Traffic-related air quality assessment for open road tolling highway facility

Abstract

Open road tolling (ORT) design has been considered as an effective means of smoothing highway traffic and reducing travel delay on toll highways. In this paper it is demonstrated that ORT can also achieve significant air quality benefits over the conventional toll plaza design. The near roadside carbon monoxide (CO) concentration levels can be reduced by up to 37%, and diesel particulate matter (DPM) emissions can decrease by as much as 58%. These large expected air quality benefits have great implications to the regional efforts of reducing mobile source air pollution toward achieving attainment status and healthier living environment.

Full text

Journal of Environmental Management 88 (2008) 962–969
Traffic-related air quality assessment for open road
tolling highway facility
Jie Lin

, Dan Yu
Department of Civil and Materials Engineering, Institute of Environmental Science and Policy, University of Illinois at Chicago, Chicago, IL 60607, USA
Received 10 November 2006; received in revised form 12 March 2007; accepted 5 May 2007
Available online 21 June 2007
Abstract
Open road tolling (ORT) design has been considered as an effective means of smoothing highway traffic and reducing travel delay on
toll highways. In this paper it is demonstrated that ORT can also achieve significant air quality benefits over the conventional toll plaza
design. The near roadside carbon monoxide (CO) concentration levels can be reduced by up to 37%, and diesel particulate matter (DPM)
emissions can decrease by as much as 58%. These large expected air quality benefits have great implications to the regional efforts of
reducing mobile source air pollution toward achieving attainment status and healthier living environment.
r2007 Elsevier Ltd. All rights reserved.
Keywords: Open road tolling; Traffic queuing; Carbon monoxide; Diesel particulate matter; Toll payment type
1. Introduction
Open road tolling (ORT) design has been considered as
an effective means of smoothing highway traffic and
reducing travel delay on toll highways. The Florida
Expressway Authority’s toll plazas are scheduled for
conversion to ORT by 2007. The Illinois State Toll
Highway Authority (ISTHA) has begun a 5-year 5-billion-
dollar system-wide conversion plan of all of the conven-
tional toll plazas to ORT by 2010. By design, an ORT lane
is an electronic toll express lane, which allows vehicles
equipped with an electronic transponder to pay toll
without slowing down, while cash paying customers are
directed to manual tollbooths off the mainline and re-enter
the mainline to resume the rest of their journey.
1
ORT is
expected to reduce pollutant hotspots due to vehicle idling,
deceleration, and acceleration. Therefore, quantifying the
potential air quality benefits of ORT is of our particular
interest. These potential large air quality benefits have
great implications to the regional efforts of reducing
mobile source air pollution toward achieving attainment
status and healthier living environment. This is particularly
relevant to the Chicago metropolitan region where motor
vehicles account for 75% of total emissions in the region
(U.S. Environmental Protection Agency, 2006).
Recent toll plaza air quality studies focused primarily on
carbon monoxide (CO) emissions at toll plazas. Sisson
(1995) analyzed the potential air quality benefits of
replacing traditional toll collection with electronic toll
collection (ETC) and predicted relatively small reduction of
emissions that could still be economically significant.
Washington and Guensler (1994) presented a statistical
‘‘modal’’ model that simulates CO generated from electro-
nic toll traffic under different operating scenarios (idle,
acceleration, deceleration, and cruise). Coelho et al. (2005)
compared the estimated emissions with field measurement
at a toll plaza in Lisbon, Portugal and found that the 99%
of CO emissions were associated with the final acceleration
back to normal travel speed after the vehicle left the
tollbooth and that 61–84% CO reduction could result if
switching conventional tolls entirely to ETC.
ARTICLE IN PRESS
www.elsevier.com/locate/jenvman
0301-4797/$ - see front matter r2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jenvman.2007.05.005

Corresponding author. Tel.: +1 312 996 3068; fax: +1 312 996 2426.
E-mail address: janelin@uic.edu (J. Lin).
1
In the Illinois State Toll Highway System (ISTHS), ORT is
implemented through a transponder called I-PASS inside a vehicle and
a sensor to detect I-PASS on I-PASS express (IPX) lanes. Hence, ORT is
called IPX in the Illinois Tollway system. There are also I-PASS only
(IPO) lanes, which still require significant slow down of vehicle passing-
through speeds. Other payment types in ISTHS are manual and automatic
(i.e., exact coin) methods.

This study focuses on CO hotspots and diesel particulate
matter (DPM) emissions in various ORT implementation
scenarios on the Illinois toll highways. DPM is one of the
six priority mobile source air toxics (MSATs) identified by
the US Environmental Protection Agency (US Environ-
mental Protection Agency, 2000). DPM is emitted from
diesel engines primarily in the form of sulfate, elemental
carbon and organic carbon, which have been found to link
to health problems such as childhood asthma (e.g., Lin
et al., 2002;Zmirou et al., 2004) and infant prematurity
(e.g., Wilhelm and Ritz, 2003). DPM is of particular
interest also because of the large volume of daily truck
traffic on the Illinois toll highways. In 2003, the reported
commercial vehicle traffic on the Illinois toll highways was
between 38.9% and 50.9% (Illinois State Toll Highway
Authority, 2004). The study findings will have direct
implications to the region’s transportation and air quality
planning and public health.
The paper is organized in two parts. The first part
describes the modeling approach and validation results; the
second part presents the results of the ORT scenario
analysis. In the first part (Sections 2–4), roadside air
dispersion modeling is introduced followed by a modified
approach proposed for this study. The modified dispersion
model is then applied to predicting the roadside CO
concentrations (ppm) at a selected toll plaza and the results
are validated with the actual monitoring data. In the
second part (Sections 5 and 6), seven toll traffic scenarios
are established to assess their impacts on roadside CO
concentration levels and total DPM emissions. Finally,
implications to the regional transportation and air quality
planning practice are discussed.
2. Overview of roadside air dispersion modeling
Generally speaking, there are four types of roadside air
dispersion models: Gaussian plume dispersion model,
atmospheric box model, source apportionment, and
computational fluid dynamics (CFD). Gaussian plume
models assume Gaussian distribution of pollutant disper-
sion and no chemical or removal processes take place
during dispersion. Such models include the US EPA’s
CALINE family models (i.e., CALINE4, CAL3QHC, and
CAL3QHCR). CALINE4 primarily predicts pollutant
(CO, NO
2
, and aerosols) concentrations from uninter-
rupted highway traffic (California Department of Trans-
portation, 1989). CAL3QHC (and CAL3QHCR)’s core
dispersion algorithm is built on CALINE3 (the predecessor
of CALINE4) and predicts signalized intersections’ 1 and
8-h pollutant concentrations (US Environmental Protec-
tion Agency, 1995). Box models are based on mass
conservation inside a virtual three-dimensional box in the
atmosphere. An example of a box model is STREET-SRI
(Johnson et al., 1973). Source apportionment models, also
known as receptor models, are based on detailed analysis
of source contribution to the pollutant of interest collected
at a monitoring site. The constrained physical receptor
model (COPREM) is one such example (Wahlin, 2003).
CFD models, such as FLUENT and PHOENICS, are
numerical solutions of fluid flow and dispersion equations.
Other modeling techniques include Lagrangian models
(e.g., Thomson, 1987;Tinarelli et al., 1998;Oettl et al.,
2001) and empirical models (e.g., Chock, 1977;Dirks et al.,
2002, 2003).
Current dispersion models require simplified traffic-
related parameters such as average hourly traffic volume
and average speed. While this may be sufficient for
estimating emissions from smooth, un-interrupted traffic
flow, the calculation of emissions in interrupted traffic flow
requires more detailed traffic information to accurately
account for emissions due to traffic queuing. For example,
in CAL3QHC the intersection traffic queue length is
estimated before the idle emissions at the intersection can
be determined. The intersection queue length is calculated
by using the Transportation Research Board (TRB)’s
Highway Capacity Manual (HCM) methods shown in
Eqs. (1) and (2).
When traffic is unsaturated, i.e., volume-to-capacity
ratio below one:
Q¼max qDþr
2q;qr
hi
, (1)
where Qis the average queue per lane at the beginning of
green traffic light (vehicles/lane), qis the vehicle arrival rate
per lane (vehicles/lane/s), Dis the average vehicle approach
delay (s/vehicle), and ris the length of the red phase (s).
Average vehicle approach delay (D) measures the
smoothness of traffic progression at the signalized inter-
section. It is a function of a complex form that involves the
length of green time, signal cycle length, and the volume-
to-capacity ratio at the intersection (Transportation
Research Board (TRB), 1985).
When traffic is over-saturated, i.e., volume-to-capacity
ratio above one:
Q0¼max qDþr
2q;rq

þ1
2ðVCÞ, (2)
where Q
0
is the average queue per lane at the beginning of
the green phase in an over-saturated condition (vehicles),
q
*
is the vehicle arrival rate per lane at capacity (i.e. V/C¼
1.0) (vehicles/lane/s), D
*
is the average vehicle delay at
capacity (i.e. V/C¼1.0) (s/vehicle), and ris the length of
the red phase (s).
These models are not suitable for toll plaza air
quality analysis due to the noticeably different physical
configurations and traffic characteristics at toll plazas and
on uninterrupted highways or at intersections. The
CAL3QHC User’s Guide acknowledges ‘‘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, etc., Future research and
testing is necessary to adapt this program for such
situations’’ (US Environmental Protection Agency, 1995).
ARTICLE IN PRESS
J. Lin, D. Yu / Journal of Environmental Management 88 (2008) 962–969 963

Traffic before a tollbooth follows a sequence of driving
modes consisting of deceleration before the tollbooth, stop-
and-go at the tollbooth, and final acceleration after the
tollbooth to merge back to the main line traffic. Traffic
queues form before tollbooths when arrival traffic exceeds
the toll lane capacity. Toll plaza queues are much more
complex than intersection queues because a toll plaza
consists of various lane configurations, toll payment
methods (electronic, automatic, manual, etc.), and types
of customers (transponder users, trucks/buses, etc.).
Arrival traffic and service time are often stochastic in
nature. There is a body of research in developing empirical
traffic delay functions (Lin, 2001;Fambro and Rouphail,
1997), stochastic queuing models (Newell, 1982), and
microscopic traffic simulation models (Al-Deek et al.,
2000;Saka et al., 2000;Robinson and Van Aerde, 1995;
Zarrillo et al., 1997). While traffic simulation models can
provide detailed traffic parameters including individual
vehicle speeds and queue length they are often data
demanding and require extensive model calibration.
3. Modified approach
Roadside pollutant concentrations at a toll plaza can be
estimated with a CAL3QHC-type Gaussian plume disper-
sion model with modified traffic queuing estimation
algorithms.
2
Gaussian plume models assume no chemical
reactions of the pollutants during dispersion in the study
range typically within a few hundred meters. The core of
the dispersion module is the finite line source Gaussian
plume model, which has the following form:
dCðx;y;zÞ
¼mdy
2psyszuexp y2
2s2
y
!
exp ðzhÞ2
2s2
z

þexp ðzþhÞ2
2s2
z

,
ð3Þ
where mis pollutant mass per unit length of the road (g/m),
uis the wind speed (m/s), his the released plume height (m),
s
y
,s
z
are standard deviations in the crosswind (y) and
vertical (z) spread of the concentration distribution (m).
The coordinate system (x,y,z) is defined such that xfollows
the wind direction, yis perpendicular to the wind direction,
and zis the vertical height of the receptor (m).
Tollbooth traffic queue estimation employs stochastic
queuing models to estimate queue lengths at tollbooths
under both under- and over-saturated traffic conditions.
Average queue length is estimated by assuming Poisson
arrival, deterministic departure for unsaturated traffic and
normal arrival, deterministic departure for over-saturated
traffic, as proven empirically and theoretically (e.g., Edie,
1954;Kingman, 1962). Hence, the unsaturated average
queue length (Q
u
) and average waiting time in queue ( ¯
w)
are determined by the following equations:
Qu¼X2
2ð1XÞand ¯
w¼X
2mð1XÞ, (4)
where Xis volume-to-capacity ratio ¼V/C, and mis the
traffic throughput rate (i.e., number of vehicles passing
through the toll plaza in vehicles/s).
When traffic is over-saturated (i.e., X¼V/C41) arrival
traffic follows normal distribution. Assumed a constant
vehicle discharge rate, the expected queue length is
EðQÞ¼QuþtðVCÞ(5)
for the study period t(usually in an hour).
4. Model validation
Model validation was carried out through the following
case study at Plaza 33 at Irving Park, IL on North–South
Tri-State Tollway I-294 because of the availability of
air monitoring data nearby. Plaza 33 is a southbound
mainline plaza, consisting of 18 toll lanes from shoulder to
center including seven manual lanes, seven automatic
(i.e., exact coin) lanes, and four passenger car I-PASS
Only (PC IPO)lanes (Fig. 1). The total length of the
expanded portion at the toll plaza is about 518 m, of which
274 m is upstream and 244 m downstream. The opposite
traffic has four through lanes with speed limit of 88.5 km/h,
which is considered as free flow travel speed.
Based on the actual transaction counts recorded at Plaza
33 during January 2005 (Fig. 2) 2 weekdays were chosen to
represent light and heavy peak-hour traffic conditions:
(1) Wednesday, January 5, 2005, 9–10 a.m., representing a
light traffic condition, referred to as Case 1, and (2)
Monday, January 31, 2005, 5–6 p.m., representing a heavy
traffic condition, referred to as Case 2.
A receptor was set up at the Illinois Environmental
Protection Agency’s (IEPA) Schiller Park air monitoring
station so that the estimated pollutant concentrations
could be compared with the observed at the station.
ARTICLE IN PRESS
Fig. 1. Aerial photo of Plaza 33 on I-294 toll highway.
2
A detailed description of the modified dispersion model can be found in
Lin and Yu (2006). A version of the model (called Toll Plaza CO Screening
Tool or TPCOST) is currently used by the Illinois State Toll Highway
Authority (ISTHA) for CO screening analysis.
J. Lin, D. Yu / Journal of Environmental Management 88 (2008) 962–969964

The monitoring station is 226 m northwest of Plaza 33
centerline. It records hourly local CO and NO
x
concentra-
tion levels (in ppm). Fig. 3 displays the average hourly CO
concentrations (ppm) observed at the station and traffic
(vehicles/h) at Plaza 33 during January 2005. Generally
speaking, the CO concentrations followed the bimodal
traffic pattern, suggesting the influence from nearby traffic.
It is worth noting that the traffic volumes shown in Fig. 3
included only the southbound traffic (tolled traffic).
Because no PM concentrations were recorded at this
station, the modified dispersion model was only validated
for CO. Background CO concentration was set at 0.2 ppm
based on the readings from a background monitoring
station 16.1 km southeast of Plaza 33 in Cicero, IL.
CO running exhaust and start emission factors were
generated from US EPA’s MOBILE6.2. Many of the input
parameters to MOBILE6.2 were local specific values
provided by the IEPA. These parameters were fuel type,
oxygenated fuel parameters, I/Mprograms, vehicle miles
traveled (VMT) by vehicle type, facility type and speed,
ARTICLE IN PRESS
January 2005
31302928272625242322212019181716151413121110987654321
Average traffic count (vehicles per hour)
4500
4000
3500
3000
2500
2000
1500
Mon
Sat
Fig. 2. Average hourly traffic at Plaza 33 during January 2005.
0
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
hour
CO concentration (ppm)
0
1000
2000
3000
4000
5000
6000
7000
8000
traffic (vehicles/hr)
CO
toll traffic
Fig. 3. Average hourly CO concentrations (ppm) at Schiller Park and traffic at Plaza 33 (southbound only) during January 2005.
J. Lin, D. Yu / Journal of Environmental Management 88 (2008) 962–969 965

vehicle age distribution, etc. All other inputs (e.g., fuel
economy, technological improvement trend) used the
MOBILE6.2 defaults.
Meteorological parameters such as wind speed and
direction, temperature, and cloud cover were obtained
from the O’Hare climatic station within 8 km from Plaza 33
and Schiller Park air monitoring station. The values of
these parameters were averaged over a 3-h interval. It was
assumed that constant meteorological conditions during
the 3-h interval. It was also assumed that the observed
conditions at the O’Hare meteorological station represent
the local, surface meteorological conditions at the toll
plaza. Air stability class (A to F) was determined based on
wind speed and solar strength/cloud cover (Turner, 1994).
The surrounding land use at Plaza 33 was primarily
residential, thus surface roughness was set at 108, adapted
from CALINE3 definition of surface roughness (Benson,
1979). Table 1 summarizes the observed traffic and
meteorological conditions of the two cases.
Table 2 summarizes the estimated traffic queue lengths
and CO levels at the receptor. The traffic volume-to-
capacity ratios (r¼V/C) in Case 2 were much higher than
those in Case 1, causing longer average traffic queues
before the tollbooths. The estimated CO concentrations at
the receptor were 0.201 ppm compared to the observed
0.3 ppm in Case 1, and 0.854 ppm compared to the
observed 0.9 ppm in Case 2. The roadside CO levels did
not exceed the 8-h or 1-h national ambient air quality
standards in either case.
Overall the model predictions are reasonably close to the
observed values. It is naturally expected that the estimated
are less than the observed because there are other sources
contributing to the local pollution level at the receptor site
such as industrial combustion, residential activity, and even
local traffic, all of which are excluded from the model. The
slightly larger differences between the estimated and
observed levels in the light traffic condition (i.e., Case 1)
can be explained by the dispersion behavior of the
pollutants. As shown in Fig. 4, the effect of traffic on
roadside CO levels quickly becomes negligible as the
distance from the source increases beyond 170 m in the
light traffic condition. The relatively strong wind also adds
to fast decline in pollution levels away from the mainline
ARTICLE IN PRESS
Table 1
Input parameters at Toll Plaza 33 on Tri-State Tollway in January 2005
Case Traffic flow (vehicles/h) Service rate
(s/vehicles)
# Lanes Wind direction (1) Wind speed
(m/s)
Stability
class
(1) 9–10 a.m., January 5,
2005
Manual 713 8 7 50 6.7 D
Auto 571 5.5 7
PC IPO 2465 2 4
(2) 5–6 p.m., January 31,
2005
Manual 1990 8 7 80 2.7 D
Auto 1846 5.5 7
PC IPO 5152 2 4
Table 2
Model estimation and comparison with field observations
Case r¼V/CAverage queue
length (vehicles)
CO (ppm)
Observed Estimated
(1) 9–10 a.m., January 5, 2005 Manual 0.2 0 0.300 0.201
Auto 0.13 0
IPO 0.46 0.2
(2) 5–6 p.m., January 31, 2005 Manual 0.55 0.3 0.900 0.854
Auto 0.4 0.1
IPO 0.95 10.0
0
0.5
1
1.5
2
2.5
0 50 500 550400 450300 350200 250100 150
Distance from toll road centerline (meters)
CO concentration (ppm)
5-Jan
31-Jan
Fig. 4. Dispersion of CO (model estimation) generated by traffic at
Plaza 33.
J. Lin, D. Yu / Journal of Environmental Management 88 (2008) 962–969966

traffic. In Case 2, the impact of heavy traffic appears to be
more noticeable with the presence of milder wind.
However, the impact becomes trivial after 350 m, which is
the range a Gaussian plume dispersion model can generally
apply.
5. Impacts of ORT on roadside CO concentrations and
DPM emissions
The air quality assessment of ORT includes CO hotspot
analysis and DPM emission estimation of seven scenarios
of toll traffic composition by payment type, summarized in
Table 3. The payment types considered are manual,
automatic, IPO and IPX (ORT). The total traffic volume
is fixed at 1000 vehicles/h for all scenarios.
The maximum possible hourly CO pollution level per
1000 vehicles at the current toll highway traffic split (15/45/
10/30) is estimated to be 1.475 ppm, of which roughly 10%
each is contributed by IPO and IPX traffic, 30% by manual
toll traffic, and about 50% by exact coin (automatic)
traffic. The maximum possible hourly CO levels of the
other six toll traffic scenarios are compared to the current
situation in Fig. 5. The vertical axis is the CO ratio to the
current level, expressed in terms of percentage; the
horizontal axis is the seven scenarios of toll payment split.
There are noticeable reductions in CO due to the increase
of I-PASS shares including both IPX (ORT) and IPO.
The CO level would have been 22% higher from the
current situation if there had been no I-PASS in place. It is
predicted a 36.7% reduction from the current condition
when the Tollway’s goal of 90% I-PASS users is achieved.
Quantifying the impact on DPM is less straightforward
for two reasons. Firstly, the above dispersion model is not
appropriate for DPM prediction because the model is not
validated for DPM. Secondly, MOBILE6.2’s estimated
DPM emission factors are highly insensitive to traffic
speeds (Fig. 6). Similar patterns are also found in gasoline
PM. This may attribute to the non-modal based, composite
estimation approach in MOBILE6.2 and relatively few on-
road diesel emissions data available (Lloyd and Cackette,
2001). MOBILE6.2 is known to be less accurate at smaller
ARTICLE IN PRESS
Table 3
Seven scenarios of toll traffic composition
Scenario Manual (%) Automatic (exact coin) (%) IPO (%) IPX (ORT) (%)
(1) 45/55/0/0 45 55 0 0
(2) 35/65/0/0 35 65 0 0
(3) 25/65/10/0 25 65 10 0
(4) 15/65/10/10 15 65 10 10
(5) 15/45/10/30—current condition 15 45 10 30
(6) 15/25/20/40 15 25 20 40
(7) 5/5/10/80—future goal by 2010 5 5 10 80
92.7%
100%
114.6%
122.0%
122.0%
122.0%
63.3%
0%
20%
40%
60%
80%
100%
120%
140%
45/55/0/0
35/65/0/0
25/65/10/0
15/65/10/10
15/45/10/30
15/25/20/40
5/5/10/80
Payment method Split (%)
Percent CO level to current
Auto
IPO
IPX
Manual
scenario (%)
Fig. 5. Percent CO level to current scenario versus percent traffic by payment type.
0.1
0.12
0.14
0.16
0.18
0.2
08871543720 105 122
Speed (km hr-1)
DPM EF (g km-1)
Fig. 6. DPM emission factors by speed.
J. Lin, D. Yu / Journal of Environmental Management 88 (2008) 962–969 967

scales than on a regional level. On the other hand, a
number of studies have demonstrated the changes of PM
emissions, in terms of mass, number of particles, and size
distribution, over speeds (Jones and Harrison, 2006;
Kittelson et al., 2004;Shah et al., 2004).
Hence, DPM mass emissions are quantified instead. The
modal-dependent DPM emission rates use the measured
values of a fleet consisting of 11 diesel vehicles by Shah
et al. (2004). In their study, Shah et al. quantified the DPM
emission rates by driving mode. Four modes were
considered: cold start/idle—a cold start followed by a
10-min idle, creep—stop-and-go driving, transient—light
to medium arterial driving with an average travel speed of
15.4 mph, and cruise—highway driving with and average
travel speed of 39.9 mph (Shah et al., 2004). Considering all
possible driving modes at a toll plaza, we simplified the
movements by assuming a creep phase before the manual
and automatic coin tollbooth and cruise for all I-PASS
traffic. Shah et al. (2004) found that the fleet average DPM
emission rates associated with these two driving modes
were 1016 and 215 mg mi
1
, respectively.
In the current situation, the total DPM emission per
1000 vehicles is 695.6 mg, of which 22% is due to manual
toll traffic, 66% due to exact coin traffic, 3% by IPO
traffic, and 10% by IPX traffic. For the other six scenarios,
the relative DPM emissions to the current situation (in
percentage) are shown in Fig. 7. The total DPM emission
would have been 46% more than the current level if no
I-PASS had been in place. With 90% I-PASS traffic the
DPM emissions are expected to be reduced by over 57%
from the current level, which represents large potential
environmental and public health benefits to be achieved by
the Illinois toll highway system after the conversion.
6. Conclusions
ORT design has been considered as an effective means of
smoothing highway traffic and reducing travel delay on toll
highways. In this paper, it has been further demonstrated
that ORT can also achieve significant air quality benefits
over the conventional toll plaza design. The near road-
side CO and DPM pollution can be reduced by as much
as 37% of CO concentration levels to 58% of DPM
emission if 50% of the conventional toll (manual and
automatic) traffic switches to ORT. These potential large
air quality benefits have great implications to the regional
efforts of reducing mobile source air pollution toward
achieving attainment status and healthier living environ-
ment. Effective traffic smoothing and regulation methods
especially those involving intelligent transportation tech-
nologies will likely result in large reductions in vehicular
emissions from highway traffic as demonstrated in this
study.
It is worth noting the limitations of the study. This study
could be strengthened by having roadside measurements of
pollutants in addition to the data from the nearby air
monitoring station. The monitoring station data, although
useful, are of somewhat limited value because the station is
more than 200 m away from the mainline traffic. Traffic
impact on pollution levels is diminished quickly as the
distance increases. Actual field measurement of DPM will
be a valuable source of DPM data for understanding on-
road DPM emissions characteristics and model validation.
Finally, the finer temporal resolution of the meteorological
parameters is needed to be consistent with the model
resolution of 1 hour.
Acknowledgments
We thank the Illinois State Toll Highway Authority
for providing the necessary data for this study. We also
thank the Illinois Environmental Protection Agency for
providing the local input parameters to MOBILE6.2.
Finally, we thank the anonymous reviewers for their
comments.
ARTICLE IN PRESS
146.1% 146.1%
134.5%
123.0%
100%
77.0%
42.4%
0%
30%
60%
90%
120%
150%
percent DPM emissions to
current scenario
IPX
IPO
Auto
Manual
45/55/0/0
35/65/0/0
25/65/10/0
15/65/10/10
15/45/10/30
15/25/20/40
5/5/10/80
Payment method Split (%)
Fig. 7. Percent DPM emissions to base scenario for various payment method splits.
J. Lin, D. Yu / Journal of Environmental Management 88 (2008) 962–969968

References
Al-Deek, H.M., Mohamed, A.A., Radwan, E.A., 2000. New model for
evaluation of traffic operations at electronic toll collection plazas.
Transportation Research Record 1710, National Research Council,
Transportation Research Board, pp. 1–10.
Benson, P.E., 1979. CALINE3—a versatile dispersion model for predict-
ing air pollutant levels near highways and arterial streets. Report No.
FHWA/CA/TL-79/23, November.
California Department of Transportation, 1989. CALINE4—a dispersion
model for predicting air pollutant concentrations near roadways. Final
Report prepared by the Caltrans Division of New Technology and
Research, Report No. FHWA/CA/TL-84/15.
Chock, D.P., 1977. A simple line-source model for dispersion near
roadways. Atmospheric Environment 12, 823–829.
Coelho, M.C., Farias, T.L., Rouphail, N.M., 2005. Measuring and
modeling emission effects for toll facilities. Transportation Research
Record 1941, National Research Council, Transportation Research
Board, pp. 136–144.
Dirks, K.N., Johns, M.D., Hay, J.E., Sturman, A.P., 2002. A simple semi-
empirical model for predicting missing carbon monoxide concentra-
tions. Atmospheric Environment 36, 5953–5959.
Dirks, K.N., Johns, M.D., Hay, J.E., Sturman, A.P., 2003. A semi-
empirical model for predicting the effect of changes in traffic flow
patterns on carbon monoxide concentrations. Atmospheric Environ-
ment 37, 2719–2724.
Edie, L.C., 1954. Traffic delays at toll booths. Journal of the Operation
Research Society of America 2 (2), 107–138.
Fambro, D.B., Rouphail, N.M., 1997. Generalized delay model for
signalized intersections and arterial streets. Transportation Research
Record 1572, National Research Council, Transportation Research
Board, Washington, DC, pp. 112–121.
Illinois State Toll Highway Authority, 2004. 2003 Traffic data report for
the Illinois tollway system, prepared by Wilbur Smith Associates, May
2004.
Johnson, W.B., Ludwig, F.L., Dabberdt, W.F., Allen, R.J., 1973. An
urban diffusion simulation model for carbon monoxide. Journal of Air
Pollution Control Association 23, 490–498.
Jones, A.M., Harrison, R.M., 2006. Estimation of the emission factors of
particle number and mass fractions from traffic at a site where mean
vehicle speeds vary over short distances. Atmospheric Environment 40,
7125–7137.
Kingman, J.F.C., 1962. On queues in heavy traffic. Journal of the Royal
Statistical Society, Series B (Methodological) 24 (2), 383–392.
Kittelson, D.B., Watts, W.F., Johnson, J.P., 2004. Nanoparticle emissions
on Minnesota highways. Atmospheric Environment 38 (1), 9–19.
Lin, F.B., 2001. A delay model for planning analysis of main-line
toll plazas. Transportation Research Record 1776. National
Research Council, Transportation Research Board, Washington,
DC, pp. 69–74.
Lin, J., Yu, D., 2006. CO hotspot modeling at highway toll facility. In:
Proceedings of the Air and Waste Management Association’s 99th
Annual Conference and Exhibition, New Orleans, 20–23 June 2006.
Lin, S., Munsie, J.P., Hwang, S.-A., Fitzgerald, E., Cayo, M.R., 2002.
Childhood asthma hospitalization and residential exposure to state
route traffic. Environmental Research, Section A 88 (2), 73–81.
Lloyd, A.C., Cackette, T.A., 2001. Diesel engines: environmental impact
and control. Journal of Air and Waste Management 51, 809–847.
Newell, G.F., 1982. Applications of queueing theory. 2nd ed. Chapman
and Hall Ltd, London, UK, ISBN:0412245000.
Oettl, D., Almbauer, R.A., Sturm, P.J., 2001. A new method to estimate
diffusion in stable, low-wind conditions. Journal of Applied Meteor-
ology 40 (2), 259–268.
Robinson, M., Van Aerde, M., 1995. Examining the delay and
environmental impacts of toll plaza. In: Proceedings of the 1995
Vehicle Navigation and Information Systems Conference—Sixth
International VNIS, Washington State Convention and Trade Center,
Seattle, WA, USA, 30 July–2 August.
Saka, A.A., Agboh, D.K., Ndiritu, S., Glassco, R.A., 2000. An estimation
of mobile emissions reduction from using electronic toll collection in
the Baltimore metropolitan area: a case study of the Fort McHenry
tunnel toll plaza. National Transportation Center, Morgan State
University.
Shah, S.D., Cocker, D.R., Miller, J.W., Norbeck, J.M., 2004. Emission rates
of particulate matter and elemental and organic carbon from in-use
diesel engines. Environmental Science and Technology 38 (9), 2544–2550.
Sisson, M., 1995. Air quality benefits of electronic toll collection.
Transportation Quarterly 49 (4), 93–101.
Thomson, D.J., 1987. Criteria for the selection of stochastic models of particle
trajectories in turbulent flows. Journal of Fluid Mechanics 180, 529–556.
Tinarelli, G., Anfossi, D., Bider, M., Ferrero, E., Trini Castelli, S., 1998.
A new high performance version of the Lagrangian particle dispersion
model SPRAY, some case studies. Preprints of the 23rd CCMS-NATO
Meeting, Varna, DEMETRA Publication, 28 September–2 October,
pp. 343–350.
Transportation Research Board (TRB), 1985. The Highway Capacity
Manual. TRB Special Report 209, Washington, DC.
Turner, D.B., 1994. Workbook of Atmospheric Dispersion Estimates: An
Introduction to Dispersion Modeling, second ed. Lewis Publishers,
Boca Raton, FL.
U.S. Environmental Protection Agency, 1995. User’s guide to CAL3QHC
Version 2.0: A Modeling Methodology for Predicting Pollutant
Concentrations Near Roadway Intersections. Report No. EPA-454/
R-92-006. Office of Air Quality Planning and Standards, Research
Triangle Park, NC.
U.S. Environmental Protection Agency, 2000. Technical support docu-
ment: control of emissions of hazardous air pollutants from motor
vehicles and motor vehicle fuels. Report No. EPA-420-R-00-023.
Assessment and Standards Division, Office of Transportation and Air
Quality, December.
U.S. Environmental Protection Agency, 2006. National Emission In-
ventory (NEI) database. /http://www.epa.gov/air/data/geosel.htmlS
(last accessed November 2006).
Wahlin, P., 2003. COPREM—a multivariate receptor model with a
physical approach. Atmospheric Environment 37 (35), 4861–4867.
Washington, S., Guensler, R., 1994. Carbon monoxide impacts of
automatic vehicle tolling operations. In: Proceedings of the National
Policy Conference on ITS and the Environment, Hubert Humphery
Institute of Public Affairs, University of Minnesota, 6–7 June.
Wilhelm, M., Ritz, B., 2003. Residential proximity to traffic and adverse
birth outcomes in Los Angeles county, California, 1994–1996.
Environmental Health Perspectives 111 (2), 207–216.
Zarrillo, M.L., Radwan, A.E., Al-Deek, H.M., 1997. Modeling traffic
operations at electronic toll collection and traffic management systems.
Computers and Industry Engineering 33 (3–4), 857–860.
Zmirou, D., Gauvin, S., Pin, I., Momas, I., Sahraoui, F., Just, J., Le
Moullec, Y., Bremont, F., Cassadou, S., Reungoat, P., Albertini, M.,
Lauvergne, N., Chiron, M., Labbe, A., 2004. Traffic related air pollution
and incidence of childhood asthma: results of the Vesta case-control
study. Journal of Epidemiology and Community Health 58 (1), 18–23.
ARTICLE IN PRESS
J. Lin, D. Yu / Journal of Environmental Management 88 (2008) 962–969 969