United States
Environmental Protection
Agency
Environmental Monitoring
Systems Laboratory
Research Triangle Park NC 27711
Research and Development
EPA-600/S4-84-088 Mar. 1985
EPA Project Summary
Inhalable Particulate
Network Report: Operation and
Data Summary (Mass
Concentrations Only)
David 0. Hinton, Jose M. Sune, Jack C. Suggs, William F. Barnard
This report is intended to serve as an
operations overview and data summary
covering the operation of the 157
Inhalable Participate (IP) Network sites
within the United States. Volume I
discusses the scope of the Network and
instrumentation utilized in the
Network. Data (mass only) are traced
from measurement through processing
and storage to routine reporting.
Quality assurance practices are also
given. Data summaries are provided.
Volume II is a list of individual data upon
which Volume I is based.
Analyses, conclusions, and examples,
either listed or indicated by reference,
should provide the reader with both
suggested uses and possible limitations
of the data. Chemical analysis of the
collected particulate (sulfate, nitrate,
and selected metals) is part of IP
Network objectives but those data will
be the subject of a separate report.
This Project Summary was developed
by EPA's Environmental Monitoring
Systems Laboratory. Research Triangle
Park, NC, to announce key findings of
the research project that is fully
documented in a separate report of the
same title (see Project Report ordering
information at back).
Introduction
The 1977 Clean Air Act Amendment
requires a reappraisal of the National
Ambient Air Quality Standard for particu-
late matter. In order to meet this require-
ment, information regarding both Total
Suspended Particulate (TSP) and smaller
inhalable particles was required.
EPA's Environmental Monitoring Sys-
tems Laboratory (EMSL), Research
Triangle Park, NC, in conjunction with
EPA's Office of Air Quality Planning and
Standards was given the responsibility of
providing ambient air data for the small
particle size range. The exact value for the
upper limit, however, was and still
remains, controversial. Therefore, data
from both the original 0-15 fjm samplers
and the subsequent 0-10 /urn samplers
are included. In 1977-78 when the
Inhalable Particulate (IP) Network was
being planned, the major monitoring
emphasis was on collection devices
which could provide measurements of
ambient air concentrations attributed to
15 fjm (and smaller) particles and 2.5 (im
(and smaller) particles. Rpr these reasons
instruments using inlets providing a
single sample (15 fjm and smaller) and
dual samples (2.5 to 15 fjm and below 2.5
fjm) were evaluated and utilized. Both are
referred to as PM,6 to designate the upper
limit of 15 fjm.
Procedure
In 1978, the data from the Inhalable
Particulate Network were anticipated to
be used primarily to assist in a revision of
the existing Total Suspended Particulate
Standard. The revised standard was to be
based on the specific particle size range
of 15 fjm mean aerodynamic diameter
and below, and to a lesser degree to
provide information on the possible
sources of the particles for subsequent
control strategy implementation. To
accomplish this, establishment of a
nationwide network of 200 air monitoring
-------
sites over a three-year period was
planned. However, due to resource
constraints, only 157 sites were placed
on line.
The following specific Network
objectives and design criteria were
provided by the Office of Air Quality
Planning and Standards (OAQPS): (a)
conduct a pilot program to demonstrate
that the monitoring technology was
adequate to proceed with the study
(technology to make routine size-specific
aerosol measurements had only recently
become commercially available), (b)
provide monitoring support to on-going
epidemiology studies wherever possible,
(c) provide background data for non-
urban and rural sites, (d) monitor fugitive
dust locations, (e) select urban sites with
priorities primarily for population density
and non-attainment of the current TSP
standard, (f) at all sites, measure the
mass concentration of TSP and IP, (g) at
selected sites, measure the fine and
coarse components of IP (i.e., PM15), and
(h) provide for a limited component
analysis scheme beyond mass concen-
tration to further characterize the data
base. Later, a final objective was added: (i)
incorporate PM,0 technology into the
network for data collection intheO-10/^m
size range.
OAQPS specified the candidate cities.
With OAQPS approval, EMSL and/or an
EMSL contractor made the specific site
selection within the city, based on desired
site classification (Commercial,
Residential, Industrial, Rural, etc.) and
specific site availability.
All of the objectives and constraints
were combined into a protocol of network
operations, which was prepared prior to
network implementation. This protocol
included the various aspects of network
design and setup, sample collection,
analyses, quality assurance, mainte-
nance, and data processing and analyses.
All operations except the actual collection
of samples would be provided by EPA.
Manpower was to be provided by State
and local agency personnel to implement
the operation of the sampling equipment.
Because of the limited manpower
available within EPA, contractor support
was also planned. A Quality Assurance
program was planned and budgeted at 5-
10% of resources.
Although each sampling site location
was physically evaluated against the
siting criteria given in the Inhalable
Paniculate Network Operations and
Quality Assurance Manual, March, 1983,
administratively the selection process
was quite variable. Land owner permis-
sion, local agency approval, Regional
Office concurrence, OAQPS recommend-
ations/concurrence all had to occur in
order for a specific site to begin and
continue sample/data collection.
Further, since more than 1,000 people
were eventually involved directly in the
data gathering activities, their
performance, interest, and assistance
directly affected the amount and quality
of data collected. In spite of the diverse
demands on time, personnel and
resources, EMSL received excellent
cooperation from local. State, and
Regional personnel. This cooperation
resulted in data collection from 525
sampler-years from 1 57 sites.
All sites provide routine TSP data from
a Hi-Volume sampler and PM,5 data from
either a Size Selective Sampler (SSS) or
Dichotomous Sampler. In addition to
routine sampling requirements, EMSL
utilized selected sites for intercompari-
son of instruments. At various times a
given site became one or more of the
following:
1. Comparison Site: In addition to the
instrument complement of a PM15
and a TSP Hi-Vol for routine
sampling, some of the initial sites
were provided with additional PM15
instruments. These special sites
provided data for comparison of
SS S - to - D i ch ot o m o u s, etc.
Eventually 128 sites had both SSS
and Dichotomous 15 instruments.
2. Collocated Site: A site containing
duplicate instruments of the same
type and usually by the same
manufacturer. Duplicates include
Dichotomous PM15 to Dichotomous
PM15, TSP-to-TSP, SSS,5-to-SSS15,
Dichotomous PM10to Dichotomous
PM10. Twelve sites were utilized for
collocated data collection
3. Key Site: An existing PM,5 site
which was augmented with a PM10
monitor. The objective for a key site
is to provide data for both PM16 and
PM10. Nine sites were designated as
key sites.
In early 1978, when the IP Network
was being planned, a recently developed
dichotomous sampler was available and
was incorporated into the network. This
sampler provided two particle size
fractions. The larger size fraction (Coarse)
included particles from 2.5 to 15 fjm
mean aerodynamic diameter. The smaller
size fraction (Fine) included particles
below 2.5 /ym. When added together, the
Fine and Coarse fractions give a "Total"
inhalable concentration in the 0-15 ^m
range (PM15). While the small fraction,
"Fine," is not a requirement for defining
an Inhalable Paniculate Standard perse,
it is useful in determining the origin of
particulates.
The dichotomous sampler was
therefore selected as the initial PM15
sampler because of availability and dual
size range fractions. It was (and is)
suitable for providing IP concentrations
and, when paired with the standard Hi-
Vol, IP/TSP relationships can be
developed. The dichotomous sampler is
more complex than the Hi-Vol and the
two sample fractions (Coarse and Fine)
require twice the sample handling,
weighing, calculation, etc., as the Hi-Vol.
Alternate samplers were therefore inves-
tigated. One PM15 sampler, The Size-
Selective Sampler (SSS), was developed
as a modification to a standard Hi-Vol and
tested at 50 of the first field sites. This
modified Hi-Vol sampler is identical to the
TSP Hi-Vol except that the gable roof is
replaced with a special mono-cut sampler
offering ease of operation, single sample,
large sample size, and associated cost
savings. For TSP, the High-Volume
sampler was used. Later a dichotomous
sampler modified to cut at 10 /jm was
added.
Results and Discussion
The IP data base includes 12, 385 TSP
Hi-Vol; 7,363 Size-Selective Inlet (SSS)
Hi-Vol; and 11,056 Dichotomous
Sampler 24-hour measurements,
collected on an every-sixth-day schedule.
Table 1 is a list of site locations. Volume I
of the report is a summary of mass data
collected. Volume II is a listing of individ-
ual 24-hour mass concentrations.
Chemical analysis is not a subject of this
report. Because of an initial emphasis on
sites in the Philadelphia area to support
other projects, a large percentage of the
total data collection is from this area.
Because of staggered sampler set up
schedules and/or sampler downtime,
TSP, SSS, and Dichotomous samplers
were not always operated simultaneously.
Therefore, care must be taken when
comparing means of different sampler
types. The number of samples, means,
standard deviation, minimum, maximum,
start date, and stop dates, are given in the
main report.
Precision estimates were made by
comparing duplicate, collocated like
instruments. Like instruments are
defined as similar instruments, or
dissimilar instruments designed to do the
same thing. Instruments were collocated
-------
Table 1. Inhalable Paniculate Network Site Locations
standard. A close approximation to equa-
tion (1) is:
Region State
Number
of Sites Location
4 Alabama 7 S. Birmingham, N. Birmingham. Inglenook,
Huffman, Mobile, Mtn. Brook. Tarrant
10 Alaska 1 Anchorage
9 Arizona 3 Carefree, Phoenix, N. Phoenix
6 Arkansas 1 Little flock
9 California 15 Azusa, Bakersfield, Chico, San Diego, Fresno, Five
Points, Livermore (2), Lompoc, W. Los Angeles,
Pasadena, Richmond, Rubidoux, San Francisco,
San Jose
8 Colorado 5 Denver (3), Pueblo, Ft. Collins
1 Connecticut 2 Hartford, Morris Dam
3 Delaware 2 Dover, Wilmington
3 D.C. 2 Washington (2)
4 Florida 1 Tampa
4 Georgia 3 Atlanta (2), Savannah
9 Hawaii 1 Pearl City
W Idaho 1 Boise
5 Illinois 4 Chicago f4)
5 Indiana 3 Gary, Indianapolis, Jeffersonville
7 Iowa 2 Marshal/town (2)
7 Kansas 3 Kansas City, Topeka, Wichita
4 Kentucky 2 Ashland, Louisville
1 Maine 1 Acadia
3 Mary/and 5 Baltimore (3). Rockville (2)
1 Massachusetts 4 Boston (2), Springfield, Worcester
5 Michigan 7 Detroit (2), Duluth, International Falls. Minneapolis
(21 St, Paul
4 Mississippi 1 Jackson
7 Missouri 3 St. Louis, Kansas City, E. St. Louis
8 Montana 2 Butte, Missoula
7 Nebraska 1 Omaha
9 Nevada 2 Reno, Winnemucca
2 New Jersey 3 Camden. Livingston. Jersey City
6 New Mexico 2 Albuquerque, Bayard
2 New York 7 Buffalo (2), Angola. Buffalo, NYC (3)
4 North Carolina 4 Charlotte, Durham, Res. Tri. Park 12)
5 Ohio 12 Akron, Cincinnati. Cleveland(3), Columbus, Dayton,
Ironton, Medina, Middletown,
Steubenville, Youngstown
6 Oklahoma 1 Oklahoma City
10 Oregon 3 Sauvie Island, Eugene, Portland
3 Pennsylvania 15 Bethlehem, Philadelphia (9), Pittsburgh (5)
1 Rhode Island 1 Providence
4 South Carolina 1 Charleston
4 Tennessee 2 Chattanooga. Nashville
6 Texas 6 Dallas, El Paso (2), Houston (3)
8 Utah 2 Magna, Salt Lake City
3 Virginia 6 Arlington, Hampton, Hopewell, Norfolk, Fairfax,
Richmond
10 Washington 3 Seattle 12). Spokane
3 West Virginia 3 Charleston. Weirton, Wheeling
5 Wisconsin 2 Beloit, Green Bay
Total 157
at selected sites and duplicate samples
were taken.
In the report, both bias and precision
are addressed using formulas for
summarizing paired data. The formulas
are the "percent difference" ratio type
commonly applied to collocated air
pollution data.
When one compares actual measure-
ments to a reference value or standard,
the signed value of the percent difference
is normally represented by:
Percent Difference =
Measurement-Reference
Reference
X 100
(D
Percent Difference =
Measurement 1 - Measurement 2
Average of the 2 measurements
(2)
X 100
If y, = measurement 1, and y2 =
measurement 2, equation 2 may be
rewritten as:
Vi - V2
When comparing two field instruments,
however, neither instrument is, in fact, a
Percent Difference = y, + y2 X 100
The results of solving this equation for
each pair of collocated sample measure-
ments is the signed percent difference,
R (CV/ A/~2 ). For analysis each resultant
(the signed percent difference R) was
treated as a statistical sample. The
hypothesis tested is that the average
difference (i.e., the relative bias) is zero
over each data set.
Data processing consists of:
1. Calculating the signed difference
expressed as a percent (R explained
previously) for each data-pair collec-
ted at each site (i.e.. Coarse Fraction
Dichot Mass collected in
Birmingham, Alabama).
2. Calculating % Standard Deviation
for each measurement pair.
3. Applying the Dixon Ratio to R for
each data set to test for outliers.
4. Removing outliers.
5. Calculating the coefficient of varia-
tion where CV (%) = Standard
Deviation of R divided by V2 .
6. Testing the Null Hypothesis that
R=O using the t-Test at a = .05.
7. Testing the homogeneity of
variances by applying Bartlett's test
to variance of each mass measure-
ment pair for each pollutant and
measurement method. The test
showed that a pooled estimate of
variance was generally not possible
although there were exceptions. It
appears that for all the Hi-Vol data
(TSP and 15 pm SSI) as well as
dichotomous data the variances are
nonhomogeneous.
Conclusions
Data from the EMSL, RTP Inhalable
Paniculate Network are described in the
-------
report. Individual values for TSP Hi-Vol;
PM15 Dichotomous Coarse, Fine, and
Total; and PM15 SSS mass are presented.
Ratios of Dichotomous Total-to-TSP Hi-
Vol, and SSS-to-TSP Hi-Vol are
summarized for PM,5 mass. Similar data
are presented for PMi0 sampling but on a
smaller number of samples. (More PM,n
samples will become available as PM,0
sampling continues throughout 1984).
Quality Control and Quality Assurance
procedures and results are presented in
the report and used to estimate sampling
accuracy by examining sample flow
rate, weighing accuracy, etc. Overall
sampling accuracy is difficult to
determine directly because the
measurement requires the production of
accurately known concentrations of
paniculate matter of a wide variety of
sizes.
Data precision is discussed using
paired data obtained from collocated
instrument sampling. The signed percent
difference of the two measurements
(expressed as R) was obtained by dividing
the difference between the data pair by
the average of the two measurements
and multiplying by 100. Student'ststatis-
tic was used to test the Null Hypothesis
that R = 0 (i.e., that the relative bias is
zero over each data set).
The value of t was statistically signifi-
cant at the 5% level for one or more sites
within each sampling class (Hi-Vol, SSS,
Dichoti5) meaning that the differences
between paired instruments is probably
real. Conversely, at least one site within
each class was not significant at the 5%
level. Overall there is substantial
variability but little bias across the entire
collocated data set.
The general contention that suspended
particulates are a complex mixture of
large and small particles, both naturally
occurring and man made, is supported by
the absence of a simple, consistent ratio
of IP to TSP. If the IP were a simple frac-
tion of TSP, a consistent ratio would be
expected and estimates of IP from past
TSP would have been possible. That this
is not the case and that IP is a complex
fraction of TSP is supported by the data
presented in the report.
The authors do not infer that for a
specific site, a consistent ratio of IP-to-
TSP is impossible. If a given site is
influenced by particulates originating
from a specific source, then the inhalable
fraction may possibly be a consistent sub-
set of TSP.
Certainly, at any given site, a ratio of IP-
to-TSP is mathematically possible, but
the actual value isdependent upon which
(if any) outliers are identified as flawed
.and not used in the computations.
Further experience with the operation of
these samplers and with the
interpretation of the resulting data will be
needed to resolve the questions raised
and to expand upon the conclusions that
can be drawn.
D. Hinton (also the EPA Project Officer, see below). J. Sune, J. Suggs, and W.
Barnard are with Environmental Monitoring Systems Laboratory, U.S. Envi-
ronmental Protect/on Agency, Research Triangle Park, NC 27711.
The complete report consists of two volumes:
"Inhalable Particu/ate Network Report: Operation and Data Summary (Mass
Concentrations On/y)~Vol. I., April 1979 - December 1982." (Order No. PB
85-148 682/AS; Cost: $19.00. subject to change).
"Inhalable Particulate Network Report: Data Listing (Mass Concentrations
Only)-Vol. II., AprH1979 - December 1982, "(OrderNo. PB85-148 690/AS;
Cost: $34.00, subject to change).
The above reports will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Environmental Monitoring Systems Laboratory
U.S. Environmental Protection Agency
Research Triangle Park. NC 27711
U.S. GOVERNMENT PRINTING OFFICE: 1985-559-016/27005
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Official Business
Penalty for Private Use S300
------- |