AIRBORNE PARTICULATE MATTER WITHIN SCHOOL
ENVIRONMENTS IN THE UNITED STATES
B. Ligman1, M. Casey2, E. Braganza2, A. Coy2, Y. Redding2, S. Womble1
'Indoor Environments Division, U.S. Environmental Protection Agency, USA
2 Radiation and Indoor Environments National Laboratory, U.S. Environmental Protection
Agency, USA.
ABSTRACT
The U.S. Environmental Protection Agency (EPA) has been conducting indoor environmental
studies to characterize the impact different interventions have on the indoor environment of
school buildings. Several indoor environmental measurements have been conducted including
measuring time-weighted gravimetric airborne particulate matter (PM). Presented here are the
results of PM data that are lOum aerodynamic diameter or less (PM10) and those with 2.5um
aerodynamic diameter or less (PM2 5). These data were collected in elementary and secondary
school (kindergarten - grade 12) indoor environments across the United States. Comparisons are
made between these data and the PM data collected in the Building Assessment and Survey
Evaluation (BASE) study, also conducted by the EPA. The BASE data represent typical
concentration levels found in U.S. office buildings. In general, PM results were higher in schools
than in office buildings. In addition, the school data show higher concentrations indoors than
outdoors.
INTRODUCTION
Ambient airborne particulate matter has received recent attention in the U.S. due to the EPA's
promulgation of a new National Ambient Air Quality Standards (NAAQS) for airborne particles
smaller than 2.5 um in aerodynamic diameter (PM25).
Despite the data that indicate Americans spend approximately 90% of their time indoors [1]
limited research, relative to ambient air research, has been conducted to characterize exposure to
indoor environmental pollutants such as PM. In response to a U.S. Congressional request, the
Committee on Research Priorities for Airborne Particulate Matter was formed under the auspices
of the National Academy of Sciences (NAS). The committee has been tasked with identifying the
most important research priorities relevant to setting ambient air PM standards and to develop and
monitor progress on a conceptual plan for the research. The committee's first report [2] (four
more are expected over the next five years) presents 10 priority research topics. The data
presented in this paper is relevant to one of these priority research areas, investigating the
breathing-zone exposures of individuals to PM, taking into account indoor pollutant sources. In
addition, the committee also recommends immediate research attention for potentially susceptible
sub-populations such as children.
To date, limited pollutant data have been collected for the indoor environments of U.S. buildings.
EPA's Building Assessment Survey and Evaluation (BASE) study is the only sizable study that
attempts to characterize indoor air quality for a specific building type and usage [3]. Fewer data
-------
have been collected on pollutant concentrations in school environments. Presented here are PM
concentration data that have been collected as part of EPA's indoor air quality studies in
elementary and secondary schools (kindergarten - grade 12). EPA has been conducting limited
research in school buildings since 1989. PM data collected during the earlier period ('89 - '95)
were obtained from the School Evaluation Program [4]. Also presented, for comparison, are U. S.
office building PM data collected as part of the BASE study.
METHODS
Two data sets are presented representing 1) BASE PM results, and 2) School Intervention (SI)
PM results. All data was obtained using similar sampling equipment and following similar
protocols. All data were collected during a normally occupied classroom school day or occupied
office area (nominal sampling period of approximately. 8-10 hours).
The BASE data represent a sample of randomly chosen office buildings across the U.S. All BASE
data were collected following a standardized protocol [5]. The study is designed to collect large
data sets of indoor pollutant concentrations (and other information such as building characteristics
and occupant perceptions) representing baseline indoor air quality in large office buildings. PM10
and PM2 5 data are two of the pollutants collected at each of the buildings. The BASE PM data
set presented here was collected from three randomly selected areas of each randomly selected
office building as well as one outdoor measurement for each building as specified in the protocol.
The SI data are comprised of two sets of data. The original data set was collected from 1989 to
1995 and consist of PMI0 data from 10 school buildings. Data were collected from one classroom
per building. The classroom selection was typically based on the highest radon concentration for
the building. These data have been reported previously [4] and are not discussed here in detail.
All of the SI buildings studied prohibited smoking within the building. The remaining SI data
were collected as part of EPA's intervention studies following a modified BASE protocol [6].
The data were collected from two ongoing intervention studies that evaluate the impact of energy
retrofits [7] and of implementing EPA's indoor air guidance for schools [8] on the quality of the
indoor air. These data were collected from four classrooms per building. The selection was purely
subjective with an attempt to select two special use classrooms (i.e., art, sciences, etc.) and two
"normal" use classrooms for each building studied. Selecting special use classrooms proved
difficult since most elementary schools lack these types of classrooms. Therefore, the data more
closely represent normal use classrooms.
The entire SI data set does not represent random building selection or random sampling locations
but rather schools that either a) had elevated radon gas concentrations (the original data collected
from 10 school buildings), or b) willingness to participate in an indoor air quality intervention
study. The selections were made with no known predisposition to high or low PM. All the school
data represent buildings before any intervention activities were conducted.
RESULTS AND DISCUSSION
Table 1 contains specific details for the data being presented. Table 1 also presents the coefficient
of skewness resulting from performing the Davies' Test for Logarithmic Distribution [9]. The
Davies' test states that if the coefficient is less than 0.20, the data are approximately logarithmic
in distribution. With the exception of SI PM2.5 indoor and the SI PM2 5 outdoor data, the
-------
distributions are logarithmic. The coefficients greater than 0.20 may be due to the smaller sample
size of the data sets which affect the parameters for the Davies' test. Since the distributions are
logarithmic, the geometric means are presented in Table 1 and used for the comparisons below.
Table 1. Summary Information on Data
Data
Set
No.
Bldgs.
Location (States)
No. Samples
Davies'
skewness
coef.
Geometric
Mean
(ug/m3)
BASE
71 (PM10
AZ, CA, CO, FL,
PM10 indoor
208
0.02
12
& PM2.5)
GA, LA, MA,
PM2 5 indoor
73
-0.30
8
MD, MI, MN,
PM10 outdoor
71
-0.12
26
MO, NE, NV,
PM2 5 outdoor
71
0.04
16
NY, OR, PA, SC,
TN, TX, WA
SI
20 (PM10)
10 (PM2.5)
CA, CO, FL, KS,
MN, NJ, NM,
NY, TX, WA
PM10 indoor
PM2 5 indoor
PM10 outdoor
PM2 5 outdoor
50
40
20
10
0.06
0.22
-0.10
0.48
46
13
19
9
Table 2 presents the results of performing the z Test for Measurements (zM Test) for determining
whether the differences in geometric means are statistically significant. The table contains the z
values for the data sets as well as the indoor/outdoor data for each set. According to the
important probability levels (calculated from the Normal Probability Formula) [10], a z value
greater than 2.58 indicates a probability of less than 1% in stating that the means are not
statistically different, (i.e., mean differences are statistically significant). A z value less than 1.64
indicates a probability greater than 10% in stating that the means are not statistically different,
(i.e. mean differences are not statistically significant). As Table 2 shows, all geometric mean
comparisons are significantly different with the exception of the SI PM2 5 indoor and outdoor
comparison.
Figures 1,2, and 3 present the results of the BASE PM, SI PM and both data sets respectively.
The Figures contain box plots of the data plotted on a logarithmic scale with the whiskers
representing the 5th and 95th percentile and the ends of the box representing the 25th and 75th
percentiles. The solid line in the box is the median (50th percentile). The black dots represent the
data that fell outside of the 5th and 95th percentiles.
While differences in the collection of SI PM and BASE PM (e.g., random vs. non-random
building selection, school buildings vs. office buildings) preclude any rigorous comparisons of
the data, BASE PM data are used to lend perspective to the SI PM data. Comparing the BASE
indoor PM mean to the outdoor PM mean shows that the indoor concentration data are
approximately 53% lower than the outdoor means for both PM10 & PM2 5. However, the same
comparison for the SI data reveals just the opposite for school environments. The SI data's
outdoor PM10 mean is 59% lower than indoor PM10 and the outdoor PM2 5 is 45% lower than
indoor PM2 5. However, as noted above, the z value for the SI PM2 5 comparison indicates the
difference is not statistically significant.
-------
Table 2. zM-Test Results (z value)
Data Compared
z value
BASE outdoor PM10 to BASE indoor PM10
13.24
BASE outdoor PM2 5 to BASE indoor PM25
10.30
SI outdoor PM10 to SI indoor PM10
6.46
SI outdoor PM2.5 to SI indoor PM25
1.30
SI indoor PM10 to BASE indoor PM10
16.79
SI indoor PM2 5 to BASE indoor PM2 5
6.60
Figure 3 presents the indoor data for both data sets. Comparison of the geometric means show
higher concentrations in these school environments than in office buildings with the mean PM10
concentration in offices to be 73% lower than in schools and 43% lower in offices for PM2 5
These results are suggestive that exposures to PM may be higher in schools than in office
buildings and may be of concern for school-aged children. Possible explanations for these results
include poor filtration of the air, poor housekeeping (such as floor cleaning), and unknown indoor
sources related to the uniqueness of school environments (such as chalk dust) and possibly the
deteriorating building structures [11].
Figure 1. Gravimetric PM Concentrations (BASE data)
Indoor PM10 Outdoor PM10 Indoor PM2.5 Outdoor PM2.5
I 1 PMto
vzvk pm2.5
The results presented here support the need for additional personal exposure research. There is
also a need for research on PM composition. It is not possible to determine the health effects
associated with PM exposure unless PM composition is characterized and quantified.
Understanding PM personal exposures must include indoor as well ambient air characterization.
Indeed, the data presented here indicate that the characterization of the indoor school environment
may be a higher priority than ambient air for children. These data provide additional information
for consideration as PM research is prioritized.
-------
In conclusion, the authors support the NAS's recommendation for determining actual personal
exposures to airborne PM by characterizing the indoor environment, ambient air, and human
time-activity patterns. Ambient PM has historically received more attention due to the regulatory
concerns. However, the importance of the school indoor environment on children's exposure to
PM may be greater than that of the ambient air as demonstrated in the data presented here. In the
absence of specific data on PM composition, it is prudent to determine how to reduce PM
concentrations in schools and disseminate this information to those who can act accordingly. This
could include intervention studies that observe the effects different interventions have on PM
concentrations in schools and prioritize those that show the greatest reduction.
O
o
Figure 2. Gravimetric PM Concentrations (SI data)
100 ¦
100
o
Indoor PM10 Outdoor PM10 Indoor PM2.5 Outdoor PM2.5
l=l ™10
wza PM2S
Figure 3. Indoor Gravimetric PM Concentrations (all data sets)
l=l PM10
mm pm2.5
-------
ACKNOWLEDGMENTS
This study was supported by the U.S. Environmental Protection Agency but was not subjected
to the U.S. Environmental Protection Agency's peer review. The conclusions in this paper are
those of the authors and not necessarily those of the U. S. Environmental Protection Agency.
REFERENCES
1. National Research Council. 1981.; Spengler and Sexton. 1983.; Robinson and Nelson. 1995.
2. National Research Council, 1998. Research Priorities for Airborne Particulate Matter I.
3. Womble S., Girman J., Ronca E., et al. 1995. Developing Baseline Information on Buildings
and Indoor Air Quality (BASE '94): Part I -Study Design, Building Selection, and Building
Descriptions. Healthy Buildings '95.
4. Shaughnessy R., Brennan T., Ligman B., et al. Indoor and Outdoor Respirable Suspended
Particles in Schools Across the United States.
5. U. S. Environmental Protection Agency. 1994. A standardized EPA Protocolfor Characterizing
Indoor Air in Large Office Buildings. Office of Radiation and Indoor Air.
6. U.S. Environmental Protection Agency. 1994. Preliminary Draft Conceptual Standardized
EPA Protocolfor Characterizing Indoor Air in School Buildings. Office of Radiation and Indoor
Air.
7. Redding Y., Harrison J. 1999. Indoor Air Quality and Performance Contracting in Schools.
Indoor Air '99. to be presented at.
8.U.S. Environmental Protection Agency. 1995. Indoor Air Quality Toolsfor Schools. EPA-402-
K-95-001. Washington, DC.
9. Davies G. 1929. Journal of American Statistics Association pp. 349-66
10. Pearson E., Hartley H. 1966. Adapted from Biometrika Tables for Statisticians. Vol. 1, Table
4.
11. U. S. General Accounting Office. 1995. School Facilities: Condition of America's Schools.
GAO/HEHS-95-61.
------- |