v>EPA
United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park NC 27711
EPA-450/4-86-015
September 1986
Air
A REVIEW OF
NMOC, NOxAND
NMOC/NOx
RATIOS
MEASURED IN
1984 AND 1985
-------
EPA-450/4-86-015
A REVIEW OF NMOC, NOX AND
NMOC/NOX RATIOS MEASURED IN
1984 AND 1985
By
Keith Baugues
U.S. ENVIRONMENTAL PROTECTION AGENCY
Air Management Technology Branch
Office of Air Quality Planning And Standards
Research Triangle Park, North Carolina 27711
-------- u.S. Environmental Protection Agency
September 1986 Region V, Library
230. South Dearborn Street
Chicago, Illinois 60604
-------
This report has been reviewed by the Office Of Air Quality Planning And Standards, U.S. Environmental
Protection Agency, and approved for publication. Any mention of trade names or commercial products is
not intended to constitute endorsement or recommendation for use.
EPA-450/4-86-015
-------
TABLE OF CONTENTS
Page
List Of Tables v
List Of Figures vii
1.0 Intrbduction 1
2.0 NMOC Data 4
2.1 1984 NMOC Data 4
2.2 1985 NMOC Data 4
2.3 Comparison Of 1984 And 1985 NMOC Levels 5
2.4 Continuous Hydrocarbon Data Versus PDFID Data 6
3.0 NOX Data 19
3.1 1984 NOX Data 19
3.2 1985 NOX Data 19
3.3 Comparison Of 1984 And 1985 NOX Levels 20
4.0 NMOC/NOX Ratio 30
4.1 1984 NMOC/NOX Ratios 30
4.2 1985 NMOC/NOX Ratios 31
4.3 Comparison Of 1984 And 1985 NMOC/NOX Ratios 32
4.4 NMOC/NOX Ratio Versus Ozone Level 33
4.5 Comparison Of NMOC/NOX Ratios - Historical Vs. Recent 33
5.0 Carbon Bond Splits For The Carbon Bond-4 Mechanism 48
5.1 1984 Carbon Fractions 48
5.2 1985 Carbon Fractions 48
5.3 Comparison Of 1984 And 1985 Carbon Fractions 49
5.4 Selection Of New Default Carbon Fractions For CB4 49
-------
Page
6.0 Mobile Source Contributions... 54
6.1 1984 Mobile Source Contributions 54
6.2 1985 Mobile Source Contributions 54
6.3 Comparison Of 1984 And 1985 Mobile Source Contributions 55
i "
6.4 Mobile Source Contributions From Emission Inventories 56
7.0 Biogenic NMOC Data 59
7.1 1984 Riogenic NMOC nata 59
7.2 1985 Biogenic NMOC Data 60
7.3 Comparison Of 1984 And 1985 Riogenic NMOC Data 60
8.0 NMOC Versus Ozone 68
9.0 Concl usi ons 73
9.1 NMOC 73
9.2 NOX 73
9.3 NMOC/NOX Ratios 73
9.4 Carbon Fractions 74
9.5 Mobile Source Contributions 74
9.6 Estimated Biogenic NMOC Data 74
9.7 NMOC Versus Ozone 74
9.8 Additional Data Needs 74
References 76
-------
LIST OF TABLES
Number Page
2-1 1984 NMOC ft
2-2 1985 NMOC 9
2-3 ' Comparison Of 1984 And 1985 NMOC 10
2-4 Median NMOC Increases Or Decreases 11
2-5 Results Of Linear Regression Analyses. 12
3-1 1984 NOX 22
3-2 1985 NOX 23
3-3 Comparison Of 1984 And 1985 NOX 24
3-4 Median NOX Increases Or Decreases 25
4-1 1984 NMOC/NOX Ratio 35
4-2 Percent Of NMOC/NOX Ratios Within Specified
Range (1984 Data) 36
4-3 1985 NMOC/NOX Ratio 37
4-4 Percent Of NMOC/NOX Ratios Within Specified
Range (1985 Data) 38
4-5 Comparison Of 1984 And 1985 NMOC/NOX Ratios
Measured At The Same Site 39
4-6 Median NMOC/NOX Ratio Increases Or Decreases
Measured At The Same Site 40
4-7 Median NMOC/NOX Ratio Vs. Ozone Level
Dallas 1984 41
4-8 Comparison Of Median NMOC/NOX Ratios (Historical Vs.
Recent) 42
-------
Number Page
5-1 Median Carbon Fractions For EKMA CR-4 Analysis
(Based Upon 1984 Data) 50
5-2 Median Carbon Fractions For EKMA CR4 Analyses
(Based Upon 1985 Data) 51
i *
5-3 Comparison of Median Carbon Fractions For EKMA CB4
Analyses 52
5-4 Overall Carbon Fractions (1984 And 1985) 53
6-1 Percent Mobile Source Contributions In 1984 And 1985
(Median Values) 57
6-2 Comparison Of Mobile Source Contributions From
Ambient Data Vs. Emission Inventory Data.. 58
7-1 Biogenic NMOC Data (1984) 62
7-2 Biogenic NMOC Data (1985) 63
7-3 Comparison Of Biogenic NMOC Data (1984 And 1985) 64
7-4 1984 And 1985 Biogenic NMOC Data Expressed As A
A Percent Of Total NMOC 65
8-1 Comparison Of 1984 And 1985 Ozone Data
(June-September Kansas City Sites) 70
VI
-------
LIST OF FIGURES
Number Page
2-1 1984 Median NMOC Levels 13
2-2 1984 Median NMOC Levels (Low To High) 14
2-3 ! 1985 Median NMOC Levels 15
2-4 1985 Median NMOC Levels (Low To High) 16
2-5 NMOC Data For Philadelphia 17
2-6 NMHC Vs. NMOC - Washington, D. C 18
3-1 1984 Median NOX Levels 26
3-2 1984 Median NOX Levels (Low To High) 27
3-3 1985 Median NOX Levels 28
3-4 1985 Median NOX Levels (Low To High) 29
4-1 1984 Median NMOC/NOX Ratios 43
4-2 1984 Median NMOC/NOX Ratios (Low To High) 44
4-3 1985 Median NMOC/NO Ratios 45
4-4 1985 Median NMOC/NOX Ratios (Low To High) 46
4-5 Population (SMSA) Vs. NMOC/NOX Ratio - 1984 47
7-1 Median Riogenic NMOC Concentrations - 1984 66
7-2 Median Biogenic NMOC Concentrations - 1985 67
8-1 Distribution of NMOC Plata - Kansas City 71
8-2 Kansas City Ozone - Site 2F01 72
-------
1.0 Introduction
During the summers of 1984 and 1985, morning (6-9am) measurements of
ambient nonmethane organic compounds (NMOC) were collected at 22 and 19 urban
sites, respectively. The data were collected by State and local agencies
which contributed grant funds and personnel. The EPA managed the analysis
i '
and provided NMOC sampling equipment. The method of determining NMOC levels
was the cryogenic preconcentration direct flame-ionization detection (PDFID)
described by McElroy et al .1 Data were collected to provide input values for
the Empirical Kinetic Modeling Approach (EKMA),2-6 a computer program which
estimates hydrocarbon control requirements necessary to attain the National
Ambient Air Quality Standard (NAAQS) for ozone. One of the key inputs to
EKMA is the Nonmethane Organic Compound/Nitrogen Oxides (NMOC/NOX) ratio.
Thus, a collocated NOX instrument was operated at each NMOC site.
Generally, NMOC sites were located in an attempt to determine "city-
wide" values. However, some sites, such as Houston, were located in indus-
trial areas while others are located in small urban areas which have a large
industrial component. The degree to which each site reflects city-wide
conditions affects conclusions regarding all the variables considered in this
report.
This report is broken into chapters which discuss various components of
the data base. Each chapter will describe results for 1984 and 1985 and then
compare statistics from the two years to indicate changes if any. Chapter 2
will discuss NMOC data. The NOX results from continuous monitors collocated
with the NMOC samplers are described in Chapter 3. Chapter 4 outlines results
for the NMOC/NOX ratios. Carbon Bond splits used in the Carbon Bond-3 mechanism
(CB-3) of EKMA are discussed in Chapter 5. Chapter 6 describes estimates of
-------
mobile source fractions, while findings regarding biogenic NMOC data are
outlined in Chapter 7. In Chapter 8, a case study is presented which compares
reductions in NMOC with changes in ozone levels. Overall conclusions are
found in Chapter 9 and are briefly summarized below.
Summary of Conclusions
Because of known deficiencies in the quality of most historical ambient
NMOC data, the high quality NMOC data collected during the summers of 1984
and 1985 provides a fairly unique opportunity to examine 6-9 a.m. summertime
concentrations of NMOC, NOX and NMOC/NOX ratios in several urban areas. The
analysis of these data indicates the following:
(1) In almost all cases, considerable day-to-day variability was found in
the magnitude of ambient concentrations of NMOC and NOX and in the NMOC/NOX
ratios measured at a given site within an urban area. A key implication of
this finding is that the NMOC/NOX ratio measured at a single site may not be
adequate to determine the NMOC/NOX ratio over an entire urban area. This is of
concern for two reasons. First, most States measure NMOC/NOX ratios at a
single site. Secondly, and more importantly, the NMOC/NOX ratio is a critical
parameter in determining the level of VOC control needed to attain the ozone
standard. For these reasons, it appears that the fewer the number of NMOC
monitors, the greater the magnitude of uncertainty in estimated control
estimates. At least two NMOC sites (and possibly more) should be operated to
more adequately characterize the NMOC/NOX ratio within an urban area. This
finding also suggests that NMOC monitors should be operated more frequently
(e.g., each summer) than the occasional sampling which is typically the case
in many urban areas.
-------
(2) The impact of mobile sources on measured ambient NMOC concentrations
was estimated using NMOC species data for each of the cities. The estimates
were based on an assumed relationship between the concentration of measured
acetylene and total NMOC in an urban area. Using this approximation procedure,
it is estimated that the mobile source contribution typically ranges from 45-70%
(and, in1some cases, 85-90%) of the total NMOC measured in most urban areas.
Although the data may only be characteristic of conditions at the monitoring
site (rather than of the entire urban area), they nevertheless suggest that
mobile sources may have a larger impact on ambient NMOC concentrations than
indicated by past VOC emission inventories. This is consistent with recent
findings which indicate that the magnitude of mobile source emissions may have
been underestimated in the past due to (a) larger than expected tampering rates
and (b) increased volatility of gasoline which increases the magnitude of evapo-
rative emissions. These findings suggest the need for updating VOC emission
inventories, since they are used as the basis for control strategy decisions
in SIPs.
(3) An analysis of the NMOC species data was performed to estimate the
contribution of biogenic sources to ambient NMOC concentrations in urban areas.
The estimation procedure assumes that reported concentrations of isoprene and
a-pinene approximate the contribution of biogenic sources to NMOC concentrations.
Using this procedure, it is estimated that biogenic emissions typically are less
than 1% in most urban areas. Although the procedure may tend to understate the
contribution of biogenic sources (since it does not include all species emitted
by biogenic sources), the analysis lends support to the position that anthropogenic
sources are the predominant source of ambient NMOC concentrations in urban areas.
(4) In many analyses, the Beaumont data appear to be anomalous. The data
are valid. However, the reasons for this site exhibiting trends counter to all
other sites are not known.
3
-------
2.0 NMOC Data
2.1 1984 NMOC Data
NMOC was measured from 6-9 a.m. (local clock time) at each site, Monday
through Friday. The sampling schedule varied slightly from site to site but,
in general, covered the time period from mid-June until the end of September.
The data consist of 3-hour integrated averages.
Table 2-1 lists key statistics for each city. Included are the number
of samples, 10th & 90th percentiles, median, mean, and standard deviation.
Figure 2-1 illustrates the median NMOC values measured at each of the 21
sites. A site was operated in Philadelphia, but the data have not been included
due to ethylene contamination. In Figure 2-2, these values are displayed
graphically in ascending order. The lowest median NMOC level (0.39 ppmC) was
recorded in Charlotte, North Carolina, while the highest median NMOC level
(1.27 ppmC) was measured in Memphis, Tennessee. No obvious geographical
patterns are apparent. The overall median of all 21 cities is approximately
0.72 ppmC.
The range of NMOC values within a given city is large. Typical
minimum NMOC levels for each city are on the order of 0.1-0.2 ppmC. Maximum
NMOC values usually exceeded 2 ppmC. In fact, 19 of the 21 sites measured
maximum NMOC levels greater than 2 ppmC, while 11 of the 21 sites measured
maximum NMOC levels in excess of 3 ppmC.
2.2 1985 NMOC Data
In 1985, 19 sites were operated in 18 cities. Two sites were operated
in Philadelphia. The sampling schedule varied slightly from site to site
but, in general, covered the time period from the beginning of June until the
end of September.
-------
Table 2-2 lists key statistics for each city. Included are the number
of samples, 10th & 90th percentiles, median, mean, and standard deviation.
Figure 2-3 illustrates the median NMOC values measured at each of the 19
sites. On Figure 2-4, these values are displayed graphically in ascending
order. The lowest median NMOC level (0.38 ppmC) was recorded in Boston,
t "
Massachusetts, while the highest median NMOC level (1.63 ppmC) was measured
in Beaumont, Texas. No obvious geographical patterns are present. The
overall median of all 19 sites is approximately 0.60 ppmC.
The range of NMOC values within a given city is large. Typical minimum
NMOC levels for each city are on the magnitude of 0.1-0.3 ppmC. Maximum NMOC
values usually exceeded 2 ppmC. In fact, 14 of the 19 sites measured NMOC
levels greater than 2 ppmC, while 4 of the 19 sites measured NMOC levels in
excess of 3 ppmC.
A comparison of the median NMOC levels for the two Philadelphia sites
was made. The median NMOC level for Site 1 is 0.49, while the median NMOC
level for Site 2 is 0.65. The Mann Whitney ll-test of the median NMOC levels
to determine if a significant difference exists, indicated that at the 95%
confidence level, no significant difference in the medians existed. This is
due to the large scatter in the NMOC levels. Figure 2-5 shows a frequency
distribution plot of both sites. Site 1 is shown above the zero line, while
Site 2 is shown below the zero line.
2.3 Comparison Of 1984 And 1985 NMOC Levels
It is inappropriate to compare overall medians from 1984 to 1985 and
draw conclusions regarding NMOC levels since samples were not collected in
the same cities in both years. However, 10 cities had NMOC measured at the
same sites in both 1984 and 1985. Comparisons of levels for both years can
be made and conclusions drawn for each city.
5
-------
Table 2-3 lists these cities, along with the number of samples, 10th &
90th percentiles, median, mean, and standard deviation, for both years. A
simple comparison of median NMOC levels would indicate that NMOC levels in 9
of the in cities decreased from 1984 to 1985. However, with the large spread
in the data, such conclusions may not be correct. A statistical test of the
medians (Mann Whitney ll-test) was performed to determine whether changes from
1984 to 1985 were statistically significant at the 95% confidence level. Of
the nine cities with decreases in NMOC concentrations, the decreases were
statistically significant in five cities, but not in the other four cities.
The NMOC increase in Beaumont was found to be statistically significant.
Table 2-4 shows the results of the Mann Whitney test, along with the percent
decrease (or increase) in NMOC. It should be noted that reductions > 20%
occurred at three cities.
While this analysis indicates that 1985 NMOC levels at several cities
were lower than 1984 NMOC levels, it does not explain the cause of these
reductions. Lower NMOC levels may be due to varying meteorology, local
economic conditions or Volatile Organic Compound (VOC) control programs.
2.4 Continuous Hydrocarbon Data Versus PDFID Data
Continuous hydrocarbon analyzers measure total hydrocarbons and methane
and take the difference to determine nonmethane hydrocarbon concentrations.
Typically total hydrocarbon and methane are large values of similar magnitude,
The difference of these two large values is a smaller value than either total
hydrocarbon or methane and is subject to errors. [The data from continuous
hydrocarbon samplers are referred to as Nonmethane Hydrocarbons (NMHC) in
this paper.] Comparisons of NMHC data with collocated NMOC data (from the
-------
PDFID) indicate poor agreement between the two instruments. Table 2-5 shows
comparisons for five cities where collocated NMOC/NMHC data were available
for 1984. Also shown is a comparison between data from analysis of samples
by both the PDFID and the fias Chromatograph (GC). A GC determines the concen-
trations of individual species such as propane, butane, etc. Concentrations
j
of these individual species are summed to arrive at a total NMOC level. This
technique is considered to be the most accurate method of determining NMOC
levels.
Figure 2-fi is a plot of NMHC versus NMOC for the Washington, D.C. case
listed in Table 2-5. The good agreement between the NMOC and GC data and the
poor agreement between the NMOC and NMHC data indicate that the NMHC data from
continuous monitors are suspect. In fact, based upon this information, the
Office of Air Duality Planning and Standards (OAOPS) has indicated that
ambient hydrocarbon data measured with the continuous technique are no longer
adequate for EKMA modeling analyses, unless the NMHC data are shown to be
comparable with GC measurements.7
-------
Table 2-1
1984 NMOC (ppmC)
6-9 AM
3-Hour Averages
City
Akron, OH
Atlanta, GA
Beaumont, TX
Birmingham, AL
Charlotte, NC
Chattanooga, TN
Cincinnati, OH
Clute, TX
Dallas, TX
El Paso, TX
Fort Worth, TX
Indianapolis, IN
Kansas City, MO
Memphis, TN
Miami , FL
Richmond, VA
Texas City, TX
Washington, DC
West Orange, TX
West Palm Beach, FL
Wilkes, Barre, PA
n 10th
Percentile
65
52
64
55
56
44
64
67
70
65
66
56
66
51
28
62
65
60
67
70
61
0.35
0.33
0.26
0.40
0.22
0.72
0.37
0.33
0.48
0.47
0.49
0.41
0.38
0.74
0.48
0.27
0.32
0.46
0.26
0.25
0.15
90th
Percentile
1.41
1.46
1.54
2.21
0.83
2.61
1.83
1.47
1.52
1.57
1.63
1.24
1.56
2.42
2.24
0.85
1.82
1.33
1.27
0.86
0.75
Median
0.60
0.60
0.75
0.73
0.39
1.16
0.74
0.70
0.89
0.82
0.80
0.69
0.62
1.27
1.03
0.50
0.78
0.71
0.65
0.49
0.43
Mean
0.75
0.80
0.89
1.00
0.55
1.36
0.93
0.83
0.95
0.92
0.97
0.80
0.79
1.44
1.31
0.54
0.94
0.82
0.69
0.55
0.45
Standard
Deviation
0.51
0.73
0.68
0.68
0.45
0.72
0.71
0.61
0.44
0.45
0.57
0.42
0.54
0.85
0.85
0.24
0.80
0.42
0.47
0.35
0.23
8
-------
Table 2-2
1985 NMOC (ppmC)
6-9 AM
3-Hour Averages
City
Raton Rouge, !LA
Beaumont, TX
Boston, MA
Cleveland, OH
Clute, TX
Dallas, TX
Fl Paso, TX
Fort Worth, TX
Houston, TX
Kansas City, MO
Lake Charles, LA
Philadelphia 1, PA
Philadelphia 2, PA
Portland, ME
Richmond, VA
St. Louis, MO
Texas City, TX
Washington, DC
West Orange, TX
n 10th
Percent! le
79
83
43
73
80
76
82
76
72
84
79
63
58
54
60
76
78
56
84
0.32
0.95
0.17
0.44
0.23
0.41
0.25
0.40
0.35
0.25
0.26
0.27
0.37
0.13
0.28
0.35
0.18
0.38
0.26
90th
Percentile
1.87
2.60
0.70
1.55
1.42
1.57
1.29
1.36
1.55
1.13
1.08
1.18
1.04
0.82
0.86
1.22
1.29
1.06
0.97
Median
0.60
1.63
0.38
0.78
0.63
0.73
0.67
0.63
0.74
0.41
0.55
0.49
0.65
0.43
0.45
0.57
0.42
0.60
0.52
Mean
0.81
1.76
0.41
0.87
0.72
0.88
0.72
0.76
0.92
0.53
0.62
0.65
0.72
0.49
0.52
0.71
0.61
0.68
0.58
Standard
Deviation
0.64
0.78
0.23
0.44
0.45
0.50
0.40
0.40
0.61
0.36
0.35
0.45
0.38
0.37
0.28
0.41
0.52
0.28
0.29
-------
Table 2-3
Comparison Of 1984 And 1985 NMOC (ppmC)
(Same Site)
City
Beaumont, TX
Clute, TX
Dallas, TX
El Paso, TX
Fort Worth, TX
Kansas City, MO
Richmond, VA
Texas City, TX
Washington, DC
West Orange, TX
Year
1984
1985
1984
1985
1984
1985
1984
1985
1984
1985
1984
1985
1984
1985
1984
1985
1984
1985
1984
1985
n 10th
Percenti le
64
83
67
80
70
76
65
82
66
76
66
84
62
60
65
78
60
56
67
84
0.26
0.95
0.33
0.23
0.48
0.41
0.47
0.25
0.49
0.40
0.38
0.25
0.27
0.28
0.32
0.18
0.46
0.38
0.26
0.26
90th
Percentile
1.54
2.60
1.47
1.42
1.52
1.57
1.57
1.29
1.63
1.36
1.56
1.13
0.85
0.86
1.82
1.29
1.33
1.06
1.27
0.97
Median
0.75
1.63
0.70
0.63
0.89
0.73
0.82
0.67
0.80
0.63
0.62
0.41
0.50
0.45
0.78
0.42
0.71
0.60
0.65
0.52
Mean
0.89
1.76
0.83
0.72
0.95
0.88
0.92
0.72
0.97
0.76
0.79
0.53
0.54
0.52
0.94
0.61
0.82
0.68
0.69
0.58
Standard
Deviation
0.68
0.78
0.61
0.45
0.44
0.50
0.45
0.40
0.57
0.40
0.54
0.36
0.24
0.28
0.80
0.52
0.42
0.28
0.47
0.29
10
-------
Table 2-4
Median NMOC Increases Or Decreases
(Same Site)
City
Beaumont, TX
Clute, TX
Dallas, TX
El Paso, TX
Fort Worth, TX
Kansas City, MO
Richmond, VA
Texas City, TX
Washington, DC
West Orange, TX
Direction Of NMOC
i
4-
4-
4-
4-
4-
4-
4-
4-
4-
% Change In NMOC (1984-85)
117%
-10%*
-18%*
-18%
-21%
-34%
-10%*
-46%
-15%
-20%*
* Not significant at the 95% confidence level
11
-------
Table 2-5
Results Of Linear Regression Analyses
NMHC Versus NMOC (1984)
City
Indianapolis, IN
Indianapolis, IN
(excluding 1 outlier)
Kansas City, MO
Richmond, VA
Washington, DC
Wiles Rarre, PA
n Slope
37 1.34
36 1.02
49 .26
38 .91
59 1.28
33 .08
Intercept r
-.07 .88
.11 .69
.48 .50
.46 .56
.56 .41
.39 .35
r2
.77
.47
.25
.32
.16
.12
/////////////////////////////////////
GC Versus NMOC
n Slope
(1984)
Intercept r
r>2
All Samples
336
1.08
.015
.97
.95
12
-------
13
-------
o
E
CvJ &
| w
CM2
u
o
DH z
D!
^2s
u_ 2
10
N\X\\\XX\X\XX\\\\\V^H\\N-
X\\\\\\\\\\\\\\ .
\\\\\\\\\V\ «ft?
sX\\\\\\
Sv\\s\\\
qoeag
3jjeg
\\\\\\\\\ MR
I 1 1 1 1
1 1 1 1 1
o
f
§
OOOOOOOOO
(Ouudd)
14
-------
15
-------
O
E
a
CL
3
LJ o
s
Q
in
00
en
\SVV\\\>\SV>\S>^ foH*^"
Rvx^^:
N
QQ
\\ aSno
\X\NV
v
5 jo/ A,
y
X
,
\
\.
X
X X V
V
V
X
x V
v x v X
v
\ uo^sog
!
I I I I I I I
QOONtOinrH"^CNT-O
666666666
o
(Otudd)
16
-------
Figure 2-5
T.
Q_
_J
Ld
Q
Q
m
o:
<8
Q
U
O
z
I "
i
e=
e=
e=z=
E=T
y ^ i
r i
y
r=3
CQ
: : i
D if) O if) O u
1 1
T O LO C
- m
- (N
-
-------
Figure 2-6
O
O
00
01
O
01
i r
oo co
*
IN tN
01
01
CN
I
oo
I 1
CO «
i
IN
i
CO
O
oo
O
(O
O
O
E
a
a
t S
O 2
CN
d
CO
O
"t
O
N
O
(Otudd)
18
-------
3.0 NOX Data
3.1 1984 N0y Data
At each NMOC site, a continuous NOX analyzer was operated. Values
discussed and analyzed include only NOX samples which were measured along
with the, 6-9 a.m. (local clock time) NMOC samples. If an NMOC sample was
missed on a sample day and NOX was available, the NOX data were not included
in the analysis.
Table 3-1 lists key statistics for each city. Included are the number
of samples, 10th and 90th percentile, median, mean, and standard deviation.
Figure 3-1 illustrates the median NOX values measured at each of the 22 sites.
On Figure 3-2, the medians are displayed graphically in ascending order. The
lowest median NOX level (0.010 ppm) was measured in West Orange, Texas, while
the highest median NOX level (0.088 ppm) was recorded in Memphis, Tennessee.
No obvious geographical patterns are present. While several Texas cities
show low NOX levels, these cities are also the least populated cities analyzed,
The levels at West Orange are considered so low as to be questionable, since
the median value is within the noise level of the instrument.
The range of 6-9 AM average NOX values within a given city is fairly
large. Typical minimum NOX levels for each city are in the range of 0.000-
0.030 ppm. Maximum NOX levels usually exceed 0.100 ppm. Seventeen of the 21
sites recorded maximum NOX levels above 0.1 ppm, while 10 of the 21 sites
recorded maximum NOX levels above 0.2 ppm.
3.2 1985 N0y Data
Table 3-2 lists key statistics for each city. Included are the number
of samples, 10th and 90th percentiles, median, mean, and standard deviation.
Figure 3-3 illustrates the median NOX values measured at each of the 19
19
-------
sites. On Figure 3-4, the medians are displayed graphically in ascending
order. The lowest median NOX level (0.005 ppm) was measured in West Orange,
Texas, while the highest median NOX level (0.100 ppm) was recorded in Cleveland,
Ohio. No obvious geographical patterns are present. Once again, the NOX
levels at West Orange are considered so low as to be questionable. Since
the values are less than the noise level of the instrument.
The range of NOX values within a given city is fairly large. Typical
minimum NOX levels for each city are in the range of 0.000-0.030 ppm. Maximum
NOX levels usually exceed 0.1 ppm. Fifteen of the 19 sites recorded NOX
maximum levels above 0.1 ppm, while seven of the 19 sites recorded maximum
NOX levels above 0.2 ppm.
3.3 Comparison Of 1984 And 1985 N0y Levels
Ten cities measured NOX levels in both 1984 and 1985 at the same site.
Table 3-3 lists these cities, along with the number of samples, 10th and 90th
percentiles, median, mean, and standard deviation for both years. A quick
review of mean NOX levels would indicate that mean NOX levels decreased in
eight of the ten cities, went up in one, and remained constant in the tenth
city. However, with the large spread in the data, such conclusions may not
be correct. Since the median NOX levels at West Orange are so low, no com-
parison was made for that city. A statistical test of the medians (Mann
Whitney U-test) was performed to determine whether changes from 1984 to 1985
were statistically significant at the 95% confidence level. The results of
the test are shown in Table 3-4, along with the percent decrease (or increase)
in mean NOX. Of the seven cities which showed decreases in median NOX, the
-------
decreases were significant at only three of the cities. Of the two cities
which showed increases in the median NOX, the increase was significant at
only one city.
21
-------
Table 3-1
1984 NOX (ppm)
6-9 AM Averages
City
Akron, OH
Atlanta, GA
Beaumont, TX
Birmingham, AL
Charlotte, NC
Chattanooga, TN
Cincinnati , OH
Clute, TX
Dallas, TX
El Paso, TX
Fort Worth, TX
Indianapolis, IN
Kansas City, MO
Memphis, TN
Miami, FL
Richmond, VA
Texas City, TX
Washington, DC
West Orange, TX
West Palm Beach, FL
Wilkes Barre, PA
n
49
52
45
51
55
35
51
52
69
60
58
31
61
35
15
62
52
54
41
60
53
10th
Percent! le
0.028
0.029
0.010
0.038
0.020
0.036
0.040
0.017
0.030
0.025
0.037
0.037
0.035
0.032
0.025
0.023
0.010
0.047
0.005
0.010
0.005
90th
Percenti le
0.128
0.123
0.050
0.149
0.086
0.117
0.170
0.054
0.100
0.110
0.170
0.142
0.116
0.168
0.191
0.087
0.038
0.150
0.037
0.100
0.074
Median
0.049
0.055
0.030
0.061
0.037
0.069
0.070
0.025
0.050
0.050
0.070
0.058
0.063
0.083
0.070
0.047
0.021
0.076
0.010
0.029
0.030
Mean
0.065
0.071
0.028
0.082
0.054
n.073
0.091
0.032
0.059
0.059
0.087
0.081
0.070
0.088
0.086
0.051
0.022
0.094
0.016
0.040
0.035
Standard
Deviation
0.054
0.060
0.013
0.057
0.049
0.035
0.059
0.019
0.032
0.034
0.054
0.066
0.03Q
0.054
0.061
0.026
0.011
0.062
0.012
0.034
0.026
22
-------
Table 3-2
1985 NOX (ppm)
6-9 AM Averages
City
Raton Rouge, LA
Beaumont, TX
Boston, MA
Cleveland, OH
Clute, TX
Dallas, TX
El Paso, TX
Fort Worth, TX
Houston, TX
Kansas City, MO
Lake Charles, LA
Philadelphia 1, PA
Philadelphia 2, PA
Portland, ME
Richmond, VA
St. Louis, MO
Texas City, TX
Washington, DC
West Orange, TX
n
67
62
43
72
72
68
76
54
69
84
77
55
55
52
58
73
68
48
78
10th
Percentile
0.023
0.020
0.028
0.054
0.010
0.037
0.020
0.030
0.027
0.025
0.013
0.047
0.021
0.013
0.022
0.036
0.008
0.039
0.005
90th
Percentile
0.080
0.040
0.081
0.202
0.047
0.120
0.107
0.138
0.117
0.121
0.039
0.125
0.150
0.087
0.085
0.109
0.030
0.123
0.010
Median
0.039
0.027
0.04fi
0.100
0.020
0.063
0.051
0.057
0.050
0.049
0.022
0.067
0.065
0.031
0.038
0.062
0.017
0.067
0.005
Mean
0.047
0.028
0.049
0.114
0.027
0.071
0.057
0.069
0.065
0.065
0.025
0.081
0.083
0.044
0.046
0.069
0.018
0.078
0.006
Standard
Deviation
0.025
0.009
0.021
0.059
0.022
0.037
0.035
0.044
0.035
0.052
0.010
0.042
0.087
0.032
0.026
0.031
0.008
0.037
0.003
23
-------
Table 3-3
Comparison Of 1984 And 1985 NOX (ppm)
6-9 AM Averages (Same Site)
City
Beaumont, TX
Clute, TX
Dallas, TX
El Paso, TX
Fort Worth, TX
Kansas City, MO
Richmond, VA
Texas City, TX
Washington, DC
Year
1984
1985
1984
1985
1984
1985
1984
1985
1984
1985
1984
1985
1984
1985
1984
1985
1984
1985
n 10th
Percentile
45
62
52
72
69
68
60
76
58
54
61
84
62
58
52
68
54
48
0.010
0.020
0.017
0.010
0.030
0.037
0.025
0.020
0.037
0.030
0.035
0.025
0.023
0.022
0.010
0.008
0.047
0.039
90th
Percentile
0.050
0.040
0.054
0.047
0.100
0.120
0.110
0.107
0.170
0.138
0.116
0.121
0.087
0.085
0.038
0.030
0.150
0.123
Median
0.030
0.027
0.025
0.020
0.050
0.063
0.050
0.051
0.070
0.057
0.063
0.049
0.047
0.038
0.021
0.017
0.076
0.067
Mean Standard
Deviation
0.028
0.028
0.032
0.027
0.059
0.071
0.059
0.057
0.087
0.069
0.070
0.065
0.051
0.046
0.022
0.018
0.094
0.078
0.013
0.009
0.019
0.022
0.032
0.037
0.034
0.035
0.054
0.044
0.039
0.052
0.026
0.026
0.011
0.008
0.062
0.037
24
-------
Table 3-4
Median NOX Increases Or Decreases
6-9 AM Averages
City Direction Of NOX
Beaumont, TX +
t
Clute, TX +
Dallas, TX t
El Paso, TX t
Fort Worth, TX +
Kansas City, MO *
Richmond, VA +
Texas City, TX +
Washington, DC +
% Change In NOX (1984-85)
-10%*
-20%
26%
2%*
-19%*
-22%
-19%*
-19%
-12%*
* Not significant at the 95% confidence level.
25
-------
26
-------
CM
n
E
Q.
Q.
I/)
_J
u
( n Q
w LJ
00
C5
O
\\\\\\\\\\\\\\\\XV^ .N-
OD
O
o
«
o
\-
kVO\^0\V^\\ ?*»
\\N\\\\\V
-
\-
sa-[Tf\
X. -
\-
o
o
m
o
i
o
«
o
O
o
o
CM
O
O
o
o
(uudd) XQN
27
-------
28
-------
LL
i
O)
O
o
X\\\\\\\\\\\\\\\\\\\>"*^P>-
;gftN-
^d\-
Z emaispBTTqd S -
^\^\\\\^\N ?N
£ K\\\\\\\\\':^^
a
i±j
(ij
-i
2 - K\\\\\\\N"H^
< . , , . N ^
m
oo
O)
O
o
1
N
O
o
O
o
I
lf)
O
o
O
o
O
o
O
o
O
*
o
(aidd) XQN
<-
29
-------
4.0 NMOC/Nf)Y Ratio
4.1 1984 NMOC/NOX Ratios
Perhaps the most important variable in an EKMA analysis is the NMOC/NOX
ratio. All other factors being equal, the higher the NMOC/NOX ratio, the
more control is necessary to reduce VOC to a level that will ensure attainment
t
of the NAAQS for ozone. In most EKMA analyses for ozone SIPs, a median
NMOC/NOX ratio is used for all days modeled (generally 5 days per site). The
median NMOC/NOX ratio is determined by calculating a NMOC/NOX ratio for each
day and then ranking those values and selecting the median. The median
NMOC/NOX ratio is not the ratio of the median NMOC to the median NOX level
for all days.
Table 4-1 lists key statistics for each city. Included are the number
of samples, 10th and 90th percentiles, median, mean, and standard deviation.
Data for West Orange have been excluded due to the questionable NOX values.
Figure 4-1 illustrates the median NMOC/NOX ratios for each of the 20 cities.
[Philadelphia is not shown, since the NOX data are missing.] On Figure 4-2,
those values are displayed graphically in ascending order. The lowest median
NMOC/NOX ratio (9.1) was observed in Cincinnati, Ohio, while the highest
median NMOC/NOX ratio (37.7) was observed in Texas City, Texas. No obvious
geographical patterns are present. While the Gulf Coast of Texas has higher
NMOC/NOX ratios than elsewhere, this difference is due to city size and
therefore low NOX levels, instead of location.
The range of daily NMOC/NOX ratios within a given city is large. Typical
minimum day specific NMOC/NOX ratios for each city are in the range of 2-8.
Maximum NMOC/NOX ratios usually exceed 40. Of the 20 sites, 15 had maximum
NMOC/NOX ratios greater than 40. Eleven of the sites had maximum NMOC/NOX
ratios greater than 60.
30
-------
With the large spread in the range of the NMOC/NOX ratios within a given
city, it is debatable whether the median ratio is a reliable representative
value for these urban areas. Table 4-2 lists the percent of NMOC/NOX ratios
within a given range (i.e., the median plus or minus a given percent). For
example, 65% of the Akron NMOC/NOX ratios are contained in a range of the
median ratio, plus or minus 30%. In order for the range to encompass 50% of
the NMOC/NOX ratios, a range of the median _+ 30% (or more) is usually necessary
for nearly all cities. The typical median ratio is approximately 13. Thus,
the range is 9.1-16.9 (13 +_ 30%).
4.2 1985 NMOC/NOy Ratios
In 1985, 19 sites operated in 18 cities. Two sites were operated in
Philadelphia. NMOC/NOX ratios will be discussed for only 18 of these sites
since the West Orange NOX data is questionable. Table 4-3 lists key statistics
for each city. Figure 4-3 illustrates the median NMOC/NOX ratios for each of
the 18 sites. On Figure 4-4, those values are displayed graphically in ascending
order. The lowest median NMOC/NOX ratio (6.5) was observed in Philadelphia,
Pennsylvania (Site 1), while the highest median NMOC/NOX ratio (53.2) was
observed in Beaumont, Texas. No obvious geographical patterns are present.
The range of daily NMOC/NOX ratios within a given city is large. Typical
minimum NMOC/NOX ratios for each city are in the range of 2-8. Maximum daily
NMOC/NOX ratios usually exceed 40. Of the 18 sites, 16 observed maximum daily
NMOC/NOX ratios equal to or greater than 40. Nine of the sites observed
maximum daily NMOC/NOX ratios greater than or equal to 60.
Table 4-4 lists the percent of NMOC/NOX ratios within a given range for
each city. Similarly to 1984, in order to encompass 50% of the NMOC/NOX
ratios a range of the median _+ 30% (or more) is necessary.
31
-------
From the data presented it is doubtful that the median ratio at a single
site would provide a representative ratio for an urban area. For this reason,
EPA is recommending that a minimum of two NMOC sites be operated in a city to
determine the appropriate NMOC/NOX ratio. Additional sites are desirable.
More specific guidance on the appropriate number of NMOC sites should be
available in the Fall of 1987, at the completion of a special study designed
to address this particular issue. This study will also examine the question
of whether a median value or day-specific ratios are more suitable for
characterizing a city's NMOC/NOX ratios for EKMA modeling.
4.3 Comparison Of 1984 And 1985 NMOC/NOy Ratios
Nine cities had NMOC/NOX ratios in both 1984 and 1985 for the same site.
Table 4-5 lists these cities, along with the number of ratios, 10th & 90th
percentile range, median, mean, and standard deviation for both years. A
quick review of the median NMOC/NOX ratios would indicate that median NMOC/NOX
ratios increased in four cities and decreased in five other cities. However,
with the large spread in the data, such conclusions may not be correct. A
statistical test of the median NMOC/NOX ratios (Mann-Whitney U Test) was
performed to determine whether changes from 1984 to 1985 were statistically
significant at the 95% confidence level. The results of the tests are shown
in Table 4-6, along with the percent decrease (or increase) in median NMOC/NOX
ratio. Of the five cities with decreases in median NMOC/NOX ratio, the decreases
were statistically significant in four of the cities. Of the four cities
with increases in median NMOC/NOX ratio, the increases were statistically
significant at only one site. In Beaumont, the median NMOC/NOX ratio increased
significantly from 1984 to 1985.
32
-------
4.4 NMOC/NOX Ratio Versus Ozone Level
The Dallas NMOC site was chosen to compare maximum ozone concentrations
to corresponding 6-9am NMOC/NOX ratios using 1984 data. All maximum daily
ozone concentrations within a 50 mile radius of the NMOC sampling station
were used for days when NMOC samples were taken. These ozone concentrations
were compared with resultant wind directions from the 8 am - 3 pm period.
to determine ozone readings that would correspond with NMOC parcels moving
from the NMOC monitoring site to the ozone monitors. The maximum ozone level
of all sites that met the above criteria was compared with the NMOC/NOX ratio
for that day. Table 4-7 shows the results of this analysis. Median NMOC/NOX
ratios associated with increasing ozone levels are shown. From the table it
appears that the median NMOC/NOX ratios at the Dallas site do not significantly
increase as the ozone level increases.
The failure of NMOC/NOX ratios to correlate with concurrent ozone levels
is not surprising. Meteorological conditions play a critical part in the
formation of ozone. Days with high 6-9 am NMOC/NOX ratios may be cloudy or
have high winds which would result in low maximum ozone levels. Also, days
with high maximum ozone levels may have large contributions from ozone aloft.
These factors would tend to confuse a simple comparison of NMOC/NOX ratios
with concurrent ozone maxima. Consideration of these variables in the
comparison was beyond the scope of this analysis.
4.5 Comparison of NMOC/NOX Ratios - Historical vs. Recent
One area of special interest is how the recent NMOC/NOX ratios compare
to NMOC/NOX ratios used in past analyses (State Implementation Plans). The
NMOC/NOX ratios for the cities involved in the 1984 and 1985 studies were
33
-------
compared to data reported in earlier SIPs. These cities and ratios are
listed in Table 4-8.
Of the seven cities shown in Table 4-8, higher median NMOC/NOX ratios
are observed in 1984 and 1985 than were used in previous SIP analyses in six
cases. Boston is the lone exception. Since significant reductions in NMOC
t
emissions should have occurred between approximately 1980 and 1984 or 1985
due to the Federal Motor Vehicle Control Program (FMVCP) and VOC control
plans, NMOC/NOX ratios should have decreased over this time period. Two
caveats should be mentioned, however. First, the NMOC/NOX ratios used in the
SIP analyses were based upon continuous NMHC sampling and are therefore
suspect. Secondly, expected emission reductions may have been offset by
growth and higher than expected motor vehicle tampering rates.
Figure 4-5 shows a plot of 1984 NMOC/NOX ratios versus population.
While there appears to be a negative relationship between the NMOC/NOX ratio
and population (the ratio decreases as population increases) the correlation
is poor.
Higher ratios may also be due to improved sampling and analysis procedures
for NMOC and NOX used in the 1984 and 1985 studies. Emission patterns for
both NMOC and NOX may have also changed.
34
-------
Table 4-1
1984 NMOC/NOX Ratio
City
Akron, OH
Atlanta, GA
t "
Reaumont, TX
Birmingham, AL
Charlotte, NC
Chattanooga, TN
Cincinnati , OH
Clute, TX
Dallas, TX
EL Paso, TX
Fort Worth, TX
Indianapolis, IN
Kansas City, MO
Memphis, TN
Miami, FL
Richmond, VA
Texas City, TX
Washington, DC
West Palm Reach, FL
Wilkes Barre, PA
n
49
52
45
51
55
35
51
52
69
60
58
31
61
35
15
62
52
54
60
53
10th
Percentile
8.0
7.0
13.8
7.0
6.6
11.9
6.0
13.4
11.0
10.5
6.9
6.7
6.4
7.5
9.5
6.1
20.2
5.1
7.0
6.1
90th
Percentile
18.4
19.8
46.5
25.4
22.2
30.5
15.4
58.8
28.1
31.5
20.2
16.6
19.5
63.5
20.5
16.9
79.8
14.8
38.0
56.7
Median
12.8
10.4
25.3
11.7
10.4
16.7
9.1
23.7
16.0
15.1
11.5
10.9
9.9
13.9
13.3
10.5
37.7
9.3
14.2
14.3
Mean
13.1
13.3
30.8
13.4
11.8
22.3
14.2
33.7
17.9
19.8
13.4
11.6
12.2
25.1
13.7
12.1
54.3
10.6
19.2
23.5
Standard
Deviation
4.7
11.1
22.2
7.6
5.6
21.3
19.2
28.9
7.7
16.3
9.6
4.5
8.3
30.4
3.8
7.2
66.5
7.7
15.1
22.4
35
-------
Table 4-2
Percent Of NMOC/NOX Ratios Within Specified Range
(1984' Data)
City
Akron, OH
Atlanta, GA
Reaumont, TX
Birmingham, AL
Charlotte, NC
Chattanooga, TN
Cincinnati, OH
Clute, TX
Dallas, TX
El Paso, TX
Fort Worth, TX
Indianapolis, IN
Kansas City, MO
Memphis, TN
Miami, FL
Richmond, VA
Texas City, TX
Washington, DC
West Palm Beach, FL
Wilkes Barre, PA
Median
Ratio
12.8
10.4
25.3
11.7
10.4
16.7
9.1
23.7
16.0
15.1
11.5
10.9
9.9
13.9
13.3
10.5
37.7
9.3
14.2
14.3
Median
+_ 10%
26
25
11
20
29
20
29
19
23
15
22
19
21
11
27
26
19
19
12
6
Median
+_ 20%
45
40
38
35
45
40
43
29
49
40
40
39
48
20
47
39
37
41
17
21
Median
+_ 30%
65
63
47
53
64
63
61
35
70
60
57
61
61
31
67
61
56
56
30
30
Median
_+ 40%
84
77
62
71
76
86
71
54
80
67
69
74
79
46
87
68
62
70
45
43
Median
i 50%
92
79
67
75
78
89
80
63
86
75
79
87
84
57
87
84
71
78
57
55
Median
+_ 60%
92
83
76
78
80
89
84
71
88
83
83
97
87
63
100
89
79
89
73
64
36
-------
Table 4-3
1985 NMOC/NOX Ratio
City
Baton Rouge, LA
Beaumont, TX
Boston, MA
Cleveland, OH
Clute, TX
Dallas, TX
El Paso, TX
Fort Worth, TX
Houston, TX
Kansas City, MO
Lake Charles, LA
Philadelphia 1, PA
Philadelphia 2, PA
Portland, ME
Richmond, VA
St. Louis, MO
Texas City, TX
Washington, DC
n
67
62
43
72
72
68
76
54
69
84
77
55
55
52
58
73
68
48
10th
Percentile
8.9
28.6
4.5
5.2
11.7
6.8
9.1
6.6
5.6
4.6
14.8
4.2
5.5
6.1
7.4
6.4
12.9
5.3
90th
Percentile
27.6
123.8
14.7
11.3
77.8
29.2
21.2
20.6
36.7
15.3
35.8
14.8
27.5
20.8
15.2
18.2
91.2
14.2
Median
14.9
53.2
7.6
7.5
24.6
11.8
11.9
11.8
12.9
8.5
23.7
6.5
9.5
11.6
11.2
9.6
28.7
8.7
Mean
17.7
69.1
8.9
7.9
37.6
15.2
14.3
13.1
19.1
11.2
26.2
8.6
12.3
13.5
12.2
11.2
41.7
9.8
Standard
Deviation
11.1
47.6
5.0
2.6
34.7
13.2
7.2
8.1
21.5
13.8
13.2
6.1
9.5
9.6
5.7
5.9
32.9
5.4
37
-------
Table 4-4
Percent Of NMOC/NOX Ratios Within Specified Range
(1985 Data)
City
Baton Rouge', LA
Beaumont, TX
Boston, MA
Cleveland, OH
Clute, TX
Dallas, TX
El Paso, TX
Fort Worth, TX
Houston, TX
Kansas City, MO
Lake Charles, LA
Philadelphia 1, PA
Philadelphia 2, PA
Portland, ME
Richmond, VA
St. Louis, MO
Texas City, TX
Washington, DC
Median
Ratio
14.9
53.2
7.6
7.5
24.6
11.8
11.9
11.8
12.9
8.5
23.7
6.5
9.5
11.6
11.2
9.6
28.7
8.7
Median
+_ 10%
22
11
21
32
15
32
41
19
16
13
26
25
7
19
29
26
9
25
Median
+_ 20%
33
29
33
54
31
43
57
31
28
37
44
45
18
35
53
40
16
46
Median
+_ 30%
54
40
56
74
43
53
74
54
38
50
61
55
47
42
76
58
24
67
Median
+_ 40%
61
52
65
81
51
63
78
65
43
65
75
62
65
63
84
67
43
73
Median
i 50%
79
61
77
89
60
78
79
74
57
77
86
71
71
73
90
78
50
79
Median
+_ 60%
82
73
84
94
65
82
87
87
68
83
88
82
75
81
93
84
60
85
38
-------
Table 4-5
Comparison Of 1984 And 1985 NMOC/NOX Ratios
Measured At The Same Site
City
Beaumont, TX
Clute, TX
Dallas, TX
El Paso, TX
Fort Worth, TX
Kansas City, MO
Richmond, VA
Texas City, TX
Washington, DC
Year
1984
1985
1984
1985
1984
1985
1984
1985
1984
1985
1984
1985
1984
1985
1984
1985
1984
1985
n
45
62
52
17.
69
68
60
76
58
54
61
84
62
58
52
68
54
48
10th
Percenti le
13.8
28.6
13.4
11.7
11.0
6.8
10.5
9.1
6.9
6.6
6.4
4.6
6.1
7.4
20.2
12.9
5.1
5.3
90th
Percentile
46.5
123.8
58. R
77.8
28.1
29.2
31.5
21.2
20.2
20.6
19.5
15.3
16.9
15.2
79.8
91.2
14.8
14.2
Median
25.3
53.2
23.7
24.6
16.0
11.8
15.2
11.9
11.5
11.8
9.9
8.5
10.5
11.2
37.7
28.7
9.3
8.7
Mean
30.8
69.1
33.7
37.6
17.9
15.2
19.8
14.3
13.4
13.1
12.2
11.2
12.1
12.2
54.3
41.7
10.6
9.8
Standard
Deviation
22.2
47.6
28.9
34.7
7.7
13.2
16.3
7.2
9.6
8.1
8.3
13.8
7.2
5.7
66.5
32.9
7.7
5.4
39
-------
Table 4-6
Median NMOC/NOX Ratio Increases Or Decreases
Measured At The Same Site
City
i
Beaumont, TX
Clute, TX
Dallas, TX
El Paso, TX
Fort Worth, TX
Kansas City, MO
Richmond, VA
Texas City, TX
Washington, DC
Direction Of Median
NMOC/NOX Ratio Change
h
t
4-
4-
t
4-
t
4-
4-
% Change In
NMOC/NOX Ratio
110%
4%*
-26%
-22%
3%*
-14%
7%*
-24%
- 6%*
Median
(1984-85)
* Not significant at the 95% confidence level.
40
-------
Table 4-7
Median NMOC/NOX Ratio Vs. Ozone Level
Dallas 1984
Ozone Level
All Data
1 -1
2. -11
1 .12
_> .13
1 -14
_> .15
>_ .16
1 -17
> .18
Number of
NMOC/NOy Ratios
69
32
26
18
13
9
9
6
1
1
Median
NMOC/NOV Ratio
16.0
15.0
15.0
16.1
17.0
17.0
17.0
14.9
17.0
17.0
41
-------
Table 4-8
Comparison of Median NMOC/NOX Ratios
(Historical vs Recent)
Past
City
Boston, MA
Cincinnati , OH
Cleveland, OH
Houston, TX
Philadelphia, PA
St. Louis, MO
Washington, DC
Analyses*
NMOC/NOX
Ratio
9.4
3.9
5.8
5.8
8.2
6.0
8.0
1984 Data
NMOC/NOX
Ratio
9.1
~
~
~
9.3
1985 Data
NMOC/NOX
Ratio
7.6
7.5
12.9
6.5/9.5
9.6
8.7
*From 1982 Ozone SIP Data Base Status and Summary Report,
Air Management Technology Branch, February 1983.
42
-------
43
-------
CN
O
o
5
cr
Dl
OS
03
o
\
in
to
\\NK\X
Sv\\
<\xv
\
x
\\
\
V. X \.
s-
\\
\ X
s-
s-
s-
s-
ipesg
X-
\\ ^ST
\\
\-
ST-[odBueTpui
puouiqoty
x ^ sr
13 SBSUB^
N
T^BUUTOUIQ
m
Cvl
o
CM
in
u
44
-------
45
-------
Kf
1
^
I
DC.
D
U
^^(i
LL
1985 MEDIAN NMOC/NOx RATIOS
s^
i
N\\\>
\\\\
\\\\
N\\\
\\
X
Od
1
I
X \ \ V x x x X
\ \ \ \ ^uouinBag N
XXXXxv xv
xx. N/XA^TQ SBX9X >
XXX x 1 'x V V V
\\\\x
\x x X
v N sanxo >
X X X X
\ \\S3T-IBMO s^Bq x
"X. ^<. 1 x x x X
x \xxxx
^
^
x x x x
\ X X X
X "> "^ x
x. OSBd 13 s
\ V X X
H^JO^ ^JtOj v
^ I X V V V
^
\
V
[^TQ sesuB)!
UO^SO
pUBXSA3-[
BTqd^ape^i
X x x X
x SBTiea S
X X X X
X X X X
v pUBT^ JOJ v
X X V X
X X X X
V \ V >.
x x X X
x v \ *V
\XxS>x>
\\\x
\xs>
3 \vv
0 xX^X^
-Md \N^
o
o
(0
O
lO
O
"fr
46
-------
Figure 4-5
O
01
x
O
O
O
CO
Ul 2
O
5
D
CL
O
CL
i
o
o
n
i
o
CM
I
O
Z
O
CD
UN-ii
o
XON/OCWN
47
-------
5.0 Carbon Bond Splits For The Carbon Bond-4 Mechanism
EKMA contains a procedure for considering the reactivity of ambient
NMOC. This is done by subdividing measured NMOC into various groups. The
exact groupings depend on the chemical mechanism employed by the model. One
commonly employed mechanism is the CR-4 mechanism. This mechanism requires
i
carbon fractions for eight carbon bond groups: paraffins, ethylene, olefins,
toluene, xylene, formaldehyde, higher aldehydes (ALD2) and unreactive.4
5.1 1984 Carbon Fractions
Two hundred and three (203) samples were analyzed to determine the carbon
fractions for 1984. Table 5-1 lists the median carbon fractions for each
city, based upon GC analyses. At the bottom of the table, an overall median
of the 1984 data and the currently recommended default values are listed.5
Default values for the EKMA model are numbers derived from data for numerous
cities and can be used for a city where these data are missing. These defaults
are based upon GC analyses of NMOC samples.
In general, the overall median values agree with the default values,
with the exception being that the 1984 samples show consistently higher
carbon fractions of paraffins and consistently lower carbon fractions of
unreactives. EPA plans followup modeling runs to determine the impact on
estimated VOC control levels due to varying carbon fractions.
5.? 19R5 Carbon Fractions
Three hundred and four (304) samples were analyzed to determine the carbon
fractions for 19R5. Table 5-2 lists the median carbon fractions for each
city, based upon GC analyses. At the bottom of the table, an overall median
of the 1985 data and the currently recommended default values are listed.
48
-------
In general, the overall median values agree fairly well with the default
values. The median of the 1985 samples show higher carbon fractions of
paraffin and higher aldehydes and lower carbon fractions of unreactives than
the current default values.
5.3; .Comparison of 1984 and 1985 Carbon Fractions
Ten cities were involved in both the 1984 and 1985 NMOC studies. Carbon
fractions for both years are compared for each city in Table 5-3. In general,
the carbon fractions do not vary significantly from 1984 to 1985. Beaumont,
Clute and Texas City appear to be exceptions. All are small cities which may
be impacted by major industrial sources.
5.4 Selection of New Default Carbon Fractions for CR4
In Table 5-4 the median carbon fractions for 1984 and 1985 are listed
along with the current default values. Data sets for 1984 and 1985 were
combined to obtain an overall median set of carbon fractions. From this data
set values for Beaumont, Clute, Texas City, West Orange and Lake Charles were
deleted since it is likely that these small cities are heavily influenced by
a particular industry and are not representative of values which would be
seen by larger cities.
The default values selected (shown in Table 5-4) are based upon 1984 and
1985 with the cities excluded as explained above. In any event there is
little difference between the overall data set and the "refined" data set.
The new default carbon fractions for use in EKMA CR4 analyses are those shown
at the bottom of Table 5-4.
49
-------
CO
r
CO
re
et
CC
I
LjO
re
4->
re
«- oo
o en
to c
c o
O Q.
o -o
re cu
(- CO
u_ re
CO
c ^^ *
o
.e
c
re
at
a;
O
re
c:
OJ
CO
cu
t- -o
»>,
CT CU
e- -O
Z r-
Lf5
CV CO
c o
c r-.
^~4
<:
C£
zc
c^ *
CD
*4->
C C
C- r-
,xx 4-3
iJO
CO *S"
I-H O
Lf) OO
o o
t-H IT)
Lf) ^J"
o o
o o
CM CO
0 0
CM 00
CM CO
vc in
^* cs
1C rv, ^j-
CO CM CO
£2 £
c^ o*1* cj*i
r**^ cc cvj
IT) LO VO
CO OO CO
C C C
VCCM^-
z.
I :r
C-5 O
z. «
re «
* O) *r~
4J O | *
4-> o re
4-> C C
o re c
C .(-> ^J
re re c
f~ f *^
0 C_5 0
«d- O
^D 00
^H O
^C CO
c o
0 C
o o
o o
CM CM
0 C
01 C
^f O
O"! 1"^
01 C
CTt C^
LT) t£>
CO CO
o c
0«
X
X H~
(_
** (/)
OJ *O
4-3 r-"
3 r
r TO
C_> C
VO CM
r^« CTi
o o
^d" ^"
0 0
o o
o o
ID C
CM CV
0 0
I-H CM
s m
1 *
^H C
«5 CV
o^ ^^
in (£:
O^^
»^
CO CO
o c
ccc^
X
1
X «
H~ --CT
1 »
* c_
o o
re
CL 4J
c_
, 0
U-1 U.
cn .-i
tc oo
o o
CO CO
o o
0 0
o c
0 0
CO CV
o c
CO i I
£ ^
CO .-<
«-< o
^ ^
00 CVJ
tr> u:
/T^ *""^
W I ^v
CM CO
o o
Cr-H
z
o
y
A
CO «
"- >>
r- +>
fY, ^^
*o
C! t/>
ro ro
r 0
OC f~H
u-> i£
1_O ^^
U ) \^J
CO CM
o o
coco
z.
h-
^ J
* 1-*-
t/)
F *
^~ !»
j^v ^
E. re
^ 2;
^-1 Lf) r- 1
vo vo m
O r-l O
co cvj ^~
o o o
o o o
o o o
O O Lf!
CO CM CM
o o o
CM CV .-H
r^ CM I-H
i-t 1^- CM
en o> vo
vc cvj co
LT ^C Lf)
O. (>
^^ t ^
CO CO CO
o c: o
ocoo
11
X O
1 C
«t
> . 0
* 1 % O
^3 *^ 4-^
C O Ol
o c
E co -i-
O X CO
r- cu re
cer t S
Lf! i i cr
CO LO f^»
rH O O
CO CO CO
o o o
o c o
o o o
o o o
CO CO CO
o o o
*d- r-- en
r^ <£; en
£ -10
r- en «*
r__| QQ ^^
<£> LO VC
_^. < » *-^
^"i ^_^ ^_^
cv co co
O O C
cccen
^^
1
u.
ft
x jc <:
t o ca-
re
* QJ A
0) CD 0)
CD C.
c E c.
re r~ re
t- re cc
f~*. r\
CO
4* 4* CU
CO CO i
o aj -i
3 3 S
CM
r~ in
O < I
<£>
CO «*
o o
t-H
LT>
O CO
O
o
CM CM
0 C
CO
CM O
I-H r-H
CO
t-H CV
1-H t-H
* *
ur>
O I-H
ic in
CO CO
0 C
CO
ex
OJ
3
t
*HJ
OJ
c
c
r 4-> CU
r- c E
re cu E
c_ c_ o
CO £-0
> 3 OJ
C O CC
Jr
4->
CO
3 t-
. O
CL C
c c,
O "H
i- U_
(J O.
re
c- 01
re j=
o <-
i. CU
3
= re
^
re 4->
c o
01 re
CO <4-
t. cu
CL-C
CO 4->
o
c
o cu
i- 3
O
re co
t_ M-
co
CO -I-
%£.
-C
QJ
"C C
re !-
4->
o> re
JC t_
re
T3 C
C O
re !-
4->
CU -r-
^ -o
>>T3
jz: res
O)
"^3 TJ3
r CU
re T3
E C
t- O)
o e
<4- E
Q
0) o
jc a;
t i_
*
cu
T3
JC
cu
u
re
cu
e
3
CO
re
cu
E
"U
O
JC
4->
CU
E
t_2
c^
50
-------
CD
O
(O
c_
c
0)
0)
to
ec
0
UJ
CM
I
in
>
jn jz
01 0)
OJ
o
IO
E
O
ooO'-«3-crvoccMoj o
*j-{Oir>^-«d-ir>«J-Lr>Lr>«d-*j-ir)ir>voir>*3-«!i-LO*3- in co
ooooooooooooooooooo o o
coc\jcocvjcsjCMco^rev/CMcvcMCA;cx'cocMroooir> CM
OJCMCMCMCMCMCVICMCMCVJCMCMCMCMCMCMCMCV4CM CM CVJ
ooooooooooooooooooo o o
O O <<«I O «ititi
r-~ in
r^ O O
O ii «i
O O «itlOiiiiti O <i O «-H «<<i i~-i i»O«IO
CVJ
ooooooooooooooooooo
CO
cr
in
CO
X O
O) I X
o> xi
I «< iI ii iI tI O
CO
«C ct
«o »ooc j=«/l3OO'C'O
c o
O) >O
tn U-
TD
u
10 in
(- <-
«4-
tn
in -r-
ai
O) -i- t-
TJ -D 3
10 10
ai a>
C -C E
i ai
10 T3 T3
ECO
£_ OJ -C
O E -U
<*- E a>
O E
O) U
J= <1J O
-------
ai
o
CO
I
If)
>
H)
cc
o
UJ
c_
o
+J
o
(O
c
o
JD
1C
o
OJ
o
V)
(O
Q.
O
O
a;
oc co
o o
VO O1
r^ cr
o o
CM O
CO
C O
1I 01
cc CTI
o o
-H CO
<& r-
o o
Lf) CC
VC CTi
1-1 O
r-H CO
un r^
c o
co co
CM ^f
ur cvj
o o
CM O
O rH
C CM
co co
c o
o cr
«d- co
o o
r-l CO
3- CO
o o
O r-l
co *d-
o o
c cr-
co co
o c
l-H CTl
CM cu
o o
o o
vc VD
co CM
o c
m co
o o
o o
^j- «*
o o
C CO
1C LO
c o
o cr>
IT) «3-
O O
O LO
ID Lf)
O O
C VO
ID *
O O
C ^f
in in
c o
C 00
^1- ^f
o o
C CM
1C Lf)
o o
C CM
«3- «ej-
O O
O CM
CM CM
O O
O CM
CM CM
O O
O CM
CVJ CM
O O
Lft CO
CVI CM
C O
o «*
CM CM
O O
O CM
CM CM
O C
o co
CO CM
C O
C co
CM CM
c o
10 co
CM CM
o c
C Lf)
CO CM
c o
CM O
vc vo
c o
a> cc
CTi Lf!
o o
C ID
CM O
I-H o cvi co
CM iI iI CM
-H O
CV; CO CM O
vc I-H r- cc
I-H II CO
r-l C
CO «*
c r-
.-H O
CM r-H
1C t^
o o
«* CTl
o- r-
o o
C CM
CTi
o
oo oo
CT cr
cr cc
co co
i-H CM
00
co cr.
oc
co
CO Lf)
O I-H O VC
«d- LO
cc cc
T± LO
CO CO
«* Lf)
CO 00
«* IT)
CO CO
^J- LO
00 00
«3- Lf)
CC 00
«i- Lf)
OC 00
^i- Lf)
OC 00
«tf- LO «d- If)
00 00 OC OC
X X
o o
X X
X X
X X
X X
+-> -I-
in
c c
o o * *
E e cu cu
^ 3 L .* -|i-*
(X3 'O 2 2 ^^ r~
cu cu i i to to
CE CC O O O O
o o o o
en in 33
-1-1 -4->
I »f
o o
r- < D- Q-
tn in
fQ fO
tn in
c c
C C
o o
o o
X X
IJ !->
^- -r
O C_5
in in
03 fo
X X
cu a;
o o
X X
i * CU CA
C C CD CD
o o c c
-(-> *J «3 (O
cr- CD c_ t_
C C CO
52
-------
o
LU
CM cr>
r*- oc
o o
in
CM 00
oc r-
o o
00
CO
c
CO
c
«3- LO
co co
o c:
Lf>
CO
o
CM
-< O
Lf) Lf) CO
O C O
LO
o
c
LT:
c
o
LO
C
CO
en
o;
C
O CM
CM CM CM
ceo
CM CM
CM CM
O O
CM
CM
O
c
ro
CO
CO
CM
LO
O
CO
CO
CM
CO
CM
i
Lf>
CU
en
c
o
O
CO
o
.o
ce
Q.
Lf)
O
(£>
oo
Lf>
1C
O
1C
00
cr
Lf!
CO
cr
IT)
-------
6.0 Mobile Source Contribution
Studies of ambient NMOC levels due solely to mobile sources have been
made in tunnels.a By comparing NMOC-acetylene ratios measured in tunnels to
NMOC-acetylene ratios in the ambient air, it is possible to estimate the
mobile source contribution to measured ambient NMOC levels. Acetylene is
i *
believed to be a good tracer of mobile source emissions since only a few
stationary sources emit acetylene (e.g., welding, carbon black plants).
Since carbon black plants and acetylene welding are not typically found in
significant amounts in urban areas, acetylene data for urban areas is probably
a good indicator of mobile source emissions. This technique of estimating
mobile source contributions is considered a fairly reliable procedure although
its accuracy has not be quantified. The equation is:
NMOC Acet
% Mobile Source Contribution = x x 100%
AcetTun NMOCAMB
An NMOC-acetylene ratio of 27.0 was used for the tunnel ratio.8
6.1 1984 Mobile Source Contributions
In Table 6-1, the estimated median mobile source contributions
(MMSCs) are listed for the 22 sites. The lowest MMSC (18%) was observed in
Texas City, Texas. The highest MMSC (88%) was observed in Miami, Florida.
The median overall MMSC is approximately 57%. Fifteen of the 22 sites had
MMSCs greater than or equal to 50%. Ten sites had MMSCs greater than or
equal to 60%. As expected mobile sources make a significant contribution to
6-9 a.m. NMOC levels.
6.2 1985 Mobile Source Contributions
The estimated median mobile source contributions (MMSCs) for the 19
sites are listed in Table 6-1. The lowest MMSC (7%) was observed in Beaumont,
54
-------
Texas. The highest MMSC (96%) was observed in Washington, DC. The median
overall MMSC is approximately 66%. Twelve of the 19 sites had MMSCs greater
than 50%. Eleven sites had MMSCs greater than 60%.
One fact that must be kept in mind is that the mobile source contribution
is dependent upon where the NMOC sites are located. For example, the 39%
median mobile source contribution for Houston would appear low. However, the
NMOC site was not a center city site, but was located in an industrial area
near the ship channel. The MMSC in Table 6-1 for Houston (39%) is probably
not representative of center city conditions.
The MMSCs should only be considered a rough estimate for two reasons:
(1) the NMOC-acetylene ratio for the tunnel was based upon 28 samples with
some scatter, and (2) the vehicle fleet mix and therefore the relationship of
NMOC to acetylene will vary from city to city.
6.3 Comparison Of 19R4 And 1985 Mobile Source Contributions
MMSCs are available for 11 sites which operated in both 1984 and 1985.
Of these 11 sites, estimates of MMSCs increased at eight sites and decreased
at the other three. In most cases, the differences between 1984 and 1985 are
not large. There are two exceptions, however. The MMSC for Beaumont was 25%
in 1984 and 7% in 1985. This significant decrease in percentage may indicate
that the activity at major industrial sources have increased from 1984 to
1985. NMOC and NMOC/NOX ratios from 1984 to 1985 in Beaumont support this
observation.
The MMSC for Kansas City was 57% in 1984 and 83% in 1985. This may be
an indication that industrial emissions decreased from 1984 to 1985. NMOC
and NMOC/NOX ratios also decreased from 1984 to 1985 to support this observation
Changes in MMSCs from 1984 to 1985 may also be due to varying meteorology,
local economic conditions or v"OC control programs.
55
-------
6.4 Mobile Source Contributions From Emission Inventories
In Table 6-2 median highway vehicle contributions for seven cities is
compared with the percent of the VOC emission inventory used in the State
Implementation Plan that was reported from highway sources. In all cases the
MMSC from ambient data is higher than the percentage obtained from the emission
inventory. This suggests that mobile source emissions may have been under-
estimated in the past. This hypothesis is supported by the recent findings
that tampering and evaporative emissions are considerably higher than believed
at the time 1982 ozone SIPs were developed. On the other hand, implementation
of industrial controls since 1980 may have resulted in higher mobile source
contributions in 1984-85. Both of these situations would lead to higher
MMSC's from the emission inventory. It should also be noted that the MMSCs
from ambient NMOC data represent 6-9 AM averages for summer conditions, a
time of peak traffic. MMSCs from emission inventories are based upon daily
averages and do not represent 6-9 AM conditions.
56
-------
Table 6-1
Percent Mobile Source Contributions In 1984 And 1985 (Median Values)
1984 Mobile Source 1985 Mobile Source
City Contribution Contribution
I
Akron, OH
Atlanta, GA
Baton Rouge, LA
Birmingham, AL
Boston, MA
Charlotte, NC
Chattanooga, TN
Cincinnati , OH
Cleveland, OH
Houston, TX
Indianapolis, IN
Lake Charles, LA
Memphis, TN
Miami, FL
Philadelphia 2, PA
Portland, ME
St. Louis, MO
West Palm Beach, FL
Wilkes Bar re, PA
Beaumont, TX
Clute, TX
Dallas, TX
Fl Paso, TX
Fort Worth, TX
Kansas City, MO
Philadelphia 1, PA
Richmond, VA
Texas City, TX
Washington, DC
West Orange, TX
48
68
____
46
_--_
77
67
50
67
____
79
88
____
____
____
47
60
25
37
58
78
67
57
50
53
18
87
32
____
53
____
82
__-_
____
-.-...
70
39
____
31
_ _ __
___
69
49
63
_-.__
__ _ _
7
33
71
81
74
83
66
66
24
96
29
57
-------
Table 6-2
Comparison Of Mobile Source Contributions From
Ambient Data Vs. Emission Inventory Data
Percent of 1980 VOC
i
City
Roston
Philadelphia
Washington, D.C.
Cincinnati
Cleveland
Houston
St. Louis
Emi
Point
19
38
3
31
IB
60
35
ssion Inventory
Area Highway
35 46
30 32
31 66
28 41
40 45
14 26
38 27
Mobile
Source
Contribution
(From Ambient
NMOC Data)
82
50
66
69
87
96
50
70
39
63
Year
1985
1984
1985
1984
1985
1984
1985
1985
1985
58
-------
7.0 Biogenic NMOC Data
Natural sources contribute to ambient NMOC levels. These emissions are
usually from vegetative sources. The amount of ambient biogenic NMOC can be
estimated by adding two components from the GC analysis. These are isoprene
and a-pinene. It should be noted that the isoprene and a-pinene peaks
have been identified by retention time only and have not been confirmed by
other techniques such as GC/Mass Spec. This procedure provides a rough esti-
mate of biogenic NMOC values. While few industrial activities emit isoprene or
a-pinene, it is possible to misidentify other species as isoprene or
a-pinene. (As discussed later, this misidentification is believed to
have occurred with the 1985 Beaumont data). Biogenics emissions do include
species other than isoprene and a-pinene, such as 3-pinene, A^-carene
and myrcene. However, the procedure of identifying species by retention time
only is not accurate enough to reliably estimate these compounds. As such,
the estimates provided below should be considered as fairly rough estimates.
7.1 1984 Biogenic NMOC Pata
Table 7-1 lists the estimated contribution of biogenic sources for each
of the 22 sites. Included are the number of samples, range, mean, and median
of the biogenic concentrations, and range, mean and median of the percent of
biogenics compared to total NMOC. The median biogenic concentration is
displayed graphically in Figure 7-1. The lowest median biogenic concentration
(0.70 ppbC) was measured in Miami, Florida, while the highest median biogenic
concentration (9.87 ppbC) was measured in Richmond, Virginia. No obvious
geographical patterns are present.
Estimated biogenic NMOC concentrations expressed as a percent of total
NMOC range from a low of 0.11 (median) in Miami, Florida, to a high of 2.02
in Charlotte, North Carolina.
59
-------
7.2 1985 Biogenic NMOC Data
Table 7-2 lists the estimated contribution of biogenic sources for each
of the 19 sites. The median biogenic concentration is displayed graphically
in Figure 7-2. The lowest median biogenic concentration (2.93 ppbC) was
measured in Texas City, Texas, while the highest median biogenic concentration
(142.68 ppbC) was measured in Beaumont, Texas. No obvious geographical
patterns are present.
Biogenic NMOC concentrations expressed as a percent of total NMOC range from
a low of 0.35% (median) in Cleveland, Ohio, to a high of 10.17 in Reaumont, Texas.
It is obvious that the values for Beaumont do not reflect solely biogenic
emission levels. If the values are truly biogenic levels, such large changes
should not occur from one year to the next. The Beaumont biogenic level is an
order of magnitude greater than any other city. Some industrial component
must be included in either of the peaks identified as isoprene or a-pinene.
7.3 Comparison Of 19R4 And 1985 Biogenic NMOC Data
Table 7-3 lists the estimated median biogenic NMOC concentrations for 1984
and 1985 for the 11 cities which were analyzed for both years. Also shown is
the percent change in biogenic NMOC from 1984 to 1985 and the percent change
in total (mean) NMOC from 1984 to 1985.
The most obvious change is in Beaumont, Texas, where the median biogenic
concentrations went up by a factor of 31 from 1984 to 1985. The total NMOC
in Beaumont nearly doubled from 1984 to 1985. No reason for such a large
increase in biogenic NMOC concentrations has been determined and therefore
the levels identified as biogenic are questionable.
At eight other sites, the median biogenic NMOC concentration increased
from 1984 to 1985 and decreased at the other two sites.
60
-------
At 8 of the 11 sites percent biogenic increased from 1984 to 1985, while
total NMOC decreased at 7 of these 8 sites.
Table 7-4 lists the biogenic NMOC concentrations as a percent of total
NMOC for 1984 and 1985. Once again, Beaumont shows a large increase from
1984 to 1985.
I
With the exception of questionable estimates for Beaumont, most of the
estimates of biogenic contributions to ambient NMOC concentrations are less
than 1%. These data support the position that anthropogenic sources are the
predominant cause of ambient NMOC concentrations in urban areas.
61
-------
Table 7-1
Riogenic NMOC nata (1984)
Estimated
City
Akron, OH
Atlanta, GA
Beaumont, TX
Birmingham, AL
Charlotte, NC
Chattanooga, TN
Cincinnati, OH
Clute, TX
Dallas, TX
El Paso, TX
Fort Worth, TX
Indianapolis, IN
Kansas City, MO
Memphis, TN
Mi ami , FL
Philadelphia 1, PA
Richmond, VA
Texas City, TX
Washington, DC
West Orange, TX
West Palm Beach, FL
Wilkes Barre, PA
n
10
7
9
6
16
12
7
10
13
8
13
10
11
8
3
7
10
13
10
10
8
9
Concentration (ppbC)
Range Mean Median
0.68- 2.03
2.17-12.13
2.96- 8.04
3.44- 8.94
0.81-27.07
3.76-34.14
1.80-26.62
0.36- 4.87
0.56-11.19
0.48- 3.52
2.03- 7.50
0.56- 6.03
0.50- 7.62
3.93-12.82
0.51- 2.00
1.40- 5.08
1.71-13.07
0.27-23.99
2.50-11.66
0.83-11.00
1.93- 4.31
2.35- 9.55
1.31
5.80
5.18
5.89
9.95
11.04
8.97
2.64
4.06
2.08
4.88
2.25
2.86
6.38
1.07
3.28
8.06
4.89
5.40
4.77
2.95
4.67
1.30
4.41
4.51
5.55
6.79
8.85
3.53
2.75
3.85
1.95
5.03
1.76
1.87
5.55
0.70
2.88
9.87
3.02
3.84
3.66
2.91
4.56
% Of Total NMOC
Range Mean Median
0.15-0.52
0.44-1.29
0.17-1.08
0.38-1.34
0.53-3.63
0.16-1.59
0.27-1.20
0.03-0.90
0.21-1.19
0.09-0.56
0.30-1.16
0.14-0.31
0.15-0.50
0.15-0.77
0.11-0.23
0.19-0.96
0.49-3.26
0.04-2.06
0.42-0.87
0.27-1.63
0.18-1.46
0.55-2.30
0.29
0.82
0.70
0.70
1.91
0.62
0.54
0.53
0.59
0.25
0.56
0.23
0.29
0.42
0.15
0.44
1.88
0.55
0.57
0.85
0.82
1.02
0.27
0.78
0.71
0.55
2.02
0.53
0.44
0.56
0.57
0.24
0.43
0.25
0.31
0.39
0.11
0.38
1.72
0.44
' 0.51
0.82
0.97
0.74
62
-------
Table 7-2
Biogenic NMOC Data (1985)
Estimated
City
Baton Rouge, LA
Beaumont, TX
Boston, MA
Cleveland, OH
Clute, TX
Dallas, TX
El Paso, TX
Fort Worth, TX
Houston, TX
Kansas City, MO
Lake Charles, LA
Philadelphia 1, PA
Philadelphia 2, PA
Portland, ME
Richmond, VA
St. Louis, MO
Texas City, TX
Washington, DC
West Orange, TX
n
16
19
a
17
18
23
17
19
22
17
16
13
11
14
14
18
15
11
16
Concentration (ppbC)
Range Mean Median
0.65-22.76
51.40-313.42
1.50- 5.43
0.61- 9.32
0.23-17.70
1.04-24.84
1.37-10.45
1.10-16.45
0.73-30.23
0.28- 8.84
0.68-17.15
1.31-18.75
1.82-12.91
0.29-43.17
2.17-13.35
0.52-18.80
0.22-38.39
2.96-24.00
3.69-13.97
4.89
154.75
3.35
3.55
4.15
6.42
4.29
6.09
9.64
4.18
6.88
4.86
5.07
9.61
6.88
5.15
4.86
6.98
7.96
2.98
142.68
3.43
3.04
3.29
4.20
3.20
6.54
6.36
3.73
6.13
4.13
4.06
4.45
6.64
4.01
2.93
4.97
7.45
% Of Total NMOC
Range Mean Median
0.17-2.09
1.21-16.65
0.19-2.14
0.14-0.88
0.16-2.26
0.17-1.93
0.17-1.26
0.18-2.24
0.11-3.41
0.10-1.66
0.32-1.89
0.31-0.84
0.38-0.83
0.06-4.00
0.19-2.49
0.22-2.22
0.03-1.43
0.54-2.09
0.33-7.45
0.72
9.92
1.25
0.39
0.62
0.80
0.56
0.89
1.06
0.67
0.87
0.51
0.55
1.29
1.27
0.78
0.63
1.00
1.94
0.60
10.17
1.35
0.35
0.40
0.75
0.53
0.87
0.73
0.66
0.72
0.44
0.54
0.89
1.29
0.68
0.43
0.90
1.47
63
-------
Table 7-3
Comparison Of Riogenic NMOC Data (1984 And 1985)
Estimated
Median 1984
City Biogenic NMOC
(ppbC)
Beaumont, TX
Clute, TX
Dallas, TX
El Paso, TX
Fort Worth, TX
Kansas City, MO
Philadelphia 1, PA
Richmond, VA
Texas City, TX
Washington, DC
West Orange, TX
4.51
2.75
3.85
1.95
5.03
1.87
2.88
0.87
3.02
3.84
3.66
Median 1985
Biogenic NMOC
(ppbC)
142.68
3.29
4.?0
3.20
6.54
3.73
4.13
6.64
2.93
4.97
7.45
% Change
Riogenic
NMOC
3064% +
20% +
13% t
64% +
30% t
99% t
43% t
33% 4-
3% 4-
29% t
104% t
% Change
Total Median
NMOC
+
4-
4-
1
4-
4-
4-
4-
4-
4-
4-
117%
10%
18%
18%
21%
34%
47%
10%
46%
15%
20%
64
-------
Table 7-4
1984 And 1985 Biogenic NMOC Data
(Expressed As A Percent Of Total NMOC)
City
;
Beaumont, TX
Clute, TX
Dallas, TX
El Paso, TX
Fort Worth, TX
Kansas City, MO
Philadelphia 1, PA
Richmond, VA
Texas City, TX
Washington, DC
West Orange, TX
Median 1984
Riogenic %
0.71
0.56
0.57
0.24
0.43
0.31
0.38
1.72
0.44
0.51
0.82
Median 1985
Biogenic %
10.17
0.40
0.75
0.53
0.87
0.66
0.44
1.29
0.43
0.90
1.47
65
-------
66
-------
67
-------
8.0 NMOC Versus Ozone
This chapter describes a case study which compares reductions in ambient
NMOC levels with ozone data. The city selected for analysis is Kansas City,
Missouri. As discussed in Chapter 2, median NMOC levels dropped 34% from 1984
to 1985.i Did ozone levels also decrease and, if so, how much? Five ozone
sites were analyzed. These were 2fi0840002F01, 261020003F01, 261020005F01,
262380023H01, and 2fi2380025H01. These sites are referred to as 2F01, 3F01,
5F01, 23H01, and 25H01 throughout the remainder of this chapter. One other
site was considered but not included in the analysis (171800001F01) since this
site located in Kansas City, Kansas, missed considerable ozone data in June
and August 1985.
Figure 8-1 shows the distribution of NMOC data in Kansas City for 1984
and 1985. The data for 1984 are displayed above the zero line, while 1985 is
plotted below the zero line. The distribution of 1985 NMOC has shifted
substantially to the left, indicating lower NMOC levels in 1985 than in 1984.
The hypothesis that mean NMOC levels were lower in 1985 than 1984 was tested
(see Chapter 2) and found to be true at the 95% confidence level. The mean
NMOC level dropped 33% from 1984 to 1985.
In Figure 8-2 the distribution of ozone at the 2F01 site is shown. The
data analyzed included June through September for 1984 and 1985. Once again,
1984 values are shown above the zero line and 1985 values below the zero
line. The 1985 distribution has shifted to the left, indicating that ozone
levels have decreased from 1984 to 1985. The other four sites showed similar
results. This was further confirmed by performing t-tests, using mean ozone
levels. In each case, a reduction in ozone from 1984 to 1985 was indicated
at the 95% confidence level.
68
-------
Table 8-1 compares 1984 and 1985 ozone statistics for each site. The
mean ozone levels decreased from 7 to 23%. Maximum ozone levels increased 1%
at one site, hut decreased at the other four sites between 9 and 22%. The
number of hours with ozone levels above .125 ppm is also reduced at each
site.
i
It would appear that ozone levels decreased in Kansas City as NMOC
levels decreased. No attempt has been made to determine the cause of the
NMOC reductions.
69
-------
Table 8-1
Comparison Of 1984 And 1985 Ozone Data
(June-September Kansas City Sites)
2F01 3F01 5F01 23H01
25H01
Percent Decrease In Mean
Ozone (1984-1985)
7% 23%
18%
10%
15%
Maximum Ozone (ppm)
1984
1985
.149 .160
.116 .145
.167 .133 .115
.132 .114 .116
Percent Decrease In Maximum
Ozone (1984-1985)
22%
9%
21%
14%
t 1%
Number Of Hours With Ozone
>_ .125 ppm
1984
1985
1
0
4
1
5
1
3
0
0
0
70
-------
Figure 8-1
<
<
^^
Q
O
O
*> E
-^- =S
Zo
* in
to
U_ 2
O t
h-
O
en
O|
i <
r- 5
D
OD
rv
uL
\
1
( n
(J )
Q
C
C
:
-
D u
vl
T C
T
D u
_
-
,
=
=
-
C=B=
B=
^=c=
B=
0 C
K3
E3
^
1 1 1 C3
E3
D u
T C
T
D U
~ T
) c
c
D U
M C
") C
M V
N
^
CM
-------
Figure 8-2
UJ
Z
O
N
O
>-5
bfc
H
(/)
(/I
s
D C
0 C
L^ ^ V y
1^ X
D C
M
C
C ^ ^
r 'v
ez
D C
*
3
3
^VX'VXXNJ
\\\ \"\J
D C
T
t-XM
>.\l
D C
C
D C
NJ r
m
r-
o
(N <
_ r- Q
. 6 m
00
0)
^-\
^
u O til
z
0
KJ
0
' *° <
-°- 5
. O %i
Q
^
oo
Oi
T
" fO
- o p
^ l^
- o
D
0
72
-------
9.0 Conclusions
9.1 NMOC
1. 6-9 AM NMOC levels vary from day to day at a given site.
2. Median NMOC levels decreased from 1984 to 1985 in many cities.
Decreases.were statistically significant at five of the nine cities which
showed a decrease.
3. Continuous NMHC measurements and those made with the PDFID do
not agree. Continuous data are suspect.
9.2 NOX
1. 6-9 AM NOX levels vary widely from day to day at a given site.
2. Median NOX levels decreased in most cities from 1984 to 1985.
Decreases were statistically significant at three of the seven cities which
showed a decrease. Only one city showed a statistically significant increase
in median NOX.
9.3 NMOC/NOX Ratios
1. 6-9 AM NMOC/NOX ratios vary widely from day to day at a given site.
2. Further investigation is underway to determine whether use of
robust statistics (e.g., medians) or day specific data is most suitable for
characterizing a city's NMOC/NOX ratios.
3.- Median NMOC/NOX ratios decreased at five of the nine sites from
1984 to 1985 and increased at four. The reductions are statistically signifi-
cant at four of the sites, while the increases are statistically significant
only in Beaumont.
4. A comparison of NMOC/NOX ratios with ozone level indicates a poor
correlation. The relationship is probably overwhelmed by the importance of
variable meteorological parameters.
73
-------
9.4 Carbon Fractions
1. 1984 and 1985 carbon fractions match fairly well with current
default conditions, except that 1984 and 1985 fractions of paraffins are
slightly higher and unreactives lower. New default values are listed (in
Table 5-4).
9.5 Mobile Source Contributions
1. Estimated Mobile source contributions to ambient NMOC levels
varied from 7 to 96%. Contributions on the order of 45-75% appear most
common at the sites in the 1984-85 network.
?.. The mobile source contributions based upon 1984. and 1985
ambient NMOC data are significantly different from values determined from
emission inventories. Recause these ratios were determined in different ways
and for different time periods, there may be logical reasons for the differences.
9.6 Estimated Riogenic NMOC data
1. Excluding Beaumont, median Biogenic (natural) NMOC levels make
up between .1 and 1.5% of total NMOC. Most sites are on the order of 0.3 -
0.7 % biogenic of total NMOC.
9.7 NMOC Versus Ozone
1. A 34% reduction in ambient median NMOC levels from 1984 to 1985
was seen in Kansas City.
?.. Hourly mean ozone levels from June-September 1984 decreased
from 7 to 23% at the five Kansas City ozone sites.
3. Maximum hourly ozone levels decreased at four of five sites.
9.8 Additional Data Needs
1. Many cities showed substantial reductions in mean NMOC levels
from 1984 to 1985, but two points do not define a long term trend. It would
be interesting and useful to observe mean NMOC levels over many years to see
74
-------
if a trend exists. Unfortunately, NMOC ambient data are not routinely collected
from year to year. There will be few sites where long term trends can be
determined. In some cases, these will be small Texas cities that may not
represent "typical" size cities. States are encouraged to establish fixed
networks of NMOC monitors in large cities and operate them for several years.
2. Median NMOC/NOX ratios decreased significantly from 1984 to
1985 at nearly half of the sites. It does not appear reasonable for States
to use an NMOC/NOX ratio for a State Implementation Plan (SIP), unless it
were current (no more than one to two years old). It is recommended that
States sample NMOC concentrations for more than one summer to provide reliable
estimates.
3. Since 6-9 AM NMOC levels are significantly impacted by mobile
sources, States are encouraged to focus increased attention on mobile source
VOC emissions. Analysis of hourly traffic counts may also be necessary to
properly define mobile source contributions during the early morning rush
hour.
75
-------
REFERENCES
1. F. F. McElroy, V. L. Thompson, D. M. Holland, W. A. Lonneman and R. L
Seila, "Cryogenic Preconcentration-Direct FID Method for Measurement of
Ambient NMOC: Refinement and Comparison with GC Speciation," Journal of
the Air Pollution Control Association, p 710-714, June 1986.
2. G. L. Gipson, W. P. Freas, R. F. Kelly and E. L. Meyer, "Guideline For
Use Of City-Specific EKMA In Preparing Ozone SIPs," U. S. Enviromental
Protection Agency, EPA 450/4-80-027, March 1981.
3. G. Z. Whitten and H. Hogo, "User's Manual For Kinetics Model And Ozone
Isopleth Plotting Package," U.S. Environmental Protection Agency, EPA
600/8-78-014a, July 1978.
4. G. L. Gipson, "Guideline For Using The Carbon Bond Mechanism In City-
Specific EKMA," U. S. Environmental Protection Agency, EPA 450/4-84-005,
February 1984.
5. H. Hogo and G. Z. Whitten, "Guidelines for Using OZIPM-3 with CBM-X or
Optional Mechanisms-Volume 1," U.S. Environmental Protection Agency,
January 1986.
6. G. L. Gipson, "User's Manual for OZIPM-2: Ozone Isopleth Plotting With
Optional Mechanisms/Version 2," U. S. Environmental Protection Agency,
EPA 450/4-84-024, August 1984.
7. Written communication from Richard G. Rhoads, OAQPS, EPA to Regional
Offices, March 10, 1986.
8. W. A. Lonneman, R. L. Seila and S. A. Meeks, "Nonmethane Hydrocarbon
Composition in the Lincoln Tunnel," Environmental Science and
Technology, p 790-796, August 1986.
76
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO. 2.
EPA-450/4-86-015
4. TITLE AND SUBTITLE
A Review Of NMOC, NO And NMOC/NO Ratios Measured In
1984 And 1985
7. AUTHOR(S)
Keith Baugues
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Monitoring and Data Analysis Division
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
12. SPONSORING AGENCY NAME AND ADDRESS
3. RECIPIENT'S ACCESSION NO.
5. REPORT DATE
September 1986
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT NO
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
During the summers of 1904 and 1985, morning (6-9am) measurements of ambient
nonmethane organic compounds (NMOC) were collected at 22 and 19 urban sites,
respectively. The data were collected by State and local agencies which contributed
grant funds and personnel. The EPA managed the analysis and provided NMOC sampling
equipment. The method of determining NMOC levels was the cryogenic preconcentration
direct flame-ionization detection (PDFID) described by McElroy et a!. Data were
collected to provide input values for the Empirical Kinetic Modeling Approach (EKMA),
a computer program which estimates hydrocarbon control requirements necessary to
attain the National Ambient Air Quality Standard (NAAQS) for ozone. One of the key
inputs to EKMA is the Nonmethane Organic Compound/Nitrogen Oxides (NMOC/NO ) ratio.
Thus, a collocated NOV instrument was operated at each NMOC site.
A
17. KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
18. DISTRIBUTION STATEMENT
b. IDENTIFIERS/OPEN ENDED TERMS
19. SECURITY CLASS (This Report)
20 SECURITY CLASS (This page)
c. COSATI I-'ield/Group
21 NO. OF PAGES
82
22. PRICE
EPA Form 22201 (Rev. 4-77) PREVIOUS EDITION is OBSOLETE
-------
U.S. Environmental Prctectt5n Agency.
Region V, Library
230 South Dearborn S'
Chicago, !!!!no:s 60604
------- |