EPA/AA/CTAB/PA/81-14
TECHNICAL REPORT
Review of the Literature and On-going EPA Projects
Comparing Portable Dosimeters and Fixed Site Monitors as
Accurate Indicators of Exposure to Carbon Monoxide
by
Allan W. Ader
May, 1981
NOTICE
Technical reports do not necessarily represent final EPA decisions or
positions. They are intended to present technical analyses of issues using
data which are currently available. The purpose in the release of such
reports is to facilitate the exchange of technical information and to inform
the public of technical developments which may form the basis of a final EPA
decision, position or regulatory action.
Control Technology Assessment and Characterization Branch
Emission Control Technology Division
Office of Mobile Source Air Pollution Control
Office of Air, Noise and Radiation
U.S. Environmental Protection Agency
2565 Plymouth Road
Ann Arbor, Michigan 48105
-------
Introduction
The primary source of carbon monoxide (CO) present in the atmosphere is the
incomplete combustion of gasoline-fueled cars with carbureted,
spark-ignitipn reciprocating engines. The CO emitted is a function of
concentration of CO in the exhaust gases, flow rate of exhaust gases and
duration of operation (1). Future exposure to ambient CO concentrations
will clearly depend on future amounts of CO emitted into the atmosphere and
future CO emission patterns. As mandated by Section 319 of the Clean Air
Act, substantial effort has been expended by the U.S. Environmental Pro-
tection Agency (EPA) to quantify the extent of CO exposure to the U.S. popu-
lation. EPA and various state and local environmental quality agencies have
promulgated standards for CO reduction based on measurement of CO at fixed
site monitors located throughout these areas. Whether CO levels measured at
these fixed site monitors are indicative of personal CO exposure has recent-
ly been examined by several investigators (2,3,4). These studies raise some
important questions as to whether personal dosimeters may be a better indi-
cator of exposure and biological dose than fixed site monitors and whether
fixed site monitor concentrations correlate with biological dose.
This report will be primarily concerned with reviewing and evaluating previ-
ous studies comparing personal dosimeter readings and fixed site monitors.
The secondary objective will be to assess the correlation, if any, of either
or both of these monitoring techniques to various biological measurements of
CO exposure including "end-expired" breath analysis and blood carboxy-
hemoglobin (COHb). An accurate assessment of the above parameters is es-
sential especially in urban areas where CO levels above the National Ambient
Air Quality Standard (NAAQS) of 10 mg/m^ (9 ppm) for an eight-hour limit
and 40 mg/m3 (35 ppm) for a maximum one-hour limit have been reported. i
The following studies in the literature related to personal dosimetry and
fixed site monitors have been evaluated in depth:
- Cortese and Spengler (Boston) study (2)
- Jabara, et al. (Denver) study (3)
- Wallace (Washington, D.C.) study (4)
- Wilson and Schweiss (Seattle) study (5)
- Wilson and Schweiss (Boise) study (6)
Where necessary, the significance of the findings in terms of comparing the
two methods of CO exposure monitoring has been reassessed by additional sta-
tistical analysis*. In addition, the on-going short-term and long-term EPA
projects on CO monitoring will be described. Hopefully these EPA projects
will provide a more complete data base in this area than the aforementioned
studies.
*(Note: We wish to thank Mr. James Jabara of the U.S. Army who conducted
the Denver study for providing us with the original data on which further
calculations were made.)
-------
-3-
Boston Study;
Cortese and Spengler (2) conducted a monitoring study of sixty-six (66)
nonsmoking individuals who commuted by several means of transportation to
different locations in the Boston metropolitan area. The primary purpose of
the study was to determine whether 1 hour concentrations of CO during com-
muting exceeded the. NAAQS. Eight-hour CO dosimeter readings were also
taken. Results from personal dosimeter monitors were used to determine com-
muting exposures and integrated 1 hour and 8 hour exposures were de-
termined. Personal dosimeter exposure data were compared to concentrations
measured at fixed location monitors operated by the Massachusetts Bureau of
Air Quality Control at urban locations near the employment .of most of the
commuters and suburban locations. Alveolar air samples were taken by each
participant before and after each commuting trip to estimate the amount of
CO absorbed by the blood during commuting.
The mean personal 1-hour commuting exposures measured by dosimeters was 1.4
times greater than the hourly reading of the fixed site monitors at urban
locations (10.8 ±,5.1 ppm for a mean personal exposure 1-hour concentration
compared to a mean hourly fixed location concentration at the urban stations
of 7.8 _+ 4.9 ppm). The mean maximum 1 hour personal exposure concentration
was 2.1 times the mean concentration (5.1 _+ 2.9 ppm) of the fixed site read-
ings at suburban locations. Cortese and Spengler also examined the re-
lationship at the range of concentrations closest to the NAAQS (the upper
5-7% of the personal exposure and fixed location measurements). The mean
personal exposure concentration (25.3 ppm) was 1.6 times the mean concen-
tration at all fixed stations (15.6 ppm) and 1.3 times the mean concen-
tration at urban stations (19.9 ppm).
In contrast to the 1-hour exposure data, a comparison of maximum 8 hour mean
personal exposure readings with those measured at fixed site monitors showed
the mean 8 hour fixed site urban station readings (6.6 ppm) to be 1.6 times
higher than the mean 8 hour personal dosimeter hourly average (4.2 + 1.9
ppm). Average concentrations at suburban fixed site monitors were similar
to the 8 hour average personal dosimeter readings. Table 1 summarizes the
results of the Cortese and Spengler study.
-------
-4-
Table 1
Summary of Cortese and Spengler Study
Measurement From Compared To Result
Commuter Urban Fixed-Site Personal Dosimeter
Personal Dosimeter Monitor, 1-hr average higher by a factor
1-hr average of 1.4 (ave.) to
2.1 (max.)
Commuter Urban Fixed-Site Personal Dosimeter
Personal Dosimeter Monitor, 8-hr lower by a factor
8-hr average average of 0.63
Commuter Suburban Fixed-Site Personal Dosimeterr
Personal Dosimeter Monitor, 8-hr about the same
8-hr average average as fixed-site
monitor
The alveolar air sample measurements showed no relationship with commuting CO
exposure measured by dosimeters. Although actual data on alveolar "end-
expired" breath samples are not presented by Cortese and Spengler, they imply
that there was essentially no difference between before and after commuting
alveolar air concentrations of CO. The basic premise upon which these
results are valid concerning biological dose of CO is if "end-expired" breath
samples are truly indicative of blood COHb and/or dose. Cortese and Spengler
state the following on this subject concerning the relatively low CO ex-
posures measured:
"It is also difficult to measure blood COHb levels below 2% [references
deleted] and the data relating alveolar air concentration to blood COHb
at such levels are sparse and difficult to interpret. Therefore,
alveolar air samples may not reflect the slight changes in COHb concen-
tration that occur from ambient CO exposure." (2)
Denver Study;
Jabara, et al. (3) evaluated the occupational exposure to CO of Denver
traffic officers during eight hour work shifts by comparing personal
dosimetry, before and after workshift breath samples analyzed for CO, and
ambient CO levels at fixed site monitors. The study design was a structured
random sampling of volunteers separated into control and exposed groups by
shift and job classification. At the beginning of each shift, dosimeters
were distributed to three exposed and one control volunteer and breath
samples were taken. The controls were police department employees who worked
at the traffic bureau but remained indoors during working hours. The exposed
group consisted of traffic officers working at urban traffic locations. At
the end of the work shift, breath samples from all participants were col-
lected, dosimeter readings were recorded and questionnaires concerning
-------
-5-
smoking habits were returned and the information recorded. Breath samples
were analyzed by the method of Jones, et al. (7) and were collected in the
same manner as in the Boston study. The average ambient CO levels monitored
at fixed sites were also recorded on the same day as the personal monitoring
took place. Ambient concentrations of CO were determined by reference to the
three nearest fixed site monitors within the Denver metropolitan area to the
traffic officers in the field.
Figure 1 presents the average daily dosimeter readings for the subjects and
controls compared to the ambient CO concentrations. Statistical analysis by
Jabara, et al. showed a significant difference between subjects and
controls. Of particular interest to this paper, however, was whether the
dosimeter readings of controls who may be more representative of the environ-
mental exposures expected in the general community (the majority of their
time is indoors), was significantly different than the ambient readings at
fixed site monitors*. Jabara1s survey data were obtained and statistical
analysis was performed on the matched sets of data by the "Student" t-test,
which is used to detect significant differences between independent groups or
several different observations for each individual within a group. A sig-
nificant difference (0.0005 •* P<0.005)** was found between control dosimetry
values and ambient CO from fixed-site monitors, with as shown in Figure 1,
the dosimeters indicating a higher value than the fixed-site monitors. As
could be expected from Jabara1 s initial analysis of variance of the subject
and control data, the "Student" t-test showed a signficiant difference
(P<0.0005) between matched observations of subject dosimetery values and
ambient CO from fixed site monitors. Table 2 summarizes the exposure data
from the Denver study.
* The author acknowledges that indoor air pollution factors such as under-
ground garages and cigarette smoking may have been a factor in these results,
but the controls were not exposed to CO as commuters, therefore reducing this
aspect of their exposure; it is not known if these balance out.
** The author would like to thank Mr. Martin Atherton of the EPA Office of
Mobile Source Air Pollution Control for performing the statistical analysis
of the original data provided by Mr. Jabara.
-------
-6-
Ambient CO fixed
site monitors-
8-hour average
Dosimeter
Readings - 8-hour
average - All
participants*
Dosimeter Readings
8-hour average
Subjects
Dosimeter Readings
8-hour averages-
Controls
Table 2
Summary of Denver Exposure Data
Number Mean PPM
98
98
6.4
18.9
Standard
Deviation
3.2
10.4
75
19
21.7
9.9
9.9
5.9
4 Dosimeter readings could not be classified as either subjects or
controls because of mixed exposures but were included for this data
analysis.
Jabara determined the correlation coefficients between the different
parameters measured in the study for smoking and nonsmoking subjects. These
data, shown in Table 3, show that smoking greatly influenced COHb levels
determined by before and after shift breath analysis. The non-smoking sub-
jects correlations did show some significant results. The relationship be-
tween dosimeter and after shift breath analysis (r = .82) showed that
dosimeter readings are a good indicator of COHb at the end of the shift.
This result was in contrast to the previously mentioned Boston study. The
dosimeter versus change in breath (after-shift breath CO level minus the be-
fore shift breath CO level) relationship for nonsmoking subjects in the
Denver study also indicated a significant correlation (r = .64).. These
relationships for nonsmokers are important in terms of EPA's proposed re-
visions of the CO standard (45 Federal Register 55066, August 18, 1980)
because EPA has considered only nonsmokers in evaluating COHb data. Another
important relationship to point out is that the dosimeter readings and fixed
site monitor data do not correlate well for nonsmokers (r = .3990) and
smokers (r = .2119).
-------
-7-
Table 3
Correlation Coefficients Between Measured Data for
Smoking and Nonsmoking Subjects for Denver Study
Smokers Nonsmokers
1. Dosimeter vs. After .1829 .8228
2. Dosimeter vs. Change .4947 .6431
3. Before vs. After .8171 .3091
4. Before vs. Ambient -.1483 .3631
5. Dosimeter vs. Before -.1292 .2177
6. Dosimeter vs. Ambient .2119 .3990
7. Before vs. Change .0409 -.4051
8. After vs. Ambient .0139 .3018
9. After vs. Change .6094 .7443
10. Ambient vs. Change .2280 .0381
NOTES: Dosimeter: Personal breathing zone sampling for CO, with an
integration of the CO dose over the entire workshift (8-hour time-weighted
average). Ambient: Stationary site sampling for CO, 8-hour moving
average. Before (breatn): Personal measurement of CO in the workers'
expired breath before the workshift. After (breath): Personal measurement
of CO in the workers' expired breath after the workshift. Change (in
breath): After-shift breath CO level minus the before-shift breath CO
level. (Adapted from Reference 2)
Further statistical analysis was performed with Jabara's data to determine
whether work-related proximity to traffic congestion was related to
dosimeter readings. Traffic officers who Jabara had separated by job clas-
sification were grouped in ascending order of proximity to auto traffic con-
gestion and coefficients of correlation were determined for exposure class
versus dosimeter and before and after breath analysis. Work related proxi-
mity to traffic congestion correlated fairly well (r = .625) with the 8 hour
average dosimeter reading but did not correlate as well, if at all, with
after shift breath analysis (r = .140) and change in breath (r = .360).
Washington, D.C. Study:
Wallace of EPA1 s Office of Monitoring and Technical Support reported CO con-
centrations inside vehicles of commuters travelling to and from Washington,
D.C. from Reston, Va. during the summer of 1978 (4). Dosimeter
-------
-8-
readings taken inside the vehicle (most often on the rear seat) were
compared to ambient CO readings obtained from a Washington, D.C. fixed site
monitor. Although the intent of the Wallace study was to examine intrusion
of CO into automobile and bus interiors, the results of the study point out
the difference in results between dosimeters located fairly close to
commuters (perhaps more indicative of personal exposure) and fixed site
monitors.
Mean CO concentration for all vehicles (buses and automobiles) was 11.7 _+
4.9 ppm. The values in the interiors of the vehicles during city driving
averaged 50% higher than during suburban driving. The difference was sig-
nificant at the p ^.01 level. Ambient levels measured by the fixed site
monitor averaged 3.5 ^1.6 ppm in the city. Even vehicles in the suburbs,
where ambient levels are presumed to be lower, contained concentrations more
than twice as great as ambient levels measured by the fixed site montior in
the city. It should also be noted that although a considerable portion of
commuting time was in the suburbs, mean in-vehicle concentrations were still
3-4 times the values recorded by the fixed site monitor. Figure 2 shows
that ambient concentrations measured by the fixed site monitor and interior
concentrations measured by the dosimeters are not related (r = .1). How-
ever, the five days when ambient levels were high coincided with the high
exposures inside the vehicles probably indicating the communter's high
exposure as well.
Seattle and Boise Studies;
Wilson and Schweiss of EPA's Surveillance and Analysis Division Region X
prepared reports on the work performed in Seattle (5) and Boise (6) of the
extent of CO in these urban areas. Although these studies cannot be used to
compare fixed site monitors and personal dosimeters, they did examine the
correlation between ambient CO at fixed site monitors and CO concentrations
at street level bag samplers and indoor locations which were believed to
more accurately reflect actual human exposure.
The Seattle study involved a 20-day monitoring for CO at 36 outdoor sites,
five indoor sites, and two pedestrian walking routes in downtown Seattle.
For the 36 outdoor sites, bag samplers were placed at representative lo-
cations some of which were expected to yield low CO values because of re-
moteness from cars. Locations where idling vehicles occur were deliberately
avoided. At each site, the sample inlet was about 3.5 meters above the
sidewalk, more than 10 meters from an intersection and more than 2 meters
from a vertical walk. Most samples were over one meter from the street
curb; two were at a much greater distance (in parks). These bag sample
results were compared to the five permanent monitoring sites located in the
downtown area. The Boise study design was similar and involved 40 outdoor
sites, six indoor sites, and two pedestrian walking routes. In Boise, there
is only one continuous monitor to compare CO data. Indoor sites in the
Boise study, as in the Seattle study, were chosen within a few blocks of the
permanent monitor(s) and data from these sites were compared to the adjacent
outdoor sites. Indoors sites were equipped with bag samplers and continuous
analyzers to monitor the daily pattern of indoor CO levels.
The Seattle data show that the permanent monitoring network was not repre-
sentative of the highest frequency of exceedences of the 9 ppm NAAQS within
-------
-9-
the study area for the twenty day period. The eight-hour standard was
exceeded at one or more sites on 80% of the study days. Exceedences at one
or more of the permanent monitors occurred on 45% of the sampled days. Each
day the highest eight-hour average for any of the five permanent monitors
was compared to the highest eight-hour average at any of the study sites.
On most days, the maximum CO average at the study site was less than 1.5
times that at the maximum permanent monitor, but on six days was over 1.5
times as great. On four of the days when no violations were observed at the
permanent monitors, the maximum study site's eight-hour average exceeded 9.0
ppm and was more than 1.5 times greater than at the permanent monitor.
The Seattle data was also analyzed to look at the 20 day average CO con-
centrations measured by stationary monitors and those by the two nearest
experimental samplers. Table 4 presents these data and shows that seven of
the ten averaged values are only marginally higher than their matched
counterparts. Linear regression analysis of these averaged values revealed
fairly good correlation (r^ = .6) between the experimental and fixed site
measures.
Table 4
Twenty Day Average CO Concentrations
Stationary Monitors vs. Experimental Samplers
Seattle, Washington
CO Concentration
PPM
Station Site
Stationary Monitor
x s
EPA Site #1
Pike Street Station
University St. Station
James Street Station
Fire Station
Smejcor Street Station
EPA Site #2
8.0
6.0
4.4
2.3
5.5
4.5
3.5
4.2
2.3
3.6
5.5
6.4
3.9
5.1
5.4
3.4
4.0
3.0
2.8
3.0
8.2
6.6
5.5
2.9
6.0
3.0
3.0
2.6
1.1
1.7
Paired average concentration values for Seattle indoor/outdoor measures
appear in Table 5. Three of the five grouped averages are higher for indoor
measures than those outdoors. Wilson and Schweiss concluded that changes in
CO concentrations at outdoor sites frequently coincided with changes in CO
concentrations at indoor sites, but the relationship between indoor and
outdoor values was not constant.
-------
-10-
Table 5
Average Concentration CO Values for Indoor/Outdoor Sites
Seattle, Washington
Sample Site Pairs CO Concentration Means (PPM)
Indoor776
Outdoor 6.6
Indoor 5.4
Outdoor 5.5
Indoor 6.0
Outdoor 4.5
Indoor 5.2
Outdoor 6.3
Indoor 8.2
Outdoor 5.5
The Boise study data indicate that the fixed site permanent monitors gave CO
readings generally lower than the experimental stations. On every day but
one of the twenty day study, the eight-hour concentration at the permanent
monitor was exceeded at one or more study site. Wilson and Schweiss
conclude, "the permanent monitor was generally representative of the higher
concentrations but did not represent the highest concentrations or frequency
of exceedences within the study area." On 95% of the study days (19 of 20)
the eight-hour NAAQS was exceeded at one or more experimental sites.
Exceedences at the permanent monitor occurred on 47% (9 of 19) of the
sampled days. The Boise study found that changes in CO concentrations at
outdoor sites frequently coincided with changes in CO concentrations at
indoor sites, but the relationship between indoor and outdoor values was not
constant. Concentrations were usually lower indoors than at the adjacent
outdoor site in this limited study.
Other Considerations in the Literature:
In addition to the studies described above, several other studies have re-
ported on the relationship between fixed site monitors and personal
dosimeters. Studies by Ott and Eliassen (8), Godin et al. (9) and Wright et
al. (10) have shown that pedestrians on downtown urban streets are exposed
to CO concentrations that exceed the NAAQS that are not being observed at
fixed site monitoring stations. CO dosimeter concentrations were sub-
stantially higher than concentrations measured by fixed site monitors.
Other work has been performed on CO "hot spots" (11), areas such as major
urban intersections where CO levels may be highest. Further analysis of
-------
-11-
this study performed in San Jose, Chicago, Seattle and Phoenix is required
to determine the relationship between CO hot spots and fixed site monitor
readings.
Another consideration is the sampling height of air quality monitors and
their effect on CO'concentration. Johnson et al (12) discovered that ele-
vating the sampling height from the breathing zone (5.5 ft) to the height of
most fixed site monitors will produce decreases in CO concentration of
between 5 and 15%. A further study in this area by Brice and Roessler (13)
on horizontial and vertical CO concentrations found that mobile individuals
are exposed to CO concentrations different than those measured at fixed
sites. In the study of six cities of the Continuous Air Monitoring Program
(CAMP; of EPA, integrated 1/2 hour CO samples taken within automobiles in
traffic were compared to fixed site monitor readings. The ratio of simul-
taneous concentrations of CO in traffic to concentrations at fixed site
monitors ranged from 1.3 to 6.8.
On-Going EPA Projects:
To establish a more complete data base on fixed site monitor readings,
personal dosimeters and general population CO exposure and applicability of
these measurements to biological dose, EPA has established a multi-faceted
research program to answer the questions raised in the previously mentioned
studies. Short-term studies have recently been completed or are near com-
pletion to examine individual variation of CO exposures and microenvironment
exposures to CO. A microenvironment is defined as a discrete place (e.g.
underground parking garage, shopping center) or location (urban street)
which may have a specific CO level different from other microenvironments.
A person's total exposure to CO would be the result of the CO exposures
received in each microenvironment. Longer term studies will be larger in
scope and will try to relate individual activity patterns to CO exposure in
determining a complete profile of CO exposure to the general population.
The initial short-term monitoring study was conducted in Los Angeles by an
EPA contractor, Science Applications, Inc. (SAI). Nine volunteers carried
dosimeters in their daily activities to determine the reliability of state-
of-the-art personal exposure monitors. In the first phase of the project
conducted September-December 1980, 9 volunteers carried CO detectors to
various locations during the course of ordinary workday and weekend activi-
ties for 45 days each, recording their activity and integrated CO ex-
posures. Preliminary analysis indicated that automated data loggers were
required for the general public to use these monitors in a large scale
study. A total of 10 were designed and constructed by SAI and appear to
satisfy this need. Further work indicated that the CO detectors were usable
and the integrators provided the ability to collect data in environments
with rapidly varying CO levels. The data collected will be compared to
ambient CO levels measured at fixed site monitors. A draft report on SAI1 s
work is expected in June, 1981.
-------
-12-
In addition to the CO monitoring work in Los Angeles, other contractors have
recently completed the experimental parts of monitoring studies from
January-March 1981 in Denver, Co., Phoenix, Az., and Stamford, Ct. These
cities were selected because of previous histories of high CO levels above
the NAAQS. One of the main objectives of these monitoring studies is to
accurately assess CO exposures to individuals in the population during their
daily activities. Contractor personnel took dosimeter readings in simulated
human activities by driving in commuter traffic and entering offices, shops
and public areas, and in residential areas. The dosimeter readings taken in
the downtown areas will be in close proximity to the CO fixed site monitor
and will be compared to determine if the fixed site monitors are repre-
sentative of CO concentrations along commuting routes and other monitored
locations. These data will serve to expand the information data base of ex-
posure to CO temporally and in various activities and locations. GCA Inc.
is obtaining these measurements in Stamford while PEDCo Environmental, Inc.
and Systems Sciences Software, Inc. are performing these tasks in Denver and
Phoenix, respectively. The contractors will obtain the measurements and do
limited analysis. EPA will perform additional data analysis and prepare a
comprehensive report that will be released later this year.
Other data related to CO exposures in microenvironments have been collected
by Dr. Wayne Ott of Stamford University and EPA1s Office of Research and
Development. He has conducted field studies of CO exposures of occupants of
motor vehicles on El Camino Real in California, a one year study that will
give data on urban arterial highway CO exposure. Dr. Ott also has collected
field data on CO levels in commercial settings.
The objective of EPA"s long term CO studies is to develop a methodology that
can identify a distribution of the number of people exposed to various
levels of CO averaged over appropriate times, e.g., one hour or eight
hours. Two approaches are being taken to address this problem. The first
approach, being coordinated by Dr. Wayne Ott is the development of a model
that will predict, based on field and activity pattern data, the CO exposure
of individuals in the urban population. CO exposure data of microenviron-
ments collected from the four contractor studies and subsequent work by Dr.
Ott and ORD staff will be applied to the Simulation of Human Air Pollution
Exposures (SHAPE) computer model. The program combines activity profiles
with data on CO concentrations in specific microenvironments and urban back-
ground concentrations in order to calculate integrated exposures of a large
number of persons over a 24-hour period. The second approach is a compre-
hensive monitoring program of an urban population with approximately 100
volunteers or more participating.* The approach of tnis study would be to
identify representative people, equip them with personal CO exposure
monitors, have them record their CO exposure and then take the results of
these exposures and extrapolate the exposures to the urban population. This
study is planned to be performed over several years in different urban areas
to estimate the frequency distribution of CO exposures during a full year
for the subject population. These data will be compared to fixed site
ambient CO readings to determine the relationship of these two parameters
for different activities.
* As of May, 1981, EPA has not formalized the details of this work.
-------
-13-
Several parts of the long-term program are presently underway. The EPA
Office of Research and Development (ORD) is developing screening techniques
to select typical volunteers to carry CO monitors. ORD is also evaluating
whether it is better to have many individuals carry the monitors for short
periods of time or fewer individuals carry them for shorter periods of
time. This evaluation will depend on whether daily CO variation is greater
or less than person-to-person CO variation. ORD is also developing a
questionnaire to administer to people carrying the CO monitor so that they
may accurately and easily record their daily activities. Evaluation of CO
monitoring instrumentation is also presently being performed by EPA1s
Environmental Monitoring and Support Division. The contract for actual
monitoring of volunteers is expected to be awarded in 1982 or 1983.
-------
-14-
CONCLUSIONS
Accurate determinations of community exposure to carbon monoxide are needed
to determine wnether the NAAQS is being met and whether CO levels as
measured by fixed- site monitors reflect its demonstrated health risk.
Review of the data on the relationship of fixed site monitor concentrations
and personal dosimeter readings and the accuracy of these readings to human
exposure has led to the following conclusions.
1) Fixed site air quality monitors underestimate short term variations in
CO exposure. Fixed site monitors are not good measurements of
exceedences of EPA1s one-hour NAAQS for CO. This conclusion is in
general agreement with Goldstein and Landovitz's study of fixed site
air quality monitors in New York City (14). Cortese and Spengler (2)
found that mean personal 1-hour commuting exposures measured by
dosimeters were 1.4 times greater than the hourly readings of fixed
site monitors in urban locations of Boston, Massachusetts.
2) For commuting and urban pedestrian exposures urban fixed site monitors
underestimate CO exposure. Studies by Wallace (4) and several authors
(8,9,10) support this conclusion. Further work is being performed by
EPA to quantify this relationship.
3) Data on eight-hour CO exposures tend to snow that fixed-site monitors
underestimate personal exposures by dosimeters but some evidence is
contradictory. Data from EPA studies now underway at Stamford Ct.,
Los Angeles, Ca., Phoenix, Az., and Denver, Co., should provide a more
definitive statement on this topic.
It is realized that the 8-hr CO NAAQS is the critical issue with
respect to the level of control needed from mobile sources. The 1-hr
CO level is assumed by the EPA Office of Air Quality Planning and
Standards (OAQPS) to be related to the 8-hr average CO value. Our
previous conclusions show how 1-hour CO exposures are underestimated
by the fixed site monitors. However, no advice can be given as to the
ratio of fixed site monitor 8-hour average readings to dosimeter
8-hour averages. The studies in this report put some doubt on whether
this ratio is unity.
4) The fixed-site monitoring system does not accurately assess the extent
of CO problems in U.S. cities. CO levels at urban locations often
exceed the NAAQS and are not observed at the fixed site monitors,
5) The relationship of fixed site monitors and/or personal dosimeter
readings to biological dose is unclear. Smoking has been shown to
interfere with any relationship that may exist between personal
dosimetry readings and biological dose. The studies on alveolar
"end-expired" breath samples and dosimetry readings are
contradictory. It is recommended that any furtner EPA work in this
area include a complete analysis of fixed site and personal dosimetry
readings, along with blood COHb and "end-expired" breath samples to
find the relationship between these parameters.
-------
-15-
References
1. U.S. Environmental Protection Agency, Air Quality Criteria for Carbon
Monoxide, EPA-600/8-79-022, Environmental Criteria and Assessment
Office, Office of Research and Development, Washington, D.C., October,
1979.
2. A.D. Cortese, and J.D. Spengler, "Ability of Fixed Monitoring Stations
to Represent Personal Carbon Monoxide Exposure", J. Air Poll. Control
Assoc., ^6, 12, 1144-1150, 1976.
3. J.W. Jabara, T.J. Keefe, H.J. Beaulieu and R.M. Buchan, "Carbon
Monoxide Dosimetry in Occupational Exposures in Denver, Colorado",
Arch. Env. Health, 15. 198-204, 1980.
4. L.A. Wallace, "Use of Personal Monitor to Measure Commuter Exposure to
Carbon Monoxide in Vehicle Passenger Compartments", Paper No.
79-59.2. Presented at the 72nd Annual Meeting of the Air Pollution
Control Association, Cincinnati, Ohio, June, 1979.
5. U.S. Environmental Protection Agency, Carbon Monoxide Study, Seattle,
Washington, October 6 - November 2, 1977, Prepared by C.B. Wilson and
J.W. Schweiss, EPA 910/9-78-054. U.S. Environmental Protection
Agency, Region X, Seattle, Washington, December, 1978.
6. U.S. Environmental Protection Agency, Carbon Monoxide Study, Boise,
Idaho, November 25 - December 22, 1977. Prepared by C.B. Wilson and
J.W. Schweiss, EPA 910/9-78-055, U.S. Environmental Protection Agency,
Region X, Seattle, Washington, December, 1978.
7. R.H. Jones, M.R. Elliott, J.B. Cadigan and E.A. Gaensler, "The
Relationship Between Alveolar and Blood Carbon Monoxide Concentrations
During Breath Holding: Simple Estimation of COHb Saturation", J. Lab.
Clin. Med., _5_1, 553-564, 1958.
8. W. Ott and R. Eliasson, "A Survey Technique for Determining the
Representativeness of Urban Air Monitoring Stations with Respect to
Carbon Monoxide", J. Air. Poll. Control Assoc., 23, 8, 685-690, 1973.
9. G. Godin, G.R. Wright and R.J. Shephard, "Urban Exposure to Carbon
Monoxide", Arch. Environ. Health, 25, 305, 1972.
10. G.R. Wright, S. Jewczyk, J. Onrot, P. Tomlison and R. Shepherd,
"Carbon Monoxide in the Urban Atmosphere: Hazards to the Pedestrian
and Street Workers", Arch. Environ. Health, 30, 123, 1975.
11. Analysis of Pollutant and Meteorological Data Collected in the
Vicinity of Carbon Monoxide "Hot Spots", SRI International Discussion
Draft Report for EPA Contract No. 68-03-2545, May, 1979.
-------
-16-
12. W.B. Johnson, W.F. Dabberdt, F.L. Ludwig and R.J. Allen, "Field Study
for Initial Evaluation of an Urban Diffusion Model for Carbon
Monoxide", Report for Contract CAPA 3-68 (1-69) prepared by Stanford
Research Institute for the Coordinating Research Council and the U.S.
Environmental Protection Agency, 1971.
13. R.M. Bruce and J.F. Roessler, "The Exposure to Carbon Monoxide of
Occupants of Vehicles Moving in Heavy Traffic", J. Air Poll. Control
Assoc., 16, 597, 1966.
14. I.F.Goldstein and L. Lanovitz, "Analysis of Air Pollution Patterns in
New York City - I. Can One Station Represent the Large Metropolitan
Area?", Atmos. Envir., 11, 47-52, 1977.
-------
36
32
28
24
o
o
20
OL
12
8
O1
^ Control Dosimeter ppm CO/hr
* Subject Dosimeter ppm CO/hr
n Ambient ppm CO/hr
« * •
N
8 10 12 14 16
Day Sampled
13
20 22 24 2f
Figure 1. Average Dally Dosimeter Readings for Subjects and Controls Compared
To Ambient CO Concentrations for Denver Study (From Reference 3)
-------
, 30 h
• 25h
Conc*nlra!lorn In City
During Commuting Hours
I .13 h
—- Carbon Monoxld«
Concentritlcni In Vahicla
» , _
789 10)1112i13|l4|l6ilQ 17 W\1Z\202J22 23 / 3 4- 6l6 7)8 6 101112131141516
""
Figure 2. CO Concentrations in Vehicle Compared to Ambient CO Levels
in Washington, D.C. (From Reference 4)
------- |