process is described in the Memorandum of Understanding (MOU) between ODOT and
FHWA, provided as an attachment to PB-02. Through this process, the alternative mobility target
may be adopted as an amendment to the HDM.

Analysis Procedure Manual Version 2 9-10 Last Updated 12/2019

9.2.5 Supplemental Vehicle Mobility Measures

Many of the mobility analysis procedures summarized in the APM have direct (or equivalent) v/c
ratio results for performance assessment. The compliance with the appropriate target (maximum
v/c ratio thresholds defined in the OHP) is the first tier of the evaluation. The other category of
performance measures focuses on travel time/speed, including progression analysis, arterial
analysis and selected outputs of many simulation models. The vehicle speed outcomes can be
compared to target or design speeds to assess relative benefit, but there is no direct comparison
with v/c ratio in these analyses. It is recommended that these types of measures be used in
conjunction with either intersection or segment analyses that do have v/c ratio related outcomes
to compare to mobility targets.

Typical travel demand model-based performance measures are calculated using model generated
outputs that yield general system performance of the scenario. Scenarios would be considered
relatively the same if there is no significant difference in the performance measure (less than
10%) because of the model’s limited accuracy. Performance measures can be system-wide or
segregated into select facilities, corridors, areas, or zones.

Quantity of Travel

Quantity of travel represents the amount of use of a facility or service. It is both a performance
measure and an input into the calculation of other performance measures. Quantity of travel is
usually expressed as the number of motorized vehicles, persons, pedestrians, bicyclists, or transit
vehicles per unit of time. Methods to estimate the quantity of travel range from simple historical
trends to cumulative analysis to complex urban, regional or statewide travel demand models.

For more information refer to APM Chapters 6 and 7 and the ODOT Planning Section Technical
Tools webpage.

Percent Change in Volume

Percent change in motor vehicle volume is used to compare different scenarios or alternatives
such as No Build versus Build, different land use scenarios, or multiple Build alternatives. Build
alternatives that increase capacity and/or reduce travel time often result in network volume
changes, for example due to demand shifting between competing routes. Volume change is
commonly obtained by comparing similar segments across screenlines or difference plots
between two scenarios from a travel demand model, or can be estimated using deterministic
methods. Volume change can be used as a high level analysis or preliminary screening measure.
Changes of more than 10% are generally considered the minimal level of significance. Changes
of more than 20% have a large impact, depending on the absolute volume level.

Performance measures
• Design hour volume on segment or screenline

Analysis Procedure Manual Version 2 9-11 Last Updated 12/2019

Example evaluation criteria
• Percent change in volume crossing screenline
• Percent change in volume on link
• Diversions or neighborhood cut-through traffic due to temporary/permanent lane closures
including road diets, work zones, congestion, incidents.

Vehicle-Miles of Travel (VMT)

VMT is the amount of vehicle travel on a system in terms of both vehicle volume and distance.
VMT is the relationship of the total vehicle volume on the specified links multiplied by the total
link lengths. VMT is typically a system performance measure reported for large-scale, regionally
significant changes or regional/MPO areas and should generally only be used in high-level
planning analyses. Although VMT can be calculated on any facility, it is not typically reported
out. VMT is also calculated as part of energy analyses in Environmental Impact Statements.

Performance measure
• Vehicle-miles of travel on segment or facility

Example evaluation criteria

• Change in area/region VMT
• Change in facility/corridor VMT
• Change in segment VMT

𝑉𝑉𝑉𝑉𝑉𝑉 = 𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 × 𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙ℎ

VMT is typically reported as a daily value but may be specified as an average annual
value based on 365 days a year. The analyst should be aware that VMT can be
calculated based on different data sources, tools or methodologies. For example,
gasoline sales based VMT (when combined with average vehicle MPG), official VMT
from HPMS used in HERS (link-level, statewide, state-owned facilities only), RSPM
(all days average, household-based, all roads), or from a travel demand model
(average weekday, mostly state system, within model area only).

VMT is typically reported as an annual average daily value per segment:

𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷 𝑉𝑉𝑉𝑉𝑉𝑉 = 𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 × 𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿ℎ

Analysis Procedure Manual Version 2 9-12 Last Updated 12/2019

VMT is also commonly reported for an entire facility, system or subset of roads by summing
individual segment VMTs:

𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹 𝑜𝑜𝑜𝑜 𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆 𝑉𝑉𝑉𝑉𝑉𝑉 = � 𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆 𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉

For trucks,

𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 = 𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 × 𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙ℎ × % 𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡

Oregon historical VMT data at a state facility level or broad regional level, reported as part of the
Highway Performance Monitoring System (HPMS), may be obtained by contacting ODOT Road
Inventory & Classification Services.

Regional VMT within an urban area is a common travel demand model measure. Reporting can
be for the entire model area or for roads within a sub-area (e.g., UGB, MPO boundary), for all
trips or a portion of the trips (e.g., internal-internal (I-I) trips only, truck-only). Model produced
VMT may be reported by mode and by trip purpose. Even where total demand is the same, VMT
can increase due to changes in trip lengths, such as a scenario where trips lengthen when land use
growth is mainly on the fringe of the urban area, or increasing road congestion may result in
either shorter trips or forces trips to take alternate routes which may be longer.

VMT is closely related to both the demand and the supply side of the urban setting. Levels are
lower in communities that are more walkable and compact and in communities that have a strong
public transport system. Increasing population density can lower VMT as well, although
increased density may increase the VMT in the local area but may reduce the overall system
VMT. VMT can also drop due to economic downturns, when unemployment is high and people
have a smaller shopping budget. Vehicle operating costs including fuel costs, per mile fees and
vehicle MPG, can also significantly impact VMT. Population shifts or new population estimates
can change VMT trends significantly. Many of these factors are outside the agency’s control.

VMT results can be subject to misinterpretation as many factors can contribute to a particular
increase or decrease in the value. For example, a VMT increase could be due to more people
driving, but it also could be due to new growth on the fringe of an area with a subset of the
population having to drive longer distances.

The TPR requirement of VMT per capita is limited to internal trips only, even though models
can produce VMT per capita for all trips. The TPR measure can be skewed based on the relative
size of the model area, the proportion of external trips, or other individual characteristics of the
urban area such as demographics (i.e., high retiree population). The measure can also be skewed
when population forecasts change.

Person-Miles Traveled (PMT)

PMT is similar to VMT except that person travel is measured rather than vehicle travel. Person
travel includes motorized vehicle drivers, passengers, transit riders, rail passengers, pedestrians

Analysis Procedure Manual Version 2 9-13 Last Updated 12/2019

or bicyclists. For motorized vehicle travel, person-miles traveled is typically calculated as
follows:

𝑃𝑃𝑃𝑃𝑃𝑃 = 𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 × 𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙ℎ × 𝑣𝑣𝑣𝑣ℎ𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜

Performance measure
• Person-miles of travel

Example evaluation criteria

• Change in area/region PMT
• Change in facility/corridor PMT
• Portion of drive-alone mode (SOV) trips

Total PMT on a facility would need to add non-motorized vehicle person trips. The typical
method of calculation involves use of travel demand model VMT divided by mode share.

The amount of person travel a corridor or system serves, PMT is directly related to VMT as it is
VMT multiplied by a vehicle or transit occupancy factor. PMT should only be used for high-
level planning processes because of the high level of estimation required. PMT can also be
calculated for modes on a regional basis if the mode split is known like from a MPO travel
demand model. Bicycle and pedestrian counts could also be used to determine PMT if trip
lengths are known on a facility basis. PMT has the same limitations as VMT. Calculating PMT
may be difficult as occupancy factors may not be available or not enough bicycle/pedestrian
counts may be available. OSUM models assume a static value for auto occupancy by trip
purpose. In JEMnR and SWIM models, auto occupancy reacts to land use and transportation
policies and projects and can be reported. The analyst should coordinate with the modeler as to
the applicability of its use.

A commonly reported mode share performance measure is the portion of travel by drive alone
mode, or single occupant vehicle (SOV). This can be reported for a region or corridor.

𝑆𝑆𝑆𝑆𝑆𝑆 𝑉𝑉𝑉𝑉𝑉𝑉
𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 𝑜𝑜𝑜𝑜 𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷 𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 =
𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝑉𝑉𝑉𝑉𝑉𝑉
Portion of SOV trips can be used to evaluate alternatives that encourage non-drive alone trips,
such as park and ride lots.

Throughput
Throughput is the hourly volume of traffic that a facility serves or discharges.

Performance measure
• Throughput on segment or intersection

Analysis Procedure Manual Version 2 9-14 Last Updated 12/2019

Example evaluation criteria
• Change in facility/corridor throughput
• Segments and time periods where demand is metered due to upstream bottlenecks
• Intersections and time periods or number of cycle lengths where green time is starved due
to upstream bottlenecks

Vehicle throughput is also a calibration measure used in microsimulation. Refer to APM version
1 Chapter 8. Vehicle throughput as reported in SimTraffic is known as “Vehicles Exited”.

Throughput is sometimes confused with capacity. Even where peak hour demand equals or
exceeds capacity, vehicle throughput is often less than capacity for several reasons, including:

• Metered volume – volume at an approach to a segment or intersection may be metered or

constrained due to an upstream bottleneck or similar condition, so fewer vehicles can be
served than otherwise could be.
• Congested bottleneck – once traffic flow has broken down on a free-flow facility, the
queue discharge flow rate from the bottleneck is less than the capacity of the freeway.
• Other temporary conditions such as inclement weather, incidents, and work zones.

Degree of Utilization/Congestion

Degree of utilization is the percent of a facility’s capacity that is being used by the traffic
volume, typically for a peak hour. The most commonly used measure is the v/c ratio. As the
degree of utilization increases, mobility (freedom of movement) and speed decrease and density
increases. Eventually, as volumes increase beyond a certain level, vehicles become impeded
enough that traffic flow breaks down, and speeds drop to near zero and the facility is considered
congested.

Degree of utilization is sometimes reported for other modes such as pedestrians, bikes, and
transit. In Oregon, with a few exceptions, pedestrian and bicyclist degree of utilization is not
typically reported because most pedestrian and bicycle volumes do not typically approach the
physical capacity of the facility.

Duration of Congestion
The measures discussed in this section evaluate recurring congestion. See Travel Time
Reliability section for measures that evaluate non-recurring congestion. Duration of congestion
reflects the temporal extent of congestion. Hours of congestion has been used as an alternative
mobility performance measure per OHP 1F.3. It is the period of time, that a segment, facility or
area is congested. A facility or area may experience multiple recurring periods of congestion,
such as an AM period and a PM period. Refer to APM Chapter 8 for procedures. Duration of
congestion may be visualized with exhibits such as contour diagrams or heat maps, see example
in Exhibit 9-2.

Analysis Procedure Manual Version 2 9-15 Last Updated 12/2019

Performance measure
• d/c ratio above 1.0
• Speed below an agreed-upon threshold
• Excess/unserved demand
• Queue on uninterrupted flow facility
• Average Daily Traffic to Capacity Ratio (ADT/C)

Example evaluation criteria

• Number of hours facility exceeds capacity, v/c ratio > 1.0
• Number of hours the facility is rated at LOS F
• Number of hours that speed is below a designated threshold

Exhibit 9-2: Sample Heat Map

Source: Traffic Congestion and Reliability: Trends and Advanced Strategies for Congestion Mitigation, FHWA,
2005

For analysis purposes non-recurring congestion is typically assumed to start when the demand in
the analysis time period exceeds capacity. The congested period ends when there is no longer
excess or unserved demand in the analysis time period. Other threshold definitions of congestion
are sometimes used for other purposes such as for performance monitoring or investment
decisions, such as using a speed threshold and speeds obtained from a travel demand model.

Analysis Procedure Manual Version 2 9-16 Last Updated 12/2019

Travel speeds from a travel demand model are approximate and should only be used on a relative
basis to compare alternatives/scenarios.

Average Daily Traffic to Capacity Ratio (ADT/C)

The ADT to C ratio (ADT/C) is the average daily traffic divided by the peak hour capacity.
ADT/C has been used by ODOT as a rough indicator of the level of congestion. The ADT/C
methodology was developed as part of studies prepared for FHWA (2) and has been used by
ODOT as part of the statewide congestion management system. It is a way to estimate peak
spreading on a road system. It is a higher planning level rating of the level of congestion as
compared to duration of congestion or queueing. ADT/C can be used for segments or
intersection approach analysis for planning estimates of congestion. For segments ADT/C should
be reported out by direction. For intersections sum the sum the lane group capacities for each
approach. The highest approach ADT/C is reported for the intersection.

Performance measure
• ADT/C

Example evaluation criteria

• Model links with ADT/C ratio exceeding threshold
• Alternative change in ADT/C ratio

Model capacities can be coded as a segment capacity or as a lane capacity depending

on the model platform. The per lane capacity must be multiplied by the number of
lanes before calculating the ADT/C ratio.

Analysis Procedure Manual Version 2 9-17 Last Updated 12/2019

ADT/C thresholds are offered (from) as shown in Exhibit 9-3 below:

Exhibit 9-3 ADT/C Congestion Level Thresholds

Level Condition Description Lower Upper
ADT/C ADT/C

1 Uncongested No decrease in speeds during the peak 0.00 6.75

hour.

2 Uncongested to 6.75 8.25

Moderately

3 Moderately Speeds decrease slightly during 8.25 9.25

Congested portions of the peak hour.

4 Moderately to 9.25 9.75

Congested

5 Congested Speeds decrease significantly during 9.75 10.75

portions of the peak hour.

6 Congested to Very 10.75 12.25

7 Very Congested Speeds decrease substantially for 12.25 13.75

substantial portions of the peak hour.

8 Very to Extremely 13.75 15.25

9 Extremely Speeds decrease substantially for more 15.25 24.00

Congested than the peak hour.

Queue Length
Motor vehicle queue length is typically a peak period performance measure. Queues occur in both
under and over-saturated conditions. Undersaturated queues occur on interrupted flow facilities at
traffic control locations. Both segments and intersections will experience oversaturated queuing
when demand exceeds capacity. Oversaturated queue lengths measure the spatial extent of
congestion in length (typically feet). Normally free-flow segments can experience queues when
oversaturated.

Performance measures
• 95th percentile queue length

Analysis Procedure Manual Version 2 9-18 Last Updated 12/2019

Example evaluation criteria
• 95th percentile queue length by approach (refer to
APM v1 Chapter 7 and v2 Chapter 8)
• Queue blocking of turn or through lane
• Intersection queue blocking percentage exceeding 5% of peak hour
o Queue spillback to railroad crossing
o Queue spillback to functional area of intersection
o Queue on exit ramp extending to deceleration area or mainline as calculated using
design speed
o Queue occurring in area with insufficient sight distance

Undersaturated queueing at a signalized intersection approach tends to build and dissipate with
every green phase, with a maximum value reached during the peak period. Undersaturated
queueing at a stop controlled intersection approach tends to gradually build to a maximum value
during the peak period and then dissipate.

A queue blockage or spillback condition should be reported when the duration exceeds five percent
of the peak hour. See Chapters 12 and 13 of APM. Queue spillbacks need to be evaluated with
other contextual information to determine the extent and nature of the problem. Spillback queues
can reduce both safety and capacity. Spillback occurs when a queue at one intersection extends
into a second signalized intersection. This is typically reported as the total length of
oversaturated queue beginning from bottleneck where the queue started.

Queuing is usually reported as the number of vehicles or length of vehicles in queue at the 95th
percentile. 95th percentile queues are typically used to identify the extent of queuing problems
and to evaluate alternatives that reduce queue lengths.

95th percentile queues are calculated using deterministic tools following HCM methods, or by
microsimulation. Depending on the solutions being evaluated, microsimulation is typically needed
for final design in congested conditions. Methods of calculation vary by facility type and level of
analysis detail. Refer to APM Chapter 12 and 13 for deterministic queue calculation procedures.
Queuing is provided by microsimulation models where v/c ratios are high or conditions are
congested (refer to APM Chapter 15).

Demand to Capacity Ratio

When the estimated v/c ratio exceeds 1.0, it is referred to as a demand to capacity (d/c) ratio.
Travel demand models generate demand which can be used to calculate d/c ratios.

Typically a travel demand model run would be a constrained run. An unconstrained

(infinite capacity) run can be requested that will show the full desired demand on a
facility.

This means that for a given time period, there are more vehicles desiring to use a facility than it
can accommodate. This is also known as oversaturation. The actual volume will never exceed the
capacity of the facility. Instead, the excess demand (unserved trips) may do one or more of the

Analysis Procedure Manual Version 2 9-19 Last Updated 12/2019

following: divert to other routes; change the time of the trip; distribute to other destinations;
change the travel mode; or queue up to be served in following time periods (incurring additional
delay).

Performance measure
• d/c ratio

Example evaluation criteria

• Travel demand model links with directional link peak hour d/c ratio exceeding 1.0
• Number of locations on state highways with a d/c ratio of 0.90 or higher
• Number of urban area lane-miles over 1.0 d/c ratio

Travel demand model d/c ratios are link-based and can only be relatively compared on a large-
scale basis such as below, at, or over capacity. They cannot be compared with the Oregon
Highway Plan or Highway Design Manual volume-to-capacity ratios as these require that
volumes are based on the 30th highest hour from actual ground counts, while raw (not post-
processed) model volumes typically only represent an average weekday condition and have been
calibrated to the facility level. Also, model capacities are generically estimated based on
functional class and speed rather than using HCM methods. Model d/c ratios represent a full 60-
minute period rather than the peak 15-minute period. Model d/c ratios provide a planning level
indication of the extent of demand on segments, including the level of potential congestion,
without pinpointing specific intersection bottlenecks. For preliminary screening purposes model
d/c ratios may be reported as below, near, or over capacity rather than reporting specific values.

• Over capacity: d/c > 1.10

• At capacity: d/c between .90 and 1.10
• Near capacity: d/c between .80 and .89
• Below capacity: d/c < 0.80

The d/c ratio can be used to evaluate and rank or prioritize oversaturated links, and to evaluate
alternatives that reduce demand or increase capacity.

Travel Time

Travel time is a measure of the length of time a segment, facility or route can be traversed in a
given time period. It is most often reported for a given direction during the peak period and
expressed as the average travel time of all vehicles. Influences include design speed
(encompassing facility geometrics), free flow speed, control delay, traffic volume, and travel
distance.

Performance measures
• Average travel time during peak period
• Freight travel time
• Emergency services response time

Analysis Procedure Manual Version 2 9-20 Last Updated 12/2019

Example evaluation criteria
Average peak period travel time is the sum of travel times for all trips on the segment over a
given time period, divided by the traffic volume, typically expressed as minutes or seconds.
𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷
𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 =
𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆

Speed is based on segment running speed or field or archived speed data between representative
locations such as intermodal facilities, employment centers, CBDs, medical centers, park and
ride lots, or transit centers.

As an input to noise analysis, speed is calculated as the speed at LOS C, except is

reported as the posted speed if volumes are less than volumes at LOS C.

Field or archived speed data such as from private sector probe data may be used as a measure of
existing travel times as well as for reasonability checking of modeled speeds. Differences in
methodologies need to be taken into account when comparing existing speeds from different
sources and modeled speeds.

Total vehicle travel time is the average travel time per vehicle multiplied by the vehicle volume
over the analysis period. At a very basic level:
𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆 𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿ℎ
𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 = × 60
𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆

Where

Travel Time = Average travel time of all vehicles traversing segment (min)
Travel Speed = Average travel speed of vehicles (mi/hr)
Segment Length = Length of section (mi)

Travel time may be developed from travel demand models, based on zone to zone travel for an
origin-destination (O-D) matrix or by summing link travel times between major intersections.
MPO model travel times may be produced for a variety of modes such as SOV, HOV, freight,
and transit. Travel times for all modes are used as inputs into measures of accessibility. Model
based travel time travel time information can also be classified by trip type (i.e., work-based
trips). Travel times from a travel demand model are approximate and should only be used on a
relative basis to compare alternatives/scenarios.

For corridor or facility level analysis travel time may be developed from operational models such
as HERS and HCM methods.

Travel time can be used on a relative basis to evaluate emergency services by making
assumptions about faster speeds for an emergency vehicle to travel a given O-D path under

Analysis Procedure Manual Version 2 9-21 Last Updated 12/2019

emergency conditions. Contributing factors include the ability of the emergency vehicle to move
through congestion and traffic control devices and the provision of emergency pre-emption.

Average Delay
Performance measures
• Average delay per vehicle (sec/veh)

Example evaluation criteria

• Average delay for intersection
• Average delay for lane group

Delay is the additional vehicle travel time beyond the free-flow travel time for a given facility.
Free-flow travel time is defined differently depending on the tools used. For reliability it could
be based on empirically determined speeds, or posted speeds can be used in some HCM
deterministic procedures. The analysis period is typically the peak 15-minute period of the
design hour. In some instances free-flow conditions may be replaced by a designated acceptable
target travel time or speed. Delay is typically calculated using HCM procedures, which also
include Level of Service thresholds based on delay for many facility types.

Delay is also calculated by travel demand models and microsimulation methods. These delay
outputs must be post-processed in order to compare with HCM delay values.
1 ℎ𝑜𝑜𝑜𝑜𝑜𝑜
𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷 = [𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝑇𝑇𝑇𝑇𝑚𝑚𝑒𝑒 − 𝑇𝑇ℎ𝑟𝑟𝑟𝑟𝑟𝑟ℎ𝑜𝑜𝑜𝑜𝑜𝑜 𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇] ×
60 𝑚𝑚𝑚𝑚𝑚𝑚

Where

Delay = Delay for all vehicles the segment over the study period (hours)

Actual Travel time = (minutes)

For uninterrupted flow facilities, threshold travel time may be defined in different ways; free-
flow travel time, travel time at posted speed limit, or a policy definition of congested speed
(minutes). For interrupted flow, delay is computed by HCM methodologies and is the sum of
segment delay and control delay. Control delay is delay occurring due to traffic control devices
such as signalized intersections, roundabouts, or stop signs. Delay per vehicle does not account
for the total number of vehicles being delayed which can result in underestimation of the impact,
as compared to using vehicle hours of delay, which is generally a better performance measure.

Total Delay (Vehicle Hours of Delay)

Vehicle hours of delay is the delay per vehicle for a given segment, multiplied by the total
number of vehicles in the study period, typically daily or annual. The delay per vehicle is based
on the travel time minus the free-flow or threshold travel time. Unlike delay per vehicle, vehicle-
hours of delay evaluates the total number of vehicles that are delayed.

Analysis Procedure Manual Version 2 9-22 Last Updated 12/2019

Performance measures
• Daily vehicle hours of delay (veh-hr/day)
• Annual vehicle hours of delay (veh-hr/yr)

Example evaluation criteria

• Peak period vehicle hours of delay for segment
• Annual hours of delay per 1000 VMT

[𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡 𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡 − 𝑡𝑡ℎ𝑟𝑟𝑟𝑟𝑟𝑟ℎ𝑜𝑜𝑜𝑜𝑜𝑜 𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡 𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡] × 𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡 𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣

𝑉𝑉𝑉𝑉ℎ𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 𝐻𝐻𝐻𝐻𝐻𝐻𝐻𝐻𝐻𝐻 𝑜𝑜𝑜𝑜 𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷 =
60𝑚𝑚𝑚𝑚𝑚𝑚
ℎ𝑜𝑜𝑜𝑜𝑜𝑜

Where
Delay = vehicle-hours of delay per given time period for the study segment

The threshold travel time is typically based on a threshold or target speed which could be free-
flow speed, posted speed, or some other policy speed determined to be the minimum desirable
operating speed. Several methods are available for estimating vehicle hours of delay, including
the HCM, PPEAG, HERS-ST, and microsimulation. Facility VHD is obtained as the sum of the
VHD of the individual segments. VHD is useful for evaluating an entire study area across
multiple segments and is useful in sketch-level cost estimation.

In-vehicle person-hours of delay is VHD multiplied by the average vehicle occupancy.

Analysis Procedure Manual Version 2 9-23 Last Updated 12/2019

9.3 Travel Time Reliability

Travel time reliability considers (1) the range of potential travel times roadway users may
experience, (2) the consistency of travel times, and (3) the ability of a roadway to provide a
desired travel time. Traditional measures of roadway operations, such as volume-to-capacity
ratios or average travel speeds, reflect conditions during a design or analysis hour, such as the
30th-highest volume hour of the year. However, demand variation is just one of a number of
factors that affect roadway operations. The effects of severe weather, incidents (e.g., stalls,
debris), crashes, construction and maintenance activities, and special events (e.g., festivals,
college football games) can all contribute to roadway operations that are different (and generally
worse) than the average condition, as illustrated in Exhibit 9-4.

Exhibit 9-4 Difference Between Traditionally Reported and Actual Roadway Operations

Source: Derived from FHWA Travel Time Reliability website,

https://ops.fhwa.dot.gov/publications/tt_reliability/brochure/index.htm

Measures of travel time reliability incorporate all of the factors influencing roadway capacity and
free-flow speed to describe the variability of travel times along a particular roadway section or
facility, or for a given trip. This variability affects roadway users in important ways, including:

• Commuters, who must plan extra time into their commute trip to avoid arriving late at
work, even though they may not need that extra time most days;
• Freight shippers, who incur extra costs when shipments take longer to reach their
destination, as well as their customers, whose supply chains may be disrupted by late
deliveries; and
• Transit operators, who may need to add buses and drivers (at a significant added cost)
to ensure frequent, reliable bus schedules that attract and retain customers.

Evaluating travel time reliability can also help roadway agencies better evaluate the effects of
traffic operations strategies, such as ramp metering, dynamic part time shoulder use, and freeway
service patrols. As illustrated in Exhibit 9-5, these strategies may produce relatively small effects
on roadway capacity and average travel speed under normal conditions, but can have much
greater effects on travel time reliability. For example, ramp metering may delay the onset of
freeway breakdowns, or even reduce the number of days when freeway operations break down.

Analysis Procedure Manual Version 2 9-24 Last Updated 12/2019

Traffic management centers and service patrols can help roadway agencies more quickly
identify, respond to, and clear incidents, minimizing the effects of the incident both spatially and
temporally.

Exhibit 9-5 Difference Between Traditionally Reported and Actual Roadway Operations

Source: Derived from FHWA Travel Time Reliability website,

https://ops.fhwa.dot.gov/publications/tt_reliability/brochure/index.htm

The remainder of this section uses the term reliability as shorthand for travel time reliability.
This section introduces methods of evaluating reliability, describes potential applications for a
reliability analysis, and presents data sources and analysis tools currently available for evaluating
and forecasting reliability. Methods for evaluating reliability on freeways and other
uninterrupted-flow facilities are presented in APM Section 11.5.

9.3.1 Applications

Travel time reliability analysis has a number of potential applications, including:

• Performance reporting,
• Project planning, and
• Traffic management planning.

Performance Reporting

The federal MAP-21 and FAST Act transportation funding legislation requires states and
metropolitan planning organizations (MPOs) to measure roadway performance. The FHWA’s
final rule implementing this legislation defines four reliability-related system performance
measures as part of the set of National Performance Management Measures (Federal Register,
Vol. 82, No. 11, January 18, 2017, 23 CFR Part 490). These performance measures can be
evaluated using travel time data contained in the National Performance Management Research
Data Set (NPMRDS) maintained by FHWA and made available to states and DOTs, or can be
evaluated using an equivalent travel time dataset acceptable to FHWA.

Although performance management is outside the scope of this chapter, one potential analysis
application is to forecast the contribution of project alternatives toward meeting roadway system

Analysis Procedure Manual Version 2 9-25 Last Updated 12/2019

performance targets. Furthermore, because the most difficult part of conducting a travel time
reliability analysis is assembling a travel time dataset, the work that ODOT and MPOs invest in
calculating FHWA’s required performance measures can readily be extended to other
applications, such as those described next.

Project Planning

Typical planning applications include problem identification, project evaluation, and project
prioritization. The first of these requires (desirably) actual travel time data, while the latter two
require both a reliability analysis model and actual travel time data for use in calibrating the
model. Sources of travel time data are discussed in APM Section 9.3.5, while descriptions of
currently available analysis models are provided in APM Section 9.3.6.

Problem Identification
In a typical problem identification application, reliability performance measures are evaluated for
a defined roadway network (e.g., all freeways in a metropolitan area, all Interstate highways in
Oregon). Roadway sections where the performance measure exceeds a threshold value, or
alternatively, the worst X% of all roadway sections, are then flagged for further analysis to
identify the cause(s) of the unreliability and, subsequently, potential projects or operational
strategies to improve reliability. Any of the four primary travel time data sources available to
ODOT (described later in Section 9.3.5) can be used to assemble a travel time dataset. Once this
dataset has been created, the full range of reliability performance measures can be calculated
from it. In addition, as described later in Section 9.3.6, planning-level estimates of some
reliability performance measures can be developed without having a travel time dataset
available. These planning methods require estimates of a roadway section’s free-flow speed,
average travel speed, and volume-to-capacity ratio.

ODOT has not yet set any targets or thresholds for reliability performance measures;
doing so will require additional investigation and experience using these measures.
The FHWA’s National Performance Management Measures and the HCM’s
reliability rating (described below) are examples of measures with built-in threshold
values for unreliable travel.

Project Evaluation
Individual projects can be evaluated by comparing the values of one or more reliability
performance measures with and without the project, following this general process:

1. Evaluate reliability performance for existing conditions using actual travel time data.
2. Calibrate a reliability-capable analysis tool, such as those described in Section 9.3.7, to
replicate existing conditions.
3. Adjust the model parameters to reflect the project aspects that influence reliability.
4. Re-run the model to forecast future reliability performance with the project.

Analysis Procedure Manual Version 2 9-26 Last Updated 12/2019

The additional effort required for the application is the effort to code the roadway facility in the
analysis tool and to then calibrate the analysis output to reasonably match existing conditions.
Once the existing facility is coded, it is relatively quick to forecast the reliability performance of
a potential project in software. Exhibit 9-6 lists a variety of potential roadway capacity,
modernization, and operations projects, along the factors influencing reliability that these
projects affect.

Exhibit 9-6 Reliability Factors Influenced by Roadway Infrastructure Projects and

Operations Strategies
Project Type Reliability Factors Influenced

Add general-purpose (GP) lane Capacity, the timing and amount of demand
Modernization Free-flow speed, capacity
Ramp metering Capacity, demand
Traffic management center Incident detection and response times
Road patrols Incident response and clearance times
Speed harmonization Free-flow speed
Managed lanes Capacity, demand in GP and managed lanes; benefits
reflected in lowered person delay
Bus-on-shoulder No bus volume in GP lanes; benefits reflected in lowered
person delay
Part-time shoulder use Capacity, possibly incident clearance times
Traveler information Timing and location of demand
Traffic demand management Timing and amount of demand

The effects of many operational strategies have not yet been well-quantified; therefore, it may be
desirable to test a range of values for how a given strategy may affect reliability, to determine the
sensitivity of the result to the assumptions used. Appendix B of the IDAS User’s Manual
(Cambridge Systematics and ITT Industries 2000), no longer supported but available in the
Technical Reference Library section of HCM Volume 4 (http://hcmvolume4.org), provides
default values for the effects of a number of operational strategies, although the information is
somewhat dated at this point.

Project Prioritization
Once the effects of a given project or strategy have been forecasted, this information can be
incorporated into a prioritization process, for example by considering both the magnitude of the
reliability improvement and the number of vehicles or people that would benefit.

Traffic Management Planning

Reliability can also be incorporated into the development of various types of traffic management
plans, such as:
• Incident management planning—forecasting the relative benefits of different strategies
under consideration
• Work zone planning—identifying suitable work zone start and end times and number of
lanes closed

Analysis Procedure Manual Version 2 9-27 Last Updated 12/2019

 • Special event planning—evaluating different event schedules, evaluating special traffic
operations strategies for freeway off-ramps (e.g., temporary lane controls, signal timing
adjustments, traffic control officers)

Considerations for Performing a Reliability Analysis

Evaluating reliability is most useful when a roadway facility operates, or is forecast to operate,
over capacity on a regular basis, leading to highly variable travel times. In these cases, even if it
is not financially or physically feasible to provide extra capacity through road widening, the
effects of incremental improvements can still be evaluated in terms of reducing worst-case travel
times, providing more consistent travel times, and/or reducing overall person delay.

For future-year forecasting, the additional effort required to conduct a reliability analysis using
default values is minimal, once the facility has been coded and calibrated in an analysis tool that
implements the HCM freeway facilities method. In other words, if a project would require a
facility analysis using the core freeway facility methodology anyway, there is little reason not to
go ahead and generate a set of reliability performance measures at the same time.

When forecasting the effects of project alternatives on a roadway’s reliability, it is desirable to

incorporate local reliability-related input values to the extent that the alternatives affect those
inputs. For example, if an intersection improvement would be expected to affect the
intersection’s crash rate, using a local existing-conditions crash rate in lieu of a national default
value is desirable. Similarly, when comparing and prioritizing potential projects on different
roadways, it is desirable to account for differences in local traffic demand patterns. If the projects
are located in different parts of the state with different climates, then using local weather data
would also be desirable. Developing local input data for reliability methods is discussed in APM
Chapter 11, Appendix 11F.

9.3.2 The Travel Time Distribution

A travel time distribution is a collection of travel time observations or forecasts for a defined
roadway section (e.g., segment, facility) over a relatively long period of time (e.g., all nighttime
time periods over the course of a year; all weekday time periods between 6:00 and 10:00 a.m.).
Each observation represents the average travel time to traverse the road section during a defined
time period, typically 5 or 15 minutes.

Once a travel time distribution has been created, nearly any reliability performance measure can
be directly developed from it, except for certain measures where the measure’s travel time
reliability component is weighted by another variable (e.g., traffic volume, truck volume, person
trips, regional population). The travel time distribution can be developed through direct
observation of travel times (see Section 9.3.5) or by forecasting travel times using analysis tools
(see Section 9.3.6).

Exhibit 9-7 illustrates travel time distributions developed for northbound I-5 in the Portland area
between the Highway 217 and I-405 (south) interchanges, for all weekday a.m. peak (6:00 to
9:00 a.m.) time periods during February 2017. One distribution was developed using 5-minute

Analysis Procedure Manual Version 2 9-28 Last Updated 12/2019

data, while the other was developed using 15-minute data. The data were originally acquired for
NCHRP Project 07-22 (Planning and Preliminary Engineering Applications Guide to the
Highway Capacity Manual) from the PORTAL database maintained by Portland State
University. Both of these graphs are frequency distributions: the x-axis represents travel times in
15-second bins, while the y-axis represents the number of travel time observations associated
with each bin (i.e., the number of 5- or 15-minute time periods experiencing an average travel
time within the 15-second range represented by the bin).

Exhibit 9-7 Examples of Travel Time Distributions Developed from 5- and 15-Minute Data

(a) 5-minute Data (b) 15-minute Data

Both distributions show a peak on the left side, corresponding to free-flow or nearly free-flow
conditions. Both also show a secondary peak in the left-center area of the distribution,
corresponding to typical peak-period traffic congestion. Finally, both distributions have a long
tail to the right, corresponding to conditions during severe weather (e.g., freezing rain) and/or
when incidents occur (e.g., crashes, stalls, water on the roadway). The 15-minute distribution is
more compact than the 5-minute distribution, as the extremely low speeds reported during
individual 5-minute periods occur less often over a longer 15-minute period.

Fifteen-minute data are generally adequate for performing reliability analyses and have the
following advantages over 5-minute data:
• One-third the amount of data must be manipulated
• Reduced quality-control effort, due to fewer time periods with missing data or outlier
travel times
• Compatible with FHWA requirements for National Performance Management System
reporting
• Compatible with HCM analysis output

The greater detail provided by 5-minute data can be useful for diagnosing the causes of
unreliability along a roadway. Diagnosing reliability problems is beyond the scope of this
chapter, but is addressed in the HCM Planning Guide workshop material on performance
management, available on HCM Volume 4, www.hcmvolume4.org.

Analysis Procedure Manual Version 2 9-29 Last Updated 12/2019

9.3.3 Reliability Performance Measures

Identifying Key Travel Times from a Travel Time Distribution

The starting point for measuring reliability is identifying the travel times required to traverse a
roadway section under specified conditions. These travel times can represent a fixed value (e.g.,
the travel time required to traverse the section at the posted speed limit), a percentile value (e.g.,
the 95th percentile highest travel time, a difference between two other travel times (e.g., the
difference between the 50th and 95th percentile travel times), or statistical descriptors of the
distribution such as the standard deviation. Exhibit 9-8 depicts common types of travel time
values that can be obtained from the travel time distribution shown in Exhibit 9-7(b). These
travel time values are described in the subsections that follow.

Exhibit 9-8 Examples of Travel Time Values Obtained from a Travel Time Distribution

Free-flow Travel Time

Free-flow travel time is the time required to travel a roadway section under low-volume
conditions. It is preferably calculated as the average vehicle speed during low-volume periods
(i.e., 500 pc/h/lane or less), with good weather and no work activity or incidents. Alternatively,
when the study roadway is a freeway, multilane highway, or two-lane highway (i.e.,
uninterrupted flow without traffic signals), and the distribution clearly contains congestion-free
periods, free-flow travel time can also be estimated as the 5th-percentile travel time, as shown in

Analysis Procedure Manual Version 2 9-30 Last Updated 12/2019

Exhibit 9-8. Typically, free-flow travel time is not reported by itself, but is used instead to
calculate other reliability measures, such as the travel time index, discussed later. Highway
Capacity Manual (HCM) methods also calculate delay based on the difference between the
actual travel time and the free-flow travel time.

Travel Time at the Speed Limit

The time required to travel a roadway section at the speed limit can be used as an alternative
starting point for calculating delay, and as an input to reliability measures based on the
percentage of time the roadway operates at or above a target percentage of the posted speed. This
value can also be used as a check that the free-flow travel time estimate is accurate; the free-flow
travel time will normally be slightly less (i.e., faster) than the travel time at the speed limit.

Target (Policy) Travel Time

This is the time required to travel a roadway section at a designated speed (e.g., free-flow speed,
posted speed, speed producing maximum vehicle throughput, speed considered “congested”,
speed at which greenhouse gas or particulate emissions significantly increase). It is typically not
reported by itself, but is used in calculating other reliability measures, such as the policy travel
time index and the percent of time or percent of travelers experiencing conditions where the
target travel time is achieved.

ODOT uses the travel time at the posted speed limit rather than at the free-flow speed
as the basis for calculating reliability performance measures for ODOT purposes.
Different travel time targets may be required for FHWA reporting purposes.

Average (Mean) Travel Time

This is the average time to travel a roadway section during a given time period. HCM segment
and facility methods predict average 15-minute travel times for a particular set of conditions.

Percentile Travel Time

A percentile travel time is the travel time over a roadway section achievable a given percentage
of the time. Percentile travel times may be reported by themselves, but are also often used in
calculating other reliability measures. The most common percentile travel times are:
• 50th-percentile (median) travel time—this time typically will be slightly lower than the
mean travel time, due to the influence of exceptionally long (outlier) travel times on the
mean travel time;
• 80th-percentile travel time—the travel time achievable 80% of the time; research has
shown that the 80th-percentile time is more sensitive to roadway operational changes than
the 95th-percentile time, making it useful for evaluating project effects on reliability; and
• 95th-percentile (planning) time—for a segment or facility, the travel time achievable
95% of the time; for a trip, the travel time one would need to budget to ensure an on-time
arrival 95% of the time (e.g., late to work approximately once a month when
commuting).

Analysis Procedure Manual Version 2 9-31 Last Updated 12/2019

Percentile Truck Travel Time
This measure, not depicted in Exhibit 9-8, is similar to percentile travel time, but is calculated
from a distribution of truck travel times. These times may be different from overall vehicle travel
times due to lower truck speed limits, severe roadway geometry (e.g., steep grades), presence of
truck weigh stations, etc. FHWA’s Freight Reliability performance reporting measure
incorporates 50th- and 95th-percentile truck travel times.

Buffer Time
Buffer time is calculated as the 95th-percentile travel time minus the average travel time. It
represents the extra amount of time a traveler would need to budget for a trip to ensure an on-
time arrival 95% of the time.

Misery Time
Misery time is the average of the highest 5% of travel time observations in the distribution,
approximating a reasonable worst-case condition.

Standard Deviation of Travel Times

This is a statistical measure of how much travel times may vary from the average travel time.
The larger the standard deviation, the greater the variability of travel times from day to day along
the roadway.

Semi-Standard Deviation of Travel Times

This is a statistical measure of how much travel times may vary from the free-flow travel time. It
is calculated from the set of travel time observations slower than the free-flow speed as follows:

𝑛𝑛
1
𝑆𝑆𝑆𝑆𝑆𝑆 = � × �(𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹 − 𝑇𝑇𝑇𝑇𝑖𝑖 )2
𝑛𝑛
𝑖𝑖=1

where
SSD = semi-standard deviation,
n = number of travel time observations slower than the free-flow speed,
FFTT = free-flow travel time (s); and
TTi = travel time observation i (s).

Ratio-Based Reliability Performance Measures

Travel time values are frequently used to create other measures of reliability. For example, one
travel time value can be divided by another to create a ratio. When an observed travel time is
divided by the free-flow travel time, the resulting ratio is known as a travel time index (TTI).
The TTI indicates how much longer the observed travel time was, relative to the free-flow travel
time. Exhibit 9-9 provides examples of ratio-based performance measures derived from the
travel time distribution shown in Exhibit 9-7(b). The travel time distribution in Exhibit 9-9 is
depicted as a cumulative distribution, with the x-axis containing TTI values and the y-axis
showing the percentage of travel time observations occurring at or below a given TTI.

Analysis Procedure Manual Version 2 9-32 Last Updated 12/2019

Exhibit 9-9 Examples of Ratio-Based Reliability Measures
100.0% Misery index = 3.60

TTI95 = 3.38 (planning index)

90.0%

TTI80 = 2.34
80.0%

70.0%
Percent of 15-minute Periods

95th percentile

TTI equivalent to misery time

60.0%

TTI50 = 1.83
50.0%

80th percentile
40.0%

50th percentile
30.0%
Reliability rating (fwy) = 27%

TTI = 1.33

20.0%

10.0%
TTI =1.00 (free-flow)

0.0%
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00 2.20 2.40 2.60 2.80 3.00 3.20 3.40 3.60 3.80 4.00

Travel Time Index

Travel Time Index (TTI)

A TTI is calculated as a travel time divided by the free-flow travel time. A TTI value of 1.00
indicates travel at the free-flow speed, while a TTI value of 2.00 indicates travel that is twice as
long, compared to free-flow conditions. Commonly reported TTIs include the 50th-percentile TTI
(TTI50, the 50th-percentile travel time divided by the free-flow travel time), the 80th-percentile
TTI (TTI80), the 95th-percentile TTI (TTI95, also known as the planning index), and the mean (or
average) TTI (TTImean, not pictured in Exhibit 9-9).

Policy Travel Time Index (TTIP)

ODOT’s policy TTI is calculated as a travel time divided by the travel time at the posted speed
limit. A TTIP value of 1.00 indicates travel at the posted speed, while a TTIP value of 2.00
indicates travel that is twice as long as travel at the posted speed limit. Similar to the TTI, a
variety of percentile values can be reported, including TTIP50 (the 50th-percentile travel time
divided by the travel time at the posted speed limit), TTIP80, and TTIP95.

ODOT uses TTIP instead of TTI for ODOT reporting purposes. Analysts should be
aware that software packages may report TTI by default.

Analysis Procedure Manual Version 2 9-33 Last Updated 12/2019

Level of Travel Time Reliability (LOTTR)
The LOTTR (not shown in Exhibit 9-9) is defined as the 80th-percentile travel time divided by
the 50th-percentile travel time. The greater the LOTTR value, the longer travel times are on
relatively poor (but not uncommon) travel days, compared to travel times on typical days. The
FHWA incorporates LOTTR into its Interstate Travel Time Reliability measure, described
below. The FHWA considers an LOTTR value less than 1.50 as indicating “reliable” conditions
for reporting purposes.

Truck Travel Time Reliability (TTTR) Index

The TTTR Index (not shown in Exhibit 9-9) is defined as the 95th-percentile truck travel time
divided by the 50th-percentile truck travel time. The FHWA uses this measure as the basis for the
Freight Reliability component of its National Performance Management measures. For each
roadway section, a TTTR Index is calculated for each of the following five reliability reporting
periods:
1. All weekday a.m. peak periods (6 a.m. to 10 a.m.) during a calendar year
2. All weekday midday periods (10 a.m. to 4 p.m.) during a calendar year
3. All weekday p.m. peak periods (4 p.m. to 8 p.m.) during a calendar year
4. All weekend daytime periods (Saturday and Sunday, 6 a.m. to 8 p.m.) during a calendar
year
5. All nighttime periods (Sunday through Saturday, 8 p.m. to 6 a.m.) during a calendar year

FHWA’s Freight Reliability measure is calculated as the length-weighted average of the

maximum of the five TTTR Index values for each Interstate roadway segment in the state.

Misery Index
The misery index indicates how much longer a reasonable-worst-case travel time is, relative to
the free-flow travel time. It is computed as the misery time divided by the free-flow travel time.

Buffer Index
The buffer index (not shown in Exhibit 9-9) is the 95th-percentile travel time divided by the
average travel time. Although this measure appears in the reliability literature, the HCM 6th
Edition recommends against using it for tracking travel time trends “because it is linked to two
factors that can change: average and 95th percentile travel times. If one factor changes more in
relation to the other, counterintuitive results can appear.” This same issue applies to other ratio-
based measures incorporating the 50th-percentile or mean travel time, such as LOTTR and the
TTTR Index.

Percentage-Based Reliability Measures

Percentage-based measures can indicate the percentage of time that a roadway operates at or
better than a specified travel time or TTI. Percentage-based measures can also indicate the
percent of people experiencing a specified condition (e.g., the percentage of people that were
able to travel a roadway at 45 mph or faster).

Analysis Procedure Manual Version 2 9-34 Last Updated 12/2019

On-time Percentage
The on-time percentage is the percentage of time periods when travel can occur at or above a
specified target speed. Failure rate is a similar, but opposite, measure of the percentage of time
periods when the target speed is not achieved. The target speed could be the posted speed, a
percentage of the posted speed, the speed at capacity (i.e., the speed that maximizes throughput),
or any other speed that makes sense for a particular analysis need. The duration of congestion
(with “congestion” defined as travel slower than the target speed) is the study period length
multiplied by the failure rate.

Reliability Rating
The HCM defines the reliability rating as the percentage of time periods where the TTI is no
greater than a threshold value of 1.33 for freeways and 2.50 for urban streets. The threshold
value represents the point at which facility operations typically break down; thus, the reliability
rating approximates the percentage of time that a roadway operates below capacity.

Interstate Travel Time Reliability

The FHWA uses this measure, along with a companion National Highway System (NHS) Travel
Time Reliability measure, as the reliability components of its System Performance measures for
performance reporting. For each directional Interstate roadway section, a LOTTR is calculated
for each of the following four reliability reporting periods:
1. All weekday a.m. peak periods (6 a.m. to 10 a.m.) during a calendar year
2. All weekday midday periods (10 a.m. to 4 p.m.) during a calendar year
3. All weekday p.m. peak periods (4 p.m. to 8 p.m.) during a calendar year
4. All weekend daytime periods (Saturday and Sunday, 6 a.m. to 8 p.m.) during a calendar
year

Interstate Travel Time Reliability is then calculated as the length- and person trip–weighted
percentage of Interstate roadway sections that have LOTTRs less than 1.50 during all four
periods.

Delay-Based Reliability Measures

All measures of delay require defining a threshold where delay starts. Possible thresholds
include:
• Free-flow travel time—the HCM uses free-flow travel time as the starting point for
delay (i.e., any travel time slower than the free-flow travel time is considered to be
delayed). This approach allows an apples-to-apples comparison of delay between
roadways with different speed limits. However, this approach also allows delay to include
travel times faster than the travel time at the posted speed, but less than the free-flow
travel time, which may be inconsistent with both agency and traveler expectations.
• Travel time at the posted speed—any travel time slower than the travel time at the
posted speed is considered to be delayed. This approach is probably the most consistent
with traveler expectations for freeways and rural highways, but may not be particularly
helpful in identifying or prioritizing problem areas, as higher-volume roadways with
relatively high speeds will produce as much or more delay as lower-volume roadways

Analysis Procedure Manual Version 2 9-35 Last Updated 12/2019

 with lower speeds. ODOT uses the posted speed as the starting threshold for calculating
delay.
• Travel time at a target speed—a minimum speed is specified as the target for
satisfactory operations, and the portion of any travel time longer than the travel time at
the target speed is considered to be delayed. For national performance reporting purposes,
the FHWA defines a speed of 60% of the posted speed as the point where “excessive
delay” begins.

Vehicle Hours of Delay

Delay is defined as the larger of (1) the actual travel time during a given time period minus the
threshold travel time, or (2) zero. Vehicle hours of delay (VHD) is then:

𝑑𝑑𝑖𝑖 × 𝑉𝑉𝑖𝑖
𝑉𝑉𝑉𝑉𝑉𝑉 = �
3,600
𝑖𝑖

where
VHD = vehicle hours of delay (veh-h),
di = delay during time period i (s),
Vi = volume during time period i (veh), and
3,600 = number of seconds in one hour (s/h).

Person Hours of Delay

Person hours of delay (PHD) is calculated similarly to VHD, but accounts for the vehicle
occupancy of each mode using the roadway. Roadway operations strategies that reduce delay for
higher-capacity modes (e.g., carpools, vanpools, transit) will show a greater percentage
improvement in PHD than in VHD.

For roadways where all travel modes experience identical delays, PHD is calculated as:

𝑑𝑑𝑖𝑖 × 𝑉𝑉𝑖𝑖 × 𝑂𝑂𝐹𝐹

𝑃𝑃𝑃𝑃𝑃𝑃 = �
3,600
𝑖𝑖

where
PHD = person hours of delay (person-h),
OF = average vehicle occupancy (persons/veh), and
all other variables are as defined previously.

Analysis Procedure Manual Version 2 9-36 Last Updated 12/2019

For roadways where different travel modes experience different delays (e.g., facilities with
managed lanes or bus-on-shoulder operations), PHD is calculated as:

∑𝑚𝑚 𝑑𝑑𝑖𝑖,𝑚𝑚 × 𝑉𝑉𝑖𝑖,𝑚𝑚 × 𝑂𝑂𝑂𝑂𝑚𝑚

𝑃𝑃𝑃𝑃𝑃𝑃 = �
3,600
𝑖𝑖

where
di,m = delay of mode m during time period i,
Vi,m = volume of mode m during time period i,
OFm = average vehicle occupancy of mode m (persons/veh), and
all other variables are as defined previously.

The Oregon default for average vehicle occupancy of private vehicles is 1.4 persons per vehicle,
based on the 2009–2011 Oregon Household Activity Survey.

Recommended Reliability Performance Measures

The following performance measures provide a good starting point for evaluating reliability:
• 80th-percentile TTIP—this measure reports the upper limit of commonly occurring (e.g.,
once a week) travel conditions. This measure is more sensitive to roadway operations
strategies such as ramp metering and road patrols than is the 95th-percentile TTIP. This is
because the longest travel times in the travel time distribution tend to be associated with
major crashes and/or severe weather, both of which are less affected by operations
strategies.
• 95th-percentile TTIP—this measure reports uncommonly poor, but not worst-case,
conditions that roadway users would account for as part of their trip planning (e.g., a
once-a-month occurrence on a commute trip). The planning time associated with this
measure can be valued in terms of commuter time that could have been spent at home,
extra freight shipment time that must be planned for, and longer transit trips that must be
scheduled (possibly requiring additional vehicles and drivers). However, the use of an
index rather than a pure travel time allows facilities with different lengths and different
free-flow speeds to be compared on an apples-to-apples basis.
Additional reliability measures, such as TTIP50, person delay, and reliability rating, can also be
evaluated, depending on the specific needs of the analysis. For example, the FHWA national
performance management measures would be forecasted if the purpose of the analysis was to
investigate the potential contribution of different project alternatives toward meeting state or
metropolitan system performance targets.

9.3.4 Reliability Reporting Periods

Reliability quantifies the uncertainty in travel times that a traveler might experience from day to
day, across different times of day, over a period of time from a few months up to a year. Key
reliability time periods are defined below.

1. The reliability analysis period is the smallest time unit for which the analysis procedure
is applied. In the case of freeway and urban street facility analysis, the analysis period is

Analysis Procedure Manual Version 2 9-37 Last Updated 12/2019

 typically 15 min, although it can be of greater or lesser duration, at the discretion of the
analyst. Alternative tools may define different analysis period lengths.
2. The study period is the sum of the consecutive analysis periods for which the facility
analysis procedure is applied (e.g., an a.m., midday, or p.m. peak period). The study
period is defined by the analyst for each specific application. A study period of multiple
hours is preferred, as a single congested peak hour could be very reliable but with poor
travel times, while the shoulder hours could be much less reliable but with better travel
times.
3. The reliability reporting period is the period over which reliability is to be estimated
(e.g., the 250 non-holiday weekdays in a year). In essence, the reliability reporting period
specifies the days within the year for which the reliability analysis is to be performed.

Exhibit 9-10 depicts these different time periods. The y-axis in the figure represents the time
dimension on a given day, with each vertical cell representing one analysis period and the
combination of all the individual analysis periods representing the length of the study period.
The x-axis in the figure represents the facility’s spatial dimension, with each horizontal cell
representing one roadway section or HCM segment and the combination of all the individual
sections or segments forming the length of the study facility.

Exhibit 9-10 Reliability Analysis Time−Space Domain

Source: HCM 6th Edition, Exhibit 11-1

Analysis Procedure Manual Version 2 9-38 Last Updated 12/2019

 The combination of the study period length and the facility length should be large
enough to contain all analysis periods and segments where demand exceeds capacity
(for example, the red and yellow areas shown in Exhibit 9-10). If congestion spills
out of the study facility length and/or the study period length, the analysis results will
understate the facility’s operational performance and reliability.

The study period length should be carefully scoped when the analysis is not intended
to cover a full day. Including numerous time periods that rarely, if ever, experience
congestion will tend to lower reliability performance measure values, which in turn
may mask problems on the facility or fail to show much benefit from operational
treatments.

The z-axis in Exhibit 9-10 introduces the reliability dimension. The facility analysis is repeated
for each day of the year represented in the reliability reporting period, each of which
experiences, to a greater or lesser degree, a different set of conditions. The travel times required
to travel the facility during each analysis period of each day in the reliability reporting period are
then aggregated into a travel time distribution.

The days to include in the reliability reporting period will depend on the type of facility being
analyzed and the purpose of the analysis. For example, a study of the reliability of a major
commute route within a metropolitan area might define a reliability reporting period of all non-
holiday weekdays during the year. In contrast, a study of the reliability of a highway leading
from the Willamette Valley to the Oregon Coast might define a reliability reporting period
consisting of Saturdays, Sundays, and holidays during the summer.

9.3.5 Travel Time Data Sources

Travel time data are most commonly obtained from online databases of probe-vehicle speed data.
The data are generated from commercial vehicle fleets and users of cell phone–based navigation
systems, with the probe devices recording speeds that are reported to a central database. Through
post-processing, speeds are attributed to a reporting segment called a TMC (Traffic Message
Channel). On freeways, a TMC is typically defined from ramp gore to ramp gore. The recorded
speed data are then converted to travel times across the TMC and stored in online archival
databases.

Analysis Procedure Manual Version 2 9-39 Last Updated 12/2019

ODOT has access to four primary sources of travel time data:
• Iteris Performance Measurement System (iPeMS),
• HERE Traffic Analytics,
• National Performance Management Research Data Set (NPMRDS), and
• Portland, Oregon Regional Transportation Archive Listing (PORTAL).

ODOT is subscribed to the Iteris Performance Measurement System (iPeMS), which is a web-
based database for speed, travel time, and other data for Oregon roadways. The iPeMS system
collects, filters, processes, aggregates and visualizes speed and travel time data derived from the
probe data collected by HERE. Access to the Oregon iPeMS database is available at
https://odot.iteris-pems.com; TPAU approval is required for access by non-ODOT staff. ODOT
also has access to raw HERE travel time data, which can be downloaded for automobiles only,
trucks only, or both automobiles and trucks. APM Section 18.2.3 provides more information
about iPeMS and HERE data.

The NPMRDS dataset is provided by FHWA to MPOs and state DOTs without charge. The
NPMRDS dataset is accessed through the RITIS (Regional Integrated Transportation
Information System) online data portal (www.ritis.org) and can be downloaded for a region or
corridor of interest using the Massive Data Downloader. The download is in the form of a CSV
(comma separated value) file containing 5-minute sample summary data for the selected TMCs
over the time period of interest. The NPMRDS further contains separate records for passenger
cars and freight traffic. Additional information on the NPRMDS is available in APM Section
18.2.3 and through FHWA at
http://www.ops.fhwa.dot.gov/freight/freight_analysis/perform_meas/vpds/npmrdsfaqs.htm. The
NPMRDS dataset only covers the National Highway System.

PORTAL is the data archive for the Portland metropolitan region, which has been a collaborative
effort between ODOT and Portland State University’s Intelligent Transportation Systems (ITS)
Laboratory, with additional data supplied by WSDOT, PBOT, TriMet, and Clark County, among
others. PORTAL archives speed and count data from approximately 500 inductive loop detectors
in the Portland metropolitan region dating back to July 2004. PORTAL has a web-based
interface that provides performance metrics designed to assist practitioners and researchers.
More information on the PORTAL system can be found on the Portland State University website
at http://portal.its.pdx.edu and in APM Section 18.2.4.

In addition to these online databases, travel time reliability data can be obtained from other
devices that allow longitudinal measurements of speeds and travel times. For example, Bluetooth
or Wi-Fi readers can be used to monitor individual vehicle travel times over extended periods of
time. These raw travel time data can be aggregated to derive a travel time reliability distribution.

9.3.6 Methods for Forecasting Reliability

Differences Between Reporting Existing Reliability and Forecasting Future Reliability

Existing travel time reliability is normally determined from actual travel time data from one of
the sources described in the previous section. Reliability performance measures derived from

Analysis Procedure Manual Version 2 9-40 Last Updated 12/2019

these travel time distributions thus report the actual conditions that occurred during the period of
time covered by the data.

In the absence of travel time data, it is also possible to forecast various measures of travel time
reliability using one of the analytical methods described in this section. Analytical methods will
tend to predict somewhat worse reliability performance than would typically occur in any given
time period (e.g., 1 year). This is because analytical methods account for very rare events (e.g.,
unusually severe weather) that have very large travel time impacts. These events may not occur
in any given reporting year, and therefore are not necessarily used in planning decisions, but
nevertheless are the events that “travelers remember,” as was highlighted in Exhibit 9-4.

When reporting travel time reliability, the majority of the effort involves manipulating the travel
time data and (potentially) matching the data to information from other databases, such as traffic
volumes. Some travel time data sources provide an analysis tool that performs this data
manipulation and analysis, while other sources provide only the raw travel time data, which
analysts must manipulate themselves.

When performing a detailed forecast of travel time reliability, the majority of the effort involves
coding and calibrating the facility in the analysis tool. The analysis tool then takes care of
creating various reliability scenarios, generating the travel time database, and reporting reliability
performance.

Categories of Reliability Forecasting Methods

Reliability forecasting methods can be divided into three main groups: (1) sketch-planning
methods developed through the SHRP 2 program, (2) the detailed HCM freeway and urban
streets reliability methods, and (3) Oregon’s implementation of HERS-ST, which incorporates
elements of the other two methods.

Although in theory microsimulation can also be used to estimate reliability, it is not currently
practical to do so in a way that addresses the multitude of potential scenarios the way the HCM
or HERS-ST can, because of the time required to develop, code, run, and analyze the many
different reliability scenarios that would be required to accurately estimate reliability. For
example, the HCM method allows random variation in the location, severity, and time of day of
incidents; severity and start time of severe weather events; and so on. HCM-implementing
software can evaluate hundreds of scenarios for a facility covering up to 24 hours a day for an
entire year in a matter of seconds. In contrast, FHWA’s pilot tests of evaluating reliability using
simulation used only 8 or 9 scenarios (combinations of demand and incidents) in two cities to
represent relatively common peak-period conditions, and without consideration of weather
effects. Such an approach may be sufficient to demonstrate some benefit from traffic
management strategies, but not to forecast future reliability.

SHRP 2 Sketch-Planning Methods

The SHRP 2 program developed planning-level methods for estimating selected travel time
reliability measures. Unlike reporting methods and the detailed HCM method, these methods do
not assemble a travel time distribution. Instead, they use equations to estimate what a roadway’s
reliability performance would be, using a minimum number of inputs: free-flow speed, volume-

Analysis Procedure Manual Version 2 9-41 Last Updated 12/2019

to-capacity ratio, and number of lanes. These equations were developed from research into the
reliability performance of a variety of roadways in different parts of the U.S.

SHRP 2 Project C11 Method

This method estimates delay due to recurring and nonrecurring congestion using just two inputs:
volume-to-capacity ratio and facility type (freeway, arterial, collector, ramp, local road). Facility
type is used as a proxy for free-flow speed. Predictive equations are then used to estimate
common reliability performance measures. The method is capable of forecasting reliability
impacts and costs for individual projects, and can be applied to any roadway type.

Roadway segments are the basic unit of analysis. Segments can be of any length, but it is
recommended that they not be so long that their characteristics change dramatically along their
length. Reasonable segment lengths would be:
• Freeways: between interchanges;
• Signalized highways: between signals; and
• Rural highways (non-freeways): 2–5 miles.

The method first estimates the mean TTI. The mean TTI then becomes an input to other
predictive equations for estimating:
• Recurring delay (hours)
• Incident delay (hours)
• Total delay (hours)
• 95th-percentile TTI
• 80th-percentile TTI
• 50th-percentile TTI
• Percent of trips < 45 mph
• Percent of trips < 30 mph
• Cost of recurring delay
• Cost of unreliability
• Total congestion cost

The reported reliability values apply to a single weekday analysis hour (the hour used in calculating
the volume-to-capacity ratio supplied to the method) over the course of a year. The results from
multiple calculations can be combined and weighted to produce reliability values for longer
weekday study periods.

HCM Planning Guide Method

The Planning and Preliminary Engineering Applications Guide (PPEAG) to the HCM presents a
method for estimating freeway reliability. It is based on the SHRP 2 C11 method, but allows
specific roadway characteristics to be used to estimate the free-flow speed, and it simplifies the
calculation of incident-related delay. Because the HCM does not currently provide reliability
methods for multilane and two-lane highways, the PPEAG limits itself to forecasting freeway
reliability. However, because the underlying SHRP 2 C11 equations can be applied to any
facility type, the PPEAG method can also be applied to any facility type.

Analysis Procedure Manual Version 2 9-42 Last Updated 12/2019

Required inputs to the method are:
• Free-flow speed: Estimated using Appendix 11A of APM Chapter 11.
• Analysis-hour speed: Estimated using the screening method in APM Section 11.3.5 (for
freeways and multilane highways) or from the appropriate PPEAG method for other
roadway types.
• Number of directional lanes: 2 to 4 (if less than 2 lanes, use 2; if more than 4 lanes,
use 4)
• Volume-to-capacity ratio: Estimated using APM Chapter 11 screening-level methods.

The method predicts the same performance measures described above for the SHRP 2 C11
method. The reliability reporting period is also the same: one or more weekday analysis hours
over an entire year.

Oregon HERS-ST Method

The HERS-ST software does not directly calculate reliability performance measures. However,
ODOT has used HERS-ST to generate the inputs required for the SHRP 2 C-11 mean TTI
equation, namely: free-flow speed, recurring delay rate, and incident delay rate. Once the mean
TTI has been determined, all of the other performance measures described above for the SHRP 2
C11 method can also be predicted.

ODOT has also demonstrated the application of HERS-ST for developing reliability scenarios
combining a variety of severe weather, incident, and work zone events. Appropriate demand and
capacity, and free-flow speed adjustments for a given scenario are made in HERS-ST before re-
running the model. The individual scenario results are then weighted by their probability of
occurrence when calculating an overall performance measure result. Because HERS-ST results
apply to individual roadway sections, they may not fully reflect the delay associated with queue
spillback from one section into other upstream sections.

The HERS-ST method can be applied to any roadway type, for a reliability reporting period
consisting of the weekday peak hour over an entire year.

HCM Freeway Reliability Method

The HCM freeway reliability analysis methods are described in Chapters 11 and 25 of the HCM
6th Edition. A reliability analysis starts by coding a base scenario for the facility, consisting of all
the data normally entered for an HCM operations analysis using the HCM’s core freeway facility
methodology (described in APM Chapter 11). The HCM reliability method then creates a series
of scenarios representing various combinations of demand, severe weather, incidents, work
zones, and special events, along with a probability of occurrence for each scenario. Each
reliability scenario adjusts the base scenario’s demand, capacity, and/or free-flow speed in some
way, resulting in a different set of performance results (e.g., travel times) for each scenario.
Finally, a travel-time distribution is generated based on the weighted probability of each scenario
occurring.

Analysis Procedure Manual Version 2 9-43 Last Updated 12/2019

Exhibit 9-11 illustrates the method’s flow from scenario generation through outputting a
distribution of travel times.

Exhibit 9-11. Flowchart of the HCM Freeway Reliability Analysis Method

The HCM provides national default values for incident probabilities and durations by incident
severity, and demand variations by day of week and month of year. It also provides probabilities
of 10 categories severe weather by month for the 101 largest metropolitan areas around the U.S.
(Portland is the only Oregon metropolitan area represented in the HCM’s default weather data)
The analyst can choose to replace any or all of the default values with local values, and can also
optionally provide data regarding long-term work zones and special events that significantly alter
traffic demand and/or traffic operations strategies.

The analyst must supply the following: the day of year represented by the base scenario’s traffic
volume (so that each scenario’s demand adjustment can be applied relative to that day), the study
period length coded in the base scenario (e.g., 6–10 a.m.), and the days to include in the
reliability reporting period.

The HCM does not provide much guidance on time periods to include in a reliability analysis,
other than to state that reliability reporting periods spanning one year are most common and that
the study period length should be long enough to allow queues to dissipate by the end of the
study period. The choice of days to include in the reliability reporting period will depend in part

Analysis Procedure Manual Version 2 9-44 Last Updated 12/2019

on the use of the facility—a commuter route might analyze weekdays, while a recreational route
might analyze weekends. For national reporting purposes, the FHWA defines four study periods
for the Interstate Travel Time Reliability measure (weekday a.m. peak, weekday midday,
weekday p.m. peak, weekend daytime) and five for the Truck Travel Time Reliability Index (the
four listed above, plus nighttime periods).

Method Comparison

The reliability forecasting methods discussed above vary in the following respects:
• Input data requirements
• Ability to be adapted to local conditions
• Number of scenarios used to model travel time variability
• Facility types covered
• Types of events modeled that influence reliability

All of the methods have tools available to assist in applying the method. Exhibit 9-12 compares
the capabilities of the different methods.

Analysis Procedure Manual Version 2 9-45 Last Updated 12/2019

Exhibit 9-12. Comparison of Travel Time Reliability Analysis Methods
Oregon
SHRP 2 C11 PPEAG HERS-ST Simulation HCM

Scenarios used 1 1 1/100s* ≤10 100s to 1,000s

Scenario
generation NA NA NA/Manual* Manual Automated
process
Freeways
Facility types Freeways,
All (extendable All All
covered urban streets
to all)
FFS, v/c, All required All required for
FFS, v/c, Obtained from
Required inputs # lanes, by simulation freeway facility
# lanes HPMS
average speed tool analysis
Values used to Inputs,
Local adjustment Scenario Inputs, scenario
No generate input scenario
capability generation generation
data generation
Reliability Most common/
Most common Most common Any Any
measures output any*
Creates sub-
Creates travel distributions
No No No/Yes* Yes
time distribution for each
scenario
Typically, 1+
Single analysis 1–24 analysis
Weekday peak analysis
Reliability hour for all hours for all Any,
hour for one hours for all
reporting period weekdays in weekdays in up to one year
year weekdays in
one year** one year
one year
Models weather
No No No/Yes* No Yes
impacts
Models incident If included as
Indirectly Indirectly Indirectly/Yes* Yes
impacts scenarios
Models work If included as
No No No/Yes* Yes
zone impacts scenarios
Notes: NA = not applicable, FFS = free-flow speed, v/c = volume-to-capacity ratio.
*In a batch-processing application using multiple scenarios.
**Calculations can be repeated for additional weekday analysis hours if desired.

The number of scenarios used by a method affects (1) the variety of conditions analyzed that can
impact roadway operations and (2) the ability to incorporate local conditions into the analysis.

Sketch-planning methods produce a single estimate of reliability, based on regression equations

developed from nationally representative travel time datasets. These methods do not account for
the effects of local weather conditions, differences in incident frequencies or detection and
clearance times, or other local factors. The base SHRP 2 C11 method is also insensitive to
differences in roadway characteristics that would affect the roadway’s FFS or capacity and thus
the reliability result. In contrast, the PPEAG method can account for these differences.

Analysis Procedure Manual Version 2 9-46 Last Updated 12/2019

As the number of reliability scenarios increases, the greater the analyst’s ability to account for
the various factors affecting reliability, but also the greater the effort required—either up front or
for each analysis—to develop the scenarios. Even the method using the greatest number of
scenarios, the HCM, places constraints on the types of situations considered in order to reduce
analysis complexity. (For example, the HCM limits consideration of weather to weather events
that decrease capacity by at least 4%.) Similarly, FHWA’s pilot tests of simulation used just 8 or
9 scenarios, because of the effort required to develop individual simulation models for each
scenario, along with determining the probability of each scenario. Although multiple simulation
runs can be performed for each scenario and the results compiled into a travel time distribution,
what the analyst ends up with is a single mean travel time for each scenario, each with its own
distribution around the scenario mean. This collection of sub-distributions does not match the
full spectrum of travel time observations that would be measured in the field.

HERS-ST offers the option of producing a single estimate of travel time reliability, using the
SHRP 2 C11 equations, or using its batch-processing feature to generate a true travel time
distribution from a series of reliability scenarios. For example, an analysis of a section of US 97
between Sunriver and LaPine incorporated 8 demand levels and 89 capacity-reducing events
(combinations of severe weather, incidents, and/or work zones) were included, for a total of 712
reliability scenarios. Capacity reductions for each event were derived from the default values
given in the HCM 6th Edition. The probabilities of each demand level and capacity-reducing
event occurring were also determined and used to weight the scenario’s resulting travel time.

The scenario-generation approach taken by the HCM is different than that used by simulation or
HERS-ST. Rather than rely on the analyst to define scenarios and decide which ones to include
or exclude, the analyst provides information on demand variability by day of week and month of
year, the probabilities of various types of severe weather by month, and probabilities of various
types of incidents. This information can come from the HCM’s national defaults, from a one-
time effort to create local or regional defaults, or from location-specific data. The analyst also
specifies the number of replications of each day–month demand combination; the HCM suggests
4 for a reliability reporting period spanning one year, corresponding to each day being modeled
approximately four times in a given month. If a shorter reliability reporting period is used, the
HCM recommends increasing the number of replications so the total number of scenarios
(replications × months × days) generated is at least 240. The HCM method then randomly
assigns weather and incident events (or non-events) to each scenario, along with random start
times for each event and (for incidents) random locations. This process recognizes, for example,
that heavy rain that occurs in the middle of the night will have a different impact on roadway
operations than a downpour in the middle of rush hour. The process also allows rarer events to be
considered as part of the overall analysis, without needing to arbitrarily decide which events to
include or exclude—it may not snow in Portland every winter, but ODOT prepares for the
possibility of snow anyway because of its severe impacts on roadway operations.

9.3.7 Tools for Forecasting Reliability

This section introduces software tools available to predict travel time reliability for freeways and
uninterrupted flow facilities. Three of the tools implement the HCM 6th Edition method, while
the other three implement versions of the SHRP 2 C11 planning-level equations.

Analysis Procedure Manual Version 2 9-47 Last Updated 12/2019

FREEVAL

FREEVAL is the official computational engine of the HCM 6th Edition freeway facilities and
freeway reliability chapters. It can be downloaded for free on the HCM Volume 4 website
(http://www.HCMVolume4.org). A FREEVAL reliability analysis builds on a calibrated and
completed freeway facilities analysis (described in APM Chapter 11), and then adds the
reliability dimension. FREEVAL applies user input or national defaults for incident distributions
and day-of-week and month-of-year demand variability, along with historical weather data and
user-specified work zone inputs. FREEVAL further integrates the HCM 6th Edition method on
Active Travel and Demand Management (ATDM) with methods for evaluating impacts of traffic
system management and operations strategies such as ramp metering, part-time shoulder use, and
managed lanes.

Highway Capacity Software (HCS)

HCS is commercial software for the Windows operating system that is developed, distributed,
and supported by the McTrans Center at the University of Florida. Similar to the process used by
FREEVAL, HCS builds from a calibrated freeway facilities analysis by adding the HCM’s
reliability method and (optionally) the HCM’s ATDM method. Users can apply the national
default values for demand variability, weather patterns, and incidents, or supply their own local
values. Users supply facility-specific work zone information.

TTR/ATDM

TTR/ATDM is a reliability analysis tool based on the HCM 6th Edition developed by
SwashWare and the University of Florida Research Foundation. The tool is an extension of the
HCM Calc tool for freeway facility analysis (described in APM Section 11.2.5). TTR/ATDM
implements the HCM reliability and ATDM methodologies, similar to what was described for
FREEVAL above. The tool can be downloaded for free through the Microsoft store.

PPEAG

The PPEAG’s freeway computational engine, available on HCM Volume 4, can be used to
estimate a freeway segment or facility’s volume-to-capacity ratio and average speed, given a
user-provided free-flow speed and number of directional lanes. These results can then be used
with the PPEAG reliability equations (either manually or by setting up a simple spreadsheet) to
estimate any of the reliability performance measures supported by the SHRP 2 C11 method.

HERS-ST

HERS-ST can estimate any roadway section’s free-flow speed directly, while HERS-ST output
can be used to develop the section’s recurring delay rate and incident delay rate. These results
can then be used to estimate (either manually or by setting up a simple spreadsheet) any of the
reliability performance measures supported by the SHRP 2 C11 method. ODOT has
demonstrated the ability to model weather and work zone scenarios with HERS-ST to develop
estimates of travel time reliability for non-freeway roadways and corridors containing a mix of
facility types. See APM Section 7.3 for more information about HERS-ST.

Analysis Procedure Manual Version 2 9-48 Last Updated 12/2019

SHRP 2 C11 Spreadsheet Tool

The C11 reliability spreadsheet tool is an Excel spreadsheet that calculates all of the reliability
measures supported by the C11 method. The spreadsheet also calculates the value of reliability
improvements based on the following assumptions, which can be changed within the
spreadsheet:
1. For passenger travel, it assumes a $19.86/hour average value of time multiplied by a 0.8
reliability ratio (i.e., hours of delay are multiplied by 0.8 when calculating the value of
changes in reliability),
2. For commercial travel, it assumes a $36.05/hour average value of time multiplied by a 1.1
reliability ratio.

For ODOT projects, values of travel time should be consistent with the most recent version of
“The Value of Travel Time: Estimates of the Hourly Value of Time for Vehicles in Oregon,”
available at https://www.oregon.gov/ODOT/Data/Pages/Economic-Reports.aspx. APM Section
10.6.8 provides more information about economic analysis.

Analysis Procedure Manual Version 2 9-49 Last Updated 12/2019

Tool Comparison

Exhibit 9-13 summarizes key features of software tools that implement the HCM’s reliability
method. Exhibit 9-14 provides similar information for tools that are based on the planning-level
SHRP 2 C11 reliability equations.

Exhibit 9-13 HCM-Implementing Tool Comparison

Overview HCS FREEVAL TTR/ATDM

Tool Overview
University of
McTrans hcmvolume4.org
Source Florida
Cost License Fee Free Free
Operating system Windows Windows/Mac Windows 10
Installation required Yes No (need Java) Yes
Widespread use High Medium Low
Staff and Support Needs
Learning curve Medium Medium Medium
Complexity Medium Medium Medium
Training available  ◐ 
User guide   
Instructional videos   
Technical support  ◐ ◐
User Experience
Copy/paste  ◐ ◐
Load/save   
Import/export   
Auto-fill ◐  
Specialized Features
Charts/visualizations (reliability)   
Charts/visualizations (ATDM)   
Automated report generation   ◐
Built-in scenario comparison   
Calibration (adjustment factors)   
Built-in weather adjustments   ◐
Incident scenario analysis  ◐ ◐
Work zone scenario analysis  ◐ ◐
ATDM method   
Ramp metering  ◐ 
Notes:  = fully supported, ◐ = partially supported,  = not supported.

Analysis Procedure Manual Version 2 9-50 Last Updated 12/2019

Exhibit 9-14 SHRP 2 C11 Implementing Tool Comparison
SHRP 2 C11 ODOT
Overview Reliability Tool PPEAG Tool HERS-ST

Tool Overview
Source tpics.us/tools hcmvolume4.org ODOT
Cost Free Free Free
Operating system Windows/Mac Windows/Mac Windows
Installation required No (need Excel) No (need Excel) Yes
Widespread use Low Low Low
Defaults or Imported from
Data source for reliability inputs Calculated
another tool HPMS
Manual or Manual or
Reliability calculations Automated separate separate
spreadsheet spreadsheet
Staff and Support Needs
Learning curve Low Low Medium
Complexity Low Medium Medium
Training available  
User guide  ◐
Instructional videos   
Technical support  
Specialized Features
Congestion cost estimates   
Notes:  = fully supported, ◐ = partially supported,  = not supported.

9.4 Level of Service (LOS)

Level of Service (LOS) and quality of service (QOS) are indicators that cannot be measured
directly in the field and are a letter grade based on an underlying performance measure value.

9.4.1 Motorized Vehicle Level of Service

Motorized vehicle Level of Service is a commonly used performance measure computed

following Highway Capacity Manual methodologies. It is a rating of the level of mobility
(typically as a function of delay or density) of a facility, segment, intersection or approach on a
scale of A-F. LOS A, B, and C indicate conditions where traffic moves without significant delays
over periods of peak hour travel demand. LOS D and E are progressively worse operating
conditions. LOS F represents conditions where average vehicle delay has become excessive and
demand has exceeded capacity. This condition is typically evident in long queues and delays.

Performance measure
• Level of Service letter grade A-F

Analysis Procedure Manual Version 2 9-51 Last Updated 12/2019

Example evaluation criteria
• Unsignalized intersection approach LOS compared to local jurisdiction LOS standard
• Freeway segment LOS

Motorized vehicle LOS is determined for the following facility types using the following
quantitative measures (all specified in the HCM):

• Freeway segments, facilities, merge, diverge and weaving segments

o Density – specifically, average number of vehicles per lane mile (pc/mi/ln). LOS
F where demand exceeds the capacity of the segment.
• Two-lane highway segments
o Density – specifically, follower density (veh/mi/ln) of directional segment (refer
to APM v2 Addendum 11B).
• Intersections – signalized, unsignalized, by approach, lane group or intersection as a
whole
o Delay – specifically, average delay (sec/veh) (by approach, lane group or
intersection as a whole).
• Urban Streets – segment or facility
o Speed (mi/hr)

Many local jurisdictions have adopted LOS as a performance measure for facilities under their
jurisdiction and have adopted LOS thresholds as standards. The analyst needs to evaluate LOS
and compare to the adopted local standards when analyzing those facilities. Some jurisdictions
have dual performance thresholds for both v/c ratio and LOS in general or by facility type.
Reporting LOS for state highways is optional, although reporting LOS on state highways is best
practice to obtain a complete picture of operations versus reporting v/c ratio alone. Facilities with
low v/c ratio could still have high delays. Refer to the HCM 6th Edition for specific calculations
and LOS thresholds for each facility type.

• Basic freeway segments – Chapter 12

• Two lane highways – Chapter 15
• Signalized intersections – Chapter 19
• Unsignalized intersections – Chapters 20-22

9.4.2 Multimodal Level of Service (MMLOS)

MMLOS is a Quality of Service (QOS) measure. QOS measures the perceived level of comfort
by the user, which could be a pedestrian, a bicyclist, or a transit rider. While vehicular LOS
includes factors for the effects of pedestrians on vehicular mobility, pedestrian/bicycle/transit
LOS reflects the point of view of the pedestrian, bicyclist or transit rider. The methodology
creates a score which is equated to a Level of Service rating. Refer to APM Chapter 14 for
procedures.

Analysis Procedure Manual Version 2 9-52 Last Updated 12/2019

A qualitative multimodal methodology is also available as an alternative to the full HCM
MMLOS method. It uses the same data categories as the HCM method, but is a qualitative
assessment which can be used where HCM methods do not apply or where data are not available.

Pedestrian and Bicycle Level of Service

The APM methodologies to calculate Pedestrian and Bicycle LOS are simplified
versions of the HCM Pedestrian and Bicycle LOS. Refer to APM Chapter 14 for
detailed procedures.

The APM Pedestrian and Bicycle LOS are based on user perception scores of the level of
comfort a user would experience on a given facility. Performance ratings for pedestrians are
provided for roadways with and without sidewalks and multi-use paths. PLOS evaluates
sidewalk width, posted speed, number of through traffic lanes and vehicle traffic volume.
Additional performance measure methods are under development for midblock crossings,
signalized intersections and unsignalized intersections.

Performance measure
• Level of Service letter grade A-F
• Qualitative MMLOS (good/fair/poor)

Example evaluation criteria

• Pedestrian facility LOS
• Bicycle facility LOS
• Multi-use facility LOS
• Signalized intersection ped or bike LOS (TBD)
• Unsignalized intersection LOS (TBD)

Performance ratings for bicyclists are provided for roadways with and without bike facilities,
separated paths, and intersections. Bike facilities evaluated include shared lanes, bike lanes,
buffered bike lanes, protected bikeways, and bike signals. Bicycle LOS evaluates the number of
through travel lanes, presence of bike lane or paved shoulder, posted speed and number of
unsignalized intersections and driveways.

Pedestrian and bicycle LOS can be used to evaluate walk and bike networks such as for a TSP to
identify needs, as well as to evaluate alternatives affecting sidewalk width, bike facility type,
volumes, lanes, posted speeds, and driveways.

Capacity-based HCM pedestrian performance measures evaluate the utilization of available

space. These measures are generally not used in Oregon due to the lack of pedestrian density.

Analysis Procedure Manual Version 2 9-53 Last Updated 12/2019

Transit Level of Service

The APM methodology to calculate Transit LOS is a streamlined version of the HCM
Transit LOS. Refer to APM Chapter 14 for detailed procedures.

The APM Transit LOS is based on user perception scores of transit service on a segment. Transit
LOS relates to passengers’ perception of walking to a transit stop on the street, waiting for the
transit vehicle, and riding on the transit vehicle. The method applies to buses, street cars, and
other types of transit vehicles operating with mixed traffic on the roadway. The measure does not
apply to transit operating in separated right-of-way. Transit LOS is a function of transit schedule
speed, transit frequency and pedestrian LOS. Transit LOS can be used to evaluate alternatives
that affect route speed, frequency, and pedestrian LOS.

Performance measures
• Segment Transit LOS letter grade A-F

Example evaluation criteria

• Transit LOS letter grades by segments along a transit route

Transit LOS is not an indicator of ridership, which may involve several contributing factors such
as land use density, transit frequency, reliability, wait time, walk time, transfers, fares, bus stop
amenities, and parking availability and cost.

9.4.3 Truck Level of Service Index

Truck Level of Service is a recently developed measure of the quality of service provided by a
facility for truck hauling of freight, as perceived by shippers and carriers. Truck LOS was
developed as part of NCFRP Report 31 (3). It is a composite index based on the percentage of
ideal truck operating conditions achieved by a facility. Ideal conditions are defined as a facility
usable by trucks with legal size and weight loads, with no at-grade railroad crossings, that
provide reliable truck travel at truck free-flow speeds, at low costs (i.e., no tolls). Truck Level of
Service (TLOS) Index is the ratio of the actual utility to the utility for ideal conditions (free-flow
speed and no tolls). Methodology details are found in the HCM Planning & Preliminary
Engineering Applications Guide (PPEAG) to the HCM 6th Edition.

Performance measure
• Truck LOS on highway facility

Example evaluation criteria

• Relative difference in facility Truck LOS letter grade

Analysis Procedure Manual Version 2 9-54 Last Updated 12/2019

 Truck Level of Service Index

1
%𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 =
(1 + 0.10𝑒𝑒 −200𝑈𝑈(𝑥𝑥) )

Where
%TLOS = truck LOS index as a percentage of ideal conditions (decimal),
U(x) = truck utility function, and
e = exponential function.
Truck Utility Function

U(x) = A × (POTA−1)+ B × (TTTI −1)+C × (Toll mi)+ D× (TFI −1)

Where
U(x) = utility of facility for truck shipments,
A = weighting parameter for reliability, sensitive to shipping distance = 5 / ASL,
for Oregon = 0.025
B = weighting parameter for shipment time, sensitive to free-flow speed = -0.32 /
FFS, FFS = free-flow speed,
C = weighting parameter for shipment cost = -0.01,
D = weighting parameter for the facility’s truck friendliness = 0.03,
POTA = probability of on-time arrival = 1 if the mixed flow (autos and trucks)
travel time index is ≤1.33 (freeways and highways) or ≤3.33 (urban streets),
TTTI = truck travel time index for the study period, the ratio of truck free-flow
speed to actual truck speed,
Toll/mi = truck toll rate (dollars per mile), a truck volume–weighted average for
all truck types, and
TFI = truck friendliness index, where 1.00 = no constraints or obstacles to legal
truck load and vehicle usage of facility and 0.00 = no trucks can use the
facility.

The truck utility function is based on several parameters including the probability of on-time
arrival for the truck shipment, the travel time index for trucks, tolls paid by trucks, and the truck
friendliness index.

The Truck LOS index is focused on the heavier long-haul trucks that travel intercity (externals in
travel demand models), rather than commercial vehicles that are typically lighter (can be pickups
and panel vans) and are used to distribute goods within an urban area. The response of these two
groups can be significantly different. For example, long haul trucks have more potential to be
shifted to rail. Commercial vehicles are often fleets of vehicles whose owner can be influenced

Analysis Procedure Manual Version 2 9-55 Last Updated 12/2019

by local policies, e.g., availability of CNG fueling stations or EV charging infrastructure, or
simply levels of local congestion.

Level of Service thresholds are based on %TLOS and class of freight facility. Three classes of
freight facility are defined. Truck LOS can be used to evaluate facilities of a uniform freight
class, and alternatives that affect travel time reliability, weight or dimensional restrictions, or
tolls.

For oversize/overweight vehicles, the ORS 366.215 approval process must be

followed if considering alternatives that reduce roadway widths on certain freight
routes. For more information refer to ORS 366.215 Implementation Guidance
(4)

9.5 Accessibility

Accessibility is a point, zonal, district or area-wide measure of the availability of a range of

opportunities such as employment, schools, shopping, medical, recreation, etc, by mode.
Accessibility measures typically include travel distance, travel time, population and employment
data. Other factors such as level of congestion, parking availability/cost, tolls, and safety may be
included directly or indirectly. A destination may be physically close by but if obstructed by a
freeway or river or other restrictions, it may not be very accessible. No single accessibility
measure captures all possible factors. Accessibility is typically an area or point measure and
requires a network. However, the Place Types land use methodology (see Chapter 7) produces a
rough accessibility measure of jobs within five miles on a straight line basis. As such, it does not
require a network, and is a rough measure of centrality of the location.

Accessibility is important for the bike, walk and transit modes as a travel option and equity
measure. The automobile mode generally has good accessibility in most areas. Bike, walk and
transit often do not have good accessibility due to incomplete networks or services. Maximizing
travel options is more likely to focus on those modes.

9.5.1 Accessibility for Motorized Vehicles, Pedestrians and Bicyclists

JEMnR (MPO-level) travel demand models include accessibility utilities. Accessibility can be
analyzed by mode, trip purpose or time of day. OSUM (non-MPO) models can produce
accessibility information by trip purpose only. Zone to zone demand and travel time is available
from trip matrices, which can also be mapped. Accessibility from travel demand models
identifies the potential level of interaction between zones. Applications include evaluating a
zone’s potential for development. Thematic or heat maps may be produced identifying the most
likely locations for development or transit for example. Accessibility can also be used for equity
analysis, based on household income, race, limited English proficiency and other socio-economic
factors.

Analysis Procedure Manual Version 2 9-56 Last Updated 12/2019

Performance measure
• Area that can be accessed within a given travel time, by mode of travel

Example evaluation criteria

• Percent change in regional accessibility (to jobs or shopping) in each TAZ stratified by
walk, transit and auto (JEMnR models)
• Number of essential destinations within a buffer travel time such as 20-minutes by any
mode. Essential destinations are defined as hospitals, grocery stores, parks, schools,
major retail, etc.
• Number of essential destinations or daily needs accessible within a certain travel time by
all modes to low-income, people of color, or limited English proficiency households
• Number of households within a market area such as a travel time radius of a commercial
or shopping zone, by mode. Percent of jobs or households within walking or bicycling
distance or travel time.
• Number of jobs within certain travel times for all modes accessible to low income, people
of color, and limited English proficiency households
• Percent of households/population located with ¼ mile of a bikeway or transit stop.
• Percent of population with X minutes between work and home
• Percent of population located within a mixed use or transit oriented development
• An index of the ratio of direct travel distances to actual travel distances. Well connected
streets result in a high index. Less connected streets with large blocks result in a lower
index.

The area or distance that can be accessed within a given travel time can be shown using
isochrone lines or by shading of TAZs to reflect numbers or percent changes in accessibility
variables. Isochrones are contour lines which show the spatial extent of the area or network that
can be accessed from a given location given different travel time thresholds, such as 5 minutes,
10 minutes or 20 minutes. Travel demand models and GIS tools are typically used to produce
isochrones lines for motorized vehicle travel, whereas GIS tools are typically used to measure
bicycle and pedestrian accessibility. It is critical to measure actual network travel time and
distance, not “as the crow flies”, especially for pedestrian and bicycle accessibility. Pedestrian
network analysis which codes each side of a street separately and which includes improved
crossings is superior to street centerline-based analysis, since lack of improved street crossings
can be a major barrier to safe pedestrian travel. See Exhibit 9-15 below for an example of an
isochrone map.

Analysis Procedure Manual Version 2 9-57 Last Updated 12/2019

Exhibit 9-15 Isochrones

Source: Wikipedia

Accessibility may be used to evaluate land use scenarios/changes, such as those that increase
density and diversity, for example Transit Oriented Developments (TODs). Travel demand
models can evaluate these changes as well as RSPM for use in scenario planning (refer to
Chapter 7).

When evaluating pedestrian or bicycle accessibility, any unofficial routes such as

trails or unofficial crossing locations should be noted as they indicate actual usage
and shortcuts which have potential to be improved as official walkways, bikeways or
crossings. For example, Strava’s heat map can be used to find non-traditional
bicycle pathways.

Accessibility for pedestrians and bikes may be used to compare alternatives or scenarios that
affect densification of land use or walk or bike connections, such as completing paths, adding
new paths, or improving crossings. Accessibility does not typically address facility adequacy.
Accessibility is one of many contributing factors that affect the amount of walking or biking
trips. Other contributing factors include level of comfort, completeness, and safety on the
facilities being traveled. Pedestrian and bike travel typically works best for travel distances under
one mile for pedestrians or five miles for bicyclists. Refer to Chapter 14 for multimodal
considerations.

Bike, Walk and Transit Scores

Bike score, walk score, and transit score are types of accessibility ratings, by mode of travel, of
locations based on the number and variety of nearby activities/amenities and the travel time to
access them. For example, for a given point (origin), what is the area that can be accessed within
a walking distance of ¼ mile, and what amenities are available within that area. A highly
accessible location for a given mode of travel would be one with a variety of amenities available

Analysis Procedure Manual Version 2 9-58 Last Updated 12/2019

within a reasonable travel time for that mode. A less accessible area would be one that has
limited or no amenities located within a reasonable travel time. These types of scores typically
do not take into account the comfort, safety or quality of the travel facilities or services being
used. The scoring method can be expressed in terms of accessibility to important destinations
such as schools, shopping areas, parks, medical facilities and transit facilities.

Such scoring methods may be identified using GIS, or using commercial products such as Walk
Score, an application used in evaluating the accessibility of candidate residential properties.
Commercial tools may create a combined measure or index that accounts for factors other than
just travel time. For example, a more complete bicycle score might include consideration of
topography. Other non-travel time based measures or indices may be included as well, such as
crime statistics, or hilliness of an area. Scoring methods may be aggregated into a rating for an
entire city, region or neighborhood, or can be localized to individual properties by address. Heat
maps may be created to visualize variations in accessibility throughout an area.

Accessibility scores may be useful in a sketch planning level analysis, but may be limited to
existing conditions only. Network-based accessibility measures can show improvements in
accessibility when certain links or crossings are added to the network. Commercial software is
generally used, as GIS analysis effort can be high, although there may be a cost associated with
obtaining the data. Commercial software may not share the complete algorithm or data sources
the score is based upon. ODOT has no proposed way to calculate these scores.

9.5.2 Accessibility for Transit Riders

Accessibility for transit riders measures the proximity of transit service available. It is used to
evaluate areas with or considering transit service. It may be used to evaluate or prioritize
alternatives that affect land use proximity via transit, such as land use densification, adding new
routes, or increasing frequency or span of service.

Performance measures
• Proximity of households/population to transit stops
• Proximity of households/population to destinations via transit

Example evaluation criteria

• Percent of population living within "X" miles or "Y" minutes that can access fixed route
transit.
• Percent of jobs or households within ¼ mile walking distance of transit stops.
• Percent of households in environmental justice (EJ) communities within half mile of high
capacity transit or quarter mile of frequent bus service. As used in equity analysis, such
as for environmental projects or TSPs, for identification of affected populations such as
minorities, income level, age, etc.

Analysis Procedure Manual Version 2 9-59 Last Updated 12/2019

Some general rules of thumb for transit corridors to be potentially viable are those with the
following characteristics

• Walk distance to/from transit stops less than or equal to ¼ mile, or ½ mile to high
capacity transit stations
• Residential density greater than 4-5 dwelling units/acre for local bus service (1 bus per
hour)
• No more than one transfer required

MPO travel demand models include transit lines, fares, and transit stops, and assign transit trips
to routes. These models can calculate accessibility to potential transit stops. Small urban area
models do not model transit. Metro’s model is more sophisticated with the ability to estimate
transit loadings at stops. Activity-based models promise a more dis-aggregate treatment of
transit, which is likely to be significantly more detailed and accurate, leading to more flexibility
in terms of transit performance measures.

Other transit accessibility measures include the use of isochrones to visualize how far a transit
rider can get from a given starting point within 15, 30, 45, and 60 minutes of travel using only
transit and walking. These illustrate the extent of activities that can be reached from points on
transit at different times. For information on other transit tools see the ODOT Public Transit
tools webpage.

Transit accessibility is just one of many contributing factors that may affect potential transit
ridership. Other factors include land use density, transit coverage, span, frequency, total travel
time, pedestrian level of stress/comfort, transit stop amenities, safety, transit fare, transfers
required, accommodation of bicycles, and bus occupancy.

Accessibility to frequent transit service may address equity by measuring the ease of access to
transit by specific groups such as lower income households. It may be part of the environmental
analysis process or may also be performed in some planning studies. GIS databases are able to
provide distance information. Factors include travel distance, level of comfort, safety, traveler
demographics, and frequency of service.

9.6 Safety

Safety performance measures evaluate historical or are predictors of future potential of crashes
on networks and facilities, including crash type and severity. Crashes or crash rates can be
displayed using GIS or other mapping tools to identify hot spots for network screening. Detailed
procedures are provided in APM Chapter 4.

9.6.1 Crash Rate

Crash rates are a commonly used safety performance measure for a wide range of planning and
project analysis studies as part of identifying safety improvement needs. Crash rates are easy to
calculate and require little data. Crashes should be based on the official data published by

Analysis Procedure Manual Version 2 9-60 Last Updated 12/2019

ODOT’s Crash Analysis and Reporting (CAR) Unit. AADTs are required for segments and Total
Entering Volume (TEV) AADTs are required for intersections. See APM Chapters 3 and 4 for
detailed information.

Performance measure
• Intersection crashes per million entering vehicles (MEV)
• Segment crashes per million vehicle-miles traveled (MVMT)
• Fatal and Severe Injury crashes
• Fatal and Severe Injury crash rates per 100 million vehicle miles traveled

Example evaluation criteria

• Segment crash rate exceeding average for similar facility type from Oregon State
Highway Crash Rate Tables
• Intersection crash rate exceeding published 90th percentile intersection crash rate for
similar intersection type (APM Exhibit 4-1)
• Crash rate exceeding site critical crash rate based on reference population of similar sites
• Pedestrian and/or bicycle involved fatal and severe injury crashes

The critical crash rate is a Highway Safety Manual screening method of the likelihood that a site
crash rate is high as compared to a reference population of similar site types. Critical crash rate is
used to flag and prioritize high crash rate locations for further study. See APM Section 4.3.4.

Requirements/Limitations
• Segment crash rates can be heavily influenced by the length of the segment.
• Lack of crashes inhibits usefulness of the measure for evaluating pedestrian and bicycle
safety
• Does not account for regression to the mean (RTM) (See Chapter 4 for definition)
• Critical rate requires sufficient number of sites in reference population

Crash Severity – indicator of need and priority based on the level of injury of crashes, the highest
priority being the reduction of fatal and severe injury crashes.
• Does not account for regression to the mean
• Requires AADT volumes
• Does not address future safety performance or alternatives

9.6.2 Safety Priority Index System (SPIS)

SPIS is a screening method developed by ODOT that computes a safety index based on crash
and volume history on segments. The SPIS index is a function of crash frequency, rate and
severity. The Traffic-Roadway Section (TRS) calculates SPIS numbers annually for the entire
public road system in Oregon. SPIS sites exceeding threshold scores based on the top 5% or 10%
percentile are identified and flagged for further safety investigation. SPIS site maps are available
including on TransGIS. The annual SPIS index is calculated based on the last 3 years of reported
crash history. Refer to APM Chapter 4 for more detailed information.

Analysis Procedure Manual Version 2 9-61 Last Updated 12/2019

Example evaluation criteria
• Top five and ten percent SPIS locations

9.6.3 Change in Crash Frequency Using Crash Modification Factors (CMFs) or Crash
Reduction Factors (CRFs)

CMFs and CRFs are typically used to evaluate candidate countermeasures for safety solutions.
The initial source for countermeasures should be the ODOT-approved set of proven
countermeasures and associated CRFs that are used for the All Roads Transportation Safety
(ARTS) Program. See Chapter 4 for detailed procedures.

Example performance measures

• Reduction or percent change in average annual crash frequency, type and/or severity by
application of a countermeasure, as calculated using CRFs or CMFs

9.6.4 Excess Proportions of Specific Crash Types

Excess proportion of specific crash types is an HSM screening measure of the extent that a crash
type (for example, fatal and serious injury, or pedestrian or bicycle crashes) at a site is
overrepresented. Crash sites can be intersections or segments. This is based on a comparison to a
reference population of similar sites. Excess proportion of crash types is an indicator of the
likelihood that a site will benefit from a countermeasure targeted at the collision type under
consideration.

Example evaluation criteria

• Target crash type or severity exceeding threshold

Excess proportion is most frequently used in large area studies such as TSPs. Refer to APM Section
4.3.5 for procedures.

The method does not account for regression to the mean. It does not require traffic volumes. It
does not address future safety performance or alternatives. It requires a sufficient number of sites
of a similar type in the reference population.

9.6.5 Expected or Predicted Crash Frequency

Expected or predicted crash frequency is an HSM predictive measure of long term crash
frequency. This is based on Safety Performance Functions (SPFs) which factor in geometrics,
traffic control, volumes, and operations. The Empirical Bayes adjustment methodology factors in
crash history. These methods account for RTM error, the natural fluctuation of crashes that
occurs over the long-term independent of the contributing factors the analysis is trying to review.
Predictive crash analysis is used most often for detailed analysis of alternatives. Expected or
predicted fatal and serious injury crash frequency should always be reported as a sub-category of
total crashes. Crash frequency can also be reported out by crash type such as bicycle or
pedestrian crashes if sufficient data exist. The method predicts reported crashes. There are no

Analysis Procedure Manual Version 2 9-62 Last Updated 12/2019

established thresholds but the measure may be used for ranking/prioritizing and for comparing
alternatives. Refer to APM Section 4.4 Predictive Methods for details.

Example performance measures

• Excess Expected Crash Frequency using Empirical Bayes (EB) Adjustments
• Net Change in Expected or Predicted Crashes

Excess Expected Crash Frequency using Empirical Bayes (EB) Adjustments is used to evaluate
the extent that a site’s long term average crash frequency differs from that of similar sites.

Excess Expected Average Crash Frequency = Expected Crashes – Predicted Crashes

The EB method requires a calibrated prediction model (with overdispersion factor) and
substantial similarity between the analysis period for which crash data exist and the analysis
period being used for the predictive method.

Net Change in Expected or Predicted Crashes is used to compare alternatives. Expected crashes
can be determined for alternatives if the only changes are in AADT. Otherwise, net change in
predicted crashes is used.

Net Change in Expected/Predicted Average Crash Frequency = Expected/Predicted

No Build Crash Frequency - Expected/Predicted No Build Crash Frequency

9.6.6 Conflicts

Conflicts are a measure of the number and type of locations where paths cross, merge or diverge
at an intersection or junction. Bicycle, pedestrian and transit vehicle conflicts can also be
reported as multimodal safety performance measures. Conflict points are potential crash
locations, although the number of conflict points does not indicate the probability of occurrence
of a crash, which would depend on additional factors such as traffic volumes. Paths that cross are
considered major conflicts while those that merge or diverge are considered minor conflicts.
Refer to APM Section 4.8.3 for procedures.

Example performance measures

• Number of conflict points at an intersection, total or by type of conflict

Conflicts are typically reported when analyzing alternative intersection types, alignments or lane
configurations.

Analysis Procedure Manual Version 2 9-63 Last Updated 12/2019

9.6.7 Access Spacing

Access spacing is a measure of the distance between driveways and public street intersections
along a roadway segment, or between interchanges along a freeway or expressway. ODOT
access spacing standards are provided in Appendix C of the OHP. Local jurisdictions may have
their own access spacing standards. A related measure is driveway density which is a factor in
bicycle Level of Service. Refer to OAR 734-051 and APM Chapter 4 for procedures.
Substandard access spacing can lead to safety and operational problems. Access density is a
factor in bicycle LOS.

Example performance measures

• Number of accesses not in compliance with spacing standards
• Percent of roadway in compliance with spacing standards
• Percent deviation from spacing standard
• Number of deviations required

Access spacing is commonly evaluated in corridor plans or refinement plans such as IAMPs or
AMPs, in approach permitting, and in projects considering new or modified accesses or roadway
connections.

Functional Area

The functional area of an intersection is a measure of the adequacy of spacing between

intersections and/or access points to accommodate vehicle paths. Functional area inputs include
speed, perception-reaction time, deceleration, lane changing, and queueing/storage lengths. Refer
to APM Chapter 4 for detailed procedures.

Performance measures
• Access or junction within functional area of an intersection

Example evaluation criteria

• Functional area of a new connection to the roadway
• Extent of overlapping functional areas

Analysis Procedure Manual Version 2 9-64 Last Updated 12/2019

9.7 Other Multimodal Performance Measures

9.7.1 Mode Share

Mode share, typically an area measure, is a function of many contributing factors. Factors
include trip purpose, travel time, level of stress/comfort, Level of Service, directness of route,
route completeness/connectivity, safety, accessibility, land use, travel costs, and household
characteristics. Typical automobile cost factors include auto ownership, maintenance, fuel,
parking, and tolls, and is highly influenced by the vehicle’s fuel efficiency (e.g., electric vehicles
can cost a fraction to fuel relative to internal combustion vehicles, with hybrids somewhere in-
between, depending upon the share of miles driven with electricity). Typical transit cost factors
include bus fares and subsidies. Bike mode share is also affected by topography, and increasingly
bike-share programs (e.g., Portland and Rogue Valley) and their cost schedule.

Performance measure
• Mode share

Example performance measures

• SOV mode share
• Change in mode share
• Percent share of total trips by mode – pedestrian, bicycle, transit, auto
• Percent share of total VMT by drive alone mode (SOV)

Mode share would typically be evaluated for transportation system plan performance or
scenarios or alternatives that may significantly change mode share. Examples include transit
route changes, transit subsidies, or parking availability/cost.

Mode share is typically obtained from a travel demand model as an estimate that may not
represent observed data and is not calibrated. In Oregon there are two levels of travel demand
models. In small urban models mode share is assumed from the household survey used to build it
(observed travel behavior). It is static and does not react to land use and transportation policies /
projects. MPO models have a mode choice model that does react to policies and projects and is
an important measure to be aware of and should be requested for all MPO-level model runs.

9.7.2 Transit Service Miles per Capita

Transit Service Miles per Capita is a measure of transit service coverage. It is calculated as fixed
route transit revenue service miles divided by area population. Data sources include the National
Transit Database (NTD), local transit agency plans and the General Transit Feed Specification
(GTFS). RSPM uses this measure and it is also an Oregon Statewide Transportation Strategy
(STS) monitoring measure. It can be used as a screening or supplemental for RTPs and TSPs.

Performance measure
• Transit service miles per capita

Analysis Procedure Manual Version 2 9-65 Last Updated 12/2019

For the base year, transit service is provided in units of bus-equivalent fixed route transit revenue
miles (not counting miles for transit vehicles when not in service). The number of miles traveled
while in service for each fixed route transit vehicle is summed over all transit vehicles, for a
period of one year. The measure is reported in units of annual service miles per capita.

This measure allows comparisons between alternatives that involve changes in transit service in
terms of routes or frequencies, including either expansions or reductions in service. The measure
does not reflect ridership.

Transit revenue miles can typically be obtained directly or calculated from miles on various
routes combined with hours of operation and headways from the local Transit Agency. This
should only include fixed route service. The National Transit Database (NTD) also reports
annual service miles by transit provider. Future transit service inputs are provided in units of
growth of the region’s bus-equivalent revenue miles per capita. It is also important to note that
revenue miles are reported in bus-equivalent units.

9.7.3 Multimodal Mixed-Use Area (MMA)

A multimodal Mixed-Use Area (MMA) is an Oregon land use designation that may be adopted
by a local government pursuant to the Transportation Planning Rule (TPR – OAR 660-012-0060-
10)) to promote mixed-use, pedestrian-friendly, transit oriented, compact land use and
transportation activity centers. In order to encourage these types of centers, an MMA designation
allows plan or land use regulation amendments to be approved without applying performance
standards related to motorized vehicle congestion levels, including volume to capacity ratio,
delay or travel time.

Performance measures for evaluating proposed MMA designations within interchange areas are
primarily safety related
• TPR requirements
o Crash rates compared to statewide average for similar facilities
o Top ten percent SPIS locations
o 95th percentile queue lengths on freeway exit ramps
• Suggested supplementary measures
o Critical crash rate
o Excess proportion of specific crash types
o Excess expected average crash frequency

For more information including definitions and maps of mixed use areas refer to the DLCD
Place Types webpage.

Analysis Procedure Manual Version 2 9-66 Last Updated 12/2019

9.8 Infrastructure

Infrastructure performance measures evaluate the supply of transportation networks or services.

9.8.1 Network Connectivity and System Completeness

Network Connectivity and System Completeness are measures of network completeness,

redundancy, and availability of alternative routes, which could include streets, intersections,
sidewalks and bicycle facilities. Connectivity and completeness is typically evaluated in system
planning as well as when considering the potential for re-routing of trips such as for TSMO
purposes. Inventories of network elements are prepared and displayed on maps, identifying gaps.
Data sources include the FACS-STIP tool for state highways, Google maps and Google Earth,
city or regional GIS databases, and Active Transportation Needs Inventories (ATNI) in Regions
1, 4 and 5.

Performance measure
• Network connectivity – extent that the network is inter-connected
• System completeness – percent of planned facility elements such as sidewalks, bike
lanes, or improved pedestrian crossings that currently exist

Example evaluation criteria

• Percent local I-I versus regional external –internal or internal-external (E-I/I-E) versus
external-external through trips (E-E) on highway – using select links relative to study
area
• Percent completeness of bike and walk facilities within ¼ mile of transit stops or ½ mile
of schools
• Percent of planned network with sidewalks and/or bicycle facilities
• Percent of network restricted to heavy vehicles
• Capacity available on parallel local facilities
• Ratio of shortest network path distance (driving, walking, or biking) to shortest straight-
line distance (as shown in Exhibit 9-16). This is a theoretical minimum distance. Ratios
closer to 1 are preferred.
• Number of roadway links divided by the number of roadway nodes or intersections (as
shown in Exhibit 9-17). 2

2
A higher index indicates that travelers have increased route choice, allowing more direct connections for access
between any two locations. Links are the segments between intersections, nodes the intersections themselves. Cul-
de-sac heads count the same as any other link end point. A higher index means that travelers have increased route
choice, allowing more direct connections for access between any two locations. According to this index, a simple
box is scored a 1.0. A four-square grid scores a 1.33 while a nine-square scores a 1.5. Dead-end and cul-de-sac
streets reduce the index value. This sort of connectivity is particularly important for nonmotorized vehicle
accessibility. A score of 1.4 is an example threshold for a ‘walkable’ community.

Analysis Procedure Manual Version 2 9-67 Last Updated 12/2019

Exhibit 9-16 Shortest Network Path versus Straight-Line Distance

Exhibit 9-17 Links/Nodes Ratio

Analysis Procedure Manual Version 2 9-68 Last Updated 12/2019

Out-of-direction Travel

Example performance measures

• Distance out-of-direction (mi or feet), by mode
• Additional VMT

This is the amount of additional travel time and/or distance for a trip or movement due to out-of-
direction travel, as compared to a base case. In other words, this is a measure of circuitousness of
a route as compared to a direct path. An example would be the out-of-direction travel for an
indirect J-turn or at-grade jug handle alternative as compared to a direct left turn. Excess out-of-
direction travel for motorized vehicles adds to travel time and VMT and may result in driver
frustration which could lead to violations or safety problems. Excess out-of-direction travel for
pedestrians (greater than approximately 0.10 mile) may deter use or lead to improper roadway
crossings. Excess out-of-direction travel for bicyclists (greater than approximately 0.33 mile) is
likely to deter use.

Intersection Density

Intersection density or multi-modal street density are not a common performance measure but
are occasionally used as a potential indicator of urban form, i.e., network redundancy,
connectivity, or pedestrian friendly paths in an area. Intersection density would be high value for
a grid system and low for an area with cul-de-sacs or public street access control is used in
JEMnR travel demand models used in many regions of the state. The similar street density is
used in Place Types, utilizing block group level data.

Example performance measures

• Number of intersections per square mile within a region or area
• Density of pedestrian-oriented/local streets and/or multi-modal streets miles per square
mile within a region or area.

9.8.2 Bicycle or Pedestrian Level of Traffic Stress (LTS)

Bicycle or pedestrian LTS is an ODOT APM Chapter 14 methodology that rates the level of
comfort of bicyclists or pedestrians traveling along or crossing a roadway. Scores range from 1
to 4, with 1 being the most comfortable and 4 being the least comfortable. Factors for pedestrian
LTS include sidewalk width, condition, and ADA ramps. Target scores are generally either 1 or
2, depending on nearby land uses and demographics such as schools, transit stops, downtown
cores, medical facilities, etc. It is useful to display LTS on maps to identify connectivity islands
and high stress locations such as major road crossings. Such locations create discontinuities
which if fixed could improve the LTS of an entire route. Refer to APM Chapter 14 for
procedures.

LTS is not by itself an indicator of the potential use of a walk or bike facility, which would need
to take into account additional factors such as the proximity and size of land use origins and
destinations, topography, and competition with other modes.

Analysis Procedure Manual Version 2 9-69 Last Updated 12/2019

Example performance measures
• Pedestrian or bicycle LTS score on a roadway segment, intersection, approach, or
crossing
• Project study area network locations not meeting LTS threshold

Analysis Procedure Manual Version 2 9-70 Last Updated 12/2019

 Appendix 9A – Applicability of Analysis Performance Measures by Plan or
Project Type

Appendix 9B – Alternative Mobility Targets (Planning Business Line Team

Operational Notice)

References

(1) Dowling, Richard. Traffic Analysis Toolbox Volume VI: Definition, Interpretation, and Calculation of
Traffic Analysis Tools Measures of Effectiveness. No. FHWA-HOP-08-054. 2007.

(2) Estimating the Impacts of Urban Transportation Alternatives, Participant’s Notebook, FHWA/NHI
December, 1995.

(3) Dowling, R., G. List, B. Yang, E. Witzke, and A. Flannery. NCFRP Report 31: Incorporating Truck
Analysis into the Highway Capacity Manual. Transportation Research Board of the National Academies,
Washington, D.C., 2014.

(4) Guidance for Implementation of ORS 366.215 (No Reduction of Vehicle-Carrying Capacity), ODOT
Transportation Development Division, April 17, 2015

Analysis Procedure Manual Version 2 9-71 Last Updated 12/2019