905R80014
THE RATIO OF PEAK 1-HOUR AVERAGE CONCENTRATIONS TO PEAK CONCENTRATIONS
OF OTHER AVERAGING TIMES FOR VARIOUS POLLUTANTS AND DIFFERING SOURCES
by
Dennis A. Trout
Regional Meteorologist
U. S. Environmental Protection Agency, Region V
Chicago, Illinois USA
Abstract. This study compares maximum observed 1-hour aver-
aged concentrations to the maximum observed concentrations for
averaging times ranging from 5 minutes to 1 year. Pollutants
considered include CO, HC, NO, NOg, NOX, 03, and S02. The
types of sources include rural power plants and numerous urban
complexes.
The ratio Xmax(l-hour)/ xmax(t) aPPears to be represented by
the function at&, where t is averaging time in hours. The
exponent, b, appears to vary between 0.1 and 0.7 depending
primarily on averaging time and source type. Strong sim-
ilarity is shown for pollutants from similar source types.
Introduction
Numerous investigators have compared peak short term concentrations with longer
term average concentrations. Several investigators including Montgomery and
Coleman, and Martin and Reeves have compared peak S02 concentrations of differ-
ing averaging times around power pi ants. ^'.^ Larsen has developed models for
calculating statistically expected pollutant concentrations (including peak
concentrations) for monitoring site locations for which records of monitoring
data are available.3,4,5 The purpose of this paper is to compare the ratios
of peak 1-hour average concentrations to peak concentrations of other averag-
ing times for various pollutants and differing sources.
Peak to peak ratios can provide insight into various areas of current interest.
These ratios can be utilized in the assessment of the relative severity of
ambient air quality standards in an area. Such information is of value, for
example, in estimating whether a short term ambient air quality standard will be
exceeded when only a longer term average concentration is known. Peak to peak
ratios can also aid in assessing the constraining averaging time for purposes of
designing modeling and monitoring investigations.
Analysis Procedure
This paper considers two different general source categories of available data
for analysis. One category is S02 monitoring data from monitoring networks
around power plants. The other category is CO, HC, NO, N02, NOX, 03 and S02
records from single monitoring sites in the vicinity of each of 8 urban areas.
From the available data, the maximum concentration for each reported averaging
time for each pollutant and each monitoring site is selected. Only one maximum
for each reported averaging time is selected from the complete period of record
-------
available from each monitoring site. The periods of record vary depending upon
the particular monitoring site. In general, several years of data are available
from each monitoring site.
Ratios are then developed to compare the peak 1-hour average concentration to
the peak concentrations of each of the other reported averaging times. Reported
averaging times ranged from 5 minutes to 1 year. Ratios are developed for each
pollutant by monitoring site, and averaged by monitoring network for each source
category (power plants and urban areas).
As expected, the longer the averaging time (t), the lower the peak concentration
for that averaging time ( Xmax(t)). Therefore, it is also obvious that the
ratio of the peak 1-hour concentration ( Xmax(l-hour)) to tne peak concentration
for another averaging time ( Xmax(t)) will increase as t increases.
Ratios (Rt) of Xmax(l-hour)/ Xmax(t) can be grouped by averaging time, pollu-
tant, monitoring site, monitoring network, and source category. From these
groupings, the variations in ratios can be studied.
Power Plant Related Observations
TVA Power Plants
Montgomery and Coleman have compiled peak to peak ratios for 8 power plants in
the Tennessee Valley Authority (TVA) system. For the TVA power plants studied,
data were available from the one to sixteen SO? monitors in operation for two to
four years in the vicinity of each TVA power plant. In total, 39 monitors were
in operation for on the average of three years each. The ratios reported for
each power plant represent the average of the ratios obtained from all monitoring
sites in the vicinity of each power plant.
Table 1 presents the ratios reported by Montgomery and Coleman. Also indicated
on Table 1 are the ranges of ratios as well as the geometric means of the ratios.
Table 1. Peak 1-hour concentration to peak 3-hour, 24-hour, 1-month, and
annual concentration ratios for S02 measurements in the vicinity
of 8 TVA power plants.
Power Plant R^ R?a R?30 R8760
Shawnee 1.6 5.8 32 82
Kingston 1.5 6.0 51 120
Johnsonville 1.5 5.8 61 130
Colbert 1.2 5.8 50 110
Allen 1.7 6.8 64 110
Gallatin 1.3 5.1 34 120
Paradise 1.4 6.2 60 120
Bull Run 1.7 8.2 56 190
Geometric Mean 1.5 6.2 50 120
Range:
Low 1.2 5.1 32 82
High 1.7 8.2 64 190
-------
Note, Montgomery and Coleman suggest that the ratios increase as average stack
height increases. The power plants shown on Table 1 are arranged in order of
lowest (Shawnee) to highest (Bull Run) average stack height.
Table 1 indicates that the means for RS and R24 are 1.5 and 6.2 respectively;
thus, R24/R3 = Xmax (3-hour)/ Xmax (24-hour) = t'1' The 3-hour National Ambient
Air Quality Standard (NAAQS) for SO? (1300 ug/m3) divided by the 24-hour NAAQS
for S02 (365 ug/m3) is 3.6. Thus, Based on monitoring in the vicinity of TVA
power plants, it should be expected in general that the 3-hour NAAQS is the more
controlling standard with respect to regulating S02 emissions from the TVA plants,
AEP Power Plants
Martin and Reeves have provided S0Ł data from which maximum ratios can be derived
for several American Electric Power (AEP) power plants in the Ohio River Valley.
The data presented by Martin and Reeves are the observations made at the 33
monitoring sites comprising 5 monitoring networks for various AEP power plants.
Martin and Reeves provided data on peak 1-hour, 3-hour, and 24-hour concentra-
tions observed during 1975 at each of the 5 monitoring networks. Each monitor-
ing network consisted of 4 to 11 monitoring sites.
Table 2 provides information similar to that provided for the TVA power plants
identified in Table 1. Table 2 data represents the averages of ratios from the
individual monitors comprising a monitoring network. The plants shown on Table
2 are arranged in order of lowest (Tanners Creek) to highest (Big Sandy) aver-
age stack height.
Note, from the ratios indicated for the AEP power plant monitoring networks, it
does not readily appear that ratios increase with increasing stack height as
suggested by Montgomery and Coleman.
Table 2 indicates that the means for R3 and R24 are 1.3 and 3.5 respectively.
R24/R3 = 2-7» thus, based on existing monitoring data in the vicinity of the AEP
plants it should be expected that, in general, the 24-hour NAAQS is the more
controlling standard with respect to regulating S02 emissions from the AEP plants.
Table 2. Peak 1-hour concentration to peak 3-hour and 24-hour concentration
ratios for S02 measurements in the vicinity of several AEP power
plants in the Ohio River Valley.
Power Plant R$
Tanners Creek 1.3 3.3
Cardinal-Tidd 1.2 2.9
Kyger-Gavin-Sporn 1.3 4.0
Clifty Creek 1.5 4.2
Big Sandy 1.3 3.5
Geometric Mean 1.3 3.5
Range:
Low 1.2 2.9
High 1.5 4.2
3
-------
Obviously, there are several major differences to be observed between the data
presented for the TVA and AEP power plants. Such differences include the magni-
tude of the ratios, and the NAAQS for SOz which appears to be more constraining.
Ohio River Valley Power Plants
Because only 3-hour and 24-hour maximums for only one year of data were available
for the AEP power plants studied by Martin and Reeves, as well as the presently
unexplained differences noted above, further data were collected from power plant
monitoring networks in the same general area. The data collected consists of
four years of data from monitors located in the vicinity of power plants in the
Ohio River Valley.
Data were obtained from 6 power plant associated monitoring networks (identified
as monitoring networks A-F) in the Ohio River Valley (ORV). The data available
represented the period 1974 thru 1977 and came from 35 ORV monitoring sites.
There were 4 to 7 monitoring sites in each of the 6 ORV monitoring networks.
Table 3 presents categories of information for the ORV power plants similar to
that provided for the TVA plants in Table 1 and the AEP plants in Table 2. The
monitoring networks shown in Table 3 are arranged in order of lowest (A) to
highest (F) average power plant stack height in the vicinity of the monitoring
network.
Note, based on the ratios indicated for the ORV monitoring networks, it is not
apparent that ratios increase as average power plant stack height increases as
suggested by Montgomery and Coleman.
Table 3 indicates that the means for R3 and R24 are 1.4 and 4.5 respectively.
R24/R3 =3.2, thus based on the existing monitoring data in the vicinity of the
ORV power plants it should be generally expected, as was also indicated for the
Martin and Reeves data, that the 24-hour NAAQS is the more controlling standard
with respect to regulating S02 emissions from the ORV power plants analyzed here.
Table 3. Peak 1-hour concentration to peak 3-hour, 24-hour, and annual
concentration ratios for S02 measurements from 6 monitoring
networks located in the vicinity of power plants in the Ohio
River Valley.
Monitoring Network PŁ R24 R8760
A 1.5 5.0 24
B 1.3 4.3 28
C 1.3 3.8 18
D 1.4 4.7 21
E 1.4 6.0 32
F 1.4 3.6 17
Geometric Mean 1.4 4.5 23
Range:
Low 1.3 3.6 17
High 1.5 6.0 32
-------
Differences Between TVA and. ORV Ratios
It can be observed in comparison of the data presented in Tables 1,2, and 3
that monitoring in the vicinity of power plants in the Ohio River Valley does
not appear to yield ratios as high as those indicated based on monitoring in
the vicinity of TVA power plants.
Several different factors may be responsible for the differences in ratios seen
from plant to plant and more sharply between the TVA and the AEP and ORV power
plants. Montgomery and Coleman suggest that the variation among TVA plants
can be attributed to the differences in average stack height for the plants.
It was not found that average stack height alone could account for the differ-
ences among the non-TVA plants studied. Other factors influencing observed
differences in ratios may include: 1) differences in power plant design and
operation, 2) meteorological differences which result in differences in maximum
ground level concentrations, 3) terrain differences which affect the transport,
dispersion, and impacts of the power plant plumes, 4) differences in regional
background concentrations of pollutants which may make-up a larger portion of
the concentrations observed for the Ohio River Valley located monitors, and 5)
differences in the location of monitors with respect to the areas of impact due
to power plant emissions (eg. it may be possible that monitors located in the
vicinity of TVA power plants are located closer to areas of maximum short-term
(3-hour) impact than are the AEP power plant monitors).
Urban Area Related Observations
Larsen has compiled data for 8 urban areas. These data include 5-minute, 1-
hour, 8-hour, 24-hour, 1-month, and 1-year observations of CO, HC, NO, N02,
NOX, 03, and S02 from a monitoring site in each of the 8 urban areas. The
monitoring periods from which these data were compiled range from 2 to 8 years
for each pollutant in each city.
The available data were analyzed by each pollutant as monitored in each city
(tables not shown), and by the average of all pollutant ratios for each city,
(see table 4) and for each pollutant by the average from all cities (see table
5). Ratio variation was greater from pollutant to pollutant within the same
city than was the variation in ratios for a single pollutant compared to the
same pollutant's ratio in each other city.
It can be seen, by comparing tables 4 and 5, that there is less variation about
the mean for ratios from city to city than for pollutant to pollutant. Pollu-
tants with the lowest ratios in urban areas are CO and HC which presumably are
largely a result of near ground level automotive emissions over broad areas.
Table 5 indicates that the geometric mean of RS (R8 = X raax (i-hourl/X max (8-
hour)) for CO is 1.6. The 1-hour NAAQS for CO is (40 mg/ m3) divided by the
8-hour NAAQS for CO (10 mg/m3) is 4.0. Thus, based on monitoring in the vicinity
of the 8 cities analyzed, it is expected in general that the 8-hour NAAQS for CO
is the more controlling standard with respect to regulating CO emissions in urban
areas. Note, all of the cities analyzed in this study have RS for CO < 2.6;
therefore, the 8-hour NAAQS for CO appears to be the controlling standard for all
cities analyzed here.
-------
Table 4. Peak 1-hour concentration to peak 5-minute, 8-hour, 24-hour, 1-month,
and annual concentration ratios for 8 cities based on the average of
ratios of 7 pollutants measured in each city.
City Ro.na
Chicago 0.73
Cincinnati 0.69
Denver 0.68
Los Angeles 0.52
Philadelphia 0.84
St. Louis 0.69
San Francisco 0.79
Washington, D.C. 0.81
Geometric Mean 0.71
Range:
Low 0.52
High 0.84
.7
.7
.0
.8
1.6
2.0
2.0
1.7
1.8
1.6
2.0
R24
2.3
2.7
3.1
2.5
2.5
2.5
2.9
2.5
2.6
2.3
3.1
R730
4.4
7.4
6.7
5.5
7.0
6.6
6.2
6.6
6.2
4.4
7.4
R8760
6.5
12.3
10.8
10.0
12.2
9.0
9.2
9.7
9.8
6.5
12.3
Table 5.
Peak 1-hour concentration to peak 5-minute, 8-hour, 24-hour, 1-
month, and annual concentration ratios for 7 pollutants based on
the average of pollutant ratios from measurements in 8 cities.
Pollutant
CO
HC
NO
N02
NOX
03
S02
Geometric Mean
Range:
Low
High
RQ.Q8
0.72
0.70
0.87
0.68
0.90
0.62
0.55
0.71
0.55
0.90
1.6
1.5
2.1
1.9
1.9
1.7
2.0
1.8
1.5
2.1
R24
2.1
2.1
3.1
2.8
2.7
2.8
3.0
2.6
2.1
3.1
R730
4.0
4.0
9.3
6.4
7.6
6.1
8.2
6.2
4.0
9.3
R876Q
5.9
5.9
19.2
8.1
12.2
10.0
13.2
9.8
5.9
19.2
O'Donnell has reported proposed 1-hour standards for N02.6 He indicates that
the World Health Organization, (WHO), U.S. Environmental Protection Agency
(EPA), and Ford Motor Company have suggested 1-hour N02 standards in the ranges
of 0.10-0.17 ppm, 0.25-0.50 ppm, and 0.50-0.75 ppm respectively. The present
annual NAAQS for N02 is 0.05 ppm. Thus, ratios of proposed 1-hour N02 standards
to the existing annual NAAQS for N02 have ranges of 2.0-3.4, 5.0-10.0, and 10.0-
15.0 for WHO, U.S. EPA, and Ford Motor Co. suggested standards respectively.
From Table 5 it can be seen that R876Q ^or ^2 ""s 8.1. Therefore, based on the
monitoring data analyzed here, it could be expected that if the established 1-
hour standard for NO? is less than 0.4 ppm (8.1 X 0.05 ppm), in general the 1-
hour standard would become the controlling standard with respect to regulating
N02 emissions in urban areas. Note, if the established 1-hour standard for N02
is less than 0.27 ppm, the 1-hour standard would become the controlling standard
for each of the 8 cities for which monitoring data were analyzed (St. Louis had
the lowest R8760 = 5-5i 5'5 x °-05 PPm = 0.275 ppm). Similarly, if the estab-
-------
lished 1-hour standard for N02 is greater than 0.70 ppm, the annual NAAQS would
remain the controlling standard for each of the 8 cities for which monitoring
data were analyzed (Cincinnati had the highest R8760 = 14.0; 14.0 X 0.05 ppm =
0.70 ppm).
Discussion
One method of summarizing the data which has been tabulated is to plot the
ratios as a function of averaging time on log-log scale graph paper. Figure 1
graphically summarizes the geometric means of peak to peak ratios for the TVA,
ORV, and urban areas, which were tabulated in Tables 1,3, and 5 respectively.
The main features to be observed from the comparison information provided in
Figure 1 is that ratios and slopes of curves displayed are greatest for the TVA
data and lowest for the urban area data. The most probable cause for such
differences are due to differences in the distribution of concentrations to be
monitored and differences in the location of monitors to measure pollutant
concentrations.
Differences between the TVA and ORV data and potential contributing factors have
been previously presented. Several of these same potential contributing factors
would be applicable for suggested explanation of the difference between the
ratios indicated for power plant monitoring networks and the ratios indicated
for urban area monitors. Differences in the distribution of concentrations to
be monitored are in part due to differences in interaction of other contributing
sources and background. The greater the number of sources contributing to moni-
tored pollution levels or the higher the background pollution levels, the lower
the peak to peak ratios. Differences in the location of monitors also would
result in differences in observed peak to peak ratios. The further removed from
the areas of maximum short term concentrations that monitoring sites are located,
the lower the expected peak to peak ratios.
The curves shown in Figure 1 can be approximated by relationships of the form:
Rt = atb
where a is a constant, t is the averaging time, and b is the slope of the curve
(line between the ratios for two different averaging times).
Based on analyses of the data, the slopes appear to vary between 0.1 and 0.7
depending on source category and averaging period. It is also obvious that the
greater the peak to peak ratios, the greater the average slope over all averag-
ing periods. Not shown or obvious from Figure 1 is the significant variation
about the means of ratios (slopes) which combine to provide the averages dis-
played.
Also not shown, by Figure 1, but can be discerned from analyses of Tables 1-5
and the data upon which these tables are based, is the similarity of ratios
(and slopes) for similar source categories and pollutant species.
In comparing the curves shown on Figure 1 for the various source categories
indicated, the curve shown for the urban areas is the smoothest of the three.
This is as expected, as Larsen has previously shown monitored concentrations
in urban areas to be log normally distributed as a function of averaging time.
7
-------
o
VD
1^
00
o
CO
r
s-
O) +->
-C C
4J OJ
O S-
OJ
$- t»-
OM-
*f- !-
TD
to
c s-
o o
(0-
$-»->
0-M
O
O)
-^ c
(O-i-
O) CD
CL(O
O O)
t-> >
rt)
S-
3
O
ro O
+J4J
CO
^->+-> O)
4-5 4->
CC CD IB
-c o
r-
1/1 CD OJ
O 10 O
r- S- S-
»-> O) 3
«S > O
a: ro to
2!
3
cr>
o
o
o
o
o
8
-------
Some of the irregularity (changing slopes) of the curves shown for power plant
source categories in Figure 1 may be due to the limited spatial and temporal
resolution of the monitoring data.
It is currently planned that further investigation will focus on the differences
observed between the power plant data analyzed in this paper. Meteorological
data, power plant parameters, source emission and background air quality data,
and monitoring location effects should be investigated. Further analyses should
also compare ratios developed based on model calculated maximum concentrations.
Such analyses may provide further insight as to the proper development of both
air quality monitoring networks and the development and validation of air qual-
ity predictive models.
Summary and Recommendations
Peak 1-hour averaged concentrations are compared to peak concentrations for
averaging times ranging from 5-minutes to 1-year. Pollutants considered include
CO, HC, NO, N02, NOX, 03, and S02. Source categories consisted of various rural
power plants and numerous urban complexes.
The ratio, Rt (Rt = ^axO-hour)/ Xmax(t))» can ^e represented by the function
Rt = atb, where t is the averaging time in hours. The exponent, b, appears to
vary between 0.1 and 0.7 depending primarily on averaging time and source type.
Strong similarity is shown for pollutants from similar source types.
It is expected that the causes for differences in peak to peak ratios among
similar pollutants from similar sources include: differences in the distribution
of concentrations and differences in the location of monitors to measure the
pollutant concentrations. It is suggested that the differences in the distri-
bution of concentrations to be monitored are attributable to differences in
emission source characteristics (including height of release, distribution of
points or areas of emission, and modes of source operation), meteorology, terrain,
and background air quality.
It is recommended that the causes of the observed ratio differences be the focus
of future investigation. It is also recommended that future investigations
should also seek to compare peak to peak ratios developed based on model cal-
culated maximum concentrations. Such further analyses may provide additional
guidance as to the proper development of air quality monitoring networks and the
development of validation tools for air quality predictive models.
Acknowledgements
The author wishes to thank Dr. Ralph Larsen, Dr. Lowell Smith and Dr. Michael
Mills for data provided for this investigation, as well as for discussions
concerning the monitoring networks from which their data were developed.
References
1. T. L. Montgomery and J. H. Coleman, "Empirical Relationships Between
Time-Averaged SO? Concentrations," Environmental Science and Technology.
Vol. 9, No. 10: 953-957, October
o
-------
J. R. Martin and R. W. Reeves, "Relationships Among Observed Short-Term
Maximum Sulfur Dioxide Concentrations Near Coal-Fired Power Plants,"
Proceedings of the 70th Annual Meeting of the Air Pollution Control Associ-
ation, Session 77-29.5, Toronto, June 1977.
R. I. Larsen, "A New Mathematical Model of Air Pollutant Concentration
Averaging Time and Frequency," APCA Journal, Vol. 19, No. 1: 24-30,
January 1969.
R. I. Larsen, " A Mathematical Model for Relating Air Quality Measurements
to Air Quality Standards," U.S. EPA, Office of Air Programs, AP-89,
February 1973.
R. I. Larsen, "An Air Quality Data Analysis System for Interrelating
Effects Standards, and Needed Source Reductions: Part 4. A Three Parameter
Averaging-Time Model," APCA Journal; Vol. 27, No. 5: 454-459, May 1977.
F. J O'Donnell; "Washington Report," APCA Journal; Vol. 28, No. 6: 566,
June 1978.
10
-------
Captions
Tables
Table 1.
Table 2.
Table 3.
Table 4.
Table 5.
Figures
Figure 1.
Peak 1-hour concentration to peak 3-hour, 24-hour, 1-month, and
annual concentration ratios for S02 measurements in the vicinity
of 8 TVA power plants.
Peak 1-hour concentration to peak 3-hour and 24-hour concentration
ratios for S02 measurements in the vicinity of several AEP power
plants in the Ohio River Valley.
Peak 1-hour concentration to peak 3-hour, 24-hour, and annual
concentration ratios for 502 measurements from 6 monitoring
networks located in the vicinity of power plants in the Ohio River
Valley.
Peak 1-hour concentration to peak 5-minute, 8-hour, 24-hour, 1-month,
and annual concentration ratios for 8 cities based on the average of
ratios of 7 pollutants measured in each city.
Peak 1-hour concentration to peak 5-minute, 8-hour, 24-hour, 1-month,
and annual concentration ratios for 7 pollutants based on the average
of pollutant ratios from measurements in 8 cities.
Ratios (R^) 9f peak 1-hour concentrations to peak concentrations for
other averaging times plotted as a function of averaging time (t) for
different source categories.
------- |