Personal exposure to fine particulate air pollution while commuting: An examination of six transport modes on an urban arterial roadway
Personal exposure to fine particulate air pollution while commuting: An examination of six transport modes on an urban arterial roadway
Robert A. Chaney 0 1 2
Chantel D. Sloan 0 1 2
Victoria C. Cooper 0 1 2
Daniel R. Robinson 0 1 2
Nathan R. Hendrickson 0 1 2
Tyler A. McCord 0 1 2
James D. Johnston 0 1 2
0 Brigham Young University, Department of Health Science , Provo, Utah , United States of America
1 Funding: This research was funded by the Ira and Mary Lou Fulton Gift Fund at Brigham Young University
2 Editor: Roger A Coulombe, Utah State University , UNITED STATES
Traffic-related air pollution in urban areas contributes significantly to commuters' daily PM2.5 exposures, but varies widely depending on mode of commuting. To date, studies show conflicting results for PM2.5 exposures based on mode of commuting, and few studies compare multiple modes of transportation simultaneously along a common route, making inter-modal comparisons difficult. In this study, we examined breathing zone PM2.5 exposures for six different modes of commuting (bicycle, walking, driving with windows open and closed, bus, and light-rail train) simultaneously on a single 2.7 km (1.68 mile) arterial urban route in Salt Lake City, Utah (USA) during peak ªrush hourº times. Using previously published minute ventilation rates, we estimated the inhaled dose and exposure rate for each mode of commuting. Mean PM2.5 concentrations ranged from 5.20 μg/m3 for driving with windows closed to 15.21 μg/m3 for driving with windows open. The estimated inhaled doses over the 2.7 km route were 6.83 μg for walking, 2.78 μg for cycling, 1.28 μg for light-rail train, 1.24 μg for driving with windows open, 1.23 μg for bus, and 0.32 μg for driving with windows closed. Similarly, the exposure rates were highest for cycling (18.0 μg/hr) and walking (16.8 μg/hr), and lowest for driving with windows closed (3.7 μg/hr). Our findings support previous studies showing that active commuters receive a greater PM2.5 dose and have higher rates of exposure than commuters using automobiles or public transportation. Our findings also support previous studies showing that driving with windows closed is protective against trafficrelated PM2.5 exposure.
Exposure to fine particulate matter (PM2.5) in ambient air is associated with multiple adverse
health outcomes in adults and children [
]. In urban areas, motor vehicle exhaust contributes
significantly to PM2.5 air pollution [
]. There is extensive spatial and temporal variation in
intra-urban air pollution exposures , with PM2.5 often occurring at higher concentrations
near major roadways due to gasoline and diesel engine exhaust [
]. Thus, for commuters,
traffic-related air pollution can significantly contribute to daily PM2.5 exposures, especially on
arterial roads during high-traffic ªrush hourº time windows [
]. American commuters spend
approximately 26 minutes commuting one direction to work each day [
], making long-term
and acute health effects from exposure a concern [
One strategy to reduce daily PM2.5 exposure is to encourage a shift in modal transportation
toward increased use of public transit and active commuting (walking and cycling). However,
to date there are relatively few studies on PM2.5 exposures by type of commuting in the United
States. Some studies show higher PM2.5 exposures for cyclists [
] and walkers [
compared with automobile commuters, whereas others report no difference  or lower
exposures for active commuters compared to those using automobiles or public transit [
Complicating the issue, walking and cycling take longer and increase individuals' breathing
rates, potentially exposing active commuters to higher inhaled doses of PM2.5 compared to
automobile or public transit [
10, 12, 16, 17
]. There is significant variability in automobile
cabin exposure depending on window position (open or closed) and whether the car's internal
air circulation is turned off, which allows outdoor air to be entrained into the vehicle [
Thus, additional research is needed to better characterize PM2.5 exposure by transportation
Understanding inter-modal differences in commuters' PM2.5 exposures is difficult for a
number of reasons. Zuurbier et al. (2010) note that in most previous studies, particulate
exposures were not measured simultaneously by mode of commuting [
]. This is problematic due
to spatial and temporal fluctuations in background air pollution levels (e.g. due to a geographic
depression, or commuting-time vs. night-time levels) [
]. Some studies have tried to
adjust for this by making comparisons with data from fixed-site monitors (FSM) [
10, 14, 21,
]. However, there can be considerable variability in PM2.5 measures based on differences in
sampling equipment [
], making direct comparisons between FSMs and personal exposure
monitors difficult [
]. There is also considerable variability associated with commuting
microenvironments. This variability includes active commuters' flexibility in choosing a wider
variety of routes. To control for this variation, previous studies have used prescribed
commuting routes. However, many of these routes had limited comparison groups (e.g. walking vs.
], or only looked at one particular mode of transportation (e.g. bicycles) [
]. Another challenge has been that all transport modes evaluated in some studies were
unable to use the exact same route (i.e. subway vs. car) [
In this study, we examined personal exposures to PM2.5 for six different commuting types
(bicycle, walking, driving with windows open and closed, bus, and light-rail train) on a single
prescribed arterial urban route during peak ªrush hourº times. To control for variation in
background PM2.5 we used identical instruments for both personal and ambient monitoring.
Finally, we estimated the PM2.5 inhaled dose and rate of exposure for each commuting mode
based on previously published minute ventilation rates [
16, 25, 26
Materials and methods
Salt Lake City (SLC) is the capital and largest municipality in Utah, with a population of
192,672 according to the 2015 United States census [
]. Population density in SLC is 657.5
people/km2 (1,703 people/mi2) [
]. Air pollution samples were collected on a 2.7 km (1.68
mi) section of road located between 600 W and 1950 W, N. Temple in Salt Lake City, Utah.
We chose this location because it serves as a main transportation corridor leading into the
downtown SLC area. This route provides access to multiple forms of transportation, including
2 / 15
Fig 1. Designated commuting route for all forms of transportation and location of the stationary monitor.
two traffic lanes each for eastbound and westbound automobiles, bicycle lanes and sidewalks
on both sides of the roadway, diesel bus routes, and an electric light rail system located
between eastbound and westbound lanes. This route is located on the west side of SLC, where
the terrain is flat. This roadway is zoned for commercial business, and includes restaurants,
shops, office buildings, banks, and grocery stores. Several blocks on the north side of the road
border on the Utah State fairgrounds. Residential areas are located within one block of the
roadway at some locations (Fig 1).
We measured breathing zone PM2.5 concentrations on 15 study personnel using TSI SidePak
AM510 personal exposure monitors (TSI, Inc., Shoreview, MN, USA) as they travelled the 2.7
km (1.68 mi) route via walking, biking, riding the bus, riding the light rail system, driving with
windows open, and driving with windows closed. Study personnel were university faculty
(n = 3) and students (n = 12), all of which were in good physical health. For both automobiles,
the internal circulation system was turned on during all study periods. All study personnel
were non-smokers. Data collection occurred from Aug 8th±Aug 11th, 2016. Air pollution
monitoring was conducted from approximately 8:00±10:00 AM, and again from 4:00±6:00 PM
during peak commuting times. Samples were collected simultaneously for all modes of
transportation beginning at the same time. At times, study workers were shifted to less efficient
modes in order to provide enough overall sampling time per mode. However, individual data
collection events occurred more expeditiously for some modes of transportation (e.g. light rail
and driving) than for others. For the purposes of this study, a data collection event was defined
as the study worker initiating data logging on the SidePak, travelling the length of the route in
3 / 15
one direction using their designated mode of transportation, and then discontinuing data
logging. Once the data collection event was completed, the study worker crossed to the other side
of the road and repeated this process. For study workers riding the light rail, SidePak data
logging was initiated approximately 5 minutes before arrival of the train, and was stopped upon
exiting the train. The worker then reset the SidePak and rode the next available train in the
opposite direction. There were no busses that ran the entire length of the study roadway, so we
were obliged to use a route that covered approximately 44% (1.19 km) of the entire study route
on the north side of the road only (Fig 1).
SidePak AM510s (n = 9) were set to record PM2.5 measures every 1-second during
sampling. Each night prior to the next day's sampling, we reset the instruments' internal clocks to
the lab computer date/time, cleaned and greased the impaction plates, recorded post-sampling
air pump flow rates (post-calibration), calibrated pump flow rates to 1.7 L/min (pre-calibration
for the next day) using a Defender 510 volumetric flow calibrator (Mesa Labs, Butler, NJ,
USA), zeroed the 670 nm laser using an in-line HEPA filter, and charged the batteries. All
preand post-calibrations were within ±5% of the 1.7 L/min target flow rate. All of the SidePak
AM510 instruments were calibrated by the manufacturer to ISO 12103±1 A1 test dust
(Arizona road dust) prior to using them in this study.
The inlet to the SidePak AM510 was attached to the collar or shoulder of the study worker's
shirt so it was within the person's breathing zone. For the purposes of this study, we defined
the breathing zone as the sphere of air located within 25 cm (10 in) of the individual's nostrils
]. To control for differences in particle release by type of clothing, all study personnel who
wore a personal monitor wore a standardized high visibility, 100% polyester shirt over the top
of their regular clothing. Previous studies have used vests made from nylon or 50:50 cotton:
polyester blends [
]. There is little data on fabric particle release, so we chose 100%
polyester based on the assumption that synthetic fabrics emit less dander than fabrics made with
natural fibers, such as cotton.
Ambient PM2.5 was measured with one stationary SidePak AM510 instrument placed on the
south side of the road (Fig 1). The inlet for the SidePak stationary monitor was located
approximately 0.9 m (3 ft.) off the ground on a stand, and placed approximately half way between the
edge of the road and the sidewalk. Stationary monitoring was performed whenever personal
breathing zone samples were being collected. The stationary monitor was calibrated and
maintained as described for the personal monitors.
Two university-owned Toyota Sienna mini-vans, 2013 and 2016 models, were used in this
study for monitoring breathing zone exposures for driving with windows open and driving
with windows closed. For driving with windows closed, van air conditioning was used with the
air recirculating in the cabin (not drawing outside air into the vehicle). Both vehicles used the
same model of cabin air filter (Toyota part number 87139-YZZ20), which had an atmospheric
dust efficiency rating of 0.3 ~ 0.5 μm: 20±10% at 240m3/h. The same cabin air filter was used
to remove particulate matter coming into the vehicle from outside and when recirculating
air inside the vehicle. University maintenance records showed the air filters had not been
replaced, and thus were original to the van. To verify that differences in personal monitoring
between the two vans were not due to air pollution produced by either van leaking into the
cabin, we conducted personal monitoring simultaneously in both vans while driving the length
of the route with windows closed. We first evaluated whether the 1-second measurements
4 / 15
were autocorrelated using the Durbin Watson test. We found significant autocorrelation
(p = 0.023), which was expected due to the short, 1-second time interval used on the SidePaks.
Thus, to compare means between the two vans, we performed a t-test taking into account serial
correlation in the data. Our results showed a statistically significant difference between the
two vans (p = 0.029). Means and standard errors were 5.2 μg/m3 (SE = 0.17) and 4.9 μg/m3
(SE = 0.13) for the 2016 and 2013 models, respectively. Although the difference between the
two vans was statistically significant, we believe the 0.3 μg/m3 difference was not necessarily
clinically significant from an overall exposure standpoint. JMP version 12.0 (SAS Institute
Inc., Cary, NC, USA) was used for this analysis.
Relative humidity (RH) greater than 50% can cause hygroscopic enlargement of particles
less than 2.5 μm, resulting in the SidePak overestimating the concentration of PM2.5 [
Ambient hourly RH measurements over the study period were obtained from the Hawthorne
monitoring station operated by the Utah Department of Environmental Quality, Division of
Air Quality (DAQ). The Hawthorne monitoring station, the closest DAQ monitor, is located
6.79 km from the study route. For the morning and afternoon commutes, the average of the
DAQ hourly means was 30.1% (range = 15.9±48.1%) and 18.6% (range = 9.0±38.7%),
respectively. Based on the low ambient RH we expected inappreciable hygroscopic particle growth,
and thus made no corrections to the SidePak measurements.
We cleaned the data by matching each data point of commuter (personal monitoring) with the
particulate levels measured by the stationary monitor by date and second. The detection limit
of the SidePak recorded by TSI is 0.001mg/m3 with a zero stability of ±0.001 mg/m3 over 24
hours. In order to account for this known limit, readings below 0.001 mg/m3 were recoded as
0.0007 mg/m3, which is the detection limit divided by the square root of two. This allowed us
to use readings where the SidePak read 0 mg/m3 due to the monitor's inability to detect
particle counts that low (the particle count is never truly zero), and proceed with analysis at those
time points. We then calculated the ratio of each commuting type to the background levels
recorded at each second. The ratios were averaged and reported along with standard errors for
overall particulate exposures by commuting type and by commuting type and time of day
(morning or afternoon rush hour). Statistical differences between groups were calculated
using linear regression within R: A Language and Environment for Statistical Computing (v
To calculate the inhaled dose of PM2.5 per type of commuting, we used the averages of
minute ventilation rates based off of different studies. Car transportation, with both windows
opened and closed, had a minute ventilation rate of 11.8 L/min [
]. Light rail and bus
transportation were considered similar enough to use the same minute ventilation rate: 12.7 L/min
]. Cycling and walking had higher minute ventilation rates due to the physical nature of
this commuting method. Walking had the second highest minute ventilation rate of 22.8 L/
]. Cycling had the highest at 23.5 L/min [
]. The inhaled dose (μg PM2.5 per trip) was
then calculated by multiplying the mean PM2.5 concentration for each commuting type by the
minute ventilation rate (L/min) and then by multiplying by the conversion factor 1m3/1000 L
and by the average number of minutes per trip [
]. Exposure rates were calculated by dividing
the inhaled dose (μg) by the average trip time (hrs).
We recorded real-time PM2.5 concentrations for six different commuting types and one
stationary monitor in downtown Salt Lake City, Utah. We collected between 4524 (bus) and
5 / 15
39,725 (walking) data points for each type of commuting. The disparity in the numbers of
points is due to inherent differences in commuting. The bus comes infrequently, whereas
walkers were able to record values continuously. It should also be noted that due to a technical
malfunction, data for driving with windows closed was only recorded on 8/8/16 and 8/11/16.
While this malfunction was unfortunate, the particulate matter for driving windows closed
was very consistent, with a standard deviation in the raw data of only 4.6 ug/m3. Overall
standard deviations for the other modes of transportation varied between 9.9 ug/m3 (bike) and
17.2 ug/m3 (light rail). The standard deviations for driving with windows open, bus and
walking were 11.4 ug/m3, 11.6 ug/m3 and 12.6 ug/m3, respectively. It is therefore unlikely that the
comparative results were strongly affected by the omission of driving windows closed on the
Using the clean data matched by second to the ambient monitoring station, the average
concentration of PM2.5 experienced by the different commuting types was between 5.21 ug/m3
(driving with windows closed) and 15.21 ug/m3 (driving with windows open). Mean exposures
were 12.49 ug/m3 for light rail, 12.21 ug/m3 for walking, 12.62 ug/m3 for biking and 13.03 ug/
m3 for taking the bus. While these averages are of interest, they may be differentially
influenced by background pollution levels. Therefore, we focus the reporting of results below on
the ratio of personal exposure by commuting type to background levels at each recorded
The number of data points (seconds), means, medians, variances and standard errors of the
ratios of PM2.5 by commuting type and background monitor are shown in Table 1. Data are
reported by type, date and time of day. The overall average ratio of particulate matter to
background monitor was highest for commuting by bus (2.63). The lowest was for driving with the
windows closed (0.93). Driving with windows closed was the only method of commuting that
had an average ratio of less than 1 compared to the background monitor. The means were
influenced by widely different variances by day, time and commuting type, as mentioned. For
example, on the morning of 8/11, the walker experienced some very high particulate matter
levels, with a variance of 163.21. The median values for all commuting types are therefore also
informative, and ranged from 0.45 (driving windows closed) to 1.33 (bus). Ratio means and
standard errors are shown in Figs 2 and 3 for overall exposure and exposure by time of day
(morning or afternoon).
The linear regression analysis showed that all commuting types were statistically
significantly different from the reference type, driving windows closed (p < 0.001).
Heteroscedasticity was not present in the residuals as determined by the Breush-Pagan test.
As shown in Table 2, the inhaled dose (μg) was highest for active commuters. Walkers
and cyclists' inhaled doses were 21.3 and 8.7 times higher, respectively, compared with
driving with windows closed over the same 2.7 km route. Inhaled doses for driving with
windows open and light rail were 3.9 and 4.0 times higher, respectively, compared with driving
with windows closed. Although the bus did not travel the entire route, we estimated the
inhaled dose (1.23 μg) for bus commuters, based on a 7.45 min travel time, to be 3.8 times
higher than driving with windows closed. When calculated by μg/hr of commuting, active
commuting by bike and walking resulted in PM2.5 exposure rates that were 4.8 and 4.5 times
higher, respectively, than driving with windows closed. This was due largely to the elevated
minute ventilation rates associated with active commuting and the lower PM2.5
concentration inside cars with windows closed. Exposure rates for driving with windows open, light
rail, and bus were 2.9, 2.6, and 2.7 times higher, respectively, than driving with windows
6 / 15
1 SidePak AM510 personal exposure monitors (TSI, Inc., Shoreview, MN, USA) were set to record measurements every second, and therefore the total
number of minutes for each mode of commuting can be derived by dividing the number of data points by 60.
2 Drive WO and Drive WC = driving with windows open windows closed, respectively.
In this study, we examined fine particulate matter exposure for six different commuting types
simultaneously on a single prescribed urban route during ªrush hourº times. Breathing zone
PM2.5 concentrations were highest for driving with windows open and bus travel, and lowest
for driving with windows closed. However, other factors such as travel time, mode of
commuting, and minute ventilation rates can significantly influence the rate of exposure and overall
inhaled dose of particulate matter [
]. Considering these factors, our findings support
previous studies that estimate the inhaled dose of PM2.5 for active commuters to be 2±7 times
greater than for automobile and public transit commuters over comparable routes [
Similarly, McNabola et al. (2008) estimated that cyclists and pedestrians in their study would
experience greater PM2.5 lung deposition than car commuters, largely due to increased
respiration rates [
]. By comparison, the estimated inhaled doses of PM2.5 for cyclists and walkers in
Fig 2. Mean ratios of PM2.5 between personal and stationary monitor values by commuting type. Bars represent standard errors.
8 / 15
Fig 3. Mean ratios of PM2.5 between personal and stationary monitor values by commuting type and time of day. Bars
represent standard errors.
our study were approximately 9 and 21 times greater, respectively, than driving with windows
closed over the same 2.7 km route.
To date, most studies compare exposures across modes of transportation in terms of
breathing zone concentrations (μg/m3); however, inhaled dose (μg) and exposure rate (μg/hr)
may be better measures for understanding health outcomes associated with commuting.
Several previous studies show no difference or lower PM2.5 concentrations for cyclists and walkers
than for automobile commuters [13±15], but these studies did not account for the higher
respiration rate for active commuters. In our study, driving with windows open and riding the bus
exposed commuters to the highest breathing zone concentrations, but resulted in relatively
low inhaled doses and exposure rates compared to active commuters. Zuurbier (2010) found
similar results, where the highest median PM2.5, PM10, and soot concentrations were among
1 The distance for bus commuting was only 44% of the entire 2.7 km (1.68 mi) route, thus resulting in a lower inhaled dose of fine particulate matter. Had the
bus traveled the entire route, we estimated the trip time to be 7.45 min, resulting in an inhaled dose of 1.23 μg.
9 / 15
automobile and bus commuters, but the highest exposure rates were found among high-traffic
]. In their study, PM2.5 exposures during cycling occurred at a rate of 100.4 μg/hr,
compared to 51.9 μg/hr for diesel car [
]. By comparison, exposure rates in our study for
cycling, walking, and driving with windows closed were 18.0, 16.8, and 3.7 μg/hr, respectively.
Commuting accounts for 6±10% of the day for many workers, but may contribute up to 12%
of daily PM2.5 exposure [
21, 34, 35
]. Based on our results, we suggest the daily PM2.5
contribution attributable to commuting for cyclists and walkers may be 4±5 times higher than among
automobile users. There is currently little data on the long-term health effects of increased air
pollution exposure among active commuters, warranting additional research in this area [
We anticipated driving with windows closed with the car's air conditioning with air
recirculating on in the cabin (not drawing outside air into the vehicle) to result in the lowest exposure
rates. Closing the vehicle windows and turning on the air recirculation system creates a
mostly-closed, protective bubble around the automobile's occupants. However, Knibbs, de
Dear, and Atkinson (2009) described that even vehicles with closed windows experience some
air flow through ªleaksº [
]. A variety of sources can cause leaks, but they seem to be
increasingly common in older vehicles, those with poor quality seals around doors and windows,
poor quality filtration systems, and leaks through the bulkhead. Based on this, newer vehicles,
like those used in this study, seem to allow the least outside air to enter the cabin (i.e. be the
least ªleakyº), and thus may provide the greatest protection against traffic-related air pollution.
Conversly, Ott, Klepeis, and Switzer (2008) explained that opening a window significantly
increases exchange between cabin and outdoor air (e.g. opening a window just 3-inches
increased the air exchanges per hour by 8±16 times) [
]. It should be noted, that variation in
protection offered by closed windows is dependent on the vehicle's air filter, the traffic
congestion, and air circulation settings [
]. From our study, we see that driving with windows closed
with the automobile's air conditioning with air recirculating on in the cabin (not drawing
outside air into the vehicle), despite ªleaks,º protects the commuter from outdoor air pollution,
whereas driving with windows open significantly increases one's exposure.
Our findings support Briggs et al. (2008) who showed that walking exposed commuters
to 2.2 times higher mean PM2.5 concentrations compared to driving with the car windows
closed. In their study, the car's air conditioner was turned off, and the ventilation system was
set to a ªmoderateº level [
]. Our results were very similar. We found that walking and cycling
exposed commuters to 2.3 and 2.4 times higher mean PM2.5 concentrations, respectively,
compared to driving with windows closed. In our study, the car's air conditioning system was used
with the air recirculating on in the cabin (not drawing outside air into the vehicle). However,
other studies show in-car exposures can be higher than for cyclists and walkers [
Zuurbier's (2010) study , car windows were closed and the air conditioner was used on a
ªmoderateº level; and in McNabola's (2008) study [
], windows and vents were closed and
the air conditioner was not used. It is not clear from each study's methods whether the internal
air circulation system was turned on or off. We suggest this may be an important factor that
contributes to the protection provided by closed-window driving. Future studies are needed to
compare ventilation system configurations for driving in high pollution environments to
determine the optimal settings to reduce exposures.
A combination of leaks when windows are closed and infiltration of outdoor air when
windows and doors are open likely explain the high level of air pollution in busses. Sabin et al.
(2005) describe that busses, especially older ones, can self-pollute (i.e. particles enter the bus
even though the windows are closed), and passenger exchange can exacerbate this effect [
Hammond, Jones, and Lalor (2007) explain that particle exposure on busses is also affected by
bus idling and ventilation in that [
]. Data collection for our study occurred during August,
which is usually the hottest month of the year in Utah. As is typical during hot summer
10 / 15
months, the busses on the route used in this study had the air-conditioning system turned on,
but we did not systematically measure if windows were opened or closed. Given the previously
reported findings, it is likely that the high PM2.5 concentrations in busses in our study were
due to self-pollution via leaks and air exchange between the cabin and outdoor air when the
bus doors were opened during stops or from windows being opened.
We found considerable variation in PM2.5 levels among all transportation modes with the
exception of driving with windows closed. The most pronounced differences were seen for
driving with windows open and walking, which had more than 10 times greater average
variation than driving with windows closed. The high variation in exposure for driving with
windows open may be explained by greater PM2.5 concentrations entering the vehicle while idling
at stoplights and lower concentrations experienced while driving. The variability experienced
by walkers is likely due to microenvironments along the walking path [
]. Over a comparable
route, a walker would have more time to experience both extremes of very low and very high
exposure concentrations. Variation in PM2.5 measures also fluctuated significantly by day and
time of day. For instance, the highest average variation in exposure for all modes (except light
rail) occurred on the morning of August 11. Similarly, the lowest average variation for all
modes (except bus) occurred in the afternoon on August 8 (Table 1). Traffic density and wind
speed & direction influence PM2.5 exposures [
], and may account for the large variations
seen across days/times in this study. Unfortunately we did not measure these variables, and
can only speculate as to their effect. Closing car windows and turning on the internal air
circulation system appears to create a protective ªbubbleº around the vehicle occupants during
commuting, resulting in significantly lower variation in PM2.5 exposures.
There are important practical implications from this research. First, studies suggest the
health benefits of active commuting outweigh the risks of higher air pollution exposures [
]. These benefits include reduced risk for several chronic diseases such as obesity,
cardiovascular disease, and cancer [
]. However, there is little data on commuters' perceptions
about air pollution exposure and how these perceptions may influence their commute in terms
of route, time of day, and mode of transportation. Understanding these perceptions may help
guide future educational efforts aimed at reducing active commuters' air pollution exposures.
Second, public transportation (bus and light rail) disproportionately exposed commuters to
about four times greater inhaled doses of PM2.5 over the same route compared with
automobile commuters who drove with windows closed. This may pose a greater lifetime risk of
cardiopulmonary mortality and lung cancer to minority commuters, who tend to use public
transportation more often than Caucasians; likewise, roughly 65% of public transportation
users have a household income less than the national median household income [
Third, driving with windows open exposed commuters to the highest concentration of
pollutants. As with active commuting, educational campaigns directed at automobile commuters
may help lower exposures by increasing their knowledge and self-efficacy for optimizing their
vehicle's ventilation system to reduce PM2.5 concentrations in the car. Finally, safety and
infrastructure are commonly evaluated in promoting active transportation; however, air pollution
exposures should be another important consideration in city planning when selecting routes
for active commuters.
This study was limited to examining commuter data on one roadway in a single
metropolitan city during summer, and therefore may not be generalizable to other locations or
seasons. We were unable to account for the large variability in PM2.5 measures by mode of
commuting, day, and time. In retrospect, measuring traffic density and the types of vehicles
using the road may have helped in this regard. We used real-time particle counters to
measure breathing zone concentrations, which did not allow for chemical speciation of PM2.5.
Likewise, we only measured PM2.5 not ultrafine or coarse PM, or other gaseous pollutants
11 / 15
(e.g. ozone). The SidePak AM510 instruments measured PM2.5 concentrations indirectly
using 90Ê light scattering based on ISO 12103±1 A1 calibration test dust. Thus, actual
ambient PM2.5 concentrations may have differed from our readings due to differences in particle
size, shape, refractive index, and chemical composition; however, we assume that any
measurement errors were systematic and roughly proportionate among all modes of
commuting. Finally, although we did re-zero the SidePak instruments each night, we did not check
for zero drift immediately before or after sampling. It is possible that temperature
differences between the laboratory and sampling site location caused the SidePaks to read higher
in the afternoon as temperatures increased.
Studies suggest the health effects associated with PM2.5 exposure in adults and children may
depend on its chemical makeup [44±47]. Fondelli et al. (2008) found differences in fine particle
sulfur concentrations between busses and taxis [
], suggesting additional research is
warranted to understand how PM2.5 constituents vary by mode of commuting. This may provide
additional valuable information to help direct future public health policy recommendations
specific to commuting. This study was also limited in that we did not measure minute
ventilation rates, but used values reported in previous studies. This may have resulted in under or
over-estimation of actual inhaled dose and exposure rates due to differences in the fitness
levels of study participants, speed of travel, weather conditions, and grade (flat vs sloped terrain)
between this and other studies. In some settings, active transportation users choose different
routes than those using motorized transportation. This study fixed route choice in order to
make comparisons between transportation modes. Strengths of this study include the use of an
identical background monitor at the sampling location, measurement of six commuting
modes simultaneously on a common route, and controlling for dander released from clothing
by using the same over-shirt for all data collection personnel.
In conclusion, our data suggests that active commuting (walking and cycling), when
compared with other modes of transportation over comparable distances and routes, results in
significantly higher PM2.5 doses and exposure rates. The higher exposure doses and rates
estimated in our study were largely a function of increased respiration rate and time of travel for
active commuters. Furthermore, driving with windows closed with the car's internal air
circulation system turned on was protective, resulting in the lowest PM2.5 breathing zone
concentration, inhaled dose, and exposure rate. Future research should evaluate differences in fine
particle constituents based on mode of commuting, including in congested conditions. These
comparisons should also include relevant biological measurements. Furthermore, we
recommend future studies to evaluate public awareness of how transportation choices influence
commuters' daily exposures.
We would also like to thank Paige Oliver, Tristin A. Gilmore, Robert J. Graul, Aaron M. Cox,
Lexie Barrett-Graul, Parker J. Banbury, Natalia Nielsen, Elyssa K. Himmer, and Joshua P.
Toone for their help collecting data.
Conceptualization: Robert A. Chaney, Chantel D. Sloan, James D. Johnston.
Data curation: Robert A. Chaney, Chantel D. Sloan, Victoria C. Cooper, Daniel R. Robinson,
James D. Johnston.
Formal analysis: Chantel D. Sloan, Victoria C. Cooper, Daniel R. Robinson, James D.
12 / 15
Funding acquisition: Chantel D. Sloan, James D. Johnston.
Investigation: Robert A. Chaney, Chantel D. Sloan, Nathan R. Hendrickson, Tyler A.
McCord, James D. Johnston.
Methodology: Robert A. Chaney, Chantel D. Sloan, James D. Johnston.
Project administration: Robert A. Chaney, Chantel D. Sloan, Daniel R. Robinson, James D.
Resources: Robert A. Chaney, Nathan R. Hendrickson, James D. Johnston.
Software: Chantel D. Sloan.
Supervision: Robert A. Chaney, James D. Johnston.
Validation: Chantel D. Sloan.
Visualization: Chantel D. Sloan.
Writing ± original draft: Robert A. Chaney, Chantel D. Sloan, Victoria C. Cooper, Daniel R.
Robinson, Nathan R. Hendrickson, Tyler A. McCord, James D. Johnston.
Writing ± review & editing: Robert A. Chaney, Chantel D. Sloan, James D. Johnston.
13 / 15
14 / 15
1. Pope CA 3rd, Dockery DW . Health effects of fine particulate air pollution: lines that connect . J Air Waste Manag Assoc . 2006 ; 56 ( 6 ): 709 ± 42 . PMID: 16805397
2. Thurston GD , Ito K , Lall R. A Source Apportionment of U.S. Fine Particulate Matter Air Pollution . Atmos Environ ( 1994 ). 2011 ; 45 ( 24 ): 3924 ± 36 .
3. Chen LW , Watson JG , Chow JC , DuBois DW , Herschberger L. PM2 . 5 source apportionment: reconciling receptor models for U.S. nonurban and urban long-term networks . J Air Waste Manag Assoc . 2011 ; 61 ( 11 ): 1204 ± 17 .
4. Wilson JG , Kingham S , Pearce J , Sturman AP . A review of intraurban variations in particulate air pollution: Implications for epidemiological research . Atmospheric Environment . 2005 ; 39 ( 34 ): 6444 ± 62 .
5. Brugge D , Durant JL , Rioux C . Near-highway pollutants in motor vehicle exhaust: a review of epidemiologic evidence of cardiac and pulmonary health risks . Environ Health . 2007 ; 6 : 23 . https://doi.org/10. 1186/ 1476 -069X- 6 -23 PMID: 17688699
6. Laumbach R , Meng QY , Kipen H . What can individuals do to reduce personal health risks from air pollution? J Thorac Dis . 2015 ; 7 ( 1 ): 96 ± 107 . https://doi.org/10.3978/j.issn.2072- 1439 . 2014 . 12 .21 PMID: 25694820
7. Bureau USC . Commuting characteristics by sex: 2015 American Community Survey 1 - Year Estimates . 2015 (https://factfinder.census.gov/faces/tableservices/jsf/pages/productview.xhtml?pid=ACS_ 15 _ 1YR_ S0801&prodType=table (accesssed 1 Mar 2017)) .
8. Zuurbier M , Hoek G , Oldenwening M , Meliefste K , van den Hazel P , Brunekreef B . Respiratory effects of commuters' exposure to air pollution in traffic . Epidemiology . 2011 ; 22 ( 2 ): 219 ± 27 . https://doi.org/10. 1097/EDE.0b013e3182093693 PMID: 21228698
9. Stieb DM , Judek S , Burnett RT . Meta-analysis of time-series studies of air pollution and mortality: effects of gases and particles and the influence of cause of death, age, and season . J Air Waste Manag Assoc . 2002 ; 52 ( 4 ): 470 ± 84 . PMID: 12002192
10. Good N , Molter A , Ackerson C , Bachand A , Carpenter T , Clark ML , et al. The Fort Collins Commuter Study: Impact of route type and transport mode on personal exposure to multiple air pollutants . J Expo Sci Environ Epidemiol . 2016 ; 26 ( 4 ): 397 ± 404 . https://doi.org/10.1038/jes. 2015 .68 PMID: 26507004
11. Morabia A , Amstislavski PN , Mirer FE , Amstislavski TM , Eisl H , Wolff MS , et al. Air pollution and activity during transportation by car, subway, and walking . Am J Prev Med . 2009 ; 37 ( 1 ): 72 ±7. https://doi.org/ 10.1016/j.amepre. 2009 . 03 .014 PMID: 19524146
12. Briggs DJ , de Hoogh K , Morris C , Gulliver J. Effects of travel mode on exposures to particulate air pollution . Environ Int . 2008 ; 34 ( 1 ): 12 ± 22 . https://doi.org/10.1016/j.envint. 2007 . 06 .011 PMID: 17688949
13. Gulliver J , Briggs DJ . Personal exposure to particulate air pollution in transport microenvironments . Atmospheric Environment . 2004 ; 38 ( 1 ):1± 8 .
14. Adams HS , Nieuwenhuijsen MJ , Colvile RN , McMullen MAS , Khandelwal P . Fine particle (PM2.5) personal exposure levels in transport microenvironments , London, UK. Sci Total Environ . 2001 ; 279 ( 1 ± 3): 29 ± 44 . PMID: 11712603
15. Boogaard H , Borgman F , Kamminga J , Hoek G . Exposure to ultrafine and fine particles and noise during cycling and driving in 11 Dutch cities . Atmospheric Environment . 2009 ; 43 ( 27 ): 4234 ± 42 .
16. Zuurbier M , Hoek G , Oldenwening M , Lenters V , Meliefste K , van den Haze P , et al. Commuters' Exposure to Particulate Matter Air Pollution Is Affected by Mode of Transport, Fuel Type, and Route . Environmental Health Perspectives . 2010 ; 118 ( 6 ): 783 ±9. https://doi.org/10.1289/ehp.0901622 PMID: 20185385
17. McNabola A , Broderick B , Gill L . Relative exposure to fine particulate matter and VOCs between transport microenvironments in Dublin: Personal exposure and uptake . Atmospheric Environment . 2008 ; 42 ( 26 ): 6496 ± 512 .
18. Ott W , Klepeis N , Switzer P . Air change rates of motor vehicles and in-vehicle pollutant concentrations from secondhand smoke . Journal of exposure science and environmental epidemiology . 2008 ; 18 ( 3 ): 312 ± 25 . https://doi.org/10.1038/sj.jes. 7500601 PMID: 17637707
19. Hewitt CN . Spatial variations in nitrogen dioxide concentrations in an urban area . Atmospheric Environment Part B Urban Atmosphere . 1991 ; 25 ( 3 ): 429 ± 34 .
20. Zhu Y , Hinds WC , Kim S , Shen S , Sioutas C . Study of ultrafine particles near a major highway with heavy-duty diesel traffic . Atmospheric environment . 2002 ; 36 ( 27 ): 4323 ± 35 .
21. Fondelli MC , Chellini E , Yli-Tuomi T , Cenni I , Gasparrini A , Nava S , et al. Fine particle concentrations in buses and taxis in Florence, Italy . Atmospheric Environment . 2008 ; 42 ( 35 ): 8185 ± 93 .
22. Thai A , McKendry I , Brauer M. Particulate matter exposure along designated bicycle routes in Vancouver, British Columbia . Sci Total Environ . 2008 ; 405 ( 1 ): 26 ± 35 .
23. Chung A , Chang DP , Kleeman MJ , Perry KD , Cahill TA , Dutcher D , et al. Comparison of real-time instruments used to monitor airborne particulate matter . Journal of the Air & Waste Management Association . 2001 ; 51 ( 1 ): 109 ± 20 .
24. Hatzopoulou M , Weichenthal S , Dugum H , Pickett G , Miranda-Moreno L , Kulka R , et al. The impact of traffic volume, composition, and road geometry on personal air pollution exposures among cyclists in Montreal, Canada . J Expo Sci Env Epid . 2013 ; 23 ( 1 ): 46 ± 51 .
25. Zuurbier M , Hoek G , van den Hazel P , Brunekreef B . Minute ventilation of cyclists, car and bus passengers: an experimental study . Environmental health . 2009 ; 8 ( 1 ): 48 .
26. Rojas-Rueda D , De Nazelle A , TeixidoÂ O , Nieuwenhuijsen M. Replacing car trips by increasing bike and public transport in the greater Barcelona metropolitan area: a health impact assessment study . Environment international . 2012 ; 49 : 100 ±9. https://doi.org/10.1016/j.envint. 2012 . 08 .009 PMID: 23000780
27. Bureau USC . Quick Facts: Salt Lake City, Utah. www.censusgov/quickfacts/table/PST045215/ 4967000,00 2017 ; (accessed 3 Mar 2017 ).
28. Utah Mapping PortalÐThe Utah Automated Geographic Reference Center. State C , and Municipal Boundaries. https://gis.utah.gov/data/boundaries/citycountystate/ (accessed 3 Nov 2016 ). 2014 .
29. Association AIH . The Occupational Enviornment: It's Evaluation and Control . Fairfax, Virginia: AIHA Press; 1997 .
30. Williams R , Suggs J , Creason J , Rodes C , Lawless P , Kwok R , et al. The 1998 Baltimore Particulate Matter Epidemiology-Exposure Study: part 2. Personal exposure assessment associated with an elderly study population . J Expo Sci Env Epid . 2000 ; 10 ( 6 ): 533 .
31. Williams R , Suggs J , Rea A , Leovic K , Vette A , Croghan C , et al. The Research Triangle Park particulate matter panel study: PM mass concentration relationships . Atmospheric Environment . 2003 ; 37 ( 38 ): 5349 ± 63 .
32. Team RC . R: A Language and Environment for Statistical Computing . 2017 .
33. Cepeda M , Schoufour J , Freak-Poli R , Koolhaas CM , Dhana K , Bramer WM , et al. Levels of ambient air pollution according to mode of transport: a systematic review . The Lancet Public Health . 2017 ; 2 ( 1 ): e23± e34 .
34. GoÈtschi T , Garrard J , Giles-Corti B . Cycling as a part of daily life: a review of health perspectives . Transport Reviews . 2016 ; 36 ( 1 ): 45 ± 71 .
35. Karanasiou A , Viana M , Querol X , Moreno T , de Leeuw F. Assessment of personal exposure to particulate air pollution during commuting in European citiesÐRecommendations and policy implications . Science of the Total Environment . 2014 ; 490 : 785 ± 97 . https://doi.org/10.1016/j.scitotenv. 2014 . 05 .036 PMID: 24907613
36. Knibbs LD , De Dear R , Atkinson SE . Field study of air change and flow rate in six automobiles . Indoor air . 2009 ; 19 ( 4 ): 303 ± 13 . https://doi.org/10.1111/j.1600- 0668 . 2009 . 00593 . x PMID : 19500174
37. Zhu Y , Eiguren-Fernandez A , Hinds WC , Miguel AH . In-cabin commuter exposure to ultrafine particles on Los Angeles freeways . Environmental science & technology. 2007 ; 41 ( 7 ): 2138 ± 45 .
38. Sabin LD , Behrentz E , Winer AM , Jeong S , Fitz DR , Pankratz DV , et al. Characterizing the range of children's air pollutant exposure during school bus commutes . Journal of Exposure Science and Environmental Epidemiology . 2005 ; 15 ( 5 ): 377 ± 87 .
39. Hammond D , Jones S , Lalor M. In-vehicle measurement of ultrafine particles on compressed natural gas, conventional diesel, and oxidation-catalyst diesel heavy-duty transit buses . Environmental monitoring and assessment . 2007 ; 125 ( 1 ±3): 239 ± 46 . https://doi.org/10.1007/s10661-006-9515-z PMID: 17219245
40. Gomez-Perales J , Colvile R , Nieuwenhuijsen M , Fernandez-Bremauntz A , Gutierrez-Avedoy V , Paramo-Figueroa V , et al. Commuters' exposure to PM 2.5, CO, and benzene in public transport in the metropolitan area of Mexico City . Atmospheric Environment . 2004 ; 38 ( 8 ): 1219 ± 29 .
41. Prevention CfDCa. Transportation Health Impact Assessment Toolkit 2011 [https://www.cdc.gov/ healthyplaces/transportation/promote_strategy.htm.
42. Bureau USC . Income, Poverty and Health Insurance Coverage in the United States: 2015: U .S. Census Bureau; 2016 [https://www.census.gov/newsroom/press-releases/ 2016 /cb16- 158 .html.
43. Association APT . A profile of public transportation passenger demographics and travel characteristics reported in on-board surveys2007 5/1/2017 . http://www.apta.com/resources/statistics/Documents/ transit_passenger_characteristics_text_ 5 _ 29 _ 2007 .pdf.
44. Bell ML , Ebisu K. Environmental inequality in exposures to airborne particulate matter components in the United States . Environmental health perspectives . 2012 ; 120 ( 12 ): 1699 . https://doi.org/10.1289/ ehp.1205201 PMID: 22889745
45. Bell ML , Belanger K , Ebisu K , Gent JF , Lee HJ , Koutrakis P , et al. Prenatal exposure to fine particulate matter and birth weight: variations by particulate constituents and sources . Epidemiology (Cambridge, Mass). 2010 ; 21 ( 6 ): 884 .
46. Ebisu K , Bell ML . Airborne PM^ sub 2 .5^ Chemical Components and Low Birth Weight in the Northeastern and Mid-Atlantic Regions of the United States . Environmental health perspectives . 2012 ; 120 ( 12 ): 1746 . https://doi.org/10.1289/ehp.1104763 PMID: 23008268
47. Bell ML , Ebisu K , Peng RD , Samet JM , Dominici F. Hospital admissions and chemical composition of fine particle air pollution . American journal of respiratory and critical care medicine . 2009 ; 179 ( 12 ): 1115 ± 20 . https://doi.org/10.1164/rccm.200808-1240OC PMID: 19299499