United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park, NC 27711
EPA-454/R-93-052
December 1993
AIR
& EPA
COMPARISON OF DESIGN CONCENTRATIONS BASED ON
HOURLY MIXING HEIGHTS ESTIMATED BY RAMMET AND METPRO
-------
EPA-454/R-93-052
COMPARISON OF DESIGN CONCENTRATIONS BASED ON
HOURLY MIXING HEIGHTS ESTIMATED BY RAMMET AND METPRO
U. S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air Quality Planning and Standards
Technical Support Division
Research Triangle Park, NC 27711
December 1993
7' tfesi Jackson Boulevard l?fh c,
Chicago, IL 60604-3590 loor
-------
ACKNOWLEDGEMENTS
This report was prepared by Mr. James O.
Paumier and Mr. Roger W. Erode of Pacific
Environmental Services, Inc., Research Triangle
Park, North Carolina. This effort has been
funded by the U.S. Environmental Protection
Agency under Contract No. 68D00124, with Mr. C.
Thomas Coulter as the Work Assignment Manager.
Special thanks go to Mr. John S. Irwin for
helpful suggestions and Mr. Gerald Moss for
assisting in retrieving the upper air sounding
data used in this study.
DISCLAIMER
This report has been reviewed by the Office of
Air Quality Planning and Standards, EPA, and
approved for publication. Mention of trade
names or commercial products is not intended to
constitute endorsement or recommendation for
use.
11
-------
PREFACE
In this report a comparison is made of the
effects of utilizing two different mixing
height algorithms, RAMMET and METPRO, in
regulatory applications of Gaussian dispersion
models. The comparisons are for rural
environments only; a future effort is planned
to compare these algorithms to treat urban
effects.
The Environmental Protection Agency must
conduct a formal and public review before the
Agency can recommend replacement of the RAMMET
mixing height algorithm with the METPRO mixing
height algorithm. This report is being
released to establish a basis for review of the
consequences resulting from use of METPRO-
derived rural mixing height estimates in
routine dispersion modeling of air pollution
impacts. As implied above, this report is one
of two that must be considered before any
formal changes can be considered.
ill
-------
CONTENTS
Preface ii
Acknowledgements iii
Contents iv
Figures v
Tables vi
1. INTRODUCTION 1
2. METEOROLOGICAL DATA BASES AND PREPROCESSOR OPTIONS 4
3. MODEL OPTIONS 7
3.1 Modeling options 7
3.2 Source characteristics ..... 8
3.3 Receptor configuration 8
4. PERFORMANCE MEASURES 10
5. MODELING RESULTS AND DISCUSSION 11
6. CONCLUDING REMARKS 23
7. REFERENCES 24
iv
-------
FIGURES
Figure Page
5.1 Annual average mixing height by hour of day for
Pittsburgh, 1985 20
5.2 Same as Figure 5.1, except results are for Oklahoma
City, 1985 21
5.3 Same as Figure 5.1, except results are for
Brownsville, 1985 21
-------
TABLES
Table Page
2.1 Missing soundings and calm wind conditions by site
and year 5
3.1 Source characteristics 8
3.2 Terrain elevation used in modeling for the 200m stack. . 9
5.1 The estimated H1H and H2H by year, source and
averaging time for Pittsburgh, PA. XR is the
concentration estimate using the rural mixing
height estimates from RAMMET; XM is tne
concentration estimate using mixing heights from
METPRO 12
5.2 Same as Table 5.1, except results are for Oklahoma
City, OK 13
5.3 Same as Table 5.1, except results are for
Brownsville, TX 14
5.4 Ratios of the H2H concentrations for all sites,
years, sources and averaging times (concentration
values are neither paired in time nor space) 17
5.5 H2H concentrations using RAMMET mixing heights and
the analogous concentration using METPRO mixing
heights (concentrations are paired in time and
space for the indicated hours). Note that he is
the effective plume height and z; is the mixing
height 18
5.6 H2H concentrations using METPRO mixing heights and
the analogous concentration using RAMMET mixing
heights (concentrations are paired in time and
space for the indicated hours). Note that the
columns for RAMMET and METPRO are reversed from
those in Table 5.5; the ratios remain XM/XR 19
VI
-------
1. INTRODUCTION
One of the input values required by many of the Gaussian
dispersion models currently recommended by the Guideline on Air
Quality Models (Revised)1 (EPA, 1986) is the hourly mixing height,
the vertical section of the lower atmosphere with intense vertical
mixing. Historically, these values have been provided by a
meteorological preprocessor (such as RAMMET (EPA, 1977), PCRAMMET
or MPRM2 (Irwin et al., 1990)) which linearly interpolates twice-
daily values to each hour of the day. The twice-daily values are
obtained from rawinsonde observations at National Weather Service
(NWS) stations that routinely collect such data. The soundings,
in combination with NWS surface observations of dry bulb
temperature, are processed according to the development by
Holzworth (1972) to obtain the twice-daily mixing heights.
With these twice-daily estimates, RAMMET occasionally
computes unrealistically low mixing heights for rural sites. When
the preprocessor encounters stable conditions for the hour just
before sunrise, it sets the mixing height to zero at sunrise.
Hourly mixing heights then are computed by interpolating from the
surface up to the maximum afternoon mixing height. This sometimes
results in hourly mixing heights of a meter or less just after
sunrise. If these mixing heights are input to a dispersion model
to estimate concentrations from a ground-based source (e.g., area
sources), the dispersion model may yield unrealistically high
concentrations because the emissions are trapped in a very shallow
layer.
In METPRO, the meteorological preprocessor for the Complex
Terrain Dispersion Model (CTDMPLUS), hourly estimates of heat
(daytime only) and momentum (both day and night) flux are utilized
to compute the hourly mixing heights. During the daytime METPRO
uses an energy budget approach to estimate the hourly surface heat
flux. Net or total incoming radiation, either measured directly
'Hereinafter referred to as the "Guideline".
2Mixing height estimation in RAMMET, PCRAMMET and MPRM are functionally
equivalent. The term RAMMET will be applied as a generic term to include all
three preprocessors.
-------
or estimated from other parameters, is used to obtain the sensible
heat flux at the surface for the hour. The surface friction
velocity is determined through an iterative procedure that
requires only one level of wind speed. The mixing height
algorithm, originally developed by Carson (1973) and modified by
Weil and Brower (1983) (hereafter referred to as the CWB model),
uses the morning sounding and hourly integrated values of the
surface heat flux to estimate a mixing height due to convective
processes. The sounding and integrated values of the friction
velocity are used to produce a mixing height estimate due to
mechanical (shear) processes. The larger of the two estimates is
chosen for the hour's mixing height. For near-neutral conditions
during the daytime, a third estimate of the mixing height is
computed. In this case, the largest of the three estimates is
selected as the hour's mixing height. Paumier and Irwin (1991)
noted, however, that the near-neutral estimate often exceeded the
convective and mechanical estimates by an order of magnitude or
more, particularly during the morning transition period (the hours
shortly after sunrise). They found that the near-neutral estimate
was selected about 15% of the time (i.e., of all daytime hours).
When selected, this estimate frequently caused abrupt and large
increases and decreases from the gradual rise of the mixing
height.
At night, the friction velocity is estimated from cloud cover
and wind speed by a method developed by Venkatram (1980). This,
with an estimate of the Monin-Obukhov length, is used to estimate
the mixing height from a scheme developed by Zilitinkevich (1972)
and extended by Nieuwstadt (1981).
The performance of the CWB model versus the interpolation
approach was investigated by Paumier and Irwin (1991). Mixing
heights from the two methods were compared to observed mixing
heights. Both models performed comparably in the late afternoon,
but the CWB model (without the near neutral estimate) performed
better in the early morning.
The purpose of this study is to use the mixing heights
estimated by both RAMMET and METPRO in a Guideline dispersion
model and compare the resulting pollutant concentrations across a
-------
range of source characteristics and meteorological data bases.
The results of this comparison will provide an indication of the
effect the CWB mixing height algorithm would have on design
concentrations for regulatory modeling applications. The data
bases used in this comparison and the preprocessor procedures are
discussed in the next section. The modeling options are discussed
in Section 3, performance measures used in the analysis are
explained in Section 4, the results and discussion are presented
in Section 5, with concluding remarks in Section 6. References
are provided in Section 7.
-------
2. METEOROLOGICAL DATA BASES AND PREPROCESSOR OPTIONS
RAMMET requires two input data files - NWS hourly surface
observations and twice-daily mixing heights processed by the
National Climatic Data Center. METPRO requires the NWS surface
observations and 1200 GMT3 upper air sounding. Additionally
METPRO requires a profile of winds and temperature from the site
for which the data are being processed.
Two years of data for three sites were retrieved and
processed. The three sites selected for this study are
Pittsburgh, PA (WBAN station #94823), representative of an urban
east coast site; Oklahoma City, OK (WBAN station #13967)/
representative of a southern plains site; and Brownsville, TX
(WBAN station #12919), representative of a coastal environment.
Data for 1984 and 1985 were used at Pittsburgh and Oklahoma City;
for Brownsville, 1985 and 1986 data were used. The twice-daily
mixing heights and NWS surface observations were retrieved from
EPA's Office of Air Quality Planning and Standards Technology
Transfer Network electronic bulletin board. Prior to being made
available on the bulletin board, the data are checked for blank
fields (missing data) and filled by accepted procedures. No
additional modifications to the data were made after retrieving
them from the bulletin board.
For METPRO to work properly, there must be a 1200 GMT
sounding each day. The upper air data obtained for this study did
not always meet this condition, so METPRO was modified to skip the
daytime mixing height estimation whenever there was no 1200 GMT
sounding for an individual day. The sounding files were modified
by inserting one record at the point of a missing sounding to
indicate to METPRO that there were no data to process. Table 2.1
summarizes the number of missing soundings by site and by year.
METPRO also requires hourly profiles of wind and temperature
from the site being modeled. There were no on-site data, so a
1-level profile was constructed from the NWS surface wind and
temperature observations, with the height of the observations
'Greenwich Mean Time
-------
Table 2.1 Missing soundings and calm wind conditions by site and
year.
Site Year
Pittsburgh 1984
Pittsburgh 1985
Oklahoma City 1984
Oklahoma City 1985
Brownsville 1985
Brownsville 1986
Hours /year
8784
8760
8784
8760
8760
8760
Missing Soundings
0000 GMT 1200 GMT
0 0
2 2
0 0
0 1
0 0
0 6
Calm wind
conditions
858
1027
181
245
381
566
(i.e., anemometer height) set to 10m. At night, METPRO will not
compute a mixing height for a wind speed of 0 ms"1 (calm wind) . By
using NWS data to construct the profile, there were several
hundred hours of calm wind conditions for all the sites combined.
The number of hours of calm wind conditions detected in the file
of profile data is shown in Table 2.1.
METPRO requires two additional files. The first is a file of
locally observed solar and net radiation and mixing heights. None
of the parameters in this file are required for METPRO to make its
computations, but they would be used if they were available. With
no locally observed data, a file indicating that there was no
available data was created for METPRO to read. The second file is
much more important and contains information defining the physical
characteristics of the site, e.g., the surface roughness length,
albedo (a measure of the reflectivity at the surface), and Bowen
ratio (a measure of the surface moisture). These parameters are
specified by categories or sectors of wind direction (as defined
by the wind direction in the file of profile data) and by month.
METPRO allows up to eight nonoverlapping sectors covering all
possible directions. Without detailed information about each
site, the parameters for this study are defined for only one
sector (1° - 360°) and are the same for all months at all three
sites. The values used for this analysis are: 0.25m for the
surface roughness length, 0.40 for the albedo, and 0.75 for the
Bowen ratio.
-------
One additional modification was made to METPRO in accordance
with the objectives of this study. As discussed earlier, the
daytime near-neutral estimate may cause abrupt changes in the rise
of the daytime mixing height. Therefore, the daytime near-neutral
estimate of the mixing height was ignored in this study. Thus
only the convective and mechanical estimates from the CWB model
were candidate mixing heights for each hour.
With all the raw data processed and modifications made to
METPRO, both preprocessors were run on the six sets (3 sites X 2
yrs/site) of data. The output file from RAMMET, a binary file,
was converted to ASCII and the METPRO mixing heights merged into
the RAMMET files as a separate field (column). Thus, there were
three mixing heights per hour in each ASCII file - the rural and
urban estimates from RAMMET and the METPRO estimate. If any one
of the three mixing heights was missing for the hour, then all
three were set to missing when the files were merged.
-------
3. MODEL OPTIONS
The dispersion model used in this study was the Industrial
Source Complex Short Term (ISCST2) model. This version of the
model adds the capability to indicate which fields to read (in an
ASCII file) for a particular run, allowing both RAMMET and METPRO
mixing heights to reside in the same file as described above. A
new (draft) area source algorithm capable of computing concen-
tration estimates at receptor locations inside the perimeter of an
area source was used in this study. This differs from the version
currently available to the general modeling community, which sets
concentration estimates to 0.0 for receptors inside the perimeter.
ISCST2 requires an input file that defines the modeling
options, the source (emissions) information, receptor
configuration and file input/output. These options are described
next.
3.1 Modeling options
The options of most interest here control which dispersion
option (rural or urban) to use, the averaging periods, and how to
handle missing input data. For this analysis, only the rural
dispersion option (corresponding to the rural mixing heights from
RAMMET) was used. METPRO only produces one set of mixing heights,
and it is assumed that these are representative of rural sites.
The averaging periods included in this study were the 1-hr,
3-hr and 24-hr short term averages, plus the period (in this case,
the annual) average.
Another option employed in this comparison was for processing
missing data necessitated by missing METPRO mixing heights due to
missing upper air soundings. This processing option provides a
means for computing the 3-hr, 24-hr and period (annual) concen-
trations that are consistent with the regulatory treatment of calm
winds.
-------
3.2 Source characteristics
Four sources were processed for this study: three point
sources and one area source. For the point sources, constant
emissions of 100 gs'1 were assumed. The characteristics for each of
these sources is presented in Table 3.1.
Table 3.1 Source characteristics.
Point sources
Stack
height
(m)
35
100
200*
X,Y location
& base
elevation (m)
0, 0, 0
0, 0, 0
0, 0, 243.84
Emission Exit Stack
rate velocity diameter
(gs-1) (ms-r) (m)
100 11.7 2.4
100 18.8 4.6
100 26.5 5.6
Temperature
(K)
432
416
425
Ground- level area source
Area (m2)
250,000
Length of
side (m)
500
Emission
rate,
(gs-'m-2)
0.0004
*A11 model runs with the 200m stack were made with elevated terrain.
3.3 Receptor configuration
A gridded polar array of receptors was used in this analysis.
There were 36 radials (beginning at 10 degrees from north and
spaced every 10 degrees). The number and distance of the
concentric rings varied according to the source being modeled.
For the 35m and 100m stacks, distances of 400m, 800m, 1200m,
2000m, 4000m, 7000m, and 15000m were used. For the 200m stack,
which is located in elevated terrain, distances of 800m, 2000m ,
4000m, 7000m and 15000m were used. For the area source, receptor
distances were 125m (halfway between the origin and the
perimeter), 250m (touching the perimeter at the midpoint of each
side), 350m (approximately at the perimeter corners), 800m, 1200m,
2000m and 4000m.
As indicated, the 200m stack was modeled with elevated
terrain. The terrain elevations are given in Table 3.2 by radial
and distance.
8
-------
Table 3.2. Terrain elevation used in the modeling for the 200m
stack. All elevations are in feet.
Radial ( ° )
10.00
20.00
30.00
40.00
50.00
60.00
70.00
80.00
90.00
100.00
110.00
120.00
130.00
140.00
150.00
160.00
170.00
180.00
190.00
200.00
210.00
220.00
230.00
240.00
250.00
260.00
270.00
280.00
290.00
300.00
310.00
320.00
330.00
340.00
350.00
360.00
800m
940.00
1040.00
960.00
1020.00
1160.00
1200.00
1180.00
1100.00
1040.00
960.00
780.00
680.00
700.00
740.00
740.00
740.00
700.00
660.00
620.00
610.00
610.00
660.00
860.00
900.00
1020.00
1000.00
940.00
920.00
740.00
610.00
610.00
640.00
640.00
640.00
680.00
780.00
2000m
1100.00
1100.00
1060.00
1040.00
1040.00
1160.00
1200.00
1140.00
1020.00
860.00
1040.00
900.00
780.00
760.00
660.00
800.00
800.00
660.00
620.00
630.00
1020.00
1100.00
1120.00
1160.00
1240.00
1260.00
1100.00
1180.00
1100.00
760.00
840.00
640.00
640.00
640.00
680.00
840.00
Distance
4000m
640.00
720.00
1040.00
1120.00
1040.00
1140.00
1280.00
1220.00
1180.00
1280.00
1280.00
1160.00
880.00
1040.00
1060.00
860.00
1200.00
1200.00
620.00
700.00
900.00
980.00
1040.00
1200.00
1240.00
1240.00
1100.00
1180.00
1220.00
1280.00
1260.00
1220.00
1200.00
1160.00
1060.00
700.00
7000m
1000.00
900.00
1160.00
1160.00
1280.00
1080.00
1320.00
1300.00
1180.00
900.00
1240.00
1100.00
940.00
1260.00
1300.00
1120.00
1140.00
1260.00
880.00
640.00
1200.00
1220.00
1300.00
1240.00
1260.00
1280.00
1280.00
1280.00
1360.00
1320.00
1220.00
1300.00
1100.00
660.00
760.00
680.00
15000m
1000.00
720.00
1120.00
1200.00
1280.00
1220.00
1280.00
1300.00
1140.00
1360.00
1240.00
1180.00
1040.00
1320.00
1260.00
1120.00
1200.00
1280.00
1200.00
1140.00
840.00
1240.00
1300.00
1220.00
1280.00
1280.00
1220.00
1300.00
1260.00
1200.00
1180.00
1000.00
780.00
1060.00
1020.00
900.00
-------
4. PERFORMANCE MEASURES
To compare the effects of mixing heights on concentration,
the high-first-high (H1H) and high-second-high (H2H) are reported
by site, by year, by source and by averaging time. The H2H was
chosen because it is the design concentration on which judgement
is based as to whether an existing or proposed pollutant source
will violate a National Ambient Air Quality Standard. The
Guideline recommends that five years of NWS meteorological data be
used for such an analysis. If less than five years of data are
available, then the Guideline recommends using the H1H
concentration.
To compare the concentration estimates, the ratio of the
concentration using METPRO-generated mixing heights (XM) to the
concentration using RAMMET-generated mixing heights (XR) / XM/XR»
is computed. With this ratio, the concentrations using RAMMET
mixing heights are regarded as the "observed" concentrations
because they are the standard input values to ISCST2.
10
-------
5. MODELING RESULTS AND DISCUSSION
The H1H and H2H are tabulated by site, by year, by source, by
averaging time, and by mixing height method in Tables 5.1 - 5.3.
The ratio XM/XR is also shown, and is averaged for the H1H and H2H
across all sources.
There are several overall impressions from looking at the
tables. First, the concentration estimates from the area source
are orders of magnitude larger than the point source estimates.
Second, there are numerous cases where the concentration estimate
using the two different mixing heights are equal, i.e., the ratio
is identically equal to 1.00. Third, for all but the area source,
the concentration estimates using the METPRO-generated mixing
heights are generally larger than the estimates using the RAMMET
mixing heights.
The first point is not too surprising with a ground-level
release area source and a portion of the receptor array within the
perimeter of the source. The H1H and H2H concentrations occurred
at the 250m arc for Pittsburgh and Brownsville. At Oklahoma City,
most of the H1H and H2H estimates were at the 250m arc, with a few
at the 125m and 350m arcs. The maximum period (annual) averages
all occurred at the 125m arc at all the sites for both years. The
most interesting point regarding the area source is the difference
between the 1-hr H1H concentrations using RAMMET and METPRO mixing
heights and, also, the difference between the H1H and H2H for
RAMMET alone, most notably for Pittsburgh 1984, Oklahoma City
1985, and Brownsville 1986. The unusually large H1H concentration
for the RAMMET estimates is due to an extremely small mixing
height at or just after sunrise. In each of the three cases, the
mixing height was less than 1m (only O.lm for the Brownsville
case). The large 3-h H1H for Oklahoma City 1985 and Brownsville
1986 are due to the dominance of the 1-h H1H in the 3-h average.
As stated in the introduction, it is this type of behavior that is
of concern with the RAMMET procedure. When small mixing heights
such as these are used in ISCST2, pollutants are trapped in an
extremely shallow layer, causing an extremely large concentration.
11
-------
Table 5.1 The estimated H1H and H2H by year, source and averaging
time for Pittsburgh, PA. XR is tn® concentration
estimate using the rural mixing height estimates from
RAMMET; XM ig tne concentration estimate using mixing
heights from METPRO.
1984
XR
(/ignr3)
35m stack
l-h HlH
H2H
3-h HlH
H2H
24 -h HlH
H2H
Annual
100m stack
l-h HlH
H2H
3-h HlH
H2H
24 -h HlH
H2H
Annual
200m stack
l-h HlH
H2H
3-h HlH
H2H
24 -h HlH
H2H
Annual
247
238
194
187
58
53
5
81
50
30
25
7
5
0
.20
.28
.70
.72
.03
.64
.00
.06
.98
.85
.51
.65
.96
.396
XM
{/jgm'3)
315
305
194
187
58
52
4
94
61
39
27
7
7
0
.89
.51
.70
.72
.03
.70
.98
.02
.29
.41
.97
.88
.04
.380
Avg.
XM/XR by
Source
1.28
1.28
1.00
1.00
1.00
0.98
1.00
HlH: 1.07
H2H: 1.09
1.16
1.20
1.28
1.10
1.03
1.18
0.96
HlH: 1.11
H2H: 1.16
1985
XR
255
242
199
171
63
56
6
72
54
28
23
6
5
0
.,
.56
.33
.53
.25
.42
.87
.42
.90
.95
.81
.12
.59
.83
.454
XM
378
302
199
171
63
57
6
85
58
34
24
6
5
0
>>
.06
.63
.53
.25
.42
.01
.43
.03
.45
.15
.33
.75
.52
.436
XM/XR
1.48
1.25
1.00
1.00
1.00
1.00
1.00
HlH:
H2H:
1.17
1.06
1.19
1.05
1.03
0.95
0.96
HlH:
H2H:
Avg.
by
Source
1.12
1.08
1.09
1.02
with terrain
34
26
18
17
4
3
0
500m X 500m area
l-h HlH
H2H
3-h HlH
H2H
24-h HlH
H2H
Annual
Average
133769
46544
50539
41696
24299
22485
10638
.86
.59
.43
.65
.57
.63
.317
38
30
18
17
4
3
0
.38
.99
.82
.65
.57
.28
.282
1.10
1.17
1.02
1.00
1.00
0.90
0.89
HlH: 1.00
H2H: 1.02
38
31
15
13
4
4
0
.56
.06
.69
.06
.17
.14
.397
36
30
13
11
4
3
0
.34
.63
.42
.98
.17
.90
.352
0.94
0.99
0.86
0.92
1.00
0.94
0.89
HlH:
H2H:
0.92
0.95
source
.88
.84
.97
.00
.63
.08
.29
46544
46544
44961
41535
24299
22482
10631
.85
.84
.72
.11
.63
.52
.30
0.35
1.00
0.89
1.00
1.00
1.00
1.00
HlH: 0.81
H2H: 1.00
HlH: 1.00
H2H: 1.07
69466
68483
45150
41454
25947
21757
10266
.44
.44
.68
.49
.32
.81
.72
69466
68483
45150
41454
25947
21757
10225
.44
.44
.68
.49
.32
.82
.05
1.00
1.00
1.00
1.00
1.00
1.00
1.00
HlH:
H2H:
H1H:
H2H:
1.00
1.00
1.03
1.01
12
-------
Table 5.2 Same as Table 5.1, except results are for Oklahoma
City, OK.
1984
35m stack
1-h H1H
H2H
3-h H1H
H2H
24 -h H1H
H2H
Annual
100m stack
1-h H1H
H2H
3-h H1H
H2H
24 -h H1H
H2H
Annual
200m stack
1-h H1H
H2H
3-h H1H
H2H
24 -h H1H
H2H
Annual
500m X 500m
1-h H1H
H2H
3-h H1H
H2H
24 -h H1H
H2H
Annual
Average
XR
(MSP
265
255
200
180
87
68
8
85
53
38
27
6
5
0
XM
•3) (/ignr3)
.13
.05
.19
.68
.84
.71
.12
.79
.91
.19
.91
.27
.58
.632
371
342
200
180
87
68
8
95
69
43
29
6
5
0
.21
.16
.19
.68
.84
.71
.05
.71
.42
.95
.46
.35
.25
.559
Avg.
XM/XR by
Source
1.40
1.34
1.00
1.00
1.00
1.00
0.99
H1H: 1.10
H2H: 1.11
1.12
1.29
1.15
1.06
1.01
0.95
0.88
H1H: 1.04
H2H: 1.10
1985
XR
(jigm
276
263
208
175
87
80
8
69
65
35
25
6
5
0
XM
~3) (/igm'3)
.13
.74
.77
.34
.78
.69
.61
.17
.80
.08
.69
.30
.82
.612
324
300
208
175
87
80
8
97
70
35
27
6
5
0
.69
.76
.77
.34
.78
.69
.67
.67
.66
.08
.14
.29
.82
.592
XM/XR
1.18
1.14
1.00
1.00
1.00
1.00
1.01
H1H:
H2H:
1.41
1.07
1.00
1.06
1.00
1.00
0.97
H1H:
H2H:
Avg.
by
Source
1.05
1.05
1.10
1.04
with terrain
34
22
17
14
5
3
0
area
77739
46521
45041
41791
21741
19813
8592
.41
.26
.23
.46
.96
.80
.514
39
28
17
13
5
3
0
.38
.55
.23
.90
.96
.80
.477
1.14
1.28
1.00
0.96
1.00
1.00
0.93
H1H: 1.02
H2H: 1.08
29
21
15
13
5
5
0
.43
.52
.32
.29
.62
.06
.438
34
25
17
16
5
5
0
.44
.22
.71
.02
.62
.22
.395
1.17
1.17
1.16
1.21
1.00
1.03
0.90
H1H:
H2H:
1.06
1.14
source
.06
.00
.38
.86
.98
.41
.98
46544
46521
45041
41791
21742
19909
8588
.84
.00
.38
.86
.17
.74
.95
0.60
1.00
1.00
1.00
1.00
1.00
1.00
H1H: 0.90
H2H: 1.00
H1H: 1.02
H2H: 1.07
499111
56094
166370
40946
23747
20331
8414
.81
.27
.61
.88
.79
.11
.03
46524
46370
40946
38395
20367
18883
8383
.45
.65
.88
.19
.43
.18
.18
0.09
0.83
0.25
0.94
0.86
0.93
1.00
H1H:
H2H:
H1H:
H2H:
0.55
0.90
0.94
1.03
13
-------
Table 5.3 Same as Table 5.1, except the results are for
Brownsville, TX.
35m stack
1-h H1H
H2H
3-h H1H
H2H
24-h H1H
H2H
Annual
100m stack
1-h H1H
H2H
3-h H1H
H2H
24-h H1H
H2H
Annual
200m stack
1-h H1H
H2H
3-h HIM
H2H
24-h H1H
H2H
Annual
500m X 500m
1-h H1H
H2H
3-h H1H
H2H
24-h H1H
H2H
Annual
Average
XR
(Aignr3)
271.29
269.54
226.60
210.33
111.56
78.90
14.58
59.69
49.05
35.34
23.16
8.23
5.73
1.04
1984
XM
(/ignr3)
362.31
323.96
226.60
210.33
111.56
78.90
14.63
77.16
58.18
35.38
23.16
8.23
5.73
1.04
Avg.
XM/XR by
Source
1.34
1.20
1.00
1.00
1.00
1.00
1.00
H1H: 1.09
H2H: 1.07
1.29
1.19
1.00
1.00
1.00
1.00
1.00
H1H: 1.07
H2H: 1.06
1985
XR
(jigm'3)
296.51
270.14
200.46
188.18
85.04
82.35
13.40
79.49
51.36
26.50
23.46
6.63
5.67
0.973
XM
(/zgnf3)
364.30
270.14
200.46
188.18
85.04
82.97
13.40
77.15
52.61
32.83
23.38
6.25
6.12
0.957
XM/XR
1.23
1.00
1.00
1.00
1.00
1.01
1.00
H1H:
H2H:
0.97
1.02
1.24
1.00
0.94
1.08
0.98
H1H:
H2H:
Avg.
by
Source
1.06
1.00
1.03
1.03
with terrain
34.71
28.29
17.78
14.99
6.49
6.29
0.950
38.55
36.80
17.06
14.99
6.33
6.29
0.915
1.11
1.30
0.96
1.00
0.98
1.00
0.96
H1H: 1.00
H2H: 1.10
33.34
25.31
17.12
15.43
5.66
5.33
0.851
39.06
29.62
17.12
15.46
5.66
5.33
0.836
1.17
1.17
1.00
1.00
1.00
1.00
0.98
H1H:
H2H:
1.04
1.06
area source
69466.44
68942.10
59532.11
55198.79
24867.44
22010.57
9490.38
69466.44
68942.10
59532.11
55198.79
24867.44
22010.58
9489.02
1.00
1.00
1.00
1.00
1.00
1.00
1.00
H1H: 1.00
H2H: 1.00
H1H: 1.04
H2H: 1.06
486653.59
69815.01
166256.86
44901.20
23801.00
22065.44
10051.20
69815.01
66808.31
45567.40
44703.58
23801.00
22065.44
10012.65
0.14
0.96
0.27
1.00
1.00
1.00
1.00
H1H:
H2H:
H1H:
H2H:
0.60
0.99
0.93
1.02
14
-------
Very near the surface, small increases in mixing height can result
in much smaller concentration estimates. For example, increasing
the mixing height from O.lm to 2.1m in the Brownsville 1986 case
decreased the maximum concentration for that hour from
486,653 ngm'3 to about 25,000 /ignr3, still quite large but
significantly less. There were no mixing heights less than 3.0m
from METPRO, therefore the concentrations for the ground-level
area source did not show the extreme variations exhibited with the
RAMMET mixing heights.
The second point, that the ratios are identically 1.00 (i.e.,
XR = XM) in many instances, is due in part to the meteorology.
Only the mixing heights differ between ISCST2 runs using RAMMET
output and METPRO output. For the hour, all other meteorology is
identical. Why a particular mix of meteorological conditions
produces the H1H or H2H concentration from an entire year of data
is beyond the scope of this comparison. Rather, the focus is on
the differences between the input data bases, i.e., the mixing
heights and (to some extent) atmospheric stability.
In examining the cases4 where XM = XR/ the estimates occurred
for the same hour or group of hours. In many cases the stability
class was either 5, 6 or 7 (stable atmosphere). ISCST2 assumes
unlimited vertical mixing in stable conditions. Computationally,
the terms for reflections at the mixing height are omitted from
the vertical term of the Gaussian plume equation. Consequently,
mixing heights from both RAMMET and METPRO, no matter how
different, will produce the same concentration. For a more
detailed discussion of the vertical dispersion term in ISCST2, see
Section 1.1.6 in the User's Guide for the Industrial Source
Complex (ISC2) Dispersion Models, Volume II (EPA, 1992).
Another consideration on this point is the effective plume
height (he) relative to the mixing height in unstable atmospheric
conditions. A general rule of thumb (R. W. Erode, Pacific
Environmental Services, Inc., personal communication) is that the
mixing height has little or no affect on the maximum concentration
"Cases where the ratio differs from 1.00 in the fourth or fifth digit to the
right of the decimal were not considered.
15
-------
estimate if the mixing height is greater than twice the effective
plume height. This was particularly evident for the 35m stack
where the effective plume height is only 100m - 200m and the
mixing height from both preprocessors is 400m or more. This
"rule" becomes less clear for the 200m stack, but may be explained
in terms of the effect of elevated terrain on he and the mixing
height. Assume that the terrain is monotonically rising in the
direction of plume transport. With the plume remaining at its
stabilization height, the effective plume height decreases
downwind. However, the mixing height is terrain-following. The
net effect is to increase the difference between he and the mixing
height, possibly to the point where the mixing height has little
or no effect on ground level concentrations. For example, if the
effective plume height in flat terrain is 350m and the mixing
height is 600m above local ground level, there may be significant
reflections from the mixed layer top that enhance the ground level
concentrations. However, for a receptor that is 140m above stack
base, the mixing height is still 600m above local ground level but
the effective plume height is now only 210m and, according to the
"rule of thumb", there are no contributions to ground-level
concentrations by reflections from the mixed layer top. With the
elevated terrain in this study, particularly at the more remote
arcs, the effective plume height can be reduced by 100m or more.
This apparently makes a sufficient adjustment such that the H1H
and H2H 24-hr average concentrations are identical for both
RAMMET- and METPRO-generated mixing heights. Section 1.1.6.2 in
the User's Guide for the Industrial Source Complex (ISC2)
Dispersion Models, Volume II (EPA, 1992) provides a more complete
discussion on the vertical dispersion term in elevated terrain.
The final point is the primary purpose of this study - to
determine what effect, if any, mixing heights estimated by METPRO
have on concentration estimates. The primary focus here is on the
H2H (design) concentrations. Table 5.4 summarizes the ratios of
the H2H estimates listed in Tables 5.1 - 5.3. The design
concentrations using mixing heights from METPRO are generally
greater by 15-30% than those using RAMMET mixing heights. The
16
-------
Table 5.4 Ratios of the H2H concentrations for all sites, years,
sources and averaging times (concentration values are
neither paired in time nor space).
AVG3TcLQi.no
Time Source
1-hr 35m
100m
200m
Area
3-hr 35m
100m
200m
Area
24-hr 35m
100m
200m
Area
Annual 35m
100m
200m
Area
1984
PA OK TX Avg.
1.28 1.34 1.20 1.27
1.20 1.29 1.19 1.23
1.17 1.28 1.30 1.25
1.00 1.00 1.00 1.00
1.00 1.00 1.00 1.00
1.10 1.06 1.00 1.05
1.00 0.96 1.00 0.99
1.00 1.00 1.00 1.00
0.98 1.00 1.00 0.99
1.18 0.95 1.00 1.04
0.90 1.00 1.00 0.97
1.00 1.00 1.00 1.00
1.00 0.99 1.00 1.00
0.96 0.88 1.00 0.95
0.89 0.93 0.96 0.93
1.00 1.00 1.00 1.00
1985
PA OK TX Avg.
1.25 1.14 1.00 1.13
1.06 1.07 1.02 1.05
0.99 1.17 1.17 1.11
1.00 0.83 0.96 0.93
1.00 1.00 1.00 1.00
1.05 1.06 1.00 1.04
0.92 1.21 1.00 1.04
1.00 0.94 1.00 0.98
1.00 1.00 1.01 1.00
0.95 1.00 1.08 1.01
0.94 1.03 1.00 0.99
1.00 0.93 1.00 0.98
1.00 1.01 1.00 1.00
0.96 0.97 0.98 0.97
0.89 0.90 0.98 0.92
1.00 1.00 1.00 1.00
Average :
2 -year
Average
1.20
1.14
1.18
0.97
1.00
1.05
1.02
0.99
1.00
1.03
0.98
0.99
1.00
0.96
0.93
1.00
1.03
difference is particularly notable for point source simulations
with 1-hr averaging times. The differences between the two sets
of estimates decrease with averaging period, but are essentially
identical for averaging times greater than 1-hr. When all sites,
years, sources and averaging times are pooled, XM/XR = 1.03.
To gain some insight into why the concentration estimates
using METPRO mixing heights are higher than those using RAMMET
mixing heights, ISCST2 was rerun for hours producing the H2H
values in the initial analyses. For the hours that produced a H2H
1-hr concentration using a RAMMET mixing height (Tables 5.1 -
5.3), the same hours were rerun using the METPRO mixing heights,
pairing in time and space. The results of this combination are
shown in Table 5.5. Similarly, for the hours that produced a H2H
concentration using a METPRO mixing height (Tables 5.1 - 5.3), the
17
-------
Table 5.5 H2H concentration using RAMMET mixing heights and the
analogous concentration using METPRO mixing heights
(concentrations are paired in time and space for the
indicated hours) . Note that he is the effective plume
height and z{ is the mixing height.
RAMMET
Site & year
Pittsburgh,
1984
Pittsburgh,
1985
Oklahoma City,
1984
Oklahoma City,
1985
Brownsville,
1985
Brownsville,
1986
Source
200m
100m
35m
200m
100m
35m
200m
100m
35m
200m
100m
35m
200m
100m
35m
200m
100m
35m
Julian
day
/hour
167/14
211/12
163/12
161/12
211/12
169/11
203/12
210/13
250/13
209/13
123/12
221/14
163/13
237/12
247/11
242/11
164/11
182/14
(m)
860
476
115
745
715
78
269
465
68
250
483
75
961
548
71
741
461
68
(m)
1052
1649
1195
970
1160
1445
1833
1981
1709
1063
718
1341
1097
1321
929
1197
834
1629
XR
(/zgm'3)
26
51
238
31
54
242
22
53
255
21
65
263
28
49
269
25
51
270
.6
.0
.3
.1
.9
.3
.3
.9
.0
.5
.8
.7
.3
.1
.5
.3
.4
.1
METPRO
(m)
830
1183
1312
829
1460
1472
844
1512
1198
724
672
1041
913
942
1138
776
939
1459
XM
(/igm~
0.
51.
238.
36.
49.
242.
22.
53.
255.
21.
70.
263.
34.
58.
269.
39.
51.
270.
.,
0
0
3
3
8
3
3
9
0
5
7
7
0
2
5
1
3
1
Average :
Ratio
0
1
1
1
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
.00
.00
.00
.17
.91
.00
.00
.00
.00
.00
.07
.00
.20
.19
.00
.55
.00
.00
.01
same hours were rerun using the RAMMET mixing heights, pairing in
time and space. The results are shown in Table 5.6.
As in the ranked comparisons listed in tables 5.1 - 5.3 (in
which concentration estimates are unpaired in time and space),
Tables 5.5 and 5.6 illustrate that the METPRO mixing heights
tended to produce higher concentrations overall by 20%. For the
36 cases in the two tables, METPRO and RAMMET produced identical
concentrations for 12 of the cases. Seven of the 12 cases were
for the 35m stack where the mixing height was often much higher
18
-------
Table 5.6 H2H concentration using METPRO mixing heights and the
analogous concentration using RAMMET mixing heights
(concentrations are paired in time and space for the
indicated hours). Note that the columns for RAMMET and
METPRO are reversed from those in Table 5.5; the ratios
remain XM/XR-
METPRO
Site & year
Pittsburgh,
1984
Pittsburgh,
1985
Oklahoma City,
1984
Oklahoma City,
1985
Brownsville,
1985
Brownsville,
1986
Source
200m
100m
35m
200m
100m
35m
200m
100m
35m
200m
100m
35m
200m
100m
35m
200m
100m
35m
Julian
day
/hour
212/13
164/12
252/09
211/12
161/12
246/07
187/14
175/13
115/08
144/13
123/11
299/09
237/12
237/12
198/08
164/11
131/11
182/14
(m)
780
467
157
1032
477
126
640
541
117
694
483
116
875
548
122
741
552
68
(m)
1153
759
158
1460
829
136
1059
735
124
1183
672
124
942
942
132
939
708
1459
XM
(/jgm-3)
31
61
305
30
58
302
28
69
342
25
70
300
36
58
324
29
52
270
.0
.3
.5
.6
.5
.6
.6
.4
.2
.2
.7
.8
.8
.2
.0
.6
.6
.1
RAMMET
(m)
1591
1328
505
1160
970
601
1943
1033
1119
2180
718
1244
1321
1321
349
834
689
1629
XR
22
44
155
38
51
184
15
49
195
14
65
174
26
49
197
33
54
270
")
.5
.3
.0
.6
.1
.1
.7
.5
.8
.0
.8
.5
.2
.1
.7
.3
.0
.1
Average :
Ratio
1
1
1
0
1
1
1
1
1
1
1
1
1
1
1
0
0
1
1
.38
.38
.97
.79
.14
.64
.82
.40
.75
.80
.07
.72
.40
.19
.64
.89
.97
.00
.39
than the effective plume height, he (recall the discussion earlier
on the relation of he to the mixing height) . In four of the 36
cases, the RAMMET mixing height produced a higher concentration.
In one of the cases, the METPRO mixing height was below he,
yielding a concentration of 0.0. In the remaining 19 cases, the
concentration using a METPRO mixing height was greater than the
concentration using the RAMMET mixing height. For the hours from
7 to 9 (the morning transition period), the concentration
19
-------
difference was as much as a factor of 2 for the 35m stack. For
those morning hours, the METPRO mixing height was about 150m and
the RAMMET mixing height was 3-10 times higher. This behavior is
not necessarily unique to these few cases.
The influence of mixing height estimates on the occurrence of
particularly high concentration ratios, and the long term
distribution pattern of such estimates, was of interest in
assessing the consequence of using the CWB model. Figures 5.1 -
5.3 show the annual average mixing height by hour of day for 1985
from each of the three sites. Each data point represents an
average of 365 values. The mixing heights from RAMMET were, on
average, about 250m - 350m higher than the annual average METPRO
mixing heights for the hours corresponding to the H2H values in
Tables 5.5 and 5.6. By late afternoon (hour 16 and later), the
differences are much less. This tendency was observed and
reported by Paumier and Irwin (1991). In their analysis they
noticed that, until the maximum mixing height was observed, the
mixing heights estimated by RAMMET significantly overpredicted the
observed values, while those estimated by METPRO showed a modest
underprediction. After the maximum mixing height was observed,
estimates from either method showed little bias.
M i -« i ncj n» i Q nt. C r
i 5 O O
1 DOD--
7 f. D - -
2 5 0 - -
-t 1 1 h
METPRO
RAMMET
1 S 3 -4 5 B ~7 8 9 ID 11 12 13 14 15 1E 17 IB 192D 21 2223
Hour-
Figure 5.1.
Annual average mixing height by hour of day for
Pittsburgh, 1985.
20
-------
Mixing neiQnt C
1500-
1250--
1 O O O - -
7 5 O - -
500 --
250 --
H 1 1 h
H (-
1 1 r-
METPRO
RAMMET
1 2 3 -4 5 6 7 B 9 -10111213141516-1718-192021222324
Hour-
Figure 5.2. Same as Figure 5.1, except results are for Oklahoma
City, 1985.
1 Q hx. C
1 5OO
125O--
1 OOO--
5 o a - -
25O --
METPRO
RAMMET
—I 1 1 i 1 1 1 1 1 1 H
I 2 3 4 5 B 7 S 3 1O-111213141516171B132O2122232-4
Hour-
Figure 5.3. Same as Figure 5.1, except results are for Brownsville,
1985.
The increase in the 1-hr concentration estimates using METPRO
mixing heights can be attributed to the mixing height being much
closer to the effective plume height. As simulated by ISCST2, all
plume material is below the mixing height when hc is less than Zj.
21
-------
Conversely, no plume material is below the mixing height when hc
is greater than z{. Such phenomena are exemplified in Table 5.5.
22
-------
6. CONCLUDING REMARKS
This comparison has shown that the 1-hr H2H design concen-
trations estimated using mixing height estimates from METPRO are
generally larger than the design concentrations using RAMMET
mixing heights for the point sources by an average of 15-20%. For
other averaging periods, the point source results appear to be
independent of mixing height algorithm. Similar conclusions are
reached when the concentrations are paired in time and space. A
notable difference is the H1H 1-hr concentration for an area
source. The potential for unusually large concentration estimates
resulting from extremely small mixing heights for a ground-level
area source is far greater when using RAMMET-generated mixing
heights. In this comparison, three of the six site-years
exhibited such an occurrence.
These results should be viewed with the reminder that the NWS
surface observations were used to create an on-site profile of
winds and temperature for METPRO. When METPRO was developed, the
intent was that the profile data would come from a multi-level,
instrumented tower (and in future years from remote sensing
capabilities) located at the site where the data are required.
The concentration estimates on an hour-by-hour basis would
certainly change if on-site data were employed, but the effect on
conclusions reached in this study would not be expected to change
significantly.
It should be emphasized that this comparison was for mixing
height estimation in rural areas only. As mentioned above, METPRO
does not estimate mixing heights for urban ares, as does RAMMET.
However, a modification to allow METPRO to emulate the urban
"switch" in RAMMET is being contemplated and will await future
testing and evaluation.
23
-------
7. REFERENCES
Carson, D., 1973. The Development of a Dry Inversion-Capped
Convectively Unstable Boundary Layer. Quart. J. R. Meteorol.
Soc., 99: 450-467.
Environmental Protection Agency, 1977. User's Manual for Single-
Source (CRSTER) Model. EPA Publication No. EPA-450/2-77-013.
U.S. Environmental Protection Agency, Research Triangle Park,
NC. (NTIS No. PB 271-360)
Environmental Protection Agency, 1986. Guideline on Air Quality
Models (Revised), and its supplements. EPA Publication No.
EPA-450/2-78-027R. U.S. Environmental Protection Agency,
Research Triangle Park, NC. (NTIS No. PB 86-245248)
Environmental Protection Agency, 1992. User's Guide for the
Industrial Source Complex (IS2) Dispersion Models., Volume II -
Description of Model Algorithms. EPA Publication No. EPA-
450/4-92-008b. U.S. Environmental Protection Agency, Research
Triangle Park, NC.
Holzworth, G., 1972. Mixing Heights, Wind Speeds, and Potential
for Urban Air Pollution Throughout the Contiguous United
States. AP-101. U.S. Environmental Protection Agency,
Research Triangle Park, NC.
Irwin, J. S. and J. 0. Paumier, 1990. Characterizing the
Dispersive State of Convective Boundary Layers for Applied
Dispersion Modeling. Bound.-Layer Meteorol., 53: 267-296.
Irwin, J. S., J. 0. Paumier and R. W. Erode, 1988. Meteorological
Processor for Regulatory Models (MPRM-1.1) User's Guide. EPA
Publication No. EPA-600/8-88-094. U.S. Environmental
Protection Agency, Research Triangle Park, NC.
Nieuwstadt, F. T. M., 1981. The Steady-State Height and Resistance
Laws of the Nocturnal Boundary Layer: Theory Compared with
Cabauw Observations. Bound.-Layer Meteorol., 20: 3-17.
Paumier, J. 0. and Irwin, J. S., 1991. Comparison of Modified
Carson and EPA Mixing Height Estimates Using Data From Five
Field Experiments. Seventh Joint Conference on Applications
of Air Pollution Meteorology with AWMA. American
Meteorological Society, Boston, MA.
Venkatram, A., 1980. Estimating the Monin-Obukhov Length in the
Stable Boundary Layer for Dispersion Calculations. Bound.-
Layer Meteorol., 19:481-485.
Weil, J. C. and R. P. Brower, 1983. Estimating Convective Boundary
Layer Parameters for Diffusion Applications. Environmental
Center, Martin Marietta Corp., PPSP-MP-48.
Zilitinkevich, S. S., 1972. On the Determination of the Height of
the Ekman Boundary Layer. Bound. Layer Meteorol., 3: 141-145.
24
-------
TECHNICAL REPORT DATA
(Please read Instructions on reverse before completing)
i. REPORT NO.
EPA-454/R-93-052
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
5. REPORT DATE
December 1993
Comparison of Design Concentrations Based on Hourly Mixing
Heights Estimated by RAMMET and METPRO
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
James O. Paumier
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Technical Support Division
Research Triangle Park, NC 27711
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68D00124 -
12. SPONSORING AGENCY NAME AND ADDRESS
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
A consequence analysis was conducted to investigate the effect on design concentration values
resulting from using two different methods for estimating mixing height. Two years of meteorological
data from Pittsburgh, PA, Oklahoma City, OK and Brownsville, TX were processed through RAMMET
and METPRO meteorological preprocessors. The mixing heights from METPRO were merged with the
RAMMET output and separate ISCST2 model runs were made using the two sets of mixing heights from
each site and year. The effects of the two mixing height algorithms on predicted high second-high
pollutant concentrations (design concentrations) were compared. The 1-hr design concentrations using
mixing height estimates from METPRO were generally larger than those using RAMMET mixing
heights for the three point sources by about 20%. Similar results were obtained when the data were
paired in time and space. This tendency was not as apparent for other averaging times. For the ground-
level area source, the potential for unusually large concentrations from extremely small mixing heights is
far greater when using RAMMET mixing heights.
17.
KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
Air Pollution
Atmospheric Dispersion Modeling
Meteorological Preprocessors
Mixing Height Estimation
18. DISTRIBUTION STATEMENT
Release Unlimited
b. IDENTIFIERS/OPEN ENDED TERMS
19. SECURITY CLASS (Report)
Unclassified
20. SECURITY CLASS (Page)
Unclassified
c. COSATI Field/Group
21. NO. OF PAGES
24
22. PRICE
EPA Form 2220-1 (Rev. 4-77) PREVIOUS EDITION IS OBSOLETE
U.S. tnvironmcnt-i: V.^ction Agency
Region 5, Library .PL-12J)
77 West Jackson Boulevard 12th
Chicago, IL 60604-3590
------- |