United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park NC 27711
EPA-450/4-83-01:
August 1983
Air
Analysis Of
Organic Compound
Data Gathered
During 1980 In
Northeast
Corridor Cities
-------
EPA-450/4-83-017
August 1983
Analysis Of Organic
Compound Data Gathered
During 1980 In
Northeast Corridor Cities
By
Harold G. Richter
Prepared For
U.S ENVIRONMENTAL PROTECTION AGENCY
Office Of Air, Noise And Radiation
Office Of Air Quality Planning And Standards
Research Triangle Park, North Carolina 2771 1
August 1983
-------
Analysis of Organic Compound Data Gathered During 1980
in Northeast Corridor Cities
Introduction
During the summer of 1980, an extensive air monitoring program was
carried out as part of the Northeast Corridor Regional Modeling Project
(NECRMP). The primary purpose of the monitoring program was to compile an
air quality and meteorological data base for use in photochemical models.
Since most of these models require ambient concentrations of organic
compounds as part of the input information, one of the principal efforts
of the study was to collect samples of air from major urban areas of the
Corridor and to analyze them for their organic species composition.
These data can be used for purposes other than modeling. They can
be used to characterize the area surrounding the point of collection, for
example (i.e., the emission sources in the vicinity), or they can be used
to estimate the contribution of mobile sources to the hydrocarbon burden
of the air. To whatever use they might be put, however, the data can
contribute to reliable conclusions only insofar as the user knows the
qualifications and peculiarities of the data. It is the purpose of this
report to acquaint data users with the large body of ambient hydrocarbon
compositional data acquired during the NECRMP, to alert the user to
limitations in some of the data from certain samples, and to show how the
data can be used for certain analyses. Complete data tapes are available
as part of the 1980 Northeast Regional Oxidant Study (NEROS) data compi-
lations from J. H. Novak, Meteorology and Assessment Division, Environ-
mental Sciences Research Laboratory, U.S. Environmental Protection Agency,
Research Triangle Park, North Carolina 27711.
-------
Site Description and Analytical Technique
The Northeast Corridor extends from Washington, DC, in the south to
Boston in the north. Samples of air for hydrocarbon analyses were collected
at surface sites in each of four metropolitan areas (Washington, Baltimore,
New York City-New Jersey, and Boston), as well as above those areas by air-
craft. Two surface sampling sites were chosen in each area, one in a high
traffic density, center-city location, and the other in an industrial location.
Figures 1 thru 4 show the locations of these sites in each metropolitan area.
Two one-hour integrated samples of air were collected in Tedlar and/or
Teflon bags at each site on most days of the monitoring program, one from
0600-0700 EOT and the other from 0800-0900 EDT. Aircraft grab samples were
collected in stainless steel canisters at various times and locations during
the day, as prescribed by a protocol determined by the meteorological con-
ditions prevailing on the individual days. Detailed records were made of the
flight paths and the locations, altitudes and collection times of these samples.
When time permitted, a third kind of sample was collected. These were called
"roving" samples and were grab samples collected at various surface sites
downwind of the Washington-Baltimore areas.
Analysis of a sample was carried out as soon as possible after collection,
in order to avoid loss of compounds from the containers. The elapsed time was
never more than 24 hours. Separate chromatographic and data report systems
were maintained for each metropolitan area (although the Washington and Balti-
more systems were housed in the same laboratory). Battelle-Columbus Labs
collected and analyzed samples in the Boston area, while We-jnington State
University collected and analyzed samples from the other three areas. Analysis
-------
Figure 1 Map showing sampling sites in the Washington, DC area
-3-
-------
BALTIMORE AREA
Figure 2. Map showing sampling sites in the Baltimore area,
-4-
-------
Figure 3- Map showing sampling sites in the New York City area
-------
rigure 4. Mao showing sampling sites in the Boston area
-------
was by GC-FID (gas chromatography with fl ame-ionization detection). Details
of the columns and procedure have been presented elsewhere.1-5
In all, more than 1300 air samples were collected during the 45-day
intensive data collection period from July 15 to August 30, 1980. Only 1120
samples were analyzed, however, because of the limited number of samples which
could be analyzed in a single day. Table 1 summarizes the number of samples
analyzed from each site.
Concurrent with the analysis of ambient samples, an extensive quality
assurance/quality control (QA/QC) program was carried out. This included
regular analyses of a pentane or neo-hexane (2, 2-dimethyl butane) standard.
In addition, each of the four analytical groups had a canister containing a
mixture of 14 organic species commonly found in urban atmospheres. For
statistical purposes, each group analyzed the mixture at least eight times
during the July 15 - August 30 period. Finally, a "round robin" sampling/
analysis program was also included in the QA/QC program. This consisted of
each analytical group collecting an ambient sample in a stainless steel
canister. After analyzing a fraction of the sample, the collecting group then
sent the canister to one of the other groups. That group, in turn, analyzed
a portion of the sample and sent the remainder to a third group which repeated
the operation. In all, five analytical groups participated, and each started
at least four samples. [The fifth group was from the Environmental Sciences
and Research Laboratory, Research Triangle Park, North Carolina (ESRL/RTP),
located in Columbus, Ohio, at the time.]
-------
Table 1
Summary Of Samples From NECRMP Sites
oo
Site
West End Library, Washington, DC
Takoma Park, Washington, DC*
University of Maryland
Read Street, Baltimore, MD
Essex, MD
Linden, NJ
Newark, NJ
East Boston, MA
Watertown, MA
Location
Washington - Baltimore
New York - New Jersey
Boston, MA
Location
Baltimore, MD - environs
Aircraft Samples
Roving Samples
Number of Samples^
60
16
44
82
90
81
71
85
90
619
Number of Samples
286
66
86
438
Number of Samples
63
Because uf space conflicts, the Takoma Park site in Washington, DC, had to be vacated early in the study period.
The university of Maryland site was then chosen to serve as a suburban site for the DC area.
-------
Results and Discussion
A. QA/QC Program
The results of the QA/QC program are reported in detail elsewhere.6»7
They are summarized here, however, because they are important to the discus-
sion of the data obtained during MECRMP, which follows. In brief, the coef-
ficient of variation for 20-25 analyses of calibration gas (pentane or neo-
hexane) at each laboratory fell between five percent and seven percent. This
is very satisfactory. Analyses of the standard mixture of 14 representative
organic compounds showed that the individual light compounds (carbon numbers
of C-2 to C-5) could be measured with a coefficient of variation of 15 per-
cent or less, but heavier species, particularly aromatic compounds, were
subject to much greater errors. Analysis of data from the round robin samples
shows that, although the various laboratories generally agreed (+_ 15 percent
coefficient of variation) as to the total concentration of compounds present
(in parts per billion carbon, ppb C), there was frequent disagreement as to
the identity of certain of the minor constituents, as well as sometimes large
disagreement (2-fold) in the concentration of total identified olefinic and/
or total aromatic compounds present.
These results show that light organic species can be measured with
very satisfactory precision. Heavier species are measured with less precision.
The reasons for this are not clear at the present time. Some evidence
(unpublished) has been gathered in the past, suggesting that passivated stain-
less steel containers may affect the behavior of heavy species, particularly
aromatics, but inconsistent data were obtained from the aircraft samples and
the mixture of 14 species during this study to corroborate the point.
-------
From one point of view, the round robin data are very encouraging.
They show that skilled analysts can measure ambient hydrocarbon species
rather precisely at the ppb C level. Different analysts usually agreed
fairly well on the individual concentrations of most of the lighter species
(hexane and below) which usually make up about 50 percent of the total concen-
tration. When the data user needs to know only the sum of hydrocarbon species
concentrations, he can be confident of a fairly reliable value. Much work
needs to be done, however, on quantification of the heavier species, identi-
fication of certain unknown peaks, and the frequent irreproducibility of
aromatic species data. And, although unrelated to the present discussion,
all studies to date of organic compounds in ambient air have almost completely
neglected separation and identification of the oxygenated and substituted
species known or thought to be present. Whereas these may be in low concen-
tration, a complete characterization of the pollutant burden must include
them.
Nevertheless, in spite of the above restrictions, because of the
excellent accuracy and precision in measuring the standard sample and the
hydrocarbon mixtures, these QA/QC data from NECRMP should allow data users to
have confidence in the GC analytical technique and in the data obtained by
experienced analysts.
B. Characteristics of the Data
The chromatographic procedure used in analysis of NECRMP samples is
capable of separating many compounds in the individual samples. Some of these
are identifiable on the basis of retention time. That is, they appear on the
chromatogram at times identical to those for known compounds. Frequently,
10
-------
other peaks appear on the chromatogram at times different from those times at
which known compounds appear. These peaks are called "unidentified compounds,"
and the individual concentrations are summed in order to estimate the total
concentration of all organic compounds (excluding methane) in the sample
(sum-of-species, ppb C).
During this study, 60 or more compounds were identified in the
samples. These are usually divided into three major chemical groups when
working with the data: olefins, paraffins and aromatics. One of the first
properties of hydrocarbon data sets which many analysts look at is the frac-
tional distribution of total carbon concentration among the chemical groups.
These distributions give some information as to the major sources of hydro-
carbons influencing the receptor, as well as pennitting a rough comparison of
data obtained during studies in other cities.
Table 2 summarizes these distributions, as well as listing two other
properties of the samples which warrant discussion: anomalous values of C2H4
(ethylene) and the fraction of unknowns in the samples. First, however, the
reader may notice that the number of samples from each site listed in Table 2
may be slightly lower than the number listed in Table 1. This is due to the
fact that duplicate samples were collected occasionally, and their analyses
are not included in the calculations of Table 2.
The percentages of the species were calculated by summing the concen-
trations of all members of the group (e.g., olefins) and dividing by the sum
of all identified species concentrations. (The difference between the sum of
these three percentages and 100 percent at any given site is the percent of
acetylene in the sample.) From the standard deviations, it is seen that the
11
-------
Table 2
Characteristics Of Data From Surface Samples
1
Site
West End Library,
Washington, DC
Takoina Park,
'Washington, DC*
University of Maryland
Read Street, Baltimore
Essex, MD
Linden, NJ
Newark, NJ
East Boston, MA
Watertown, MA
Average - all sites
Number
of
Sampl es
58
16
43
80
89
71
71
85
90
603
1
[Average
Percent
Olefins* °
9.1 6.1
5.8 5.6
9.2 2.9
10.6 5.3
6.2 2.8
8.7 3.6
9.3 3.0
8.4 3.7
14.2 6.2
9.0 4.5
Average
Percent
Paraffins °
58.7 5.6
63.0 11.6
54.0 8.8
59.1 7.3
60.6 7.9
64.3 11.0
57.5 5.7
65.7 10.5
58.3 7.7
60.1 8.5
Average
Percent
Aromatics CT
28.8 7.0
28.0 9.5
33.8 9.7
27.5 7.7
30.9 8.4
25.4 12.3
30.3 6.9
23.2 7.4
24.3 5.7
28.0 8.3
Average
Percent
Unidentified
17.5
19.4
18.4
12.3
15.2
12.1
14.4
15.4
18.7
15.9
C2H4
Anomalies
# of Samples
36
12
2
27
51
42
15
* Calculated from data for all samples, including those samples for which ethylene
concentrations were recorded as zero. See discussion on C2H4 anomalies.
-12-
-------
samples vary appreciably in their fractional composition of some of the
groups, particularly the olefins. The paraffin fraction, on the other hand,
is much less variable (coefficient of variation) among samples from a given
site than are the olefins or aromatics.
The penultimate column of Table 2, called "Average Percentage of
Unidentified," presents a characteristic of the data which should be of
interest to data users. As described above, unidentified compounds are those
which appear on the chromatogram at retention times different from those
times at which known compounds appear. Little or nothing is known about the
chemical structure of these compounds; thus, they cannot be placed in the
olefin, paraffin or aromatic categories. If the data user is interested in
knowing the total concentration of olefins (or paraffins or aromatics) in a
sample, he must consider the probability that some of the compounds in the
unidentified fraction are olefinic. That is, the total concentration of
olefins calculated from a simple summation of the identified olefins in a
sample is less than actually existed in the sample as collected. Without
more information about the distribution of hydrocarbon species in ambient
air, it would seem that the only alternative a data user has is to distri-
bute the unidentified compounds among the three major groups in the same pro-
portion as these individual identified groups stand to the total identified
hydrocarbon concentrations in the sample. Put more simply, if the concen-
tration of olefins were found to be "X" percent of the concentration of
identified compounds in a given sample, then the data user should assume, as
a first approximation, that the fractional concentration of olefins in the
unidentified compounds is also "X" percent. Whereas this will not change the
-13-
-------
fractional relationships among the major groups, it will change the absolute
concentrations of the group sums, as well as the calculated group fraction of
the recorded sum-of-species. Table 3 shows how these concentrations change,
using data from an actual sample (Read Street, Baltimore, 8/13/80, 0600-0700
EOT).
Comparing columns 2 and 6, or columns 3 and 7, shows that there is a
25 percent increase in the recorded concentration (or fraction) of the group
sums if the unidentified compounds are distributed proportionately among the
three groups. Acetylene, of course, remains the same concentration and frac-
tion, since it is an identified species. Such an increase in group concen-
trations would very probably have an appreciable impact on some uses of the
data.
The last column of Table 2, "C2H4 Anomalies," requires discussion,
too, since it may be of importance for some data uses. After analysis of the
samples, the analysts reported a zero value for ethylene concentration in
each sample for which there was evidence of ethylene contamination. Whenever
this contamination is found, it is almost always in samples from monitoring
stations housing chemiluminescent ozone analyzers. The reported sum of ole-
fins in samples which were contaminated is, therefore, much too small, if
only the concentrations of identified olefins are summed, since previous
studies3'9 have shown that ethyl ene frequently makes up about half of the total
olefin concentration in an uncontaminated sample. Table 2 shows that many
samples seem to have been contaminated with ethylene in some of the stations.
The problem with ethylene contamination affects other aspects of the
analysis of hydrocarbon data, although the impact may be less than it is when
examining only the olefin data. The total sum-of-species concentration,
-14-
-------
riroup
Table 3
Summary Of Data From Surface Samples
Read St., Baltimore 0600-0700 EOT, 8/13/80
1
fin
tic
lene
if ied
it if ied
ppb C
in
Sample
53.5
306.5
145.5
12.5
518.5
123.8
641.8
Fraction
of
Total
0.083
0.478
0.227
0.019
0.807
0.193
1.000
Fraction of
Identified
(less acetylene)
0.106
0.606
0.288
-
-
-
-
ppb C
Unidentified
(fraction x 123.8)
13.1
75.0
35.7
-
-
-
-
Corrected
ppb C
66.6
381 .5
181.2
12.5
-
-
641 .8
Corre
Fract
0.1
0.5
0.2
0.0
-
-
1.0
Olefin
Paraf
Aroma
Acety
IdiMit
i Un idei
Tolxl
-------
as recorded, is less than it actually was in the sample by the amount of the
missing ethylene concentration. This will typically cause an underestimation
of the sum-of-species concentration by about five percent, which may be
relatively unimportant for most uses of the data, but it may be important for
some. It will affect the calculated paraffin and aromatic fractions (of
total hydrocarbon concentation) to some extent also.
If the data user needs to estimate the value of a missing ethylene
concentration, he has two alternatives. He can either assume the missing
concentration is equal to that of the sum of all other olefin concentrations,
or he can calculate the average C2H4/C2H2 (ethylene/acetylene) ratio in uncon-
taminated samples from the same sampling site and use that ratio and the
acetylene concentration in the sample of interest to approximate the missing
ethylene concentration. Both methods usually give similar concentrations,
although the latter method is probably more correct.
A different anomaly in ethylene concentrations is noted in data from
samples collected and analyzed at the two Boston sites. In many of these
samples, ethylene is found to make up 70-90 percent of the total olefin
compounds. This cannot be attributed to a source of ethylene near the Boston
sites, however, as shown by an inspection of data from the round-robin sam-
ples. (Round-robin samples were grab samples -- not 1-hour integrations --
collected usually in the vicinity of the monitoring stations and were thus
similar in species concentration to the daily samples.) In those round-robin
samples collected in Boston, and analyzed by the other laboratories, ethylene
made up about 50 percent or less of the total olefin concentration, similar
to that found in samples from other areas. Furthermore, a close analysis of
-16-
-------
the round-robin data shows that the Boston analysts reported ethylene concen-
trations very similar to those concentrations found by other analysts of the
same samples. It thus seems that the Boston analysts reported fewer or
smaller concentrations of other olefins in the samples, thereby making the
ethyl ene/total olefin ratios high. Whatever the reason, these developments
introduce uncertainty as to the reliability of the olefin group fractions
and, to a lesser extent, of other group fractions and sum-of-species con-
centrations for the Boston sites.
C. Uses of the Data
In addition to using the NECRMP hydrocarbon species data for modeling
purposes, there are other uses which make the data base valuable, in spite of
its limitations. Some of these are discussed below. Other uses and analyses
will undoubtedly be made in the future.
First, it is necessary to state, explicitly, a rule which all data
users should follow when using these data, as well as when using all other
data bases: Simple summaries of data may be very misleading. Data users
should become very familiar with the details of the data base they work with.
Whereas it may be time consuming, it is imperative that users examine individ-
ual sample data, rather than summaries, in order to avoid arriving at wrong
conclusions. The rule seems obvious, but it is too frequently ignored.
The value of this rule should be clear from the discussion above about
anomalies in olefin concentrations in this data base. Table 4 shows how a
simple summary of average olefin fraction (all samples) at the various NECRMP
sites differs from a summary of average olefin fraction of only those samples
in which ethylene concentrations were reported. Large errors in calculated
-17-
-------
Table 4
Comparison of Olefin Fractions
All Samples Vs. Only Those Reporting Ethylene
Site
West End Library,
Washington, DC
Read Street, Baltimore, MD
Essex, MD
Linden, NJ
Newark, NJ
East Boston, MA
Water town, MA
mean
0.091
0.106
0.062
0.087
0.093
0.084
0.142
Al 1 Sampl es
a
0.061
0.053
0.028
0.036
0.030
0.037
0.062
n
58
80
89
71
71
85
90
mean
0.139
0.122
0.077
0.100
0.097
0.084
0.142
Only Those With C2H4
a
0.048
0.053
0.033
0.035
0.026
0.037
0.062
n
22
53
38
29
56
85
90
CD
I
-------
olefin fractions are thus possible at some of the sites if the ethylene
problem is unknown or ignored. Olefin fractions at the two Boston sites are
included in Table 4 for comparison with the other sites. It is apparent that
a significant difference in the spectrum of hydrocarbon species exists among
the several sites, although more data are necessary in order to determine how
statistically important this is.
1. Compound/Acetylene Ratios
One use to which hydrocarbon species data is sometimes put is to
estimate the contribution of automotive emissions in the samples. This is
done by assuming - and it is usually a sound assumption - that the only source
of acetylene is from automobiles. If one knows the fraction of acetylene in
auto exhaust, then the olefin/acetylene ratio or the paraffin/ acetylene
ratio in an ambient sample is a measure of automobile contribution to the
hydrocarbons in the sample. These fractions have been measured in both pure
exhaust^ and in highway tunnels.3,9,11 Table 5 summarizes these NMHC/acety-
lene ratios.
There is some agreement among ratios, but the contribution of
auto exhaust to observed hydrocarbon levels in ambient samples cannot be
asserted too rigidly. Using Lonneman's ratio from the recent Lincoln Tunnel
study for total NMHC/acetyl ene of 29.7, which is believed to be the most
representative of exhaust from the current mobile source fleet, the percent
of automobile contribution to observed hydrocarbon concentrations at the
various NECRMP sites can be estimated by multiplying the average acetylene
concentration by 29.7 (giving an estimate of the HC sum-of-species concen-
tration contributed by mobile sources) and dividing that product by the
average HC sum-of-species concentration at the site. Table 6 shows these
-19-
-------
Table 5
Nonmethane Hydrocarbon/Acetylene Ratios in Auto Exhuast
1
Ratio
01 ef in/Acetylene
Paraffin/Acetylene
Aromatic/Acetyl ene
Total NMHC/Acetylene
Black8
(exhaust)
4.7
10.2
3.0
22.0
Lonneman^
(tunnels)
3.24
6.81
3.87
13.90
Lonneman^
(Houston)
3.08
7.59
5.22
15.90
Lonnemanll
(Lincoln Tunnel )
6.05
13.1
9.53
29.7
I
ro
CD
i
-------
Table 6
Estimated Automobile Contribution To Observed HC Sum-of-Species
IXi
I
Site
West End Library,
Washington, DC
University of Maryland
Read Street, Baltimore
Essex, MD
Linden, NJ
Newark, NJ
East Boston, MA
Watertown, MA
Acetylene
ppb C , average
17.8
16.5
15.6
11.1
10.6
12.6
12.1
7.7
HC Sum-of-Species
(ppb C, average)
714
666
639
606
961
510
777
352
Percent
Auto
74
74
72
54
33
73
46
60
-------
results.
At four of the sites, one may conclude that more than 70 percent
of the observed hydrocarbons come from auto exhaust. The Linden, New Jersey,
site shows a smaller contribution from automobiles. However, there is a large
refinery nearby, and a more detailed analysis of the data would be required to
establish the impact from this source more firmly. Similarly, a petroleum tank
farm was located upwind (on certain days) from the East Boston site. On these
days, the total hydrocarbon concentration in the samples was several ppm C, most
of it paraffinic. Excluding these data from the calculation would increase
the estimated mobile source contribution to nearer 70 percent. This higher
number for automobile contributions to ambient hydrocarbons would then reflect
the situation which prevails around the sampling site most of the time.
There is no rigorous proof of the correctness of this method for
estimating the fraction of mobile source contribution to ambient hydrocarbons.
The basis seems sound, and the calculated fraction agrees with one's impression
of mobile sources in the neighborhood, but an experimental program to test the
the assumptions would be very difficult to mount.
Another example of compound/acetylene ratio analysis can show how
this technique may help characterize the vicinity around a sampling site, as
well as to illustrate how the use of data summaries without a critical review
may be mi sleadi ng.
Some hydrocarbon compounds in the ambient air usually come from
a limited number of sources. Large concentrations of ethane, for example,
probably have commercial natural gas as their source. Large concentrations
of propane probably come from liquefied petroleum gas sources. Benzene and
toluene generally come from evaporated solvents, while n-butane and i-pentane
22
-------
come from gasoline vapors. Since the absolute concentrations of any of these
species may vary from site to site (because they also are found in auto
exhaust), the best way to compare one site with another is by comparing the
compound/acetylene ratios (i.e., acetylene is a normalizing factor).
The average n-butane/acetylene and i-pentane/acetylene ratios
in atmospheres not dominated by point sources of these two compounds is about
3 and 5, respectively.13 These ratios are found at all of the NECRMP sites,
except East Boston and Linden, where they were 12.5 and 15.7, respectively.
This indicated a probable source of gasoline evaporative loss in the vicinity
of the site. Closer examination of the individual sample data for the East
Boston site, however, showed that in only a few samples were these ratios
elevated. All others were in the customary range. On those certain days,
the ratios were very high (100-300). These dominated the averaging process,
potentially leading to an erroneous conclusion that the site was continually
surrounded by such a source.
Examination of wind directions on those days of high ratios showed
that the wind was always from the NW, 310°-350°. The sum-of-species concentra-
tions on these days always exceeded 1 ppm C. The conclusion seems now very
clear: there was a large source of butanes and pentanes, probably from evapo-
rative emissions, NW of the East Boston site. On further investigation, it
was found that a gasoline storage tank farm was located at some distance NW
of the site. In retrospect, the choice of the location for the monitor proved
to be a less than ideal location at which to measure hydrocarbons. However,
this was suspected when the siting decisions were being made, but other factors
dictated locating the hydrocarbon monitor at this site.
-23-
-------
2. Sum-of-Species/NMOC Comparisons
Another use to which species data are put is to compare sum-of-
species concentrations with concentrations of total NMOC recorded by collocated
continuous instruments. At seven of the sampling sites (Read St., Baltimore;
Essex, MD; West End Library, DC; Linden, NJ; Newark, NJ; East Boston; and
Watertown, MA), continuous NMOC analyzers (MSA 11-2) were set up to measure
ambient hydrocarbons during the NECRMP- These insruments were calibrated with
great care (propane standard) and maintained by experienced technicians in
order to gather the best data possible. They sampled ambient air from the
same manifold which supplied air to the Teflon/Tedlar bags for species analysis.
Analysis of data from these simultaneous hourly measurements may be interpreted
in more than one way: a comparison of how well (or poorly) behaved are contin-
uous NMOC analyzers compared with the sum-of-species; how comparable are the
data from different instruments of the same make; or how confident a modeler
might be in using NMOC instrument data for ozone modeling.
Figures 5 thru 11 show the results of comparisons of NMOC data
from analyzers with sum-of-species. Data have been plotted on log-log scales
in order to simplify comparison of the different sites. Such plots provide
greater resolution at the lower end of the scale than do linear plots, as well
as allowing percentage differences to be read directly from the plot. Sum-of
species includes the unidentified compounds, as well as the identified species.
As noted previously, no correction was made to the sum-of-species data for
missing ethylene concentrations in those samples where ethylene contamination
was evident (and the ethyl ene concentration was recorded as zero). To do so
would increase all sum-of-species concentrations by a few percent, thereby
translating some of the points in the figures slightly to the right.
-24-
-------
5.0
4.0
3.0
2.0
O 1-0
£
o.
a
cc
in
£ 0.5
_l
I 0.4
<
O 0.3
O
1 0.2
0.1
0.05
0.1
0.2 0.3 0.4 0.5 1.0
SUM-OF-SPECIES, ppm C
2.0 3.0 4.0 5.0
Figure 5. Hest End Library: NMOC Analyzer Sum-of-Soecies Conpariscn
-------
5.0
4.0
3.0 —
2.0
1.0
0.
Q.
OC
LU
£ 0.5
< 0.4
<
O 0.3
O
1 0.2
0.1
0.05
0.1
0.2 0.3 0.4 0.5 1.0
SUM-OF-SPECIES. ppmC
2.0 3.0 4.0 5.0
Figure 6. Read Street, Baltimore: NMOC Analyzer - Sum-of-Species Comparison
-i.0-
-------
5.0
4.0
3.0
2.0
a 1.0
EC
UJ
N
2 0.5
O 0.4
O
2 0.3
0.2
0.1
0.05
0.1
0.2 0.3 0.4 0.5 1.0
SUM-OF-SPECIES, ppm C
2.0 3.0 4.0 5.0
Figure 7. Essex, MD: NMOC Analyzer - Sum-of-Species Comparison
-------
0.1 0.2 0.3 0.4 0.5 1.0 2.0 3.0 4.0 5.0
NMOC ANALYZER, ppm C
Figure 8. Linden, NJ: NMOC Analyzer - Sum-of-Species Comparison
-28-
-------
5.0
4.0
3.0
2.0
o 1.0
Q.
o.
cc
01
N
>
_J
Z
0
o
Z
0.5
0.4
0.3
0.2
0.1
0.05
l!_3L
0.1
0.2 0.3 0.4 0.5 1.0
SUM-OF-SPECIES, ppm C
2.0 3.0 4.0 5.0
Figure 9. Newark, NO: N'MOC Analyzer - Sum-of-Species Comparison
-29-
-------
5.0
4.0
3.0
2.0
O
i—
I
ec
in
M
2 0.5
O 0.4
O
I 0.3
0.2
0.1 —
0.05
1:1
0.1
0.2 0.3 0.40.5 1.0
SUM-OF-SPECIES, ppm C
2.0 3.0 4.0 5.0
Figure 10. East Boston, MA: NMOC Analyzer - Sum-of-Species Comparison
-3Q-
-------
0.05
0.1
0.2 0.3 0.4 0.5 1.0
SUM-OF-SPECIES, ppm C
2.0 3.0 4.0 5.0
Figure II. Watertown, MA: MMOC Analyzer - Sum-of-Species Comparison
-------
Perfect agreement between all of the measurements would have all
of the data lying on the 1:1 line. There is no absolute way to determine which
of the measurements in such a comparison is correct. There is some justifica-
tion, too, for thinking that the two procedures do not measure equally well the
same species in all the samples. Because of the known difficulties with con-
tinuous NMOC instruments,10.12 however, and the satisfactory results which the
five analytical groups obtained in analyzing the round-robin samples (_+ 15 per-
cent coefficient of variation in sum-of-species, 65 analyses), it seems reason-
able to assume, for this study, that the sum-of-species procedure yields more
uniformly reliable data than do the NMOC instrument readings.
Little can be generalized about these comparisons. It must be
borne in mind that the NMOC instruments were not maintained as well as they
would have been in a well controlled laboratory study. Thus, there could be
some slight inaccuracies in sample timings (both NMOC data and species collec-
tion timing), calibration, voltage and/or temperature fluctuations, or perhaps
other problems. Nevertheless, the data obtained from the instruments is
thought to be the result of above average attention to technical details and
to be an accurate representation of the data obtainable from these instruments.
At most of the sites, the data scatter widely. Only at the East
Boston site (Figure 10) does there seem to be consistent agreement between
the two methods, although there is a positive bias of about 15 percent in the
NMOC continuous instrument data. Table 7 summarizes statistical information
about the least squares analysis of these natural logarithmic data. As the
table shows, only data from the East Boston site exhibit a high r-square value.
At five of the sites, half or more of the samples contained less
than 600 ppb C of hydrocarbons. This is of some concern, since the comparisons
-32-
-------
Table 7
Sum-of-Species/NMOC Comparisons, Least Squares Regressions
OJ
(.0
Site
West End Library,
Washington, DC
Read Street,
Baltimore, MD
Essex, MD
Linden, NJ
Newark, NJ
East Boston, MA
Watertown, MA
Slope
0.552
0.113
0.835
0.531
0.987
1.108
0.750
Intercept
- 0.552
- 0.283
- 0.101
- 0.277
+ 0.095
- 0.568
Standard Error
0.672
0.713
0.599
0.865
0.574
0.327
0.574
r-square
0.169
0.0077
0.354
0.141
0.467
0.887
0.475
-------
of NMOC data and sum-of-species show that there can be large differences in
concentrations, as measured by the two techniques. Data from the NMOC analy-
zers were all carefully examined by the contractor responsible for their
operation. All questionable data were discarded; thus, only "verified" NMOC
data - that which might be used by an agency for its regulatory purposes have
been archived. There is good reason to expect no more accurate data from
NMOC instruments in routine field monitoring programs which are likely to be
maintained with less care than were the instruments in this study. The poten-
tial for large error in hydrocarbon concentration data is then reflected in
greater uncertainty in modeling predictions, if the continuous instrument
data were to be used for this purpose.
The NMOC instruments should not have been affected by ethylene
contamination, as were the samples destined for species analysis, for the
following reason. Most, if not all, of the collocated ozone analyzers were
fitted with ethylene oxidizers which were designed to destroy the excess
ethylene after it left the ozone instrument. Thus, assuming properly func-
tioning oxidizers, no ethylene should have been exhausted to outside where it
could be entrained in manifold air from which all instruments (including both
the Tedlar bags and the NMOC device) sampled. Moreover, if ethylene had
entered the NMOC instruments through entrainment, concentrations of NMOC from
these monitors should have been surprisingly large, most if not all the time,
since the ozone and NMOC instruments were running continuously. On the other
hand, leakage of ethylene at the instrument and from plumbing connections
inside the monitoring shelters could have exposed the Tedlar bags to about 24
hours of a low level (ppm) ethylene atmosphere. Small molecules, ethylene
particularly, are known to diffuse readily through thin Tedlar membranes. No
34
-------
such diffusion is possible into the NMOC instrument. This is not a completely
satisfying explanation, because it is difficult to believe, with the air con-
ditioning maintained in the sampling stations, that much ethylene could collect
and diffuse into the bags. In summary, the source of the ethylene contamina-
tion of GC samples at certain times is unknown.
In a laboratory studylO comparing continuous NMOC analyzer data
with sum-of-species data, smaller scatters of data were observed. In that
referenced study, more than 80 percent of the NMOC data from the MSA 11-2
instrument lay above the 1:1 line. The slope was 1.06 and intercept 0.079,
n = 48, and r^ = 0.951. These studies show that much more work is needed to
resolve the differences or to get better, consistent agreement between these
two measurement techniques.
3. Sum-of-Species/NOx Ratios
At some of the sites, NOX chemiluminescent analyzers were used
to measure ambient NOX concentrations. Not as many data were collected from
these instruments as was desired because of malfunctions. Table 8 shows the
average sum-of-species/NOx ratios. The hourly sum-of-species averages have
been listed for all sites studied, as well as the (simultaneous) hourly NOX
averages. (Note that the sum-of-species/NOx ratio recorded for the sites in
Table8 is not obtained by dividing the average sum-of-species by average
NOX. The ratios in column 3 were obtained by averaging all the individual
sum-of-species/NOx ratios recorded at those sites.)
Except as described below, all data pairs were used for these
averages, including those hydrocarbon samples which exhibited C2H4 anomalies.
Correction of these samples for missing C2'ri4 concentrations would increase the
ratios of Table 8 by less than five percent.
-35-
-------
Table 8
Sum-of-Species/NOx Ratios
Site
West End Library,
Washington, DC
University of Maryland
Read Street, Baltimore
Essex, MD
Linden, NJ
Newark , NJ
East Boston, MA
Watertown, MA
Sum-of -Species
(Average)
0.714
0.666
0.639
0.606
0.961
0.510
0.777
0.352
NOX
(Average)
0.094
0.095
0.074
0.059
0.063
Ratio (Average)
+ Standard Deviation
13.5 + 10.7
9.1 + 5.4
10.7 + 5.4
23.9 + 28.5 (16.2 + 9.9)*
( n ^ 61 )
31.3 + 79.7 ( 9.6 + 6.3)*
( n ^ 61 )
n
56
--
49
28
65
--
78
OJ
en
* Certain high ratios eliminated for these calculations (see text).
-------
The high ratios in Linden, NJ and East Boston, MA stand out,
including the high standard deviations for data from these two sites. Since
there is a petroleum refinery near the Linden site and a tank farm upwind (on
certain days) of the East Boston site, the individual sample data were inspec-
ted to determine if more information could be gleaned from an analysis of less
source-specific data.
Data from the Linden site show that two of the individual sum-
of-species/NOx ratios were above 134, and two others were above 80. In three
of these four, the NOX concentrations seemed very low (around 0.01-0.02 ppm),
while the fourth had a sum-of-species concentration of more than 6.8 ppm C.
Eliminating these questionable samples from the calculations reduces the sum-
of-species/NOxratio for the Linden site to 16.2 +_ 9.9 (n = 61), which is a
ratio comparable to the other cities. On the other hand, it should be borne
in mind that relatively large sum-of-species /NOX ratios may be characteristic
of highly industrialized areas.
Similarly, eliminating from the calculations those samples col-
lected at the East Boston site during which the wind was from the direction of
the tank farm gives a sum-of-species/NOx ratio of 9.6 _+ 6.3 (n = 61). This
ratio is now in line with the other three sites in Table 8.
All data pairs were used for the calculations of Table 8, except
those noted in the above two paragraphs. A more informative number is the sum-
of-species/NOx ratio on those days which experienced high ozone values later
in the day at a downwind monitoring site (although it is recognized that other
factors - mainly meteorological - also contribute to the formation of ozone).
Using only these days of high ozone for the HC/NOX ratio calculation eliminates
periods when other factors (1 ow temperature, cloud cover, adverse winds, etc.)
-------
mitigate against ozone formation. Table 9 summarizes these data for all days
on which an ozone concentration of 0.15 ppm, or greater, was recorded at a
site within about 50 miles of the station where hydrocarbon and NOX concen-
trations were measured. Table 9 also shows NMOC/NOX ratios on high ozone
days together with sum-of-species/NOx and NMOC/NOX ratios (from Table 8) on all
days of observation. These NMOC/NOX ratios were included in order to gain
some impression of the differences a photochemical modeler (particularly with
the EKMA) might find if he had only continuous NMOC instrument data to use.
The comparison of ratios on high ozone days does not seem as bad
as does the comparison of ratios on all days. But, as stated earlier, statis-
tical summaries should be used with great caution. If the individual data
which make up these summaries are examined, the photochemical modeler would
have reason to be concerned. Table 10 shows the sum-of-species, NMOC, and
NOX data on high ozone days during the NECRMP in Essex, MD together with the
calculated ratios. A similar picture of the data is seen at all the other
sites on high ozone days: there is frequently a great difference between
sum-of-species/MOx ratios and NMOC/MOX ratios on days of interest to photo-
chemical modelers. Clearly, more work is needed to reconcile the differences
in data obtained from these two measurement techniques.
More can be gleaned from Table 9. Both the Essex and the Linden
sites show sum-of-species/NOx ratios quite different from the other three sites
on high ozone days. These two sites would probably be classified as industrial
sites on the basis of these data, both exhibiting an excess of hydrocarbon as
compared with sites influenced mostly by automobile exhaust. This is borne out
to some extent by the data in Table 6. There, it is estimated that only 54
percent of the measured hydrocarbons come from automobiles at the Essex site,
-38-
-------
Table 9
Hydrocarbon/N0x Ratios On High Ozone Days
Site
West End Library,
Washington, DC
Read Street,
Baltimore, MD
Essex, MD
Linden, NJ
East Boston, MA
Ratio - Higf
Sum-of-Species/NOx
mean a n
7.8 3.6 16
6.9 2.7 15
12.8 8.9 10
13.5 5.6 33
6.0 0.5 4
i 03 Days
NMOC/NOX
mean o n
6.6 2.8 16
10.8 5.3 15
12.0 7.3 10
11.6 10.5 29
7.4 0.9 4
Ratio -
Sum-of-Speci es/NOx
mean a n
13.5 10.7 56
9.1 5.4 49
10.7 5.4 28
23.9 28.5 65
31.3 79.7 78
All Days
NMOC/NOX
mean 0 n
6.2 3.2 51
12.9 12.6 36
11.1 5.7 27
14.2 20.2 60
19.6 29.1 73
I
CO
-------
Table 10
Sum-of-$pecies/NOx And NMOC/NOX On High Ozone Days In Essex, MD
Date
8/14
8/24
8/26
8/27
8/29
1
Samp! i ng
Time
6-7
8-9
6-7
8-9
6-7
8-9
6-7
8-9
6-7
8-9
Sum-of-Species
ppmC
0.676
0.593
0.339
0.236
0.944
0.793
2.755
0.984
0.516
0.552
NMOC
ppmC
0.90
0.78
0.27
0.18
1.14
1.17
2.22
0.60
0.45
0.48
NOX
ppmC
0.099
0.120
0.050
0.013
0.081
0.106
0.159
0.137
0.039
0.016
Sum-of-Speci es/NOx
Ratio
6.8
4.9
6.8
18.2
11.6
7.5
17.3
7.2
13.2
34.5
NMOC/NOX
Ratio
9.1
6.5
5.4
13.8
14.1
11.0
14.0
4.4
11.5
30.0
Ozone
ppm
0.15
0.15
0.17
0.18
0.15
-------
and only 33 percent at the Linden site. In contrast, hydrocarbons at the
West End Library site were estimated as 74 percent automotive; the University
of Maryland site also as 74 percent; Read St, Baltimore as 72 percent; and
Newark, NJ as 73 percent. The two Boston sites were lower, particularly the
East Boston site, where the estimation was probably heavily influenced by the
few days when wind was from the direction of a source to the northwest.
4. Aircraft and Roving Van Sample Data
Only a few words need be said about samples collected by air-
craft and roving vans. The data user would be well advised to examine the
individual sample data, rather than summaries of the data. Because of the
frequently very low concentrations of species in the samples (80 percent of
the aircraft samples and 20 percent of the roving van samples had less than
100 ppb C total hydrocarbon), there is great opportunity for large errors in
the group fractions. The unidentified components of some aircraft samples
exceed 50 percent of the total. Some aircraft samples are reported to have
no olefins, others no aromatics. In some samples, the sum of aromatics is
reported to exceed the sum of paraffins.
The roving van data, too, must be examined carefully. It may be
that the chromatograph was not functioning properly during analysis of some
of the samples, because some of the most commonly occurring aromatics are
reported missing. In other samples, ethylene is the only reported olefin.
These results are very unusual, and the user must assure himself, by using
other information or by consulting chemists, that the reported numbers are
reliable.
41
-------
Summary
1. During the 1980 Northeast Corridor Regional Modeling Project (NECRMP)
about 1120 ambient air samples were collected and analyzed from surface sites
and by aircraft for hydrocarbon species analysis. The surface sites were in
the metropolitan areas of Washington, DC, Baltimore, New York-New Jersey, and
Boston.
2. An extensive QA/QC program carried out during the NECRMP showed that
five participating laboratories could analyze standard gases repetitively with
a coefficient of variation of 5-7 percent. C-2 to C-5 compounds in a standard
mixture of 14 compounds were measured repetitively with a coefficient of vari-
ation of 15 percent. Heavy hydrocarbons, particularly aromatic species, were
measured with much less precision. Round-robin samples (ambient samples col-
lected and analyzed by one participating laboratory, then sent for consecutive
analyses by all other participants) showed reasonable agreement (+_ 15 percent
coefficient of variation) as to total concentration of hydrocarbons in the
samples, but there was frequent disagreement as to identification and quantity
of some of the minor species.
3. There was an average of 16 percent (in concentration) of unidentified
species in the 619 surface samples analyzed. Aircraft samples averaged much
more. About 45 percent of the surface samples analyzed at six of the eight
sites evidenced anomalous ethylene concentrations or ethylene contamination.
4. Data users are advised to examine data from the individual samples,
rather than rely on summary statistics of those data, because of the character-
istics summarized in "3" above.
-42-
-------
5. Mobile source contribution to the observed hydrocarbon at each site
was estimated. Automobiles were calculated to contribute more than 70 per-
cent of the hydrocarbons at four of the eight sites, but the method of esti-
mation is difficult to verify, thus the reliability of the estimate may have
1imited useful ness.
6. A comparison of sum-of-species concentration with data from collocated
NMOC continuous analyzers was made. There was reasonable agreement among the
data points at only one of the seven sites having NMOC instruments. Assuming
sum-of-species data are more uniformly reliable than data from the continuous
instruments, these latter did not agree very well with GC measurements under
field conditions on this project. Even at sum-of-species concentrations exceed-
ing 0.5 ppm C, there was reasonably good agreement between the two methods at
only three of the sites.
7. Sum-of-species/NOx ratios were computed for sites where NOX instru-
ments were collocated with species samplers. Computations were made, both
for all days and for those days when ozone concentrations greater than 0.15
ppm were recorded downwind of the metropolitan areas. The average of this
latter ratio differs by more than 2-fold between certain cities, but the
number of high-ozone days was relatively few, and thus the statistical signi-
ficance of the data is open to question.
8. Sum-of-species/NOx ratios were compared with NMOC/NOX ratios, both on
high-ozone days and on all days for which data were available. The average of
the ratios by these two methods on high ozone days seem to agree better than
the average of the ratios on all days for which data are available, but compar-
ison of individual day's ratios show large differences.
9. The fractions of species groups (olefins, paraffins, and aromatics)
were shown to be fairly uniform among the four metropolitan areas.
-43-
-------
References
1. W. A. Lonneman, "Ozone and Hydrocarbon Related Measurements in Recent Oxidant
Transport Studies," International Conference on Photochemical Oxidant Pollution
and Its Control, Proceedings: Volume I, EPA-6QU/3-77-001a, U.S. Environmental
Protection Agency, Research Triangle Park, NC 27711, January 1977, pages 211-223.
2. H. H. Westberg, R. A. Rastnussen, and M. Holdren, "Gas C'nromatographic Analysis of
Ambient Air for Light Hydrocarbons Using a Chemically Bonded Stationary Phase,"
Anal . Chem. 46_ (1974).
3. W. A. Lonneman, S. L. Kopczynski, P. E. Darley, and F. D. Sutterfield, "Hydro-
carbon Composition of Urban Air Pollution," Env. Sci. Tech. 8(3) 229 (1974).
4. H. Westberg, K. Allwine, and E. Robinson, "Measurement of Light Hydrocarbons and
Oxidant Transport, Houston Area, 1976," EPA-600/3-78-062, July 1978.
5. R. L. Seila, "Nonurban Hydrocarbon Concentrations in Ambient Air North of Houston,
Texas," EPA-600/3-79-010, February 1979.
6. H. H. Westberg, W. A. Lonneman, and M. Holdren, "Analysis of Individual Hydro-
carbon Species in Ambient Atmospheres - Techniques and Data Validity," American
Chemical Society Annual Meeting, Kansas City, MO, September 1982.
7. W. A. Lonneman — to be published.
8. F. Black and L. High, "Automotive Hydrocarbon Emission Patterns and the Measure-
ment of Nonmethane Hydrocarbon Emission Rates," Paper 77-0144, International
Automotive Engineering Congress and Exposition, Detroit, February 28 - March 4,
1977.
9. W. A. Lonneman, G. R. Namie, and J. J. Bufalini, "Hydrocarbons in Houston Air,"
EPA-600/3-79-018, February 1979.
10. F. W. Sexton, R. M. Michie, Jr., F. F. McElroy, and V. L. Thompson, "A Compara-
tive Evaluation of Seven Automated Ambient Nonmethane Organic Compound Analyzers,"
EPA-600/54-82-046, August 1982, Environmental Monitoring Systems Laboratory,
Research Research Triangle Park, NC 27711.
11. W. A. Lonneman, Lincoln Tunnel Studies, 1982 --- unpublished data.
12. L. R. Rechner, "Survey of Users of the EPA Reference Method for Measurement of
Nonmethane Hydrocarbons in Ambient Air," EPA-650/4-75-008, December 1974;
F. F. McElroy and V. L. Thompson, "Hydrocarbon Measurement Discrepancies Among
Various Analyzers Using Flame-Ionization Detectors," EPA-600/4-75-010, September
1975; J. W. Harrison, M. L. Timmons, R. B. Denyszyn, and C. E. Decker, "Evalua-
tion of the EPA Reference Method for the Measurement of Nonmethane Hydrocarbons
- Final Report," EPA-600/4-77-033, June 1977.
13. J. R. Martinez, F. L. Ludwig, and C. Maxwell, "1978 Houston Oxidant Modeling
Study, Volume I: Data Evaluation and Analysis," March 1982, (SRI Project No.
7938) p. 57ff.
-44-
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse betore con pictinz;
1 REPORT NC 2. |3. RECIPIENT'S ACCESSION NO
EPA-450/4-83-017
1. TITLE AND SUBTITLE 5. RE
Analysis Of Organic Compound Data Gathered During
1980 In Northeast Corridor Cities S-FE
7. AUTHOR(S) 8. PE
Harold G. Richter
PORT DATE
RFORMING ORGANIZATION CODE
RFORMING ORGANIZATION REPORT \iO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS ! 10. PROGRAM ELEMENT NO.
Office Of Air Quality Planning And Standards
U . S. Environmental Protection Agency 11. c
Research Triangle, NC 27711
12. SPONSORING AGENCY NAME AND ADDRESS 13. T
14. Sf
15. SUPPLEMENTARY NOTES
ONTRACT GRANT NO.
YPE OF REPORT AND PERIOD COVERED
'ONSORING AGENCY CODE
16. ABSTRACT
During the summer of 1980, an extensive monitoring proiect was carried out
as part of the Northeast Corridor Regional Modeling Project (NECRMP) , the primary
purpose of which was to compile an air quality and meteoroloeical data base for use
in photochemical models. Since most of these models require ambient concentrations
of organic compounds as part of the input information, one of the principal efforts
of the study was to collect samples of air from major urban areas of the Corridor
and to analyze them for their organic species composition.
17. KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS b. I OENTI Fl E RS. OPEN EN
1
2C SEC.s T. c.-S; .-•
DED TERMS C. COSATI 1 idd'fjruup
i
aR..p.r: :1 ;-^.,f r-ES
'-' •"-'-'-' ' - - ''^ - =-
-------
INSTRUCTIONS
1. REPORT NUMBER
Insert the EPA report number as it appears on the cover of the publication.
2. LEAVE BLANK
3. RECIPIENTS ACCESSION NUMBER
Reserved for use by each report recipient.
TITLE AND SUBTITLE
"itle should indicate clearly and briefly the subject coverage of the report, and be displayed prominently. Set subtitle, if used, in smaller
pe or otherwise subordinate it to main title. When a report is prepared in more than one volume, repeat the primary title, add volume
Tiber and include subtitle for the specific title.
5. REPORT DATE
Each report shall carry a date indicating at least month and year. Indicate the basis on which it was selected (e.g., date of issue, date of
.:pi. oval, date of preparation, etc.}.
6. PERFORMING ORGANIZATION CODE
Leave blank.
7. AUTHOR(S)
Give name(s) in conventional order (John R. Doe, /. Robert Doe, etc.]. List author's affiliation if it differs from the performing organi
zation.
8. PERFORMING ORGANIZATION REPORT NUMBER
Insert if performing organization wishes to assign this number.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Give name, street, city, state, and ZIP code. List no more than two levels of an organizational hirearchy.
10. PROGRAM ELEMENT NUMBER
Use the program element number under which the report was prepared. Subordinate numbers may be included in parentheses.
11. CONTRACT/GRANT NUMBER
Insert contract or grant number under which report was prepared.
12. SPONSORING AGENCY NAME AND ADDRESS
Include ZIP code.
13. TYPE OF REPORT AND PERIOD COVERED
Indicate interim final, etc., and if applicable, dates covered.
14. SPONSORING AGENCY CODE
Insert appropriate code.
15. SUPPLEMENTARY NOTES
Enter information not included elsewhere but useful, such as: Prepared in cooperation with, Translation of, Presented'at conference of,
To be published in, Supersedes, Supplements, etc.
16. ABSTRACT
Include a brief (200 words or less) factual summary of the most significant information contained in the report. If the report Contains a
significant bibliography or literature survey, mention it here.
17. KEY WORDS AND DOCUMENT ANALYSIS
(a) DESCRIPTORS - Select from the Thesaurus of Engineering and Scientific Terms the proper authorized terms that identify the major
concept of the research and are sufficiently specific and precise to be used as index entries for cataloging.
(b) IDENTIFIERS AND OPEN-ENDED TERMS - Use identifiers for project names, code names, equipment designators, etc. Use open-
ended terms written in descriptor form for those subjects for which no descriptor exists.
(c) COSATI FIELD GROUP - Field and group assignments are to be taken from the 1965 COS ATI Subject Category List. Since the ma-
jority of documents are multidisciplinary in nature, the Primary Field/Group assignment(s) will be specific discipline, area of human
endeavor, or type of physical object. The application(s) will be cross-referenced with secondary Field/Group assignments that will follow
the primary postingls).
18. DISTRIBUTION STATEMENT
Denote releasability to the public or limitation for reasons other than security for example "Release Unlimited." Cite any availability to
the public, with address and price.
19. & 20. SECURITY CLASSIFICATION
DO NOT submit classified reports to the National Technical Information service.
21. NUMBER OF PAGES
Insert the total number of pages, including this one and unnumbered pages, but exclude distribution list, if any.
22. PRICE
Insert the price set by the National Technical Information Service or the Government Printing Office, if known.
EPA Form 2220-1 (Rev. 4-77) (Reverse)
------- |