Estimating Ozone Exposures for the General Population in Eight Urban Area Using a Probabilistic Version of NEM


ESTIMATING OZONE EXPOSURES FOR THE GENERAL

   POPULATION IN EIGHT URBAN AREAS USING A

          PROBABILISTIC VERSION OF NEM
                    Thomas McCurdy
              U.S. Environmental Protection Agency
               Research Triangle Park, NC 27711
                        1994

-------
                              DISCLAIMER

      The attached papers and memo were prepared by staff at the U.S. Environmental
Protection Agency. The opinions, findings, and conclusions expressed are those of the
author and are not necessarily those of the Environmental Protection Agency. This draft
document  is being circulated for review and public comments.  Comments should be
addresed to Mr.  Harvey Richmond, U.S.  EPA, Office of Air  Quality Planning and
Standards, MD-15, Research Triangle Park, North Carolina  27711; Telephone (919) 541-
5271 or  FAX (919) 541-0824.

-------
ESTIMATING OZONE EXPOSURES IN THE CHICAGO URBAN AREA USING A
       SECOND-GENERATION PROBABILISTIC VERSION OF NEM
                           Thomas McCurdy
                   U.S. Environmental Protection Agency
                    Research Triangle Park, NC 27711
                             March 1994
File: CHICAGO.

-------
INTRODUCTION

The U S Environmental Protection Agency (EPA) currently is reviewing the ozone
(00 primary national ambient air quality standard (NAAQS) to determine if the existing
standard protects public health with an adequate margin of safety. In order to evaluate
whether alternative O3 NAAQSs provide an adequate margin of safety, the EPA Office of
Air Quality Planning and Standards (OAQPS) assesses such factors as the nature and seventy
of health effects associated with O3 exposure, the degree of total human exposure (i.e.,
indoor and outdoor) to O3, and the risks (probabilities) of ozone-related health effects
occurring in the exposed population when alternative O3 NAAQS are just attained. This
paper discusses an EPA analysis of O3 exposure estimates for parts of the New York
metropolitan area which employed the pNEM/O3 model developed by International
Technology-Air Quality Services (IT-AQS). A report by Johnson, et al. (1) describes this
model in detail and presents initial results of applying it to nine U.S. urban areas, including
Chicago Exposure estimates for parts of the Los Angeles area are discussed in an EPA
memorandum (2) that was reviewed by the Clean Air Science Advisory Committee (CASAC)
at a public meeting. Exposure estimates for the Philadelphia area are described in an
AWMA paper (3). Exposure estimates for Houston are described in an AWMA specialty
conference paper (4). New York results are contained in an EPA paper (5). Additional
analyses of O3 exposure estimates for the remaining urban areas that were modeled will be
forthcoming.

OVERVIEW

Pursuant to section 108 of the Clean Air Act, the Environmental Criteria Assessment
Office (ECAO) within the Office of Research and Development (ORD) is preparing a criteria
document (6) for O3. This document will be a comprehensive summary and critical
assessment of the latest scientific research relevant to assessing the effects of a criteria
pollutant. The document will be reviewed by CASAC, part of the EPA's independent
Science Advisory Board, and by the general public. Upon completion of the criteria
document OAQPS will prepare a Staff Paper that summarizes and integrates scientific
information presented in the criteria document considered to be most relevant to decisions on
alternative NAAQS and various analyses (e.g., air quality, exposure, and risk).

An important aspect of health risk assessment is an analysis of population exposure
(7) Such an analysis provides the population basis for "headcount risk" estimates developed
as part of EPA's NAAQS risk assessment work. Estimates of population exposure also
provide important information regarding the number of people that may experience varying
levels of O3 as they go through their daily activities.

Since 1980, OAQPS has used the NAAQS Exposure Model (NEM) to analyze human
exposure associated with the "criteria air pollutants" covered by Title I of the Clean Air Act,
as amended. Recent versions of this model use Monte Carlo analyses to define numeric
values for certain variables inherent in the model. These version of NEM are called pNEM,

-------
an acronym for "probabilistic" NEM. To differentiate among the various pollutant-specific
versions of pNEM, the one described here is denoted as pNEM/O3.

Analysis of population exposure under alternative O3 NAAQS requires that significant
factors contributing to total human exposure be taken into account. These factors include the
temporal and spatial distribution of people and O3 concentrations throughout an urban area,
the variation of ozone levels within each microenvironment, and the effects of exercise
(increased ventilation) on ozone uptake in exposed individuals.

To date, the most recent version of pNEM/O3~the "second generation" version-has
been applied to nine major urban areas. These areas vary in size from Denver with 1.4
million people to New York with over 10.6 million people. pNEM/O3 partitions all land
within a selected urban area into large "exposure districts." The number of exposure
districts defined for each urban area varies from 6 to 16 in the set of nine areas recently
modeled; the number of districts reflects the number of monitors having valid air quality data
in a study area. Most of the nine urban areas have 10 or more districts within their
boundaries.

People living within each exposure district, as estimated by the U.S. Bureau of
Census in 1990, are assigned to a single discrete point, the population centroid. O3 air
quality within each exposure district is estimated for each hour of the O3 season using data
from a nearby monitoring site. Because O3 in the ambient air generally is modified
considerably when entering a building or vehicle, outdoor O3 estimates are adjusted using a
mass-balance model to account for 3 indoor microenvironments. These microenvironments
are: indoors-at-home, indoors-other, and within a motor vehicle. The mass-balance model
used in pNEM/O3 is a simplified version of the generalized Nagda, Rector, and Koontz
(NRK) model (8). For our use, this model was revised to incorporate the assumption that
indoor decay rate is proportional to indoor O3 concentration. It was further revised to
incorporate assumptions concerning ozone decay rates suggested by Weschler, et al (9). The
resulting revised mass-balance model is fully described in Ref. 1.

Because both the intake dose received and susceptibility to effects of pollution may
vary with age, occupation, and intensity of exercise, the total population of each study area is
divided into 9 age-occupation (A-0) groups. Each A-O group is further subdivided into
cohorts depending upon (1) the type of air conditioning system present in the home, if any,
and (2) home-to-work commuting patterns.

While many types of exposure metrics are provided by pNEM/O3 only those for the
non-outdoor working population in Chicago are discussed in this paper. Exposure estimates
for children are included in the results discussed here. Estimates for outdoor workers—a
major population sub-group of concern from an O3 exposure perspective-will be completed
in the Fall of 1994.

-------
It should be noted that all versions of pNEM systematically underestimate the highest
part of the exposure distribution. (See Ref. 1 for an evaluation of pNEM/O3 results against
real-time personal exposure monitoring data.) Thus, results presented here have a downward
bias.

AIR QUALITY SCENARIOS INVESTIGATED

Six air quality scenarios were modeled for the Chicago urban area (and the other 8
urban areas); the scenarios are all on a daily maximum basis. They are:

1. 1991 air qualiry-the "as is" scenario
2. Attain Ih lEx 0.12 ppm-the current standard
3. Attain 8h lEx 0.08 ppm
4. Attain 8h lEx 0.10 ppm
5. Attain 8h 5Ex 0.06 ppm
6. Attain 8h 5Ex 0.08 ppm

Ex stands for "expected exceedances," as defined by EPA in 40 Code of Federal Regulations
50.9, and h is the usual symbol for hour(s). Thus, a full description of the first alternative
NAAQS scenarios is: attain the 1 hour daily maximum-one expected exceedance NAAQS of
0.12 ppm. The other standards follow the same format. The 5Ex NAAQS would allow 5
daily exceedances per year, on average, if they were implemented.

The iterative adjustment procedure used by IT-AQS to simulate just-attaining the 5
alternative NAAQS listed above involves a complex procedure of fitting Weibull equations to
the 1991 O3 data, modifying parameters of the equations so that the "worst" monitor just
achieves the desired standard, and then re-specifying the entire O3 pattern for all districts in
Chicago. The procedure is described in detail by Johnson, et al. (1).

CHICAGO RESULTS

Results are based upon 10 runs (realizations) of the pNEM/O3 model for each air
quality scenario. Based on a previous analysis of sets of 10-run results versus a 108-run
result (10), the author believes that results from only 10 runs of the model will adequately
predict the mean and variance observed in 100 or more runs of pNEM/O3. Ten runs of the
model will not, however, adequately represent the entire range of possible outcomes.
Additional runs can only increase this range. This is an obvious shortcoming of the analyses
presented here, but limited resources preclude undertaking more runs.

-------
hundred person-occurrence could be 10 people experiencing the exposure 10 times, 100
people experiencing it once, one person experiencing it 100 times, or any combination
thereof.) Even though data are discussed here regarding the absolute and relative number
(percentage) of people exposed, it must be noted that "people" should not be used as an
exposure metric. The cohort-specific activity-days data base used for pNEM/O3 modeling
(and for aU other exposure models that use human activity information) provides no more
than three days of data for a single individual. Thus, rigorous inferences from these data can
be made only to "person-days" of exposure or "person-occurrences" of exposure, where an
occurrence can be Ih, 8h, or 24h time period in pNEM/O3. Only when activity data are
available sequentially for an entire year for individuals in a cohort can inferences be made to
people exposed over the year (or to children, or to outdoor workers as the case may be).
This point usually is ignored in the general exposure assessment literature.

All exposure estimates presented in this paper pertain to "daily maximum dose,"
(MAXD) where dose is defined as the product of ozone concentration and ventilation rate
over a defined time period. Note that maximum daily dose does not necessarily occur during
the time period of maximum ozone concentration in a given urban area.

Ventilation rate is a very important factor in estimating total dose for any air pollutant
(11). The ventilation rate metric used in pNEM/O3 is "equivalent ventilation rate" (EVR),
or ventilation rate (in liters per minute) divided by body surface area (BSA, in units of m2).
EVR has units of 1 min"1 m"2. It is a surrogate indicator for the number of O3 molecules that
enter the oral-nasal cavities per unit time period.

The Chicago urban area population analyzed here includes
6.2 million NOW (non-outdoor worker) people and 1,156,500 children. There are 1,321
million possible daily maximum NOW person-occurrences (NOW people x 214 days in the
O3 season). The equivalent figure for children is 247,491,000.

Exposure results appear in Tables 1 and 2. Table 1 includes Ih daily maximum dose
exposure estimates, while Table 2 provides similar results for 8h exposure estimates.

The results indicate that:

1. Currently (1991), fewer than 0.5% of the NOW population experience a daily
maximum dose (MAXD) O3 exposure @ >0.12 ppm @ an EVR of 30 1 min"1 m"2 or
higher, which is thought by many researchers to be a dose rate of concern (12). All
of these exposures occur only on one day during the O3 season.

2. Currently (1991), about 2% of the children in Chicago exceed the 0.12 ppm/30 EVR
max dose exposure level. This is about 80% of the NOW exposed population. All of
the children exposed at that level are predicted to experience only 1 day/season.

-------
3 Such MAXD exposures are reduced whenever one of the alternative standards are
attained. For the relatively few people expected to have a MAXD exposure >0.12
ppm @ >i30 EVR when an alternative NAAQS is attained, all of them are predicted
to experience only one such daily Ih exposure per season.

While not shown in Table 1, the results of O3 exposures >0.12 ppm at any. exercise
level are considerably higher than the numbers presented in that Table. For instance,
currently (1991) about one million adults and 312,500 children are predicted to experience
one or more daily exposure >0.12 ppm when exercise level is ignored. This is about 16%
of the adult population and 27% of the children in the Chicago urban area. Thus, exercise
level, and its associated ventilation rate, at maximum dose greatly affects pNEM/O3 MAXD
exposure estimates.

The same general findings apply to the 8h results (Table 2), although relatively more
people may be affected. The level of concern here is a MAXD exposure of >0.08 ppm for
8h @ an EVR >_15 1 min'1 m'2. (Obviously, the EVR of concern drops as the averaging
period increases, just as the O3 concentration of concern drops. People cannot hold an EVR
of 30 1 min'1 m"2 for an extended period of time(ll).)

The 8h results indicate that:

1. Currently (1991), about 1% of the NOW population are estimated to experience a
daily maximum dose O3 exposure level of concern. About 3% of these people may
experience such a MAXD exposure on more than one day during the O3 season; no
one is estimated to experience this exposure on more than two days.
2. Currently (1991), over 4% of children in the Chicago area may experience the 8h
MAXD of concern. In fact, about 83% of the total population experiencing this 8h
MAXD are children. About 97% of the children may experience this MAXD on
only one day, while 3% may experience it on 2 days.

3 Attaining any of the alternative NAAQS reduces the 8h MAXD exposure of concern,
although almost 4% of children may experience that exposure if the 8h lEx 0.10 ppm
NAAQS is attained (5% of them on 2 days). Attaining the current NAAQS (Ih lEx
0.12 ppm) may also result in almost 4% of children experiencing the 8h MAXD of
concern one or two times during the O3 season in Chicago. In addition, attaining the
8h 5Ex NAAQS of 0.08 ppm is estimated to leave a small percentage (0.1%) of
children at or above the 8h MAXD of concern-all on one day, however.

Consistent with the Ih results, it is estimated that approximately 32% in the NOW
population of Chicago currently (1991) experiences one or more 8h exposures at maximum
daily dose when ventilation level is ignored. (This result is not shown in Table 2.) This
contrasts with the 1% of the NOW population who may experience the same 8h O3
concentration at exercise. Again, exercise levels-expressed as EVR-greatly affect
pNEM/O3 exposure results. This same pattern holds true for children in Chicago, but the

-------
percentage is higher: 50% of all children may experience an 8h 0.08 ppm exposure if
breathing rate is ignored, as compared to 4% at exercise.

Note that all of the estimates have variability, reflecting the Monte Carlo sampling
process that produces a different estimate for each run. The range in the estimates are
between a factor of 2-to-3 in most cases, which seems reasonable.

Statistical tests of the exposure distributions were undertaken using a two-sample
Kolmogorov-Smirnov (K-S) non-parametric test. The null hypothesis was that the cumulative
frequency exposure distributions for the various air quality scenarios are the same for each
exposure metric being tested. The test statistic was the maximum difference between the two
distributions (Dn); it was evaluated using a large-sample F statistic. The significance level is
0.05, using the chi-square approximation. The test indicates that the exposure indicators
shown in Tables 1 and 2 vary significantly across the six air quality scenarios with the
following exceptions.

For the Asls and 8h lEx 0.10 ppm pair:

NOW people exposed, NOW person-occurrences, and number of people exposed on
one day/ozone season at Ih 0.12 ppm @ exercise and at 8h 0.08 ppm @ exercise and
at any exercise level.

Children exposed, children occurrences, and number of children exposed on one
day/season at Ih 0.12 ppm @ exercise and at 8h 0.08 ppm @ exercise and at any
exercise level.

For the Asls and current NAAQS (Ih lEx 0.12 ppm) pair:

NOW people and children, NOW person- and children-occurrences, and number of
people and children exposed on one day/O3 season to 8h 0.08 ppm @ exercise and at
any exercise level.

For the current standard (Ih lEx 0.12 ppm) and 8h lEx 0.08 ppm pair:

NOW people exposed, NOW person-occurrences, and number of people exposed on
one day/O3 season at Ih 0.12 ppm or 8h 0.08 ppm @ exercise and at any exercise
level.

Children exposed, children-occurrences, and number of children exposed on one
day/O3 season at Ih 0.12 ppm for any exercise level and at 8h 0.08 ppm @ exercise
and at any exercise level.

Note that these are not strong statistical tests and that some of the K-S test
requirements were not met. Random sampling was not used. The sample "observations"

-------
simply are results of a simulation model; the population of interest, then, can only be the set
of possible outcomes produced by the model.  A large-sample test statistic  was applied to
small samples (10 runs).  These caveats aside, 43 "cannot rejects" out of 118 tests is three
times as many as are expected based on a "pure-chance" probability of rejecting a true
hypothesis (5%)  This 36% cannot-reject rate is more than quadruple similar rates seen in
most of the urban areas that were analyzed (2-5). The conclusions that have to be reached in
the Chicago urban area are:  (1) attaining a Ih lEx NAAQS of 0.12 ppm-the current
standard-may not  greatly affect exposure @ exercise for non-outdoor workers  or children,
(2) attaining an 8h lEx NAAQS of 0.10 ppm may also not greatly affect exposure @
exercise for the  same two groups,  and  (3) the current standard cannot be distinguished from
an 8h lEx 0.08  ppm standard on the basis of exposures  reduced.


SUMMARY
       The analyses of 5 alternative standards ozone NAAQS in the Chicago area (Ih lEx
0 12 ppm-  8h lEx 0.10 pm; 8h 5Ex 0.08 ppm; 8h lEx 0.08 ppm; and 8h 5Ex 0.06 ppm)
indicate that attainment of the current standard or the 8h lEx 0.10 ppm alternative  may not
significantly reduce exposures at exercise in the Chicago urban area.  This finding  applies to
both the Ih and 8h estimates, but are more important for the 8h results since more people
and children are currently estimated to experience 8h MAXD levels of concern than the Ih
MAXD levels.

       The reader should remember the early caveat about the  downward bias inherent in
 pNEM results.  While every effort has been made to present "best estimates," it is  inherently
 difficult to predict the upper end of the population exposure distribution using human activity
 data bases that consist of only 1-3 days of information and that also under-represent adults
 and children who exercise  regularly. Year-long activity data from all segments  of the
 population are needed to adequately estimate the true population exposure distribution.

-------
                                 REFERENCES

1     T. Johnson, J. Capel, and M. McCoy.  Estimation of Ozone Exposures
      Experienced by Urban Residents Using a Pro-
      ballistic Version of NEM and 1990 Population Data.  Durham, NC: IT-AQS,
      1993 (Draft).

2.    T. McCurdy.  "Los Angeles Ozone Exposure Estimates," EPA Memorandum;
      November 4, 1993.

3     T McCurdy et al  "Estimating Ozone Exposures in Philadelphia Using a Second-
      Generation Probabilistic Version of NEM." Paper A210 presented at the 87th
      Annual Meeting of AWMA; Cincinnati, June 1994.

4    T McCurdy et al  "Estimating ozone Exposures in Houston Using  a Second-
      Generation Probabilistic Version of NEM." Paper presented at the Troposphenc
      Ozone Specialty Conference, Air and Waste Management Association; Orlando,
       1994.

 5    T McCurdy   "Estimating Ozone  Exposures in the New York Urban Area Using a
       Second-Generation Probabilistic Version of NEM."  EPA Paper, February  1994.

 6     Environmental Criteria and Assessment  Office.  Air Quality Criteria for Ozone and
       Photochemical Oxidants.  Research Triangle Park, NC: 1994 (Draft).

 7     H M  Richmond and T. McCurdy. "Use of Exposure Analysis and Risk Assessment
       in the Ozone NAAQS Review."  Paper 88-121.3 presented at the  81st Annual
       Meeting of APCA; Dallas, June 19-24,  1988.

  8.     N.L. Nagda, H.E. Rector, and M.D. Koontz. Guidelines  for Monitoring Air
        Quality. Washington,  DC: Hemisphere Pub. Co., 1987.

  9      C J  Weschler, H.C. Shields, and D.V.  Nike.  "Indoor Ozone: Recent Findings," in:
        R. Burglund (ed.), Tropospheric Ozone in the Environment II.  Pittsburgh.  Air &
        Waste Management Association,  1992.

  10.    T. McCurdy. "Testing pNEM/O3 Runs to Determine if a  Set of 10 Runs is
        'Representative'," EPA Memorandum; April 19, 1994.

  11    W.C.  Adams. Measurement of Breathing Rate and Volume in Routinely
        Performed Daily Activities.  Davis C A:  University of California, 1993.

  12    M  Absil P  Narducci, R. Whitfield, and H.M. Richmond.  "Chronic Lung Injury
        Risk Estimates for Urban Areas Having Ozone Patterns Similar to Those in the

-------
Northeast." Paper presented at the AWMA Specialty Conference on Tropospheric
Ozone and the Environment II; Atlanta, November 4-7, 1991.
                                 10

-------
                                     Table I. ESTIMATES OF Ih "MAXIMUM DOSAGE" EXPOSURES IN CHICAGO TO O3 CONCENTRATIONS
                     EXCEEDING 0.12 PPM AND EVR" EQUALED OR EXCEEDED 30 ONE OR MORE TIMES PER OZONE SEASON UNDER SIX AIR QUALITY SCENARIOS
Statistic
TOTAL EXPOSED POPULATION
Mean Estimate of the Number of People
Percent of Total Population
Range in this percentage for all runs
Mean Est. of Person-Occurrences
Percent of Total Per-Occurrences
Range in this percentage for all runs
Mean Est. of Occurrences/Person Exposed
Number of Days/Season Exposed (% Exposed)
1 Day
2 Days
3 Days
>3 Days
CHILDREN
Mean Estimate of the Number of Children
Percent of Children Population
Range in this percentage for all runs
Percent of Total Exposed Population
Mean Estimate of Children-Occurrences
Percent of total Child-Occurrences
Range in this percentage for all runs
Mean Estimate of Occurrences/Child Exposed
Number of Days Exposed/Season (% Exposed)
1 Day
2 Days
3 Days
>3 Days
As Is
Situation
(199.1)

27,000
0.4
*-2.0
27,000
@
@-@
1.00
100
0
0
0

21,700
1.9
0.0-10.5

80.4
21,700
@
@-@
1.00

100
0
0
0
Attain
8h lEx
0.10 ppm

3,500
0.1
0.0-0.2
3,500
@
@-@
1.00
100
0
0
0

2,900
0.3
0.0-1.1

82.9
2,900
®
@-@
1.00

100
0
0
0
Attain
8h5Ex
0.08 ppm

0
0
-
0
0
-
-
-
-
-
-

0
0
-

-
0
0
-
-

-
-
-
-
Attain
Ih 1 Ex
0.12 ppm

10,300
0.2
0.0-1.7
10,300
@
@-@
1.00
100
0
0
0

10,200
0.9
0.0-8.8

99.0
10,200
@
@-@
1.00

100
0
0
0
Attain
8h 1 Ex
0.08 ppm

0
0
-
0
0
-
-
-
-
-
-

0
0
-

-
0
0
-
-

-
-
-
-
Attain
8h 5 Ex
0.06 ppm

0
0
-
0
0
-
-
-
-
-
-

0
0
-

-
0
0
-
-

-
-
-
-
Notes:HEquivalent ventilation rate.  It is equal to ventilation rate/body surface area; units are in 1  mm"1 m"
    *Less than 0.05% but > 0.005%
    @Less than 0.005% but >0.

-------
                                      Table 2. ESTIMATES OF 8h "MAXIMUM DOSAGE' EXPOSURES IN CHICAGO TO O3 CONCENTRATIONS
                EXCEEDING 0.08 PPM AND EVR1 EQUALED OR EXCEEDED 15 ONE OR MORE TIMES PER
                         OZONE SEASON UNDER SIX AIR QUALITY SCENARIOS
Statistic
TOTAL NON-OUTDOOR WORKER EXPOSED POPULATION
Mean Estimate of the Number of People
Percent of Total Population
Range in this percentage for all runs
Mean Est. of person-Occurrences
Percent of total Per.-Occurrences
Range in this percentage for all runs
Mean Est. of Occurrences/Person Exposed
Number of Days/Season Exposed (% Exposed)
1 Day
2 Days
3 Days
4 Days
>4 Days
CHILDREN
Mean Estimate of the Number of Children
Percent of Children Population
Range in this percentage for all runs
Percent of Total Exposed Population
Mean Estimate of Children-Occurrences
Percent of Total Child-Occurrences
Range in this percentage for all runs
Mean Estimate of Occurrences/Child Exposed
Number of Days/Season (% Exposed)
1 Day
2 Days
3 Days
4 Days
>4 Dnys
As Is
Situation
(1991)

60,400
1.0
*-2.0
62,200
@
@-@
1.02

97
3
0
0
0

50,400
4.4
0.0-9.0
83.4
52,200
*
0.0-*
1.04

97
3
0
0
0
Attain
8h lEx
O.lOppm

45,100
0.7
0.2-1.5
47,200
@
@-@
1.05

95
5
0
0
0

41,800
3.6
1.0-7.0
92.7
43,900
*
@-*
1.05

95
5
0
0
0
Attain
8h5Ex
0.08 ppm

3,200
0.1
0.0-0.3
3,200
@
@-@
1.00

100
0
0
0
0

2,100
0.1
0.0-0.8
65.6
2,100
®
0.0-@
1.00

100
0
0
0
0
Attain
Ih 1 Ex
0.12 ppm

49,500
0.8
0.5-1.2
50.800
0.8
0.4-1.2
1.03

97
3
0
0
0

42,000
3.6
2.1-5.6
84.8
43,300
*
*_*
1.03

97
3
0
0
0
Attain
8h 1 Ex
0.08 ppm

0
0
-
0
0
-
-

-
-
-
-
-

0
0
-
-
0
0
-
-

-
-
-
-
-
Attain
8h5 Ex
0.06 ppm

0
0
~
0
0
-
-

-
-
-
-
-

0
0
-
-
0
0
-
-

-
-
-
-
-
Notes:  ''Liquivalcnt ventilation rate.  It is equal to ventilation rate/body surface area; units arc in I   mm '  m
          "Less than 0.05% but  > 0.005%
          ©Less than O.Q05% butt>0.

-------
ESTIMATING OZONE EXPOSURES IN THE DENVER URBAN AREA USING A
      SECOND-GENERATION PROBABILISTIC VERSION OF NEM
                           Thomas McCurdy
                    U.S. Environmental Protection Agency
                     Research Triangle Park, NC 27711
                               March 1994
  File: DENVER.

-------
INTRODUCTION

The U.S. Environmental Protection Agency (EPA) currently is reviewing the ozone
(O3) primary national ambient air quality standard (NAAQS) to determine if the existing
standard protects public health with an adequate margin of safety. In order to evaluate
whether alternative O3 NAAQSs provide an adequate margin of safety, the EPA Office of
Air Quality Planning and Standards (OAQPS) assesses such factors as the nature and severity
of health effects associated with O3 exposure, the degree of total human exposure (i.e.,
indoor and outdoor) to O3, and the risks (probabilities) of ozone-related health effects
occurring in the exposed population when alternative O3 NAAQS are just attained. This
paper discusses an EPA analysis of O3 exposure estimates for parts of the Denver
metropolitan area which employed the pNEM/O3 model developed by International
Technology-Air Quality Services (IT-AQS). A report by Johnson, et al. (1) describes this
model in detail and presents initial results of applying it to nine U.S. urban areas, including
Denver. Exposure estimates for parts of the Los Angeles area are discussed in an EPA
memorandum (2) that was reviewed by the Clean Air Science Advisory Committee (CASAC)
at a public meeting. Exposure estimates for the Philadelphia area are described in an
AWMA paper (3). Exposure estimates for Houston are described in an AWMA specialty
conference paper (4). New York and Chicago results are contained in EPA papers (5,6).
Additional analyses of O3 exposure estimates for the remaining urban areas that were
modeled will be forthcoming.

OVERVIEW

Pursuant to section 108 of the Clean Air Act, the Environmental Criteria Assessment
Office (ECAO) within the Office of Research and Development (ORD) is preparing a criteria
document (7) for O3. This document will be a comprehensive summary and critical
assessment of the latest scientific research relevant to assessing the effects of a criteria
pollutant. The document will be reviewed by CASAC, part of the EPA's independent
Science Advisor}' Board, and by the general public. Upon completion of the criteria
document, OAQPS will prepare a Staff Paper that summarizes and integrates scientific
information presented in the criteria document considered to be most relevant to decisions on
alternative NAAQS and various analyses (e.g., air quality, exposure, and risk).

An important aspect of health risk assessment is an analysis of population exposure
(8). Such an analysis provides the population basis for "headcount risk" estimates developed
as part of EPA's NAAQS risk assessment work. Estimates of population exposure also
provide important information regarding the number of people that may experience varying
levels of O3 as they go through their daily activities.

-------
values for certain variables inherent in the model. These version of NEM are called pNEM,
an acronym for "probabilistic" NEM. To differentiate among the various pollutant-specific
versions of pNEM, the one described here is denoted as pNEM/O3.

To date, the most recent version of pNEM/O3—the "second generation" version—has
been applied to nine major urban areas. These areas vary in size from Denver with 1.4
million people to New York with over 10.6 million people. pNEM/O3 partitions all land
within a selected urban area into large "exposure districts." The number of exposure
districts defined for each urban area varies from 6 to 16 in the set of nine areas recently
modeled; the number of districts reflects the number of monitors having valid air quality data
in a study area. Most of the nine urban areas have 10 or more districts within their
boundaries.

People living within each exposure district, as estimated by the U.S. Bureau of
Census in 1990, are assigned to a single discrete point, the population centroid. O3 air
quality within each exposure district is estimated for each hour of the O3 season using data
from a nearby monitoring site. Because O3 in the ambient air generally is modified
considerably when entering a building or vehicle, outdoor O3 estimates are adjusted using a
mass-balance model to account for 3 indoor microenvironments. These microenvironments
are: indoors-at-home, indoors-other, and within a motor vehicle. The mass-balance model
used in pNEM/O3 is a simplified version of the generalized Nagda, Rector, and Koontz
(NRK) model (9). For our use, this model was revised to incorporate the assumption that
indoor decay rate is proportional to indoor O3 concentration. It was further revised to
incorporate assumptions concerning ozone decay rates suggested by Weschler, et al (10).
The resulting revised mass-balance model is fully described in Ref. 1.

Because both the intake dose received and susceptibility to effects of pollution may
vary with age, occupation, and intensity of exercise, the total population of each study area is
divided into 9 age-occupation (A-O) groups. Each A-O group is further subdivided into
cohorts depending upon (1) the type of air conditioning system present in the home, if any,
and (2) home-to-work commuting patterns.

While many types of exposure metrics are provided by pNEM/O3 only those for the
non-outdoor working population in Denver are discussed in this paper. Exposure estimates
for children are included in the results discussed here. Estimates for outdoor workers—a
major population sub-group of concern from an O3 exposure perspective-will be completed
in the Fall of 1994.

AIR QUALITY SCENARIOS INVESTIGATED

Six air quality scenarios were modeled for the Denver urban area (and the other 8
urban areas); the scenarios are all on a daily maximum basis. They are:

1. 1990 air quality-the "as is" scenario
2. Attain Ih lEx 0.12 ppm--the current standard
3. Attain 8h lEx 0.08 ppm
4. Attain 8h lEx 0.10 ppm
5. Attain 8h 5Ex 0.06 ppm
6. Attain 8h 5Ex 0.08 ppm

The iterative adjustment procedure used by IT-AQS to simulate just-attaining the 5
alternative NAAQS listed above involves a complex procedure of fitting Weibull equations to
the 1990 03 data, modifying parameters of the equations so that the "worst" monitor just
achieves the desired standard, and then re-specifying the entire O3 pattern for all districts in
Denver. The procedure is described in detail by Johnson, et al. (1).

Denver had relatively good O3 air quality in 1990 in that it had few days with a Ih
daily maximum concentration >0.11 ppm. The area's "design value"-the second-highest Ih
daily value measured at any site in the urban area-was 0.115, which is lower than the
current standard. Denver's 1990 monitored 8h values were also correspondingly low. Thus,
just-attaining some of the alternatives listed above can result in higher Ih and 8h O3 peak
concentrations being included than were monitored in 1990. Exposure estimates for these
alternatives similarly can be higher than those predicted for 1990. These results follow
directly from the logic used to evaluate alternative NAAQS (8), and are quite feasible in
reality.

DENVER RESULTS

Results are based upon 10 runs (realizations) of the pNEM/O3 model for each air
quality scenario. Based on a previous analysis of sets of 10-run results versus a 108-run

-------
result (11), the author believes that results from only 10 runs of the model will adequately
predict the mean and variance observed in 100 or more runs of pNEM/O3. Ten runs of the
model will not, however, adequately represent the entire range of possible outcomes.
Additional runs can only increase this range. This is an obvious shortcoming of the analyses
presented here, but limited resources preclude undertaking more runs.

This paper includes tabular data on (1) the number and percent of people and children
exposed, and (2) the number and percent of person-occurrences for children and the adult,
non-outdoor worker population in Denver. (A person-occurrence is the metric that focuses
on an event: when the exposure-of-concern is reached regardless of who reaches it. One
hundred person-occurrence could be 10 people experiencing the exposure 10 times, 100
people experiencing it once, one person experiencing it 100 times, or any combination
thereof.) Even though data are discussed here regarding the absolute and relative number
(percentage) of people exposed, it must be noted that "people" should not be used as an
exposure metric. The cohort-specific activity-days data base used for pNEM/O3 modeling
(and for all other exposure models that use human activity information) provides no more
than three days of data for a single individual. Thus, rigorous inferences from these data can
be made only to "person-days" of exposure or "person-occurrences" of exposure, where an
occurrence can be Ih, 8h, or 24h time period in pNEM/O3. Only when activity data are
available sequentially for an entire year for individuals in a cohort can inferences be made to
people exposed over the year (or to children, or to outdoor workers as the case may be).
This point usually is ignored in the general exposure assessment literature.

Ventilation rate is a very important factor in estimating total dose for any air pollutant
(12). The ventilation rate metric used in pNEM/O3 is "equivalent ventilation rate" (EVR),
or ventilation rate (in liters per minute) divided by body surface area (BSA, in units of m2).
EVR has units of 1 min"1 m"2. It is a surrogate indicator for the number of O3 molecules that
enter the oral-nasal cavities per unit time period.

The Denver urban area population analyzed here includes
1.5 million NOW (non-outdoor worker) people and 260,000 children. There are 317.7
million possible daily maximum NOW person-occurrences (NOW people x 214 days in the
O3 season). The equivalent figure for children is 55.6 million.

Exposure results appear in Tables 1 and 2. Table 1 includes Ih daily maximum dose
exposure estimates, while Table 2 provides similar results for 8h exposure estimates.

The Ih results indicate that:

-------
1 Currently (1990), it is estimated that no one in Denver would experience a daily
maximum dose (MAXD) O3 exposure @ >0.12 ppm @ an EVR of 30 1 min'1 nV or
higher, which is thought by many researchers to be a dose rate of concern (12).

2. One hour MAXD exposures of concern are estimated to increase if two of the
alternative NAAQS standards that were analyzed were just attained. The two
NAAQS are the current O3 standard (Ih lEx 0.12 ppm) and an 8h lEx 0.10 ppm.
Just-attaining the latter NAAQS shows more exposures, but the percent of the NOW
population exposed is quite low: <0.1%. The percentage of children so exposed to
the Ih MAXD of >0.12 ppm @ _>_30 EVR is higher however; it is estimated to be
1.1%. The reason for the difference in relative magnitude is that over 93% of people
exposed at the MAXD of concern are children. (At the current Ih NAAQS, only
12.3% of MAXD exposures are attributed to children so their relative impacts are
small: <0.5%.)

3. None of the Ih MAXD exposures of concern occur on more than one day/O3 season
for the two alternative NAAQS that are estimated to have them.

While not shown in Table 1, the results of O3 exposures >0.12 ppm at any exercise
level are considerably higher than the numbers presented in that Table. For instance,
currently (1990) about 61,000 adults and 19,000 children are predicted to experience one or
more daily exposure >0.12 ppm when exercise level is ignored. This is about 5% of the
adult population and 7% of the children in the Denver urban area. (As compared to no adult
or children exposure @ 30 EVR exercise.) The corresponding estimates for non-exercising
exposure when the current standard is just attained is 85,000 (7%) for adults and 28,000 for
children (11%). Similar estimates for attainment of the 8h lEx 0.10 ppm NAAQS
alternative are 118,000 (10%) for adults and 72,000 for children (28%). Both of the just-
attaining analyses for these two alternatives indicate that exposures >0.12 ppm @ any
exercise level would occur on more than 1 day/season, unlike the situation that occurs when
exercise @ 30 EVR is considered. Thus, exercise level, and its associated ventilation rate, at
maximum dose greatly affects pNEM/O3 MAXD exposure estimates.

The 8h results (Table 2), indicate that relatively more people may be exposed. The
level of concern here is a MAXD exposure of >0.08 ppm for 8h @ an EVR >J5 1 muV
nV2. (Obviously, the EVR of concern drops as the averaging period increases just as the O3
concentration of concern drops. People cannot hold an EVR of 30 1 min'1 nV for an
extended period of time(12).)

The 8h results indicate that:

1. Currently (1990), < 1% of the NOW population are estimated to experience a daily
maximum dose O3 exposure level of concern. All of these exposures are estimated to
occur on 1 day during the O3 season, and all of them involve children.

-------
2. Attaining any of the alternative NAAQS-except the 8h 5Ex 0.06 ppm~still results in
some people being exposed to the 8h MAXD of concern. The absolute and relative
numbers of people involved are quite small for the 8h lEx 0.08 ppm and 8h 5Ex 0.08
ppm alternatives, however. In both cases, about 0.1% of NOW people are involved,
and 0.2-0.3% of the children. It should be noted that only one run of the 10
undertaken for each of these two standards had non-zero exposures; thus, their
estimates should be viewed with caution.

3. The absolute and relative number of NOW people and children estimated to be
exposed to the 8h MAXD of concern is rather large for the 8h lEx 0.10 ppm and Ih
lEx 0.12 ppm NAAQS alternatives. For the Ih alternative-the current standard--
over 1% of the NOW population and almost 7% of children may receive an 8h
MAXD of concern, over 10% of them receiving two days of this exposure during the
O3 season. The absolute estimates and relative proportion of people and children are
even higher for the 8h 1 Ex 0.10 ppm alternative. It is estimated that 3.2% of NOW
people and 13.4% of children will experience the 8h MAXD of concern for this
standard when it is just attained. About 8-9% of these exposures may occur on 2
days during the O3 season.

Putting the 8h MAXD estimates into perspective, the results indicate that currently
(1990) over 117,000 NOW people might experience an 8h exposure >0.08 ppm @ any
exercise level. (In other words, when exercise is ignored; this estimate is not shown in
Table 2.) This contrasts to the 100 NOW people estimated to experience the same 03 air
quality concentration @ 15 EVR or higher. The corresponding estimates for children are
38,600 @ any exercise level versus 100 @ 15+ EVR. Obviously, exercise levels-expressed
as EVR~greatly affect 8h pNEM/O3 exposure estimates.
are:
For the record, the 8h any-exercise estimates for the non-zero alternative NAAQS
8h lEx
0.10
623,300
42.0
142,200
54.7
8h 5Ex
0.08
31,200
2.1
10,400
4.0
Ih lEx
0.12
489,800
33.0
132,900
51.1
8h lEx
0.08
44,500
3.0
10,600
4.1
NOW People
(%)

Children
(%)

Contrast these estimates with those appearing in Table 2 for insight into the impact of
exercise level on exposure estimates. The any-exercise estimates are 1-2 orders of magnitude
higher than the _>.15 EVR estimates.

-------
Note that all of the estimates have variability, reflecting the Monte Carlo sampling
process that produces a different estimate for each run. The range in the estimates are
between a factor of 2-to-3 in most cases, which seems reasonable.

Statistical tests of the exposure distributions were undertaken using a two-sample
Kolmogorov-Smirnov (K-S) non-parametric test. The null hypothesis was that the cumulate
frequency exposure distributions for the various air quality scenarios are the same for each
exposure nJric being tested. The test statistic was the maximum difference between toe two
dSudom (Dn); it was evaluated using a large-sample F statistic The sigmficanc. level is
0 05 using the chi-square approximation. The test indicates that the exposure mdicators
shown in Tables 1 and 2 vary significantly across the six air quality scenarios with the
following exceptions.

For the Asls and 8h lEx 0.08 ppm pair:

Children exposed and children-occurrences at 8h 0.08 ppm for any exercise level.

For the Asls and In lEx 0.12 ppm (the current NAAQS) pair:

NOW people and children, NOW person occurrences, and children occurrences at Ih
0.12 ppm for any exercise level.

For the current standard and 8h lEx 0.10 ppm pair:

NOW people, person-occurrences, and number of people exposed on one day/season
at Ih 0.12 ppm during >:30 EVR exercise.

Children at 8h 0.08 ppm at any exercise level.

NOW persons and number of people exposed at 8h 0.08 ppm during >_15 EVR
exercise. (Note the person-occurrences metric at this MAXD was not statistically
different for this pair of NAAQS scenarios.)

For the 8h lEx 0.08 ppm and 8h 5Ex 0.08 ppm pair:

NOW people and person-occurrences, and children and children-occurrences at 8h
0.08 ppm at any exercise level.

Compared with the K-S analyses undertaken for the previous urban areas (2-6), the
above results are unusual. Most of the non-rejects involve the "any exercise" metrics, which
are not the main exposure/exercise conditions of concern. The only exercising metric that
could not be rejected as coming from the same distribution was the current standard/8h lEx
0 10 ppm pair, and not all exercising metrics could be so treated. In other words, the

-------
reject/cannot reject pattern was not consistent across all of the exercising exposure metrics.
(Nor was it consistent for the any-exercise metrics.)

Note that these are not strong statistical tests and that some of the K-S test -
requirements were not met. Random sampling was not used. The sample "observations"
simply are results of a simulation model; the population of interest, then, can only be the set
of possible outcomes produced by the model. A large-sample test statistic was applied to
small samples (10 runs). These caveats aside, 14 "cannot rejects" out of 126 tests is twice as
many as are expected based on a "pure-chance" probability of rejecting a true hypothesis
(5 %). This 11.1% cannot-reject rate is about the same at the rates seen in most of the urban
areas that were analyzed (2-6). Unlike most of these areas, there is not a clear-cut picture
produced by the K-S testing exercise in Denver. There is not a consistent pattern-especially
at exercise-associated with the alternative NAAQS investigated.

SUMMARY
The analyses of 5 alternative standards ozone NAAQS in the Denver area (Ih lEx
0.12 ppm; 8h lEx 0.10 pm; 8h 5Ex 0.08 ppm; 8h lEx 0.08 ppm; and 8h 5Ex 0.06 ppm)
indicate that attainment of the current standard or the 8h lEx 0.10 ppm alternative may not
minimize exposures at exercise in the Denver urban area. This finding applies to both the Ih
and 8h estimates, but are more important for the 8h results since more people and children
are estimated to experience 8h MAXD levels of concern than the Ih MAXD levels under
these alternative NAAQS.

The reader should remember the early caveat about the downward bias inherent in
pNEM results. While every effort has been made to present "best estimates," it is inherently
difficult to predict the upper end of the population exposure distribution using human activity
data bases that consist of only 1-3 days of information and that also under-represent adults
and children who exercise regularly. Year-long activity data from all segments of the
population are needed to adequately estimate the true population exposure distribution.

-------
                                 REFERENCES

1.     T. Johnson, J. Capel, and M. McCoy.  Estimation of Ozone Exposures
      Experienced by Urban Residents Using a Pro-
      babilistic Version of NEM and 1990 Population Data.  Durham, NC:  IT-AQS,
      1993 (Draft).

2.     T. McCurdy.  "Los Angeles Ozone Exposure Estimates," EPA Memorandum;
      November 4, 1993.

3     T McCurdy et al. "Estimating Ozone  Exposures in Philadelphia Using a Second-
      Generation Probabilistic Version of NEM."  Paper A210 presented at the 87th
      Annual Meeting of AWMA; Cincinnati, June 1994.

4     T McCurdy et al. "Estimating ozone Exposures in Houston Using a Second-
      Generation Probabilistic Version of NEM."  Paper presented at the Tropospheric
      Ozone Specialty Conference, Air and  Waste Management Association; Orlando,
       1994.

5     T McCurdy   "Estimating Ozone  Exposures in the New York Urban Area Using a
       Second-Generation Probabilistic  Version of  NEM."  EPA Paper, February 1994.

6      T McCurdy   "Estimating Ozone  Exposures in the Chicago Urban Area Using a
       Second-Generation Probabilistic Version of NEM."  EPA Paper, March 1994.

 7      Environmental Criteria and Assessment Office.  Air Quality Criteria for Ozone and
       Photochemical Oridants.  Research Triangle Park, NC:   1994 (Draft).

 8     H M  Richmond and T. McCurdy.  "Use of Exposure Analysis and Risk Assessment
       in the Ozone NAAQS Review."  Paper 88-121.3 presented at the 81st Annual
       Meeting of APCA; Dallas, June 19-24, 1988.

 9.    N.L. Nagda, H.E. Rector, and  M.D.  Koontz.  Guidelines for Monitoring Air
       Quality. Washington, DC: Hemisphere Pub. Co., 1987.

 10.    C.J. Weschler, H.C. Shields, and D.V. Nike.   "Indoor Ozone: Recent Findings," in:
        R. Burglund (ed.), Tropospheric Ozone in the Environment II.  Pittsburgh: Air &
        Waste Management Association,  1992.

 11.    T. McCurdy.  "Testing pNEM/O3 Runs to Determine if a Set of 10 Runs is
        'Representative'," EPA Memorandum; April 19, 1994.

  12.    W.C.  Adams. Measurement of Breathing Rate and Volume in Routinely
        Performed Daily Activities. Davis CA:  University of California, 1993.


                                         10

-------
13.    M. Absil, P. Narducci, R. Whitfield, and H.M.  Richmond.  "Chronic Lung Injury
      Risk Estimates for Urban Areas Having Ozone Patterns Similar to Those in the
      Northeast."  Paper presented at the AWMA Specialty Conference on Tropospheric
      Ozone and the Environment II; Atlanta, November 4-7, 1991.
                                        11

-------
  Table 1. ESTIMATES OF Ih 'MAXIMUM DOSAGE" EXPOSURES IN DENVER TO O3 CONCENTRATIONS EXCEEDING
0.12 PPM AND EVRa EQUALED OR EXCEEDED 30 ONE OR MORE TIMES PER OZONE
                            SEASON UNDER SIX AIR QUALITY SCENARIOS
Statistic


TOTAL EXPOSED POPULATION
Mean Estimate of the Number of People
Percent of Total Population
Range in this percentage for all runs
Mean Est. of Person-Occurrences
Percent of Total Per-Occurrences
Range in this percentage for all runs
Mean Est. of Occurrences/Person Exposed
Number of Days/Season Exposed (% Exposed)
1 Day
2 Days
3 Days
>3 Days
CHILDREN
Mean Estimate of the Number of Children
Percent of Children Population
Range in this percentage for all runs
Percent of Total Exposed Population
Mean Estimate of Children-Occurrences
Percent of total Child-Occurrences
Range in this percentage for all runs
Mean Estimate of Occurrences/Child Exposed
Number of Days Exposed/Season (% Exposed)
1 Day
2 Days
3 Days
>3 Days
As Is
Situation
(1991)

0
0
-
0
0
-
-

-
-
-
-

0
0
-
-
0
0
-
-

-
-
-
•
Attain
8h lEx
O.lOppm

2,900
0.1
0.0-0.9
2,900
@
0.0-®
1. 00

100
0
0
0

2,700
1.1
0.0-4.9
93.4
2,700
@
0.0-®
1.00

100
0
0
0
Attain
8h 5Ex
0.08 ppm

0
0
-
0
0
-
-

.
-
-
-

0
0
-
-
0
0
-
-

-
-
-
•
Attain
Ih 1 Ex
0.12 ppm

700
*
0.0-0.4
700
@
0.0-@
1. 00

100
0
0
0

100
*
0.0-0.3
12.3
100
@
0.0-@
1.00

100
0
0
0
Attain
8h 1 Ex
0.08 ppm

0
0
-
0
0
-
-

-
-
-
-

0
0
-
-
0
0
-
-

-
-
-
-
Attain
8h5 Ex
0.06 ppm

0
0
-
0
0
-
-

-
-
-
-

0
0
-
-
0
0
-
-

.
_
-
-
otcs:aEquivalent ventilation rate. It is equal to ventilation rate/body surface area; units are in 1 min"1 m"2
"Less than 0.05% but > 0.005%
@Lcss than 0.005% but >0.

-------
          Table 2. ESTIMATES OF 8h "MAXIMUM DOSAGE" EXPOSURES IN DENVER TO O3 CONCENTRATIONS EXCEEDING
       0.08 PPM AND EVR» EQUALED OR EXCEEDED 15 ONE OR MORE TIMES PER OZONE SEASON
Statistic
TOTAL NON-OUTDOOR WORKER EXPOSED POPULATION
Mean Estimate of the Number of People
Percent of Total Population
Range in this percentage for all runs
Mean Est. of person-Occurrences
Percent of total Per. -Occurrences
Range in this percentage for all runs
Mean Est. of Occurrences/Person Exposed
Number of Days/Season Exposed (% Exposed)
1 Day
2 Days
3 Days
4 Days
>4 Days
CHILDREN
Mean Estimate of the Number of Children
Percent of Children Population
Range in this percentage for all runs
Percent of Total Exposed Population
Mean Estimate of Children-Occurrences
Percent of Total Child-Occurrences
Range in this percentage for all runs
Mean Estimate of Occurrences/Child Exposed
Number of Days/Season (% Exposed)
1 Day
2 Days
3 Days
4 Days
>4 Days
As Is
Situation
(1991)

100
@
100
@
@-@
1.00
100
0
0
0
0

100
0.1
0.0-0.5
100.0
100
@
0.0-®
1.00

100
0
0
0
0
Attain
8h lEx
O.lOppm

47,200
3.2
0.9-4.8
50,100
*
@-*
1.06
92
8
0
0
0

34,800
13.4
3.2-20.2
73.7
37,700
0.1
*-0.1
1.08

91
9
0
0
0
^^^^^=^^=^^=
Attain
8h5Ex
0.08 ppm

800
0.1
0.0-0.6
800
@
0.0-@
1.00
100
0
0
0
0

800
0.3
0.0-3.2
100.0
800
@
0.0-*
1.00

100
0
0
0
0
^^_^^^^^^— ^J»^ 11-
Attain
Ih 1 Ex
0.12 ppm

20,800
1.4
0.0-3.6
23,200
@
0.0-*
1.12
89
11
0
0
0

17,100
6.6
0.0-13.1
82.2
19,500
*
0.0-0.1
1.14

86
14
0
0
0
Attain
8h 1 Ex
0.08 ppm

1,600
0.1
0.0-1.1
1.600

0.0-*
1.00
100
0
0
0
0

500
0.2
0.0-1.8
31.3
500
@
0.0-*
1.00

100
0
0
0
0
Attain
8h5 Ex
0.06 ppm


0
0


-
.
•
—
—


0
0
~
-
0
0
-


-
-
-
-

                                                                  m
*Less than 0.05% but > 0.005%
©Less than 0.005% but >0.

-------
    ESTIMATING OZONE EXPOSURES IN HOUSTON USING A SECOND-
            GENERATION PROBABILISTIC VERSION OF NEM
                            Thomas McCurdy
                   U.S. Environmental Protection Agency
                     Research Triangle Park, NC 27711
                               Ted Johnson
                                Jim Capel
                            Michael McCoy, Jr.
                  International Technology-Air Quality Services
                         Durham, NC  27707-6208
                              February 1994

FILE: HOUSTON.2

-------
INTRODUCTION7

       The U S  Environmental Protection Agency (EPA) is currently reviewing the ozone
(O,) primary national ambient air quality standard (NAAQS) to determine if the existing
standard protects public health with an adequate margin of safety.  In order to evaluate
whether alternative O3 NAAQSs provide an adequate margin of safety, the EPA Office  ot
Air Quality Planning and Standards (OAQPS) assesses  such factors as the nature and seventy
of health effects associated with O3 exposure, the degree of total human exposure (i.e.,
indoor and outdoor) to O3, and  the risks (probabilities) of ozone-related health effects
occurring in the exposed population when alternative O3 NAAQS are just attained.  This
paper discusses an EPA analysis of 03 exposure estimates for parts of the Houston
metropolitan area which employed the pNEM/O3 model developed by International
Technology - Air Quality Services  (IT-AQS). A report by Johnson, et al. (1) describes this
model in detail and presents  initial  results of applying it to nine U.S. urban areas, including
Houston  Exposure estimates for parts of the Los Angeles area are discussed in an EPA
memorandum (2) that was recently reviewed by the Clean Air Science Advisory Committee
 (CASAC) at a public meeting.  Exposure estimates for the Philadelphia area are described in
 an AWMA paper (3).  Additional analyses of O3 exposure estimates for the remaining  six
 urban areas that were modeled  will be forthcoming.

 OVERVIEW

       Pursuant to section 108 of the Clean Air Act, the Environmental Criteria Assessment
 Office (ECAO) within the Office of Research and Development (ORD) is preparing a criteria
 document (4) for O3.  This  document will be a comprehensive summary and critical
 assessment of the latest scientific research relevant to  assessing  the effects of a criteria
 pollutant  The document will be reviewed by CASAC, part of the EPA's independent
 Science Advisory Board, and by the general public.  Upon completion of the criteria
 document  OAQPS will prepare a  Staff Paper that summarizes and integrates scientific
 information presented in the criteria  document considered to be most relevant to decisions on
 alternative NAAQS and various analyses (e.g., air quality, exposure, and risk).

        An important aspect of health risk assessment  is an analysis of population exposure.
 Such an analysis provides the population basis for "headcount risk" estimates developed as
 part of EPA's NAAQS risk assessment work.  Estimates of population exposure  also provide
 important information regarding the number of people that may experience varying levels of
 O3 as they go through their daily  activities.


        Since 1980, OAQPS has used the NAAQS Exposure Model (NEM) to analyze  human
  exposure associated with the  "criteria air pollutants" covered by Title I of the Clean Air Act,
  as amended   Recent versions of this model use Monte  Carlo analyses to define numeric
  values for certain variables inherent  in the model.  These version of NEM are called pNEM,

-------
an acronym for "probabilistic" NEM. To differentiate among the various pollutant-specific
versions of pNEM, the one described here is denoted as pNEM/O3.

To date, the most recent version of pNEM/O3~the "second generation" version-has
been applied to nine major urban areas. These areas vary in size from Denver with 1.4
million people to New York with over 10.6 million people. pNEM/O3 partitions all land
within a selected urban area into large "exposure districts." The number of exposure
districts defined for each urban area varies from 6 to 16 in the set of nine areas recently
modeled; the number of districts reflects the number of monitors having valid air quality data
in a study area, most of the nine urban areas have 10 or more districts within their
boundaries.

People living within each exposure district, as estimated by the U.S. Bureau of
Census in 1990, are assigned to a single discrete point, the population centroid. O3 air
quality within each exposure district is estimated for each hour of the O3 season using data
from a nearby monitoring site. Because O3 in the ambient air generally is modified
considerably when entering a building or vehicle, outdoor O3 estimates are adjusted using a
mass-balance model to account for 3 indoor microenvironments. These microenvironments
are: indoors-at-home, indoors-other, and within a motor vehicle. The mass-balance model
used in pNEM/O3 is a simplified version of the generalized Nagda, Rector, and Koontz
(NRK) model (6). For our use, this model was revised to incorporate the assumption that
indoor decay rate is proportional to indoor O3 concentration. It was further revised to
incorporate assumptions concerning ozone decay rates suggested by Weschler, et al (7). The
resulting revised mass-balance model is fully described in Ref. 1.

While many types of exposure metrics are provided by pNEM/O3 only those for the
non-outdoor working population in Houston are discussed in this paper. Exposure estimates
for children are included in the results discussed here. Estimates for outdoor workers—a
major population sub-group of concern from an O3 exposure perspective-will be completed
in the Fall of 1994.

AIR QUALITY SCENARIOS INVESTIGATED

Six air quality were modeled for Houston (and the other 8 urban areas); the scenarios
are all on a daily maximum basis. They are:

1. 1990 air quality--the "as is" scenario
2. Attain Ih 1 Ex 0.12 ppm--the current standard
3. Attain 8h lEx 0.08 ppm
4. Attain 8h lEx 0.10 ppm
5. Attain 8h 5Ex 0.06 ppm
6. Attain 8h 5Ex 0.08 ppm

Where Ex stands for "expected exceedances," as defined by EPA in 40 Code of
Federal Regulations 50.9, and h is the usual symbol for hour(s). Thus, a full description of
the first alternative NAAQS scenarios is: attain the 1 hour daily maximum-one expected
exceedance NAAQS of 0.12 ppm. The other standards follow the same format. The 5Ex
NAAQS would allow 5 daily exceedances per year, on average, if they were implemented.

The iterative adjustment procedure used by IT-AQS to simulate just-attaining the 5
alternative NAAQS listed above involves a complex procedure of fitting Weibull equations to
the 1990 O3 data, modifying parameters of the equations so that the "worst" monitor just
achieves the desired standard, and then re-specifying the entire O3 pattern for all districts in
Houston. The procedure is described in detail by Johnson, et al. (1).

HOUSTON RESULTS

Results are based upon 10 runs (realizations) of the pNEM/O3 model for each air
quality scenario. Based on a previous analysis of sets of 10-run results versus a 108-run
result (8), the author believes that results from only 10 runs of the model will adequately
predict the mean and variance observed in 100 or more runs of pNEM/O3. Ten runs of the
model will not, however, adequately represent the entire range of possible outcomes.
Additional runs can only increase this range. This is an obvious shortcoming of the analyses
presented here, but limited resources preclude undertaking more runs.

-------
hundred person-occurrence could be 10 people experiencing the exposure 10 times, 100
people experiencing it once, one person experiencing it 100 times, or any combination
thereof.) Even though data are discussed here regarding the absolute and relative number
(percentage) of people exposed, it must be noted that "people" should not be used as an
exposure metric. The cohort-specific activity-days data base used for pNEM/O3 modeling
(and for all other exposure models that uses human activity information) provides no more
than three days of data for a single individual. Thus, rigorous inferences from these data can
be made only to "person-days" of exposure or "person-occurrences" of exposure, where an
occurrence can be Ih, 8h, or 24h time period in pNEM/O3. Only when activity data are
available sequentially for an entire year for individuals in a cohort can inferences be made to
people (or children, or outdoor workers as the case may be). This point is usually ignored in
the general exposure assessment literature.
All exposure estimates presented in this paper pertain to "daily maximum dose,"
(MAXD) where dose is defined as the product of ozone concentration and ventilation rate
over a defined time period. Note that maximum daily dose does not necessarily occur during
the time period of maximum ozone concentration in a given urban area.

Ventilation rate is a very important factor in estimating total dose for any air pollutant
(9). The ventilation rate metric used in pNEM/O3 is "equivalent ventilation rate" (EVR), or
ventilation rate (in liters per minute) divided by body surface area (BSA, in units of m2).
EVR has units of 1 min"1 m"2. It is a surrogate indicator for the number of O3 molecules that
enter the oral-nasal cavities per unit time period.

The Houston population analyzed here includes 2,370,510 NOW (non-outdoor
worker) people and 489,900 children. There are 865,236,900 possible daily maximum NOW
person-occurrences (NOW people x 365 days in the O3 season). The equivalent figure for
children is 178,813,140.

Exposure results appear in Tables 1 and 2. Table 1 includes Ih daily maximum dose
exposure estimates, while Table 2 provides similar results for 8h exposure estimates.

The results indicate that:

1. Currently (1990), fewer than 3% of the NOW population experience a daily
maximum dose (MAXD) O3 exposure @ >0.12 ppm @ an EVR of 30 1 min"1 m"2,
which is thought by many researchers to be a dose rate of concern (9). Almost all—
96%—of these exposures occur only on one day during the O3 season.

2. Currently (1990), about 7% of the children in Houston exceed the 0.12 ppm/30 EVR
max dose exposure level. This is about 53% of the NOW exposed population. 93%
of the children exposed at that level are predicted to experience only 1 day/season,
but 6% may experience two such exposures.

-------
3. Such MAXD exposures are greatly reduced whenever one of the alternative standards
are attained. For the relatively few children or adults having a MAXD exposure
>0.12 ppm @ _>_30 EVR when an alternative NAAQS is attained, all of them are
predicted to experience only one such daily Ih exposure per season.

The same general findings apply to the 8h results (Table 2), although relatively more
people may be affected. The level of concern here is a MAXD exposure of >0.08 ppm for
8h @ an EVR ^15 1 min'1 m'2. (Obviously, the EVR of concern drops as the averaging
period increases, just as the O3 concentration of concern drops. People cannot hold an EVR
of 30 1 min'1 m'2 for an extended period of time(9).)

The 8h results indicate that:

1. Currently (1990), over 5% of the NOW population are estimated to experience a daily
maximum dose O3 exposure level of concern. About 11% of these people may
experience such a MAXD exposure on more than one day during the O3 season. Two
percent of the applicable population may experience the exposure of concern on 3
days.

2. Currently (1990), over 19% of children in Houston may experience the 8h MAXD of
concern. In fact, about 79% of the total population experiencing this 8h MAXD are
children. Almost 86% of the children may experience this MAXD on only one day;
12% may experience it on 2 days, and 2% may see it on 3 days.

3. Attaining any of the alternative NAAQS reduces the MAXD exposure of concern,
although about 1 % of the children population might experience that exposure even if
the current NAAQS (In lEx 0.12 ppm) standard is attained. All of these exposures
are estimated to occur only on one day, however. The 8h lEx standard of 0.10 ppm
allows almost 5% of the children to experience the 8h MAXD exposure of concern--
all on only one day of the season, however.

Consistent with the Ih results, it is estimated that almost everyone (85%) in the NOW
population of Houston currently (1990) experiences one or more 8h exposures at maximum
daily dose when ventilation level is ignored. (This result is not shown in Table 2.) This
contrasts with the 5.0% of the NOW population who may experience the same 8h O3
concentration at exercise. Again, exercise levels-expressed as EVR-greatly affect

-------
pNEM/O3 exposure results. This same pattern holds true for children in Houston, but the
percentage is higher: 93% of all children may experience an 8h 0.08 ppm exposure if
breathing rate is ignored, as compared to 19% at exercise.

Statistical tests of the exposure distributions were undertaken using a two-sample
Kolmogorov-Smirnov (K-S) non-parametric test. The null hypothesis was that the cumulative
frequency exposure distributions for the various air quality scenarios are the same for each
exposure metric being tested. The test statistic was the maximum difference between the two
distributions (Dn); it was evaluated using a large-sample F statistic. The significance level is
0.05, using the chi-square approximation. The test indicates that all of the exposure
indicators shown in Tables 1 and 2 vary significantly across the six air quality scenarios with
the following exceptions, all for the 8h 5Ex 0.08 ppm/lh lEx 0.12 ppm pair of standards.

NOW adults and children with a IhDM exposure >0.12 ppm at any exercise level.

NOW- and children-occurrences at a MAXD exposure >0.12 ppm at exercise levels
of concern. (Surprisingly, the corresponding people or children indicators were not
found to have statistically significant distributions.)

NOW-people, children, NOW-occurrences, and children-occurrences for the 8h
MAXD indicator (>0.08 ppm @ _>_15 EVR).

Note that these are not strong statistical tests and that some of the K-S test
requirements were not met. Random sampling was not used. The sample "observations"
simply are results of a simulation model; the population of interest, then, can only be the set
of possible outcomes produced by the model. A large-sample test statistic was applied to
small samples (10 or 11 runs). In addition, 8 "cannot rejects" out of 152 tests is about 5%
of the total tests made—about the same value as the "pure-chance" probability. It is of
interest, however, to note that all 8 exposure metrics that were not rejected using the K-S test

involved the same paired air quality scenarios, which supports the finding that those
scenarios do not differ significantly with respect to estimated O3 exposures.
SUMMARY

The analyses of 5 alternative standards ozone NAAQS in Houston (Ih lEx 0.12 ppm;
8h lEx 0.10 pm; 8h 5Ex 0.08 ppm; 8h lEx 0.08 ppm; and 8h 5Ex 0.06 ppm) indicate that
attainment of each standard probably will reduce the percentage of the exposed population-

-------
both adults and children-experiencing Ih ozone exposures thought to increase health risks.
The analyses also found that for all air quality scenarios that were examined, just attaining an
alternative O3 standard reduces multiple-day Ih exposures for  the air quality/exercise levels
of interest.

       The picture is not quite so rosy,  however, for 8h MAXD exposures of concern,
especially for children.  Attaining a 8h lEx 0.10 ppm is predicted to reduce 8h MAXD
exposures of concern to 1% of the NOW population, but 96% of this population is composed
of children. This results in almost 5% of all children being exposed-albeit only on one day-
to the 8h MAXD of >0.08 ppm @ ^.15 EVR for this standard.   In addition, attaining the
current Ih standard results in  0.6% of children being exposed to the 8h MAXD of concern.
These  8h results are quite different than those for Los Angeles (2) and Philadelphia (3).

       The  reader should  remember the early caveat about the downward bias inherent in
the pNEM results.  While every effort has been made to present  "best estimates," it is
inherently difficult to predict the upper end of the population exposure distribution using
human activity data bases that consist of only 1-3 days of information and that also under-
represent adults and children who exercise regularly. Year-long activity data from all
segments of the population are needed to adequately estimate the  true population exposure
distribution.

-------
                                 REFERENCES

1.     T. Johnson, J. Capel, and M. McCoy.  Estimation of Ozone Exposures
      Experienced by Urban Residents Using a Pro-
      babilistic Version of NEM and 1990 Population Data.  Durham, NC:  IT-AQS,
      1993 (Draft).

2.     T. McCurdy.  "Los Angeles Ozone Exposure Estimates,11 EPA Memorandum;
      November 4, 1993.

3     T McCurdy, et al. "Estimating Ozone Exposures in Philadelphia Using a Second-
      Generation Probabilistic Version of NEM."  Paper A210 presented at the 87th
      Annual Meeting of AWMA; Cincinnati, June 1994.

4     Environmental  Criteria and Assessment Office.  Air Quality Criteria for Ozone and
      Photochemical Oxidants. Research Triangle Park, NC:  1994 (Draft).

5     H.M. Richmond and T. McCurdy.  "Use of Exposure Analysis and Risk Assessment
      in the Ozone NAAQS Review."  Paper 88-121.3 presented at the 81st Annual
      Meeting of APCA; Dallas, June 19-24, 1988.

6.    N.L. Nagda, H.E. Rector, and M.D. Koontz. Guidelines for Monitoring Air
      Quality.  Washington, DC: Hemisphere Pub. Co., 1987.

7.    C.J. Weschler, H.C. Shields, and D.V. Nike.  "Indoor Ozone: Recent Findings," in:
       R. Burglund (ed.), Tropospheric Ozone in the Environment II.  Pittsburgh: Air &
       Waste Management Association, 1992.

 8.     T. McCurdy.  "Testing pNEM/O3 Runs to Determine if a Set of 10 Runs is
       'Representative'," EPA Memorandum; April 19, 1994.

 9.     W.C. Adams. Measurement of Breathing Rate and Volume in Routinely
       Performed Daily Activities.  Davis CA: University of California, 1993.

 10.   M. Absil, P. Narducci, R. Whitfield, and H.M. Richmond.  "Chronic Lung Injury
       Risk Estimates for Urban Areas Having Ozone Patterns Similar to Those in the
        Northeast." Paper presented at the AWMA Specialty Conference on Tropospheric
        Ozone and the Environment U, Atlanta, November 4-7, 1991.

-------
                  Table 1.  ESTIMATES OF Ih "MAXIMUM DOSAGE" EXPOSURES IN HOUSTON TO O3 CONCENTRATIONS EXCEEDING
                0.12 PPM AND EVRa EQUALED OR EXCEEDED 30 ONE OR MORE TIMES PER OZONE SEASON UNDER
                               SIX AIR QUALITY SCENARIOS
Statistic
TOTAL EXPOSED POPULATION
Mean Estimate of the Number of People
Percent of Total Population
Range in this percentage for all runs
Mean Est. of Person-Occurrences
Percent of Total Per-Occurrences
Range in this percentage for all runs
Mean Est. of Occurrences/Person Exposed
Number of Days/Season Exposed (% Exposed)
1 Day
2 Days
3 Days
>3 Days
CHILDREN
Mean Estimate of the Number of Children
Percent of Children Population
Range in this percentage for all runs
Percent of Total Exposed Population
Mean Estimate of Children-Occurrences
Percent of total Child-Occurrences
Range in this percentage for all runs
Mean Estimate of Occurrences/Child Exposed
Number of Days Exposed/Season (% Exposed)
1 Day
2 Days
3 Days
>3 Days
As Is
Situation
(1991)

66,100
2.8
1.6-4.3
68,100
0.01
@-.01
1.03

96
3
1
0

35.000
7.1
2.1-15.0
53.0
36,900
*
*_*
1.05

93
6
1
0
Attain
8h lEx
0.10 ppm

3,700
0.2
0.1-0.7
3,700
@
@-@
1.00

100
0
0
0

1.600
0.3
0.0-2.6
43.2
1,600
@
0.0-*
1.00

100
0
0
0
Attain
8h 5Ex
0.08 ppm

1,600
0.1
0.0-0.7
1,600
@
0.0-®
1.00

100
0
0
0

0
0
-
-
0
0
-
-

-
-
-
™
Attain
Ih 1 Ex
0.12 ppm

0
0
-
0
0
-
-

-
-
-
-

0
0
-
-
0
0
-
-

-
-
-
m
Attain
8h 1 Ex
0.08 ppm

0
0
-
0
0
-
-

-
-
-
-

0
0
-
-
0
0
-
-

-
-
-
"
Attain
8h5 Ex
0.06 ppm

0
0
-
0
0
-
-

-
-
-
-

0
0
-
-
0
0
-
-

-
-
-
"
Notcs^Equivalent ventilation rate.  It is equal to ventilation rate 4- body surface area; units arc in I min"1 m"1
    *Lcss than 0.05% but > 0.005%
       @Lcss than 0.005% but >0.

-------
                   Table 2. ESTIMATES OF 8h "MAXIMUM DOSAGE" EXPOSURES IN HOUSTON TO O3 CONCENTRATIONS EXCEEDING
               0.08 PPM AND EVRa EQUALED OR EXCEEDED 15 ONE OR MORE TIMES PER OZONE SEASON UNDER
                               SIX AIR QUALITY SCENARIOS
Statistic
TOTAL NON-OUTDOOR WORKER EXPOSED POPULATION
Mean Estimate of the Number of People
Percent of Total Population
Range in this percentage for all runs
Mean Est. of person-Occurrences
Percent of total Per. -Occurrences
Range in this percentage for all runs
Mean Est. of Occurrences/Person Exposed
Number of Days/Season Exposed (% Exposed)
1 Day
2 Days
3 Days
>3 Days
CHILDREN
Mean Estimate of the Number of Children
Percent of Children Population
Range in this percentage for all runs
Percent of Total Exposed Population
Mean Estimate of Children-Occurrences
Percent of Total Child-Occurrences
Range in this percentage for all runs
Mean Estimate of Occurrences/Child Exposed
Number of Days/Season (% Exposed)
1 Day
2 Days
3 Days
>3 Days
As Is
Situation
(1991)

119,500
5.0
2.1-7.5
134,300
0.02
0.01-0.03
1.15

89
9
2
0

94,100
19.2
8.4-28.8
78.7
108,800
0.06
*-0.12
1.16

86
12
2
0
Attain
8h5Ex
0.08 ppm

1,000
*
0.0-0.4
1,000
@
0.0-®
1.00

100
0
0
0

1,000
0.2
0.0-1.8
100.0
1,000
@
0.0-®
1.00

100
0
0
0
Attain
8h lEx
0.10 ppm

23,800
1.0
*-3.1
23,800
0.01
@-0.01
1.00

100
0
0
0

22,800
4.7
0.0-14.7
95.8
22,800
0.01
0.0-0.04
1.00

100
0
0
0
Attain
8h 1 Ex
0.08 ppm

800
*
0.0-0.3
800
@
0.0-®
1.00

100
0
0
0

800
0.2
0.0-1.6
100.0
800
@
0.0-®
1.00

100
0
0
0
Attain
Ih 1 Ex
0.12 ppm

2,800
0.1
0.0-0.6
2,800
@
0.0-®
1.00

100
0
0
0

2.800
0.6
0.0-2.7
100.0
2,800
@
0.0-®
1.00

100
0
0
0
Attain
8h5Ex
0.06 ppm

0
0
-
0
0
-
-

-
-
-


0
0
-
-
0
0
-
-

-
-
-
~
Notes:  HEquivalent ventilation rate.  It is equal to vcntilntion rate/body surface area; units arc m 1 mm'
        *Less than 0.05% but > 0.005%
        ©Less than 0.005% but >0.
m"

-------
                        November 4,  1993
                     (Revised March 9,  1994
MEMORANDUM

SUBJECT:  Los Angeles Ozone Exposure Estimates

FROM:     Tom McCurdy
          Ambient Standards Branch  (MD-12)

TO:       Interested Parties


SUMMARY

     This memo summarizes analyses  that I have done on ozone  (O3)
exposure estimates for Los Angeles  produced by International
Technology - Air Quality Services  (IT-AQS).  The estimates  are  a
result of applying our pNEM/O3 exposure model to six air quality
scenarios representing 1991 "as is" conditions and five
hypothetical situations when alternative  Ih and 8h O3 standards
are  just attained.  Unless otherwise noted, the exposure
estimates provided in this summary  represent the mean value of
ten  runs of the model.

     The analyses were undertaken  for  the total population--
excluding outdoor workers--and for  children.  Outdoor worker
exposures will be evaluated in the  next few months when a better
human  activity data base can be constructed for outdoor workers
than we  currently have.

     Analyses of  "as  is" air quality conditions for  the year  1991
indicate that:

     1)    80% of  the  total non-outdoor working population at
           "any" breathing  rate experiences one or more  exposures
           >0.12 ppm on a Ih daily  maximum basis.

     2)    While at a  high  exercise level  (equivalent  ventilation
           rate  (EVR)  of  30 or higher), only  9% of the  total non-
           outdoor working  population experiences  one  or more
           exposure  >0.12 ppm  (Ih maximum  daily basis).   For
           children only, the percentage who  experiences one or
           more  exposures >0.12 ppm (Ih maximum daily basis) at
           this  exercise  level  is  27%.

-------
     3)   11% of the total non-outdoor working population
          experiences one or more exposures >0.08 ppm 8h daily
          maximum average when at a breathing rate of 15 EVR or
          higher; 37% percent of the children experiences such an
          exposure at the same breathing rate.

     4)   21% percent of the total non-outdoor working population
          experiences one or more exposure >0.06 ppm 8h daily
          maximum average when at a breathing rate of 15 EVR or
          higher; 69% of the children experiences such an
          exposure at the same breathing rate.

     The analyses of 5 alternative standards (Ih 1 Ex 0.12 ppm;
8h 1 Ex 0.10 ppm; 8h 5 Ex 0.08 ppm; 8h 1 Ex 0.08 ppm; and 8h 5 Ex
0.06 ppm) indicate that when any alternative standard that was
examined is just attained, the percentage of the exposed
population is sharply reduced.  For all cases,  except the 8h 1 Ex
0.10 alternative, the percentage of both the total non-outdoor
work population and children alone exposed at exercise one or
more times to the level of concern is well below 1%.  The 8h
exposures to 0.08 ppm or higher for the 8h 1 Ex 0.10 alternative
are 1% for the total non-outdoor working population and 3% for
children.  These analyses also found that for all air quality
scenarios that were examined, just attaining an alternative 03
standard reduces multiple-day exposures for the air quality/
exercise levels of interest.

DETAILS

     IT-AQS has completed its modeling of 03 exposures estimates
for six air quality scenarios.  Those scenarios are:

     1.   1991 air quality—the "as is" scenario
     2.   Attain In lEx 0.12 ppm--the current standard
     3.   Attain 8h lEx 0.08 ppm
     4.   Attain 8h lEx 0.10 ppm
     5.   Attain 8h 5Ex 0.06 ppm
     6.   Attain 8h 5Ex 0.08 ppm

     The model used for exposure assessment is the most recent
version of pNEM/O3--"probabilistic" NAAQS Exposure Model  applied
to O3.   The model is described in a draft IT-AQS report entitled:

          Estimation of Ozone Exposures Experienced by Urban
          Residents Using a Probabilistic Version of NEM and 1990
          Population Data (September 1993) .

The report was written by Ted Johnson, project manager for IT-
AQS, and two of his staff members:  Jim Capel and Mike McCoy.

-------
     Modeling results exist for all nine urban areas that we plan
to evaluate.   We wanted to model a year that allowed us to
reasonably match up 1990 census data with air quality
information.   We used 1990 or 1991 as the year of analysis in an
area, using the year with the greatest number of O3 monitors
reporting valid data.  The following situations were modeled:

          Urban Area	Year	Population
Chicago
Denver
Houston
Los Angeles
Miami
New York
Philadelphia
St . Louis
Washington, DC
1991
1990
1990
1991
1991
1991
1991
1990
1991
6,175,000
1,485,000
2,371,000
10,371,000
1,942,000
10,658, 000
3,786,000
1,707, 000
3, 085, 000
     All of these areas violated the existing O3 standard in the
year analyzed except Miami.  The areas are located in every EPA
region, except 10--the northwest, which has only one non-
attainment area  (Portland).

     While modeling results exist for the above areas, they have
not yet been analyzed or summarized.  Only the Los Angeles
results have been thoroughly evaluated, and these results are
described in this memo.

     Estimates for two population groups will be described:  (1)
all people except outdoor  workers--who constitute about  10% of
the adult working population, and  (2) children—everyone younger
than 19 years of age.  Outdoor worker exposure estimates will be
evaluated after  IT-AQS can construct a larger file of human
activity patterns that better reflect the amount of time that
true outdoor workers  (i.e., those individuals who work outdoors
at  least 6 hours per day,  on average) spend outside in their
daily  occupation.  In developing this new data base, we  will
probably focus on people working in the following industrial
classification categories:

     1.  agriculture,  forestry and  fisheries  (101-031)
     2.  mining  (040-050)
     3.  construction  (060)

These  workers currently are under-represented in our NEM human
activity data base.

     Our results are based upon  10  runs  (realizations) of pNEM/03
for each air quality scenario.   Based on my previous analyses of
sets of  10-run results versus a  108-run result  (see the  attached
April  19,  1993 memo),  I feel that the estimates presented here

-------
adequately represent the mean and overall distribution expected
if more runs of pNEM/O3 were undertaken.  What 10-runs do not
adequately represent, however, is the entire range of possible
outcomes.  More runs could only increase this range.

     This memo includes tabular and graphical data on  (1) the
number and percent of people exposed, and (2) the number and
percent of person-occurrences for the two population groups
mentioned above.  Even though data are discussed here regarding
the absolute and relative  (percentage) number of people exposed,
it must be remembered that we really should not be using "people"
as an exposure metric.  The cohort-specific activity-days data
base used for pNEM/O3 modeling (and for all  other exposure models
that uses human activity information) is based upon one, two, or
three days of data from a single individual.  Thus, inferences
from these data rigorously can only be made to "person-days" of
exposure, or "person-occurrences" of exposure, where an
occurrence can be Ih, 8h, or a day in pNEM/03.  This metric also
applies to children-occurrences or children-days of exposure.

     Only if activity data are available sequentially for an
entire year for individuals in a cohort can inferences be made to
people (or kids, or outdoor workers as the case may be).  This
point is never discussed in the general exposure assessment
literature but should be.

     Data for three exposure indicators are included.

     1.   daily maximum dose Ih exposures >.12 ppm at an
          equivalent breathing rate  (EVR)  of 30 liters or higher
          per minute per square meter of body surface area  (1
          min'1  m"2) .  This is about  55-60 1 min'1 for adults.
          These results are shown in Table 1 and Figure I.

     2.   daily maximum dose 8h exposures >0.08 ppm at 15 EVR.
          This is about 20-30 1 min"1 for  adults  and 12-25  1 min"
          1 for children.   These results are depicted in Table 2
          and Figure 2.

     3.   daily maximum dose 8h exposures >0.06 ppm at 15 EVR.
          See Table 3 and Figure 3 for results.

Other indicators were evaluated and many others can be evaluated.
One of the purposes of this memo is to obtain definitive feedback
on what indicators should be presented in the O3  staff paper.
Therefore, please review the information presented to see if all
exposure indicators of interest and importance are covered.

     As an example of an available but unreported indicator,
daily max dose 8h exposures to 0.10 ppm or higher were analyzed.
However,  exposures at this high an 8h average were seen only
under the "as is" air quality scenario.   Hence, I do not provide

-------
a table/figure for this indicator.  For the record, however,
pNEM/0, indicates that currently 630,700 people on average are
exposed to 8h 0.10 ppm or higher for an average of 1.4 times per
year   This is about 6.1% of the LA population that was analyzed.
The range in population estimate is 527,100-814,700, or 5.1-7.9%.
The range in occurrences per exposed person is 1.3-1.7 times per
year.

     To put the numbers and percentages that follow into
perspective, the baseline "denominators" used  in the Los Angeles
analyses are:

     Total population  :       10,371,115
     Children population  :     1,954,886

     Total daily maximum person-occurrences: 3,785,456,975
     Total daily children-occurrences:         713,533,390

     Total hourly person-occurrences:       90,850,867,400

     Exposures to the  air quality  levels of interest  are  much
higher, of course,  at  "any" breathing  rate than at  the relatively
high breathing rates  of  interest.   The following  statistics are
for  exposures to Ih and  8h  concentrations of interest at  any
breathing  rate for  the "as  is"  air quality scenario:
Ih 0.12
ppm
8h 0.10
ppm
8h 0.08
ppm
8h 0.06
ppm
 Mean  Number  of
 People  Exposed   8,305,200     4,556,100     7,196,600     9,872,400

 Mean  %  of Tot.
 Population        80.1           43.9         69.4         95.2

 Range in this
 Percentage      76.7-82.7     40.8-51.2     67.3-71.2     94.1-97.2

      Note the relatively high mean percentage estimates and the
 quite narrow range in these estimates.   The Ih 0.12  ppm estimate
 of 80.1% of  the total population is about 15% lower  than
 comparable estimates from previous REHEX and SAI/NEM 03 modeling
 analyses.  Their runs,  however,  are for the late 1980 time period
 when 03 air  quality was worse than it currently is (i.e.  than
 1991, our "as is" year).

      In fact, two districts in the Los Angeles area did not
 exceed a 0.12 ppm Ih O3 concentration level in 1991.   These
 districts are Long Beach and Hawthorne.  They comprise about 15%
 of the modeled urban area population base and include major
 employment centers that boost daytime inhabitants to over that
 relative level.  Thus,  the 80% estimate is plausible given the

-------
amount of time spent indoors by most people, especially in air
conditioned spaces where O3 levels are a fraction (1/4 to 1/2) of
that seen outside at the same time.

     Other modeling differences can account for the disparity in
people estimated to experience a 0.12 ppm O3 level.   For
instance, the two non-EPA studies noted above used modeled O3 air
quality data, whereas ours used transformed ambient monitoring
data.  In addition, our pNEM/O3 analyses explicitly incorporates
sequential time spent indoors as a function of time-of-day,
whereas REHEX does not and SAI/NEM used the old 1982 NEM activity
patterns that are out of date.

     A more in-depth review of the outputs follows.

Table I/Figure 1

     pNEM/O3 estimates that 9.1% of the total population
currently experiences 03 exposures >0.12 ppm 1-h at  high exercise
(.> 30 EVR) .   Over half  (56%) of these people are children, and
over 25% of children (27.1%) experience one or more of these
exposures per 03 season (i.e.,  per year in LA).   About 31% of the
exposed children experience more than 1 day of exposure at a Ih
30 EVR or higher.

     Attaining any of the alternative ambient standards listed in
Table 1 dramatically reduces Ih O3 exposures at  high exercise.
Attaining the last two standards  (the 8h lEx 0.08 and 8h SEX 0.06
scenarios) reduces Ih high-exercise exposures to zero.  Note that
no child or person would be exposed to more than one day of 0.12
Ih @30 EVR exposure if any alternative O3 NAAQS  analyzed here is
attained.

     Some of the data shown in Table 1 is depicted in Figure 1.
The figure is good for depicting the range in estimates for any
scenario.  Note that the middle-three scenarios have 3-5 orders
of magnitude of range in their estimates of Ih 0.12 ppm @30 EVR
exposures.  This is a much larger range than seen in previous NEM
exposure analyses, and probably reflects the higher--and in my
opinion, better--ventilation rate data recently developed by IT
for the high exercise level.

Table 2/Figure 2

     More people, both absolutely and relatively,  are exposed to
0.08 ppm 8h @ 15 EVR than to 0.12 ppm Ih @ 30 EVR for all the air
quality scenarios except the last one.  The differences are
larger for children than for the entire population.   For
instance, Table 2 indicates that almost 37% of children currently
see 0.08 8h @15 EVR one or more times/year (range: 31-44%).
About 51% of these exposed children see more than one
exposure/year.

-------
     The estimates of exposure to 0.08 ppm 8h @15 EVR drop rather
rapidly with attainment of the alternative NAAQS.   See Figure 2.
Note again the rather large range in estimates for the middle
three scenarios.

Table 3/Figure 3

     The air quality indicator of interest here is 0.06 ppm 8h @
15 EVR.  Because  this is a relatively low 8h average for Los
Angeles, a large  number and percentage of people experience this
exposure for all  of the scenarios evaluated.  Contrary to the
previous indicators, except for the last NAAQS air quality
scenario, the indicator range estimated for each scenario is
quite narrow.

     Most of the  people exposed to 0.06 ppm 8h @ 15 EVR or above
are children.  The proportion of children only varies between 61-
74% for the first 5 scenarios.   (This jumps to 86% for the last,
quite low 8h 5Ex 0.06 ppm scenario.)

     Table 3 does not include an estimate of the percent of
exposed children or the total population that sees more than one
day/season over the indicator of interest.  It just was too much
work to evaluate that data for this preliminary analysis; the
data are available to obtain that estimate, however. A
superficial examination of this exposure indicator, however,
indicates that all  (100%) of children exposed to 0.06 ppm for 8h
at 15+ EVR saw more than one day per year under the first two air
quality scenarios.  While this percentage drops off for the next
3 scenarios, some children and adults obviously saw more than 1
day per year  (because mean occurrences/personal exposed is >1.0).
No one saw more than 1 day for the 8h SEX 0.06 ppm scenario, the
alternative that allows up to five exceedances per year.

FINAL COMMENTS

     The results presented here represent only a fraction of the
data available from pNEM/O3 modeling exercise.  Comments and
recommendations are hereby solicited regarding the type of
exposure indicators evaluated, the format of presentation, and
emphasis.  All comments and questions regarding this memo and the
pNEM/O3 modeling effort should be directed to Tom McCurdy at the
address given above or by calling  (919)541-5658.

-------
Table 1.
ESTIMATES OF Ih OZONE EXPOSURES IN LOS ANGELES TO OZONE CONCENTRATIONS EXCEEDING 0.12 PPM AND EVR" EQUALED OR EXCEEDED


Statistic
TOTAL EXPOSED POPULATION
Mean Estimate of the Number of People
Percent of Total Population
Range in This Percentage
Mean Est. of Person-Occurrences
Percent of Total Per. -Occurrences
Mean Est. of Occurrences/Person Exposed
Percent of People > 1 day/Year Exposure
Number of Days Exposed/Season (% Exposed)
1l"\n< ,
Day
2 Days
3 Days
4 Days
5 Days
>5 Days
EXPOSED CHILDREN
Percent of total Exp. Pop. Who are Children
Mean Estimate of the Number of Children
Percent of Children
Range in this Percentage
Mean est. of Children-Occurrences
Percent of total Child. -Occurrences
Mean Est. of Occurrences/Child Exposed
Percent of Children > 1 Day/Year Exposure

Number of Days Exposed/Season (% Exposed)
1 Day
2 Days
3 Days
4 Days
5 Days
>5 Dnys
As Is Situation
(1991)

947,400
9.1
7.2-11.4
1,097,500
0.03
1.2
22.0
78

16
4
1
1
0
56.0
530,400
27.1
18.8-31.8
949.200
0.1

1.8
42.0

69
21
7
2
1
0
• ii i IP -
Attain
8h IE*
0.10 ppm
23,100
0.2
0.0-0.7
23,100
@
1.0
0
100

0
0
0
0
0
85.0
19,700
1.0
0.0-3.4
19,700
@

1.0
0

100
0
0
0
0
0
Attain
8h5Ex
0.08 ppm
3,000
*
0.0-0.2
3,000
@
1.0
0
100

0



0
53.0
1,600
0.1
0.0-0.8
1.600
@

1.0
0

100
0
0
0
0
0
. . ' i i 1 i i .U
Attain
Ih 1 Ex
0.12 ppm
900
^
0.0-*0.1
900
@
1.0
0
100

0



0
100.0
900
*
0.0-0.5
900
@

1.0
0

100
0

•
0
Attain
8h 1 Ex
0.08 ppm
0
0

0


"




-
0
0

0
0




-



-
Attain
8h 5 Ex
0.06 ppm
0
0

0
„






-
0


0





-



-
Notes: ^Equivalent ventilation rate. It is equal to ventilation rale -=- body surface area; units are in 1 mm m
*Lcss than 0.05% but > 0.
©Less than 0.005% but >0.

-------
Table 2.  ESTIMATES OF 8h OZONE EXPOSURES IN LOS ANGELES TO OZONE CONCENTRATIONS EXCEEDING AND EVR" EQUALED OR EXCEEDED 0.08


Statistic
TOTAL EXPOSED POPULATION
Mean Estimate of the Number of People
Percent of Total Population
Range in This Percentage
Mean Est. of Person-Occurrences
Percent of Total Per. -Occurrences
Mean Est. of Occurrences/Person Exposed
Percent of People > 1 day/Year Exposure
Number of Days Exposed/Season (% Exposed)
1 Day
2 Days
3 Days
4 Days
5 Days
>5 Days
EXPOSED CHILDREN
Percent of total Exp. Pop. Who are Children
Mean Estimate of the Number of Children
Percent of Children
Range in this Percentage
Mean est. of Children-Occurrences
Percent of total Child.-Occurrences
Mean Est. of Occurrences/Child Exposed
Percent of Children > 1 Day/Year Exposure
Number of Days Exposed/Season (% Exposed)
1 Day
2 Days
3 Days
4 Days
5 Days
>5 Days
As Is Situation
(1991)

1,122,800
10.8
9.3-13.5
1,990,300
0.05
1.8
45.0
55
25
12 .
5
2
1

63.9
718,000
36.7
30.8-43.5
1,389,300
0.2
1.9
51.3

49
26
14
6
3
2
Attain
8h lEx
O.lOppm
106,300
1.0
0.6-1.7
116,200
@
1.1
7.7
93
5
2
0
0
0

61.3
65,200
3.3
0.5-7.6
72,500
it
1.1
7.9

92
5
3
0
0
0
Attain
8h5Ex
0.08 ppm
7,700
0.1
0.0-0.2
7,700
@
1.0
0
100
0
0
0
0
0

49.4
3,800
0.2
0.0-0.7
3,800
*
1.0
0

100
0
0
0
0
0
Attain
Ih 1 Ex
0.12 ppm
5,000
*
0.0-0.1
5,000
@
1.0
0
100
0
0
0
0
0

0
0
0
•
0
0
0
0

-
-
•
-
-
•
Attain
8h I Ex
0.08 ppm
1,700
^
0.0-0.1
1.700
@
1.0
0
100
0
0
0
0
0

41.2
700
*
0.0-0.4
700
*
1.0
0

100
0
0
0
0
0
Attain
8h5 Ex
0.06 ppm
0

0
0
0

-
~
"
"
"
~

"
0
0
™
0
0
0
0

~
•
"
"
"
"
Notes: "Equivalent ventilation rate. It is equal to ventilation rate + body surface area; units are in 1 min ' m *•
*Less than 0.05% but > 0.
©Less than 0.005% but >0.

-------
ESTIMATING OZONE EXPOSURES IN THE MIAMI URBAN AREA USING A
    SECOND-GENERATION PROBABILISTIC VERSION OF NEM
                           Thomas McCurdy
                   U.S. Environmental Protection Agency
                     Research Triangle Park, NC 27711
                              March 1994
 File: MIAMI.

-------
INTRODUCTION

The U.S. Environmental Protection Agency (EPA) currently is reviewing the ozone
(O3) primary national ambient air quality standard (NAAQS) to determine if the existing
standard protects public health with an adequate margin of safety. In order to evaluate
whether alternative O3 NAAQSs provide an adequate margin of safety, the EPA Office of
Air Quality Planning and Standards (OAQPS) assesses such factors as the nature and severity
of health effects associated with O3 exposure, the degree of total human exposure (i.e.,
indoor and outdoor) to O3, and the risks (probabilities) of ozone-related health effects
occurring in the exposed population when alternative O3 NAAQS are just attained. This
paper discusses an EPA analysis of O3 exposure estimates for parts of the Miami
metropolitan area which employed the pNEM/O3 model developed by International
Technology-Air Quality Services (IT-AQS). A report by Johnson, et al. (1) describes this
model in detail and presents initial results of applying it to nine U.S. urban areas, including
Miami. Exposure estimates for parts of the Los Angeles area are discussed in an EPA
memorandum (2) that was reviewed by the Clean Air Science Advisory Committee (CASAC)
at a public meeting. Exposure estimates for the Philadelphia area are described in an
AWMA paper (3). Exposure estimates for Houston are described in an AWMA specialty
conference paper (4). New York, Chicago, and Denver results are contained in EPA papers
(5-7). Additional analyses of O3 exposure estimates for the remaining urban areas that were
modeled will be forthcoming.

OVERVIEW

Pursuant to section 108 of the Clean Air Act, the Environmental Criteria Assessment
Office (ECAO) within the Office of Research and Development (ORD) is preparing a criteria
document (8) for O3. This document will be a comprehensive summary and critical
assessment of the latest scientific research relevant to assessing the effects of a criteria
pollutant. The document will be reviewed by CASAC, pan of the EPA's independent
Science Advisory Board, and by the general public. Upon completion of the criteria
document, OAQPS will prepare a Staff Paper that summarizes and integrates scientific
information presented in the criteria document considered to be most relevant to decisions on
alternative NAAQS and various analyses (e.g., air quality, exposure, and risk).

An important aspect of health risk assessment is an analysis of population exposure
(9). Such an analysis provides the population basis for "headcount risk" estimates developed
as part of EPA's NAAQS risk assessment work. Estimates of population exposure also
provide important information regarding the number of people that may experience varying
levels of O3 as they go through their daily activities.

-------
an acronym for "probabilistic" NEM. To differentiate among the various pollutant-specific
versions of pNEM, the one described here is denoted as pNEM/O3.

To date, the most recent version of pNEM/O3-the "second generation" version-has
been applied to nine major urban areas. These areas vary in size from Denver with 1.4
million people to New York with over 10.6 million people. pNEM/O3 partitions all land
within a selected urban area into large "exposure districts." The number of exposure
districts defined for each urban area varies from 6 to 16 in the set of nine areas recently
modeled; the number of districts reflects the number of monitors having valid air quality data
in a study area. Most of the nine urban areas have 10 or more districts within their
boundaries.

People living within each exposure district, as estimated by the U.S. Bureau of
Census in 1990, are assigned to a single discrete point, the population centroid. O3 air
quality within each exposure district is estimated for each hour of the O3 season using data
from a nearby monitoring site. Because O3 in the ambient air generally is modified
considerably when entering a building or vehicle, outdoor O3 estimates are adjusted using a
mass-balance model to account for 3 indoor microenvironments. These microenvironments
are: indoors-at-home, indoors-other, and within a motor vehicle. The mass-balance model
used in pNEM/O3 is a simplified version of the generalized Nagda, Rector, and Koontz
(NRK) model (10). For our use, this model was revised to incorporate the assumption that
indoor decay rate is proportional to indoor O3 concentration. It was further revised to
incorporate assumptions concerning ozone decay rates suggested by Weschler, et al (11).
The resulting revised mass-balance model is fully described in Ref. 1.

While many types of exposure metrics are provided by pNEM/O3 only those for the
non-outdoor working population in Miami are discussed in this paper. Exposure estimates
for children are included in the results discussed here. Estimates for outdoor workers~a
major population sub-group of concern from an O3 exposure perspective—will be completed
in the Fall of 1994.

AIR QUALITY SCENARIOS INVESTIGATED

Six air quality scenarios were modeled for the Miami urban area (and the other 8
urban areas); the scenarios are all on a daily maximum basis. They are:

1. 1991 air quality-the "as is" scenario
2. Attain Ih 1 Ex 0.12 ppm--the current standard
3. Attain 8h lEx 0.08 ppm
4. Attain 8h lEx 0.10 ppm
5. Attain 8h 5Ex 0.06 ppm
6. Attain 8h 5Ex 0.08 ppm

The iterative adjustment procedure used by IT-AQS to simulate just-attaining the 5
alternative NAAQS listed above involves a complex procedure of fitting Weibull equations to
the 1991 O3 data, modifying parameters of the equations so that the "worst" monitor just
achieves the desired standard, and then re-specifying the entire O3 pattern for all districts in
Miami. The procedure is described in detail by Johnson, et al. (1).

Miami had relatively good O3 air quality in 1991 in that it had few days with a Ih
daily maximum concentration >0.11 ppm. The area's "design value"-the second-highest Ih
daily value measured at any site in the urban area-was 0.123, which is very close to the
current standard. Miami's 1991 monitored 8h values were also correspondingly low. Thus,
just-attaining some of the alternatives listed above can result in higher Ih and 8h O3 peak
concentrations being included than were monitored in 1991. Exposure estimates for these
alternatives similarly can be higher than those predicted for 1991. These results follow
directly from the logic used to evaluate alternative NAAQS (9), and are quite feasible in
reality.

MIAMI RESULTS

Results are based upon 10 runs (realizations) of the pNEM/O3 model for each air
quality scenario. Based on a previous analysis of sets of 10-run results versus a 108-run

-------
result (12), the author believes that results from only 10 runs of the model will adequately
predict the mean and variance observed in 100 or more runs of pNEM/O3. Ten runs of the
model will not, however, adequately represent the entire range of possible outcomes.
Additional runs can only increase this range. This is an obvious shortcoming of the analyses
presented here, but limited resources preclude undertaking more runs.

This paper includes tabular data on (1) the number and percent of people and children
exposed, and (2) the number and percent of person-occurrences for children and the adult,
non-outdoor worker population in Miami. (A person-occurrence is the metric that focuses on
an event: when the exposure-of-concern is reached regardless of who reaches it. One
hundred person-occurrence could be 10 people experiencing the exposure 10 times, 100
people experiencing it once, one person experiencing it 100 times, or any combination
thereof.) Even though data are discussed here regarding the absolute and relative number
(percentage) of people exposed, it must be noted that "people" should not be used as an
exposure metric. The cohort-specific activity-days data base used for pNEM/03 modeling
(and for aH other exposure models that use human activity information) provides no more
than three days of data for a single individual. Thus, rigorous inferences from these data can
be made only to "person-days" of exposure or "person-occurrences" of exposure, where an
occurrence can be Ih, 8h, or 24h time period in pNEM/O3. Only when activity data are
available sequentially for an entire year for individuals in a cohort can inferences be made to
people exposed over the year (or to children, or to outdoor workers as the case may be).
This point usually is ignored in the general exposure assessment literature.

Ventilation rate is a very important factor in estimating total dose for any air pollutant
(13). The ventilation rate metric used in pNEM/O3 is "equivalent ventilation rate" (EVR),
or ventilation rate (in liters per minute) divided by body surface area (BSA, in units of m2).
EVR has units of 1 min"1 m"2. It is a surrogate indicator for the number of O3 molecules that
enter the oral-nasal cavities per unit time period.

The Miami urban area population analyzed here includes
1.9 million NOW (non-outdoor worker) people and 327,000 children. There are 709 million
possible daily maximum NOW person-occurrences (NOW people x 365 days in the O3
season). The equivalent figure for children is 119.5 million.

Exposure results appear in Tables 1 and 2. Table 1 includes Ih daily maximum dose
exposure estimates, while Table 2 provides similar results for 8h exposure estimates.

The Ih results indicate that:

-------
1. Currently (1991), it is estimated that very few (<_0.1%) of NOW people and children
would experience a daily maximum dose (MAXD) O3 exposure @ >0.12 ppm @ an
EVR of 30 1 min'1 m'2 or higher, which is thought by many researchers to be a dose
rate of concern (13). About 42% of the NOW people so exposed are children, and
all MAXD exposures occur only on one day/O3 season.

2. Attainment of the two 8h 0.08 ppm NAAQS, both 1 Ex and 5 Ex, are expected to
decrease NOW people exposures to the Ih MAXD of concern, and eliminate children
exposures to that dose rate. Just-attaining the current standard might result in a small
increase in the number and proportion of NOW people exposed to the MAXD of
concern. However, no children are expected to be so exposed for this NAAQS.
None of the Ih MAXD exposures of concern occur on more than one day/O3 season
for these two alternative NAAQS.

While not shown in Table 1, the estimates of O3 exposures >0.12 ppm at any
exercise level are considerably higher than the numbers presented in that Table. For
instance, currently (1991) about 20,000 adults and 8,000 children are predicted to experience
one or more daily exposure >0.12 ppm when exercise level is ignored. This is about 1% of
the adult population and 2% of the children in the Miami urban area. (As compared to
<0.1% adult and 0.1% children exposure, on average, @ 30 EVR exercise.) The
corresponding estimates for non-exercising exposure when the current standard is just
attained is 62,000 (4%) for adults and 12,000 for children (4%). Similar estimates for
attainment of the 8h lEx 0.10 ppm NAAQS alternative are 137,000 (8%) for adults and
32,000 for children (10%), and are 21,000 (1%) for NOW adults and 4,000 (1%) for
children when the 8h 5Ex standard of 0.08 ppm is just attained. Thus, exercise level, and its
associated ventilation rate, at maximum dose greatly affects pNEM/O3 MAXD exposure
estimates.

The 8h results (Table 2), indicate that relatively more people may be exposed. The
level of concern here is a MAXD exposure of >0.08 ppm for 8h @ an EVR ^.15 1 min"1
m'2. (Obviously, the EVR of concern drops as the averaging period increases, just as the O3
concentration of concern drops. People cannot hold an EVR of 30 1 min"1 m'2 for an
extended period of time(13).)

The 8h results indicate that:

1. Currently (1991), no one in the NOW population-including children-is estimated to
experience an 8h MAXD of concern.

2. Attaining the next 3 NAAQS listed in Table 2 indicates that NOW-people may be
exposed to the 8h MAXD of concern. A goodly proportion of these people are
children, with the range in the estimate of the children's proportion of this exposed
population being 39.5-88.9%. The absolute and relative numbers of NOW-people
being so affected are relatively small (_<_0.4%) for the current standard (Ih lEx 0.12

-------
ppm) and the 8h 5 Ex 0.08 ppm alternative standard. The proportion of children so
affected is estimated to be 1% for these two NAAQS.

3. Attaining the 8h 1 Ex 0.10 ppm NAAQS is estimated to result in a significant number
of children being exposed at the 8h 0.08 ppm MAXD of concern: about 13,400
children (4.1% of children in the modeled part of Miami). It should be noted that a
very small percentage (0.1%) of the NOW-people exposed at this MAXD may
experience two days of exposure during the O3 season, which is the entire year in
Miami.

Putting the 8h MAXD estimates into perspective, the results indicate that currently
(1991) only 3,500 NOW people might experience an 8h exposure >0.08 ppm @ any
exercise level. (In other words, when exercise is ignored; this estimate is not shown in
Table 2.) This compares to no one estimated to experience the same O3 air quality
concentration @ 15 EVR or higher. The corresponding estimates for children are no
exposures at any exercise or at 15+ EVR. Exercise level in Miami's As Is results hardly
make a difference in the number of people exposed to the 0.12 ppm concentration level.
This contrasts to all of the other urban areas previously modeled (2-7), which had far fewer
exposures at exercise than when exercise level is ignored.
For the record, the 8h any-exercise estimates for the non-zero alternative NAAQS
are:

8h lEx 8h 5Ex Ih lEx
0.10 0.08 0.12

NOW People 428,100 90,900 242,800
(%) 22.0 4.7 12.5

Children 92,800 24,100 73,300
(%) 28.3 7.4 22.4

Contrast these estimates with those appearing in Table 2 for insight into the impact of
exercise level on exposure estimates. The any-exercise estimates are 1-1.5 orders of
magnitude higher than the _>.15 EVR estimates.

-------
exposure metric being tested. The test statistic was the maximum difference between the two
distributions (Dn); it was evaluated using a large-sample F statistic. The significance level is
0.05, using the chi-square approximation. The test indicates that the exposure indicators
shown in Tables 1 and 2 vary significantly across the six air quality scenarios with the
following exceptions.

For the Asls and 8h 5Ex 0.08 ppm pair:

NOW people and occurrences and children and children-occurrences at Ih >0.12
ppm exposures at any exercise level.

For the As Is and the current standard (Ih lEx 0.12 ppm) pair:

Children and children-occurrences at >0.12 ppm at any exercise level.

For the current standard and 8h 5Ex 0.08 ppm pair:

NOW person-occurrences, children exposed, and children-occurrences at Ih >0.12
ppm at any exercise level.

NOW people, NOW person-occurrences, and the number of people exposed on one
day/season at Ih >0.12 ppm at ^30 EVR.

For the current standard and 8h lEx 0.10 ppm pair:

Children exposed, children-occurrences, and the number of children exposed on one
day/season at 8h >0.08 ppm at _>_15 EVR and at any exercise level.

Children exposed and children-occurrences at Ih >0.12 ppm at any exercise level.

Compared with the K-S analyses undertaken for most of the previous urban areas (2-
7), the above results are unusual. Most of the non-rejects involve the "any exercise"
metrics, which are not the main exposure/exercise conditions of concern. The only
exercising metrics that could not be rejected as coming from the same distribution involved
the current standard and two alternative NAAQS. The current standard/8h SEX 0.08 pair
could not be rejected for the Ih >0.12 ppm at 30-1- EVR exposure metric. The current
standard/8h lEx 0.10 ppm pair could not be rejected for the 8h >0.08 ppm at _>_15 EVR.
However, there is no real consistent pattern among the exposure metrics and air quality
standard pairs that were tested.

-------
small samples (10 runs). These caveats aside, 19 "cannot rejects" out of 92 tests is four
times as many as are expected based on a "pure-chance" probability of rejecting a true
hypothesis (5%). This 21% cannot-reject rate is about quadruple the rates seen in most of
the urban areas that were analyzed (2-7). Unlike most of these
areas, there is not a clear-cut picture produced by the K-S testing exercise in Miami. There
is not a consistent pattern-especially at exercise-associated with the alternative NAAQS
investigated.
SUMMARY

The analyses of 5 alternative standards ozone NAAQS in the Miami area (Ih lEx
0.12 ppm; 8h lEx 0.10 pm; 8h 5Ex 0.08 ppm; 8h lEx 0.08 ppm; and 8h 5Ex 0.06 ppm)
indicate that attainment of the current standard or the 8h lEx 0.10 ppm and 8h 5Ex 0.08
ppm alternatives may not minimize exposures at exercise in the Miami urban area. These
findings apply to both the Ih and 8h estimates, but are more important for the 8h results
since more people and children are estimated to experience 8h MAXD levels of concern than
the Ih MAXD levels under these alternative NAAQS.

-------
                                 REFERENCES

1.     T. Johnson, J. Capel, and M. McCoy.  Estimation of Ozone Exposures
      Experienced by Urban Residents Using a Pro-
      babilistic Version of NEM and 1990 Population Data.  Durham, NC: IT-AQS,
      1993 (Draft).

2.     T. McCurdy.  "Los Angeles Ozone Exposure Estimates," EPA Memorandum;
      November 4, 1993.

3.     T. McCurdy, et al.  "Estimating Ozone Exposures in Philadelphia Using a Second-
      Generation Probabilistic Version of NEM." Paper A210 presented at the 87th
      Annual Meeting of AWMA; Cincinnati, June 1994.

4.     T. McCurdy, et al.  "Estimating ozone Exposures in Houston Using a Second-
      Generation Probabilistic Version of NEM." Paper presented  at the Tropospheric
      Ozone Specialty Conference, Air and Waste Management Association; Orlando,
      1994.

5.     T. McCurdy.  "Estimating Ozone Exposures in the New York Urban Area Using a
      Second-Generation Probabilistic Version of NEM."  EPA Paper, February 1994.

6.     T. McCurdy.  "Estimating Ozone Exposures in the Chicago Urban Area Using a
      Second-Generation Probabilistic Version of NEM."  EPA Paper, March 1994.

7.     T. McCurdy.  "Estimating Ozone Exposures in the Denver Urban Area Using a
      Second-Generation Probabilistic Version of NEM."  EPA Paper, March 1994.

8.     Environmental Criteria and Assessment Office.  Air Quality Criteria for Ozone and
      Photochemical Oxidants. Research Triangle Park, NC: 1994 (Draft).

9.    H.M. Richmond and T.  McCurdy.  "Use of Exposure Analysis and Risk Assessment
      in the Ozone NAAQS Review."  Paper 88-121.3 presented at the 81st Annual
      Meeting of APCA; Dallas, June 19-24, 1988.

 10.   N.L. Nagda, H.E. Rector, and M.D. Koontz.  Guidelines for Monitoring Air
      Quality.  Washington, DC: Hemisphere Pub. Co., 1987.

 11.   C.J. Weschler, H.C. Shields, and D.V. Nike.  "Indoor Ozone: Recent Findings," in:
      R. Burglund (ed.), Tropospheric Ozone in the EnTironment n. Pittsburgh: Air &
      Waste Management Association, 1992.

 12.   T. McCurdy.  "Testing pNEM/O3 Runs to Determine if a Set of 10 Runs is
       'Representative'," EPA Memorandum; April 19, 1994.

                                        10

-------
13.    W.C. Adams.  Measurement of Breathing Rate and Volume in Routinely
      Performed Daily Activities.  Davis CA:  University of California, 1993.

14.    M. Absil, P. Narducci, R. Whitfield, and H.M. Richmond.  "Chronic Lung Injury
      Risk Estimates for Urban Areas Having Ozone Patterns Similar to Those in the
      Northeast."  Paper presented at the AWMA Specialty Conference on Tropospheric
      Ozone and the Environment IT, Atlanta, November 4-7, 1991.
                                         11

-------
               Table 1.  ESTIMATES OF Ih "MAXIMUM DOSAGE" EXPOSURES IN MIAMI TO O3 CONCENTRATIONS EXCEEDING
            0.12 PPM AND EVRa EQUALED OR EXCEEDED 30 ONE OR MORE TIMES PER OZONE SEASON UNDER
                                                SIX AIR QUALITY SCENARIOS
Statistic
TOTAL EXPOSED POPULATION
Mean Estimate of the Number of People
Percent of Total Population
Range in this percentage for all runs
Mean Est. of Person-Occurrences
Percent of Total Per-Occurrences
Range in this percentage for all runs
Mean Est. of Occurrences/Person Exposed
Number of Days/Season Exposed (% Exposed)
1 Day
2 Days
3 Days
>3 Days
CHILDREN
Mean Estimate of the Number of Children
Percent of Children Population
Range in this percentage for all runs
Percent of Total Exposed Population
Mean Estimate of Children-Occurrences
Percent of total Child-Occurrences
Range in this percentage for all runs
Mean Estimate of Occurrences/Child Exposed
Number of Days Ex posed /Season (% Exposed)
1 Day
2 Days
3 Days
>3 Dnys

As Is
Situation
(1991)

500
*
0.0-0.1
500
@
0.0®
1. 00

100
0
0
0

200
0.1
0.0-0.6
41.8
200
@
0.0-@
1.00

100
0
0
0

Attain
8h lEx
0.10 ppm

300
*
0.0-0.1
300
©
0.0-®
1.00

100
0
0
0

0
0
-
0
0
0
-
-

-
-
-
-

Attain
8h 5Ex
0.08 ppm

300
*
0.0-0.1
300
@
0.0-@
1.00

100
0
0
0

0
0
-
0
0
0
-
-

-
-
-
-

Attain
Ih 1 Ex
0.12 ppm

900
+
0.0-0.3
900
@
0.0-@
1.00

100
0
0
0

0
0
-
0
0
0
-
-

-
-
-
—

Attain
8h 1 Ex
0.08 ppm

0
0
~
0
0
~
-

-
-
-
"

0
0
~
-
0
0
~
-

-
•
-
"

Attain
8h 5 Ex
0.06 ppm

0
0

0
0
"
-

-
-
-


0
0
"
-
0
0
™
-

-
•
-


*Less than 0.05% but > 0.005%
©Less than 0.005% but >0.

-------
          Table 2  ESTIMATES OF 8h "MAXIMUM DOSAGE' EXPOSURES IN MIAMI TO 03 CONCENTRATIONS EXCEEDING
       0.08 PPM AND EVR" EQUALED OR EXCEEDED 15 ONE OR MORE TIMES PER OZONE SEASON UNDER
oist ms\ \^wnu*» • v*^a-*i " ••»«vw
Statistic


TOTAL NON-OUTDOOR WORKER EXPOSED POPULATION
Mean Estimate of the Number of People
Percent of Total Population
Range in this percentage for all runs
Mean Est. of person-Occurrences
Percent of total Per. -Occurrences
Range in this percentage for all runs
Mean Est. of Occurrences/Person Exposed
Number of Days/Season Exposed (% Exposed)
1 Day
2 Days
3 Days
4 Days
>4 Days
CHILDREN
Mean Estimate of the Number of Children
Percent of Children Population
Range in this percentage for all runs
Percent of Total Exposed Population
Mean Estimate of Children-Occurrences
Percent of Total Child-Occurrences
Range in this percentage for all runs
Mean Estimate of Occurrences/Child Exposed
Number of Days/Season (% Exposed)
1 Day
2 Days
3 Days
4 Days
>4 Days
	 ^^
As Is
Situation
(1991)
0
0
0
0
-
-



-
0
0
-
0
0
-
-
.
-
-
-
-
-^
Attain
8h lEx
O.lOppm
23.900
1.2
0.5-2.2
23,900
@
@-@
1.00
100



0
13,400
4.1
0.0-11.9
56.1
13,400
*
0.0-*
1.00
100
0
0
0
0
-
Attain
8h5Ex
0.08 ppm
3,600
0.2
0.0-1.5
3,600
@
0.0-®
1.00
100
o
\J
n
vJ
o
\J

3,200
1.0
0.0-9.0
88.9
3,200
@
0.0-@
1.00
100
0
0
0
0
	 !^^
Attain
In 1 Ex
0.12 ppm
8.600
0.4
M.I
8,600
@
@-@
1.00
100
o
o
o


3,400
1.0
0.0-6.5
39.5
3,400
©
0.0-@
1.00
100
0


	 -
=====
Attain
8h 1 Ex
0.08 ppm
0
0
0
0
-
-




0
0
-
0
0

-
-
"


=
=====
Attain
8h5 Ex
0.06 ppm
0
0
0
0
-
-
.
m


0

-
0


-
-



—
Notes: 'Equivalent ventilation rate. It is equal to ventilation rate/body surface area; units are in 1 mm m
*Less than 0.05% but > 0.005%
©Less than 0.005% but >0.
                                                    13

-------
                                                                A-210
ESTIMATING OZONE EXPOSURES IN PHILADELPHIA USING A SECOND-
   GENERATION PROBABILISTIC VERSION OF NEM
                             Thomas McCurdy
                     U.S. Environmental Protection Agency
                      Research Triangle Park, NC 27711
                               Ted Johnson
                                Jim Capel
                            Michael McCoy, Jr.
                  International Technology-Air Quality Services
                          Durham, NC 27707-6208
                               February 1994
  FILE:PHILA.

-------
A-210
INTRODUCTION

The U.S. Environmental Protection Agency (EPA) is currently reviewing the ozone
(03) primary national ambient air quality standard (NAAQS) to determine if the existing
standard protects public health with an adequate margin of safety. To evaluate whether
alternative O3 NAAQSs provide an adequate margin of safety, the EPA Office of Air Quality
Planning and Standards (OAQPS) assesses such factors as the nature and severity of health
effects associated with O3 exposure, the degree of total human exposure (i.e., indoor and
outdoor) to O3 and the risks (probabilities) of ozone-related health effects occurring in the
exposed population when alternative O3 NAAQS are just attained. This paper discusses an
EPA analysis of O3 exposure estimates for parts of the Philadelphia metropolitan area which
employed the pNEM/O3 model developed by International Technology - Air Quality Services
(IT-AQS) A report by Johnson, et al. (1) describes this model in detail and presents initial
results of applying it to nine U.S. urban areas, including Philadelphia. Exposure estimates
for parts of the Los Angeles area are discussed in an EPA memorandum (2) that was recently
reviewed by the Clean Air Science Advisory committee (CASAC) at a public meeting.
Additional analyses of O3 exposure estimates for the remaining seven urban areas (see
below) that were modeled will be forthcoming.

OVERVIEW

Pursuant to section 108 of the Clean Air Act, the Environmental Criteria Assessment
Office (ECAO) within the Office of Research and Development (ORD) is preparing a criteria
document (4) for O3. This document will be a comprehensive summary and critical
assessment of the latest scientific research relevant to assessing the effects of a criteria
pollutant. The document will be reviewed by CASAC, part of the EPA's independent
Science Advisory Board, and by the general public. Upon completion of the criteria
document, OAQPS will prepare a Staff Paper that summarizes and integrates scientific
information presented in the criteria document considered to be the most relevant to decisions
on alternative NAAQS and various analyses (e.g., air quality, exposure, and risk). Figure 1
illustrates the relationships among scientific research, criteria document development, the
OAQPS Staff Paper, and exposure and risk assessment.

An important aspect of health risk assessment is an analysis of population exposure.
Such an analysis provides the population basis for "headcount risk" estimates developed as
part of EPA's NAAQS risk assessment work. Estimates of population exposure
also provide important information regarding the number of people that may experience
varying levels of O3 as they go through their normal daily life.

-------
A-210

values for certain variables inherent in the model. These versions of NEM are called
pNEM, an acronym for probabilistic" NEM. to differentiate among the various pollutant-
specific versions of pNEM, the one described here is denoted as pNEM/O3.

People living within each exposure district, as estimated by the U.S. Bureau of
Census in 1990, are assigned to a single discrete point, the population centroid. O3 air
quality within each exposure district is estimated for each hour of the O3 season using data
from a nearby monitoring site. Because O3 in the ambient air generally is modified
considerably when entering a building or vehicle, outdoor O3 estimates are adjusted using a
mass-balance model to account for 3 indoor microenvironments. These microenvironments
are: indoors-at-home, indoors-other, and within a motor vehicle. The mass-balance model
used in pNEM/O3 is a simplified version of the generalized Nagda, Rector, and Koontz
(NRK) model (5). For our use, this model was revised to incorporate the assumption that
indoor decay rate is proportional to indoor O3 concentration. It was further revised to
incorporate assumptions concerning ozone decay rates suggested by Weschler, et al (6). The
resulting revised mass-balance model is fully described in Ref. 1.
Because both the intake dose received and susceptibility to effects of pollution may
vary with age, occupation, and intensity of exercise, the total population of each study area is
divided into 9 age-occupation (A-O) groups. Each A-O group is further subdivided into
cohorts depending upon (1) the type of air conditioning system present in the home, if any,
and (2) home-to-work commuting patterns.

While many types of exposure metrics are provided by pNEM/03, only those for the
non-outdoor working population in Philadelphia are discussed in this paper. Exposure
estimates for children in Philadelphia are not included in the results discussed here.
Estimates for outdoor workers—a major population sub-group of concern from an O3
exposure perspective—will be completed in the Fall of 1994.

-------
A-210

AIR QUALITY SCENARIOS INVESTIGATED

Six air quality scenarios were modeled for Philadelphia (and the other 8 urban areas);
the alternative NAAQS scenarios all are on a daily maximum basis:

1. 1991 air quality-the "as is" scenario
2. Attain In lEx 0.12 ppm~the current standard
3. Attain 8h lEx 0.08 ppm
4. Attain 8h lEx 0.10 ppm
5. Attain 8h 5Ex 0.06 ppm
6. Attain 8h 5Ex 0.08 ppm

The iterative adjustment procedure used by IT-AQS to simulate just-attaining the 5
alternative NAAQS listed above involves a complex procedure of fitting Weibull distributions
to the 1991 O3 data, modifying parameters of the equations so that the "worst" monitor just
achieves the desired standard, and then re-specifying the entire O3 pattern for all districts in
Philadelphia. The procedure is described in detail by Johnson, et al. (1).

PHILADELPHIA RESULTS

Results are based upon 10 runs (realizations) of the pNEM/O3 model for each air
quality scenario. Based on a previous analysis of sets of 10-run results versus a 108-run
result (7), the authors think that results from only 10 runs of the model will adequately
predict the mean and variance observed in 100 or more runs of pNEM/O3. Ten runs of the
model will not, however, adequately represent the entire range of possible outcomes.
Additional runs can only increase this range. This is an obvious shortcoming of the analyses
presented here, but limited resources preclude undertaking more runs.

This paper includes tabular data on (1) the number and percent of people exposed,
and (2) the number and percent of person-occurrences for the adult, non-outdoor worker
population in Philadelphia. (A person-occurrence is the metric that focuses on an event:
when the exposure-of-concern is reached regardless of who reaches it. One hundred person-
occurrences could be 10 people experiencing the exposure 10 times, 100 people experiencing
it once, one person experiencing it 100 times, or any combination thereof.) Even though
data are discussed here regarding the absolute and relative number (percentage) of people

-------
A-210

exposed, it must be noted that "people" should not be used as an exposure metric. The
cohort-specific time/activity data base used for pNEM/O3 modeling (and for aU other
exposure models that uses human activity information) provides no more than three days of
data from a single individual. Thus, rigorous inferences from these data rigorously can be
made only to "person-days" of exposure or "person-occurrences" of exposure, where each
occurrence relates to a Ih, 8h, or 24h time period in pNEM/O3. Only when activity data
are available sequentially for an entire year for individuals in a cohort can inferences be
made to people (or children, or outdoor workers as the case may be). This point is usually
ignored in the general exposure assessment literature.

Ventilation rate is a very important factor in estimating total dose for any air pollutant
(8). The ventilation rate metric used in pNEM/O3 is "equivalent ventilation rate" (EVR), or
ventilation rate (in liters per minute) divided by body surface area (BSA, in units of m2).
EVR has units of 1 min"1 m'2. It is a surrogate indicator for the number of O3 molecules that
enter the oral-nasal cavities per unit time period.

The Philadelphia population analyzed here includes 3,785,810 NOW (non-outdoor
worker) people and 810,163,340 possible daily maximum NOW person-occurrences (NOW
people x 214 days in the O3 season).

Exposure results appear in Tables 1 and 2. Table 1 includes Ih daily maximum dose
exposure estimates, while Table 2 provides similar results for 8h exposure estimates.

The results indicate that:

1. Currently (1991), fewer than 2% of the NOW population experience a daily
maximum dose (MAXD) 03 exposure @ >0.12 ppm @ an EVR of 30 1 min"1 m"2 or
higher. This exposure/breathing rate combination is thought by many researchers to
be a dose rate of concern (9). Almost all--96%~of these MAXD exposures occur
only on one day during the O3 season.

2. Such MAXD exposures are greatly reduced whenever one of the alternative standards
are attained. For the relatively few people having a MAXD exposure >0.12 ppm @
J>_30 EVR when an alternative NAAQS is attained, all of them experience only one
such Ih exposure per season.

-------
A-210
While not shown in Table 1, the results of O3 exposures >0.12 ppm at any exercise
level are considerably higher than the exposure numbers presented in the Table. For
instance, currently (1991) about 2.7 million people are predicted to experience one or more
daily exposures >0.12 ppm when exercise level is ignored. This is about 70% of the total
NOW population in Philadelphia. Thus, exercise level, and its associated ventilation rate, at
maximum dose greatly affects pNEM/O3 MAXD exposure estimates.

The same findings apply to the 8h results (Table 2), although relatively more people
are affected. The level of concern here is MAXD exposure @ >0.08 ppm for 8h @ an
EVR _>_15 1 min"1 m"2. (Obviously, the EVR of concern decreases as the averaging period
increases, just as the O3 concentration of concern decreases.) People cannot maintain an
EVR of 30 1 min'1 m'2 for an extended period of time (8).

The 8h results indicate that:

1. Currently, over 6% of the NOW population experience a daily maximum dose O3
exposure level of concern. About 15% of these people may experience such a
MAXD exposure on more than one day during the O3 season. One percent of the
applicable population may experience the exposure of concern on 3 days.

2. Attaining any of the alternative NAAQS reduces the MAXD exposure to the level of
concern, although about 1 % of the NOW population would experience that exposure
even if the current NAAQS (Ih lEx 0.12 ppm) or the 8h lEx 0.10 ppm standards
was attained; however, almost all of these exposures occur on only one day.

Consistent with the Ih results, almost everyone (90%) in Philadelphia currently
(1991) experiences one or more 8h exposures >0.08 ppm at maximum daily dose when
ventilation level is ignored. (This result is not shown in Table 2.) This contrasts to the
6.4% of the NOW population who experience the same 8h O3 concentration at exercise.
Again, exercise levels-expressed as EVR-greatly affect pNEM/O3 exposure results.

Note that all of the estimates exhibit variability, reflecting the Monte Carlo sampling
process that produces a different estimates for any run. The range in the estimates are
between a factor of 2-to-3 in most cases, which seems reasonable.

Statistical tests of the exposure distributions were undertaken using a two-sample
Kolmogorov-Smirnov (K-S) non-parametric test. The null hypothesis was that the cumulative
frequency distributions for the various air quality scenarios are the same for each exposure
metric being tested. The test statistic was the maximum difference between the two
distributions (Dn); it was evaluated using a large-sample F statistic. The significance level
was 0.05, using the chi-square approximation. The tests indicate that all of the exposure

-------
A-210

indicators shown in Tables 1 and 2 vary significantly across the six air quality scenarios with
the following exceptions:

Persons with an 8h MAXD exposure _>0.08 ppm @ _>_15 EVR for the Ih lEx 0.12
ppm and 8h lEx 0.10 ppm NAAQS.

Person-occurrences for the same exposure condition and air quality scenarios as
above.

The number of people with 1 day of exposure >0.08 ppm @ ^.15 EVR for the same
two air quality scenarios.

Note that these are not strong statistical tests and that some of the K-S test
requirements were not met. Random sampling was not used. The sample "observations"
simply are results of a simulation model; the population of interest, then, can only be the set
of possible outcomes produced by the model. A large-sample test statistic was applied to
small samples (10 or 11 runs). In addition, three "cannot rejects" out of 74 tests is about
4% of the total tests made—about the same value as the "pure-chance" probability. It is of
interest, however, to note that all three exposure metrics that were not rejected using the K-S
test involved the same paired air quality scenarios, which supports the finding that those
scenarios do not differ significantly with respect to estimated O3 exposures.

SUMMARY

The analyses of 5 alternative standards ozone NAAQS (Ih lEx 0.12 ppm; 8h lEx
0.10 pm; 8h 5Ex 0.08 ppm; 8h lEx 0.08 ppm; and 8h 5Ex 0.06 ppm) indicate that
attainment of each standard probably will sharply reduce the percentage of the exposed
population experiencing ozone exposures thought to increase health risks. With the exception
of the Ih lEx 0.12 and 8h lEx 0.10 alternatives, the percentage of non-outdoor worker
population exposed at relatively high exercise levels one or more times to the level of
concern is well below 1%. The 8h exposures to 0.08 ppm or higher for the NOW
population are 1.1% for the 8h 1 Ex 0.05 standard and 0.9% for the 0.12 standard. A K-S
test cannot distinguish between these two standards for most exposure indicators. (This
finding is not applicable to the Los Angeles results (2).) The analyses also found that for all
air quality scenarios that were examined, just attaining an alternative O3 standard reduces
multiple-day exposures for the air quality/exercise levels of interest.

-------
                                                                       A-210

                                 REFERENCES

1.     T. Johnson, J. Capel, and M. McCoy. Estimation of Ozone Exposures
      Experienced by Urban Residents Using a Pro-
      babilistic Version of NEM and 1990 Population Data.  Durham, NC:  IT-AQS,
      1993 (Draft).

2.     T. McCurdy.  "Los Angeles Ozone Exposure Estimates," EPA Memorandum;
      November 4, 1993.

3.     H.M. Richmond and T. McCurdy. "Use of Exposure Analysis and Risk Assessment
      in the Ozone NAAQS Review."  Paper 88-121.3 presented at the 81st  Annual
      Meeting of APCA; Dallas, June 19-24, 1988.

4.     Environmental Criteria and Assessment Office.  Air Quality Criteria for Ozone and
      Photochemical Oxidants.  Research Triangle Park,  NC: 1994 (Draft).

5.     N.L. Nagda, H.E. Rector, and M.D. Koontz. Guidelines for Monitoring Air
      Quality.  Washington, DC: Hemisphere  Pub. Co., 1987.

6.     C.J. Weschler, H.C. Shields, and D.V. Nike. "Indoor Ozone: Recent Findings," in:
      R. Burglund (ed.), Tropospheric Ozone in the Environment II.  Pittsburgh: Air &
      Waste Management Association, 1992.

7.     T. McCurdy.  "Testing pNEM/O3 Runs to Determine if a Set of 10 Runs is
      'Representative',"  EPA Memorandum;  April 19, 1994.

8.    W.C. Adams.  Measurement of Breathing Rate and Volume in Routinely
      Performed Daily Activities.  Davis CA: University of California, 1993.

9.    M. Absil, P. Narducci, R. Whitfield, and H.M. Richmond.   "Chronic Lung Injury
      Risk Estimates for Urban Areas Having  Ozone Patterns Similar to Those in the
      Northeast."  Paper presented at the AWMA Specialty Conference on Tropospheric
      Ozone and the Environment II; Atlanta, November 4-7, 1991.

-------
         Table 1  ESTIMATES OF Ih "MAXIMUM DOSAGE" EXPOSURES IN PHILADELPHIA TO O3 CONCENTRATIONS EXCEEDING
         0 12 PPM AND EVR" EQUALED OR EXCEEDED 30 ONE OR MORE TIMES PER OZONE SEASON UNDER
                                               SIX AIR QUALITY SCENARIOS
Statistic
TOTAL EXPOSED POPULATION
Mean Estimate of the Number of People
Percent of Total Population
Range in this percentage for all runs
Mean Est. of Person-Occurrences
Percent of Total Per. -Occurrences
Range in this percentage for all runs
Mean Est. of Occurrences/Person Exposed
Number of Days Exposed/Season (% Exposed)
1 Day
2 Days
>2 Days
As Is
Situation
(1991)

70,000
1.8
0.6-4.1
72,800
.01
@-.02
1.02

96
4
0
Attain
8h lEx
0.10 ppm

4,100
0.1
0.0-0.7
4,100
@
0.0-®
1.00

100
0
0
Attain
8h 5Ex
0.08 ppm

0
0
~
0
0
-
-

-
-

Attain
Ih 1 Ex
0.12 ppm

300
*
0.0-*
300
@
0.0-®
1.00

100
0
0
Attain
8h 1 Ex
0.08 ppm

0
0

0
0
~
-

-
-

Attain
8h5 Ex
0.06 ppm

0
0

0
0
~
-

-
•

Notes:
'Equivalent ventilation rate.
*Less than 0.05% but > 0.
©Less than 0.005% but  >0.
It is equal to ventilation rate -r body surface area; units are in 1  min"1 m"

-------
         Table 2  ESTIMATES OF 8h "MAXIMUM DOSAGE" EXPOSURES IN PHILADELPHIA TO 03 CONCENTRATIONS EXCEEDING
        0.08 PPM AND EVRa EQUALED OR EXCEEDED 15 ONE OR MORE TIMES PER OZONE SEASON UNDER
*JL/\ mix \^\j r\t^i i t *j^»i^i it »mx»vi~»
Statistic
TOTAL NON-OUTDOOR WORKER EXPOSED POPULATION
Mean Estimate of the Number of People
Percent of Total Population
Range in this percentage for all runs
Mean Est. of Person-Occurrences
Percent of Total Per. -Occurrences
Range in this percentage for all runs
Mean Est. of Occurrences/Person Exposed
Number of Days Exposed/Season (% Exposed)
1r\i,i
Day

2 Days
3 Days

>3 Days
-
As Is
Situation
(1991)
243.700
6.4
4.7-9.6
289,100
0.04
0.03-0.05
1.19
85






—
Attain
8h lEx
0.10 ppm
41,700
1.1
0.2-2.2
42,500
0.01
1.01
99

i
i




Attain
8h5Ex
0.08 ppm
1.100
*
0.0-0.1
1,100
@
0.0-®
1.00
100


n
\J


Attain
Ih 1 Ex
0.12 ppm
35,000
0.9
*-2.4
35,000
@
@-@
1.00
100

o
n
V
o

=====
Attain
8h 1 Ex
0.08 ppm
500
*
0.0-0.1
500
0.0-®
1.00
100


o



=====
Attain
8h5 Ex
0.06 ppm
0
0
0

-
.





=====
Notes: . . . . i .1
"Equivalent ventilation rate.  It is equal to ventilation rate/body surface area; units are in 1  min'1 nV2
*Less than 0.05% but > 0.
©Less than 0.005% but >0.

-------
ESTIMATING OZONE EXPOSURES IN THE ST. LOUIS URBAN AREA USING
    A SECOND-GENERATION PROBABILISTIC VERSION OF NEM
                            Thomas McCurdy
                    U.S. Environmental Protection Agency
                     Research Triangle Park, NC 27711
                              March 1994
 File: STLOUIS

-------
INTRODUCTION

The U.S. Environmental Protection Agency (EPA) currently is reviewing the ozone
(O3) primary national ambient air quality standard (NAAQS) to determine if the existing
standard protects public health with an adequate margin of safety. In order to evaluate
whether alternative O3 NAAQSs provide an adequate margin of safety, the EPA Office of
Air Quality Planning and Standards (OAQPS) assesses such factors as the nature and severity
of health effects associated with O3 exposure, the degree of total human exposure (i.e.,
indoor and outdoor) to O3, and the risks (probabilities) of ozone-related health effects
occurring in the exposed population when alternative O3 NAAQS are just attained. This
paper discusses an EPA analysis of O3 exposure estimates for parts of the St. Louis
metropolitan area which employed the pNEM/O3 model developed by International
Technology-Air Quality Services (IT-AQS). A report by Johnson, et al. (1) describes this
model in detail and presents initial results of applying it to nine U.S. urban areas, including
St. Louis. Exposure estimates for parts of the Los Angeles area are discussed in an EPA
memorandum (2) that was reviewed by the Clean Air Science Advisory Committee (CASAC)
at a public meeting. Exposure estimates for the Philadelphia area are described in an
AWMA paper (3). Exposure estimates for Houston are described in an AWMA specialty
conference paper (4). New York, Chicago, Denver, and Washington, DC results are
contained in EPA papers (5-9).

OVERVIEW

Pursuant to section 108 of the Clean Air Act, the Environmental Criteria Assessment
Office (ECAO) within the Office of Research and Development (ORD) is preparing a criteria
document (10) for O3. This document will be a comprehensive summary and critical
assessment of the latest scientific research relevant to assessing the effects of a criteria
pollutant. The document will be reviewed by CASAC, part of the EPA's independent
Science Advisory Board, and by the general public. Upon completion of the criteria
document, OAQPS will prepare a Staff Paper that summarizes and integrates scientific
information presented in the criteria document considered to be most relevant to decisions on
alternative NAAQS and various analyses (e.g., air quality, exposure, and risk).

An important aspect of health risk assessment is an analysis of population exposure
(11). Such an analysis provides the population basis for "headcount risk" estimates
developed as part of EPA's NAAQS risk assessment work. Estimates of population exposure
also provide important information regarding the number of people that may experience
varying levels of O3 as they go through their daily activities.

-------
an acronym for "probabilistic" NEM. To differentiate among the various pollutant-specific
versions of pNEM, the one described here is denoted as pNEM/O3.

People living within each exposure district, as estimated by the U.S. Bureau of
Census in 1990, are assigned to a single discrete point, the population centroid. O3 air
quality within each exposure district is estimated for each hour of the O3 season using data
from a nearby monitoring site. Because O3 in the ambient air generally is modified
considerably when entering a building or vehicle, outdoor O3 estimates are adjusted using a
mass-balance model to account for 3 indoor microenvironments. These microenvironments
are: indoors-at-home, indoors-other, and within a motor vehicle. The mass-balance model
used in pNEM/O3 is a simplified version of the generalized Nagda, Rector, and Koontz
(NRK) model (12). For our use, this model was revised to incorporate the assumption that
indoor decay rate is proportional to indoor O3 concentration. It was further revised to
incorporate assumptions concerning ozone decay rates suggested by Weschler, et al (13).
The resulting revised mass-balance model is fully described in Ref. 1.

Because both the intake dose received and susceptibility to effects of pollution may
vary with age, occupation, and intensity of exercise, the total population of each study area is
divided into nine age-occupation (A-O) groups. Each A-O group is further subdivided into
cohorts depending upon (1) the type of air conditioning system present in the home, if any,
and (2) home-to-work commuting patterns.

While many types of exposure metrics are provided by pNEM/O3, only those for the
non-outdoor working population in St. Louis are discussed in this paper. Exposure estimates
for children are included in the results discussed here. Estimates for outdoor workers~a
major population sub-group of concern from an O3 exposure perspective—will be completed
in the Fall of 1994.

AIR QUALITY SCENARIOS INVESTIGATED

Six air quality scenarios were modeled for the St. Louis urban area (and the other
eight urban areas); the scenarios are all on a daily maximum basis. They are:

1. 1990 air quality-the "as is" scenario
2. Attain Ih lEx 0.12 ppm-the current standard
3. Attain 8h lEx 0.08 ppm
4. Attain 8h lEx 0.10 ppm
5. Attain 8h 5Ex 0.06 ppm
6. Attain 8h 5Ex 0.08 ppm

The iterative adjustment procedure used by IT-AQS to simulate just-attaining the 5
alternative NAAQS listed above involves a complex procedure of fitting Weibull equations to
the 1990 O3 data, modifying parameters of the equations so that the "worst" monitor just
achieves the desired standard, and then re-specifying the entire O3 pattern for all districts in
St. Louis. The procedure is described in detail by Johnson, et al. (1).

St. Louis had relatively good O3 air quality in 1990 in that its "design value"-the
second-highest Ih daily value measured at any site in the urban area-was 0.130 ppm. This
value is only slightly above the current O3 standard concentration of 0.12 ppm.

ST. LOUIS RESULTS

Results are based upon 10 runs (realizations) of the pNEM/O3 model for each air
quality scenario. Based on a previous analysis of sets of 10-run results versus a 108-run
result (14), the author believes that results from only 10 runs of the model will adequately
predict the mean and variance observed in 100 or more runs of pNEM/O3. Ten runs of the
model will not, however, adequately represent the entire range of possible outcomes.
Additional runs can only increase this range. This is an obvious shortcoming of the analyses
presented here, but limited resources preclude undertaking more runs.

-------
       This paper includes tabular data on (1) the number and percent of people and children
exposed, and (2)  the number and percent of person-occurrences  for children and the adult,
non-outdoor worker population in St. Louis.  (A person-occurrence is the metric that focuses
on an event: when the exposure-of-concern is reached regardless of who reaches  it.  One
hundred person-occurrence could be 10 people experiencing the  exposure 10 times, 100
people experiencing it once, one person experiencing it 100 times, or any combination
thereof.)  Even though data are discussed here regarding  the absolute and relative number
(percentage) of people exposed,  it must be noted that "people" should  not be used as an
exposure metric.  The cohort-specific activity-0.12 ppm @ an EVR
       of 30 1 min"1 m"2 or higher, which is thought by many researchers to be a dose rate
       of concern (15).  About 57% of the NOW people so exposed are children, and all
       MAXD exposures occur only on one day/O3 season.

-------
2. Attainment of the alternative NAAQS investigated is expected to substantially
decrease or eliminate NOW people and children exposures to the Ih MAXD of
concern. None of the Ih MAXD exposures of concern occur on more than one
day/O3 season for the alternative NAAQS.

The 8h results (Table 2), indicate that many more people may be exposed but that the
overall proportion of the total population exposed to the MAXD is small. The level of
concern here is a MAXD exposure of >0.08 ppm for 8h @ an EVR _>_15 1 min'1 m'2.
(Obviously, the EVR of concern drops as the averaging period increases, just as the O3
concentration of concern drops. People cannot hold an EVR of 30 1 min"1 m"2 for an
extended period of time (16).)

The 8h results indicate that:

1. Currently (1990), < 1 % of the NOW population and about 4% of the children are
estimated to experience the 8h MAXD of concern. These exposures occur on only
one day/season.
2. Attaining the alternative NAAQS that were investigated do not guarantee that there
will be no 8h MAXD exposures of concern-except for the last two alternatives shown
in Table 2. Attaining the 8h lEx 0.10 ppm O3 standard in St. Louis could result in
almost 2% of children exposed to the concern dose rate. Attaining the current
standard (Ih lEx 0.12 ppm) could result in almost 2% of children being so exposed at
exercise. Both of these estimates are relatively high when compared to results from
some of the other urban areas than were modeled (2-9).

Putting the 8h MAXD estimates into perspective, the results indicate that currently
(1990) 332,700 NOW people might experience an 8h exposure >0.08 ppm @ any exercise
level. (In other words, when exercise is ignored; this estimate is not shown in Table 2.)
This compares to 13,200 people estimated to experience the same O3 air quality
concentration @ 15 EVR or higher. The corresponding estimates for children are 51,700
exposures at any exercise level versus 11,700 at 15 + EVR. Thus, exercise level certainly
makes a large difference when 8h exposures to specific O3 concentration levels are being
estimated.

-------
For the record, the 8h any-exercise estimates for the non-zero alternative NAAQS
are:
NOW People
(%)

Children
8h lEx
0.10
262,900
15.4
64,100
20.5
8h 5Ex
0.08
78,200
4.6
25,900
8.3
Ih lEx
0.12
280,900
16.5
73,300
23.4
8h lEx
0.08
3,300
0.2
200
0.1
Contrast these estimates with those appearing in Table 2 for insight into the impact of
exercise level on exposure estimates. The any-exercise estimates are 1-1.5 orders of
magnitude higher than the _>.15 EVR estimates.

For the Asls and current standard (Ih 1 Ex 0.12 ppm) pair:

All exposure indicators that were tested except for the NOW population and NOW-
occurrences exposed to Ih >0.12 ppm at any exercise level.

For the Asls and 8h lEx 0.10 ppm pair:

All exposure indicators that were tested except for the Ih >0.12 ppm exposures for
the NOW population and for children at any exercise level, and the 8h >0.08 ppm
exposures for children at any exercise level.

For the current standard (Ih lEx 0.12 ppm) and 8h lEx 0.10 ppm pair:

All exposure indicators that were tested except for the NOW population and NOW-
occurences to 0. 12 ppm at any exercise level and at exercise _X30 EVR, and for

-------
children and children-occurrences exposed to a MAXD of 0.12 ppm at exercise >30
EVR.

For the 8h lEx 0.08 ppm and 8h 5Ex 0.08 ppm pair:

Children and children-occurrences exposed to the Ih MAXD of 0.12 ppm at any
exercise level.

The results of the K-S testing procedure undertaken in St. Louis are similar to those
for Miami (8) and Washington, DC (9). Many exposure indicator pairs could not be
distinguished with respect to attaining some of the alternative NAAQS in both areas. These
standards do not make much impact on human exposure for the indicators listed above.
However, there is no real consistent pattern among the exposure metrics and air quality
standard pairs that were tested.

Note that these are not strong statistical tests and that some of the K-S test
requirements were not met. Random sampling was not used. The sample "observations"
simply are results of a simulation model; the population of interest, then, can only be the set
of possible outcomes produced by the model. A large-sample test statistic was applied to
small samples (10 runs). These caveats aside, 32 "cannot rejects" out of 103 tests is six
times as many as are expected based on a "pure-chance" probability of rejecting a true
hypothesis (5%). This 31% cannot-reject rate is about quadruple the rates seen in many of
the urban areas that were analyzed (2-9). Unlike these
areas, there is not a clear-cut picture produced by the K-S testing exercise in St. Louis.
There is not a consistent pattern among the exposure metrics associated with the alternative
NAAQS investigated.
SUMMARY

The analyses of 5 alternative standards ozone NAAQS in the St. Louis area (Ih lEx
0.12 ppm; 8h lEx 0.10 pm; 8h 5Ex 0.08 ppm; 8h lEx 0.08 ppm; and 8h 5Ex 0.06 ppm)
indicate that attainment of any of the alternatives will reduce-but not completely eliminate-
exposures to the 8h MAXD of concern. In particular, attaining the current standard (Ih lEx
0.12 ppm) and the 8 1 Ex 0.10 ppm alternative NAAQS does not protect a small, but
important, percentage of children from experiencing the 8h MAXD of 0.08 ppm at _>_15
EVR.

The reader should remember the early caveat about the downward bias inherent in
pNEM results. While every effort has been made to present "best estimates," it is inherently
difficult to predict the upper end of the population exposure distribution using human activity
data bases mat consist of only 1-3 days of information and that also under-represent adults
and children who exercise regularly. Year-long activity data from all segments of the
population are needed to adequately estimate the true population exposure distribution.

-------
                                 REFERENCES

1.     T. Johnson, J. Capel, and M. McCoy.  Estimation of Ozone Exposures
      Experienced by Urban Residents Using a Pro-
      babilistic Version of NEM and 1990 Population Data.  Durham, NC: IT-AQS,
      1993 (Draft).

2.     T. McCurdy.  "Los Angeles Ozone Exposure Estimates," EPA Memorandum;
      November 4, 1993.

3.     T. McCurdy, et al.  "Estimating Ozone Exposures in Philadelphia Using a Second-
      Generation Probabilistic Version of NEM."  Paper A210 presented at the 87th
      Annual Meeting of AWMA; Cincinnati, June 1994.

4.     T. McCurdy, et al.  "Estimating ozone Exposures in Houston Using a Second-
      Generation Probabilistic Version of NEM."  Paper presented  at the Tropospheric
      Ozone Specialty Conference, Air and Waste Management Association; Orlando,
      1994.

5.     T. McCurdy.  "Estimating Ozone Exposures in the New York Urban Area Using a
      Second-Generation Probabilistic Version of NEM."  EPA Paper, February' 1994.

6.     T. McCurdy.  "Estimating Ozone Exposures in the Chicago Urban Area Using a
      Second-Generation Probabilistic Version of NEM."  EPA Paper, March 1994.

7.     T. McCurdy.  "Estimating Ozone Exposures in the Denver Urban Area Using a
      Second-Generation Probabilistic Version of NEM."  EPA Paper, March 1994.

8.     T. McCurdy. "Estimating Ozone Exposures in the Miami Urban Area Using a
      Second-Generation Probabilistic Version of NEM."  EPA Paper, March 1994.

9.    T. McCurdy.  "Estimating Ozone Exposures in the Washington, DC Urban Area
      Using a Second-Generation Probabilistic Version of NEM." EPA Paper, March
       1994.

 10.   Environmental Criteria and Assessment Office.  Air Quality Criteria for Ozone and
      Photochemical  Oridants. Research Triangle Park, NC:  1994 (Draft).

 11.   H.M.  Richmond and T. McCurdy.  "Use of Exposure Analysis and Risk Assessment
      in the Ozone NAAQS Review." Paper 88-121.3 presented at the 81st Annual
      Meeting of APCA; Dallas, June 19-24, 1988.

 12.    N.L. Nagda, H.E.  Rector, and M.D. Koontz.  Guidelines for Monitoring Air
       Quality.  Washington, DC: Hemisphere Pub. Co., 1987.

-------
13.    C.J. Weschler, H.C. Shields, and D.V. Nike.  "Indoor Ozone: Recent Findings," in:
      R. Burglund (ed.), Tropospheric Ozone in the Environment II. Pittsburgh:  Air &
      Waste Management Association, 1992.

14.    T. McCurdy.  "Testing pNEM/O3 Runs to Determine if a Set of 10 Runs is
      'Representative'," EPA Memorandum; April 19, 1994.

15.    W.C. Adams.  Measurement of Breathing Rate and Volume in Routinely
      Performed Daily Activities.  Davis CA:  University of California, 1993.

16.    M. Absil, P. Narducci, R. Whitfield, and H.M. Richmond.  "Chronic Lung Injury
      Risk Estimates for Urban Areas Having Ozone Patterns Similar to Those in the
      Northeast."  Paper presented at the AWMA Specialty Conference on Tropospheric
      Ozone and the Environment II; Atlanta, November 4-7,  1991.
                                        10

-------
                  Table 1. ESTIMATES OF Ih "MAXIMUM DOSAGE" EXPOSURES IN ST. LOUIS TO O3 CONCENTRATIONS EXCEEDING
                0.12 PPM AND EVR" EQUALED OR EXCEEDED 30 ONE OR MORE TIMES PER OZONE SEASON UNDER
                              SIX AIR QUALITY SCENARIOS
Statistic
TOTAL EXPOSED POPULATION
Mean Estimate of the Number of People
Percent of Total Population
Range in this percentage for all runs
Mean Est. of Person-Occurrences
Percent of Total Per-Occurrences
Range in this percentage for all runs
Mean Est. of Occurrences/Person Exposed
Number of Days/Season Exposed (% Exposed)
I Day
2 Days
3 Days
>3 Days
CHILDREN
Mean Estimate of the Number of Children
Percent of Children Population
Range in this percentage for all runs
Percent of Total Exposed Population
Mean Estimate of Children-Occurrences
Percent of total Child-Occurrences
Range in this percentage for all runs
Mean Estimate of Occurrences/Child Exposed
Number of Days Exposed/Season (% Exposed)
1 Day
2 Days
3 Days
>3 Days
As Is
Situation
(1991)

5,100
0.3
0.0-1.3
5,100
@
0.0-*
1.00

100
0
0
0

2,900
0.9
0.0-4.6
56.9
2,900
@
0.0-@
1.00

100
0
0
0
Attain
8h lEx
0.10 ppm

200
*
0.0-0.1
200
@
0.0-®
1.00

100
0
0
0

0
0
-
0
0
0
-
-

-
-
-

Attain
8h 5 Ex
0.08 ppm

0
0
-
0
0
-
-

-
-
-
-

0
0
-
0
0
0
-
-

-
-
-

Attain
Ih 1 Ex
0.12 ppm

0
0
-
0
n
\j
~
-

-
-
-
-

0
0
•
0
0
0
-
-

-
-
-

Attain
8h 1 Ex
0.08 ppm

0
0
~
0
0
-
-

-
-
-
~

0
0
*
-
0
0
-
-

-
-
-

Attain
8h 5 Ex
0.06 ppm

0
0
"
0
0
*
-

-
-
-
"

0
0
-
-
0
0
-
-

-
-
-

NolcsrEquivaTcnt ventilation rate. It is equal to ventilation rate/body surlacc area; units are in I  mm"
    "Less than 0.05% but > 0.005%
    ©Less than 0.005% but >0.

-------
                  Table 2  ESTIMATES OF 8h "MAXIMUM DOSAGE" EXPOSURES IN ST. LOUIS TO O3 CONCENTRATIONS EXCEEDING
                                0 08 PPM AND EVRa EQUALED OR EXCEEDED 15 ONE OR MORE TIMES PER OZONE SEASON UNDER
                                                                      SIX AIR QUALITY SCENARIOS
Statistic
TOTAL NON-OUTDOOR WORKER EXPOSED POPULATION
Mean Estimate of the Number of People
Percent of Total Population
Range in this percentage for all runs
Mean Est. of person-Occurrences
Percent of total Per. -Occurrences
Range in this percentage for all runs
Mean Est. of Occurrences/Person Exposed
Number of Days/Season Exposed (% Exposed)
1 Day
2 Days
3 Days
4 Days
>4 Days
CHILDREN
Mean Estimate of the Number of Children
Percent of Children Population
Range in this percentage for all runs
Percent of Total Exposed Population
Mean Estimate of Children-Occurrences
Percent of Total Child-Occurrences
Range in this percentage for all runs
Mean Estimate of Occurrences/Child Exposed
Number of Days/Season (% Exposed)
1 Day
2 Days
3 Days
4 Days
>4 Days
As Is
Situation
(1991)

13,200
0.8
0.2-1.7
13,200
@
@-*
1.00

100
0
0
0
0

11,700
3.7
0.0-9.2
88.6
11,700
*
0.0-*
1.00

100
0
0
0
0
Attain
8h lEx
O.lOppm

6,700
0.4
0.0-1.1
6,700
@
@-@
1.00

100
0
0
0
0

5,100
1.6
0.0-5.7
76.1
5,100
@
0.0-*
1.00

100
0
0
0
0
	 =
Attain
8h5Ex
0.08 ppm

2,400
0.1
0.0-1.1
2,400
@
0.0-®
1.00

100
0
0
0
0

2,400
0.8
0.0-5.9
100.0
2,400
@
0.0-*
1.00

100
0
0
0
0
Attain
Ih 1 Ex
0.12 ppm

5,300
0.3
0.1-0.6
5.300
@
@-@
1.00

100
0
0
0
0

5,100
1.6
0.6-3.1
96.2
5,100
*
@>
1.00

100
0
0
0
0
Attain
8h 1 Ex
0.08 ppm

0
0
-
0
0
-


-
-
-
-
"

0
0
•
-
0
-
-
-

-
-
-
-
-
Attain
8h5 Ex
0.06 ppm

0
0
"
0

•


~
•
*
-


0
0
~
-
0
0
-
-

-
-
-
-
-
Notes:  aEquivalent ventilation rate.  It is equal to vend
        *Less than 0.05% but > 0.005%
        ©Less than 0.005% but >0.
lation rate/body surface area; units are in 1 mm"1 m

-------
 ESTIMATING OZONE EXPOSURES IN THE WASHINGTON, DC URBAN AREA
    USING A SECOND-GENERATION PROBABILISTIC VERSION OF NEM
                           Thomas McCurdy
                   U.S. Environmental Protection Agency
                    Research Triangle Park, NC 27711
                             March 1994
File: WASH.DC

-------
INTRODUCTION

The U.S. Environmental Protection Agency (EPA) currently is reviewing the ozone
(O3) primary national ambient air quality standard (NAAQS) to determine if the existing
standard protects public health with an adequate margin of safety. In order to evaluate
whether alternative O3 NAAQSs provide an adequate margin of safety, the EPA Office of
Air Quality Planning and Standards (OAQPS) assesses such factors as the nature and severity
of health effects associated with O3 exposure, the degree of total human exposure (i.e.,
indoor and outdoor) to O3, and the risks (probabilities) of ozone-related health effects
occurring in the exposed population when alternative O3 NAAQS are just attained. This
paper discusses an EPA analysis of O3 exposure estimates for parts of the Washington, DC
metropolitan area which employed the pNEM/O3 model developed by International
Technology-Air Quality Services (IT-AQS). A report by Johnson, et al. (1) describes this
model in detail and presents initial results of applying it to nine U.S. urban areas, including
Washington, DC. Exposure estimates for parts of the Los Angeles area are discussed in an
EPA memorandum (2) that was reviewed by the Clean Air Science Advisory Committee
(CASAC) at a public meeting. Exposure estimates for the Philadelphia area are described in
an AWMA paper (3). Exposure estimates for Houston are described in an AWMA specialty
conference paper (4). New York, Chicago, Denver, and Miami results are contained in EPA
papers (5-8). Additional analyses of O3 exposure estimates for St. Louis, the last urban area
modeled, will be forthcoming.

OVERVIEW

Pursuant to section 108 of the Clean Air Act, the Environmental Criteria Assessment
Office (ECAO) within the Office of Research and Development (ORD) is preparing a criteria
document (9) for O3. This document will be a comprehensive summary and critical
assessment of the latest scientific research relevant to assessing the effects of a criteria
pollutant. The document will be reviewed by CASAC, part of the EPA's independent
Science Advisory Board, and by the general public. Upon completion of the criteria
document, OAQPS will prepare a Staff Paper that summarizes and integrates scientific
information presented in the criteria document considered to be most relevant to decisions on
alternative NAAQS and various analyses (e.g., air quality, exposure, and risk).

An important aspect of health risk assessment is an analysis of population exposure
(10). Such an analysis provides the population basis for "headcount risk" estimates
developed as part of EPA's NAAQS risk assessment work. Estimates of population exposure
also provide important information regarding the number of people that may experience
varying levels of O3 as they go through their daily activities.

People living within each exposure district, as estimated by the U.S. Bureau of
Census in 1990, are assigned to a single discrete point, the population centroid. O3 air
quality within each exposure district is estimated for each hour of the O3 season using data
from a nearby monitoring site. Because O3 in the ambient air generally is modified
considerably when entering a building or vehicle, outdoor O3 estimates are adjusted using a
mass-balance model to account for 3 indoor microenvironments. These microenvironments
are: indoors-at-home, indoors-other, and within a motor vehicle. The mass-balance model
used in pNEM/O3 is a simplified version of the generalized Nagda, Rector, and Koontz
(NRK) model (11). For our use, this model was revised to incorporate the assumption that
indoor decay rate is proportional to indoor O3 concentration. It was further revised to
incorporate assumptions concerning ozone decay rates suggested by Weschler, et al (12).
The resulting revised mass-balance model is fully described in Ref. 1.

Because both the intake dose received and susceptibility to effects of pollution may
vary with age, occupation, and intensity of exercise, the total population of each study area is
divided into nine age-occupation (A-O) groups. Each A-O group is further subdivided into
cohorts depending upon (1) the type of air conditioning system present in the home, if any,
and (2) home-to-work commuting patterns.

While many types of exposure metrics are provided by pNEM/O3, only those for the
non-outdoor working population in Washington, DC are discussed in this paper. Exposure
estimates for children are included in the results discussed here. Estimates for outdoor
workers~a major population sub-group of concern from an O3 exposure perspective-will be
completed in the Fall of 1994.

AIR QUALITY SCENARIOS INVESTIGATED

Six air quality scenarios were modeled for the Washington, DC urban area (and the
other eight urban areas); the scenarios are all on a daily maximum basis. They are:

1. 1991 air quality--the "as is" scenario
2. Attain Ih lEx 0.12 ppm—the current standard
3. Attain 8h lEx 0.08 ppm
4. Attain 8h 1 Ex 0.10 ppm
5. Attain 8h 5Ex 0.06 ppm
6. Attain 8h 5Ex 0.08 ppm

The iterative adjustment procedure used by IT-AQS to simulate just-attaining the 5
alternative NAAQS listed above involves a complex procedure of fitting Weibull equations to
the 1991 O3 data, modifying parameters of the equations so that the "worst" monitor just
achieves the desired standard, and then re-specifying the entire O3 pattern for all districts in
Washington, DC. The procedure is described in detail by Johnson, et al. (1).

Washington, DC had moderate O3 air quality in 1991 in that its "design value"--the
second-highest Ih daily value measured at any site in the urban area—was 0.174 ppm. This
value is considerably above the current O3 standard concentration of 0.12 ppm. The area
also had a number of relatively high 8h daily maximum concentrations in 1991.

WASHINGTON, DC RESULTS

Results are based upon 10 runs (realizations) of the pNEM/O3 model for each air
quality scenario. Based on a previous analysis of sets of 10-run results versus a 108-run
result (13), the author believes that results from only 10 runs of the model will adequately
predict the mean and variance observed in 100 or more runs of pNEM/O3. Ten runs of the
model will not, however, adequately represent the entire range of possible outcomes.
Additional runs can only increase this range. This is an obvious shortcoming of the analyses
presented here, but limited resources preclude undertaking more runs.

-------
This paper includes tabular data on (1) the number and percent of people and children
exposed, and (2) the number and percent of person-occurrences for children and the adult,
non-outdoor worker population in Washington, DC. (A person-occurrence is the metric that
focuses on an event: when the exposure-of-concern is reached regardless of who reaches it.
One hundred person-occurrence could be 10 people experiencing the exposure 10 times, 100
people experiencing it once, one person experiencing it 100 times, or any combination
thereof.) Even though data are discussed here regarding the absolute and relative number
(percentage) of people exposed, it must be noted that "people" should not be used as an
exposure metric. The cohort-specific activity-days data base used for pNEM/O3 modeling
(and for all other exposure models that use human activity information) provides no more
than three days of data for a single individual. Thus, rigorous inferences from these data can
be made only to "person-days" of exposure or "person-occurrences" of exposure, where an
occurrence can be Ih, 8h, or 24h time period in pNEM/O3. Only when activity data are
available sequentially for an entire year for individuals in a cohort can inferences be made to
people exposed over the year (or to children, or to outdoor workers as the case may be).
This point usually is ignored in the general exposure assessment literature.
All exposure estimates presented in this paper pertain to "daily maximum dose,"
(MAXD) where dose is defined as the product of ozone concentration and ventilation rate
over a defined time period. Note that maximum daily dose does not necessarily occur during
the time period of maximum ozone concentration in a given urban area.

Ventilation rate is a very important factor in estimating total dose for any air pollutant
(14). The ventilation rate metric used in pNEM/O3 is "equivalent ventilation rate" (EVR),
or ventilation rate (in liters per minute) divided by body surface area (BSA, in units of m2).
EVR has units of 1 min'1 m'2. It is a surrogate indicator for the number of O3 molecules that
enter the oral-nasal cavities per unit time period.

The Washington, DC urban area population analyzed here includes 3.1 million NOW
(non-outdoor worker) people and 488,000 children. There are 660 million possible daily
maximum NOW person-occurrences (NOW people x 214 days in the O3 season). The
equivalent figure for children is 104 million.

Exposure results appear in Tables 1 and 2. Table 1 includes In daily maximum dose
exposure estimates, while Table 2 provides similar results for 8h exposure estimates.

The Ih results indicate that:

1. Currently (1991), it is estimated that 1.5% of NOW people and children would
experience a daily maximum dose (MAXD) O3 exposure @ >0.12 ppm @ an EVR
of 30 1 min"1 m"2 or higher, which is thought by many researchers to be a dose rate
of concern (14). About 16% of the NOW people so exposed are children, and most
MAXD exposures occur only on one day/O3 season.

-------
2. Attainment of the alternative NAAQS investigated are expected to substantially
decrease NOW people exposures to the Ih MAXD of concern, and eliminate children
exposures to that dose rate. None of the Ih MAXD exposures of concern occur on
more than one day/O3 season for the alternative NAAQS.

The 8h results (Table 2), indicate that many more people may be exposed. The level
of concern here is a MAXD exposure of >0.08 ppm for 8h @ an EVR J>_15 1 min"1 m"2.
(Obviously, the EVR of concern drops as the averaging period increases, just as the O3
concentration of concern drops. People cannot hold an EVR of 30 1 min"1 m"2 for an
extended period of time (15).)

The 8h results indicate that:

1. Currently (1991), about 3% of the NOW population and 15% of the children are
estimated to experience the 8h MAXD of concern. These exposures mostly occur on
only one day/season, but some adults and children could see up to 3 days of MAXD
exposures >0.08 ppm @ 15+ EVR.

2. Attaining the alternative NAAQS that were investigated do not guarantee no 8h
MAXD exposures of concern-except for the very stringent 8h 5Ex 0.06 ppm
NAAQS. Attaining the 8h lEx 0.10 ppm 03 standard in Washington, DC could
result in over 2% of children exposed to the concern dose. Many of these exposures
could occur on more days-about 15%. Attaining the current standard (Ih lEx 0.12
ppm) could result in over 1% of children being so exposed at exercise. Both of these
estimates are relatively high when compared to results from the other urban areas than
were modeled (2-8).

Putting the 8h MAXD estimates into perspective, the results indicate that currently
(1991) 2.4 million NOW people might experience an 8h exposure >0.08 ppm @ any
exercise level. (In other words, when exercise is ignored; this estimate is not shown in
Table 2.) This compares to 91 thousand people estimated to experience the same O3 air
quality concentration @ 15 EVR or higher. The corresponding estimates for children are
443,000 exposures at any exercise level versus 73,900 at 15+ EVR. Thus, exercise level
certainly makes a large difference when 8h exposures to specific O3 concentration levels are
being estimated.

-------
For the record, the 8h any-exercise estimates for the non-zero alternative NAAQS
are:
NOW People

Children
8h lEx
0 .10
623,200
20.2
152,000
31.2
8h 5Ex
0.08
197,600
6.4
57,100
11.7
Ih lEx
0. 12
459,900
14 .9
109,600
22.5
8h lEx
0.08
44,700
1.4
9,000
1.8
Contrast these estimates with those appearing in Table 2 for insight into the impact of
exercise level on exposure estimates. The any-exercise estimates are 1-1.5 orders of
magnitude higher than the .>_15 EVR estimates.

Statistical tests of the exposure distributions were undertaken using a two-sample
Kolmogorov-Smimov (K-S) non-parametric test. The null hypothesis was that the cumulative
frequency exposure distributions for the various air quality scenarios are the same for each
exposure metric being tested. The test statistic was the maximum difference between the two
distributions (Dn); it was evaluated using a large-sample F statistic. The significance level is
0.05, using the chi-square approximation. The test indicates that the exposure indicators
shown in Tables 1 and 2 vary significantly across the six air quality scenarios with the
following exceptions.

For the Asls and 8h lEx 0.10 ppm pair:

Number of NOW people and children expected to be exposed to 8h 0.08 ppm levels
at 15+ EVR exercise levels.

For the current standard and 8h lEx 0.10 ppm pair:

All exposure indicators that were investigated except for NOW people and NOW-
occurrences at 8h 0.08 ppm at any exercise level.

For the current standard and 8h 5Ex 0.08 ppm pair:

NOW people, NOW-occurrences, and the number of people exposed on 1 day/O3
season at Ih >0.12 ppm at .X30 EVR.

Children and children-occurrences exposed to Ih >0.12 ppm at any exercise level.

-------
For the 8h lEx 0.08 ppm and 8h 5Ex 0.08 ppm pair:

NOW people, NOW-occurrences, and the number of people exposed on 1 day/O3
season at Ih >0.12 ppm at >3Q EVR.

The results of the K-S testing procedure undertaken in Washington, DC are similar to
those for Miami (8). Many exposure indicator pairs could not be distinguished with respect
to attaining some of the alternative NAAQS in both areas. These standards do not make
much impact for the indicators listed above. However, there is no real consistent pattern
among the exposure metrics and air quality standard pairs that were tested.

Note that these are not strong statistical tests and that some of the K-S test
requirements were not met. Random sampling was not used. The sample "observations"
simply are results of a simulation model; the population of interest, then, can only be the set
of possible outcomes produced by the model. A large-sample test statistic was applied to
small samples (10 runs). These caveats aside, 24 "cannot rejects" out of 145 tests is three
times as many as are expected based on a "pure-chance" probability of rejecting a true
hypothesis (5%). This 17% cannot-reject rate is about double the rates seen in most of the
urban areas that were analyzed (2-8). Unlike most of these
areas, there is not a clear-cut picture produced by the K-S testing exercise in Washington,
DC. There is not a consistent pattern—especially at exercise—associated with the alternative
NAAQS investigated.

SUMMARY

The analyses of 5 alternative standards ozone NAAQS in the Washington, DC area
(Ih lEx 0.12 ppm; 8h lEx 0.10 pm; 8h 5Ex 0.08 ppm; 8h lEx 0.08 ppm; and 8h 5Ex 0.06
ppm) indicate that attainment of any of the alternatives will sharply reduce-but not
completely eliminate—exposures to Ih and 8h MADs of concern. In particular, attaining the
current standard (Ih lEx 0.12 ppm) and the 8 1 Ex 0.10 ppm alternative NAAQS does not
protect a small, but important, percentage of children from experiencing the 8h MAXD of
0.08 ppm at J>_15 EVR. The current standard is more protective than the 8h lEx 0.10 ppm
alternative.

-------
                                 REFERENCES

1.     T. Johnson, J. Capel, and M. McCoy.  Estimation of Ozone Exposures
      Experienced by Urban Residents Using a Pro-
      babilistic Version of NEM and 1990 Population Data.  Durham, NC: IT-AQS,
      1993 (Draft).

2.     T. McCurdy.  "Los Angeles Ozone Exposure Estimates," EPA Memorandum;
      November 4, 1993.

3.     T. McCurdy, et al.  "Estimating Ozone Exposures in Philadelphia Using a Second-
      Generation Probabilistic Version of NEM."  Paper A210 presented at the 87th
      Annual Meeting of AWMA; Cincinnati, June 1994.

4.     T. McCurdy, et al.  "Estimating ozone Exposures in Houston Using a Second-
      Generation Probabilistic Version of NEM."  Paper presented at the Tropospheric
      Ozone Specialty Conference, Air and Waste Management Association; Orlando,
      1994.

5.     T. McCurdy.  "Estimating Ozone Exposures in the New York Urban Area Using a
      Second-Generation Probabilistic Version of NEM."  EPA Paper, February  1994.

6.     T. McCurdy.  "Estimating Ozone Exposures in the Chicago Urban Area Using a
      Second-Generation Probabilistic Version of NEM."  EPA Paper, March 1994.

7.     T. McCurdy.  "Estimating Ozone Exposures in the Denver Urban Area Using a
      Second-Generation Probabilistic Version of NEM."  EPA Paper, March 1994.

8.     T. McCurdy.  "Estimating Ozone Exposures in the Miami Urban Area Using a
      Second-Generation Probabilistic Version of NEM."  EPA Paper, March 1994.

9.     Environmental Criteria and Assessment Office. Air Quality Criteria for Ozone and
      Photochemical Oxidants.  Research Triangle Park, NC:  1994  (Draft).

10.   H.M. Richmond and T. McCurdy.  "Use of Exposure Analysis  and Risk Assessment
      in the Ozone NAAQS Review."  Paper 88-121.3 presented at the 81st Annual
      Meeting  of APCA; Dallas, June 19-24, 1988.

11.   N.L. Nagda, H.E. Rector, and M.D. Koontz. Guidelines for Monitoring Air
      Quality.   Washington, DC: Hemisphere Pub. Co., 1987.

-------
12.    C.J. Weschler, H.C. Shields, and D.V. Nike.  "Indoor Ozone: Recent Findings," in:
      R. Burglund (ed.), Tropospheric Ozone in the Environment EL Pittsburgh: Air &
      Waste Management Association, 1992.

13.    T. McCurdy. "Testing pNEM/O3 Runs to Determine if a Set of 10 Runs is
      'Representative'," EPA Memorandum; April 19, 1994.

14.    W.C.  Adams.  Measurement of Breathing Rate and Volume hi Routinely
      Performed Daily Activities.  Davis CA:  University of California, 1993.

15.    M. Absil,  P. Narducci, R. Whitfield,  and H.M. Richmond.   "Chronic Lung Injury
      Risk Estimates for Urban Areas Having Ozone Patterns Similar to Those in the
      Northeast."  Paper presented at the AWMA Specialty Conference on Tropospheric
      Ozone and the Environment II;  Atlanta, November 4-7, 1991.
                                       10

-------
Table 1.  ESTIMATES OF Ih "MAXIMUM DOSAGE" EXPOSURES IN WASHINGTON, DC TO O3 CONCENTRATIONS EXCEEDING
0.12 PPM AND EVR" EQUALED OR EXCEEDED 30 ONE OR MORE TIMES PER OZONE SEASON UNDER
               SIX AIR QUALITY SCENARIOS
Statistic


TOTAL EXPOSED POPULATION
Mean Estimate of the Number of People
Percent of Total Population
Range in this percentage for all runs
Mean Est. of Person-Occurrences
Percent of Total Per-Occurrences
Range in this percentage for all runs
Mean Est. of Occurrences/Person Exposed
Number of Days/Season Exposed (% Exposed)
1 Day
2 Days
3 Days
>3 Days
CHILDREN
Mean Estimate of the Number of Children
Percent of Children Population
Range in this percentage for all runs
Percent of Total Exposed Population
Mean Estimate of Children-Occurrences
Percent of total Child-Occurrences
Range in this percentage for all runs
Mean Estimate of Occurrences/Child Exposed
Number of Days Exposed/Season (% Exposed)
1 Day
2 Days
3 Days
>3 Days
As Is
Situation
(1991)

45,800
1.5
0.7-2.0
46.000
*
@-*
1.0

99
1
0
0

7,300
1.5
0.0-4.8
15.9
7,300
*
0.0-*
1.00

100
0
0
0
Attain
8h lEx
0.10 ppm

1,900
O.I
0.0-0.1
1,900
@
0.0-@
1.00

100
0
0
0

0
0
-
0
0
0
-
-

-
-
-
"
Attain
8h5Ex
0.08 ppm

400
*
0.0-0.1
400
@
0.0-@
1.00

100
0
0
0

0
0
-
0
0
0
-
-

-
-
-
"
Attain
Ih 1 Ex
0.12 ppm

800
*
0.0-0.1
800
@
0.0-@
1.00

100
0
0
0

0
0
-
0
0
0
-
-

-
-
-

Attain
8h 1 Ex
0.08 ppm

0
0
-
0
0
-
-

-
-
-
-

0
0
-
-
0
0
-
-

-
-
-

Attain
8h 5 Ex
0.06 ppm

0
0
-
0
0
-
-'

-
-
-
-

0
0
-
-
0
0
-
-

-
-
-

otes:aEquivalent ventilation rate. It is equal to ventilation rate/body surface area; units are in 1 min"1 m"2
*Less than 0.05% but > 0.005%
@Lcss than 0.005% but >0.

-------
              Table 2.  ESTIMATES OF 8h "MAXIMUM DOSAGE" EXPOSURES IN WASHINGTON. DC TO O3 CONCENTRATIONS EXCEEDING
               0.08 PPM AND EVRa EQULAED OR EXCEEDED 15 ONE OR MORE TIMES PER OZONE SEASON UNDER
                              SIX AIR QUALITY SCENARIOS
Statistic
TOTAL NON-OUTDOOR WORKER EXPOSED POPULATION
Mean Estimate of the Number of People
Percent of Total Population
Range in this percentage for all runs
Mean Est. of person-Occurrences
Percent of total Per. -Occurrences
Range in this percentage for all runs
Mean Est. of Occurrences/Person Exposed
Number of Days/Season Exposed (% Exposed)
1 Day
2 Days
3 Days
4 Days
>4 Days
CHILDREN
Mean Estimate of the Number of Children
Percent of Children Population
Range in this percentage for all runs
Percent of Total Exposed Population
Mean Estimate of Children-Occurrences
Percent of Total Child-Occurrences
Range in this percentage for all runs
Mean Estimate of Occurrences/Child Exposed
Number of Days/Season (% Exposed)
1 Day
2 Days
3 Days
4 Days
>4 Days
As Is
Situation
(1991)

90,600
2.9
1.6-5.0
96,500
*
@-*
1.06

94
5
1
0
0

73,900
15.2
7.2-27.0
81.6
79,800
0.1
*-0.1
1.08

93
6
1
0
0
Attain
8h lEx
O.lOppm

12,300
0.4
0.1-0.0
13,900
@
@-@
1.13

87
13
0
0
0

11,000
2.3
0.3-5.3 '
89.4
12,600
+
@-*
1.15

85
15
0
0
0
Attain
8h5Ex
0.08 ppm

1,300
*
0.0-0.2
1,300
@
0.0-@
1.00

100
•
0
0
0
0

1,300
0.3
0.0-1.5
100.0
1,300
@
0.0-®
1.00

100
0
0
0
0
Attain
Ih 1 Ex
0.1 2 ppm

9,100
0.3
0.0-1.0
9,100
@
0.0-®
1.00

100
0
0
0
0

5,800
1.2
0.0-5.5
63.7
5,800
*
0.0-*
1.00

100
0
0
0
0
Attain
8h 1 Ex
0.08 ppm

400
*
0.0-0.1
400
@
0.0-@
1.00

100
0
0
0
0

400
0.1
0.0-0.8
100.0
400
@
0.0-@
1.00

100
0
0
0
0
Attain
8h5Ex
0.06 ppm

0
0
-
0
0
-
-

-
-
-
-
-

0
0
-
-
0
0
-
-

-
-
-
-
-
Notes:  "Equivalent ventilation rate.
  *Less than 0.05% but  > 0.005%
  ©Less than 0.005% but >0.
It is equal to venti    lation rate/body surface area; units are in 1  mm"1 m''

-------