EPA-454/R-92-006
User's Guide to CAL3QHC Version 2.0:
A Modeling Methodology for Predicting
Pollutant Concentrations Near
Roadway Intersections
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Technical Support Division
Research Triangle Park, NC 27711
November, 1992
-------
DISCLAIMER
This report has been reviewed by the Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency and has been approved for publication. Any mention of
trade names or commeraal products is not intended to constitute endorsement or
recommendation for use.
-------
TABLE OF CONTENTS
LIST OF FIGURES
• • v
LIST OF TABLES
vii
ACKNOWLEDGEMENTS
"" *' • vui
1 INTRODUCTION
2 BACKGROUND . . .
• •• 3
3 MODEL DESCRIPTION •
3.1 Overview _
3.2 Site Geometry .
3.2.1 Free Flow Links
3.2.2 Queue Links r
3.2.3 Receptor Locations -.'..........!..!..!.. 11
3.3 Emission Sources
3.3.1 Free Flow Links
. 3.3.2 Queue Links ' ]?
13
3.4 Queuing Algorithm
15
3.4.1 Overview
3.4.2 Queue Estimation for Under-Saturated'Conditions ' Jq
3.4.3 Queue Estimation for Over-Saturated Conditions '.'.'.'..'.'.'.'.'.'.'.'.'.'.'. 21
3.5 Dispersion Component
3.6 Future Research Areas
4 USER INSTRUCTIONS .-
4.1 Data Requirements
4.2 Limitations and Recommendations 00
• £o
4.3 Input Description . .
iii
-------
TABLE OF CONTENTS (Continued)
Pace
4.4 Run Procedure .......................... .................. 39
4.5 Output Description .................... ^ ................... 3g
4.6 Examples ........................... , ............... 40
4.6.1 Example 1: Two-way Signalized Intersection
(Under-Capacity) ................................ 41
4.6.2 Example 2: Two-way Multiphase Signalized ......................
Intersection (Over-Capacity) ................. 41
4.6.3 Example 3: Urban Highway ....... .......... '. \ . . . . . . .' ..' ', '. '. [ ', [ ' .' 42
5 SENSITIVITY ANALYSIS .." ...... '7«
.................. 'J ....... ......... /O
5.1 Overview .................... T0
' .............................. /o
5.2 Signal Timing ............... . ....................... 75
5.3 Traffic Volume on the Queue Link .... ........ . ............. 75
5.4 Traffic Lanes in the Queue Link ..... .................... 78
*
5.5 Optional Traffic Parameters ............................... 78
6 MODEL VALIDATION .... ......
................. ..................... ol
6.1 Overview .............. 0 .
........... • .......................... o i
6.2 The New York City Database .................. . g1
6.3 Modeling Methodology ............................. 82
6.4 Model Evaluation Results - ....................... ;; 83
*
6.4.1 Regulatory Default Analysis ........ 83
6.4.2 Scoring Scheme Results ............ ....................' 83
References ........
89
IV
-------
LIST OF FIGURES
Figure Title and Descrlotlon
2 Link and receptor geometry
Page
1 Flowchart for CAL3QHC routines Q
• o
10
3 Flowchart for queue link calculations 16
4 Queue and delay relationships for a near-saturated
signalized intersection 1Q
5 Queue and delay relationships for an over-saturated
signalized intersection
6 Example 1: Geometric configuration for a two-way
intersection (units are in feet) 43
7 Example 2: Geometric configuration for a two-way
multiphase intersection (units are in meters) 52
8 Example 3: Geometric configuration for an urban
highway (units are in feet) ' CA
1 • • 64
9 Sensitivity analysis example run
74
10a Variation of CO concentrations (ppm) at receptor 1
(comer) versus wind angle for three different
values of signal timing: 30% red time (V/C = 0 75
queue = 5.6), 40% red time (V/C = 0.88, queue = 9 0V
and 50% red time (V/C = 1.08, queue = 42.9).....'... 76
10b Same as Figure 10a except at receptor 2 (mid-block) 76
11a Variation of CO concentrations (ppm) at receptor 1
(comer) versus wind angle for three different
values of approach traffic volume: 1000'vph
(V/C = 0.59, queue =5.0), 1500 vph (V/C = 0.88
queue = 9.0), and 2000 vph (V/C = 1.18, queue = 93.5) -. 77
11b Same as Figure 11 a except at receptor 2 (mid-block)
77
-------
12b
LIST OF FIGURES (Continued)
Title and Description
Variation of CO concentrations (ppm) at receptor 1
(comer) versus wind angle for different number of
traffic lanes: two traffic lanes (V/C = 0.88,
queue « 9.0) and three traffic lanes (V/C = 0.59,
queue m. 5.0) ; 79
Same as Figure 12a except at receptor 2 (mid-block) 79
13 The composite model comparison measure (CM) with 95%
confidence limits using CPM statistics 87
14 CM with 95% confidence limits using the AFB of
scientific category 1 afl
VI
-------
UST OF TABLES
Table Title and Description
1 Surface Roughness for Various Land Uses 30
2 Description of Type of Variables 38
3 Example-1: Two way Signalized intersection
(Under-Capacity) 44
4 Example-2: Two way Multiphase Signalized Intersection
(Over-Capacity) 53
—5 Example-3: Urban Highway 65
6 Comparison of Top-Ten Observed Concentrations with
CAL3QHC Predicted Concentrations 84
VII
-------
-------
ACKNOWLEDGEMENTS
This report was prepared for the United States Environmental Protection Agency (EPA), Office
of Air Quality Planning and Standards (OAQPS) under contract No. 68-D90067. The authors,
Guido Schattanek and June Kahng. would like to express special acknowledgements to the '
EPA technical director, Thomas N. Braverman, for his guidance and assistance in resolving
technical issues, and to Donald C. DiCristofaro of Sigma Research Corporation, Concord,
Massachusetts, for his contribution to the update of Chapters 4, 5, and 6, and the overall'
compilation of the report.
The initial User's Guide to CAL3QHC was prepared in 1990 for the EPA/OAQPS under
Contract No. 68-02-4394 by the authors at Parsons Brinckerhoff Quade & Douglas, Inc. in
New York, New York. Special acknowledgements for their contribution to the initial report are
given to Thomas Wholley who provided the first concept of CAL3Q and offered technical
guidance; George Schewe (Environmental Quality Management) for his assistance and
direction in this effort; John Sun (Bechtel/Parsons Brinckerhoff) whose initial recommendations
led to the use of Highway Capacity Manual procedures; James Brown and Joel Soden
(Parsons Brinckerhoff) for their guidance and review of this document; and Tereza Stratou,
Steven Warshaw, and Ingrid Eng for their Fortran programming efforts.
VIII
-------
-------
SECTION 1
INTRODUCTION
CAL3QHC is a microcomputer based model to predict carbon monoxide (CO) or other inert
pollutant concentrations from motor vehicles at roadway intersections. The model includes the
CALINE-3 line source dispersion model1 and a traffic algorithm for estimating vehicular queue
lengths at signalized intersections.
•
CALINE-3 is designed to predict air pollutant concentrations near highways and arterial streets
due to emissions from motor vehicles operating under free flow conditions. However, it does
not permit the direct estimation of the contribution of emissions from idling vehicles.
CAL3QHC enhances CALINE-3 by incorporating methods for estimating queue lengths and
the contribution of emissions from idling vehicles. The model permits the estimation of total
air pollution concentrations from both moving and idling vehicles. It is a reliable tool2 for
predicting concentrations of inert air pollutants near signalized intersections. Because idle
emissions account for a substantial portion of the total emissions at an intersection, the model
•s relatively insensitive to traffic speed, a parameter difficult to predict with a high degree of
accuracy on congested urban roadways without a substantial data collection effort.
CAL3QHC requires all the inputs required for CALINE-3 including: roadway geometries,
receptor locations, meteorological conditions and vehicular emission rates. In addition
several other parameters are necessary, including signal timing data and information '
describing the configuration of the intersection being modeled.
The model has been revised to address public comments. The principal difference between
the original CAL3QHC model and the revised CA13QHC (Version 2.0) pertains to the
calculation of intersection capacity, vehicle delay, and queue length. Version 2.0 includes
three new traffic parameters that can be optionaily specifiedby the user: Saturation Row
Rate, Signal Type, and Arriva. Type. These parameters permit,more precise specification of
the operational characteristics of an intersection than in the original CAL3QHC model" If not
spec,f,ed by the user, the model defaults to a set of values;for these characteristics
representative of typical urban intersections. This revised version also replaces "stopped"
delay (used in the queue calculation) with "approach" delay. These modifications are based
-------
\ on recommendations from the 1985 Highway Capacity Manual (HCM)3. This revised version
< contains the same input/output format as the original version; the same input files may be
used with both versions of the model. CAL3QHC Version 2.0 can accommodate up to 120
roadway links, 60 receptor locations, and 360 wind angles, an increase from the original
version which could accommodate 55 links and 20 receptors. This allows the modeling of
adjacent intersections that interact with each other within a short distance.
This User's Guide is intended to provide'the information necessary to run CAL3QHC Version
2.0. Development of the model is discussed in Section 2. Section 3.contains a technical
description of how the different components and algorithms operate within the program.§ In
addition, future research areas are discussed in Section 3. Model inputs and outputs,
Instructions for executing the model on a personal computer, and example applications are
contained in Section 4. Section 5 presents a sensitivity analysis evaluating the effect of
changes in model inputs on resultant pollutant concentration estimates. Section 6
summarizes the results of model verification tests completed by the United States
Environmental Protection Agency2.
While this document includes information on CALINE-3 necessary for using the CAL3QHC
• =model, it does not describe the theory underlying CALINE-3. It is recommended that the user
consult the CALINE-3 Usei"s.Guide1 for information on the theoretical aspects of CALINE-3.
-------
SECTION 2
BACKGROUND
When originally published in 1978, Volume 9 of the EPA Guidelines for Air
Maintenance Planning and Analysis* was considered to be the most appropriate methodology
for calculating CO concentrations near congested intersections. The workbook procedure
described in Volume 9 is composed of three components: traffic, emissions, and dispersion.
Although no one model has been developed to replace all of the procedures in Volume 9, *
various procedures have been devised that have improved each component.
The manual workbook procedures included in Volume 9 are cumbersome and time consuming
to use in situations where there are numerous roadway intersections or multiple traffic
alternatives. In addition, Volume 9 utilizes an outdated modal emissions model, and its
procedures are limited to situations where the estimated volume of traffic (V) approaching an
intersection is less than the theoretical capacity (C) of the intersection (V/C<1). Consequently
during the period 1985 to 1987, Thomas Wholley and Thomas Hansen from the U.S. EPA
Regional Offices I and IV developed CAL3Q, a computer-based procedure for estimating CO
concentrations near roadway intersections. CAL3Q used the running and idling emission
rates from the U.S. EPA mobile; source emission factor model to estimate emissions, a
queuing algorithm developed by the Connecticut Department of Transportation (CONDOT) to
estimate queue lengths.and the CALINE-3 linevsource dispersion model to estimate,
dispersion.
While CAL3Q provided a means for considering the effect of queuing vehicles on pollutant
concentrations, testing of the model indicated that it failed to accurately estimate queue
lengths under near-saturated and over-saturated traffic conditions (i.e., when the approach
volume reaches or surpasses; the capacity of tharoadway). Since these conditions are
common.occurrences in many congested urban areas and are of particular concern in
determining the worst;(maximum) air quality impacts of a;Proposed action, an extensive
re-evaluation of the traffic assumptions used in determining delays and queue lengths at
congested intersections was undertaken.
3
-------
One of the principal recommendations of the re-evaluation was to replace the delay formulas
Included In CAL3Q with a hybrid methodology based on the signalized intersection analysis
technique presented in the 1985 Highway Capacity Manual (HCM)3 and the Deterministic
Queuing Theory5-8. In the hybrid methodology, a simplified 1985 HCM procedure is used to
estimate the average vehicle delay for the under-saturated condition. The additional delay
associated with over-saturation conditions is estimated based on the Deterministic Queuing.
Theory procedure. Using the average vehicle delay estimated through the hybrid
methodology, queue length is subsequently estimated based on a.queuing formula developed
by Webster7-8 and the Deterministic Queuing Theory. The revised version of CAL3Q was
named CAL3QHC, and was applied extensively to model conditions near locations where
traffic conditions were near or over the capacity of the intersection, and at complex
intersections where roadways interacted with ramps and elevated highways.
*
During 1989-1990 the U.S. EPA commissioned a performance evaluation of eight intersection
models. The results of this study indicated that of the models tested, CAL3QHC performed
well in predicting CO concentrations in the vicinity of a congested intersection. Based on the
results of that evaluation, the original CAL3QHC User's Guide was prepared for EPA OAQPS
and released in September 1990. On February 13, 1991, EPA issued a notice of proposed
rulemaking identifying CAL3QHC as the-recommended model for estimating carbon monoxide
concentrations in the vicinity of intersections.
During=199T, comments were received in response to the proposed rulemaking and as part of
the.Rfth Conference on Air Quality Modeling. Most of the commentors pointed out that; given
the great degree of variability in the operational characteristics,of a,signalized,intersection,
more consideration should be given to the calculation of delay and intersection capacity.
In order to.address these comments, the?model has:been revised to: (1) give the user more
options in determining, the capacity of an intersection, and: (2) consider the effects of different:
types of signals and arrival rates. All the changes were=based;on recommendations from the
1985 HCM.
During 1991, EPA sponsored another evaluation2 of the performance of eight different
modeling methodologies (including CAL3QHC Version 2.0) in estimating CO concentrations
-------
using both the MOBILE4 and MOBILE4.1 emission factor models. The data used for this
evaluation were collected during 1989-1990 as part of a major air quality study performed in
response to the proposed reconstruction of a portion of Route 9A in New York City, and
included traffic, meteorological, and CO data collected at six intersections during a three-
month period. The results of this evaluation indicated that CAL3QHC was one of the best
performing models.
-------
-------
\
SECTION 3
MODEL DESCRIPTION
3.1 OVERVIEW
CAL3QHC is a consolidation of the CALINE-3 line source dispersion model1 and an algorithm
that estimates the length of the queues formed by idling vehicles at signalized intersections.
The contribution of the emissions from idling vehicles is estimated and converted into line
sources using the CALINE-3 link format. CAL3QHC requires all input parameters necessary
to run CALINE-3 plus the following additional inputs: idling emission rates, the number of
"moving- lanes in each approach link and the signal timing of the intersection. Version 2.0 of
CAL3QHC also includes three additional traffic parameters that may be provided by the user-
Saturation Flow Rate, Signal Type, and Arrival Type. Figure 1 depicts the major routines of
the CAL3QHC program and how they interact. A description of these routines and how each
input parameter is used in the model is provided below.
3.2 SITE GEOMETRY
CAL3QHC permits the specification of up to 120 roadway links and 60 receptor locations
within an XYZ plane. The Y-axis is aligned due north, with wind angle inputs to the model
following.accepted meteorological convention -e.g. 270° represents a wind from the west
The positive X-axis is aligned due east. A link can be specified as either a free flow or a
queue link. The program automatically sums the contributions:from each link to each
receptor. Surface roughness and meteorological variables (such as atmospheric stability
wind speed and wind direction) are assumed to be spatially constant over the entire study
area.
-------
INPUT
SITE VARIABLES AND
RECEPTOR INFORMATION
CALCULATE THETA. THE
ANGLE FORMED BY THE
ASSUMED QUEUE LINK
AND THE COORDINATE SYSTEM
CALCULATE THE
LINK LENGTH (LL)
QUEUE LINK
CALCULATIONS
(aw FIGURE 3)
INPUT METEOROLOGICAL
DATA AND WIND
ANGLE VARIATION
INPUT NEW
[METEOROLOGICAL OATA
OUTPUT UNK co
CONCENTRATION FOR
EACH RECEPTOR AND
WINO- ANCLE
NEXT WND ANGLE RANGED
CALINE3 DISPERSION
CALCULATIONS
LAST
METEOROLOGICAL
CONDITION ?.
Figure 1. Flowchart for CAL3QHC routines.
-------
3.2.1 Free Row Links
A free flow link is defined as a straight segment of roadway having a constant width, height,
traffic volume, travel speed, and vehicle emission factor. The location of the link is specified
by its end point coordinates, X1, Y1, and X2, Y2 (see Figure 2). It is not necessary to specify
which way traffic is moving on a free flow link, but the link length must be greater than link
width for proper element resolution. A new link must be coded when there is a change in
width, traffic volume, travel speed or vehicle emission factor.
Link width is defined as the width of the travelled roadway (lanes of moving traffic only) plus 3
meters (10 feet) on each side to account for the dispersion of the plume generated by the
wake of moving vehicles. Link height cannot be greater than 10 meters (elevated section) or
less than -10 meters (depressed section), since CALINE-3 has not been validated outside of
this range. In most cases (at grade section), a link height of 0 meters should be used.
3.2.2 Queue Links
A queue link is defined as a straight segment of roadway with a constant width and emission
source strength, on which vehicles are idling for a specified period of time. The location of a
link is determined by its beginning point (i.e., X1, Y1 coordinates of the locations at which
vehicles start queuing at an intersection "stopping line") and an arbitrary end point (i.e X2 Y2
coordinates of any point along the line where the queue is forming.) (See Figure 2). The
purpose of specifying a queue link end point is to specify the direction^ the .queue. The
actual length of the queue is estimated by the program based on the traffic volume and the
capacity of the approach. (Section 3.4 describes-how queue length is estimated.)
Link width is determined by the width of the travelled roadway only (width of the lanes on
wrnch vehicles are idling). Three meters are not added on each side since vehicles are not
movmg and no wake is generated. Lane widths typically vary, between 10 feet (3 m) and 12
feet (4 m) per lane depending on site characteristics.
-------
1 1
• 0
XR.YR.ZR
(RECEPTOR COORDINATES)
~i 1
1
i
>
i
i
^
K.
>
^ — c
.s— F
. X2
s. ^
X1.Y1 (BEGINNING OF FREE FLOW LINK)
X1.Y1 (BEGINNING OF QUEUE LINK)-STOPPING LINE
QUEUE LINK WIDTH (TRAVELLED WAY ONLY)
FREE FLOW LINK WIDTH (TRAVELLED WAY+20ft or 6m)
X2.Y2 (POINT ALONG* THIS LINE, DETERMINES
DIRECTION. OF QUEUE LINK)
X2.Y2 (END OF FREE FLOW LINK)
Figure 2. Link and receptor geometry.
10.
-------
3.2.3 Receptor Locations
Receptor locations are specified in terms of X, Y, and Z coordinates. A receptor should be
located outside the "mixing zone" of the free flow links (i.e., total width of travel lanes plus 3
meters (10 feet) on each of the outside travel lanes) (See Figure 2). The mixing zone is
considered to be the area of uniform emissions and turbulence. The 10 meter (32 foot)
link"heiqht rest"c«on does not apply to receotor-heightr receptors can be specified at
elevations greater than 10 meters (32 feet) if so desired. In most applications, receptors are
entered at an assumed breathing height of 1.8 meters.
3.3 EMISSION SOURCES
*
Separate emissions estimates must be provided as input data for each free flow and queue
link. Emissions from vehicles travelling from point "A" to point "B" are calculated using the
composite emission rate for the length of the link. (This composite emission rate is the
resultant of the average speed of a driving cycle that, includes different levels of acceleration
and deceleration.) When vehicles are idling at an intersection (i.e., not moving), emissions
are calculated using the idle emission rate for the duration of the idling time. While a
sub-population of approach traffic experience idling (i.e., are queued), the number of the
queued vehicles varies significantly as discussed in section 3.4.
Although CAL3QHC can be used with any mobile source emission factor model, it is
recommended that emission source strength be estimated using the most recent version of
the U.S. EPA mobile source emission factor model (MOBILES9 is currently the most recent
version of this program), or in California, where different automobile emission standards apply
the most current version of EMFAC1' (Emission Factor program for California).
Pollutant concentration estimates are directly proportional to the emission factors used as
•nput data to the program. Consequently, the accuracy of the results.of a microscale air
quality analysis^ dependent on the accuracy of thaemission factors used. The most critical
variables affecting;the emission factors are: average link speed, vehicle operating conditions
(percent: cold/hotstarts), and ambient temperature.
11
-------
3.3.1 Free flow links
Vehicles are assumed to be travelling without delay along free:flow links. The link speed for a
free flow link represents the speed of a vehicle travelling along the link in the absence of the
delay caused by traffic signals.
It is recommended that this free flow speed be obtained either from actual field measurements
or from a traffic engineer with adequate local knowledge of the intersections under
consideration. In the absence of these information sources, the use of the free flow speeds
presented on the following page may be considered within the context of the locally posted
speed limits. However, considerable caution should be exercised in using these speeds since
they represent the traffic operating environment with minimal to moderate pedestrian/parking
frictions. In urban areas with significant pedestrian/vehicle conflicts and/or parking activities
(e.g., Central Business Districts, Fringe Business Districts), the use of substantially lower free
flow speeds (e.g., 15 mph to 20 mph) may be warranted.
Free Flow Speeds for Arteriais
(Source: 1985 Highway Capacity Manual3, Chapter 11)
Arterial Class ~~ j ~jj jjj
Range" of free flow
speeds (mph) 35 to 45 30 to 35 25 to 30
Typical free flow
speeds (mph) 40 33 27
The criteria for the classification of arterials for use in conjunction with the free flow speeds
mentioned above, are presented as follows:
12
-------
Arterial Class According to
Function and Design Category
(Source: 1985 Highway Capacity Manual3, Chapter 11)
Functional Category
n , ~ Principal Minor
Design Category ArterIPa, ™™
Suburban . „
1 II
Intermediate
(Suburban/Urban) H
Urban m
III
The composite running emission rate in "grams/vehicle mife" should be obtained for the
average link speed, operating conditions of the engine, and vehicle mix for each free flow link
using the current version of the U.S. EPA MOBILE emissions factor model, EMFAC, or other
appropnate emission estimation programs. (Appropriate inspection/maintenance program
anti-tampering program, vehicle age distribution, and analysis year must be specified to '
accurately develop emission rates.)
3.3.2 Queue Links
Vehicles are assumed to be in an idling mode of operation during a specified period of time
along a queue link. CAL3QHC assumes.that vehic.es wlM-.be-Fn.an idling mode of operation
only during the red phase of the signa. rvr,P. Based on a user-specified idling emission rate
the number of lanes of vehicles idling at the stopping line, and the percentage of red time
CAL3QHC calculates the emission source strength and converts- it to^a Iine,source value so
that the^CALINE-3 model can process it as a nominal free flow link, ^strength per unit
length of aline; source ismot dependent on ^.approach trafficvolumeor capacity These
parameters are only used to determine the Jength of the line source for the queue link.
An idle emission factor in "grams per vehicle-hour" musttie converted to "micrograms per
meter-second" to calculate linear source strength. "Grams per vehicle-hour" is converted to
"m,crograms per vehicle-hour" by multip.ying by a million., "Micrograms per vehicle-hour" is
13
-------
-\
converted to "micrograms per vehicle-second" by dividing by 3600. _ Based on the assumption
that there is a distance of 6 meters (20 feet) pen vehicle in a queue, "micrograms per
vehicle-second" is converted to "micrograms per meter- second" by dividing by 6. Thus, by
converting the units of the idling emission factor, the Linear Source Strength (Q,) for "one
traffic lane for one meter over one second" can be determined as follows:
Idle Emission factor (g/veh-hr)x106
Qi ^ • frtg/m-s]
3600 x 6
To determine the total Linear Source Strength (Q,) for a queuing link, the total number of lanes
in the queue link and the percent of time that vehicles are estimated to be idling in the queue
link must be considered. This is done by multiplying the Linear Source Strength for one lane
(Q,) by the number of traffic lanes in the link and the percent of red time during the signal
cycle. The total Linear Source Strength (Qt) for the queuing link in -micrograms per meter-
second" is calculated as follows:
Q, = Q, x number of lanes x percent red time Qig/m-s]
It is assumed that the vehicles will be in the idling mode of operation only
during the Red Time phase of the signal cycle.
CALlNE-3 estimates total Linear Source Strength (Qt) as follows:
Q, - 0.1726 x VPH x EF Qig/m-s]
•
where: VPH = Vehicles per hour
EF = Emissions factor (g/mi)
To convert the Linear Source Strength into.the^CALINE-3 format; CAL3QHC fixes one of the
two variables by assigning an arbitrary value of 10,0 to EF (as seen in the output line for the
queue link). VPH can then be calculated as follows:
14
-------
QI
VPH =
0.1726 x 100
As seen in the output line for the queue link, this VPH will give the appropriate total Linear
Source Strength for the queue link when multiplied by EF=100.
Since the current MOBILE emissions model estimates idle emission rates in "grams per
vehicle hour", CAL3QHC Version 2.0 also requires that the idle emission rate be input in
"grams per vehicle hour." (It should be noted that the original CAL3QHC required .idle
emission rate input in "grams per vehicle minute").
*
3.4 QUEUING ALGORITHM
3.4.1 Overview
Figure 3 depicts the queue length estimation procedure employed in CAL3QHC. The input
parameters required to determine the queue length are: traffic volume of the link, signal cycle
length, red time length, and clearance interval lost time. The following optional additional
parameters may also be specified:
• SFR - saturation flow rate [vehicles per hour of effective green time, vphg]
• ST - traffic signal type [pretimed (=1), actuated (=2), or semiactuated (=3)]
• AT - "arrival type" of vehicle platoon [worst (=1) through most favorable (=5)]
If any of the optional parameters are not input, the model will default to a set of conditions
typical of an urban intersection.
The capacity of an intersection approach lane is determined by applying the effective green
fme to its saturation flow rate (SFR). Saturation flow rate represents the maximum number of
verncles that:can pass through a given intersection approach lane assuming that the approach
lane had 100 percent of real time as effective: green time3. CAL3QHC Version 2.0 employs
1600:vphg;as a default saturation flow rate^to,represent an urban intersection. Saturation flow
rate may vary substantially from this'default value depending on site specific traffic conditions
and ..site geometry.
15
-------
ASSUME TIME LOST GETTING
QUEUE IN MOTION IS
MAXIMUM: K1-2.Q see.
CALCULATE:
GREEN TIME (GAVG) -
S!gnol Length (CAVG) - Red TTma (RAVG)
RED TO CYCLE RATIO (RC) -
Red Tlmo / Signal Cycle Length
INTERSECTION APPROACH
DELAY (D) CALCULATIONS
0 - d x PF x Fc
where: d » average stopped delay
PF » progression adjustment factor
Pc — stopped deiay-to-approach delay
conversion factor
, CALCULATE:
Intersection Approach Capacity (C) Vehicles/Lane/Hour
C » (3600/CAVG) . (SFR/3600) • (GAVG-K1-YFAC)
where: CAVG =• cycle length
SFR = saturation flow rate
GAVG - green time
K1 = start up delay
YFAC = clearance interval lost time
CALCULATE:
DEMAND - CAPACITY RATIO
V/C
where:
V=volume per lane
C=capac!ty per lane
CUEUE LENGTH CALCULATION
No • Mox{q x D + r/2 x q. q x r]
wh*r« q - volume per lane
D » intersection approach delay/vehicle/lane
r =• length of red phase
QUEUE LENGTH CALCULATION
FOR OVER-CAPACITY
No
Nu« + •—• (V - C)
LL =. LL«
3{V-C)
COMPUTE NEW QUEUE LINK
END COORDINATE
JL
COMPUTE NEW LINE LENGTH
ASSUMING 6m PER VEHICLE
LL - Nu «• 6
COMPUTE EMMISSION RATE FOR LINK
TER « QOLFAC . 106) . (NLANES - RC)
3600 * 6
COMPUTE THE VPL THAT
WILL. PRODUCE THET APPROPRIATE
EMISSION SOURCE
VPL -. TER/0.1726 . 100.0
SET'ASSUMED EMMISSION FACTOR
EFL = 100.0
Rgure 3, Flowchart for queue link calculations.
16,
-------
Effective green time is calculated by subtracting the amount of red time, start up delay (2.0
seconds) and the time lost during the clearance interval11 from total signal cycle length. The
clearance interval lost time represents the portion of the yellow phase (i.e. the period between
the green and red phases) that is not used by the motorists. It's value is a function of signal
timing and driver characteristics. While a clearance interval lost time of 2 seconds is
recommended as a default value to reflect "normal/average" driver behavior12, the model
permits the user to specify clearance lost time to reflect site-specific traffic conditions (e.g., 0
to 1 seconds for "aggressive" drivers and 3 to 4 seconds for "conservative" drivers)12.
Thus, the capacity of the intersection approach per lane is calculated as:
C = (SFR) x (CAVG - RAVG - K1- YFAC)
CAVG
where: C = hourly capacity per lane [veh/hr/lane]
SFR = saturation flow rate [veh/lane/hr of green time]
CAVG = cycle length [s]
RAVG = length of red phase [s]
K1 = start-up delay [s] = 2 s
YFAC = clearance interval lost time [s]
Vehicles arriving at a signalized intersection during the red phase queue-up behind the
stopping line of the approach. After the signal turns to green, the first vehicle on the queue
proceeds forward after a start-up delay of approximately 2 seconds, followed by the remaining
vehicles in the queue. This results in the propagation of a "shock-wave" traveling backwards
toward the last vehicle in the queue. Vehicles arriving during the green phase prior to the
dissipation of the queue are stopped and join.the end.of the queue.. Figure 4 illustrates:this
process, assuming a uniform vehicle arrival rate, q [vehicles/Iane/second], and a uniform
departure rate, s [vehicles/lane/second] for a near-saturated cycle (i.e., volume-to-capacity
ratio, V/C, is close to 1). In Figured, the vertical distance (Ay) between the cumulative arrival
curve, A(t), and; the cumulative:departure curve; D(t), represents the queue, on each approach
lane (i.e., the number of vehicles idling) at time t5-6. The horizontal distance (Ax) between the
two curves, t,-^, represents the stopped delay experienced by the: nth vehicle arriving at the;
17
-------
CUMULATIVE
NUMBER
OF VEHICLES
PER LANE
(vehicles/lone)
CUMULATIVE ARRIVALS PER LANE
(veh icles/lane) = A( t)
CUMULATIVE DEPARTURES'PER LANE
(vehicles/lane)=0(t)
TIME
GREEN PHASE
Figure 4.
Queue and delay relationships for a near-saturated signalized
intersection;
18
-------
\
intersection approach lane at time t= t,. The total vehicle delay for each approach lane during
the cycle is represented by the area of the triangle OCR When the approach is at a
near-saturation condition and the signal timing has a 50-50 split between red and green time,
(l.e., 50 percent of the cycle is red phase), the total vehicle delay per lane, W, may be
approximated as follows:
W = FB x OE x 1/2
= FB x OF
d)
where: W = total vehicle delay per lane during a cycle [vehicles x
second/lane]
FB = average number of vehicles queued per lane at the beginning
of the green phase [veh]
OE = cycle length [s]
OF = the duration of the red phase [s]
Since CAL3QHC assumes that the queued vehicles idle only for the duration of the red phase
(i.e., average delay is equivalent to the duration of the red phase, OF), the corresponding
queue yielding a correct estimation of total vehicle delay per lane is defined as FB, (i.e., the
number of queued vehicles at the beginning of the green phase) using the Equation (1).'
3'4'2 Queue Estimation for Under-Saturated Conditions
In the under-saturated-condition (i.e., volume to capacity ratio, v/c, is less than 1), the number
of vehicles queued.at an intersection at the beginning of the green phase is estimated based
on the following formula from Webster7-8:
FB = Nu = MAX [q x D + r/2 x q, q x r]
(2)
where: Nu =. average,queue per lane at the beginning of green phase in
under-saturated conditions [veh/lane]
q= vehicle arrival rate per lane [veh/lanes/s]
D = average vehicle approach delay [s/veh]
r= length of the red phase [s]
19
-------
For light traffic flow conditions, the second term of Equation (2), q x r gives a good
approximation of the queue at the beginning of the green phase. However, for heavier traffic
flow conditions, Webster found the first term, q x D + r/2 x q, produces a more accurate
estimate of the average queue at the beginning of the green phase. The first component of
the first term of Equation (2), q x D, represents the average queue length throughout the
signal cycle. The second component, r/2 x q, represents the average fluctuation of the queue
during the red phase. Since the queue generally reaches its maximum at the end of the red
phase (i.e., at the beginning of the green phase) in under-saturated condition, these two
components are added together in the first term to estimate the average queue at the
beginning of the green phase.
The average approach vehicle delay, D, in Equation (2) is estimated using the following
formula for signalized intersection delay given in Chapters 9 and 11 of the 1985 Highway
Capacity Manual (HCM)3:
D = dxPFxFc (3) -
where: d « average stopped delay per vehicle [s/veh]
PF = progression adjustment factor
Fc = stopped delay-to-approach delay conversion factor (= 1.3)
The first term-in Equation (3), d, the average stopped delay per vehicle, for an assumed
random arrival pattern for approaching vehicles; is estimated using the following formula from
the 1985 HCM:
d - (0.38)(G4VG) J CAVG\ +
(4)
where: GAVG=iength of green:phase, [s]
CAVG =cycle length [s]
C =hourty capacity per lane [veh/hr/lane]
X =volume-to-capacity ratio = V/C
V =houriy approach volume- per lane [veh/hr/lane]
20:
-------
The first term of Equation (4) accounts for uniform delay, (i.e., the delay that occurs if the
arrival of vehicles is uniformly distributed over the cycle). The second term of the equation
accounts for additional delay due to random arrivals and/or occasional cycle failures.
The second term in Equation (3), the progression adjustment factor (PF), is included to
account for the variation of stopped delay with traffic flow progression quality. Progression
adjustment factors are determined using the following key variables:
• Arrival Type (AT) - a general categorization of the way the platoon of vehicles
arrives at the intersection. Five arrival types are defined in
the 1985 HCM:
1 = worst platoon condition (dense platoon arriving at the beginning
of the red phase)
2 . unfavorable platoon condition (dense or dispersed platoon
arriving during the red phase)
3 = average condition (random arrivals)
4 = moderately favorable platoon condition (dense or dispersed
platoon arriving during the green phase)
5 = most favorable platoon condition (dense platoon arriving at the
beginning of the green phase)
The model uses arrival type 3 as default if it is not specified by the user.
• Signal Type (ST) - user may select one of the following three traffic signal
types:
1 = pretimed
2 = actuated
3 = semiactuated
The model assumes signal type 1 (pretimed) as default if it is: not specified by.the user In the
case of actuated or semiactuated signals, the,user must specify the estimated red time for
.each approach. . .
3'4-3 Queue Estimation for Over-Saturated Condition*
In the over-saturation condition (i.e. volume to capacity ratio, y/C, greater than one), the
queue consists of the two; components,Nr and N2, as illustrated in Figure 5. A'(t) in Figure 5
21
-------
CUMULATIVE
NUMBER .
OF VEHICLES
PER LANE
(vehicles/lane)
t=1 hour
t=2 hours
TIME
Figure o. uueue ana ae.ay relationsh.ps for an over-saturated signalized
I
intersection.
22:
-------
depicts the cumulative arrivals per lane in an over-saturated condition (i.e., V/C greater than
1)i A(t) represents the cumulative arrivals per lane during at-capacity condition (i.e., V/C
equal to 1). Other symbols are similar to those defined in Figure 4. N, is the vertical
difference between A(t) and D(t) and represents the normal fluctuation of a queue during
at-capacity conditions due to change of signal phase (i.e., from green to red, etc.). As shown
in Equation (5), the estimate of the average of N, at the beginning of the green phase,
denoted by Nu*. is identical to that of Nu, which can be estimated based on the procedures
provided in section 3.4.2.:
Nu* = MAX [q* x D* + r/2 x q*, r x q*]
(5)
where: q* = vehicle arrival rate per lane during at-capacity operating conditions
(i.e. V/C = 1.0) [veh/lane/s]
D* = average vehicle delay during at-capacity operating conditions (i.e.
V/C = 1.0) [s/veh]
r = length of the red phase [s]
N2, which is the vertical difference between A'(t) and A(t), represents the additional queue
resulting from over-saturation. In the over-saturated condition, N2 continues to grow until the
slope of A'(t)-is lower than that of A(t). Thus, the average of N2, denoted by N2*. for the first
hour can be estimated as one half of the difference between the A'(t) and,A(t) at t = 1. hour as
shown in the following equation:
N2* = 112 x [A'(t)-A(t)], at t = 1 hour
= 1/2 x (V-C) ' (6)
s
where: N2* =, average additional queue per lane due to over-saturation [veh/lane]
Av(t) ~ cumulative vehicular arrivals per !ane in over-saturated condition
[veh/lane]'
A(t) = cumulative vehicular arrivals per lane in at-capacity condition
[veh/lane]
V = hourly approach volume per lane (i.e., A'(t) at t = 1 hour) [veh/lane/hr]
C = hourly capacity per lane (i.e., A(t) at t = 1 hour) [veh/lane/hr]
23
-------
Therefore, the average queue at the beginning of the green phase during over-saturated
conditions, N0, may be approximated by the following equation:
MAX [q* x D* + r/2 q*, r x q*] + 1/2 x (V-C) (7)
where: N0« average queue per lane at the beginning of the green phase in an
over-saturated condition [yeh], •
q*, D*. r, V and C are the same as defined in Equations (5) and (6).
For both under- and over-saturated situations, the length of the queue link is calculated by
multiplying the number of vehicles in the queue by 6 m (20 ft) per vehicle. If the predicted
queue extends into the next intersection, it is recommended to stop the queue at the end of
the modeled block by adjusting the specified link endpoints.
3.5 DISPERSION COMPONENT
The dispersion component used in CAL3QHC is CALINE-3, a line source dispersion model
developed by the California Department of Transportation. CALINE-3 estimates air pollutant.
concentrations resulting from moving vehicles on a roadway based on the assumptions that
pollutants emitted from motor vehicles-travelling along a segment of roadway can be
represented as a "line source" of emissions, and that pollutants will disperse in a Gaussian
distribution from a defined "mixing zone" over the roadway being modeled. For a complete
discussion of the theory and application of CALINE-3 the user is referred to CALJNE-3: A
.Versatile Dispersion Model for Predicting Air Pollutant'Levels Near Highways and Arterial
Streets1.
3.6 FUTURE RESEARCH AREAS
While CAL3QHC includes improved procedures for estimating air pollutant levels in the vicinity
of intersections, there remain potential areas of further study which could result in higher
levels of accuracy in completing air quality studies. These include:
24
-------
The derivation of queue length for the under-saturated condition (i.e., V/C less or equal
to 1) was simplified by assuming a near-capacity (i.e., V/C approximately equal to 1)
operation and an even-split of signal timing (i.e. 50% of the cycle length is green phase).
This procedure works the best for near and over-saturated conditions (i.e., conditions of
most concern) but it could be refined to produce a more precise estimation of queue
length for cases deviating significantly from the assumed condition.
The average additional queue due to over-saturation was assumed to be idling only
during the red phase of the signal cycle. Further investigation is required to fully
validate this assumption.
While the model provides the general concept for estimating emissions at signalized
intersections, there remain other traffic controls, such as stop signs or toll plazas, where
a similar concept could be extended. Future research and testing,is necessary to adapt
this program for such situations.
•
The model assumes flat topography. Its handling of vehicular queuing could be adapted
to urban canyon situations.
25
-------
-------
SECTION 4
USER INSTRUCTIONS
4.1 DATA REQUIREMENTS
The accuracy of the results of a microscale air quality analysis is directly dependent on the
accuracy of the input parameters. Meteorology, traffic, and emission factors can vary widely
and in many situations there is a great degree of uncertainty in their estimation. The user
should have a high degree of confidence in these data before proceeding to apply the model.
It is recommended that the user contact the EPA or appropriate state or local air pollution
control agency prior to selecting meteorological parameters and estimating composite running
and idling emission factors, since these factors depend on many variables unique.to a
particular region (e.g.,* thermal state of engines, ambient air temperatures, local inspection and
maintenance program, and anti-tampering credits all vary by region).
The following parameters are required input to the program, (Section 4.2 provides
recommendations on how to use these factors and Section 4.3 describes their location in the
input file):
Meteorological Variables:
Averaging Time [min]
Surface Roughness coefficient [cm]
Settling Velocity [cm/s]
Deposition Velocity [cm/s]
Wind Speed [m/s]
Stability Class [1 to 6 = A to F]
Mixing Height [m]
Site Variables:
Roadway Coordinates [X,Y,2][m or ft]
Roadway Width [m or ft]
Receptor Coordinates [X,Y,Z] [m or ft]
27
-------
I
Traffic Variables:
Traffic Volume [each link] [veh/hr]
Traffic Speed [each link] [mi/hr]
Average Signal Cycle Length [each intersection] [s]
Average Red Time Length [each approach] [s]
Clearance Lost Time [s]
Saturation Flow Rate [veh/hr] [optional]
Signal Type [pretimed, actuated, or semiactuated] [optional]
Arrival Rate [worst, below average, average, above average, best
progression] [optional]
Emission Variables:
, Composite Running Emission Factor [each free flow link] [g/veh-mi]
Idle Emission Factor [each queue link] [g/veh-hr]
4.2 LIMITATIONS AND RECOMMENDATIONS
CAL3QHC can process up to 120 links.and 60 receptor locations for al!360 degree wind
angles. A new link is required when there is a change in link width, traffic volume, travel
speed or emission factor.
In specifying link geometry, link length must always be greater than the link width.
Otherwise, correct element resolution cannot be calculated (error message will appear).
• ^ Since emissions from idling vehicles account for a substantial portion of the total
* emissions from an intersection, it is recommended that roadway segments up to 1000
feet from the intersection of interest be included, in the site geometry. Testing of the
model indicates that links beyond 1000 feet from the receptor locations will have a minor
contribution to the results.
• In overcapacity situations, where V/C > 1, the • model predicted queue length" could be.
larger than the physical roadway configuration. The user could either revise the traffic
assumption for the link, or limit the length of the queue by running the analysis in the
1 following manner: 1) inputthe queue link as;a free flow,link; 2) specify X1, YT,-X2...Y2:
28
-------
coordinates that determine the physical limits of the queue (i.e., the physically largest
queue length); and 3) input the emission source as the equivalent VPH (from the output
run on the queue link) with an emission rate of EF=100. This will provide the
appropriate emission source for the queue link with the manually determined queue
length.
When the site specific clearance lost time (portion of the yellow phase that is not used
by motorist) is unknown, a default value of 2 seconds may be used.
Source height should be within ± 10 m (± 32 ft), (+10 m for an elevated roadway section
and -10m for a depressed section). CALINE-3 has not been validated outside this
range (error message will appear). In most applications (at-grade), a source height of 0
•m should be used. • "
Receptor height should be greater than the roadway height, except for elevated roadway
sections, since CALINE-3 assumes plume transport over a horizontal plane. The 10 m
height limitation does not apply to receptors; which may be placed at any height above
the raodway. For most applications, receptors should be placed at an assumed
breathing height of 1.8 m.
Wind speed should be at least 1 m/s. (CALINE-3 has not been validated for wind
speeds below 1 m/s).
Surface roughness coefficient ,(z0) should be within the range of 3 cm to 400 cm. Table -
1. which is reprinted from the CALINE-3 manual, provides the recommended surface
roughness coefficients for various land uses.
Averaging time should be within the range of 30 min to 60:mln. The most common
value is 60 min, since most predictions are performed for a one'hour period.
Mixing height should be generallyset at 1000 m. CALINE-3 sensitivity to mixing height
is significant only for extremely low values (much less than 100 m).
29
-------
TABLE 1
SURFACE ROUGHNESS LENGTHS (ZJ FOR VARIOUS LAND USES
Type of Surface
Smooth desert
Grass (5-6 cm)
Grass (4 cm)
Alfalfa (15.2 cm)
Grass (60-70 cm)
Wheat (60 cm)
Com (220 cm)
Citrus orchard
Fir forest
City land-use
Single family residential
Apartment residential
Office
Central business district
Park
Z0 (cm)
0.03
0.75
0.14
2.72
11.40
22.00
74.00
198.00
283.00
108.00
370.00
175.00
321.00
127.00
30:
-------
Free flow link width should be equal to the width of the traveled roadway plus 3 m (10 ft)
on each side of the roadway (to account for the mixing zone created by the dispersion
of the plume generated by the wake of moving vehicles).
Queue link width should be equal to the width of the traveled roadway only.
Receptors should always be located outside of the mixing zone (link width) of the free
flow and queue links. In the case of urban intersections/where buildings are located
closer than 3m (10 ft) from the roadway and the speed of the traffic is very slow, a
reduced mixing zone should be considered to maintain receptor locations outside of the
mixing zone.
It is recommended that the link speed information be obtained from traffic engineers
familiar with the area under consideration. The link speed for a free flow link represents
the speed experienced by drivers travelling along the link in the absence of the delay
caused by traffic signals. In the absence of recommended information from traffic
engineers, the use of the free flow speeds presented in Section 3.3.1-may be
considered. .
The saturation flow rate or the hourly capacity per lane should be determined by the
user depending on the: characteristics and operation of the intersection. The default
value is 1600 vehicles per hour which is representative of an urban intersection.
The signal type should be input as:
1 = Pretimed
2 = Actuated
3-Semiactuated
The defaultvaiue is pretimed (1). In the case of actuated or semiactuated signals, the
use? must.input the estimated red time for each approach.
31
-------
The arrival type should be input as:
1 s Worst progression (dense platoon at beginning of red)
2 m Below average progression (dense platoon during middle of red)
. 3 3 Average progression (random arrivals)
4 » Above average progression (dense platoon during middle of green)
5 m Best progression (dense platoon at beginning of green)
The default value is 3 for average progression (random arrivals).
Notes:
If a CAL3QHC file produced for the original version is run with Version 2.0, the idle
emission factor must be input in g/hr (instead of the original g/min). The rest of the
input format is the same with the only addition of the optional traffic parameters. If
the user does not specify these optional traffic parameters, the model will default to
a saturation flow rate of 1600 vph, pretimed signal type, and a progression that
assumes random arrivals. An identical file run for both versions of the program
(assuming default optional traffic parameters) should result in equal or larger queue
lengths with the associated effects in CO concentrations for Version 2.0.
If CAL3QHC were used to predict CO concentrations near highways or arterial
streets where only free flow links interact (i.e., not for a signalized intersection), it
would produce the same results as CALINE-3.
4.3 INPUT DESCRIPTION
The following is a tabular description of the CAL3QHC variables .and identifies their position in
the input data file. The "format" description of each variable is explained in Table 2.
LINE
NUMBER
VARIABLE
NAME
FORMAT COLUMNS
VARIABLE
DESCRIPTION
JOB
ATIM
ZO
A40
F4.0
F4.0
1-40
41-44
45-48
Current job title.
Averaging time [minj.
Surface roughness [cm].
32
-------
LINE VARIABLE
NUMBER NAME FORMAT
VS F5.0
VD F5.0
NR |2
SCAL F10.4
IOPT H
IDEBUG 11
2 RCP A20
XR F10.0
YR F10.0
ZR F10.0
-
COLUMI
49-53
54-58
59-60
61-70
75
80
1-20
21-30
31-40
41-50
VARIABLE
DESCRIPTION
Settling velocity [cm/s].
Deposition velocity [cm/s].
Number of receptors,max=60.
Scale conversion factor [if
units are in feet enter 0.3048,
if they are in meters enter
1.0].
Metric to english conversion
• in output option. Enter "1" for
output in feet, if left blank,
the output will be in meters. '
Debugging option. Enter "1"
for this option which will
cause the input data to be
echoed onto the screen. The
echoing process stops when
an "error is detected.
Receptor name.
X-coordinate of receptor.
Y-coordinate of receptor.
Z-coordinate of receptor.
Repeat line 2 for NR (number of receptors) times
33
-------
LINE VARIABLE :
NUMBER : NAME FORMAT
COLUMNS
VARIABLE
DESCRIPTION
RUN
NL
NM
A40
13
13
1-40
41-43
44-46
PRINT2
IQ
12
49-50
13
1-3
Current run title.
Number of links, max=120.
Number of meteorological
conditions, unlimited number.
Each unique wind speed,
stability class, mixing height,
or wind angle range
constitutes a new
meteorological condition.
Enter "1" for the output that
includes the receptor - link
matrix tables (Long format),
enter "0" for the summary
output (Short format).
Enter "1" for free flow and "2"
for queue links
Enter lines 5a and 5b for IQ=2 (queue link).
**** Enter line 5c for IQ=1 (free flow link) ****
5a
LNK
TYP
A20
A2
1-20
2-1-22
XL1
YL1
XL2
F7.0
F7.0
F7.0
23-29
30-36
37-43
Link description.
Link type. Enter "AG" for "at
grade" or "FL" for "fill," "BR"
for "bridge" and "DP" for
"depressed".
Link X-coordinate for end
point 1 at intersection
stopping line.
Link Y-coordinate for end
point 1 at intersection
stopping line.
Link X-coordinate for end
point 2.
34
-------
LINE VARIABLE
NUMBER : NAME
YL2
HL
WL
NLANES
5b CAVG
RAVG
YFAC
IV
IDLFAC
SFR
ST
FORMAT
F7.0
F8.0
F4.0
14
15
15
F5.1
IS
F7.2
14
11
COLUM
44-50
51-58
59-62
63-66
6-10
16-20
26-30
31-35
36-42
44-47
49-
VARIABLE
DESCRIPTION
Link Y-coordinate for end
point 2.
Source height.
Mixing zone width.
Number of travel lanes in
queue link.
Average total signal cycle
length [s].
Average red total signal cycle
length [s].
. Clearance lost time (portion of
the yellow phase that is not
used by motorist) [s].
Approach volume on the
queues link [veh/hr].
Idle emission factor [g/veh-
hr].
Saturation flow rate
[veh/hr/lane]. •
Signal type. Enter "1" for
pretimed, "2" for actuated, "3"
for semiactuated. Default is
"1."
AT
11
51
Arrival rate. Enter "1" for .
worst progression, "2" for
below average progression,
"3" for average progression,
"4" for above average
progression, "5" for best
progression. Default is 3.
35
-------
LINE VARIABLE
NUMBER ; NAME
5c LNK
TYP
XL1
YL1
XL2
YL2
VPHL
I
FORMAT COLUMNS
A20 1-20
A2 21-22
F7.0 23-29
F7.0 30-36
*
F7.0 37-43
F7.0 44-50;
F8.0 51-58
,
VARIABLE
DESCRIPTION
Link description.
Link type. Enter "AG" for "at
grade" or "FL" for "fill," "BR"
for "bridge" and "DP" for
"depressed".
Link X-coordinate for end •
point 1.
Link Y-coordinate for end
point 1.
Link X-coprdinate for end
point 2. "
Link Y-coordinate. for end.
point 2.
Traffic volume on link
EFL
HL
WL
F4.0
F4.0
F4.0
59-62
63-66
67-70
[veh/hr].
Emission factor [g/veh-mi].
Source height.
Mixing zone width.
Repeat lines 4 and 5 for NL (number of links) times
U
BRG
F3.0
F4.0
1-3
4-7
Wind speed [m/s].
Wind angle (0-360 degrees,
0=positive Y axis). Enter 0 if
angle variation data follow.
Enter actual wind angle, if
only one wind angle will be
used.
36
-------
LINE VARIABLE, VARIABLE
NUMBER NAME ^ FORMAT COLUMNS DESCRIPTION
CLAS I1 8 Stability class.
MIXH F6-0 9-14 Mixing height [m].
AMB F4.0 15-18 . Ambient background
concentration [ppmj.
VAR A1 '19 Enter "Y" if angle variation
data follow. Enter "N" if only
one angle [BRG] will be
considered.
DEGR l3 20-22 Increment angle [degrees].
VAI(1) l3 23-25 Lower boundary of the
variation range(First
increment multiplier).
VAI(2) l3 26-28 Upper boundary of the
variation range (Last
increment multiplier).
*** Repeat line 6 for .each time that new ***
*** meteorological conditions ***
*** are to be run ***
37
-------
•\
TABLE 2
DESCRIPTION OF TYPE OF VARIABLES
FORMAT
VARIABLE
TYPE EXPLANATION*
Ax
Ix
Fx.y
CHARACTER
INTEGER
REAL
Input a string that has a maximum of'"x" number
of characters.
Input an integer that has a maximum of "x"
number of digits. The integer should be right
justified, e.g., I3.--12
Input a real number that consists of a total of "x"
digits (including the decimal'point). The. real
number can have up to "x-1" digits to.the right of
the decimal point, e.g., F8.3.--234.156
(*) The symbol"-" denotes a blank space.
as:
-------
4.4 RUN PROCEDURE
CAL3QHC is designed to operate on any IBM compatible personal computer. A math
co-processor is not required, but its use will speed the overall program run time considerably.
The memory requirements are 512 KB. A hard disk is not needed, but if it is available, the
program should be copied onto the hard disk.
To execute the program, at the DOS prompt, type:
CAL3QHC
-------
shows the estimated number of vehicles in the queue. (This number, multiplied by
6 m/veh, determines the length of the queue as used in the program).
The second page of the output shows the queue specific,input parameters: cycle
length, red time, clearance lost time, approach volume, saturation flow rate, idle
emission factor, signal type, and arrival rate.
The second section on the second page lists the receptor locations and the X, Y, Z
coordinates (in ft or m) for each receptor.
The third page lists the model results in parts per million (ppm). Two output versions are
available. The short version of the output (summary table) lists the total CO concentration
(ppm) at each receptor for each wind angle analyzed, together with the maximum total-
concentration at each receptor with the corresponding angle. The long version of the output
prints the-same summary table with the total CO concentrations for each receptor as printed
in the short version, plus a table showing the contribution from each link to the total CO
concentration at each receptor for the angle where the maximum total CO concentration
occurs.
In the case where multiple meteorological conditions are run, one printout with all the results
will be generated for each meteorological condition. The following section describes three
examples showing the different types of output that could be generated. "
4.6 EXAMPLES
Three example cases are described in this section: 1) a signalized intersection with an
under-capacity situation where V/C ratios are less than'1.0 for all approaches; 2) a two way
multiphase intersection with an over-capacity situation, where-V/C. ratios-are above 1.0 for
some approaches; and 3) an urban highway where only free flow links interact.
In order to highlight how the model could be used, all these examples were kept as simple as
possible, however realistic values for traffic parameters, emission rates, and roadway
40
-------
configuration were used. For all cases, a map showing the geometric configuration of the
intersection being modeled is followed by a description-of all input parameters and the model
input/output formats.
4'6-1 Example 1; Two-wav Signalized intersection fUnder-Capacltvl
This intersection consists of a two-way main street intersecting a one-way local street. Figure
6 shows the geometric configuration of the site and the X, Y coordinates of each link and
receptor location. Table 3 shows all the input parameters with their corresponding units, in
the same order as they are used in the input file. This example uses default values for the
optional traffic parameters.
4'6'2 Example 2; Two-wav Multiphase Signalized Intersection (Over-Capacity)
This example consists of a two-way main street with exclusive left turning bays intersecting
with a two-way local street. The signal cycle of this intersection is considered a three phase
signal, where the left turning movements from the main street (Northbound and Southbound
left turns) have an exclusive green phase/separate from the main street green phase for the
through traffic. Figure 7 shows the geometric configuration of the site and the X, Y
coordinates of each link and receptor locations. Table 4 shows all the input parameters with
their corresponding units, in the same order as they are used in the input file. This example
uses user-specified optional traffic parameters.
In order to show a variation of the short output format, several wind angle ranges with different
wind speeds were run:
1st wind angle range from 150° to 210,° in 5° increments,
wind speed = 1 m/s
2nd wind angle range from 240° to 300° in 3° increments,
wind speed = 1 m/s
3rd,wind angle range from 330° to 70° (430°) in 10° increments,
wind speed = 2 m/s
41:
-------
4.6.3 Example 3: Urban Highway
This example consists of a two-way highway with an exit ramp, where only free flow links
Interact. Rgure 8 and Table 5 show the geometric configuration of the site and all the input
parameters with their corresponding units in the same order as they are used in the input file.
In this case the long version of the output format is printed. The second page of the output
shows the summary table with results for all wind angles, and the third page shows the
contribution from each link for the angle producing the maximum concentration at each
receptor.
42
-------
TABLE 3
EXAMPLE - 1: Two-way Signalized Intersection (Under-Capacity)
Default optional traffic parameters
Input and output in feet
Description of Parameters:
Site Variables:
Averaging time (ATIM)
Surface roughness length (z,,)
Settling velocity (VJ'
Deposition velocity (Vd)
Number of receptors
Scale conversion factor
Output in feet
Main St. NB Approach I ink-
X1, Y1 coordinates
X2, Y2 coordinates
Traffic volume
Emission factor
Source height
- Mixing zone width
Main St. NB Queue Link:
X1, Y1 coordinates
X2, Y2 coordinates
Source height
Mixing zone width
Number of travel lanes
Avg. signal cycle length
Avg. red time length
Clearance lost time
Approach traffic volume
Idle emission factor
60 min
175 cm
0 cm/s
0 cm/s
8
0.3048(units are in ft)
1.,
10,-1000 (ft)
10, 0 (ft)
1500 veh/hr
41.6g/veh-mi (*)
Oft
40ft
10,-10 (ft)
10,-1000 (ft)
0
20ft
2
90s
40s
3 s
1500 veh/hr
735.0 g/veh-hr (")
44
-------
TABLE 3 (Continued)
Main St. NB Departure Link:
X1, Y1 coordinates
X2, Y2 coordinates
Traffic volume
Emission factor
Source height
Mixing zone width
Main St. SB Approach Link;
X1, Y1 coordinates
X2, Y2 coordinates
Traffic volume
Emission factor
Source height
Mixing zone width
Main St. SB Queue Link: -
X1, Y1 coordinates
X2, Y2 coordinates
Source height
Mixing zone width
Number of travel lanes
Avg. signal cycle length
Avg. red time length
Clearance lost time
Approach traffic volume
Idle emission factor
Main St. SB Departure Link;
X1, Y1 coordinates
X2, Y2 coordinates
Traffic volume
Emission factor
Source height
Mixing zone width
10, 0 (ft)
10, 1000 (ft)
1500 veh/hr
41.6g/veh-mi (*)
Oft
40ft
-10, 1000 (ft)
-10, 0 (ft)
1200 veh/hr
41.6g/veh-mi (*)
Oft
40ft
-10, 10 (ft)
-10, 1000 (ft)
Oft
20ft
2
90s
40s
3s
1200 veh/hr
735.0 g/veh-hr (**)
-10, 0 (ft)
-10, -1000 (ft)
1200 veh/hr
41.6g/v'eh-mi
Oft
40ft
45
-------
TABLE 3 (Continued)
Local St. Approach Link-
X1, Y1 coordinates
X2, Y2 coordinates
Traffic volume
Emission factor
Source height
Mixing zone width
Local St. Queue Link:
X1, Y1 coordinates
X2, Y2 coordinates
Source height
Mixing Zone Width
Number of travel lanes
Avg. signal cycle length
Avg. red time length
Clearance lost time
Approach traffic volume
Idle emission factor
Local St. Departure Link:
X1, Y1 coordinates
X2, Y2 coordinates
Traffic volume
Emission factor
Source height
Mixing zone width
Site Meteorology
Wind speed
Wind angle
Stability class
Mixing height
Background concentrations
-1000, 0 (ft)
0, 0 (ft)
1000veh/hr
41.6g/veh-mi (*)
Oft
40ft
-20, 0 (ft)
-1000, 0 (ft)
Oft
20ft
2
90 s
50:s:; '
3s
1000 veh/hr
735.0 g/veh-hr (**)
0, 0 (ft)
1000, 0 (ft)
1000 veh/hr
41.6 g/veh-mi (*)
Oft
20ft
1 m/s
0°
4(D)
1000 m
0.0 ppm
46
-------
TABLE 3 (Continued)
Site Meteorology (Continued)
Multiple wind angles
Increment
Rrst increment multiplier
Last increment multiplier
Yes
10°
0°
36
(*) Emission factor = 41.6 g/veh-mi, obtained from MOBILE 4.1 emission factor model
assuming: average speed = 20 mph, Year 1990, ambient temperature = 30° F default for
vehicle mix and thermal states, no I/M program, no ATP program, RVP = 11.5 psi and
ASTM * C.
H Idle emission factor. 735.0 g/veh-hr obtained from MOBILE 4.1 emission factor model.
47
-------
INPUT EXAMPLE 1
6.0
6.0
6.0
6.0
6.0
910
EXAMPLE - TWO WAY INTERSECTION (EX-1) 60 175
REC 1 (SE CORNER) 45. _35 ' g*0
REC 2 (SW CORNER) -45. _35' g'Q
REC 3 (NW CORNER) -45. 35' g"0
REC 4 (NE CORNER) 45. 35'
REC 5 (E MID-MAIN) 45. _150*
REC 6 (W MID-MAIN) -45. -150*
REC 7 (N MID-LOCAL) -150. 35*
REC 8 (S MID-LOCAL) -150. -35'
MAIN ST. AND'LOCAL ST. INTERSECTION
1
Main St.NB Appr.
2
Main St.NB Queue
90
1
Main St.NB Dep.
1
Main St.SB Appr.
2
Main St.SB Queue
0. 0.
0.3048
AG
AG
40
AG
AG
AG
10. -1000.
10.
3.0
10.
-10.
-10.
-10.
1500
0.
1000.
10.
10. 0.
10. -1000.
735.0
10. 1000.
-10. 0.
-10. 1000.
0. 1500. 41.6 0. 40.
0. 20.0 2
90
40
3.0 1200 735.0
0. 1200. 41.6 0. 40.
0. 20.0 2
Main St.SB Dep. AG -10. 0.
Local St.Appr.Lnk. AG -1000. 0
2
Local St.Queue Lnk. AG -20. 0. -1000.
90 50 3.0 1000 735.0
Local St.Dep.Lnk. AG 0. 0 1000
1.000.41000. 0. Y 10 0 36
-10. -1000. 1200. 41.6 0. 40.
0. 0. 1000. 41.6 0. 40.
0. 0. 20.0 2
0. 1000. 41.6 0. 40.
48
-------
OUTPUT EXAMPLE 1 (Short Version)
OATEl 01/25/52
CAL3QHCI LINE SOURCE DISPERSION MODEL - VERSION 2.0, JANUARY'1992
11 R<™' MAIN ST. AND' LOCAL ST. INTERSECTION
"° HAY IHTER3ECTIOM
TIMBl 17(20
SITE < METEOROLOGICAL VARIABLES
V3 - O.fl CM/S VD - 0.0
0 - 1.0 H/3 CLA3 - 4
LIHK VMIABLE3
LIKK DESCRIPTION
1.
2.
3.
4.
S.
«.
7.
1.
9.
Main St. MB Appr.,
Kaln St. MB QUIUI
Main St. KB Dip.
Kiln 3t.3B Appr.
Hlin 3C.3B QU«C«
Klin 3t.3B Dtp.
Local 3t.Appc.Lnlc.
Local 3t.Quiu* Lnk.
Local st.Dip.Lnk.
XI
10.
"10.
10.
-10.
-10.
-10.
-1000.
-20.
0.
0
0
0
0
0
0
0
0
0
CH/S 20 - 175.
(D) ATIH - 60.
LINK COORDINATES (FT)
VI X2
-1000
-10
0
1000
10
0
0
0
0
.0
.0
.0
.0
.0
.0
.0
.0
.0
10.
10.
10.
-10.
-10.
-10.
0.
-165.
1000.
0
0
0
0
0
0
0
4
0
MI3CH - 1000. M AMB - 0.0 PPM
12
0
-238
1000
0
141
-1000
0
0
0
.0
.5
.0
. U
.2
.a
.a
.a
.0
LENGTH
(FT)
1000.
229.
1000.
1000.
131.
1000.
1000.
145.
1000.
BRG
(DEC)
360.
180.
360.
180.
360;
180.
90.
270.
90.
TYPE
AG
AC
AC
AG
AC
AG
AC
AG
AG
VPH
1500.
1752.
1500.
1200.
17S2.
1200.
1000.
2191.
1000.
EF
(G/MI)
41.6
100.0
41.6
41.6
100. D
41.6
41.6
100.0
41.6
(
n
0
n
n
0
0
n
0
0
H
FT)
0
.0
n
n
.0
n
0
.0
.0
H
-------
OUTPUT EXAMPLE 1 (Continued)
n™ " T"° MAY INTERSECTION (EJC-1)
DATE: 01/25/92 TIME: 17:20
ADDITIONAL QOEDE LINK PARAMETERS
HUN: MAIN ST. AND LOCAL ST. INTERSECTION
PAGE 2
LINK DESCRIPTION
2. Main St. OB Queue'
5. Main St. S3 Queue
8. Local St.Queue' Lnk.
RECEPTOR LOCATIONS
RECEPTOR
1. REC 1 (SE CORNER)
2. REC 2 (SN CORNER)
3. REC 3 (NH CORNER)
4. REC 4 (HE CORNER)
5. REC 5 IE MID-MAIN!
6. REC, 6 (1C MID-MAIN)
7. REC 7 (H MID-LOCAL)
8. REC 8 (S MID-LOCAL)
CYCLE
LENGTH
(SEC)
90
90
90
RED CLEARANCE
TIME LOST TIME
(SEC) (SEC)
40 3.0
40 3.0
50 3.0
APPROACH SATDRATION IDLE SIGNAL
•VOL FLOW RATE EM-FAC TYPE
(VPH) (VPH/L) <9B/hr)
1500 1600 735.00 1
1200 1600 735.00 1
1000 1600 735.00 1
COORDINATES (FT)
X Y z
45.0
-45.0
-45.0
45.0
45.0
-45.0
-150.0
-150.0
-35.0
-35.0
35.0
35.0
-150.0
-150.0
35.0
-35.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
ARRIVAL
RATE
3
3
3
50
-------
TABLE 4
EXAMPLE - 2: Two-way Multiphase Signalized Intersection (Over-Capacity)
User specified optional traffic parameters
Input and output in meters
Description of Parameters:
Site Variables:
Averaging time (AT1M)
Surface roughness length (z0)
Settling velocity (VJ
Deposition velocity (Vd)
Number of receptors
Scale conversion factor
Main St. NB Approach Link:
X1.Y1 coordinates
X2, Y2 coordinates
Traffic volume
Emission factor
Source height
Mixing zone width
Main St. NB Queue Link:
X1, Y1 coordinates
X2, Y2 coordinates """'""
Source height
Mixing zone width
Number of travel lanes
Avg. signal cycle length
Avg. red time length
Clearance lost time
Approach traffic volume
Idle emission factor
Saturation flow rate
Signal type
Arrival rate
60 min
175 cm
0 cm/s
0 cm/s
8
1.0 (units are in m)
4.7, -305 (m)
4.7, 0 (m)
1730veh/hr
41.6g/veh-mi (*)
0 m
12m
4.7, -6.2 (m)
4.7, -305 (m)
0 m
6.2m
2
90s
45s
2s
1500 veh/hr
720.0 g/veh-hr (**)
1700 veh/hr/lane
actuated (2)
worst progression (1)
53
-------
TABLE 4 (Continued)
Main St. NB Queue Left Turn-
X1, Y1 coordinates
X2, Y2 coordinates
Source height
Mixing zone width
Number of travel lanes
Avg. signal cycle length
Avg. red time length
Clearance lost time
Approach traffic volume
Idle emission factor
Saturation flow rate
Signal type
Arrival rate
Main St. NB Departure I ink-
X1, Y1 coordinates
X2, Y2 coordinates
Traffic volume
Emission factor
Source height
Mixing zone width
Main St. SB Approach Link-
X1, Y1 coordinates
X2, Y2 coordinates
Traffic volume"
Emission factor
Source height
Mixing zone width
Main St. SB Queue Link!
X1, Y1 coordinates
X2, Y2 coordinates
Source height
Mixing zone width
Number of travel lanes
Avg. signal cycle length
Avg. red time length
0, -6.2 (m)
0, -60 (m)
0 m
3.1 m
1
90s
75s
2s
230 veh/hr
720:0 g/veh-hr (**)
1400 veh/hr/lane
actuated (2)
average progression (3)
4.7, 0(m)
4.7, 305 (m)
1500 veh/hr
41.6g/veh-mi(*)
0 m
12m
-4.7, 305 (m)
-4.7, 0 (m)
1950. veh/hr
41.6g/veh-mi (*)
0 m
12m
-4.7, 6.2 (m)
-4.7, 305 (m)
0 m
6.2m
2
90 sec
45 sec
54
-------
TABLE 4 (Continued)
Main St. SB Queue Link (Continued):
Clearance lost time
Approach traffic volume
Idle emission factor
Saturation flow rate
Signal type
Arrival rate
Main St. SB Queue Left Turn
X1, Y1 coordinates
X2, Y2 coordinates
Source height
Mixing zone width .
Number of travel lanes
Avg. signal cycle length
Avg. red time length
Clearance lost time
Approach traffic volume
Idle emission factor
Saturation flow rate
Signal type
Arrival rate
Main St. SB Departure Link;
X1, Y1 coordinates
X2, Y2 coordinates
Traffic volume
Emission factor
Source height
Mixing zone width
Local St. EB Approach Link.-
X1, Y1 coordinates
X2, Y2 coordinates
Traffic volume
Emission factor
Source height
Mixing zone width
2s
1750veh/hr
720.0 g/veh-hr (**)
1800veh/hr/lane
actuated (2)
worst progression (1).
0, 6.2 (m)
0, 60 (m)
0 m
.3.1 (m)
1
90s
75s
.2s
200 veh/hr
720.0 g/veh-hr (**)
1400 veh/hr/lane
actuated (2)
average progression (3)
-4.7, 0 (m)
-4.7, -305 (m)
.1750 veh/hr
41.6 g/veh-mi (*)
0 m
12m
-305, -3.1 (m)
0, -3.1 (m)
450 veh/hr
41.6 g/veh-mi (*)
0 m '
12m
55
-------
TABLE 4 (Continued)
Local St. EB Queua Link;
X1, Y1 coordinates
X2, Y2 coordinates
Source height
Mixing zone width
Number of travel lanes
Avg. signal cycle length
Avg. red time length
Clearance lost time
Approach traffic Volume
Idle emission factor
Saturation flow rate
Signal fype
Arrival rate
Local St. EB Departure I ink-
X1, Y1 coordinates
X2, Y2 coordinates
Traffic volume
Emission factor
Source height
Mixing zone width
Local SL WB Approach I ink-
X1, Y1 coordinates
X2, Y2 coordinates
Traffic volume
Emission factor
Source height
Mixing zone width
Local St. WB Queue Link-
X1, Y1 coordinates
X2, Y2 coordinates
Source height
Mixing zone width
Number of travel lanes
Avg. signal cycle length
Avg. red time length
-7.8, -3.1 (m)
-305, -3.1 (m)
0 m
6.2 m
2
90s
60s
2s
450 veh/hr '
720.0 g/veh-hr (**)
1400 veh/hr/lane
actuated (2)
average progression (3)
0, -3.1 (m)
305, -3.1 (m)
680 veh/hr
41.6 g/veh-mi (*)
0 m
12m
305, 3.1 jm)
0, 3.1 (m)
510 veh/hr
41.6 g/veh-mi (*)
0 m
12m
7.8, 3.1 (m)
305, 3.1 (m)
0 m
6.2m
2
90s
60s
56
-------
TABLE 4 (Continued)
Local St. WB Queue Link (Continued);
Clearance lost time = 2s
Approach traffic volume = 510 veh/hr
Idle emission factor = 720.0 g/veh-hr (**)
Saturatfon flow rate = 1400 veh/hr/lane
Signal type = actuated (2)
Arrival rate _ average progression (3)
Local St. WB Departure Link:
X1, Y1 coordinates = o, 3.1 (m)
X2, Y2 coordinates = .305, 3.1 (m) -
Traffic volume = 710 veh/hr
Emission factor ' = 41.6 g/veh-mi (*)
Source height „ Om
Mixing zone width _. 12 m
Site Meteorology for wind anale range (150 to 210° in 5° increments)
Wind speed = 1 m/s
Wind angle _ Q°
Stability class = 4 (D)
Mixing height . = 1000m
Background concentrations = o.O ppm
Multiple wind angles = Yes
Increment _ 50
First increment multiplier = 30
Last increment multiplier = 42
Site Meteorology for wind angle range (240 to 300° in 3° incre
Wind speed = -j m/s
Wind angle _ 0°
Stability class . = 4 (D)
Mixing height = 1000m
Background concentrations = o.O ppm
Multiple wind angles = yes
Increment _ 30
First increment multiplier - QQ
Last increment multiplier - 100
57
-------
TABLE 4 (Continued)
Site Meteorology for wind anole range K330 to 70° I430°1 in 10<
\All~^ ____J ~"~"~~~~~""~~ ~~~"•
speed
Wind angle
Stability class
Mixing height
Background concentrations
Multiple wind angles
Increment
First increment multiplier
Last increment multiplier
2 m/s
0°
4(D)
1000m
0.0 ppm
Yes
10?
33
43
() Emiss.on factor . 41.6 g/veh-mi, obtained from MOBILE 4.1 emission factor model
assuming: average speed = 20 mph, Year 1990, ambient temperature = 30°F
default for vehicle mix and thermal states, no I/M program, no ATP program RVP =
11.5 psi, and ASTM « C.
(**) Idle emission factor
model.
720.0 g/veh-hr obtained from the MOBILE 4.1 emission factor
58
-------
INPUT EXAMPLE 2
REG 1 (SE CORNER)
REG 2 (SW CORNER)
REC 3 (NW CORNER)
REG 4 (NE CORNER)
REC 5 (E MID-MAIN)
REC 6 (W MID-MAIN)
REC 7 (N MID-LOCAL)
REC 8 (S MID-LOCAL)
MAIN ST. AND LOGS
1
Main St.NB Appr.
2
Main St.NB Queue
90
2
Main St.NB Q.Left
90
1
Main St.NB Dep.
Main St.SB Appr-.
2
Main'St.SB Queue
90
2
Main St.SB Q.Left
90
1
Main St.SB Dep.
Local St.EB Appr.
2
Local St.EB Queue
90
. 1
Local St.EB Dep.
Local St.WB Appr.
2
Local St.WB Queue - -AG
90
1
Local St.WB Dep. AG 0
1.000.41000. 0. Y 5 30 42"
1.000.41000. 0. Y 3 80100
2.000.41000. 0. Y 10 33 43
ULTIP
)
)
L)
L)
L ST.
AG
AG
45
AG
75
AG
AG
AG
45
AG
75
AG
AG
AG
60
AG
AG
- -AG
60
BASE INT
16.7
-16.7
-16.7
16.7
16.7
-16.7
-45.7
-45.7
. (EX-2)
-13.7
-13.7
13.7
13.7
-45.7
-45.7
13.7
. -13.7
60.175. 0. 0. 8 1.0
1.8
1.8
1.8
1.8
1.8
1.8
1.8
1.8
INTERSECTION 14 30
4.7
4.7
2.0
0.0
2.0
4.7
-4.7
-4.7
2.0
0.0
2.0
-4.7
-305.
-7.8
2.0
0.
305.
7.8
2.0
-305. 4
-6.2 4.
1500 720.0
-6.2 0.
230 720.0
0. 4.
305. -4.
6.2 -4.
1750 720.0
6.2 0.
200 720.0
0. -4.
-3.1 0
-3.1 -305
450 720.0
-3.1 305
3.1 0
3.1 305
510 720.0
.7 0. 1730. 41.6 0." 12
.7 -305. 0. 6.2 2
1700 2 1
. 0 -60. 0'. 3.1 1
1400 2 3
,7 305. 1500. 41.6 0. 12.
7 0. 1950. 4-1.6 0. 12.
7 305. 0. 6.2 2
1800 2 1
0 60. 0. 31 l
1400 2 3
7 -305. 1750. 41.6 0 . 12 .
. • -3.1 450. 41'.6 0. 12.
-3.1 0. 6.2 2
1400 2 3
-3.1 680. 41.6 0. 12.
3.1 510. 41.6 0. 12.
3.1 0. 6.2 2
1400 2 3
3.1 -305. 3.1 710.
41.6 0. 12,
59
-------
OUTPUT EXAMPLE 2 (Short Version)
CAL30JIC: Lira SOURCE DISPERSION MODEL - VERSION 2.0. JANUARY 1992
INT. (SX-2) Rra, «„,, ST. ^ 1OCM( „ IHTERSECTIOH
PAGE 1
3IT8 « KETgQ»OLOalCAL VARIABLES
VS - 0.0 CM/S
-------
OUTPUT EXAMPLE 2 (Continued)
INT-(EXr2)
RUM: MAIN ST. AND LOCAL ST. INTERSECTION
PAGE 2
ADDITIONAL QUEUE LINK PARAMETERS
2.
3.
6.
7.
10.
13.
LINK DESCRIPTION
Main St.NB Quaua
Main St.NB Q.Left
Main St. SB Queua
Main St. SB Q.Left
Local St.EB Quaue
Local St.NB Queua
CYCLE
LENGTH
(SEC)
90
90
90
90
90
90
RED '
TIME
(SEC)
45
75
45
75
60
60
CLEARANCE
LOST TIME
(SEC)
2.0
2.0
2.0
2.0
2.0
2.0
APPROACH
VOL
(VPH)-
1500
230
1750
200
450
510
SATURATION
FLOW RATE
(VPH/L)
1700
1400
1800
1400
1400
1400
IDLE
EM FAC
(gm/hr)
720.00
720.00
720.00
720.00
720.00
720.00
SIGNAL
, TYPE
2
2
2
2
2
2
ARRIVAL
RATE
1
3
1
3
3
3
RECEPTOR LOCATIONS
RECEPTOR '
1. REC 1 (SE CORNER)
2. REC 2 (SH CORNER)
3. REC 3 (NK CORNER)
4. REC 4 (NE CORNER)
5. REC 5 (E MID-MAIN)
6. REC 6 (W MID-MAIN)
7. REC 7 (N MID-LOCAL)
COORDINATES (M)
X Y 2
16.7
-16.7
-16.7
16.7
16.7
-16.7
-45.7
8. REC 8 (S MID-LOCAL) -45.7
MODEL RESULTS
-13.7 1.
-13.7 1.
13 . 7 1
13.'7 1.'
-45.7 l
-45.7 1.
13.7 1
-13.7 i'.t
REMARKS : In aaarch or th« angla corresponding to
tha maximum concsntration, only tha first
angla, of tha angles with saraa maximum
concantrations, is indicated aa maximum.
MIND ANGLE'RANGE: 150.-210.
HIND
ANGLE
(DEGR)
ISO.
1S5.
160.
165.
170.
175.
180.
185.
190.
195.
200.
205.
210.
MAX
DEGR.
CONCENTRATION
REC1
0.0
0.1
0.4
1.0
2.2
4.0
6.1
8.2
9.9
10.9
11.3
11.3
10.9
11.3
200
(PPM)
REC2
9.3
9.5
9.3
9.0
7.8
6.4
4.7
3.1
1.6
0.7
0.3
0.1
0.0
9.5
155
REC3
11.8
12.7
13.3
13.6
13.1
11.9
10.2
8.6
6.9
5.9
5.3
4.7
4.2
13.
165
6
REC4
6.
6.
7.
7.
9.
11.
13.
14.
16.
16.
15.
14.
13.
16.
195
5
8
2
7
0
0
0
9
2
5
8
8
7
5
REC5
0.0
0.1
0.4
0.9
1.9
3.2
5.1
7.1
8.6
10.0
10.6
10.7
10.6
10.7
205
REC6
8.7
8.7
8.5
7.9
6.9
5.5
4.0
2.7
•1.5
0.7
0.3
0.1
0.0
8.7
150
HEC7
6.5
6.0
5.4
4.6
3.9
2.9
2.1
1.7
1.4
1.2
1.2
1.2
1.2
6.5
150
REC8
4.0
3.8
3.4
2.8
2.'1
1.4
0.8
0.4
0.1
0.0
0.0
0.0
0.0
4.0
150
THE HIGHEST CONCENTRATION IS 16.50 PPM AT 195 DEGREES FROM REC4
61
-------
OUTPUT EXAMPLE 2 (Continued)
JOB: EXAMPLE-TWO MAY MULTIPHASE INT. (EX-2)
METEOROLOGICAL VARIABLES
O - 2.0 M/3 CLAS - 4 (D) ATIM
MODEL RESULTS
RON: MAIN ST. AND LOCAL ST. INTERSECTION
60. MINUTES MIXa - 1000. M AMB - 0.0 PPM
*n March of the angle corresponding to
the maximum concentration, only the first
angle, oC the angle* with same maximum
concentrations. Is indicated a* maximum.
WIND ANGLE RANGE! 330.-430.
HIND
ANGLE
(DEGR)
330.
340.
350.
360.
370.
380.
390.
400.
410.
420.
430.
MAX
DEGR.
•COHC
REC1
6.
7.
7.
5.
• 3.
2.
2.
2.
1.
1.
1.
7.
350
•E
3
3
4
7
a
»
4
2
7
2
0
4
HTRAl
(PPM)
REC2
3.
3.
5.
7.
9.
8.
7.
6.
5.
5.
S.
9.
370
T(
3
7
1
3
1
8
4
0
3
1
0
1
DH
REC3
0.0
0.3
1.5
3.7
5.9
6.3
5.9
5.5
5.0
4.9
4.7
6.3
380
REC4
S.O
5.3
4.7
2.8
1.0
0.2
0.0
0.0
0.0
0.0
0.0
5.3
340
RECS
6.8
7.3
6.3
4.S
2.3
1.0
o.s
0.4
0.3
0.3
0.3
7.3
340
RECS
•0.5
1.0
2.5
4.3
5.6
5.8
5.4
5.5
5.0.
4.6
4.3
S.8
380
REC7
0.0
0.0
0.1
0.7
1.7
2.5
2.7
2.7
2.4
2.3
2.2
2.7
390
REC8
0.6
0.6
0.7
1 4
3.2
3.7
4.0
4.C
4.6
4.6
420
THE HIGHEST CONCENTRATION IS 9.10 PPM AT 370 DEGREES FROM REC2
63
-------
-------
5 REC. 3
(-110.-70
. 4
.(-110.-200
"(500,0)
(0,-2pOO) m REC. 2
^(50, -200)
Flgure 8- Example 3: Geometric configuration for an urban highway (units
are in
64
-------
-------
TABLES
EXAMPLE - 3: Urban Highway
Description of Parameters:
Input and Output in meters
Site Variables:
Averaging Time (ATIM)
Surface roughness length (zj
Settling velocity (VJ
Deposition velocity (Vd)
Number of receptors
Scale conversion factor —
Northbound Link 1:
X1, Y1 coordinates
X2, Y2 coordinates
Traffic volume
Emission factor
Source height
. Mixing zone width
Northbound Link 2:
X1, Y1 coordinates
X2, Y2 coordinates
Traffic volume
Emission factor
Source height
Mixing zone width
Exit Ramp Link 3:
X1, Y1 coordinates
X2, Y2 coordinates
Traffic volume
Emission factor
Source height
Mixing zone width
60 min
175 cm
0 cm/s
0 cm/s
4
0.3048 (units are in ft)
0, -2000 (ft)
0, -50 (ft)
5000 veh/hr
29.6 g/yeh-mi (*)
Oft
60ft
0, -50 (ft)
0, 2000 (ft)
4000 veh/hr
29.6 g/veh-mi (*)
Oft
60ft
0, -50 (ft)
70, 0 (ft)
1000 veh/hr
54.0 g/veh-mi (**)
Oft
40ft
65
-------
TABLE 5 (Continued)
Exit Ramp Link 4:
X1, Y1 coordinates
X2, Y2 coordinates
Traffic volume
Emission factor
Source height
Mixing zone width
Southbound Link 5:
X1, Y1 coordinates
X2, Y2 coordinates
Traffic volume
Emission factor
Source height ——
Mixing zone width
Site Meteorology
Wind speed
Wind angle
Stability class
Mixing height
Background concentrations
, Multiple wind angles
Increment
First increment multiplier
Last increment multiplier
70, 0 (ft)
500, 0 (ft)
1000 veh/hr
54.0 g/veh-mi (**)
Oft
40ft
-60, 2000 (ft)
-60, -2000 (ft)
5000 veh/hr
29.6 g/veh-mi (*)
Oft
60ft
1 m/s
0°
4(D)
1000m
0.0 ppm
Yes
10°
0
36
O
n
Emission factor
29.6 g/veh-mi, obtained from the MOBILE 4.1 emission factor
Speed - 55 mph' Year 1990' ambjent temperature =
n
54'° 9/Veh"ml' °btained from tne MOBILE 4-1 ^ission factor
e Speed = 15 mph' Year 1990, ambient temperature =
thermal StateS' n° I/M pra^m' "o ATP program,
66
-------
OUTPUT EXAMPLE 3 (Continued)
JOB: EXAMPLE - URBAN HIGHWAY- (EX-3)
DATE: 08/25/92 TIME: 17:24
RECEPTOR - LINK MATRIX FOR THE ANGLE PHODUCIue
THE MAXIMUM CONCENTRATION TOR EA?H RECEPTOR
RUNt URBAN HIGHWAY (FREE FLOW LINKS ONLY)
c t
1
2
3 •
4
CO/LINK
(PPM)
ANGLE (DEGREES)
REC1
.200
S.2
0.0
0.0
0.0
S 2.8
REC2
200
5.2
0.0
0.0
0.0
2.8
REC3
160
2.8
0.0
0.0
0.0
S.2
REC4
160
2.8
0.0
0.0
0.0
5.2
71
-------
-------
SECTION 5
SENSITIVITY ANALYSIS
5.1 OVERVIEW
The CAL3QHC model includes the CALINE-3 line source dispersion model1 and a traffic
algorithm for estimating vehicular queue lengths at signalized intersections. Because
CAL3QHC includes CALINE3, the sensitivity analyses presented in the CALiNE-3 manual are
directly applicable to CAL3QHC. The user should refer to the CALINE-3 manual for
discussion of the sensitivity of the model results with respect to: wind speed, atmospheric
stability, highway width, highway length, surface roughness, averaging time, deposition
velocity, settling velocity, wind angle, source height, mixing height, and median width.
•
Because its difference with the CALINE-3 model relates primarily to the handling of vehicular
queues, the two areas in which CAL3QHC warrants a.separate sensitivity, discussion are: 1)
the emission source strength of the vehicles in the queue, and 2) the link length representing
the number of vehicles in a queue. The variability of these two parameters results in a
nonlinear relationship between the source strengths and the predicted concentrations - as
opposed to CALINE-3, where predicted concentrations are directly proportional to source
strengths. The three variables that directly affect the calculation of vehicular queues are:
signal timing, traffic volume on the queue link, and number of traffic lanes in the queue link.
To determine the effect of the variability of each of these parameters on resultant pollutant
levels, a series of model runs was performed in which each of these three parameters were
individually varied. The sensitivity runs were performed for a single roadway segment
representing two traffic lanes (each 10 feet wide) with one receptor near the corner and one
receptor at mid-block. Figure 9 shows the configuration of the roadway segment and the
variables used in the sensitivity run. Plots were then completed depicting CO concentrations
versus wind angle (with 180° representing a parallel wind and 270' representing a crosswind)
73
-------
-LU-
CROSS
WIND
(270 degrees)
REG. 1
(CORNER)
Site Variables
Averaging Time =
Surface Roughness :
Settling Velocity :
Deposition Velocity
Wind Speed =
Wind Direction =
Stability Class
Backround Concentration =
Mixing Height =
! 60 min
175 cm
0
0
1 m/sBC
(variable)
4 (0)
0
1000 meters
feet
feet
Receptor Locations
REC.1 (CORNER) 35,-35.6
REC.2 (MIDBLOCK) 35.-150.6
Link Variables
Approach Link: XI, Y1 coordinates = 0.-1000 feet
X2, Y2 coordinates = 0)0 feet
Source Height
Mixing Zone Width
Traffic Volume
Emission Factor
= 0
= 40 feet
= 1500 VPH
= 40.7 (gr*veh/mile)
REG. 2
(MID-BLOCK)
= 0.-10 feet
= 0.-1000 feet
= 20 feet
= 2
Queue Link: X1, Y1 coordinates
X2, Y2 coordinates
Mixing Zone Width
Number of Travel Lanes
Average Signal Cycle Length = 90 sec
Average Red Time Length = 36 sac
Clearance Lost Time
Traffic Volume
Idle Emission Factor
= 2 sec
= 1500 VPH
= 735 g/hr.
PARALLEL
WIND
(180 degrees)
Figure 9. Sensitivity analysis example run.
74
-------
5.2 SIGNAL TIMING
Signal timing affects the computation in two ways. The emission source for the queue links
depends on both the idling emission factor and the fraction of red time (the larger the fraction
of red time, the stronger the emission source). In addition, the length of the queue is
determined by the volume to capacity ratio (V/C) of the approach link. Since the capacity of
the Imk is affected by the fraction of red time, the longer the red phase, the smaller the
available capacity, and the longer the queue length.
Three cases were analyzed: 30 percent red time, 40 percent red time, and 50 percent red
fme. As seen in Figure 10, the increase in percent red time results in an increase of CO For
the corner receptor the peak concentration, which occurs under a cross wind condition in the
case of 30 percent red (low V/C and short queues), shifts toward an almost parallel wind
cond,tion for the 50 percent red case (higher V/C and longer queues). For the midblock
receptor, the CO increase is substantial when the iength of the queue reaches the midblock
location.
5.3 TRAFFIC VOLUME ON THE QUEUE LINK
An increase in the traffic volume on an approach link will result in a longer queue length but
w,H not effect the strength of the emission source for the queue link. As explained in Section
3.3.2, the strength of this line source is not dependent on the approach volume. Three
approach volumes were evaluated: 1000, 1500, and 2000 vehicles per hour (VPH) As seen
in Figure 11, an increase in traffic volume results in increased CO concentrations and a shift
•n peak CO values from a cross wind situation, in the case of short queues, to a parallel wind
condrtwn. as the queues get longer. For the midblock receptor, the CO increase is substantial
when the length of the queue reaches the midblock location.
75
-------
1SO -ISO 200 210 220 33O 2-tO 250 ISO
Wind Angl. rrilcr«M3
d 30» r«d tin* + -to* r«> time o SCSI rmd tin*
Figure 10a. Variation of CO concentrations (ppm) at receptor 1 (comer) versus wind
angle for three different values of signal timing: 30% red time (V/C = 0.75,,
queue » 5.6 vehicles), 40% red time (V/C = 0.88, queue = 9.0 vehicles),
and 50% red time (V/C = 1.08, queue = 42.9 vehicles).
1BO 1BO 300 210 330 330 340 ISO 3BO 270
Wind Anal. COWXMB3
_, a 3€» rad tin* + H0» rwl tin. o SO* r«d tine
Ffgure 10b. Same as Figure 10a except at receptor 2 (midblock)
76
-------
Figure 11 a.
Var ation of CO concentrations (ppm) atreceptor 1 (corner) versus wind
angle for three different values of approach traffic volume: 1000 vph (V/C
0.59, queue = 5.0 vehicles), 1500 vph (V/C = 0.88, queue = 90 vehices*
and 2000 vph (V/C = 1.18, queue = 93.5 vehicles)! vehicles),
220 230 240 250
Vlnd Angle Cd»or«««3
O 1000 VPH + 1300 VPH O 2000 VPH
260 270
Figure 11 b. Same as Figure 11 a except at receptor 2 (mid-block).
77
-------
5.4 TRAFFIC LANES IN THE QUEUE UNK
The number of moving lanes affects the computations in two ways. First, the strength of the
emission source for a queue link is directly proportional to the number of moving lanes (e.g.
doubling the number of lanes at an intersection will double the source strength). Second, the
addition of lanes increases capacity. Thus by adding more available lanes with the roadway
traffic volume held constant, the length of the queue is shortened. The net effect of these two
components on CO concentrations is dependent on the wind angle and the relative location of
the receptor with respect to the intersection. An increase of the number of available traffic
lanes will not necessarily result in a reduction of predicted CO concentration, since the
strength of the line source will increase (more rows of idling vehicles), but the queue will
shorten (less vehicles queuing per lane). If the receptor is very close to the intersection, with
a larger number of lanes under cross-wind conditions, higher CO levels may be predicted; but
if the receptor is further away from the intersection, a smaller number of lanes (a longer
queue) under near parallel winds will result in higher predicted CO levels. Two cases were
analyzed for two and three'traffic lanes for the approach. As seen in Figure 12, even though
the case with three traffic lanes has more capacity and shorter queues compared with that of
two traffic lanes, the cross wind condition .results in higher CO concentration at the corner '
receptor in the case of three traffic lanes. For the midblock receptor, which is farther away
from the intersection, two traffic lanes (with the longer queues) result in higher CO
concentrations.
S3 OPTIONAL TRAFFIC PARAMETERS
The three optional traffic parameters (Saturation Flow Rate, Signal Type, and Arrival Type)
affect the calculation of intersection capacity, delay, and queue length.
78
-------
I
100.
_1_
200 210 Sao.
Wind Angle
O a traffic linw + 3 traffic Ivm
2-4O 250 2BO 270
Figure 12a.
Var ation of CO concentrations (ppm) at receptor 1 (corner) versus wind
angle for different number of traffic lanes: two traffic lanes (V/C = 0 88
queue = 9.0 vehicles) and three traffic lanes (V/C = 0.59, queue = 50*
vehicles).
' 3°° aia M°230 240250iS j^-
Wlnd Anal* i
a 3 traffic Ima + 3 traffic I.
Figure 12b. Same as Figure 12a except at receptor 2
(mid-block).
79
-------
Saturation flow rate Is used in the calculation of intersection capacity and V/C ratio, having a
direct effect of the calculation of approach delay; the lower the saturation flow rate, the higher
the delay. Signal type and arrival type are used in the calculation of the progression
adjustment factor which has an effect on the approach delay but not on the V/C ratio; the
worst the progression, the higher the approach delay.
The effect of these parameters on the resulting CO concentrations is only significant when the
intersection operates at medium to high V/C ratios (near or over saturation conditions), which
are the conditions when higher delay results in longer queues and higher CO levels. In the
case of light traffic conditions (low V/C ratios), the change in approach delay has minimum
effect on the length of the queue and the resulting CO levels.
80
-------
SECTION 6
MODEL VALIDATION
6.1 OVERVIEW
The U.S. EPA completed the performance evaluation of eight intersection models in simulating
CO concentrations at the six intersections monitored as part of the Route 9A Reconstruction
Project ,n New York City2. The eight models evaluated included CAL3QHC Version 2 0
FHWAINT", GIM", EPAINT". CAUNE4". VOL9MOB4 (Volume 9< updated with MOBILE4)
TEXIN216, and IMM17. A complete phase I model evaluation study was conducted using
MOBILE4 emissions estimates. The phase I evaluation included all eight intersection models
at ail six intersections. In late 1991, the MOBILE4.1 emissions model9, an update to
MOBILE4, was released. Thus, a phase II evaluation utilizing MOBILE4.1 was conducted
us,ng a subset of the intersection models. Of the three EPA intersection models (EPAINT
VOL9MOB4, and CAU3QHC), CAL3QHC performed best using MOBILE4. Of the two models
unhang the FHWA advocated average speed approach rather than explicit queuing (FHWAINT
and GIM), GIM performed best. Therefore, the phase II MOBILE4.1 analysis was performed
for the following five models: CAL3QHC, GIM, IMM, TEXIN2, and CALINE4. When collecting
and compiling the New York City database, the best quality assurance procedures (e g
analyse and comparison of collected data) were followed at two of the six intersection sites
S,te #1 (West/Chambers) and Site #2 (34th/8th). A uniform wind analysis (e.g., similar wind
speed and direction for different monitors at the same intersection) conducted for each site
•ndicated that Sites #5 (34th/12th) and #1 are best in terms of unhindered approach wind
flows and wind field uniformity. Thus, the phase-1! MOB.LE4.1 analysis was performed for the
intersections at Sites #1, 2, and 5.
6.2 THE NEW YORK CITY DATABASE
•
A major air quality monitoring study was conducted in 1989-1990 in response to the proposed
reconstruction of a portion of Route 9A in New York City. As part of the monitoring project
meteorological and CO air qua.ity data were collected at two background sites (Battery Park
and Post Off,ce) and six different intersections.(Site #1 West/Chambers; #2 34th/8tir #3
65th/Broadway-*4 57th/7th; #5 34th/12th; #6 Battery Tunnel). These sites are al, located in
81
-------
mldtown or lower Manhattan. The meteorological data collected at each intersection included
wind direction, wind speed, temperature, and the fluctuation of the wind direction (sigma
theta). These data were measured at two towers per intersection except at site #2 where they
were measured at three towers. The meteorological measurements were taken at a height of
10m±1 m.
In order to obtain detailed information concerning the traffic characteristics, a series of video
cameras were used to film the traffic at each site. Three months of continuous traffic data
were collected at each site producing approximately 13,000 hours of video recordings. A •
limited number of videotaped hours from the Route 9A Study were examined in order to obtain
detailed information about the local traffic. The collected traffic data were concurrent with the
observed CO data. The examined traffic data are comprised of the top 50 hours of CO
concentrations observed-for each of three months at Sites #1 and 2 and the top 25 hours
observed for each of three months at the remaining sites.
All traffic data were obtained from videotapes except for the acceleration/deceleration rates
and the cruise speed. The acceleration/deceleration rates and cruise speeds were obtained
through the use of a vehicle outfitted with a travel-log machine that recorded instantaneous
speed versus time while traveling. Cruise speeds were taken directly from the strip charts
created in this way; acceleration/deceleration rates were determined from the slope of the
lines on the strip charts.
6.3 MODELING METHODOLOGY
The hourly averaged temperature data from the meteorological towers at each site were
averaged to calculate a site specific hourly value. For the remaining meteorological input
data, the meteorological tower closest to the CO monitor location was used. Mixing heights of
1000 m were used, since the results are not affected if the mixing height is between 100 and
1000 m high and mixing heights below 100 m in Manhattan do not occur on a frequent basis.
The closest background concentration (Battery Park or the Post Office Station) was subtracted
out of the observed concentration at each monitor. All modeling was performed for one hour
82
-------
averages only. After the removal of the background concentrations, a screening threshold of
0.5 ppm was used. When both the observed and predicted concentrations at a monitor are
less than 0.5 ppm that data pair was eliminated from the data set.
A surface roughness length of 3.21 m was used for approach flows over numerous city -
blocks. Lower values of the surface roughness length (0.03 m) were used at Site #5 (34th/-
12th) when the intersection was exposed to flows over the Hudson River without intervening
bu,.dmgs. For modeling CO concentrations, the settling velocity and deposition velocity were
set at zero because CO is a gaseous emission. An averaging time of 60 minutes was used
F.nally, a temperature-sensitive conversion of the modeled concentrations from mg/m' to parts
per million (ppm) was 'conducted.
6.4 MODEL EVALUATION RESULTS
6-4.1 Regulatory Default
The ten hours with the highest observed concentrations were used to compare the CAL3QHC
predicted concentrations using the regulatory default meteorology to the observed concen-
trates. The comparisons for each site are presented in Table 6. The regulatory default
meteorological conditions are defined as: Wind Speed = 1 m/s; Stability Class = D- Sigma-
Theta = 25°; Observed Temperature; and "Worst Case" Wind Direction Angle (determined '
using ten degree increments).
At Site #1, the highest observed CO concentration of 10.6 ppm is nearly matched (10.4 ppm)
™ ,™QHC UnPalred !n tlme °f SPaCe" At Site *2' the maximum Predicted concentration by
CAL3QHC of 8.0 ppm underpredicts the maximum observed concentration of 1 1 5 ppm
Finally, at Site #5, the maximum observed concentration of 15.5 ppm is nearly matched by
CAL3QHC which predicts 15.1 ppm.
6-4-2 Scoring Scheme Result*
A method for aggregating component results of model performance (using the observed
meteorology) into a single performance measure18 was used to compare the overall
performance' of the five models evaluated at three intersection sites. The bootstrap
83
-------
re-sampling technique19 was used to determine the significance of differences in composite
performance between models.
The statistical analysis uses the robust highest concentration (RHC) for one-hour averages. The
RHC is based on a tail exponential fit to the upper end of the distribution and is calculated as
follows
where x = average of the n-1 largest values
x(n) = nth largest value
number of values exceeding the threshold value (n=26 or less)
n
The size of the three intersection data sets requires the value of n to be less than 26. The value
of n was nominally set to 11 so that the number of values averaged (x) was 10. In general, the
RHCs are largest using the operational (or entire) dataset for each site. When calculating either
the fractional bias (FB) or the absolute fractional bias (AFB),
AFB = \FB\ = 2
(OB - PFh
(6)
(OB + PR)
where OB and PR refer to the averages of the observed and'predicted values, the RHC is used
rather than the mean of the highest 10 concentrations. The RHC is preferred in this type of
statical evaluation because of its stability-. Also, the bootstrap distribution of the RHCs is not
arfficially bounded at the maximum predicted or observed concentration, which allows for a
continuous range of concentrations.
•
When comparing these performance measures, one would like to know if differences are
s.gn,f,cant. Simultaneous confidence intervals for each pair of models were calculated20 in order to
ensure an adequate confidence level and to protect against falsely concluding that two mode.s are
d,fferent. A composite performance measure (CPM) is calculated for each model as a weighted
Imear combination of the individual absolute fractional bias components. The operational
85
-------
component is given a weight that is equal to the weight of the combined scientific components.
The scientific component refers to the evaluation of peak concentrations during specific
meteorological conditions and the operational component refers to the evaluation of peak averages
independent of meteorological conditions. The results from the different data bases (intersections)
are given equal weight The CPM is defined as
CPM =
where AFB(i)^ Absolute fractional bias weighted for each scientific category i,
AFB(1)= Absolute fractional bias for the operational one-hour averages.
The wind speed (u) £ 6 mph and neutral/stable category is weighted more than the other two
categories because of the importance of this category for regulatory modeling purposes. Thus, the
average of AFB(i) is
AVG(AFB(ljt) - 0.5 AFB(u * 6 mph, Neutral Stable) +
0.25 AFB(u * 6 mph, Unstable) + (8)
0.25 AFB(u > 6 mph, All stabilities
A combination of the GPM values across all three sites yields the composite model comparison
measure (CM). The CM results, shown in Figure 13, indicates that the best performing models are
CAL3QHC, TEXIN2, and CALINE4. Similarly, the AFB from scientific category 1 (u < 6 mph,
neutral/stable) can also be combined over all three sites into a single model comparison measure
(CM). This category is typically most important in terms of regulatory applications. As shown in
Figure 14, CAL3QHC has the lowest CM by a factor of two.
86
-------
.0
o.a
0 .6
o
D .
\
0.2
o. a
CAL3QHC IMM4 TEX IN2 GIM CALINE4
MODEL
Fi9Ure 13' CPMCstatS
comparison measure (CM) with 95% confidence limits using
87
-------
1 .D
o.a
Q.S
u
Q.2
Q.O
CAL3QHC IMM4 TEX I N2
MODEL
CAL I NE4
Figure 14. CM with 95% confidence limits using AFB of scientific category
88
-------
REFERENCES
MU, ;, Ponutant Leve,s
s!lm rT ?, ArtRrial — ' Ofllce °f TransP°rtation Laboratory. California DOT.
Sacramento, California. FHWA/CA/TL-79/23, Nov. 1979.
2. U.S. Environmental Protection Agency, Evaluation of en intersection
™ Specia, Report
Pl.nninn fl'nH
" EPA Office ^ Air. Nofee
o n dar°lina- EPA-450/4-78-001- OAQPS
.2~U28R, 1979. J
5. Newell G.F.. Applications of Queuing
nL...
ITE Journal Volume 58 No 3, March 1988.
7. Webster F.B. and Cobee B.M., Traffic Signal . R»ad Researnh T^hnical Paper N
Road Research Laboratory, 1966. r - ' -
ebster FB
and Signal Settings - Road
Paper No 3Q Road
Research Laboratory 1958.
MN p o MOR,LE5
Model), U.S. EPA, Ann Arbor, Michigan. 1992.
State Qf
,r Resources Board, Technical Support Division, Sacramento, CA, 1987.
' TransP°^ion and Traffic tnoin»^ u^^ 2nd
12. Wallace C.E. et al. TRANSYT 7FF Tmffic Network SHHU T.., Version 7R . , ,go,e
ty of Rorida Transportation Research Center & ^m. Highway Adminls^on
89
-------
\
13.
PE'. Development and RRVJRW of Traffic and CO Emission CQmnnnonts Qf intersection
.Modeling Techniques. U.S. EPA, Research Triangle Park, NC, 1988.
14. EMI Consultants, The Georgia Intersection Model for Air Quality Anaiyoie Knoxville, TN,
I wO*
15. Benson, P., CALINE4 - A Dispersion Model for Predicting Air Pollutant Concentrations Ne
Roadways. Report No. FHWA/CA/TL-84/14. Office of Transportation Laboratory
Sacramento, CA, 1989.
16. Bullln, G., J. Korpics, and M. Hlavlnka, User's Guide to the TEXlNig/MDBILE4 Mod»i
Research Report 283-2. Texas State Department of Highways and Public Transportation
College Station, TX, 1990.
17. NYDOT. Intersection Midblnrk Mode| User's Qnirip New York State Department of
Transportation, Albany, NY, 1982. *
18'
S0*',:!^' ** JA T^^ A Statistical Procedure for Determining the Best Performing Air
Quality Simulation Model. Atm^Biy., 24, 2387-2395, 1990.
19. Efron. B.. The Jackknife, the Rontstrap and Qthsr ResamPlinn Piana Society for Industrial
and Applied Mathematics, Philadelphia, PA, 1982.
20. Cleveland, W.S. and R. McGill, Graphical Perception: Theory, Experimentation, and
Application to the Development of Graphical Methods. J. Am. Stet. Ass™.. 79, 531-554,
^ * *
90
-------
ft. flEPOHT NO?
EPA.-454/R-92-OQ6
j*. TITLE AND SUBTITLE
£5^3 f*'3'2 to •CaL3QHC version 2.0-
£^DSL(^y f°r Pre£3icting Pollutant
Near Boadway Intersections
J7. AUTHOR(I)
OKGANIZATION NAME AND AOOR6SS-
j*»«daaaa assi EE*_ZL.
. RECIPIENTS ACCESSION NO,
• AND ADDRESS
y Planning and Standards
. Protection Agency
Sesearcn Triangle Park, NC 27711
j IS. SUPPLEMENTARY NOTES
Is. REPORT DATE
November 1992 _
6. PERFORMING ORGANIZATION CODE'
.PERFORMING ORGANIZATION REPORT NO.
10. PROGRAM ELEMENT No7
il.CONTHACI/GRANT N07
'3. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
68A
a tachn. of S
_
KEY WORDS AND DOCUMENT ANALYSIS
Carbon I-fonoxide (CD)
Intersection Modelina
CAL3QHC
Hot Spot Modeling
^
b.lDENTIFieRs/OPEN ENDEDTERMS
IB. DISTRIBUTION STATEMENT"
19.SECURIT* LLAi* (This Reporti
20. SECURH V CLASS (This page*