HighwayCapacityManual2000.doc

I. INTRODUCTION
Highway Capacity Manual 2000

In this chapter, a brief overview of the analytical procedures in this manual, their
organization into chapters, and guidance on their general application are provided.
The analytical procedures in this manual can be used for a number of applications
covering a broad range of facility types. The facility types are distributed among five
categories: urban streets, pedestrian and bicycle facilities, highways, freeways, and
transit.
In Chapters 10 through 14, for each of the five categories, general concepts are
presented, required inputs for each methodology are identified, reasonable
approximations for specific parameters are suggested for use if local data are not
available, and example service volume tables are provided. The Part II chapters also
contain special procedures used to supplement the planning applications defined in the
Part III chapters.

II. OVERVIEW OF ANALYTICAL PROCEDURES
For the analytical methods defined in Part III, the calculations of average speed,
density, and delay will provide insight into the level of service for what is considered a
steady-state condition. This means that the outputs provided by the computational
methods are considered representative for the length or area of the analysis and for the
duration of the analysis period. Thus, the Highway Capacity Manual (HCM) methods are
generally not appropriate (unless the analyst performs a special intervention) for the
evaluation of inclement weather conditions, accidents or construction activities, queues
that are building over both time and space, or the possible effects of vehicle guidance or
driver guidance systems typical of intelligent transportation systems. However, some
guidelines are identified in Chapter 22 to address these conditions.
The Part III methods have been designed to be sensitive to roadway, traffic, and
control characteristics of the facility. However, the methods cannot predict the effects of
changes in the posted speed limit, the level of police enforcement, safety features, driver
education, or vehicle performance.
A ground transportation system is composed of six modal and facility type
subsystems located in a defined study area or corridor. The six subsystems are freeway,
urban street, rural highway, transit, pedestrian, and bicycle. Each transportation
subsystem is composed of two or more individual facilities. The facilities within each
subsystem are all of a single type (freeway, urban street, rural highway) or mode (transit,
pedestrian, bicycle). Each facility is in turn made up of segments and points. For
example, a freeway contains basic, weaving, and ramp merge/diverge segments. An
urban street contains street segments and intersections (points).
A segment is a length of facility where demand and capacity are relatively constant.
Each segment begins and ends at a point. Segments are generally directional; for
example, each stretch of two-way street is composed of two one-way segments. The
exception to this is two-lane highways, where each segment is bidirectional but can be
split into two directional segments for analysis. A point is a very short length of facility
where demand or capacity changes abruptly from conditions on the upstream or
downstream segment.
Analysis of the transportation system proceeds from estimates of travel times and
delays at the segment and point levels using the methods described in Part III. Segment
and point delays and travel times are converted to total person hours of delay or travel
time and then summed to obtain facility estimates.


Chapters 10–14 of Part II
present general concepts

Part III, Chapters 15–27,
presents methodologies

A facility is composed of
segments and points

9-1


Chapter 9 - Analytical Procedures Overview
Introduction

Highway Capacity Manual 2000

Part IV, Chapters 28–30,
presents corridor and
areawide analyses
Part V contains
information on simulation
and other models


For analyses that combine facility types or that address a corridor or expanded area,
the analyst must consult Part IV. Part V contains useful information on applications of
simulation and other models to complement the use of HCM 2000 methodologies.
Exhibit 9-1 illustrates the content, by chapter, of the analytical sections of this
manual. Outputs from computations based on the methodologies are also indicated.
Most of the analytical processes require estimates of hourly demand in one direction.
The section on equivalency of hourly and daily volumes provides guidance on
determining directional hourly volumes from average daily traffic volumes. The
analytical procedures in Part III (Chapters 15 through 27) require information on the
geometric design, control, and demand for the facility being analyzed. The following
sections provide some brief guidance on the development of local default values for input
data that are difficult to obtain. Generic default values that may be used for specific
facility analyses in the absence of local values are provided in Chapters 10 through 14.
Some of the analytical procedures can be quite complex. Analysts may wish to
develop tables of maximum service volumes for typical highway facilities in their area.
The tables may be used in planning studies to roughly size a facility when resources do
not permit more detailed analyses. Guidance on the development of local service volume
tables is provided in Appendix B. Examples of service volume tables are given in
Chapters 10 through 14.

III. PRECISION AND ACCURACY OF THE MANUAL
The presentation of numerical values and calculations in this manual is based on a
long history of evolving methodologies for assessing capacity and quality of service. The
first HCM was produced in 1950. It was followed by a series of manuals, the last update
being the 1997 HCM. A large number of researchers and research projects in the past 50
years have contributed to the methodologies presented in this, the 2000 edition. To
provide a better understanding of the framework in which this edition was developed, the
accuracy and precision of numerical values are discussed.
The terms accuracy and precision are independent but complementary concepts.
Accuracy relates to achieving a correct answer, while precision relates to the size of the
estimation range of the parameter in question. As an example of accuracy, consider a
method that is applied to estimate a performance measure. If the performance measure is
delay, an accurate method would provide an estimate closely approximating the actual
delay that occurs under field conditions. The precision of such an estimate is the range
that would be acceptable from an analyst’s perspective in providing an accurate estimate.
Such a range might be expressed as the central value for the estimated delay plus or
minus several seconds. In general, the inputs used for the methodologies in this manual
are from field observations or estimates of future conditions. In either case, and
particularly for future conditions, the inputs can only be expected to be accurate to within
5 or 10 percent of the true value. Thus, the computations performed cannot be expected
to be extremely accurate, and the final results must be considered as estimates that are
accurate and precise only within the limits of the input values used.
To provide numerical values and computational results that are relatively easy to use
and that indicate the presumed accuracy and precision, a framework of guidelines was
established during preparation of this manual. In the following sections, an explanation
of this framework is given.
PRECISION AND ACCURACY FRAMEWORK
The user of the HCM should be aware of the limitations of the accuracy and
precision of the methodologies in the manual. Such awareness will help the user to

Chapter 9 - Analytical Procedures Overview
Overview of Analytical Procedures

9-2

Highway Capacity Manual 2000

interpret the results of an analysis and to use the results to make a decision on design or
operation of a transportation facility.
Many of the models in the HCM are based on theoretically derived relationships,
which include assumptions and contain parameters that must be calibrated on the basis of
field data. Other models in the HCM are primarily statistical. Both types require data
collected at a sampling of sites. The degree to which the models reflect reality is often
stated in terms of the accuracy and precision of the model. Accuracy and precision are
terms used to express the probable error associated with an estimate.
Frequently, after a model is developed, it is validated by comparing the estimates
from the model with values measured in the field from an independent set of sites. A
regression line fitted to the plot of points for field-measured versus model-estimated
values will result in a line with a slope different from 45 degrees. The difference can be
considered the relative accuracy of the model. The dispersion of the points around the
regression line can be considered the precision of the model. The measure of dispersion
with which many analysts are familiar is the R2 value. These statistics, based on field
and predicted data, indicate the limitations of the models in predicting with great
precision and accuracy.
Few of the models in the HCM have well-documented measures of accuracy and
precision. Typically, when research is completed and statistical relationships are
reported, the Committee on Highway Capacity and Quality of Service will exercise its
judgment in modifying the results.

Prediction error from other sources may also result when the user applies the HCM.
For example, the accuracy of results may be reduced by the use of default values for one
or more of the parameters in the models. In addition, there are limitations on the
accuracy and precision of traffic inputs used in these models. Traffic measurements and
predictions, including magnitude and mix of traffic, have inherent limitations on
accuracy.
The limitations on the accuracy and validity of predictions of performance measures
should be recognized in applying the results of an analysis. For instance, small
differences between the values of performance measures for alternative designs should
not always be assumed to be real (statistically significant) differences. Furthermore, if
the predicted value for a measure of effectiveness is near, but below, a critical threshold,
there is some probability that it will in fact be higher than predicted and exceed the
critical threshold. The HCM user should recognize, therefore, that judgment is required
in applying the results of analyses. One basis for that judgment is a good understanding
of the structure and basis of the models used in this manual.
Constraint of Prior Research Results
The methodologies in this manual have been developed by a number of researchers
working on many research projects. Few of these projects have presented results with
accompanying statements on precision and accuracy. Rather, most of the methods have
involved the use of mean or average values for parameters. Results have been presented
in a variety of forms with regard to the use of tables, graphs, and interpolated values. The
number of digits to the right of the decimal point in factors, calculated values of
performance measures, and threshold values used to define level of service has also
varied. In general, it was considered prudent to follow the presented results and the
significant figures used in prior research rather than to change the recommended values
arbitrarily. Whenever possible, the tabulated factors and adjustments and the final
calculated values of performance measures used in the reported research were maintained
for the methods in this manual.
Several factors result in
limitation on the accuracy and
precision of HCM analysis

Research precision and
accuracy

9-3


Chapter 9 - Analytical Procedures Overview
Precision and Accuracy of the Manual

Calculation Precision Versus Display Precision

Precision in calculation
differs from precision in
presenting final results

Conventions for display
of results in the HCM
The extensive use of personal computers has allowed calculations of capacity and
level of service to be carried to a large number of digits to the right of the decimal point.
Because of this ease of calculation, there is a need to state clearly that the final result of
calculations done manually and carried to the suggested number of significant figures
might be slightly different from the result of calculations performed on a computer. This
difference has been explicitly recognized in this manual. For example, lists of factors are
often displayed with three or four digits to the right of the decimal point to more closely
adhere to the calculation protocol inherent in computers.
Implied Precision from Displayed Results
The typical interpretation given to a value such as 2.0 is that the value is in a
precision range of two significant figures and that results from calculations should be
rounded to this level of precision. Occasionally, particularly in the running text of the
manual, editorial flexibility allows a zero to be dropped from the number of digits. In
most cases, however, the number of the digits to the right of the decimal point does imply
that a factor or numerical value has been calculated to that level of precision.
Directives from TRB Committee
Prior to publication of this manual, the Committee on Highway Capacity and Quality
of Service (A3A10) developed guidelines for the presentation of results. The guidelines
were presented in mid-1997 in the form of advice to the preparers of this manual. Several
recommendations were included and were particularly aimed at the exhibits and values
shown and used in Chapter 16, Signalized Intersections. This advice was considered,
along with the factors mentioned above, in developing the HCM.
Specific Components for Presentation Guideline
The overall objective of the guideline is to present tabular values and calculated
results in a consistent manner throughout the manual. Another objective is to use a
number of significant digits that is reasonable and indicates to the analyst that the results
are not extremely precise but take on the precision and accuracy associated with the input
variables. As stated earlier, such accuracies for traffic volume counts and measurement
of geometric conditions seldom are better than a central value plus or minus 5 percent.
Prediction to a futur