United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park NC 27711
EPA-450/3-78-042b
May 1978
Air
Emission Inventory/
Factor Workshop
Volume 2
-------
EPA-450/3-78-042b
Emission
Inventory / Factor Workshop
Volume 2
Moderator
James H. Southerland
OAQPS,
Monitoring Data and Analysis Division
Co-Moderators
Richard Burr
OAQPS,
Emission Standards and Engineering Division
Dale Denny
ORD,
Industrial Environmental Research Laboratory
Charles Masser
OAQPS,
Monitoring Data and Analysis Division
September 13-15, 1977
Raleigh, NC
Co-Sponsored by
Air Pollution Training Institute and Air Management Technology Branch
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Research Triangle Park, NC 27711
May 1978
-------
This report is issued by the Environmental Protection Agency to
report technical data of interest to a limited number of readers.
Copies are available free of charge to Federal employees, current
contractors and grantees, and nonprofit organizations - in limited
quantities - from the Library Services Office (MD-35), U. S. Envir-
onmental Protection Agency, Research Triangle Park, North Carolina
27711; or, for a fee, from the National Technical Information
Service, 5285 Port Royal Road, Springfield, Virginia 22161.
USEBV
This is not an official policy and standards docu-
ment. The opinions, findings, and conclusions are
those of the authors and not necessarily those of the
United States Environmental Protection Agency. Any
mention of products, or organizations, does not consti-
tute endorsement by the United States Environmental
Protection Agency.
ii
-------
FOREWORD
Emission inventories and emission factors are major components
of an air pollution control program. The inventory is perhaps
one of the most important planning tools available to an air
pollution control agency. Emphasis on these inventories and fac-
tors, the procedures used, and the use of the information has often
been lacking, however. On September 13-15, 1977, the Office of Air
Quality Planning and Standards hosted a workshop with both prepared
topics and open discussion in Raleigh, N. C. to focus attention to
some of the aspects of such emission inventory and factor activi-
ties particularly as related to the timely aspect of organic
emissions. This document constitutes the proceedings of that
workshop and will be distributed to the approximately 130 attendees.
Additional copies are available from EPA Library Services Office.
Papers prepared for and presented at the workshop have been
finalized by the authors and are included with no additional
editorial or technical modifications. Papers presented do not
necessarily represent policies of the Agency but may provide a
basis for development or discussion of such policies. The workshop
also provided a forum for various criticism which may appear to be
unanswered but hopefully helped to create an open atmosphere
conducive to constructive change.
The discussions during and following the papers were condensed
and edited to the extent possible. Many important discussions may
have been left out due to inadequate clarity of the recording and
-------
transcription. Some of the topics of discussion may also have been
more clearly addressed by the authors when the final manuscripts -v
were prepared. It was felt, however, to be worthwhile to include
the condensed discussions to indicate where the attendees felt
emphasis or clarification were needed.
Following this workshop the A±r Pollution Control Association's
(APCA) newly formed committee; TP-7; Emission Factors and Inven-
tories, developed plans for an APCA Specialty meeting on Inventories
and Factors which will be held in Anaheim California the week of
November 13, 1978, and hosted by the West Coast, APCA Section. Pa*tici-
pants at this workshop are especially invited to submit papers for
possible presentation at the meeting in California and/or be present
to participate in the discussion. It has been suggested that the
concept of a forum for this general topic become an annual under-
taking of EPA and/or APCA. Discussion of this point and general
comments on the content of this document or the need for an annual
conference of some sort can be addressed to the Office of Air
Quality Planning and Standards, Environmental Protection Agency,
Research Triangle Park, N. C. 27711. More detail on specific
papers would best be obtained by directly contacting the author(s).
As prime moderator of the workshop, I would like to express
my thanks to the Air Pollution Training Institute and their
contractor Northrop Services, Inc. who provided the arrangements,
taping, transcription, and related work that made the workshop
iv
-------
possible. Especially, I would like to thank the authors, co-
moderators and attendees for their hard work and participation
which made the workshop, I feel, to be a success.
James H. Southerland, Moderator.
-------
Table of Contents
Emission Inventory/Factar Workshop
Volume 1
Foreword iii
1 Organic Emission Inventory 1-1
Considerations and Purposes to
C.P. Bartosh, W.J. Moltz, and B.P. Cerepaka 1-14
2 Analysis of Data for Hydrocarbon 2-1
Sources in Non-Attainment Areas to
in Louisiana 2-40
B.C. McCoy and K.J. Guinaw
3 Documentation of Emission Inventories 3-1
in Region IX to
B.C. Henderson 3-14
4 Methodologies and Problems Encountered 4-1
in a Level 3 Multi-State/County Hydrocarbon to
Area Source Emissions Inventory 4-26
T.A. Trapasso and W.K. Duval
5 Air Force Emission Inventories 5-1 to
B.C. Grems 5-14
6 A Format for the Storage of Area 6-1
Source Emission Data to
S.R. Tate, N.L. Matthews, D.J. Ames 6-35
R.A. Bradley
7 Maryland Special Factors and 7-1
Inventory Techniques to
E.L. Carter and J.W. Paisie 7-22
8 Panel Discussion of Inventory 8-1
Methodology Procedures and Applications to
to Oxidant Control 8-26
vii
-------
9 Hydrocarbon Emissions from Households 9-1
in New York and New Jersey to
E.Z. Finfer 9-20
10 Hydrocarbon Carrier Emissions 10-1
from Atmospheric Dye Becks to
R. Hawks 10-25
11 Air Quality and Energy Conservation 11-1
Benefits From Using Emulsions to to
Replace Asphalt Cutbacks in Certain 11-21
Paving Operations
F.M. Kirwan
12 Commercial Bakeries as a Major 12-1
Source of Reactive Volatile Organic to
Gases 12-18
D.C. Henderson
13 Reactive Organic Gas Emissions 13-1
from Pesticide Use in California to
F.J. Wiens 13-53
Emission Inventory/Factor Workshop
Volume 2
14 Volatile Organic Compound Emissions 14-1
From Architectural Coatings to
R.A. Friesen, R.E. Menebroker, 14-12
D.R. Saito
15 NO Reducations in the Portland Cement 15-1
Industry with Conversion to Coal-Firing to
R.J. Hilovsky 15-26
16 Current API Emission Measurement 16-1
Programs to
J.G. Zabaga 16-26
17 Hydrocarbon Emissions From 17-1
Floating Roof Storage Tanks to
R.L. Russell 17-22
18 Emission Inventory of Petroleum 18-1
Storage and Handling Losses to
(A Case History) 18-29
J.T. Alexander
viii
-------
19 Inventorying Hydrocarbon Emissions 19-1
From Small Gasoline Bulk Plants to
R.L. Norton and R.J. Bryan 19-39
20 An Organic Specie Emission 20-1
Inventory for Stationary Sources to
in the Los Angeles Area - Methodology 20-49
H.J. Taback, T.W. Sonneschen,
N. Brunetz, and J.L. Stredler
21 Highway Motor Vehicle Emission 21-1
Factors to
Motor Vehicle Manufacturers 21-80
Association of the United States, Inc.
22 FTP Emission Factor Development: 22-1
Correction for Non-FTP Conditions to
J. Becker and M. Williams 22-34
23 Land Use Based Emissions Factors 23-1 to
F. Benesh and T. McCurdy 23-27
24 Emission Rates for Biogenic NO 24-1 to
H.C. Ratsch and D.T. Tingey X 24-28
25 Procedures for Conducting Hydrocarbon 25-1
Emission Inventories of Biogenic to
Sources and Some Results of 25-32
Recent Investigations
P. Zimmerman
ix
-------
VOLATILE ORGANIC COMPOUND
EMISSIONS FROM ARCHITECTURAL COATINGS
Prepared by,
Ronald A. Friesen
Raymond E. Menebroker
Dean K. Saito
Presentation Given at the
Emission Inventory/Factor Workshop
Sponsored by the
Environmental Protection Agency
September 14, 1977
14-1
-------
ABSTRACT
By using the data in this paper, estimates can be made of
Volatile Organic Compound (VOC) emissions from the use of
architectural coatings and associated solvents. These emissions
contribute significantly to air pollution in California. In an
effort to reduce these emissions, the staff of the California Air
Resources Board (CARB) developed a model rule to regulate the
solvent content of architectural coatings. Essential to the
development of the model rule was a data base. This data base was
established by use of responses to questionnaires that were
mailed to coatings manufacturers. These responses provided data
on the sales volume of coatings and solvents and the VOC content
of these coatings and solvents for a given year, 1975. With
this data, the CARB staff determined the level of VOC emissions in
California for 1975. In addition, by dividing the level of VOC
emissions by the 1975 population of California, the CARB staff has
devised a method to estimate VOC emissions when sales data are
not known. Thus, two methods of computing VOC emissions are given
in this document, and for each method, the effect of the model
rule is shown.
14-2
-------
VOLATILE ORGANIC COMPOUND
EMISSIONS FROM ARCHITECTURAL COATINGS
by R. Friesen, R. Menebroker, D. Saito
Introduction
Little information has been generated with regard to emissions
of Volatile Organic Compounds (VOC) from the use of architectural
coatings. The staff of the California Air Resources Board has
conducted an extensive inventory of such emissions in California,
and the data gathered are presented herein. Using the data contained
in this paper it is possible to estimate emissions from the use
of architectural coatings with varying degrees of accuracy
depending on the preciseness of information available. Such
information includes population, total gallons of architectural
coatings sold, or gallons of architectural coatings sold by coating
category.
The methods of computing VOC emissions, presented here, are
preceded by brief explanations of the research involved and the
background of the study.
Background
Volatile Organic Compound emissions from the use of architectural
coatings are a significant contributor to the air pollution problem
in California. Based on this fact the California Air Resources
Board (CARB) embarked on a program to control such emissions.
14-3
-------
As part of this program, on July 7, 1977, CARB approved a model
regulation to control emissions of VOC from the use of
architectural coatings. The data base upon which an architectural
coating rule could be developed, however, was nonexistent prior
to CARB activity in this area. The first step in developing a
rule, therefore, was to establish a sound data base. Recognizing
this need, CARB staff developed a questionnaire that would ascertain
the volumes of various waterborne and solvent-borne coatings sold
in California and the VOC content of each. The VOC was requested
by average and maximum for each coating category for the 1975 calendar
year.
Industry Survey
The questionnaire was mailed to 446 potential paint manufacturers,
260 of which responded. Of these, 122 completed the questionnaire,
and 138 reported that they did not manufacture architectural
coatings. The responses received indicated that about 50 million
gallons of architectural coatings were sold in California in 1975.
A national sales survey prepared for the National Paint and
Coatings Association indicates that about 435 million gallons of
coatings were sold throughout the United States and that
approximately 10 percent of the nationwide total, or 43.5 million
gallons, was sold in California. A comparison of the two estimates
indicates that the CARB survey is reasonably complete. The
results of the CARB conducted industry survey are presented in Table I.
14-4
-------
Table 1
Architectural Coatings Sales and VOC
Emissions In California 1n 1975
COATING
TYPt
Interior Wall
Exterior Hall
Aerosols
*
Roof
H*HO»HH^^V^^^^B.^^4^».^^^HMHMMM^«M
Clears
Stains
Primers, Sealers
and Undercoats
High Performance
Topcoats
Traffic
Sash, Trlfii
and Trellis
Metallic
Pigraented
Porch, Deck
and Floor
High Performance
Primers
Barn and Fence '
Mobile Home
Hot Classified
Subtotal
Other Solvent**
Total
1975,
SALES (103 GAL)
Solvent
Borne
2572
1694
1940
2316
^^^^^^^b^B^BM
1820
1368
1617
1678
1604
1182
492
498
294
100
48
1707
20930
?fi7Rl
Water
Borne
11445
8375
-0-
2354
18
1205
1084
27
44
891
3
339
15
298
4
2604
29216
-0-
M»1R
EMISSIONS (TOHS/DAY)*
Solvent
Borne
12.0
7.1
13.4
10,1
-9.4
8.0
7.0
7.3
7.1
4.6
2.6
2.5
1.4
0.4
0.2
8.2
101.3
69.0
170.3
Water
Borne
5.4
5.4
-0-
-0-
^^^^^^^K^M^WtBBVB
0.008
1.3
.69
0.03
0.03
1.3
.001
0.1
0.007
0.1
0.005
1.5
15.9
-0-
15.9
Total
17.4
12.3
13.4
10.1
^.^^MMMBBHV
9.4
9.3
7.7
7.3
7.1
5.9
2.6
2.S
1.4
0.5
0.2
9.7
117.2
69.0
186.2
\
\
* From thinning and cleaning up.
14-5
-------
VOC Content of Architectural .Coattogs
Average VOC content and range of VOC content by volume for
various coating categories are listed in Table 2, The average
VOC contents listed in Table 2 are weighted by sales in each
category. The ranges given are the lowest and highest maximum
VOC content indicated on the returned questionnaires.
Low-solvent coatings are essentially waterbome latex (or
emulsion) types, although water-dispersion type coatings, which
are partially soluble in water and organic solvent mixtures, are
not becoming commercially available.
While latexes do contain some VOC, concentrations are usually
much less than in solvent-type coatings. Average VOC concentrations
are eight percent by volume for latexes and 54 percent by volume for
solvent-borne coatings.
From the returned questionnaires the staff was able to use the
sales data and the weighted average solvent content to assess the
emissions of VOC from each coating category. The VOC content for
each coating category was calculated by multiplying the volume of
coatings sold by the average VOC content. For each specific coating
category the total of these estimates yields the volume of VOC
released to the atmosphere. The volume of emissions was converted
to weight by multiplying by the appropriate density (6.5 pounds per
gallon for solvent used in solvent-borne coatings; 8.6 pounds per
gallon for solvents used in waterborne coatings). The total weight
14-6
-------
Table 2
Average VOC Content and Range of
VOC Content for Various Architectural
Coating Categories
Architectural
Coating
Category
Exterior Wall
Interior Wall
'Roof
Sash, Trim, and Trellis
Barn and Fence
Porch, Deck, and Floor
Mobil Home
Traffic
Primers, Sealers and
Undercoaters
Stains
Clears
Metallic Pigmented
Aerosals
High Performance Primers
High Performance Topcoat
Not Classified
Overall Average
Average VOC (%)
Solvent
Borne
47
52
49 .
44
45
. 56
47
50
50
66
58
59
78
53
49
54
54
Mater
Borne
10
7
8
13
4
3
9
9
0
5
8
Range of VOC (%)
Solvent
Borne
17-66
30-75
33-80
17-75
26-68
48-83
35-71
25-76
20-92
50-95
35-95
43-80
48-99
36-94
49-90
Water
Borne
1-15
2-38
0-10
3-23
1-39
.3-5
.3-5
1.5-5
.1-32
2-36
2-77
1.4-74
3.6-17
1-36
14-7
-------
per year was converted to an average emission rate in tons per day
(TPD) based on 365 days per year. The total VOC emissions from the
use of architectural coatings was determined to be 117 TPD in 1975
in California.
VOC emissions from thinning and cleanup was accounted for based
on an average of one-half gallon of solvent per gallon of solvent-
borne coating (except aerosols, roof, and traffic coatings). This
figure is derived from estimates obtained from the industry and from
local air pollution control districts. Estimates from the industry
ranged from one gallon of thinner for each 10 gallons of coating to
one to one. For traffic and roof coatings a ratio of one to five
was used because these coatings are typically applied in large
quantities, and their usage does not entail much solvent for
cleanup. No solvent usage was assumed for cleanup and thinning of
aerosols. The volume of solvents used for thinning and cleanup
in California totaled over seven million gallons in 1975 or 69 TPD.
Therefore, the total emission of VOC associated with the use of
architectural coatings in California was 186 tons per day in 1975.
Emission Factors
Perhaps the most useful way of expressing findings in terms of
an emission factor is pounds of VOC emitted per capita. Official
State of California Department of Finance figures indicate that the
state population as of January 1, 1975, was 21,030,245. Based on
the 1975 data of VOC emissions of 186 tons per day, the California
14-8
-------
per capita emission factor for 1975 is therefore 6.5 pounds per
capita.
Reasonably detailed emission inventories for emissions from
architectural coatings have been conducted in California by the
San Diego County and the San Francisco Bay Area Air Pollution
Control Districts (APCDs) for 1975 and 1976 respectively. As can
be seen from Table 3 the results of the Bay Area APCD survey
compare favorably with the ARB survey. The San Diego County APCD
survey is significantly different; no explanation of this
discrepancy could be found.
Table 3
Comparison of Air Resources Board and San Diego and Bay Area
Air Pollution Control Districts Inventories for
Architectural Coatings
Population*
Total VOC
Emission
(T/D)
Yearly Per
Capita
Emissions
(pounds)
APCD
ARB
APCD
ARB
Bay Area APCD
4,767,032
= 41
42.5
6.3
6.5
San Diego County APCD
1,541,500
17.8
13.7
8.4
6.5
* Bay Area Inventory conducted for 1976, San Diego Co, Inventory
conducted for 1975, Population figures from California Department
of Finance, SB 90 population data.
14-9
-------
Effect of ARB Model Rule on _Emisgign Factor
As indicated earlier the Air Resources Board has developed a
model regulation to control VOC emissions from the use of architectural
coatings. The regulation limits the VOC content of architectural
coatings to 250 grams per liter of coating less water except that
interior coatings are permitted to contain 350 grams per liter of
container less water. The following coatings are exempt until 1982:
1. Varnishes, lacquers, or shellacs;
2. Semitransparent stains;
3. Opaque stains on bare wood, cedar, fir, and mahogany;
4. Primers, sealers, and undercoaters;
5. Wood preservatives;
6. Fire retardant coatings;
7. Tile-like glaze coatings;
8. Waterproofing coatings;
9. Industrial maintenance finishes;
10. Metallic pigmented coatings;
11. Swimming pool coatings; and
12. Graphic arts coatings.
The emission reduction expected from the model rule is 53
tons per day in 1978. Potentially an emission reduction of 156
tons per day can be achieved in 1982 if adequate substitutes are
available for those coatings which are initially exempted. These
emission figures were derived based on the assumption that all
solvent-borne coatings that are required to meet the model rule would
14-10
-------
be replaced by waterborne coatings of the same avenge VOC content
as those waterborne coatings now in use. This assumption may be
conservative, because many waterborne coatings currently contain
more VOC than is allowed by the model rule. Implementing the model
jrule wil}. bring about a reduction in the VOC content of waterborne
coatings and an accompanying reduction in emissions attributable
to such coatings. Therefore, the net reduction in VOC emissions will
be greater than 53 tons per day, but exact figures cannot be
determined at this time.
Based on a reduction of 53 tons per day, the total emissions
of VOC to the atmosphere in 1975 would have been 133 tons per day
instead of 186 tons per day if the model rule had been in effect.
Therefore, the emission factor if the ARB's model rule had been in
effect in 1975 would have been 4.6 pounds of VOC emitted per capita.
Summary
The level of VOC emissions from architectural coatings can be
determined by using sales data or per capita emission factors.
When sales data are known for a state or anyother geographical
area, the volume of solvent for each coating category can be calculated
by multiplying the volume of coatings sold by the average solvent
content. For each specific coating category the total of these
estimates yields the volume of organic released to the atmosphere.
The volume of emissions can then be converted to weight by
multiplying by 6.5 pounds per gallon for the solvents used in
14-11
-------
solvent-borne coatings or 8.6 pounds per gallon for the solvents used
in waterborne coatings. In addition, similar computations should
be made to estimate the VOC emissions from the solvents used for
thinning and cleanup, and these emissions should be added to the
emissions attributable to the coatings.
Data compiled by the GARB indicate that in 1975 the VOC emissions
attributable to architectural coatings was 117 TPD, and the VOC
emissions from solvents used for thinning and cleanup was 69 TPD,
thereby giving a total of 186 TPD. Had the model rule been in
effect, the total would have been 133 TPD.
A quicker method, and one that works in the absence of sales
data, is to compile VOC emissions on a per capita basis. According
to figures compiled by the CARS and San Francisco Bay Area APCD,
yearly per capita emissions from architectural coatings and the
solvents used for thinning and cleaning approximately 6.5 pounds per
capita without the model rule and 4.6 pounds with the model rule.
Multiplying the per capita figure by state or area population
will yield an annual emission amount that can be converted to TPD
by dividing by 365 (*g|} and 2000 founds).
Either method will yield acceptable accuracy when computing
emissions of VOC associated with architectural coatings.
-------
NOX REDUCTIONS IN THE PORTLAND CEMENT INDUSTRY
WITH CONVERSION TO COAL-FIRING
Presented at the 1977
Environmental Protection Agency
Emission Inventory/Factor Workshop
Raleigh, North Carolina - September 13-15, 1977
By
ROBERT J. HILOVSKY, P.E.
Supervisor, Source Test Section
Engineering Division, Eastern Zone
South Coast Air Quality Management District
Col ton, California
15-1
-------
Introduction
i
The cement industry is one of the nation's most energy-
intensive industries - where more energy is consumed producing a
dollar's worth of product than for any other major product. A
report issued by the Cost of Living Council in 1973 shows that
the energy cost for cement was 43 percent of the product. This
figure has continued to rise with the increasing cost of fuel.
The cement plants of Southern California have used natural gas as
fuel, with oil as a standby energy source. The high availability
of natural gas, ease of handling and its cheap cost compared to
other fuels were the major factors for continuing its use. How-
ever, with the growing shortage of natural gas, estimates by the
California Public Utilities Commission that no gas supplies will
be available to major industries by 1980 and large price increases
(38 percent in 1975) for gas, the cement industry began conversion
to fuel oil and coal.
The South Coast Air Quality Management District (SCAQMD) has
four cement companies (operating six different facilities) under
its jurisdiction. All of these facilities are located in the
Eastern Zone of the District, with five plants in San Bernardino
County and one plant in Riverside County. The SCAQMD was formed
on February 1, 1977, as a successor agency of the Southern
California APCD. That APCD, in turn, had been formed on
15-2
-------
July 1, 1975, from the Los Angeles, Orange, Riverside and San
Bernardino County APCD's. All data referenced in this report was
collected by the same group of personnel - although the organiza-
tion changed names.
Background
The San Bernardino County APCD began source testing for NOx
emissions in 1969-70 for all industries in the county for both
compliance and emission inventory information. The larger indus-
tries in the county were also tested on an annual basis, begin-
ning in 1972. Variations in NOx emissions from one facility
were observed, but investigation as to the cause was not pursued
at that time. The emission inventories showed that the cement
industry comparatively was a very large NOx emitter (Table I)
in San Bernardino County.
15-3
-------
TABLE I
NOX Emissions from
San Bernardino County Cement Plants
NOx Emissions
Facility (Tons/Day)*
California Portland Cement Co., Col ton 19.10
Riverside Cement Co., Oro Grande 25.66
Kaiser Cement & Gypsum Corp., Lucerne Valley 20.42
Southwestern Portland Cement Co., Victorville 7.0
Southwestern Portland Cement Co., Black Mountain 13.44
TOTAL 85.62
*Based on an average rate of 80% production, natural gas for fuel
NOx is reported as N02-
Fuel Changes and Effects Upon Pollutants
Riverside Cement Company and California Portland Cement
Company filed applications in 1974 with the District to convert
their rotary kilns to coal-firing. Review of these applications,
in considering the possible changes in emissions, led to the
analysis of the data collected from source tests on cement kilns.
Analysis of these data revealed:
(1) The sulfur in the fuel oil was absorbed in the
clinker manufacturing process (as sulfates or
sulfides) and only very small amounts of sulfur
dioxide would be emitted to the atmosphere. It
was expected, therefore, that the sulfur in the
coal also would be absorbed and would not cause
any violation (500 ppm limit) of the District's
S02 rules.
15-4
-------
(2) Existing air pollution control equipment could
adequately control any increase in particulate
matter expected from coal use.
(3) The use of fuel oil showed a reduction in NOx
emissions, compared to NOx from natural gas.
It is believed that when burning fuel oil in the cement kiln
that it can more readily be burned with a flame that is less
oxidizing than the flame resulting from natural gas combustion.
(It would appear to be a "lazy" flame pattern when viewed through
flame ports.) With these differences in the kinetics of combus-
tion in the kiln, the result is lower NOx generation when burning
fuel oil in the cement kiln - compared to natural gas. The use
of coal for fuel should result in an even further reduction of
NOx emissions since it typically produces a longer, "lazier"
flame (with lower temperature in the center of the flame) than
does fuel oil combustion in the cement kiln. In reviewing appli-
cations from the cement plants, the "Permits to Construct" were
approved since it was calculated that an overall reduction in
emissions into our air basin would occur.
The conversion to coal was completed by November 1974 for
the Riverside Cement Company and by May 1975 for the California
Portland Cement Company. Southwestern Portland Cement Company
and Kaiser Cement & Gypsum Corporation switched over from
15-5
-------
natural gas to fuel oil combustion in 1976. Source testing of
these units has indicated that a substantial reduction occurred
in NOx emissions into the atmosphere.
TABLE II
NOx Reductions in Cement Kilns
Due to Fuel Changes
FACILITY
California Portland Cement
Riverside Cement
kaiser Cement & Gypsum
Southwestern Portland Cement
(Victorville)
Southwestern Portland Cement
(Black Mountain)
- TOTAL
NOX EMISSIONS^1)
(TONS/DAY)
Gas
19.10
25.66
20.42
7.0
13.44
85.62
Oil Coal
4.58<2) 3.50
7.75
15.46
4.30
12.06
43.07
PERCENT
REDUCTION
76^^/81.7
69.7
24.2
38.2
10.2
49.7
U)Based on 80% production rate. NOx is measured as N02-
used at this facility since conversion to coal.
15-6
-------
Table II shows that larger reductions in N02 emissions are
accomplished with conversion to coal-firing versus oil-firing.
With the growing scarcity of petroleum products, there would be
more advantages in the long run for cement plants to convert to
coal-firing (directly from natural gas) rather than to oil-firing;
even though a conversion to oil-firing would somewhat reduce NOx
emissions into the atmosphere. Kaiser Cement & Gypsum Corporation
has filed an application for coal conversion with the District,
and Southwestern Portland Cement Company has approved funds for
coal conversion.
Test Methods and Procedures
Two test methods were used in obtaining the data (Appendix A)
presented in this report. The Phenoldisulfonic Acid (PDS) method,
which is the approved California Air Resources Board and U. S.
Environmental Protection Agency reference method, was used along
with a continuous electrochemical cell analyzer (Envirometrics)
and recorder. Both methods well complimented each other although
the analyzer was not obtained until 1972. Some early PDS data
was considered invalidated when it was indicated that NOx concen-
trations were over 1,000 ppm. For NOx values near or over 1,000
ppm, the chemist performing the PDS analysis must be aware of the
15-7
-------
potentially high concentration so proper steps in the preparation
of aliquot portions can be taken to assure accuracy in the analy-
sis.
The continuous analyzer revealed variations in emissions
throughout the process operations (Figure 1). For example, the
concentration range for one test was 950 to 1,650 ppm NOx, with
an-average of 1,490 ppm. For this example, the PDS values could
vary greatly depending upon when the "grab sample" was taken,
with respect to hitting "peaks" or "valleys" in the NOx versus
time curve.
Emission Factors
The five plants tested have different configurations of ex-
haust gas ducting and different types of control systems. This
resulted in different excess-air concentrations for each test
site. To obtain a correlation of NOx emissions into the atmos-
phere, emission factors were generated. These are listed in
Tables III, IV and V and divided into categories dependent upon
(1) fuel use, (2) type of process and (3) production rate.
15-8
-------
Conclusions and Recommendations
Table VI is a summary of the emission factors generated, and
Figures 2, 3 and 4 are plots of the emission factors versus kiln
capacity. The following conclusions are indicated from this data:
(1) Emission factors vary greatly depending upon fuel,
type of process and kiln size.
(2) There is a significant reduction in N02 emissions
when either oil or coal is used for fuel, versus
natural gas. It appears that greater reductions
in emissions are available for coal-firing versus
oil-firing (Table II).
(3) The emission factors for wet-process operations
tend to be lower than for those with dry-process
operations (Table VI).
(4) As the capacity of the kiln increases, the emission
factor decreases for dry-process operations (Figures
2 and 3) while the reverse is indicated for wet-
process operations (Figures 2 and 4). There can,
however, be a larger NOx variation between kilns
of the same size - especially the smaller units
(Figures 2 and 3).
(5) The emission factor and NOx reduction from natural
gas-firing versus oil-firing, for dry-process kilns
of 100,000 Ibs/hr of clinker, were much greater
than for a 175,000 Ibs/hr kiln (respectively 4.58
Ibs/ton and 76% reduction versus 12.06 Ibs/ton and
10.2% reduction).
(6) The NOx emission factors depend upon a number of
variables, and the use of only one factor should
be discouraged in estimating NOx emissions from
cement kilns.
15-9
-------
Some of the more important variables have been covered in
this paper although other factors, such as diameter of kiln,
length of fire zone and dwell-time before emitting into the atmos-
phere, should be investigated before developing a family of curves
for cement kiln NOx emission factors.
15-10
-------
TABLE III
Emission Factors for Cement Kilns
Using Natural Gas
Kiln
Raw Material Feed
(1,000 Ibs/hr)
Dry Process Units*
RC1
RC2
RC3
RC4
RC5
RC6
CP1
CP2
BM
BM
64
64
64
64
64
65.7
161
161
264
240
Wet Process Units
SW5
SW6
SW7
SW8
SW9
KC1
KC2
KC3
29
39
49
40
50
38
46
40
41
92.4
92.4
184
184
Emission Factor (Ibs.
Raw Material Feed |
14.3
13.9
12.6
13.7
12.5
15.8
13.6
11.9
10.9
11.7
18.7
3.9
9.5
3.3
5.2
5.6
6.5
8.3
15.3
3.2
4.1
6.6
6.0
NOx/tonl
Clinker
22.4
21.8
19.7
21.4
19.6
24.7
20.5
18.7
16.9
18.1
28.9
6.1
14.6
5.1
8.1
8.6
10.0
12.7
23.6
5.0
6.4
10.3
9.4
*RC = Riverside Cement, Oro Grande; CP = California Portland
Cement, Col ton; BM = Southwestern Portland Cement, Black
Mountain; SW = Southwestern Portland Cement, Victorville;
KC = Kaiser Cement & Gypsum, Lucerne Valley
15-11
-------
TABLE IV
Emission Factors for Cement Kilns
Using Fuel Oil
Kiln
Raw Material Feed
(1,000 Ibs/hr)
Dry Process Units
CP1
CP2
BM
168
168
168
168
240
Wet Process Units
SW7
SW8
SW9
KC1
KC2
KC3
49
49
41
92
92
92
92
184
Emission Factor (Ibs.
Raw Material Feed I
1.6
4.1
2.9
2.8
10.5
3.7
7.9
2.3
2.8
2.9
3.0
3.1
5.1
NOx/ton)
Clinker
2.6
6.9
4.9
4.6
16.1
5.7
12.2
3.5
4.4
4.5
4.7
4.8
7.9
15-12
-------
TABLE V
Emission Factors for Cement Kilns
Using Coal
Kiln
Raw Material Feed
(1,000 Ibs/hr)
Dry Process Units
RC2
RC3
RC4
RC5
RC6
CP1
CP2
64
64
64
64
64
64
64
64
65.7
161
171
159
157
Emission Factor (Ibs.
Raw Material Feed [
1.4
3.6
4.4
4.9
5.4
5.6
6.2
6.2
4.1
2.0
2.9
2.4
1.9
NOx/ton)
Clinker
2.2
5.7
6.9
7.6
8.5
8.6
9.7
9.6
6.4
3.3
4.7
3.7
3.1
15-13
-------
TABLE VI
Summary of N02 Emission Factors
for Cement Kilns (Ibs. NOx/ton of Clinker)
Type of
Cement-Manufacturi ng
Fuel Process Range Average
Gas Dry 16.9 to 24.7 20.4
Gas Wet 5.0 to 28.9 11.5
Oil Dry 2.6 to 16.1 7.0
Oil Wet 3.5 to 12.2 5.9
Coal Dry 2.2 to 9.7 6.2
15-14
-------
APPENDIX A
TEST DATA USED FOR REPORT
15-15
-------
Capacity Test
Unit Bbl/day Date
Dry Process
CP1 6,500 12/28/76
6/15/76
10/12/73
1/28/70
4/28/70
? CP2 6,500 12/28/76
t
O\
6/15/76
10/12/73
4/28/70
1/28/70
Raw Material
Ibs/hr
151,000
(+ 20,400 coal)
161,000
168,000
159,000
157,780
168,000
Clinker
Production
Tons/hr Fuel
50 Coal
50 Coal
Gas
Oil
Oil
50 Petroleum
Coke & Oil
49 Coal
Gas
Oil
Oil
Emission Factor
Flowrate
DSCFM
138,555
150,000
140,500
127,900
127,900
139,597
135,000
146,600
188,041
188,000
N02 Emi
PPM |
220
150
1,000
142
372
183
157
880
178
169
ssions
Ibs/hr
221.9
165.1
1,023.2
132.3
346.5
186.0
154.7
939.4
243.7
231.4
Raw Material
(Ibs/ton)
2.94
(2.6)
2.0
13.5
1.6
4.1
2.4
1.96
11.9
2.9
2.8
Clinker
4.7
(4.2)
3.3
20.5
2.6
6.9
3.7
3.1
18.7
4.9
4.6
-------
Emission Factor
I
II
VJ
Capacity
Unit Bbl/day
Wet Process
KC1 4,000
KC2 4,000
KC3 8,000
Test
Date
3/5/76
5/2/72
12/14/76
5/2/72
12/16/76
10/15/73
5/2/72
Raw Material
Ibs/hr
136,400
142,588
136,137
142,588
273,100
292,786
276,255
Solids
(92,300)
Solids
(92,300)
Solids
(184,615)
Fuel
Oil
Oil
Gas
Oil
Gas
Oil
Gas
Gas
Flowrate
DSCFM
77,939
78,630
60,933
57,100
57,012
55,185
119,072
108,443
99,600
90,973
N02 Emissions
PPM 1
493
503
770
710
780
1,082
1,180
2,000
1,880
Ibs/hr
279.8
288.0
341.7
294.9
324.0
434.3
1,023.0
1,450.6
1,245.5
Raw Material
(Ibs/ton)
4.1 (2.8)*
4.3
4.8
4.4 (3.0)*
4.8
6.1
7.5 (5.1)*
9.9 (6.6)*
9.0 (6.0)*
Clinker
6.4
6.7
7.5
7.5
9.5
11.69
15.4
14.0
*Raw material feed of dry product excluding water
-------
Capaci ty
Unit Bbl/day
Dry Process
RC1 2,600
RC2 2,600
RC3 2,600
>
£ RC4 2,600
oo
RC5 2,600
RC6 3,000
3,000
Test Raw Material
Date Ibs/hr
3/19/74
5/25/76 64,000
3/19/74
5/25/76 64,000
3/19/74
5/25/76 64,000
3/19/74
5/25/76 64,000
3/19/74
11/12/75 63,000
7/28/74 65,700
Clinker
Production
Tons/hr Fuel
Gas
20.51 Coal
Coal
Gas
20.51 Coal
Coal
Gas
20.51 Coal
Coal
Gas
20.51 Coal
Coal
Gas
20.19 Coal
21.05 Gas
Emission Factor
Flowrate
DSCFM
48,917
45,990
44,478
44,478
46,520
40,295
40,295
59,940
^59,000
^44,000
58,794
58,800
»- 48,900
44,462
36,710
17,997
N02 Emissions
PPM I
1,288
135
360
1,382
(1,050-1,640)
417
535
1,380
(990-1,520)
398
170
1,375
(1,160-1,400)
465
460
1,128
(920-1,200)
400
1,158
1,609
Ibs/hr
458.6
45.2
116.6
447.6
141.2
156.9
404.9
173.7
73.0
440.0
199.1
196.9
401.6
129.5
520.4
Raw Material
(Ibs/ton)
14.3
1.4
3.6
13.9
4.4
4.9
12.6
5.4
5.6
13.7
6.2
6.2
12.5
4.1
15.8
Clinker
22.4
2.2
5.7
21.8
6.9
7.6
19.7
8.5
8.6
21.4
9.7
9.6
19.6
6.4
24.7
-------
Capacity Test
Unit Bbl/day Date
Solids - Wet Process
SU5 1,300 4/26/74
SW6 2,200 3/21/74
5/12/70
SW7 2,200 5/12/70
b/1/70
4/29/76
K SW3 2,200 5/12/70
,1 3/21/74
* 4/29/76
SW9 2,200 3/21/74
4/29/76
6/11/75
Dry Process
Blk 9,500 7/12/74
Mtn 0/11/75
4/30/76
Raw Material
Ibs/hr
29,150
49,300
39,720
40,610
50,240
49,318
38,610
46,200
49,641
40,400
41,603
41,600
264,000
240,000
240,000
Clinker
Production
Tons/hr
9.5
16.0
12.9
13.1
16.2
16.0
12.5
15.0
16.1
13.1
13.5
13.5
85.3
77.5
77.9
Emission Factor
Fuel
Gas
Gas
Gas
Gas
Gas
Oil
Gas
Gas
Oil
Gas
Oil
Gas
Gas
Gas
Oil
Flowrate
DSCFM
25,319
38,373
29,681
30,948
42,821
38,240
27,747
32,500
40,900
36,333
37,459
36,200
86,340
86,340
86,340
N02 Emissions
PPM |
1,490
(950-1,650)
836
(700-900)
362
297
420
330
535
636
659
631
179
1,212
2,300
2,230
2,000
Ibs/hr
274.7
233.6
78.2
66.9
130.9
91.9
108.1
150.4
196.3
166.9
48.8
319.4
1,445.6
1,401.6
1,257.1
Raw Material
(Ibs/ton)
18.7
9.5
3.9
3.3
5.2
3.7
5.6
6.5
7.9
8.3
2.3
15.3
10.95
11.68
10.47
Clinker
28.9
14.6
6.1
5.1
8.1
5.7
8.6
10.0
12.2
12.7
3.5
23.6
16.9
18.1
16.1
-------
FIGURE 1
1 "
^
C "^
( =-.
- .- .^
f
\
s~.
( >
x
(
( -*
k e.
f ***.
V ^"^
C "T-
£'
f \-
V s-~
«
t "
f
r-
!1 N
: c
~^~
,
* *
f
V w
«
C *
C -
1
.
K "
; *"
(T
C j-j
-{.
C "*
I
f ^
^
(^
'
Cc
( s
(_
^^.
C SC
C C
k
s
11 /
^'
^
""-i.
^1
,
>
t<
y
_
v
r
)>-
i
\
i
^*
*_
-
f
br
x -
u -
c _
i; -
"
_
~* '
< '
u *
«-
- *
^i
V
X
2
£
,
o
,< ;<
f (
;'-..-
L.L-]
M->-
- K,.
-
-. ._.
_i
8 '
G' *
£ <
C <
-15
-
^.
Z=^
^ »
_
..
*'l
£
-
f
:
\
"^*k *f*
\ ^"^
if--
tfe
N N
1
C (
c <
f
c
1
)
^
^
9T
r>
*-
r
i
",<
(
t=
'] ,
i,
I '
i
|
4-
i
i
i
1
F
1
1
[_
1
1
f
I
I
1-
i
r-
C
s
o 3
(
v»
rs '
^X
r*^
~^t
C
C
,c
c
~t
1
t.
1
i
: c
! SZ
C
c
c
~f^
\J\f
-e~
<">'
j, **
f\^
^
C
c
If
c
(
'\r\
i/i
i;
«l J
'(
|i
1
-» *
j_
n ~1
-I
;I"
.l_
4-
IL
"T
"t"
1
"^
v~
C
C
r<
c
c
Y-
c
c
L
f
I
1
I
i
'
j
U A
f
\ S
(
fS
c
e?
C
f
s
c
jj
c
c
c(
S 1
(
K
r
vJ
- ^
c
z
5 i f
C
-8
/W
4
-?
C
C
C
r J
F 1 C A T
C (
>e
n
v^V '
\-*
V*
n
o
: c <
4
c ^
- 1 C A T 1
C <
n
0
Jb-fc »
cf\T^J-
_
^.- I r
0
1 . . .....
c c
C C
ON IP
: c c
1
1
1
f
1
1 ....
1
r c c
r c
L e-e
ON IN
. C C
D
ft
f
I
I
,/
.
c c
>c c
,c ,c
4 D t X
c c
/""
1
*
c c
c c
,c *
O E X
c c
1 R E
>
c c
C C
f »C
( (
c c
c c
4C *2C
c c
c ^
ft
Trrlr-r-
\*rr-f
c c
c c
H
1
-*"
f
c
h
c
f
c
N
f
c
c
c
c
5
l^"*
c
c
c
o c
'
c
c
,.<"" v.c
c c
< 52 3
; X
V-* * '
I f\
r(r/
f-
VA.A jjv
c c
c c
X
'VmjV^^
pi**^
Y ,v 1
/""
^ f _!/ / j
^7 ^7 /^
^" / lv
. /
s\
c c
c c
tjr |M
rv\
j
h\^j ^^
1
. . ....
1
C C
c c
.C U2C
c c
1
T*~&t~T~~
Ttrfi f-
f 0 -J-3'jT
Id -
*> M
/ /rr
t-'ij~r
C C
p C
C C
f
c
Fife
1
-/57T
^1
PjWf1"
X W "*
^ "
"*" 1
- C5 .
4V*/
, 0
o «
E^:
2 '
- **-H
"o~
LiJ
--CM-
"?
~ cn "
lZ ^
ul
-2-
c
c
,c
c
~s~
^_
rl<..
o
2 -
r
O
E^~
b^cj
o
-<
«3
Vt
o
1_
"r>"
C
r
c
z
*
p
^
"*
r
iv
i1)
"t
'{
n
1
z
J
X
,
'
3
^
3
-
H
_
r
o
r.
<
(
j
(*k
'//
r
^
'.
^
;
-
(
2
a
-,
£
i
(
<
<
;
j^
;
.-
^
^j
H
3
"
v,
H
*
""r %
: c
c
"O ^ ^~
> "
L .r
n C
-> ~c
= -r
T 1^ ^"
*-
f
- -L
0 ?("
^
- *J r
^
n f
*-
n fJr
c
r c
: c
: c
C~
z
o ^
N V.
" v.
1
" -C
3 (
T
- -c
r >* V-
o ^» f
Q -f
z ,
H *"
- WC
m * (
^
o ^* (
f
C
c c>;
15-20
-------
*
f "
\
C '
*
f **
^"
( r-
V ^
- -
\
t
C "
( -
*"
. . .
r «
i
c <
(^
_ _ j
-
C "
{ ~
f -
i
c c
*
'J
*-
~
*
, >
^
-
^
J
2
.
,.
J
;
-
(
(
^
'
-
(
(
C C
5 ?; , i
/
C C
( C
^c
c
( C
C (
c c
C cC
L A S S 1
( (
\
N
V-VXJ
, *_
1
.
c c
c c
cjc
F 1 C A T 1
c c
^n
n
u
V
\
\
V
A
Vv.
n
n
c c
c c
( -C
ON 1 f
( C
o
A
J^|V
r\_^
^x*
X
C C
c c
* k
M o t: x
C (
1 R C
K^_*^^»/V
"~^*
C (
C (
( C
C T
<-^V
V
1 J.
c c
c c
c
c
'* 50
1 N C)
Pt5s/-"t
C
(
c
(
c
/*
J
i
^4
C
c
, c v.c
c c
. X
J
rv- /
v_^/
JI-\
f\ ]
J--I \ i~
i > / /
O/'f /
' '
\ . j -o
-T i y < /
^" ? ^
t.s
^\ (
ti 1 1
c c
C (
( (
c c
r\yAX^-
^ ^v\ /*) *i
V t1 H- 1
-77 flj
/ J> /(/ S
-"
1
_ 1 ., ..
M 4 b; it
i i
' ul /
C (
C (
r Tf
Y F. A r
C C
If
,
^VTAK
L KEfL
L-r 1 / ;/.£.y
. | .
?,. -//;,
/A 'yt-
s
i 0/0 C
c c
c c
c I("
7
C (
,o
V
w
f^
0N~t"' '
?) ^.^...
v ff I »
U f
~\ *
-^ i^/Vl
»l f t \^ -
C (
C (
4( L
c
^ ff /
, ^
t^7\
i S J
> /T>.
ft" S
r^ixl * *
C
c
r
C
1
1
LL
C
(
fc
1 C
r~
r 5 ~
-
r i "
ji^
o
i -'
r r
i O"
r *-*
E °-
i
cu
1 -£
}C£
.-V
!±,
T
cj
1^>
Q
LU
-T
-f i-1
C
c
2
_
-
z
-I
"
,
r.
-
-
T
i
;
~
^
;
=
J
;
-
f
41
<
C C
c c
(
z ,-
** c^
^
- - f
V.
- - 1
" " C
o fr
- -^r
V.
_
*~ * C
o »-- r
^
~~ *° C
p *- ^"
^
-c -
c
: c
15-21
-------
r
Oi
ro
N>
28ZO 9V
VSDHIiaVH 'OO H3S53 a
3MI 01 X i'HDNI 3HX Oi 01 X 01 S
-------
1
4
__
.:
--
7L
- - -
._ .. _
1
. _ _
_. ....
_ ..
N02
RAW
--
-
-
j
i
-
-
-
-
-
-
-
-
-
-
-
*
--
._.
---
--
EMISSIONS
LBS/TON
MATERIAL
~:
... _
~
_. .
--
- -
.:"
.
--
-
i
-
-
--
-
-
r~
r~
....
--
-
--
---
_j
--
--
--
--
-
_
.-
-
-
-
-
't
...
~'t
-
-
-
-
--
-
-
-
y
-
OT
-
-
,_
-
_
-
-
-
-
-
-
-
-
-
-
-
!
-
-
-
-
-
-
-
-
-
-
-
...
-
-
-
-
-
-
-
-
-
-
-
-
_
-
...
-
-
-
-
-
-
-
-
~
-
_
-
-
-
-
-
-
~"
-
-
-
2.
-
-.
-
-
..
-
-
-
-
<
-
-
-
-
-
-
-
_
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
_
-
-
-
'
^
-
4
-
-
-
-
...
_
-
«
-
-
0
-
_..
-
-
^
f
-
-
-
-
-
--
-
-
-
-
-
-
-
-
:
-
-
-
-
...
-
-
-
-
-
_|_
L~
-
-I
-
-
i
i -
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
^«
i
-U
-
-
-
\~
-
--
-
-
-
-
^
-
-
...
-
-
«.
»
-
-
-
-
-
-
-
--
-
-
9
i
-
...
-
-
-
._
-
-
-
_
-
0
-
-
-
-
-
-
-
-
-
-
-
-
^
-
-
-
-
-
_
-
--
«*
:<
-
-
-
Fi
-
-
--
-
^
.
't
--
-
-
-
...
-
-
-
-
»
-
c
RAW MATERIAL
-
--
-
_|.
1
I"
1 1
-
-
--
-
gure
-
-
-
-
-
-
-
:
t
F
-
-
--
-
...
-
-
:E
-
-
-
-
--
--
*
-
-
i
E
-
-
-
-
_
-
';
D
-
*
-
-
-
-
-
1
-
-
-
-
KB.
i
(1
~! -
-
-
\
-
-
-
-
-
«
t
-
-
-
i
0
-
-
-
-
-
-
ft
.1
0
-
-
...
_
-
-
-
-
-------
:.
.-
--
--
--
... _
.__
--
.
...
-
-
-
-
-
-
-
r i
-
:
-
r1 "i
-
-
-
-
._
-
-
-
~i
.
-
-
-
-
-
-
-
-
-
-
-
-
-
-
...
r~
-
-
-
-
-
-
-
-
-
-
-
~
-
...
.-
-
-
N02 EMISSIONS
LBS/TON
RAW MATERIAL
-
-
-
-
-
-
-
-
-
--
-
-
-
-
-
-
-
--
-
-
...
-
-
-
-
P
-
-
/.
/
-
-
-
-
-
5
~
f
...
4
w
--
-
»
-
-
-
f
-
...
-
-
-
-
-
-
-
-
-
-
-
-
-
-
:
_
-
-
-
-
-
.
-
-
-
-
-
-
-
-
...
-
-
-
_
-
-
-
-
-
-
-
:
-
-
-4
-
-
-
...
-
-
!
-
1
-.
-
-
--
-
...
-
-
-
J
-
"
-
-
-
-
-
-
-
"
-
-
-
-
--
"
-
-
-
_-
"
1
-
-
:
-
-
'
?
-
-
-
-
-
-
(
0
-
-
-
-
-
-
-
-
-
-
-
-
X
-
X
-
--
-
-
-
-
-
;
-
-
-
-
-
-
i
-
-
_
...
-
-
-
-
-
**
0
F
-
_.
.-
-
-
-
-
--
-
A
:
~
-
-
...
-
w
i
_
4
-
-
-
-
...
_
-
.<
i*
il
9
V
-J
-
-
-
-
-
_
*
-
-
-
--
p
_
-
-
-
-
*
F
-
-
-
-
-
-
-
-
/
-
-
-
.
-
-
-
-
-
--
-
-
7
4
i
i .
-
-
-
-
-
'<
-
-
-
-
*
t
i.
-
-
-
-
-
^
s
.. .^ .( . ... - . .
,000 LBS/HR)
"
J
-
\
-
-
-
-.
1?
f
1
/
-
-
-.
-
-
-t
-
-
_
-
-
«
-
'i
-
-
-
-
-
/
r
-
ft
~\
-
-
-
,
«
!
-
-
-
...
-
(
/
-
~
1
^
7.
1
:
-
._..
".,(
/
* "
f
i
i
-
-------
QUESTION:
HILOVSKY:
QUESTION:
HILOVSKY:
QUESTION:
HILOVSKY:
CONDENSED DISCUSSION
What is the difference in tests on kilns
using natural gas and coal as fuel?
There is no difference in running coal
versus natural gas. The dust loading from
the cement process is so high that actually
the ashes absorb in the cement. It actually
makes a better cement product.
Do the production rates change when you change
fuels? What are the heat release rate differ-
ences?
Let's see if I can go back to the charts.
There is a slight change in the total raw
materials going in. On gas they were running
around 161 thousand pounds of raw material in
that particular kiln. They are running 151
thousand pounds of raw material on 20 thou-
sand pounds of coal. Approximately the same
out-put. Still 15 tons per kiln.
When were these tests run?
All the way from '69 through two or three
weeks ago. I might say that we run, or we
have had a policy of testing all of our
large companies at least on an annual basis
15-25
-------
QUESTION:
HILOVSKY:
QUESTION:
HILOVSKY:
and sometimes two times a year depending upon
how good they have been in terms of compliance.
We have more test on this one in the south
coast basin because it's closer to us and we
have done some research work with them. We
have tested our instruments there and a few
other things.
Do you have a significant amount of carbon
monoxide fluctuation?
We have been taking samples and didn't notice
any significant difference. I don't have that
with my data here, but when we go out and
sample, we take everything. We take carbon
monoxide, S02, and NO and particulate matter
all at the same time.
Do you have a wet process estimate?
There is no wet process on coal yet. One of
the companies running a wet process has the
application already approved to construct
coal handling facilities. At present though,
it is not in business.
15-26
-------
CURRENT API EMISSION MEASUREMENT PROGRAMS
Presented at the 1977
Environmental Protection Agency
Emission Inventory/Factor Workshop
Raleigh, North Carolina - September 13-15, 1977
By
J. G. Zabaga
Associate Engineer
Mobil Research & Development Corporation
Princeton, New Jersey 08540
16-1
-------
Introduction
My name is J. G. Zabaga. I represent the
American Petroleum Institute, and specifically, the
Committee on Evaporation Loss Measurement (CELM). I
am the Chairman of CELM.
A great deal of current interest is being focused
on the evaporation loss bulletins that have been
published by the American Petroleum Institute over a
period of years. Publication of the bulletins preceded
the Clean Air Act Amendments of 1970, and it has now
become clear that both industry and regulators need
updated emissions data to properly respond to the Act.
To start, I would like to review the function of
the Committee on Evaporation Loss Measurement. Figure
1 shows the organizational structure of the American
Petroleum Institute (API). There are a number of
technically oriented departments, one of which is
Non-Departmental/Industry Affairs. Within that
Department, as shown in Figure 2, is the Committee on
Petroleum Measurement (COPM), a free-standing API-
based committee. Titles of the standing committees
describe, in general, the scope of the Committee's
responsibilities, all ultimately resulting in the es-
tablishment of standards for measurements of all types
16-2
-------
of petroleum products throughout the industry. This
includes production, refining, distribution and
marketing activities.
The CELM is one of COPM's standing committees.
CELM deals only with the measurement of evaporation
loss, which is quite different from the finite measure-
ment considerations of the other committees in
developing standards for inventory control or custody
transfer.
Figure 3 lists the publications of the CELM that
were developed over a 12-year period from 1957-1969.
These bulletins were the first industry-published data
on quantifying evaporation loss of hydrocarbon liquids,
and they have been widely accepted. To this day they
remain, or are the basis for, virtually all data
sources dealing with hydrocarbon evaporation loss.
The several editions of the EPA's AP-42 Compilation of
Air Pollution Control Factors, have emission factors
derived from these bulletins.
It is important to realize that all of the
bulletins were published prior to the U. S. Clean Air
Act and do not necessarily comprehend the accuracy
levels required by the Act. In fact, the bulletins
were originally developed to enable oil company
16-3
-------
operating divisions to prepare cost benefit studies
for evaluation of alternative conservation techniques.
When the data and estimating methods became one of the
tools used by regulators for control strategies,
industry requested that API review the content of all
of the bulletins for suitability in this new application,
Figure 4 shows the present organization of CELM,
including active subcommittees. Figure 5 is a pictorial
view of the equipment involved within each subcom-
mittee 's responsibility- The five subcommittees and
any ad hoc task groups are staffed by 31 members, 21
from oil companies and 10 from supplier companies.
Referring to Figure 5, the petroleum liquids of
concern to the Committee, and the basic equipment and
handling characteristics are as follows:
1) Open top floating-roof tanks have been in
use since 1880 and in their present form
since 1923. The concept is to curtail
evaporation by floating a disc, the movable
roof, in the liquid. Under certain
conditions, some emissions result from the
perimeter sealing ring area. Floating-roof
tanks are usually used for liquids with
storage vapor pressures between about 1.5
16-4
-------
to 11 psia i.e., motor and aviation gasolines,
and some crude oils.
2) Internal floating covers or covered floating-
roof tanks are a logical combination of open
top and fixed-roof tanks, providing weather
protection to the roof and contents, and
evaporation control by use of a floating
deck. Liquids generally stored are the same
as open top floating-roof tanks.
3) Fixed-roof tanks have been used for many years
to store all types of petroleum. Recognition
by industry that product could be conserved
by use of one of the floating-roof systems.-
and later, regulations to limit use to liquids
with less than 1.5 psia storage vapor pressure,
have generally restricted use of fixed-roof
.feanks to low volatility liquids, i.e.,
distillates, residuals and some crude oils.
4) Tank truck and tank car considerations are
confined primarily to motor gasolines. Some
crude oil and other finished product are also
transported by these means. The method of
loading is apparently the feature most
16-5
-------
affecting evaporation loss.
5) Marine vessels transport crude oils and all
types of finished product. The depth of the
loaded compartment is a major factor affecting
evaporation loss as we will see later.
When evaporation loss data were developed by the
original committee, the resulting bulletins were not
intended for use as a means to determine emission
factors for conducting emissions inventories nor for the
evaluation of regulatory strategies. Their use for
such purposes may be both, impractical and improper.
The subcommittees that developed:and/or subsequently
revised the bulletins recognized that the estimating
methodology and data had precision appropriate only
for comparative correlations among varying hardware
designs and different hydrocarbon liquids. The
estimating techniques in the bulletins were/ for the
most part, originally developed and standardized from
data collected in the period 1930-1950. Within the
specific purposes of allowing the industry to evaluate
on a cost/benefit basis, with a limited precision,
alternative vapor conservation devices, these
techniques were entirely adequate. However, the
techniques are not satisfactory in a regulatory
16-6
-------
climate demanding data that comprehends technical
advancements in both testing methods and equipment
design.
In most applications, the present bulletins have
been found to overstate losses when compared to new,
detailed tests. This will be amplified in the
discussions that follow. I do want to emphasize now
that the API Committee on Evaporation Loss Measurement,
consistent with current Institute policy and direction,
is charged with regularly revising its publications
to incorporate technological improvements in the state
of the art on vessel and hardware design, and testing
methods.
Marine Terminal Emissions
In 1976, CELM updated Bulletin 2514A, Hydrocarbon
Emissions from Marine Vessel Loading of Gasolines.
A small but important point in the bulletin title is
the use of the term emissions. API distinguishes
between the terms evaporation loss and emissions. The
industry has traditionally required loss data
expressed as a volumetric change, i.e., barrels per
year, for use in inventory control. This volume
change is defined as evaporation loss. Regulators are
properly interested in mass quantity of hydrocarbon
16-7
-------
released per unit of time, i.e., pounds per day; this
is defined as emissions. API bulletins published
subsequent to 1976 will include factors for both
purposes.
Figure 6 shows a summary of emission factors from
83 tests performed by four companies loading motor
gasoline into ships and barges. These factors average
about 40 percent of previously published data. The
chart format also suggests the parameters that were
found to dominate the evaporation phenomena. Filling
a container with petroleum product produces emissions
by displacement of the existing vapor space. The
hydrocarbon concentration in the vessel vapor space
was found to consist of two distinct components:
vapors that existed before loading started - the
arrival vapors, and vapors generated by the loading
operation. Figure 7 depicts this phenomenon. Arrival
vapors in ships were found to be quite lean, averaging
about two to seven percent hydrocarbon, while the
average hydrocarbon content of emitted vapors for the
entire loading sequence was only five to 11 percent.
The generated vapors form a rich blanket, about
four to eight feet deep, that floats on the liquid
level in the arrival vapors. Tankers have compartments
16-8
-------
typically 50 to 60 feet deep. Loading terminates
within a few feet of the deck. Therefore, displaced
vapors reflect about 45 feet of arrival vapors and two
to six feet of the rich blanket.
Barge loading is similar to tanker loading except
that the shallow compartment depths, about 12 feet,
permit more of the rich blanket to escape. The emission
profile is similar, with slightly higher hydrocarbon
levels, but still lower than previously anticipated.
The concentration of arrival hydrocarbons reflects
the history of each vessel compartment since the
previous unloading. Low levels generally indicate that
the compartment was cleaned on the return voyage,
perhaps to accommodate an unleaded gasoline in a
compartment that previously held leaded product, to
switch to heating oil from gasoline service, or indirect
cleaning by ballasting the compartment for navigational
reasons. For any specific marine terminal, the
percentage of compartments that arrive clean annually
can be determined by analysis of ships Togs and
terminal records. With this information the factors in
Figure 6 can be used with confidence in predicting
emission levels when loading motor gasolines.
Other factors are still required for loading crude
16-9
-------
oils and for emissions resulting from water ballasting
of vessels after unloading either gasoline or crude.
The Western Oil and Gas Association has just
completed a test program on loading California crudes
and will publish a report shortly. Preliminary test
results indicate phenomena similar to gasoline loading,
with emission factors generally lower than gasoline.
Also, eight oil companies, acting in response to an
EPA 114A letter, have started a joint program to
develop new data on ballasting emissions, plus any
information necessary to fill gaps in the previously
described API and WOGA studies. Correlation of all of
these efforts should provide the data base necessary
for publication of new, comprehensive marine terminal
emission factors by mid-1978.
Floating-Roof Tanks
Emissions from floating-roof tanks have been
under intensive investigation since last summer. While
much of the data in the original API Bulletin (2517)
was about 40 years old, the relatively slow evaporation
rate from floating-roof tank seals and the difficulty
in measuring the evaporation loss have generally
frustrated any new in-depth studies.
Figure 8 depicts a floating-roof tank and the
16-10
-------
perimeter seal area through which losses occur. The
amount of evaporation occurring in a reasonable amount
of time, say several months with the tank dormant, is
too small to detect reliably with conventional
measuring techniques.
Two methods have been used within the last year in
an attempt to overcome this problem. As shown in the
figure, a field tank with product in it is taken out of
active service and permitted to weather. Separate
testing has established that any evaporation due to
tank wall wetting as the roof descends in normal
operations is insignificant. Therefore, any emissions
are those that result from loss of lighter gasoline
components through the seal area. Further, tests have
also shown that natural convective mixing maintains
a homogeneous liquid in the tank. This establishes the
format for one test approach. Periodic samples are
taken and the density of the bulk liquid remaining is
determined with an extremely precise densitometer.
Sensitivity is required, in grams per millilitre, to
five decimal places. Establishing a time rate of
density increase (increase due to loss of the lighter
components) permits calculation of decrease in bulk
volume.
16-11
-------
A test program, using the density change method
and involving 13 gasoline tanks, was completed in
January by the Western Oil and Gas Association.
Average evaporation loss was about half of previously
published values.
A second test program, also completed recently,
was conducted by Chicago Bridge and Iron Company and
Standard Oil of Ohio. Here a 20 foot covered pilot
tank, designed to simulate climatic and product
variables, was used to capture and account for all
hydrocarbon losses. This was a unique and pioneering
effort. Being capable of isolating individual
variables, the study program developed a clearer
understanding of the significance of the parameters
affecting evaporation. Wind in particular was
determined to play a dominant role in the evaporation
process. The test program was designed for a specific
client's application, and within that framework -
the latest in tank and seal technology for storage of
crude oil - indicates emission levels at about 10 to
20 percent of previously published values.
These programs cannot be extrapolated to a data
base adequate to provide a comprehensive emission
predicting method for all types of floating-roof tanks
16-12
-------
in all geographical and service conditions. CELM has
started a program to complement, correlate and conclude
the two efforts just described and to develop new
bulletins 2517, Evaporation Loss from Floating-Roof
Tanks, and 2519, Use of Internal Floating Covers and
Covered Floating Roofs to Reduce Evaporation Loss,
applicable to any open top or covered floating-roof
tank. That work is scheduled for completion by early
1979.
Fixed-Roof Tanks
Current correlations for fixed-roof tanks are
confined to products with a vapor pressure greater than
two psia, and due to a limited data base, to tanks with
diameters less than about 150 feet. Therefore, some
crude oils, residuals and distillates stored in larger
modern tanks are excluded. It is desireable to
establish emission levels for these products. The
Western Oil and Gas Association has just completed a
46-tank test program on crude and distillate liquids
that provide some new data on emission levels. The
study demonstrates the apparent significance of the
way a tank is operated, e.g., continuous in-out flow,
fast filling and-slow emptying, etc., and the
importance of vapor pressure on loss, and generalizes
16-13
-------
that emission levels are about half of API 2518
estimates. This agrees roughly with work done last
year in Germany indicating that filling losses are 88
percent and breathing losses 11 percent of API 2518
estimates. However, no new emission formulae were
developed by either program.
Two API committees, including CELM, are advancing
a joint program, building on these recent studies, to
produce a new comprehensive prediction method
applicable to all service conditions. A revised
Bulletin 2518, Evaporation Loss from Fixed-roof Tanks,
is planned for publication in 1978.
Truck and Tank Car Emissions
Emissions from these mobile sources are the final
area of CELM's immediate concern. Advancements in
loading methods, e.g., bottom filling vs. submerged
fill pipes or splash loading, and the effect of vapor
recovery units must be evaluated. The effects of
Stage 1 service station return vapors will also affect
evaporation loss. Much information on these items has
already been collected by industry and regulators.
Emission levels calculated by Bulletin 2514 again
appear to be overstated, particularly for the
increasingly popular bottom loading. Bottom loading
16-14
-------
was not common when Bulletin 2514 was published in 1959.
The industry has traditionally used correlations for
the apparently similar submerged loading which have
been found to produce emission estimates about one-
third higher than the current testing of bottom
loading systems indicates.
CELM is analyzing the latest test data to verify
its adequacy for new vehicle loading correlations.
Any additional testing required to ensure a comprehen-
sive data base should be completed this year. A
totally re-written Bulletin 2514B is scheduled for
publication in early 1978.
Summary
The CELM organization chart in Figure 9 shows
the major areas of evaporation loss activity of
immediate concern to regulators and industry-
The existing evaporation loss bulletins, originally
prepared for economic comparisons, have been shown to
generally overstate emissions, restricting their proper
use to industry and for control strategies. All of
the pertinent bulletins are in some stage of updating,
with publication schedules being expedited to make
the new information available as soon as possible.
These programs are expensive and extensive. Total
16-15
-------
cost of testing will approach $2 million. During the
data accumulation period, cooperation by industry and
government will be required to ensure a comprehensive
data bank satisfactory to each-
16-16
-------
An'if;iirini Petroleum Institute
MAI 1 OUGANIZATION
ASST. TO EXECUTIVE
VICE PRESU1CNT I
SENIOR AOVIXOH,
LABOR RELATIONS
MM Hntock
iiiiii.cr-.ii-. of
MA.'MOl'Ml'NI
AMI DUUC.ni
" " Will. I-'.
U Ki'i.li-
GOVT. AFFAIRS
«
VICE PRESIDENT
S. P.
Pottor
FEDERAL RELATIONS
C. E. Sandier
TAXATION
T. A. Mllrlin
STATE RELATIONS
F J. Jandrowitz
1
CENTRAL REGION
E. H. Slcarns
EASTERN
REGION
N C L Brown
1
II Mo
Ind Nob
lo NO
Kins SO
Mich Wise
M-nn
Ct Mass Pa
Del NH Rl
NJ VI
Mo NY Va
Md On VWo
1
SOUTHERN REGION
W. S. Haga
J
1
1
Ala Ga NC
AiK Ky SC
Fla Miss Tcnn
+
VICE PRESIDENT
C 1 Sawyer
EXPLORATION
AFFAIRS
W. M U.'J
MMIKI.IING
Q W Cycil
ni.l'INING
n. M. wnfliu
V. K Leonard
PRODUCTION
R. F. Cartoon
NCN-Di PAOTMENTAL
W. N lii-vs.irrt
VICE PRESIDENT
D. B. Ha'.hbun
ENVIRONMENTAL
AFFAIRS
A. £
FEOLRAL AGENCIES
A. I) Mooro
FINANCE AND
ACCOUNTING
H H, Slowafl
MED. & CIO. SCIENCE
N. K. Weaver
FIRE & SAFETY
COORDINATION
j 0. Mannjy
POLICY ANALYSIS
0. T. Palton
Figure 1
SEPT 1977
-------
API COMMITTEE ON PETROLEUM MEASUREMENT
MAY 31,197?
COMMITTEE ON
PETROLEUM MEASUREMENT
CHAIRMAN
fl.A HAHTMANN
API BOARD COMMITTEE ON
PUBUC ISSUES
CONTACT
R THOMAS
API NON DEPARTMENTAL
WDUSTRY AFFAIRS STAFF
HW. SEW/WO
M. WALTERS
STANDARDIZATION
VICE-CHAIRMAN
B.MESSER.JR
SPECIAL PROJECTS
VICE-CHAIRMAN
AE aflvscw
00
COMMITTEE ON
STATIC
MEASUREMENT
CHAIRMAN
U. HILLBURN
S/C ON NBS
PHYSICAL PROP.
DATA PROJECT
CHARMAN
RMECKEfl
COMMnTEE ON
DYNAMIC
MEASUREMENT
CHAIRMAN
EAfcAlUSTEH
API/ISO/OIML
S/C ON PS.5
CHAIRMAN
COMMITTEE ON
NATURAL GAS
FLUIDS
MEASUREMENT
CHAIRMAN
0. KEMP
*
S/CONAGA3
ORIFICE METER
PROJECT
CHAIRMAN
e.BUXTON
/
\
X
"N
COMMITTEE ON
EVAPORATION
LOSS
MEASUREMENT
CHAIRMAN
J.ZABABA
\L
^
**
S/C ON DATA
FLOATING
ROOF TANKS
CHAIRMAN
a GOOD
COMMITTEE ON
MARINE LOSS
CONTROL
CHAIRMAN
A. GRIFFITH
COMMITTEE ON
PERSONNEL
TRAINING
CHAIRMAN
Fl. BOYLE
ADVISOR TO US
A/COIML
Ml. HALL
COMMITTEE ON
US. INT1.
TRADE
COMMISSION
CHAIRMAN
LOOOGION
TASK GROUP
ON LONG RANGE
PLANNING
CHAIRMAN
K. BAILEY
ADVISOR TO API
METRIC
TRANSITION
COMMfTTEE
COMMITTEE ON
PROGRAM
AND AWARDS
CHAIRMAN
MMOASSIFED RESEARCH PROJECT MANAGEMENT OROUP
* *UNCLASSIHED fCSEARCH PROPOSAL GROUP
-------
API EVAPORATION LOSS BULLETINS
API BULLETIN 2512: TENTATIVE METHODS OF MEASURING EVAPORATION LOSS
FROM PETROLEUM TANKS AND TRANSPORTATION EQUIPMENT (1957)
API BULLETIN 2513: EVAPORATION LOSS IN THE PETROLEUM INDUSTRYCAUSES
AND CONTROL (1959)
API BULLETIN 2514: EVAPORATION LOSS FROM TANK CARS, TANK TRUCKS, AND
MARINE VESSELS (1959)
API BULLETIN 2515: USE OF PLASTIC FOAM TO REDUCE EVAPORATION LOSS (1961)
API BULLETIN 2516: EVAPORATION LOSS FROM LOW-PRESSURE TANKS (1962)
API BULLETIN 2517: EVAPORATION LOSS FROM FLOATING-ROOF TANKS (1962)
j? API BULLETIN 2518: EVAPORATION LOSS FROM FIXED-ROOF TANKS (1962)
API BULLETIN 2519: USE OF INTERNAL FLOATING COVERS FOR FIXED-ROOF TANKS
TO REDUCE EVAPORATION LOSS (1962)
API BULLETIN 2520: USE OF VARIABLE-VAPOR-SPACE SYSTEMS TO REDUCE EVAP-
ORATION LOSS (1964)
API BULLETIN 2521: USE OF PRESSURE-VACUUM VENT VALVES FOR ATMOSPHERIC
PRESSURE TANKS TO REDUCE EVAPORATION LOSS (1966)
API BULLETIN 2522: COMPARATIVE METHODS FOR EVALUATING CONSERVATION
MECHANISMS FOR EVAPORATION LOSS (1967)
API BULLETIN 2523: PETROCHEMICAL EVAPORATION LOSS FROM STORAGE TANKS
(1969) ,
Picture 3
-------
API STAFF
NJ
O
S/CON
FLOATING-
ROOF TANKS
2517
G.J. GOOD
S/CON
INTERNAL
FLOATING
COVERS
2519
R.C.KEftN
COMMITTEE ON
EVAPORATION
LOSS
MEASUREMENT
CHAIRMAN
J.G.ZABAGA
S/CON
FIXED-ROOF
TANKS
2518
SPECIAL LIAISON
J.R. ARNOLD
S/CON
TANK CARS,
TANK TRUCKS,
MARINE VESSELS
2514
R.L.JOHNSON
S/CON
TEST
METHODS
2515
A.D. WHITE
Figure 4
-------
FLOATING ROOF
S/C 2517
o o o o 00
INTERNAL DECK
S/C 2519
FIXED ROOF
CLOSEDTOP &
NO FLOATING ROOF
TRUCKS OR TANK CARS
I I
SPLASH
LOADING
S/C 2518
S/C 2514
OR
I I
BOTTOM
LOADING
I T
SUBSURFACE
LOADING
BARGES
TANKERS
Figure 5
16-21
-------
API 2514A
MARINE VESSEL LOADING OF GASOLINES
SUMMARY OF AVERAGE HYDROCARBON EMISSION FACTORS
T VESSELS
to
SHIPS
BARGES
ARRIVAL
CONDITIONS
CLEANED
UNCLEANED
CLEANED
UNCLEANED
NO. OF COMPART-
MENTS TESTED
50
21
1
11
EMISSION FACTORS
(lbs./1000
GALLONS LOADED)
1.3
2.5
1.2
3.8
Figure 6
-------
MASTVENT
FLAME ARRESTER
VENT
I
BARGE LOADING
SHIP LOADING
TYPICAL SHIP LOADING EMISSION PROFILE
40
VOLUME %
HYDROCARBON
20
T
T
T
T
T
AVERAGE % HC AT END
OF LOADING *=48% TO 55%
AVERAGE % HC=5.4% T011 %
AVERAGE ARRIVAL % HC=2% TO 7%
10 20 30 40 50
DISTANCE FROM TOP OF COMPARTMENT
Figure 7
16-23
-------
A. FRT EMISSIONS
FLOATING ROOF
SEAL AREA
TANK
SHELL
EMISSIONS
ROOF
LIQUID LEVEL
B. TESTING METHODS
1. DENSITY CHANGE
DORMANT FIELD TANKS
6-12 SAMPLES
2-4 WEEKS
7
DENSITY
GM/ML
TIME
2. PILOT TANK
AIR
HC&CFM
f H
VARY PRODUCT, TEMPERATURE,
PRESSURE, WIND EFFECT & SEAL
CONDITIONS
COLLECT & ANALYZE EMISSIONS
Figure 8
16-24
-------
PLANNED PUBLICATION
OF UPDATED BULLETINS
Ul
EARLY 1979
API STAFF
EARLY 1979
COMMITTEE ON
EVAPORATION
LOSS
MEASUREMENT
CHAIRMAN
J.G. ZABAGA
SPECIAL LIAISON
J.R. ARNOLD
LATE 1978
TRUCKS - EARLY 1978
MARINE-MID-1978
LATE 1979
J
I
S/CON
FLOATING
ROOF TANKS
2517
G.J. GOOD
I
S/CON
INTERNAL
FLOATING
COVERS
2519
R.C.KERN
I
. S/C ON
FIXED-ROOF
TANKS
2518
I
S/CON
TANK CARS,
TANK TRUCKS,
MARINE VESSELS
2514
R.L JOHNSON
S/CON
TEST
METHODS
2515
A.D.WHITE
' Fiqure .9
-------
HYDROCARBON EMISSIONS
FROM FLOATING ROOF STORAGE TANKS
Presented at the 1977
Environmental Protection Agency
Emission Inventory/Factor Workshop
Raleigh, North Carolina - September 13-15, 1977
By
Dr. Robert L. Russell
Research Chemist, Union Oil Research and
Chairman of Western Oil and Gas
Association Floating Roof Tank Task Force
P. 0. Box 76
Brea, California 92621
17-1
-------
Summary
The Western Oil and Gas Association has funded several projects
designed to quantify the hydrocarbon emission rate from floating-roof
tanks in which volatile petroleum products are stored. These projects
have involved measurements of emissions from petroleum product storage
tanks in Southern and Northern California, and from a pilot scale test
tank in Illinois where the environmental factors are controllable.
Several important conclusions were drawn from these studies:
(1) Emissions from in-service floating-roof tanks are lower than
predicted by the American Petroleum Institute Bulletin 2517
equation.
(2) Wind is the major driving force causing emissions from
floating-roof tanks.
(3) Good metallic shoe seals control emissions better than good
toroidal seals which do not contact the stored liquid surface.
(4) Gaps between a toroidal seal and the tank shell can sub-
stantially increase the emission rate.
(5) Gaps between a shoe seal and the tank shell result in an
appreciably increased emission rate only when the gaps
extend continuously over about 50% or more of the tanks
circumference.
(6) The inherently low emissions from floating-roof tanks can
be further reduced by placing secondary seals above the
primary seals.
17-2
-------
(7) Secondary seals reduce emissions effectively even when
gaps are present in both primary and secondary seals.
(8) With reasonably well fitting secondary seals extending
from the roof to the tank shell one can expect about the
same emission rate for equivalent tanks of either welded
or riveted construction.
History of Floating-Roof Tank Seal Controversy
in Southern California
The Los Angeles County Air Pollution Control District had a rule
for many years Which required that floating-roof tanks storing a
petroleum distillate with a vapor pressure greater than 1.5 psia must
have a seal which "closes the space" between the floating roof and
the tank shell. Early in 1976 California Air Resources Board (CARS)
Personnel inspected floating roof petroleum storage tanks at a
Southern California refinery. As a result of these inspections, CARB
alleged that these tanks were causing "massive violations" of air
pollution control laws and that the district's rule and enforcement
procedures were not adequate to deal with the situation.
The latter allegation was based on CARB's interpretation of the
phrase "closes the space". CARB interpreted this phrase to mean
that there should be no gap between the seal and the tank shell,
while the district had traditionally interpreted it to mean that a
reasonably well functioning seal was present. Following a public
hearing CARB ruled that the enforcement procedures were inadequate
17-3
-------
and consequently assumed joint enforcement authority with the
district.
The allegation of excessive emissions was based on a GARB
calculation which purported to show floating-roof tank hydrocarbon
emissions from this one particular refinery to be 3000 tons per year.
This number was obtained by using an American Petroleum Institute
(API) equation for floating-roof tank evaporation loss and
arbitrarily increasing the calculated amount of emissions obtained
for tanks with "poor" seals by up to four-fold.
Both GARB and the Southern California Air Pollution Control
District (SCAPCD) conducted public hearings which involved the
District's "Storage of Organic Liquids" rule (Rule 463). By mid-1976
GARB decided that gaps between a floating-roof seal and the tank
shell anywhere along the circumference could not be greater than 1/8"
in width. However CARB delayed enforcement of this extremely strin-
gent gap criterion (referred to as a "no-gap" criterion) until
February 1, 1977. The delay was allowed because of testimony that
the Western Oil and Gas Association (WOGA) was initiating an experi-
mental study of hydrocarbon emissions from floating-roof tanks.
CARB indicated that Rule 463 might be changed if the study demon-
strated that the CARB staff calculations were erroneous. If the
study indicated that the CARB calculations were reasonably accurate,
then all floating-roof tanks would have to immediately comply with
the extremely strict gap criteria.
17-4
-------
During the course of the WOGA studies GARB and other environ-
mental agencies personnel were invited to regular monthly status
report meetings. These meetings provided for a free and open dis-
cussion by all interested parties. All suggestions received at
these meetings were carefully reviewed and incorporated within the
program if they were reasonable and achievable within the time and
cost constraints. This free exchange of information proved to be
very beneficial in ensuring that all important factors were considered
and tested during the study.
As soon as each of the project final reports was completed it
was made available to GARB and all interested parties. The openness
with which these studies were conducted contributed to their credibil-
ity with GARB and other environmental agency personnel. GARB subse-
quently used some of the conclusions, but in the opinion of WOGA
personnel they did not use all of the pertinent conclusions, in their
justification of recommended modifications to Rule 463.
The data from the WOGA study revealed that the original GARB
calculations were erroneous. However, the study also revealed that
a secondary seal could reduce the already low emissions. Therefore
GARB decided to relax the gap criteria slightly and to require
installation of secondary seals on all floating-roof tanks. WOGA
and petroleum company representatives noted that the new GARB gap
criteria were still not justified by the.available data. GARB
subsequently proposed gap criteria equivalent to the gaps tested in
17-5
-------
a pilot scale test tank and made it clear that any further
relaxation of the gap criteria would require more data from the
pilot scale test tank. Although WOGA was convinced that the data
base was broad enough to justify larger gaps than had been tested,
they contracted for additional tests to unequivocally prove this.
The WOGA Studies
Industry personnel generally agreed that the API Bulletin 2517
equation overestimated floating-roof tank emissions for modern
floating-roof tanks. This belief was based on the following facts:
(1) the equation is based on data obtained by many different
companies between the 1920's and the 1950's, and (2) the data were
obtained to demonstrate the superiority.of floating-roof versus
j
fixed-roof storage rather than to quantify emission levels. WOGA
decided that obtaining new, hard data was the only way to ensure
adoption of a rational rule. . *
WOGA appointed a Task Force and established three objectives:
(1) to determine hydrocarbon emissions from floating-roof tanks as
a function of seal gap size, vapor pressure, wind velocity and other
important variables; (2) to determine what constitutes best available
seal technology and to estimate hydrocarbon emissions from use of
that technology; and (3) to compare hydrocarbon emissions from exist-
ing floating-roof tanks with estimated hydrocarbon emissions from
the use of best available seal technology.
17-6
-------
Part (1) of the objective was further subdivided into three
tasks: field test of actual floating-roof tanks, pilot-scale testing,
and laboratory testing. Standard Oil of Ohio (SOHIO) generously made
available data gathered in a pilot-scale study performed for it by
(2)
the Chicago Bridge and Iron Company (CBI). That study involved
measuring emissions from a model floating-roof tank, equipped with
a toroidal primary seal, under controlled conditions. WOGA subse-
quently contracted with CBI to investigate the level of emissions in
the model floating-roof tank when equipped with a shoe primary seal.
Field Testing of Floating-Roof Tanks
The WOGA Task Force decided that the density change method
(3)
described in API Bulletin 2512 would be used for the field test
work. Engineering Science, Inc. (ES) from Arcadia, California was
selected to perform the study. CARB, Environmental Protection Agency,
SCAPCD, and the San Diego APCD accepted WOGA's invitation to fully
participate in the study.
The density change method relies on the fact that lighter ends
evaporate faster than heavier ends from a hydrocarbon mixture. There-
fore in a storage tank, with no liquid flow in or out, the weathering
process increases the stored liquid's density. By use of a density-
evaporation curve and with knowledge of the initial and final density
and the Volume of stored liquid the evaporation losses can be deter-
mined. It proved to be necessary to use a highly precise Mettler-Parr
density comparator to measure densities to five decimal places.
17-7
-------
Because of the expected low emission levels it was necessary
to float a 3 to 5 foot layer of hydrocarbon on water to ensure that
the density change over the expected three to four month storage
could be observed. Extensive laboratory testing demonstrated that
the water had no significant influence on the hydrocarbon density
(4)
during the storage period .
In general, samples were taken via water displacement into
narrow neck.8-ounce bottles at 12 different roof leg support sleeves.
The 12 sampling positions were selected to represent approximately
12 equal volume elements within a tank. The data demonstrated that
vertical and horizontal stratification, if it does occur, is so
slight that it does not affect calculation of the average stock
density.
Tanks in the field study represented a variety of different
seals, products, roof heights, gaps, and tank wall types. Typical
seal designs represented in the field study are shown in Figures I
and II.
The average observed hydrocarbon emission rate from 13 of the
study tanks is compared in Figure III to the API 2517 calculated
emission rate. It can be seen that the observed emissions are
approximately 1/2 of the calculated emissions instead of the 2 to 4
times greater that the GARB staff originally estimated.
17-8
-------
.Pilot Scale Testing of Floating-Roof Tank Emissions
Chicago Bridge & Iron Company (CBI) built an insulated,
temperature-controllable, 20 foot diameter, 9 foot high tank with
(2a)
a double deck floating-roof and a cone roof cover. The tank is
designed to permit blowing air through the space between the floating
and cone roof. A flame ionization detector is used to determine the
hydrocarbon concentration in the outlet and inlet air. The differ-
ence between these two numbers is a direct emission measurement at
the preset operating conditions.
/2)
The SOHIO/CBI experiments demonstrated that wind speed had
a pronounced effect on emissions from a floating-roof tank equipped
with a toroidal type seal. As Figure IV shows, subsequent WOGA/CBI
experiments demonstrated that while emissions from a floating-roof
tank equipped with a shoe seal are wind speed dependent, the effect
is much less pronounced than with a toroidal seal.
The SOHIO/CBI and WOGA/CBI data also demonstrated the effect
of introducing gaps between the seals and the tank shell, and the
effect of placing a secondary seal from the roof to the shell above
the primary seal. (Typical examples of secondary seals are shown in
Figure V.) Figure VI shows a data summary depicting the effect that
various size gaps in primary and secondary seals have on emissions.
Several important conclusions can be drawn from the data
presented in Figure VI:
(1) If no secondary seal is present, toroidal seal gaps increase
17-9
-------
emissions substantially over a "no-gap" toroidal seal.
(2) If a secondary seal is present above a toroidal seal, the
emissions can be lower than with the "no-gap" toroidal seal
even when overlapping, diametrically opposed 1/2" wide gaps
over 6.4% of the tank circumference are present in both seals.
(3) If no secondary seal is present the emissions from a shoe seal
system are less than from a "no-gap" toroidal seal system even
when gaps up to 1-3/4" for 39.6% of the tank circumference are
present in the shoe seal system.
(4) Emissions from a shoe seal system increase appreciably only
when the gaps extend continuously for a considerable percentage
of the tank circumference.
(5) If a secondary seal is present above a shoe seal the emissions
can be lower than the "no-gap" shoe seal even when substantial
gaps are present in both seals.
(6) A secondary seal on a riveted tank is an effective emission
control device even under worst-case conditions.
Emissions Comparisons
Figure VII shows some comparisons between predicted emissions
using the API-2517 equation and the CBI data base. Bars numbered 1
through 3 and 8 are for a welded tank equipped with a shoe seal.
Bars numbered 4 through 7 are for a riveted tank equipped with a
shoe seal. Bars 9 and 10 are for a welded tank equipped with a
toroidal seal.
17-10
-------
Bars numbered 1, 4, 5 are calculated by the API-2517 equation
for a welded tank with or without a secondary seal, a riveted tank
without a secondary seal, and a riveted tank with a secondary seal,
respectively. Bars numbered 2, 3 and 6, 7 are calculated by using
appropriate emission values from the CBI data for various ranges of
circumferential gap openings between the primary shoe seal and tank
shell, and weighting these emission values by the percent of time the
respective gap openings are expected to be encountered in a field
tank. The latter numbers were determined from the tank survey data
(4)
in the Engineering Sciences reports. For bars 3 and 7, the CBI data
for 3/4" gaps over approximately 10% of the tank circumference in the
secondary seal were used. Bars numbered 8, 9 and 10 use the CBI data
for a "no-gap" shoe seal, "no-gap" toroidal seal, and a 1/2" x 6.4%
of circumference gap in both the toroidal and secondary seals,
respectively.
The following conclusions can be drawn from Figure VII:
(1) Emissions from shoe seal equipped floating-roof tanks are
considerably lower than predicted by the API-2517 equation.
(2) The average emission levels from welded or riveted tanks
equipped with shoe seals and no secondary seals are low and
differ only slightly between the two tank types.
(3) The presence of a reasonably well fitting secondary seal which
goes from the roof to the tank shell can further reduce the low
emissions of a shoe-sealing system.
17-11
-------
(4) The average emissions expected from a sealing system which
employs both a primary and a secondary seal, with gaps in both
seals, are lower than the emissions from a "no-gap" primary
seal alone.
(5) When a reasonably well fitting secondary seal, which goes from
the roof to the tank shell is used, the average emissions from
equivalent welded and riveted tanks should be about the same.
Acknowledgments
The efforts of the WOGA Task Force members are gratefully
acknowledged. These members were:
Jerry Adams
Earl K. Dewey, Jr.
Dennis Dykstra
John A Glaser
Gordon J. Good
Hayden H. Jones
Peter E. Jonker
Peter L. Mehta
Richard A. O'Hare
William J. Porter
Robert M. Stoneham
Fletcher Oil & Refining Co.
Continental Oil Company
Chevron U.S.A., Inc.
Gulf Oil Company, U.S.
The Standard Oil Co. of Ohio
Union Oil Co. of California
Union Oil Co. of California
Atlantic Richfield Company
Shell Oil Company
Chevron, U.S.A., Inc.
Texaco, Inc.
Wilmington, CA.
Ponca City, OK.
El Segundo, CA.
Santa Fe Springs, CA.
Cleveland, Ohio
Wilmington, CA
Los Angeles, CA.
Carson, CA.
Carson, CA.
El Segundo, CA.
Wilmington, CA.
17-12
-------
References
API Bulletin 2517, Evaporation Loss From Floating-Roof Tanks,
American Petroleum Institute, New York, 1962.
(a) SOHIO/CBI Floating Roof Emission Test Program,
Preliminary Information, Chicago Bridge & Iron Company,
August 27, 1976.
(b) SOHIO/CBI Floating Roof Emission Test Program, Interim
Report, Chicago Bridge & Iron Company, October 7, 1976.
(c) SOHIO/CBI Floating Roof Emission Test Program, Final
Report, Chicago Bridge & Iron Company, November 18, 1976.
API Bulletin 2512, Tentative Methods of Measuring Evaporation
Loss From Petroleum Tanks and Transportation Equipment, American
Petroleum Institute, New York, 1957.
(a) Evaluation of Hydrocarbon Emissions From Floating-Roof
Petroleum Tanks, Interim Report, Engineering Science, Inc.
December 1, 1976.
(b) Hydrocarbon Emissions From Floating Roof Petroleum Tanks,
Engineering Science, Inc. January 1977.
(a) Western Oil and Gas Association, Metallic Sealing Ring,
Emission Test Program, Interim Report, Chicago Bridge &
Iron Company, January 19, 1977.
(b) Western Oil and Gas Association, Metallic Sealing Ring,
Emission Test Program, Final Report, Chicago Bridge &
Iron Company, March 25, 1977.
(c) Western Oil and Gas Association, Metallic Sealing Ring,
Emission Test Program, Supplemental Report, Chicago Bridge
& Iron Company, June 30, 1977.
17-13
-------
(-«
JS
TANK SHELL
Figure I
TYPICAL SHOE SEAL
SHOE
SEAL FABRIC
ROOF
^ -**- -***
PANTAGRAPH HANGER
LIQUID LEVEL
COUNTER WEIGHT
-------
Figure II
TYPICAL TOROIDAL SEAL
-4
>-
Ul
TANK SHELL
SEAL ENVELOPE
RESILIENT
URETHANE FOAM
CURTAIN SEAL
LIQUID
LEVEL
ROOF
HANGER BAR
SEAL SUPPORT RING
RIM
BUMPER
-------
Figure III
EMISSIONS FROM 13 GASOLINE TANKS
WOCA /ENGINEERING - SCIENCE STUDY
AVERAGE
HYDROCARBON
EMISSION RATE
LB/DAY/TANK
125
100
75
50
25
EMISSIONS ESTIMATE USING
API 2517 EQUATION
(TIGHT FITTING SEALS)
ACTUAL MEASURED EMISSIONS
FROM ENGINEERING
SCIENCE STUDY
-------
X
o
-o
J!
o
E
to
O
u
z
g
to
.28
.24
.20
'«
.12
.08
.04
Figure IV
EFFECT OF WIND SPEED
CBI PILOT SCALE TANK
TOROIDAL
SEAL
SHOE
SEAL
2 4 6 8 10
WIND SPEED, MILES PER HOUR
12
17-17
-------
Figure V
TYPICAL SECONDARY SEALS
H-
00
TANK SHELL
TANK SHELL
SEAL FABRIC .
POLYURETHANE LOG
MINI-TOROIDAL
ROOF
1
1
RUBBER WIPER
WIPER
-------
I
vo
Figure VI
EFFECT OF GAPS AND SECONDARY SEALS CBI PILOT SCALE TANK DATA
65
~ 60
-
a -^
A ~-
M»
Z 25
0
!/>
^V
IK
a. 20
«
c
u
7 15
^
^v
4^
/>
5 10
s
1AJ
5
n
X
X
X
X
X
X
X
X
?
TOROIDAL
SEAL
^
X"
^
^
X
X
^
^
x
^
X
^^
1
X
X
X
X
2
^
x1
^
X
X
x
^
X
X
X
^
DATA
R
Ixj
bd
,_
KJ
p^
p£
SHOE SEAL DATA
Cxi
IX
|x
K
P3
X
^
x^
X
X
X
X
X
<.
txt Pxl
p^j K3
Pd 0 Ixl
^
^
i^
X
^
__
ixj
H H
WORST CASE
RIVETED
TANK DATA
X
X
x
X
X
X
x
X
^
X
X
X
X
;>
x
^
^
X
^
^
X
X
^
1
^
X ^d
X ^j
^ ^
PRIMARY SEAL
MAXIMUM GAP, IN.
/. OF CIRCUMFERENCE
NO 1 J_ NO 111
GAPS 7 2 GAPS 1222
NO
GAPS
^ ]J ^ 4^2
33 13 13 1
6.4 6.4
13.6 52.0 52.0 52.0
39.6 39.6 39.6 100 100 100
SECONDARY SEAL
MAXIMUM GAP, IN.
Y. OF CIRCUMFERENCE
Non« Non« 2 Nona None None 4 y Non« None "J" "J
6.4
9.3 7.4
11.1 4.3
Nona 4" 4"
21.2 9.5
-------
Figure VII
EMISSION COMPARISON FOR 150 FOOT DIAMETER TANK STORING 5 PSIA STOCK
EMISSIONS
(Ib/DAY)
ro
o
300 r
250
200
150
100
50
-1
M
TANK TYPE
WELDED
RIVETED
WELDED
PRIMARY SEAL
SHOE
SHOE
SHOE TOROIDAL
SECONDARY SEAL YES or NO NO YES
NO
YES
NO
YES
NO
NO
YES
CALCULATION
METHOD
API- CBI TEST DATA
2517 WEIGHTED BY
OCCURRENCE
OF GAP OPENINGS
IN FIELD TANK
PRIMARY SEALS
API- CBI TEST DATA
2517 WEIGHTED BY
v OCCURRENCE
OF GAP OPENINGS
IN FIELD TANK
PRIMARY SEALS
CBI
TEST
DATA,
NO
GAP
CBI
TEST
DATA,
NO
GAP
CBI
TEST
DATA,
1/2" GAP
6.4% IN
EACH SEAL
-------
QUESTION:
RUSSELL:
QUESTION:
RUSSELL:
QUESTION:
CONDENSED DISCUSSION
I have no problem in inspecting a floating
roof seal. I have a great deal of trouble
inspecting the seal on an internal floater.
Could you explain how you got in there
and measured all those gaps on internal
floaters?
We didn't. We haven't done it on the inter-
nal floaters. Not that I am aware of anyway.
Originally you started out and said that
because of the gaps, the California board
evaluated that emissions were four times
higher. With your new data would the emiss-
ions be two times or three times higher
now?
The new data shows that the emissions are
50% less than what would be calculated using
the straight forward API equation. CARB was
using the API equation and then increasing
it by two to four times, so the new data says
that we are considerably less than by four to
eight times what CARB originally estimated.
I noticed in one of your emission loss
factors for breathing loss that you used a
constant breathing loss factor for I assume
your used control tanks without any vapor
returns.
17-21
-------
RUSSELL:
COMMENT:
RUSSELL:
COMMENT:
RUSSELL:
QUESTION:
RUSSELL:
The vapor return wasn't installed to control
breathing losses. Breathing losses are more
or less uncontrolled.
But you used one factor and you said you based
it upon API. API is based upon the breathing
loss as a function of the amount of tempera-
ture rise during the day.
We had to make some assumptions based upon
the survey data as far as average tank size.
Most tanks were in the same size range.
Well, regardless of size it depends on what
the average temperature was.
Yes, I think we used 15 degree daily tempera-
ture variation.
In the work did you establish the relative
estimated condition level from small bulk
plants as compared to bulk terminals?
No. The studies that we performed were deal-
ing only with bulk plants - less than 20,000
gallons. This data was submitted to EPA to
formulate or to determine how proposed vapor
recovery regulations will effect small bulk
plants. So we weren't really concerned with
large terminals.
17-22
-------
EMISSION INVENTORY
OF PETROLEUM STORAGE AND HANDLING LOSSES
(A CASE HISTORY)
Presented at the 1977
Environmental Protection Agency
Emission Inventory/Factor Workshop
Raleigh, North Carolina - September 13-15, 1977
By
James T. Alexander, Jr.
Regional Engineer
Virginia State Air Pollution Control Board
Northern Virginia Office
Falls Church, Virginia 22043
18-1
-------
Abstract
Gasoline and distillate products emissions were calculated
for the Northern Virginia sector of Metropolitan Washington for
the peak oxidant month of July using factors adjusted for specific
July weather conditions. Sources included floating roof and fixed
roof storage, bulk terminal loading racks, service station deliv-
ery and sales, airports, and small bulk plants.
Calculations were made in Ibs/day and lbs/6-9 a.m. peak period
under summer weather conditions typical for the high oxidant peri-
ods experienced in the region. The basis for emission factors was
the Radian Corporation Study of August 1976 by Burklin and
Honerkamp. Their version of the API empirical equation for float-
ing roof losses was used in the absence of a better predictive
model, recognizing the lack of correlation shown in January 1977
by Engineering-Science.
As expected, this method of inventory produces a great deal
more emissions compared with year-around averaging of gasoline
throughput and nationwide temperatures. Nearly all the differ-
ences are additive. Some are quite large. For example, gasoline
vapor pressure at 79°F is 36% higher than it is at 60°F.
An earlier inventory for 1972 was re-calculated and the
emissions were 70% greater using the July-specific inventory pro-
cedure. The effect of vapor control measures was determined.
Overall reductions of 38% are shown on a Ibs/day basis and 42% on
18-2
-------
a Ibs/peak period basis.
Correction by the Radian Study of the apparently erroneous
data in AP-42 for properties of distillate fuels results in very
small emissions from those products in the Northern Virginia
region.
18-3
-------
EMISSION INVENTORY
OF PETROLEUM STORAGE AND HANDLING LOSSES
(A CASE HISTORY)
James T. Alexander, Jr., Regional Engineer
Virginia State Air Pollution Control Board
Background
The cities and counties of Northern Virginia included in the
National Capital Interstate Air Quality Control Region contain
eight pipeline terminals and fourteen small bulk plants serving a
greater metropolitan area of about three million people. Air
quality exceeds the photochemical oxidant standard by 25% or more
for about 35 days each summer. The Virginia portion of the region
has very little industrial activity and comprises suburban and
rural development characterized as "bedroom" communities. Gaso-
line vapor control measures were initiated in 1973.
Inability to apply, up to this time, an air quality simula-
tion model that relates hydrocarbon and nitrogen oxides emissions
to oxidant concentrations highlighted a need for better emission
inventory of pollutants as an interim yardstick of control effec-
tiveness during continued population growth. One portion of this
inventory effort covered petroleum storage and handling emissions
in Northern Virginia. These were calculated for 1976 and recal-
culated for 1972 using best available emission factors adjusted
for local conditions. The paper describes how this was done and
18-4
-------
compares the results with an earlier, less precise, inventory.
The work involved two steps. First, the calculation of emis-
sion factors based on local conditions and second, the inventory
of individual Virginia sources using those emission factors.
Local Meteorological Conditions
At the outset it was decided to calculate emissions in Ibs/
day for July conditions and to convert those emissions to lbs/6-9
a.m. period. July was determined to be the peak oxidant month.
Based on 30-year weather observations at National Airport, July
has 30% more days with maximum temperature 90°F or higher than
either June or August. Studies have shown maximum temperature to
be a useful surrogate for numerous meteorological variables that
relate to peak oxidant formation in the Eastern U. S. Data for
the 6-9 a.m. period was sought so as to facilitate subsequent use
of the Dodge-Demitriades smog chamber curves as an air quality
simulation model. It will be recalled that the units of NMHC
o
concentration plotted on those curves are in ppmC (6-9 a.m.)
Table I shows local July meteorological factors for Washington
National Airport compared with the nationwide annual values nor-
mally used in calculating evaporative emission factors for use in
the National Emissions Data System (NEDS). Local values for the
6-9 a.m. period were taken for July 1976, a month which happened
to conform quite closely to the 30-year mean for July.
Above ground storage temperature at 6-9 a.m. was assumed to
18-5
-------
coincide with the ambient daily mean. It was considered that
early morning temperature depression from the ambient mean is
probably offset by heating due to delivery of hot products from
the pipelines. Products are known to be received in summertime
as high as 90°F. This assumed product temperature was backed by
sampling delivery truck loading tickets.
For below ground storage, the 6-9 a.m. ambient mean was used.
That value is comparable with a number of local summertime tank
measurements made during Stage I vapor balance testing. Many ser-
vice station tanks are subject to external heating from absorption
through black-top paving, and do not cool down during the pre-dawn
period to the temperature of grass or dirt-covered subsurfaces.
Gasoline Distillation Properties
The largest distributor in the region, EXXON Company (U.S.A.),
was asked to furnish data on Reid vapor pressure and API° gravity
that could be considered typical for summertime gasoline products.
They reported data for their three grades of gasoline that were
marketed locally in July 1976. The sales-weighted values were:
RVP-9.7 and liquid density - 6.16 Ibs/gal.
Table II shows the resulting values of true vapor pressure
at the assumed temperatures shown in Table I. The column titled
"multiple" indicates the local value compared with the nationwide
value as 1. In equations where emissions vary directly with true
vapor pressure, or in the exponential form shown, the summertime
18-6
-------
temperature effect is pronounced.
Empirical Equations for Evaporative Losses
Initial attempts to use the empirical equations developed by
the American Petroleum Institute in 1962, cited in the Air Pollu-
tion Engineering Manual, AP-40 (2nd Edition),3 to derive localized
emission factors were unsuccessful. The losses expressed in the
API equations are in volumetric units (barrels of vapor emitted
per year or per million barrels throughput). Physical properties
of the vapor which are needed to convert emissions to a weight
basis were not available.
Additionally, the empirical equations cited in AP-42 (Supple-
ment One) could not be used to calculate the "standard" evapora-
tive factors contained in Section 4.3 of that document. Although
those equations express losses on a weight basis, the vapor prop-
erties used in the equations apparently differ from those used to
calculate the standard factors in Table 4.3-2. Values for true
vapor pressure of distillate products used in AP-42 are apparently
in error.
As a result of these shortcomings, the Federal EPA awarded
a contract to the Radian Corporation, Austin, Texas to upgrade and
refine the information presently contained in AP-42. Their report,
prepared by C. E. Burklin and R. L. Honerkamp,5 was obtained and
the empirical equations contained therein were used to derive July
emission factors for Metropolitan Washington. (This document is
18-7
-------
hereafter cited as "EPA-450/3").
Since the empirical equations contain factors for tank geome-
try as well as vapor pressure, local data was collected for tank
diameters and tank vapor space height and estimates were made for
paint condition. Average storage tank sizes at the eight bulk
terminals in Northern Virginia were found to be slightly smaller
than the nationwide average used for calculating standard factors
in EBk-450/3. Table III shows these differences. Since tank dia-
meter appears as an exponential function in the equations, these
tank size differences are significant.
Emission Factors for Bulk Terminals
Local emission factors were computed for floating roof stor-
age, fixed roof storage and loading rack emissions using the
pressure, density, temperature, wind speed, and tank paramenters
described above.
A comparison of the local values with the "standard" values
in the Radian study, EEA.-450/3, is shown in Table IV. The column
titled "multiple" shows the ratio of the local factor divided by
the standard factor. The local gasoline factors are 1.15 to 1.30
times the standard factors. This has an important bearing on the
precision of an overall inventory.
The local fixed roof factors for breathing loss are smaller
than standard values because of the smaller average tank diameter
used in the equations. As will be shown, jet kerosene and No. 2
18-8
-------
fuel oil emissions are quite small in comparison with gasoline, so
an emission factor refinement for those products is not as impor-
tant as it is for gasoline. This was not the case when AP-42
standard emission factors were used. The values in Table 4.3-2
of that document for distillate products are 8 times larger for
breathing and 37 times larger for working losses than correspond-
ing values in EPA-450/3.
It is recognized that floating roof emission factors based on
empirical relationships must be regarded with some skepticism. The
Radian Study did not correct basic inadequacies in the API empiri-
cal equation for floating roof standing storage emissions. Recent
tests by Engineering-Science, Inc. on thirteen tanks in California
revealed a marked lack of correlation between observed and calcu-
lated emissions. It appears likely that further testing will
show that well-designed internal floating pans result in less emis-
sions than external floating decks that are subject to more wind
effect.
Efficiency of Gasoline Vapor Recovery Systems
The empirical equations in EPA-450/3 provide a means to
include the effects of gasoline vapor recovery at a known level of
efficiency. Loading rack losses are effected by: (1) whether
trucks return vapor collected by vapor balance at service stations,
and (2) whether the loading racks themselves are controlled by on-
site processing equipment.
18-9
-------
Northern Virginia has Stage I vapor balance control on under-
ground storage tanks at nearly all service stations. It is esti-
mated that 90% of the gasoline throughput is so controlled. Vapor
processing units are installed at each bulk terminal to liquify or
incinerate the vapors displaced from truck compartments during re-
loading .
While the efficiency of these processing units was measured at
92-95%, extensive leakage is currently being experienced at truck
compartment dome covers and truck vapor manifold fittings such that
a significant amount of vapor is discharged during re-loading. For
emissions factor calculation it was assumed that only 76% vapor
recovery occurs at the loading racks. Applying .92 unit efficiency
to .76 recovery efficiency yields an estimated .70 for overall
loading rack efficiency for Northern Virginia terminals.
With that efficiency inserted in the empirical equation for
loading rack losses, the local factor with dedicated trucks in
o
vapor balance service becomes 3.3 lbs/10 gal.
The same computations for 70% loading rack efficiency, but
o
without vapor balance at the service stations, yield 2.0 lbs/10
gal for local temperature conditions.
Table V shows gasoline loading rack losses under each of the
various conditions.
Emission Factors for Service Stations
The passenger car refueling loss equation developed by Scott
18-10
-------
Laboratories in 1972 was not used by Radian Corporation in EPA-
450/3 nor are any empirical equations suggested in that report. It
simply states certain values for service station losses without
defining the standard ambient conditions for which they apply.
If it is assumed that the values in EPA-450/3 for service
station losses are more accurate than earlier references » and
that they are based on the same 60°F conditions used for above-
ground storage, one could factor them up to 75.5°F by applying the
vapor pressure multiplier (1.29), the number shown in Table II.
That of course assumes the losses are directly proportional to true
vapor pressure. For purposes of this inventory, that procedure was
followed and local values were derived as shown in Table VI.
Conversion to the 6-9 a.m. Period
The development of factors to convert emissions in Ibs/day to
lbs/6-9 a.m. period is not susceptible to rigorous derivation. It
is difficult to determine, for example, how much of the total ser-
vice station monthly throughput is handled daily, let alone during
the exact hours between 6 and 9 a.m. Table VII shows the conver-
sion factors estimated for petroleum storage and handling operations
in Northern Virginia during summertime. The percentage is the
portion of 24 hour emissions that occurs during the three hours 6
to 9 a.m.
Storage Capacities and Volume Throughput
One final step before emissions can be computed is to deter-
18-11
-------
mine, for each source and each product, the storage capacities and
volume throughput. Gasoline volumes are greater in summertime and
heating oil volumes are smaller, so for this analysis, source owners
were asked to report volumes for the 31 calendar days of July 1976.
Using July data in the Eastern U. S. has an appreciable effect.
Statewide gasoline tax data for Virginia over the past four years
shows that July sales are between 6% and 10% higher than the 12-
month average.
Volumes were converted to a daily basis by considering July
a 31-day month for 7-day a week operations and a 26-day month for
6-day operations.
Inventory Results
Table VIII shows the inventories for July 1972 and July 1976
based on the local emission factors described herein. Gasoline
sales during the four-year period were up 23%. Control measures
resulted in an overall emission decrease of 35%.
Table IX shows the comparison with an earlier 1972 inventory
based on AP-42 emission factors calculated on a tons per year
basis. The total emissions are 71% greater using the procedure
described in this paper.
Table X shows the July 1976 inventory for the three bulk
terminal locations by petroleum products. The effect of the very
small emission factors for jet kerosene, diesel and No. 2 fuel
oil is shown. Although distillate products account for 22% of the
18-12
-------
throughput they generate less than 2% of the emissions. As noted
previously, floating roof standing storage losses were calculated
from an empirical equation known to be inadequate. Those emissions
are 9% of the bulk terminal losses.
Table XI shows the effect of vapor recovery control measures
in Northern Virginia. Without Stage I vapor balance at service
stations and without vapor processing units at the terminals,
emissions would be increased by 6343 lbs/6-9. This represents
overall control of 42%.
Table XII shows the July 1976 emissions on a 24-hour basis in
Ibs/day instead of lbs/6-9 a.m. Comparing these figures with the
previous data, Table VI, indicates the degree to which a 6-9 a.m.
computation enhances overall control effectiveness. On a 6-9 basis
it is 42%, while on a 24-hour basis it is 38%.
Conclusions
Two general conclusions may be drawn from this analysis.
First, assuming the vapor properties of distillate fuels have been
correctly specified by the Radian Corp. study, evaporative emis-
sions from those products are hardly worth the effort to inventory,
at least in the quantities they are handled in Northern Virginia
in summertime.
The second conclusion is that gasoline evaporative emissions
should be computed for the same ambient temperature and weather
conditions that prevail at the time those emissions have the most
18-13
-------
pronounced,effect oh air quality during the peak oxidant season.
It is to be hoped that future revision of Sections 4.3 and
4.4 of AP-42 will a^id standardized, national average (NEDS)
emission factors for gasoline evaporative emissions and portray
the factors based on regional peak oxidant season conditions.
This could be done with temperature correction factors similar to
those used in calculating vehicle emissions.
18-14
-------
References
1. Meteorological Conditions Conducive to High Levels of Ozone,
T. R. Karl and G. A. DeMarrais, EPA (RTF) Paper Presented
Sept. 12, 1976 at International Conference on PCOX Pollution
and its Control, Raleigh, N.C.
2. Alternatives for Estimating the Effectiveness of State Imple-
mentation Plans for Oxidant. Draft Paper, January 1977, OAQPS,
EPA (RTF).
3. Air Pollution Engineering Manual (AP-40, Second Edition), J. A.
Danielson, Los Angeles APCD, May 1973, (pp. 632-642).
4. Compilation of Air Pollutant Emission Factors (AP-42, Second
Edition), February 1976, OAQPS, EPA (RTP) (Sections 4.3 and
4.4 by W. M. Vatavuk and R. K. Burr, dated July 1973).
5. Revision of Evaporative Hydrocarbon Emission Factors, (EPA-450/
3-76-039) C. E. Burklin and R. L. Honerkamp, Radian Corpora-
tion, Austin, Texas, August 1976 (EPA Project Officer: C. C.
Masser; Contract No. 68-02-1889).
6- Hydrocarbon Emissions from Floating Roof Storage Tanks,
Engineering-Science, Inc., Arcadia, California, Jan. 1977
(Report prepared for Western Oil and Gas Association).
7. Investigation of Passenger Car Refueling Losses. (APTD-1453)
M. Smith, Scott Research Labs, San Bernadino, California,
Sept. 1972, (CRC Project CAPE 9-68).
18-15
-------
TABLE I
Ambient daily mean temp.
Ambient daily A T
Ambient 6-9 am mean temp.
Ambient 6-9 am A T
Average daily wind speed
Average 6-9 am wind speed
Nationwide
Annual
60°
15°
10 mph
Local
July
78.7°
15°
75.5°
/
6°
8.1 mph
7.3 mph
Assumed aboveground
product temperature (6-9 am)
Assumed belowground
product temperature (6-9 am)
60°
60C
78.7°
75.5°
Meteorological Factors
18-16
-------
TA.BLE II
Above Ground Storage
P (psia)
f P V'7
^14.7 - Pj
Belowground Storage
P (psla)
Nationwide
60°F
5.2
.6558
60°F
i
5.2
Local
78.7°F
7.1
.9535
75.5°F
6.7
(Multiple
(1.36)
(1.45)
(1.29)
True Vapor Pressure of Gasoline (RVP 9.7)
18-17
-------
TABLE III
Nationwide Local
103 gal D (ft) 103 gal D (ft)
Floating Roof 2814 110 2298 94.2
Fixed Roof 2814 110 2098 87.1
Average Sizes of Storage Tanks
18-18
-------
TABLE IV
EPA-450/3 Local (Multiple)
Floating RooJ
Gasoline
Fixed Roof
Kerosene
No. 2 FO
Loading Rack
Gasoline
Kerosene
No. 2 FO
E
Storage
Withdrawal
Breathing
Working
Breathing
Working
(uncontrolled)
.033a
.023
.0045a
.027
.0040a
.023
5.0
.02
.01
.038a
.029
.0038a
.044
.0034a
.037
6.5
.03
.02
(1.15)
(1.26)
(.84)
(1.63)
(.85)
(1.61)
(1.30)
(1.50)
(2.00)
a - lbs/day/103 gal; all others in lbs/103 gal
Bulk Terminal Emission Factors
18-19
-------
TABLE V
(lbs/103 gal) EPA-450/3 Local (Multiple)
Uncontrolled,
w/o vapor balance 5.0 6.5 (1.30)
Uncontrolled,
with vapor balance 8.0 11.0 (1.38)
70% control
w/o vapor balance 2.0
70% control
with vapor balance 3.3
Gasoline Loading Rack Factors (Submerged Fill)
1&-20
-------
TABLE VI
(lbs/103 gal) EEA.-450/3 Local
Subm. Fill U. G. Tanks
(w/vapor balance) .30 .39
Breathing Loss U. G. Tanks
(after fill) 1.0 1.3
Vehicle Refueling 9.0 11.6
Nozzle Drip and Spill .7 .7
Service Station Emission Factors
18-21
-------
TABLE VII
Gasoline 6-9 a.m. Factor
Loading Rack 24
Filling U. G. Tanks 24
Filling Vehicle Tanks 18
Jet Kerosene
Loading Refuelers 18
Loading Aircraft 18
Jet Kero, Diesel. No. 2 FO
Loading Rack 24
Filling U. G. Tanks 24
All Products
Storage 12.5
Peak Period Conversion Factors
18-22
-------
No. Virginia
Bulk Terminals
Bulk Plants
Airports
Controlled Serv. St.
Uncontrolled Serv. St.
Totals
TABLE VIII
July 1972
lbs/6-9 °
July 1976
lbs/6-9 %
6377
349
209
0
6365
13,300
47.9
2.6
1.6
-
47.9
100.0
3764
195
37
3877
784
8657
43.5
2.3
.4
44.7
9.1
100.0
THC Evaporative Emissions
18-23
-------
TABLE IX
No. Virginia
Bulk Terminals
Bulk Plants
Airports
Controlled Serv. Sta.
Uncontrolled Serv. Sta.
Original
Annual 1972
lbs/6-9 _ %
Re-calculated
July 1972
lbs/6-9 _ %
3200
180
200
0
4220
41.0
2.3
2.6
0
54.1
6377
349
209
0
6365
47.9
2.6
1.6
0
47.9
7800
100.0 13,300
100.0
Re-calculation of 1972 Inventory
18-24
-------
TABLE X
No. Virginia Emissions (lbs/6-9) from
Bulk Terminals
Fairfax
New ing ton
Manassas
Totals
L. R. Volume (103 gals)
Gasoline
1813
1674
225
3712
4105
Distillate
28
23
1
52
1170
(1.4%)
(22.2%)
July 1976 Inventory for Bulk Terminals
18-25
-------
TABLE XI
No. Virginia
July 1976
Bulk Terminals
Bulk Plants
Air pert: ts
Controlled Serv. Sta.
Uncontrolled Serv. Sta,
Control Effectiveness
Present
Emissions
(lbs/6-9)
3764
195
37
3877
784
8657
If
Uncontrolled
(lbs/6-9)
6821
266
51
0
7862
15000
(46343)
42%
Effect of Vapor Control Measures (6-9 am basis)
18-26
-------
TABLE XII
No. Virginia
July 1976
Bulk Terminals
Bulk Plants
Airports
Service Stations
Totals
Control Effectiveness
Present
Emissions
Ibs/day
17,060
878
160
24,480
42,578
If
Uncontrolled
Ibs/day
29,800
1,157
220
37,820
68,997
(+26,419)
387.
Effect of Vapor Control Measures (24-hour basis)
18-27
-------
QUESTION:
ALEXANDER:
QUESTION:
ALEXANDER:
QUESTION:
ALEXANDER:
CONDENSED DISCUSSION
At your terminals 1n airports, you had
uncontrolled emissions. Is that looking at
a tank without an internal floating roof?
Yes, I took the Internal floaters and just
factored them off as though they weren't
there. We've got other parts of the state
in Virginia that have no hydrocarbon regula-
tions at all and it's kind of useful for them
to see what would happen if they had the same
regulations that we have in northern Virginia.
You did the analysis for July for the petrol-
eum storage, etc?
That's right.
Now to put those emissions into prospective
with other sources of hydrocarbons. Did you
also calculate hydrocarbons from other sources
for July? Namely automobiles, natural
sources, what have you?
This is a very very good point. If you are
going to run a July Inventory for one source
you should run it for all sources. I regret
to say that we simply don't have the technique
to do this.
18-28
-------
QUESTION:
ALEXANDER:
Do you record any relationship between July
inventory and the rest of the other eleven
months of that year? Is it 102, 20% higher
than an average month?
By comparing the 72 inventories that we had
made - the old July to new July 76 hydro-
carbons were up 70% on the basis of using
the July weather conditions as opposed to
averaging the month of July out of a 12
month inventory.
18-29
-------
INVENTORYING HYDROCARBON EMISSIONS
FROM SMALL GASOLINE BULK PLANTS
Presented at the 1977
Environmental Protection Agency
Emission Inventory/Factor Workshop
Raleigh, North Carolina - September 13-15, 1977
By
R.L. Norton and R.J. Bryan
Pacific Environmental Services, Inc.
1930 14th Street
Santa Monica, California 90404
19-1
-------
This paper describes the technique used to inventory hydro-
carbon emissions for nearly four hundred gasoline bulk plants in the
San Joaquin Valley and San Diego County areas of California, the
Metropolitan Denver area, and the Houston/Galveston and Baltimore/
Washington, D.C. air quality control regions. Updated throughput
information and complete inventories of bulk plant operations were
obtained. Emission estimates were generated for potential as well
as controlled hydrocarbon losses within each study region. The
paper describes the emission factors used and the methodology for
presenting the emission estimates.
Problems encountered while attempting to obtain the inventory
data are discussed and include: 1) the lack of an existing inventory,
2) the variation in the applicability of state regulations which in
turn effected the quality of the data available (i.e. if the state
did not regulate the bulk plants, there would be no information on
them in their files), 3) inability to obtain operating data from some
sources because of confidentiality claims, 4) problems in verifying
all inventory entries from the various sources as bulk plants defined
by the study.
19-2
-------
I. Introduction
Environmental Protection Agency (EPA) regulations for the
storage and transfer of gasoline require bulk plants in certain
specified air quality control regions to control hydrocarbon emis-
sions from their operations. A recent Federal Register (June 7,
1977) defines these air quality control regions and gives background
information on the proposed regulations. These regulations require
vapor recovery systems to be installed and operated in a manner
that will prevent release to the atmosphere of no less than 90 per-
cent by weight of organic compounds in vapors generated during gaso-
line transfer operations. Individual states have submitted control
strategies in State Implemention Plans and enacted laws which fre-
quently provide for the exemption of small bulk plants from hydro-
carbon emissions control regulations. Conditions for granting these
exemptions are not uniform among the states, e.g., throughput limits
often differ. The rationale for allowing exemptions has generally
been based on the anticipated adverse economic impact to the indus-
try or on the estimated minor contribution of bulk emissions to the
area wide hydrocarbon/oxidant levels.
In order to determine whether Federal vapor recovery regula-
tions need revision, the Division of Stationary Source Enforcement
(DSSE) contracted with Pacific Environmental Services, Inc. (PES)
to perform a preliminary investigation of the impact of vapor re-
covery regulations on small bulk plants. This first study focused
19-3
-------
on bulk plants in the San Joaquin Valley and San Diego areas in
California and the Denver, Colorado area. A similar survey of bulk
plant operations in the specific areas surrounding Baltimore, Mary-
land and Washington, D.C. and Houston/Galveston, Texas was also per-
formed to determine whether the descriptive, market and economic
data presented in the first report could be adequately applied to
other areas of the country. The tasks completed for both studies
were quite similar:
1. Provide an inventory of bulk plants.
2. Describe facilities and vapor recovery equipment
at the bulk plants.
3. Classify these bulk plants by throughput.
4. Determine types of customers and volume dis-
pensed to non-exempt accounts, agricultural
accounts and accounts with small tanks.
5. Determine the financial profile of typical
bulk plants.
6. Estimate the long and short term economic
effects of installing and maintaining vapor
recovery equipment, particularly with respect
to the number of anticipated plant closures
or plant start-ups.
7. Estimate emissions from bulk plant operations
and the decrease in emissions if controls are
adopted.
This paper describes the survey techniques employed to inven-
tory nearly four hundred bulk plants in the study areas. Methods
for formulating emission estimates and problems associated with
information gathering will also be discussed. Because of the
19-4
-------
current EPA interest in bulk plants, sections have been included
on definition of emission sources and applicable hydrocarbon con-
trol concepts.
II. Description of Bulk Plants
A gasoline bulk plant lies in the gasoline marketing chain
between the large terminal and the ultimate end user. The bulk
plant, normally found in the rural areas, serves small customers
where either the large truck transports cannot or will not deliver.
The studies, within which the inventories were conducted, defined
the small bulk plant to be one whose daily throughput was less than
twenty thousand gallons of gasoline and one where all deliveries
are made to and from the bulk plant by road. Although bulk plants
may handle a variety of products, including gasoline, fuel oil,
diesel fuel, etc., the derived emission estimates were only deve-
loped for gasoline.
The typical facility, as defined by the surveys, consisted of
three or four above ground storage tanks, one delivery truck load-
ing facility, two delivery trucks and a. gasoline throughput of from
five thousand to eight thousand gallons per day.
19-5
-------
III. Inventory of Bulk Plants
A. Information Sources
The first step in inventorying the bulk plants in the five
study areas was to determine the quality of the existing data base.
The applicable EPA regional offices were contracted but little or
no data on gasoline bulk plants existed in their files. Contacts
at the state agencies in the study area were obtained from the
applicable EPA offices.
The quality of the data available at these agencies was usually
dependent upon whether or not the bulk plant emissions were covered
under an applicable State Regulation. A fairly complete inventory
of bulk plants was provided by the Virginia State Air Pollution
Control Board and by the Maryland Bureau of Air Quality and Noise
Control for work in the Baltimore/Washington, D.C. AQCR. The infor-
mation provided included: 1) plant name and location, 2) plant
gasoline throughput, 3) number and types of tanks, and 4) number and
types of trucks, but the Maryland data was from 1974 arid the Vir-
ginia data was from a 1972 survey.
The California Air Resources Board was also contacted as to
the extent of data available on the number of bulk plants in the
San Joaquin Valley AQCR and in San Diego County. The Air Resources
Board supplied a summary of bulk plant operations for each county
but did not indicate specific locations. Since the purpose of the
studies was more than just an inventory and data was needed from the
19-6
-------
individual bulk plant operators, the ARE was again contacted and
asked if they could supply the data from which the summary was
made. The resultant information supplied consisted of copies of
letter heads, business cards, return addresses and hand written
notes. This information, which consisted of over 200 entries, was
then organized by counties and the totals checked against the sup-
plied summary data.
An initial inventory of bulk plants was supplied by the Colo-
rado Department of Health, Air Pollution Control Division and supple-
mented by information from the Oil Inspectors Office. PES was in-
formed that this inventory was the most up-to-date possible, but it
was later found to be inaccurate. For example, the supplied list
consisted of approximately twenty bulk plants. Four of these plants
were out of business and approximately twenty-five additional bulk
plants were identified.
The Texas Air Control Board was also contacted. There was no
data available in the Texas office since the State Regulations were
not applicable to bulk plants in the size range studied.
When there was no available information from the State offices
or it was felt that the data might not be complete, the local air
pollution agencies in the study areas were contacted. This includ-
ed fourteen agencies in California, two in Texas and one in the
Washington, D.C. area. The data obtained ranged from excellent to
none at all, aga^n, due mostly to the applicability of the local
19-7
-------
regulations. In some cases, written requests or personal visits
were required to determine the availability of the data and to en-
sure that the data obtained was pertinent to the project. Written
data requests for each bulk plant included the following:
Bulk Plant Name
Location
Plant Contact (Owner/Operator)
Gasoline Throughput
Storage Tanks: Number of Tanks
Above/Below Ground
Tank Capacity
Physical Size of Tanks
Splash/Submerged Fill
Vapor Recovery Apparatus (If Any)
Pressure-Vacuum Vent Setting & Type
Plot Layout
Delivery Trucks: Size
Vapor Return
Splash/Submerged/Bottom Loading
Loading Rack Controls: Vapor Recovery System
Size of Vapor Return Line
Accounts Breakdown (% Commercial, % Agricultural,
And Percentage Exempt Under
Vapor Recovery Regulations
Letters were sent to the San Joaquin, Stanislaus and Tuolumne
County California Agencies and visits were made to the King and
Kern County Agencies in California, The Texas Air Control Board,
Houston Office and the Galveston County Health District in Texas,
and the Colorado Health Department.
Governmental agencies were not the only sources of information
and often times not the most productive. In each state Involved,
the applicable trade associations were contacted. These included
the California Oil Marketers Association, Texas Oil Marketers
Association, Maryland Oil Jobbers Council, Virginia Petroleum
19-8
-------
Jobbers Association and the National Oil Jobbers Council. In many
cases, the trade organizations were hesitant at first to supply the
information because they were not sure as to the ultimate use of
the data. Once they were assued that the information and the pur-
pose of the project was not enforcement oriented, the data was
generally supplied on the location of bulk plant operations in their
regions. However, this data was limited to operations or operators
that were members of these organizations.
Representatives from the major oil companies operating in each
region were contacted as to the location of bulk plant operations.
In most cases, the oil companies would not release the information
because they felt that this was confidential. Others could not
release the requested data because the operations were run by pri-
vate businesses and the oil companies did not have the details of
the business operations. Where information was provided by the oil
company, (Standard Oil of California and Continental Oil Company in
Denver) the data was included in the respective county inventories.
As a supplement to the sources of information already des-
cribed, telephone directories in the study regions were consulted
to ensure that all bulk plant facilities would be included in the
inventory survey. The facilities developed from the several infor-
mation sources were arranged into county groups to constitute the
initial bulk plant inventory.
19-9
-------
B. Verification of Bulk Plant Inventory
Once the initial list had been generated, it was decided that
a verification of the operations as bulk plants, as defined by the
studies, was necessary. Types of information obtained from the
verification procedures included:
identity of bulk plants which had shut down
identity of plants included in the initial inventory
which were not bulk plants as defined by the study
identity of plants which had ceased handling gasoline
identity of plants which had moved to locations not
within the study boundaries.
Verification procedures were performed by on-site visits, by
off-site visual verification or by telephone contact. In the five
study areas, over eighty on-site visits were conducted and nearly
300 telephone contacts and visual verifications made. Visual veri-
fications were often originally scheduled to be on-site visitations
but time restraints in the field would not allow interviews to be
conducted at each location. These visual verifications were
followed up by telephone contacts to obtain additional data.
C. Data Gathering
To obtain updated information on the bulk plants and to pro-
vide a current data base for the emission estimates, data gather-
ing was performed in conjunction with the various verification
procedures described previously. The requested data included in-
formation on 1) plant gasoline throughput 2) storage tank number,
19-10
-------
size, contents and capacity, and 3) capacity and number of delivery
trucks owned by the bulk plant operator. Forms were developed to
obtain the pertinent operating data, to obtain information on con-
trol approaches employed, and to obtain other data which was neces-
sary to complete the study. A copy of the inventory questionnaires
is shown in Appendix A.
The data gathering was conducted by on-site interviews with the
bulk plant operators and through telephone conversations. As can
be expected, many of the operators and through telephone conversa-
tions. As can be expected, many of the operators were reluctant
to release operating information on their particular businesses.
This was especially true of the independent operator. The bulk
plants which were operated through the major oil companies, either
as consignees or as independent agents, were much more willing, on
the whole, to divulge throughput information. In the case of a few
major oil company operated bulk plants, though, the requested infor-
mation would only be released after the operator received clearance
from his marketing manager. This usually required PES to submit a
written description of the project and to describe the need for the
operating data. In all cases it was considerably easier to obtain
operating data than it was to obtain financial data or the finan-
cial status of the operations, which was needed to complete the
study.
When the verifications and data gathering tasks were completed,
19-11
-------
the county inventories were edited to include the most recent and
accurate operating data. Tables I, II and III indicate summaries
of the bulk plant inventories. Table I shows the average values
for all bulk plants within the California and Colorado study areas
and Tables II and III depict data from the actual interviews per-
formed by PES in the Houston/Galveston and Baltimore/Washington,
D.C. areas. Table IV summarizes the data obtained in the PES bulk
plant inventories.
D. Problem Areas
Several problems arose in obtaining the inventory data from
the numerous sources described above. The lack of an existing in-
ventory in many areas made it- difficult to generate a basis for the
surveys. The wide variability of the regulations from state-te-
state and county-to-county made data gathering very inconsistant.
While one county may have a detailed list of bulk plants, the
neighboring county may have no data at all. In many cases, the
bulk plants were identified but the operators would not release any
operating data because of confidentiality claims. Some operators
would not release data because they were unhappy with Federal vapor
recovery regulations and did not want to release information that
they felt might be used for enforcement purposes.
Other problems arose in attempting to verify the initial en-
tries into the inventories as bulk plants defined by the study.
Since a bulk plant which had gone out of business could not be
19-12
-------
Table I. SMALL GASOLINE BULK PLANT INVENTORY, SUMMARY
OF CALIFORNIA AND COLORADO STUDIES
County
California
taador
Calaveras
Fresno
Kern
tings
Hadera
Harlposa
Merced
San Joaojjln
Stanislaus
Tulare
Tuoluane
San Diego
Colorado
Utm
Arapahoe
(oulder
Denver
Jefferson
Totals
California
Colorado
Survey
No.* of
Bulk
Plants
1
4
47
35
10
11
)
18
21
24
31
3
10
7
4
19
13
2
218
45
263
Avg.
Gasoline
Through-
put
Liters/
dtyb'
16,600
20,900
26,800
13,100
11,100
20,600
12.700
23.800
11.700
34.400
25.500
14,500
17,400
19.700
38,000
13,800
23,600
8.900
21,400
19.200
21.100
GASOLINE STORAGE TANKSC
Avo.
No. Of
Tanks
3.0
3.0
3.6
2.6
3.2
2.6
3.0
3.1
2.9
3.9
3.2
3.3
2.4
3.6
5.3
3.8
2.9
4.0
3.2
3.5
3.2
I of
Plants
with
Above-
Sround
Tanks
Only
100
67
66
77
80
73
33
71
67
73
52
100
40
8C
<7
94
50
100
66
79
68
V of
Plants
with
Under
Ground
Tanks
Only
0
33
21
23
20
27
67
29
22
9
45
0
60
14
33
6
50
0
28
21
27
Avg.
Storage
Capact ty
Liters"
xlO-3
101
113
227
144
165
163
156
196
177
204
178
2M
115
117
179
142
146
155
183
142
177
I of
with
Vapor
Recovery
On In-
eating
Loads
0
0
42
91
N.O.
0
33
78
29
100'
72
50
80
20
0
14
22
0
58
18
S3
OUTGOING LOADS
t of
Top
Load-
Ing
100
100
89
100
100
100
100
89
75
N.O.
97
100
40
100
100
93
90
50
90
92
90
Bottoa
Load-
ing
0
0
11
0
0
0
0
11
25
n.o.
3
0
60
0
0
7
10
50
11
8
10
Avg.
No. of
Trucks
N.O.
3.0
2.4
2.1
2.2
1.7
2.0
2.1
2.0
2.0*
2.2
2.0
2.2
1.7
2.0
1.4
2.1
l.S
2.2
1.7
2.1
I of
Plants
with
Sub-
Fill
100
N.0.0
70
90
100*
88
50
25
78
N.O.
79
100
90
N.D.
100
75
100'
N.O.
74
86
74
t of
Plants
with
Vapor
De-
covcry
0
0
5
0
0
0
0
11
0
N.O.
7
0
60
0
0
20'
0
0
9
7
9
CUSTOMER ACCOUNTS11
Avg. t
Sull
Agri-
cultural
60
S3
67
49
73
69
75
74
13
63
73
30
57
78
IB
60
1
15
64
44
61
Avg. I
Service
Stations
40
33
27
N.O.
13
31
13
17
44
16
15
52
43
1
57
39
18
50
23
31.
2$
MTES: I - Plants stated to be operating at tlK of survey
b - Divide by 3.785 to obtain gallons
e - Where totals do not equal 100, Indicates plants with
both above- and underground tanks
t N.D. no data
> Only one response
f - Only three responses
S- Only four responses
> Where totals do not equal 100, Indicates other types of
accounts served (coaaercial. large agriculture, etc.)
19-13
-------
Table II. SUMMARY OF DATA OBTAINED ON BULK PLANTS IN HOUSTON/GALVESTON AREA
- Clint Throughout
6*1 /Day
3.200
5.700
18.200
9.100
7,600
11.400
11.400
11.400
9,100
6.400
6,800
4,500
7,400
9.100
3.200
4,500
2.300
8,300
11.400
3.200
3.400
5.000
3.000
I/Day
12.100
21.500
68,800
34,400
28,800
43.000
43,000
43,000
34,400
24,100
25,800
17,200
28.000
34,400
12,100
17.200
8.600
31,500
43,000
12.100
12,900
18.900
11.200
Exempt Accounts
Farm
X
60
60
10
20
-
40
40
50
5
70
25
50
25
1
2
5
-
33
10
SO
80
25
80
Non-Fain
Tanks
X
40
40
90
10
-
20
40
40
85
29
40
30
65
20
2
0
-
80
48
20
75
20
Storage Tanks
Tanki
No.
3A*
3A
4A
3A
4A
3A
4A
5A
4A
3A
3A
3Uf
3U
3U
2U
2U
2U
3A
3A
3A
2A
3A
2A
Capacity
Thousand
Gal
29
38
50
51
55
48
60
50
36
SO
24
30
26
36
20
20
38
42
52
35
47
56
L
110
143
190
193
208
182
227
190
136
190
91
114
98
136
76
76
143
159
197
132
178
' 212
Vapor
Recovery
III"
Ic
I. IId, III
I. II. Ill
-
I. II, III
I. Ill
I. II, III
I. II. Ill
i. nr
i. ii. in
in
i. in
i. ii. in
i
None
-
None
I, III
I
None
I. Ill
None
Account
Trucks
Hater
3
2
2S«
3 S
2
2 S
2
3 S
2 S
2
3S
2
3
2 S
1
2
-
2
2
1
1
1
1
19-14
-------
Table II. SUMMARY OF DATA OBTAINED IN BULK PLANTS IN HOUSTON/GALVESTON AREA (continued)
Plant Throughput
Gal/Day
7,700
19,000
12,500
2,300
21,600
4,500
7,300
18,200
UD«y
29,200
71.900
47,300
8,600
81,700
17,200
27,500
68.800
Exempt Accounts
Ftra
I
5
-
5
33
15
40
40
-
Non-Farm
Tanks
X
IS
-
30
67
25
0
0
-
Storage Tanks
Tanks
No.
4A
-
5A
3U, 6A
4A
4A
3A
4A
Capacity
Thousand
Gal
63
-
70
17
53
52
36
56
I
238
-
265
64
201
197
136
212
Vapor
Recovery
I, III
I. II. Ill
I. II. Ill
None
I. Ill
I, II
I, III
-
Account
Trucks
Number
2
-
3 S
1
3
2
2
-
*A Abovegrotmd tank
bIII Vapor Recovery Installed on tt leut one account truck
CI Phase I Vapor control (Control of Inconlng gasoline transfers)
dlt Hunt II Vapor control (Control of outgoing gasoline transfers)
*5 Submersed fill
fU Underground tank
-------
Table III. BULK PLANTS IN BALTIMORE/WASHINGTON D.C. AREAS INTERVIEWED BY PES
PLANT
THROUSUPUT
Gal/Day
15.400
700
700
3.300
4.500
10.800
1.600
2.700
4.800
2.000
6.800
5.800
3.600
1,900
3,000
1,250
7.700
4,800
17.300
900
1.500
1.500
1.900
8.000
5.500
1,000
8,300
I/Day
58,000
2,600
2.600
12,000
17,000
41,000
6,000
10,000
18.000
7,500
26.000
22.000
14.000
7.300
11,000
4,700
29,000
18,000
66,000
3.300
5,500
5,800
7.300
30.000
21.000
3.900
32.000
RACK SALES
FARMS
I
0
75
0
75
7
25
95
3
75
?
15
75
75
78
65
10
5
30
?
80
25
85
70
5
50
75
50
SHALL TANKS!
I
70
85
60
100
97
71
99
20
95
7
25
95
75
99
100
95
90
30
75
80
45
99
90
50
100
99
99
GASOL
TANKS
Number"
2U
2U. 2A
6U
3D
3D
3A
2U
4A
4A
2U
3A
2A, 1U
3A
1U.3A
2U
3D
2A
2U
3D
1A.1U
2U.1A
2A.1U
2A
ZA
3U
5A
3A
NE STORAGE
CAPAC
THOUS
Gal
40
41
34
22
34
61
30
80
80
24
51
58
45
50
40
26
240
30
90
35
50
34
20
40
90
69
45
ITY,
AND
1
151
155
129
83
129
231
114
303
303
91
193
219
170
190
151
98
908
114
341
133
189
129
76
151
341
261
170 ,
VAPOR h
RECOVERY"
+
-
+
-
». i
.
»
-
+
+
.
+
+
+
.
-
-
-
.
+
ACCOUNT
TRUCKS^
Nusber
2
3S
IS
IS
IS
5S
IS
2*
4S
1
5
1
IS
2S
IS
IS
45
35
IS
IS
1
2S
3S
1
2
4S
25
'Tanks less than 2,000 gal capacity (7600 1)
h/apor recovery systems for control of Incoming loads. + yes. - - no; for control of outgoing loads. I yes
CA11 plants surveyed, except one marked with asterisk, used only top-loading account trucks. S: lubccrged filling
U - underground tanks; A abovegroimd tanks
19-16
-------
Table IV. SUMMARY OF SHALL GASOLINE BULK PLANT OPERATIONS
Ami
SU UliCUVU JU-
qvln Villir Anu
OMwr Art*
liltlion/
ItttMngton. D.C
KO-.ittO*/
G*lv«ttn Art*
M. «r
n«nu
21t
M
70
feultiw Stengt Tttiki
Throughput
llun/tty*
21,400
19,200
1(,900
11 ,(00
AM.
Mo. of
unit
1.2
l.S
1,0
3.S
I dull
tilt*
Starcgt
M
n
4!
to
1 Plinu
Kith
SUrtgi
21
tl
15
to
Storigt
Ctptcltjt
lltfjl'
1B3
142
1H
IK
Ith Viper
M Inco^ng
U
11
n
?i
Outgotng lotos
1 J?
90
»1
M
M
t totui
10
I
4
M
AVI.
Ho. of
Trvckt
2.2
l.J
1.1
1.0
< Plinu
fllllnj
71
U
a
>4*
< riuitt
with Viper
fttcevtry
1
I
0
41
Cust«Mn Actounts
S Throtishptit
(4
44
44
-IS
"IT
~n
"«t
n
^>so
l fill plpt> In ait traeU w« MI M O> lu«tx ricu. If «)/
praptrly «ul««a tnicki
-------
contacted, this information had to come from other operators in
the area. Some entries derived from local telephone directories
could not be located or contacted and the information again had to
come from a cooperative bulk plant operator in the respective area.
IV. Estimating Emissions Based on Bulk Plant Inventory
A. Sources of Emissions
Before the actual emission estimates could be performed, the
sources of the emissions had to be defined. The hydrocarbon emis-
sion sources described here were determined from the numerous on-
site inspections and interviews performed by PES personnel.
Emissions from bulk plants consist of vapor which can escape
from storage tanks, even when there is no transfer activity, be-
cause of changes in temperature of the tank wall and stored mater-
ials which vary the pressure in the vapor space. Variation forces
vapor-laden air out of the tank and aspirates fresh air into the
vapor space, allowing further vaporization of gasoline into that
space. The amounts of vapor escaping under these conditions are
referred to as "breathing losses." Losses of vapors due to liquid
transfer forces air-hydrocarbon vapors out during filling of the
tank and ingests air (promoting evaporation) during draining. Mis-
cellaneous or fugitive losses are primarily related to spillage and
leakage during gasoline handling.
19-18
-------
1. Breathing Losses
Factors affecting breathing or standing losses for fixed roof
tanks include the amount and volatility of the gasoline stored,
type and condition of tanks and appendages, and the prevailing
meteorological conditions. If there are no leaks or direct open-
ings, temperature fluctuation is the major cause of breathing
losses. As the temperature of the liquid rises, the vapor pressure
increases and evaporation takes place. The overall pressure in the
gas space increases and when the vent pressure set point is exceed-
ed, a mixture of air and hydrocarbons is discharged into the air.
As the temperature decreases, gases partially condense, contract,
and fresh air is drawn into the vapor space. This permits addi-
tional hydrocarbons to vaporize. Since hydrocarbons are emitted,
but generally not drawn back into the tanks, a continued loss of
hydrocarbon results from the daily changes in ambient temperature.
2. Working Losses
The principal cause of vapor loss during liquid transfer is
displacement of the gas (air laden with hydrocarbon vapors) in the
vapor space by the liquid entering the tank. Other causes include
the entrainment of liquid droplets in the displaced gas and post-
withdrawal pressure increase caused by evaporation.
Certain operating conditions can increase or decrease these
vapor losses. Splash loading in which gasoline is dumped onto the
surface of the liquid causes turbulence which increases evaporation
19-19
-------
rates and entrainment of droplets in the vapor being displaced.
A short interval between emptying and filling of storage tanks can
decrease losses by minimizing the time allowed for evaporation.
Also, storage tanks can be emptied in increments over a period of
several days or can be emptied in one operation prior to refilling,
with resultant differences in vapor loss.
Assuming no controls, each time a gasoline tank is filled, the
vapors above the liquid surface are emitted to the atmosphere. The
quantity of hydrocarbon vapors emitted is a function of the volume
displaced, type of loading, temperature and the degree of saturation
of the vapor space with gasoline vapors. At any given temperature,
the amount of vapor in the vapor space cannot exceed a limit imposed
by the saturation pressure corresponding to that temperature. This
limit, however, increases as the temperature increases.
In a quiescent state, the approach of saturation and pressure
increase of a vapor space with gasoline vapors is a slow process.
Since hydrocarbon vapors are heavier than air and diffusion is slow,
a saturated blanket of vapor initially forms over the liquid sur-
face, decreasing the driving force for further vaporization. Also,
with evaporation of the lighter hydrocarbon molecules, the tendency
of the components in the stagnant surface to vaporize decreases.
Thus, the degree of saturation in the overall gas space of a tank
can be decreased by minimizing liquid surface and vapor space mix-
ing during the filling operation.
19-20
-------
Splash loading not only creates droplets which can be en-
trained in the outgoing effluent and exposes fresh liquid surfaces,
but results in mixing of the vapor space as well. This mixing of
the vapors disturbs the saturated blanket near the liquid surface,
increasing the driving force for further vaporization. Hydrocarbon
emissions under splash filling conditions can significantly exceed
that calculated by assuming saturation.
Another factor which can affect the quantity of hydrocarbons
emitted is the interval between drainage and filling. When a tank
is drained and immediately refilled, the air drawn into the tank
during draining may be expelled with relatively little hydrocarbon
content. In a tank allowed to sit after draining, the air drawn
into the vapor space becomes saturated with hydrocarbons, thus,
increasing pressure (and emissions) and resulting in the maximum
loss of vapor during refilling.
Another operational procedure which may increase losses is the
\
small sequential withdrawals of gasoline from a storage tank over
a period of several days rather than one continuous large with-
drawal. After a small withdrawal, the post-transfer emissions
caused by evaporation tend to be high in hydrocarbons since little
air is ingested during the withdrawal. After a large withdrawal,
the initial post transfer emissions are low in hydrocarbons since
large amounts of air are ingested during the withdrawal.
19-21
-------
3. Miscellaneous Losses
Miscellaneous losses have been found to be highly variable.
These losses Include spillage, liquid and vapor leakage and opera-
tional losses which occur when opening tank hatches for visual
inspection or measuring liquid levels with dip sticks. Leakage can
occur and has been observed at dry breaks, pressure vacuum valves,
hatches, manholes, pump seals, shut-off valves and piping joints.
It has been visually observed that some spillage (on the order of
half liter) occurs when connecting and disconnecting transfer lines.
Visible liquid leakage at dry breaks (few milliliters of gasoline)
was observed from the connections after transfer. Opening of
hatches of an empty truck to verify that they have received all the
gasoline expected.
B. Control Concepts
When estimating emissions from an operation it is imperative
that knowledge of the control concepts and how they effect the
emissions are available. The controls that were used as determined
by the survey are discussed in this section.
1. Breathing Losses
Storage tanks are subject to evaporation or standing losses
due to volatility of the material stored, type and condition of the
tank and its appendages and prevailing meteorological conditions.
The simplest methods for reducing these venting losses are to (1)
inspect and repair leaks in the tank and fittings, (2) paint the
19-22
-------
tank with white paint where possible, (3) assure that vent valves
do not leak and (4) set the pressure and vacuum relief settings to
minimize breathing. The broader the band for the vent valve set-
tings, the lower will be the breathing losses.
Another method for preventing vapor loss is to install vapor
recovery equipment at the vent valve. For vapor recovery, the
vented vapor must be able to be condensed and recycled to the tank
or be collected and regenerated. For prevention of pollution only,
the vapors can be combusted or collected and disposed of in some
approved manner.
2. Working Losses
Excluding spillage, the two major sources of loss of gasoline
vapor during transfer are .1) venting to the atmosphere the volume
of gasesair and hydrocarbonsdisplaced by the entering liquid
and 2) filling in a manner which creates turbulence which results
in increased vaporization rates and liquid droplet entrainment in
the vapor space.
The most common current methods of reducing working losses are
to use submerged filling for the loading of gasoline and to install
a vapor balance system between the vapor spaces of the tanks con-
nected during the gasoline transfer.
a. Submerged Fill
Submerged fill is the introduction of liquid gasoline into the
tank being filled with the transfer line outlet being below the
19-23
-------
liquid surface. This is compared to splash loading, where the
transfer line outlet is at the top of the tank. Submerged filling
minimizes droplet entrainment, added vaporization and turbulence.
If a fill port is located at the tank top, submerged fill is accom-
plished by either extending the nozzle (commonly referred to as
stingers) or permanently attaching to the fill port a pipe extending
to within 6 inches (15 cm) of the tank bottom. This permanent
installation is commonly referred to as a drop tube.
Aboveground storage tanks normally include submerged fill.
Submerged fill for underground storage tanks can be accomplished by
attaching a pipe to the fill port. These installations were common
on underground tanks in surveyed areas.
Bottom loaded trucks by definition include submerged filling.
Top loaded trucks utilize an extension such as a pipe or flexible
hose on the loading arm, or a pipe can be permanently attached to
the trucks.
Submerged filling of customer tanks can be accomplished with
either nozzle extension or a permanently attached drop tube. Some
difficulties have occurred with the installation and use of per-
manent drop tubes. One problem is "spit back." "Spit back" is the
return flow and spillage of gasoline at the fill port during trans-
fer. This appears to be primarily related to the smaller fill port
and drop tube sizes in customer tanks, compared to service station
tanks, and the lack of a coupling at the fill port interface. Use
19-24
-------
of a nozzle extension with or without a permanently Installed drop
tube or a coupling should solve the "spit back" problem.
b. Balance System
Probably the most common vapor recovery system currently in use
is the vapor balance system. Efficiency is good for the control of
working losses, but not significant in controlling breathing losses.
A pipeline between the vapor spaces of the truck and storage tanks
essentially creates a closed system permitting the vapor spaces of
the tank being filled and the tank being emptied to balance with
each other. The net effect of the system is to transfer vapor dis-
placed by liquid into the tank in which draining of the liquid
creates additional vapor space. This prevents the compression and
expansion of vapor spaces which would otherwise occur in a filling
operation. If a system is leak tight, very little or no air is
drawn into the system and venting dur to compression also is re-
duced substantially. The system is applicable to underground and
aboveground storage facilities equipped with either bottom or top
loading, and are applicable to both incoming and outgoing transfers.
c. Secondary Control Systems
Secondary control systems, such as refrigeration, oxidation,
adsorption, etc., were found in only one of the study areas, San
Diego. Many systems had been initially installed but only one such
system was operating. This vapor recovery system employed a re-
frigeration unit to reduce pressure in the storage tanks and there-
19-25
-------
by to minimize venting. In this system, vapors were drawn from the
storage tanks by a blower, passed over cooling coils in the re-
frigeration unit and exhausted back to the storage tanks through
an insulated return line. The system made no effort to condense
vapors but was designed strictly to maintain a constant temperature
in the storage tanks (in this case 60°F) and thereby maintain a
pressure below the venting level.
3. Efficiencies
For determination of the emission quantities, efficiencies
corresponding to the described control concepts had to be generated.
These were based upon test results, emission factors and litera-
ture. Submerged fill resulted in an emission reduction of 58.6%
from transfer losses based upon emission factors used. The balance
system efficiency was estimated at 90% based upon actual test re-
sults. The efficiency used for the secondary system was also 90%
based upon test data. This perhaps is low but only one such system
was in operation out of the nearly four hundred bulk plants surveyed
,and it was felt that this error was very minor.
C. Development of Emission Factors
The actual hydrocarbon emission factors were determined by
using emission formulas, such as API formulas for storage tank
losses, and data on typical bulk plants, as described by the inven-
tories. For example, the API equation for the breathing losses
from fixed roof tanks is:
19-26
-------
L = 2.21 X 10-*M (U.7P. p )°'68 a1'" H°-51 I°-5°Fp CK
Where:
L = Fixed roof breathing loss (Ib/day)
M = Molecular weight of vapors lost (Ib/lb mole)
P = True vapor pressure of liquid at bulk liquid temperature
(psia)
D = Tank diameter (ft)
H = Average vapor space height (ft)
T = Average daily ambient temperature change ( F)
F = Paint factor (dimensionless)
P
C = Adjustment factor for small diameter tanks (dimensionless)
K = Adjustment factor dependent on product stored
(Dimensionless = 1 for gasoline)
Typical values for bulk plants in the study areas were as follows:
M = 66
P = 42,800 newtons/sq meter (6.2 psia, for RVP-10 gasoline
at 70°F)
D = 3.2 meter (10.5 ft survey average)
H = 4.0 meter (13 feet, 1/2 survey average tank height)
T = 8.3°C
F = 1.24 for white roof and specular aluminum shell in poor
condition, condition based on survey observations)
C = 0.55
By substituting these values into the equation, an emission factor
of 6.7 Ib/day/tank was obtained. In many of the counties surveyed,
actual throughput data was not obtained for each bulk plant loca-
tion. Since the purpose of the emission inventory was to develop
total emissions from bulk plants in the study regions, an average
throughput for the particular county or region was determined from
the survey data and applied to all bulk-plant operations within the
study areas.
19-27
-------
The emission factors derived for the estimating procedures
were:
Fixed Roof Storage Tanks: Breathing Loss 6.7 Ib/tank/day
Working Loss 9.7 lb/1000 gal.
Truck Loading Losses: Splash Fill 14.0 lb/1000 gal.
Submerged Fill 5.8 lb/1000 gal.
Miscellaneous Losses: 3.2 lb/1000 gal.
D. Emission Estimates
. Emission estimates based upon the bulk plant inventories were
generated for all regions. As an example, emission estimates for
the bulk plants in the Houston/Galveston area were calculated for
uncontrolled sources, for emissions from bulk plants with current
vapor recovery controls, and emissions from bulk plants with poten-
tial control under the existing Texas SIP regulations. For esti-
mating uncontrolled emissions, it was assumed that no vapor re-
covery equipment was installed and that all tank fillings and trans-
fers to account trucks were by splash filling. This would amount
to the worst case emissions. For the purposes of these calcula-
tions, the average annual throughput of gasoline through each bulk
plant was 2,203,000 gal (8,338,000.1), based upon the PES bulk
plant sample. Also for the purposes of these calculations, seventy
bulk plants were assumed to be operating in the area.
The uncontrolled emissions from the bulk plants in the
Houston/Galveston area were:
19-28
-------
Breathing losses from aboveground tanks -
236 tons/year (214 metric tons/year)
Working losses from all bulk plant storage tanks -
737 tons/year (tt8 metric tons/year)
Transfer losses from filling account trucks -
1063 tons/year (964 metric tons/year)
Transfer losses from filling account storage tanks -
1063 tons/year (964 metric tons/year)
Miscellaneous losses -
243 tons/year (220 metric tons/year)
Total uncontrolled hydrocarbon emissions -
3342 tons/year (3031 metric tons/year)
The emission factors used for these emission estimates were the
same as those listed in the previous section.
To determine current emissions from bulk plants, data on
*
Phase I and Phase II controls installed were obtained from the
PES bulk plant survey. From this data, it was found that 75% of
all the bulk plants had installed Phase I vapor recovery. Phase
II controls were installed at 36% of all the bulk plants. Vapor
recovery had been installed on at least one truck at 68% of the
facilities. For calculation purposes, it was assumed, since most
bulk plants had one truck with vapor recovery and one without, that
the deliveries which went to exempt accounts, i.e., farm accounts
and tanks less than 2,000 gal (7,600 1.), were made with the
delivery truck which had no vapor recovery. From the PES survey,
the average bulk plant delivered 50% of its gasoline 67% of its
*Definition of Phase I and Phase II controls found in Table II
19-29
-------
accounts to exempt accounts. An efficiency of 90% for the
Phase I and Phase II controls was used.
The hydrocarbon emission estimates for the bulk plants in the
Houston/Galveston area incorporating current vapor recovery control
installations were:
Breathing losses from aboveground tanks -
236 tons/year (214 metric tons/year)
Working losses from all bulk plant storage tanks -
239 tons/year (217 metric tons/year)
Transfer losses from filling account trucks -
791 tons/year (718 metric tons/year)
Transfer losses from filling account's storage tanks -
638 tons/year (579 metric tons/year)
Miscellaneous losses -
243 tons/year (1948 metric tons/year)
Total hydrocarbon emissions incorporating current control
practices -
2148 tons/year (1948 metric tons/year)
Under current control practices, a 67.5% reduction in working
losses, a 25.6% reduction in transfer losses involving account
trucks and a 40% reduction in transfer losses involving account
tanks served have been accomplished. This is an overall hydro-
carbon emission reduction of 35.7% representing 1,194 tons/year
(1,083 metric tons/year) from all bulk plant emission sources.
To determine the amount of potential reduction of the hydro-
carbon emissions, the vapor recovery and emission controls as out-
lined in the Texas SIP regulations were extrapolated for the gaso-
19-30
-------
line throughput used above. This would mean that Phase I controls
would be installed on all bulk plants, Phase II controls on the
racks would be used on all non-exempt deliveries and submerged fill
pipes would be installed on all non-exempt tanks. The SIP regula-
tions, as written, do not require any controls to limit storage
tank breathing losses and miscellaneous losses (i.e., spillage) so
these emissions levels do not vary.
The hydrocarbon emission estimates for the bulk plants in the
Houston/Galveston area incorporating potential vapor recovery con-
trols were:
Breathing losses from aboveground tanks -
236 tons/year (214 metric tons/year)
Working losses from all bulk plant storage tanks -
74 tons/year (67 metric tons/year)
Transfer losses from filling account trucks -
585 tons/year (530 metric tons/year)
Transfer losses from filling account's storage tanks -
585 tons/year (530 metric tons/year)
Miscellaneous losses -
243 tons/year (220 metric tons/year)
Total hydrocarbon emissions incorporating potential emission
controls under Texas SIP regulations -
1723 tons/year (1562 metric tons/year)
Under potential control strategies presented in the Texas SIP
regulations, a 90% reduction in hydrocarbon working losses from
bulk plant storage tanks and a 45% reduction of transfer losses to
both account trucks and account tankage would be experienced.
19-31
-------
This would yield an overall bulk plant emissions reduction of
48.5% or 1620 tons/year (1469 metric tons/year). If it is assumed
that all deliveries made from the bulk plant would require Phase II
controls, not just non-exempt deliveries, a further reduction in
bulk plant emissions would occur. The overall bulk plant hydro-
carbon emissions would then be reduced to 1244 tons/year (1128
metric tons/year). This would then indicate an overall emissions
reduction of 63% or 2098 tons/year (1903 metric tons/year).
Not only were emission estimates supplied to show the effect
of control strategies, but also to illustrate how proposed or
possible regulatory strategies would effect the emission quantities.
An example of this is shown in Table V, illustrating how potential
regulatory actions would effect emissions from bulk plants in the
Houston/Galveston area.
19-32
-------
Table V. SUMMARY OF DECREASE IN EMISSIONS FROM BULK PLANTS IN HOUSTON/GALVESTON AREA AS FACILITIES
COMPLY WITH HYDROCARBON VAPOR RECOVERY REGULATIONS
£
U)
Facility Description
Vapor recovery on all bulk
plant operations and account
tanks
Total compliance with currant.
approved 5IP Regulations.
Total Gasoline Throughput
220,000 Gal/Mo.
Total Gasoline Throughput
24,000 Gal/Day
Throughput to Non-Exeupt
Accounts 22.000 Gal/Day
Throughput to Non-Exempt
Accounts 22,000 Gal/Day or
Total Gasoline Throughput
24,000 Gal/Day
Total Gasoline Throughput
28.000 Gal/Day
Total Gasoline Throughput
28,000 Gal/Day or Throughput
to Non-Exempt Accounts 22,000
Gal/Oay
Total Gasoline Throughput
212.000 Gal/Day
Total Gasoline Throughput
212,000 Gal/Day or Throughput
to Non-Exempt Accounts
22,000 Gal/Oay
Total Gasoline Throughput
220,000 Gal/Dayd
No Plants In Compliance
Nunber
of
Plants
Affected
70
70
70
54
36
34
29
20
11
9
2
0
EMISSION ESTIMATE
TOTAL
T/Yrb
765
1723
1723
2095
2499
2566
2678
2872
3084
3132
3294
3342
MT/Yrc
693
1563
1563
1900
2266
2326
2428
2604
2796
2840
2986
3030
Breathing
Losses
From Above
Ground Tanks
T/Yr
236
236
236
236
236
236
236
236
236
236
236
236
KT/Yr
214
214
214
214
214
214
214
214
214
214
214
214
Working
Losses from
All Storage
Tanks
T/Yr
74
74
74
226
392
419
465
545
631
651
717
737
HT/Yr
67
67
67
206
356
380
422
494
672
590
650
668
Transfer Losses To
Account Trucks
T/Yr
106
585*
585
695
814
834
867
924
987
1001
1049
1063
MT/yr
96
531
531
630
738
756
786
838
895
908
951
964
Account Tanks
T/Yr
106
585
585
695
814
834
867
924
987
1001
1049
1063
MT/yr
96
531
531
630
738
756
786
838
895
908
951
964
Losses
T/Yr
243
243
243
243
243
243
243
243
243
243
243
243
MT/Yr
220
220
220
220
220
220
220
220
220
220
220
220
'it Is assumd that deliveries to exeapt accounts are done with trucks without vapor recovery
""T/Yr Tons/Years
cMT/Yr Metric Tons/Year
''Current Texas State Regulation V Incorporates this exemption
-------
V. Conclusions
Based on the inventories completed in the five study areas,
it can be concluded that data on small bulk plants in other areas
of the county are very limited. Specific bulk plant inventories
would be necessary in these areas. Even if an inventory exists,
it may be quickly out-dated because of the rapid changes in
current gasoline marketing trends.
The direct inquiry approach to obtaining inventory data was
very successful especially for obtaining information on control
concepts currently employed in the field. Enough information to
successfully complete the study objectives was gathered without
having to resort to Form 114 letters.
Finally, the current hydrocarbon emission levels from small
bulk plants can vary considerably from region to region. (For
example, hydrocarbon emissions from the seventy bulk plants in the
Houston/Galveston area are currently 2148 tons/year while the
hydrocarbon emissions from the .fifty-six bulk plants in the Balti-
more and National Capital AQCR's totalled only 725 tons/year).
The variation is due mostly to the number of operating bulk plants,
the accompanying throughput, and the controls currently being used.
19-34
-------
Bibliography
R.J. Bryan, M.M. Yamada and R.L. Norton. "Effects of Stage I Vapor
Recovery Regulations on Small Bulk Plants and on Air Quality
in the Washington, D.C., Baltimore, Maryland and Houston/
Galveston, Texas Areas." Pacific Environmental Services, Inc.
under EPA Contract No. 68-01-3156, Task No. 28, March 1977-
R.J. Bryan, R.L. Norton and P.S. Bakshi. "Compliance Analysis of
Small Bulk Plants." Pacific Environmental Services, Inc.
under EPA Contract No. 68-01-3156, Task No. 17, October 1976.
R.J. Bryan, W.O. Jacobson and R.R. Sakaida. "Study of Gasoline
Vapor Emission Controls at Small Bulk Plants." Pacific
Environmental Services, Inc. under EPA Contract No. 68-02-
3156, Task No. 15, October 1976.
"Revision of Evaporative Hydrocarbon Emission Factors." Radian
Draft Report No. 100-086-01, June 15, 1976.
Applicable SIP Regulations on Gasoline Transfer Vapor Control:
40 CFR52 - Subpart F - California 52.255
40 CFR52 - Subpart G - Colorado 52.336
40 CFR52 - Subpart V - Maryland 52.1101
40 CFR52 - Subpart SS - Texas 52.2285
52-2286
40 CFR52 - Subpart W - Virginia
19-35
-------
APPENDIX A
SMALL BULK PLANT EVALUATION
19-36
-------
SMALL BULK PLANT EVALUATION
1. Owner Operator
Address
Contact Phone ( )
2a. Gasoline throughput gallons/month
2b. Grades of gasoline
3a. Other products sold
3b. Percentage of business selling gasoline
4. Underground tanks - number
Capacities
5. Aboveground tanks - number
Capacities
6. Top loading
Bottom loading
Submerged fill
7. Supply vehicles owned by
Number
Capacity
Frequency of delivery
8. Delivery vehicles owned by
Number
Capacity
Time to fill
19-37
-------
9. Fugitive emissions
10. Vapor Recovery, Deliver Vehicle
Date installed by whom
Cost
Bottom or Top Load
Operating Problems
11. Vapor Recovery Storage Tank/incoming loads
Date installed by whom
Cost
Operating Problems
12. Vapor Recovery, Storage Tank/Loading Rack (Delivery)
Installed by whom
Installation Cost , Operating Cost
Maintenance Cost
Operating Problems
13. Vapor Recovery, Delivery Vehicle
Installed by whom
Installation Cost Operating Cost
Maintenance Cost
Operating Problems
14. Percent Deliveries to Exempt Customers
Type of Exemption: Agricultural
Small Tanks Size
Other
19-38
-------
Of Non-exempt customers, what vapor control techniques
are being used
15. Assume vapor recovery will initially cost $20,000 and opera-
ting costs increase 20%.
a) Would you stay in business
b) Could you obtain loan Down payment required
c) Comments
Repeat for $10,000 $30,000
16. What could you sell your trucks, facility, accounts receivable
and good will for $
What did you originally invest $
17. Present margin
Annual sales $ Gallons
Profit $
Debts $
18. Debt-equity or debt-total assets ratio
Rate of return on total assets or net worth
Break even point
Assessed valuation
19. Comments - Closures, Competitors, bottom loading, etc.
19-39
-------
P-193
AN ORGANIC SPECIE EMISSION INVENTORY FOR
STATIONARY SOURCES IN THE
LOS ANGELES AREA - METHODOLOGY
By
Harold J. Taback, KVB, Inc.
Tim W. Sonnichsen, KVB, Inc.
Nicholas Brunetz, KVB, Inc.
Joan L. Stredler, Abacus Programming Corp.
For
EMISSION INVENTORY/FACTOR WORKSHOP
September 13-15, 1977
Raleigh, NC
20-1
-------
TABLE OF CONTENTS
Section Page No.
1.0 INTRODUCTION 1
2.0 INVENTORY DATA PROCESSING 4
2.1 Data Sources 4
2.2 Data Management 9
3.0 EMISSION PROFILES 13
3.1 Description 13
3.2 Methodology 14
4.0 EMISSION FACTORS 18
4.1 Point Sources 18
4.2 Area Sources 23
5.0 FIELD TESTING 31
5.1 Field Measurements and Sampling 32
5.2 Laboratory Analysis 38
5.3 Quality Control 42
REFERENCES 43
20-2 P-193
-------
SECTION 1.0
INTRODUCTION
To provide detailed data on which to model the photochemical formation
of atmospheric oxidants and haze and to provide information on which to base
comprehensive control strategy, an inventory of gaseous organic emissions
from stationary sources was conducted for a district in California known as
the South Coast Air Basin. This region includes portions of Santa Barbara/
Ventura, Los Angeles, Orange, Riverside, and San Bernardino Counties. Unlike
most organic emissions inventories in the past, this study included the de-
velopment of emission profiles, i.e., a breakdown of the individual organic
species which contributed at least 1% of the total organic emissions from
each source. From one to 30 different species were identifed in the emission
profiles which were developed for 200 sources by a comprehensive field samp-
ling and laboratory GC-MS analysis program.
The inventory accounted for all known stationary source organic
emissions including major and minor point sources and area sources (oil pro-
duction fields, architectural coatings, domestic solvent usage, etc.).* The
inventory was prepared in the EPA's Emission Inventory Subsystem (EIS) format.
All sources were located by Universal Transverse Mercator (UTM) coordinates.
A three-phase approach wastaken in conducting the program. First,
a preliminary inventory of total organics (without specific species) was pre-
pared to identify the major sources and to determine the distribution of emis-
sions among the various source types. Next, a field test program was con-
ducted to characterize emissions from sources selected on the basis of the
preliminary findings, emphasizing those source types comprising the greater
amount of the emissions. Test results were augmented by questionnaire
*A major point source was defined as any emission source belonging to a
point with total organic emissions of more than 25 tons per year.
20-3 P-193
-------
responses and literature data in establishing both emission factors and
emission profiles (% composition by weight). In the third phase, a final
inventory will be created from the preliminary inventory by the addition of
emission profiles and new sources (especially area sources) and the updating
of emission factors.
'ro appreciate the approach taken in this inventory, it is important
to understand the potential magnitude of the task and the budget which was
allocated for the program. There were over 50,000 active permits cover-
ing the sources in the various county districts. The final inventory
will account for over 8,000 major point sources of organic emissions plus
minor point sources (which include gasoline stations) and area sources. The
total effort budgeted for-this program, including field testing, laboratory
analyses, other data collecting, data analyses and processing, as well as
the overall program management, was only eight people for one year plus
expenses (computer, travel, equipment, etc.). Thus, maximum use was made of
existing information. The new research performed on the program was to:
1. Provide information where none existed
2. Check the validity of existing data, and
3. Update existing data to reflect current trends.
The characterization of refinery emissions was an example of how
judiciously the test program was designed. In the late 1950's the Los
Angeles APCD and EPA spent an 8-10 person effort for more than two
years to characterize total hydrocarbon emissions from the refineries in the
Basin. The EPA is currently sponsoring a program to measure refinery emis-
sions alone with a budget that is'three times greater than the one available
on this program in which all of the emissions in one of the largest industrial
areas of the country has to be characterized. This program afforded a total
refinery testing effort of five weeks with a crew of four. Thus testing was
primarily directed at obtaining emission profile data and checking the emis-
sion factors developed on the 1950's project.
20-4
-------
The inventory was conducted by KVB, Inc. under sponsorship of the
California Air Resources Board (ARE)* as part of a total program which also
included a study of control systems and a 10 year projection of emissions.
Preliminary inventory data, EIS format data and technical advice on emissions
and data processing were received from the South Coast Air Quality Management
District (SCAQMD) and the APCD's of Ventura and Santa Barbara Counties.
GC/MS analyses of emission samples and advice on the development of the field
sampling train were provided by Analytical Research Laboratories, Inc. (ARLI)
of Monrovia, CA, Mr. M. L. Moberg, President and Chief Chemist. Data
quality and evaluation was performed by EcoScience Systems, Inc. of Riverside,
CA, Dr. James N. Pitts, Jr., President and Principal Consultant. Data manage-
ment and computer programming were provided by Abacus Programming Corporation
of Santa Monica, CA, Mr. Calvin Jackson, Vice President and Director of Pro-
duction. The Western Oil and Gas Association provided guidance and assistance
in the petroleum industry aspects of the program. A special measurement of
refinery emissions using ambient testing and diffusion modeling techniques
was performed by Aerovironment, Inc. of Pasadena, CA, Mr. Ivar Tombach, Vice
President and Program Director.
At this writing the final emission inventory had not been compiled
pending the completion of EIS data entry and checkout by the SCAQMD and the
APCD's. In this paper the methods used to collect, formulate and process
the inventory data are presented. It is estimated that the inventory itself
will be published by the end of 1977 along with a comprehensive engineering
report.
*Dr. John R. Holmes, Research Division, Chief, and Mr. Jack Paskind, Research
Project Officer.
20-5
P-193
-------
SECTION 2.0
INVENTORY DATA PROCESSING
2.1 DATA SOURCES
The data used in this organic emission inventory were obtained from
the following sources:
1. Various government agency files
2. Field testing
3. Questionnaires
4. Literature
5. Engineering analyses
6. Personal contacts with government and industry personnel.
All county enforcement agencies were in the process of a total recom-
pilation of their permit files using the EPA's EIS/P&R* (Ref. 1) format
(referred to as EIS hereafter) during the period of this inventory. The
computerized permit file from Los .Angeles County and the permit files from
the other counties were used to obtain data for the preliminary inventory.
It was planned that the final inventory would be compiled using the new EIS
data base for the major point sources as soon as data entry was completed
and checked for all counties in the inventory. The following key data were
contained in the EIS data base:
1. Plant name, address, ID No., etc.
2. Standard Industrial Code (SIC)
3. Source Classification Codes (SCC)
4. UTM Coordinates
5. Stack Height
6. Pollutant Identification
7. Emission Factor
8. Throughput Rates
*Emission Inventory Subsystems/Permit and Registration
20-6 p-193
-------
9. Estimated emissions
10. Seasonal variations
11. Operating period (hr/day, day/week, week/yr)
For minor sources in LA County, the original permit file was used.
Also a computer tape file of gasoline station locations in LA County was
received from the SC AQMD. Both of these files had location coordinates on
a one mile square grid basis. The ARE provided an algorithm for converting
the one-mile grid to UTM coordinates (Ref. 2).
The ARE also provided a tape file of population by UTM coordinates
which was used to distribute population related area source emissions.
Field test data were used to formulate emission profiles and to
develop emission factors for new sources or check those factors on sources
already characterized by the districts or the EPA in AP-42 (Ref. 3).
Questionnaires were received from approximately 100 industrial
sources with comprehensive data on their solvent and fuel usage. Data
received were used to develop emission profiles and to check values con-
tained in the district files.
There was a great deal of activity in the area of organic emission
assessment by other agencies and contractors. A list of those programs which
provided valuable data for this inventory are summarized in Table 1. Excellent
cooperation and data exchange were maintained with those contractors and
agencies listed.
Other sources of information included personal contacts with various
industry associations (dry cleaning, refinery, asphalt, printing, etc.) and
government agencies (especially the ARB, California Division of Oil and Gas,
EPA Office of Air Quality Planning and Standards in Durham, EPA Region 9,
local air pollution districts and the Southern California Association of
Governments, SCAG).
From data received from the above sources, comprehensive analyses were
conducted to derive emission profiles in a form compatible with the inventory
format. Analyses of test data from this and related programs listed in Table
1 were performed to create or evaluate existing emission factors.
P-193
20-7
-------
TABLE 1.
RELATED STUDIES ON ORGANIC COMPOUND EMISSIONS
Project Title
Scope
Sponsor*
Contractor
Status (as of Mid 1977)
K>
?
00
Air Quality Impacts of
Outer Continental Shelf
Oil Development in the
Santa Barbara Channel
Fugitive Emissions from
Oil Field Production
Operations
Assessment of the
Environmental Effluents
from Oil Refining
Emissions from Ships
and Shipping Operation
including Transfer of
Oil
Assess the impacts of OPR
OCS development on the
environment
Determine emission API
factors on a compon-
ent basis for onshore
and offshore facilities
Determine validity of EPA
refinery emission
factors currently used
in AP-42
Determine emissions ARB
from shipping opera-
tions in the SCAB
OPR Staff
S ERT, Inc.
Rockwell
Air Monitor-
ing Center
Radianr Inc.
Scott
Research
Laboratory
Final Report Draft
issued March 1977
Work plan being
developed
Tests are currently
being conducted in
second refinery
Work plan being
developed
Hydrocarbon Emissions
from Floating Roof
Petroleum Tanks
Hydrocarbon Emissions
from Fixed Roof Tanks
Hydrocarbon Emissions
from Tanker Loading
Operations
Determine validity of WOGA
AP 2517
Assess the validity WOGA
of AP 2518
Determine HC emissions WOGA
resulting from crude oil
loading off Ventura and
Santa Barbara counties
Engineering
Science,
Inc.
Engineering
Science, Inc.
Chevron
Research,
Inc.
Final Report released
Jan. 1977
Final Report released
July, 1977
Final Report originally
scheduled for May 1977
P-193
-------
TABLE 1.
RELATED STUDIES ON ORGANIC COMPOUND EMISSIONS (Continued)
Project Title
Scope
Sponsor *
Contractor
Status (as of Mid 1977)
Floating Roof Tank
Metallic Sealing Ring
Emission Test Program
Organic Compound Emis-
sions From Natural
Sources
Emission Factors from
Burning Agricultural
Wastes Collected in
California
Determine effects of WOGA
wind, ring quality, gap
size and secondary seals
on HC emissions
Determine natural ends- EPA
sion rates from forest
vegetation
Determine emission ARE
factors from burning
31 field and orchard
crops
Chicago
Bridge and
Iron
Washington
State
University
UCR State-
wide Air
Pollution
Research
Center
Final Report issued
March 1977
Final draft completed
Final Report, January
1977
VO
ORGSOL Regulation
Study Group.
Architectural Coatings
Survey
A Methodology for
Reactive Organic Gas
Emissions: Assessment
of Pesticide Usage in
California
Gasoline Marketing
Vapor Recovery System
Development
Determine potential ARE
reduction of organic
emission using H20
borne paints
Determine HC emissions ARB
from pesticide applica-
tions
Determine test procedures SDAPCD
to assess the effective-
ness of vapor recovery
techniques applied to
petroleum marketing
ARB Staff
ARB Staff
Staff report issued
June 1977
Report due August 1977
SDAPCD
Test continuing
Continued
P-193
-------
TABLE 1. RELATED STUDIES ON ORGANIC COMPOUND EMISSIONS (Continued)
Project Title
Scope
Sponsor*
Contractor
Status (as of Mid 1977)
Measurement of Atmos-
pheric Organic Emission
from Natural Sources
Control of Volatile
Organic Emissions for
Existing Stationary
Source
Determine HC emission EPA
factors from asphalt
operations and landfills
Preparation of documents EPA
for control of organic
emissions
Midwest
Research
Institute
EPA Air
Program
Staff
Preliminary tests
underway
Volume 1 issued November
1976. Future volumes to
be released late 1977.
*Abbreviation code: OPR
API
EPA
ARE
WOGA
SDAPCD
California Governor's Office of Planning and Research
American Petroleum Institute
U.S. Environmental Protection Agency
California Air Resources Board
Western Oil & Gas Association
San Diego Air Pollution Control District
P-193
-------
2.2 DATA MANAGEMENT
The data to be processed as part of the final organic emission
included:
1. EIS data for major point sources for SCAQMD
2. Minor point source data from the SCAQMD (Metro Zone)
3. Gasoline station data for SCAQMD (Metro Zone)
4. EIS data for Ventura and Santa Barbara County (major and
minor sources including gasoline stations)
5. Petroleum production field
6. Additional area data for sources such as forests, landfills,
architectural coating, domestic solvent use, etc.
7. Emission profiles
8. Population distribution by one kilometer grid
9. Emission factor adjustments to EIS data
The available EIS data processing software was incorporated for
processing the EIS data. In this system individual sources could be modified,
added or deleted. KVB added a feature which also permitted the data to be
modified by SCC number. For example, the emissions in the EIS data base
from certain fixed roof tanks (identified by a specific SCC number) were be-
lieved to be too high based on recent test data. The emissions from those
tanks were modified by one correction factor applied to all the emissions of
that specific SCC number.
The profile data was organized with SCC number as the key. The
specific organic specie emissions for any source were determined by factoring
the total source emissions by the profile of specie weight percentages.
In describing organic emissions, the point source data were a minor
portion of the emissions, whereas natural emissions, architectural surface
coatings and gasoline marketing constitute the major portion of the emissions.
Since a standard format was not yet available for describing emissions not
meeting the EIS point source criteria, KVB choose to develop an area source
data base for this purpose. The format was designed to allow description
of emissions by their one kilometer grid location and process (or activity).
P-193
20-11
-------
Gasoline marketing data for this data base were acquired in three different
ways. In Los Angeles County, information concerning 10,000 stations in the
Basin was computerized; however, individual station through-puts were missing.
KVB chose to take the total sales for Los Angeles County and apportion the
through-put by the total storage tank volume at each station. For Ventura
County the gas station data were included in the EIS file. For consistency,
KVB chose to take these EIS data and format them as the Los Angeles County
data with the gas station locations and through-puts retained. KVB assigned
new emission factors to the data. For the remaining counties the gas station
emissions were apportioned by residential area based on the total county
sales. These gas station data were formatted in the KVB area source format.
Since the EIS. point source data base did not contain the organic
emissions for minor point sources under permit in L.A. County (Metro Zone),
these data were acquired from Metro Zone's permit files and formatted in the
KVB area source format. Data on the emissions from minor point sources not
under permit were also obtained and included in the KVB area source data
base.
Finally the KVB area source data base contained data from all area
sources identified by the KVB engineers. These emission sources included
waste disposal, petroleum operation other than refining, domestic and agri-
cultural sources, geogenic sources and natural emissions.
All sources in the inventory were given an SCC number which was
occasionally qualified by the SIC number. (SCC numbers for area sources were
created after consultation with ARB personnel.) A file was created with all
information relative to these SCC numbers, the emission correction factors to
be applied to all sources with the given SCC/SIC number, the profile key to
identify the profile for this source type, the relevant ARB application cate-
gory, and summer or winter differentials to be used to alter emissions
seasonally if warranted by the source type. This SCC file was used as the
major system link between sources and their profiles.
20-12
-------
Within each ARE application category only unique profiles were
identified. The profile records contained information concerning the method
of determining the profile and estimated error. In addition, each profile
contained the SAROAD code and percent by weight of each specie in the profile.
Where the SAROAD coding was not comprehensive enough, KVB and their subcon-
tractor, ARLI, added SAROAD codes in a logical manner. Since SAROAD codes
were the only specie identifier in the profile data base, a separate tabular
file was created to contain SAROAD codes, species name, molecular weights
and ARE reactivity class.
From the data files used in this inventory nine reports were to be
produced. They included:
1. A Total Organic Emission Report containing:
a. Source information (county, APCD Point ID No., SSC No.,
SIC No.)
b. Total organic emissions, ton/year
c. Summer emissions, ton/day, broken down into weekday
emissions and weekend emissions
d. Winter emissions, ton/day, broken down into weekday
emissions and weekend emissions
e. Emission profile key which will relate to an emissions
species breakdown in Item 2 below.
These data were reported in two sorted orders:
a. According to the ARE application categories and including
point and area sources (Report #1)
b. According to location in 10 Km UTM grid squares and
including point and area sources (Report #2)
Two plant identification indexes were also provided as in the
preliminary inventory, one in alphabetical order by company
name (Report #3) and one by point ID no. (Report #4).
2. An Emission Profile Listing which lists each organic specie
(by name and code no.) emitted by a particular source or
source type, the reactivity class (according to the ARB's
3-class system) of that specie, and the percent by weight
of the total emitted hydrocarbons that the specie contri-
butes (Report #5).
P-193
20-13
-------
3. An Emission Summary consisting of the following data for each
1 Km grid square in the Basin (Report #6) :
a. Total organic emissions, ton/year
b. Individual specie emissions by code no., Ib/year
c. Emissions for each reactivity class: I, II, and III,
ton/year
4. An Individual Organic^ Specie Report showing the emission of
each specie broken down by ARE Application Categories (Report #7).
5. An SCC report listing the profile keys, application and emission
correction factors for all SCC codes encountered in the source files.
These data were reported in two sorted orders:
a. By SCC code (Report #8)
b. By profile key in order to reference all SCC codes
attributed to a given profile (Report #9).
P-193
20-14
-------
SECTION 3.0
EMISSION PROFILES
3.1 DESCRIPTION
A unique aspect of the current program was the development of emis-
sion profiles, the identification of the organic compound species represented
by the total hydrocarbon emission rates currently given in emission measure-
ments. Only one other study (Ref. 4) had previously attempted a breakdown
into generic classes. That was done primarily for the purpose of dividing
emissions into reactivity classes. The results of that previous study have
been widely used in the Basin.
A primary objective of this program was to identify the organic
compound emissions for each stationary source type in the Basin and develop
a data management system capable of applying this information to the total
hydrocarbon emissions in order to calculate the emissions of the individual
organic compounds. Thus an emission profile was formulated for each Source
Classification Code (SCC) emitting organic compound species in the Basin.
Both point and area sources were included. In certain instances a further
breakdown was made into individual industries identified by Standard Industrial
Codes (SIC).
Another objective of this program was to predict future emission
trends. Satisfying this objective required emission profiles based on SCC
number rather than individual plant profiles based on individual plant
characteristics. All plant devices identified by the same SCC and SIC
number were given the same emission profile. Conversely, it was important
that profiles be truly representative of the device in general. Additional
advantages of developing aggregate profiles by SCC number were that:
(1) estimations based on larger data samples were more statically reliable
than single data samples, (2) the profiles were compatible with the EIS
concept by describing devices by the SCC number system, and (3) the volume
of profile data was reduced to a more manageable level.
P-193
20-15
-------
In each profile the organic species were identified by their ap-
propriate SAROAD code and molecular weight. The ARE three class reactivity .
scheme, which describes reactivity by photochemical smog formation level, was
also included. Associated with each emission profile was a subjective
estimate of its relative error. Sources of the data for the development
of these profiles included KVB test data, KVB source questionnaires, the
relevant literature, and private communications with government, industry
and academic personnel.
3.2 METHODOLOGY
Two general approaches were used to formulate the emission profiles,
one where only one data point was available to characterize many sources
and another where multiple data points were available. In cases where a
profile was available from only one source and that source was believed
to be representative of all such source types in the Basin, then that
particular source emission profile was used. An appropriate error estimate
was given to reflect the relative confidence level of these data. It was
anticipated early in the program that a significant number of source types
would fall into this category due to the limited amount of field tests
available. Therefore, test locations were carefully selected on the basis
of the representative nature of their emissions to all other devices of that
particular type. In this way, data from this source could be correctly
applied to other non-tested sources. Similarly, questionnaires were
submitted to and received from selected solvent users. Follow-ups were
made to assure that the data from these large and representative sources
were obtained.
Two examples of formulating profiles based on one data point from
a selected source are the following. The emission profile typical of
residual oil fuel combustion was obtained by (1) recognizing that 95% of
all residual oil combustion in the Basin occurs in utility boilers,
(2) selecting a boiler that was "typical" of such devices in the Basin and
finally (3) conducting a test on this unit. Multiple samples were taken
and the profile was based on an average composition. Data from questionnaires
were used similarly. One source in the Basin, according to the SCAQMD files,
was responsible for 90% of the emissions from adhesive use. A questionnaire
P-193
20-16
-------
was mailed to this source and follow-up contacts were made to assure that
information from this source was received. The questionnaire contained a
comprehensive breakdown of the composition of the solvent composition and
usage which formed the basis for the emission profile.
The second approach used was to develop emission profiles based on
data from several sources within a particular source type. This involved
(1) acquiring the data, (2) determining the relative magnitudes of each
source compared to the total emissions from the source type and finally.
(3) forming a composite profile by factoring the data from each source by
an appropriate weighing factor. In this manner, emission profiles were
developed for individual source types that in actuality represented the
average emissions from sources of that category (SCC number).
An example of this approach was the formulation of a profile for
"Miscellaneous Organic Storage" in the Basin. There were SCC numbers assigned
to storage tanks for gasoline, jet fuel, crude oil, various solvents, etc.
This miscellaneous category covered those products not specifically identified.
Table 2 presents a summary of the calculation procedures employed to determine
this profile. Listed across the top are the various organic products iden-
tified and the fraction of the emissions from fixed roof tank storage for each
based on information compiled from the SCAQMD file. Listed down the page
are the various organic species that have been identified in emissions from
these products. The weight percentages of each specie associated with the
product is listed in the appropriate column. The weight percentage for
asphalt and Stoddard solvent were determined from KVB test data. The
adhesive percentages came from questionnaire data. The remainder of the
percentages were specified (e.g. 100% for acetone) or estimated based on
\
contacts with industry (e.g. the breakdown of alcohols and ketone). The
weight percent of each organic compound in the composite profile was determined
by multiplying the weight percents by the appropriate fractions and are listed
in the right hand side.
Given in Table 3 are the emission profiles generated for this study
for 20 of the major source types in the Basin.
20-17 P-193
-------
TABLE 2. COMPOSITE PROFILE FOR MISCELLANEOUS PETROLEUM STORAGE
(Fixed Roof Tanks)
Product Stored
Acetone Adhesive Alcohol Asphalt Perchloroethyleno
Ethylene
Dichloride
Formaldehyde Xatona Stoddard Xylane Others CoppoiUe
Fraction of
Emission*
Organic Compounds
Acetone
Perchloroe thylene
Ethylene
Dlchloride
Formaldehyde
H£K
HIHK
Xylene
Toluone
Ethane
Ethylone
Propane
H-Butane
I-Butane
N-P«ntane
I-P*ntano
Htxane
I-Hexane
Heptane
I-Heptane
I-Octane
I-Monana
I-Oecane
X-Undeeane
Ethyl Acetate
C-7 Cycle-
paraffins
iBOpropyl Alcohol
Ethyl Alcohol
Iiobutyl Alcohol
0.163
100.0
0.022
4.0
0.084
5.6
84.6
5.8
40.0
10.0
30.0
0.078
1.0
2.0
13.0
16.0
8.0
18.0
2.0
12.0
14.0
11.0
1.0
0.051
100.0
0.004
0.004
0.191
0.071 0.057
0.275
100.0
100.0
65.0
35.0
100.0
31.0
12.5
25.0
0.8
27.3
69.4
2.4
1S.S
15.5
16.4
S.I
0.4
0.4
12.4
6.7
5.7
8.6
0.1
0.2
1.0
1.4
0.6
1.4
3.4
8.9
0.9
l.l
1.0
0.1
1.9
4.9
0.2
4.4
4.2
3.4
2.6
2.6
P-193-
-------
TABLE 3. EMISSION PROFILES
IO
?
\o
nj
i
to
w
O C 3 C HI
° ^2 2
rH -rl >s -H n
Id »J M 4* -<
"OS 63 M
HA -rl A V '
w E "" C .WO*
. . Q> O WO -H w
J c 5 S
u nation u c E >-i cue
I U C Cl «1 0. -rt V Hfll 3 - -I «l C « _
rH > O «l > U > UV >IU Oil C-Ht>Ol
*< >ri~te«rtu -H Q)*H i-i6 cn -HU c
Q iJ w -H u ui p n x; u oo «« i-i«i , 4Jr-l 30 UK 310
O 9ECC3.HM 'O SH VO U4J lOrt MO
O t, |3 li M p. « W Q P. iX UU) OX UK)
36.0 28.6 62.0 6.2 10.2
S.8 17.8 3.9 5.6 3.6
11.5 11.1 17.6 1.8 23.4
0.1
18.3 4.4 27.1 19.8 26.5
1.8
7-4 2.9 1.5 6.6 5.4
7.7 0.7 14.6 10.6 7.3
7-8 1.1 1.5 29.0 8.0
6.2
3'4 59.0 7.9 2.5 5.6
1.4 9.2 0,6 5,6
1-8 6.9 o.l 3.0
'0.60 Ii0
13.0 1-6 0.3 8.9
0.0 2.0
0.3
0.4 0.8 1.9
Q.I 0.1 2.1
0.1
°-5 27.3 0.2
0.3 69.3
2.6
°'2 1.3 0.2
0.5
0.1
M rH **
9g U <0 «l 3
a ." n H « o
KC C 5' ° * =
Pub) 0. ? " <* DIKE
a > U rH « 3
4 fl I8O* flM **OC fl *H rt^HCn
M M hO U^-HIQ-HM M 41O4
U U VX UOXIMW3 T) UhX
c c c«i c u uhiof c nuo
01 « QJrH VQ U3QIO Q HU«'
ij u Ob. o£>
-------
TABLE 3. EMISSION PROFILES (Continued)
Profile (» Height)
1 N
M «
rt « H D< M S
*< H S ** C « -H
ocucn -~ fl *» e -3 i K
* .3 >, 8 * " *t*°" u^§ IH * S *
S«J W *J ^ t«l > O «l > *" ** > W 01 V4* Oil S -5 t« ** *<
W II M CP ^4 ^4*4C>HU -H WiHiHC Oi -HV I; <«
^3 C3 W b «IM'^4J«3 M f]U °S| ^S M«i)«4 h
Lactol Spirits
Nothyl Alcohol 1.0
Ethyl Alcohol 0.8
Iso Propyl
Alcohol 2.8
H- Butyl Alcohol , 4.0
Iso Butyl
Alcohol
Glycol Ether 11.5
Propylene Glycol
Ethylene Glycol
Ethyl Acetate 4.1 0.9
Propyl Acetate
Q H-Butyl Acetate . 9"5
iJj Cellosolve
O Acetate
Isopropyl 0 .
Acetate
Isobutyl
Acetate
Dimethyl
Formanide
Isobutyl I«o-
butyrate
Formaldehyde 42.0 7.6 51.0
Acetone 28.0 3-° 7<1
Methyl Ethyl 7 ?
Ketone
Methyl H-Butyl
Ketone
Methyl Isobutyl 3 x
Ketone
,_, Ethylene
1 Bichloride
ID Peetnvlene 3°'° I00-0
S - "
s u * «
3 -H 41 H U
I J J I «, J
« igot aki 4JUC 4
h MO WD>.»«^i M
V WX VO X!^4^* ?
c c« cu u w n u
« vi-t VQ H30 4
(5 Ofx (3«
-------
TABLE 3. EMISSION PROFILES (Continued)
to
Profile (% Height)
Chemical Name
.-*
° 1
*S **
3 W
*O 3
11
B o
w
!«
fr M
c a
£§
u O
1! '
£ »
h
V
M
8
«i *
V C H
> 0 0
H -H C
!H
|
M
J
«) Oi
"m o
£ «
1
g
0.
M
Of
>
~4
1
0<
.5
U tl
1 S
C U
W Q
U U
rH C
° >
iJ r-l
£5
^
O »
tr
u w
M
|
V C <0
-rt W C
in u p IB
O 3? U 3
14
-H
h
Ot
1
rH
i
£
,-4
(0
«
c
w
0
1
a
rH h
a o.
S 8
is
|
*-t 4
« U
ss
«ss
«
1
u
5 2.5
JC
-------
SECTION 4.0
EMISSION FACTORS
There has been considerable interest in the development of emission
factors that can be employed to estimate emissions from specific sources
based upon a knowledge of the pertinent operating characteristics of the
source. Such procedures are in common use throughout the country by local
control agencies to estimate air pollution emission rates for point and area
sources. One of the primary objectives of the ARB organic compound emission
study was to critically evaluate these emission factors and to develop new
emission factors applicable to the South Coast Air Basin. The following
discussion outlines the methodology employed during this analysis. Emission
factors for point and area sources were separated as they represent signifi-
cantly different approaches.
4.1 POINT SOURCES
A comprehensive listing of point source emission factors was found
in the EPA publication, "Compilation of Air Pollution Emission Factors"
(Ref. 5), hereafter referred to as "AP-42." The SCAQMD had their own emis-
sion factors which had been employed in the process of estimating emission
rates for industrial point sources contained in the EIS data file. To a
certain extent, these emission factors were the same, because frequently
SCAQMD data were used as the basis for the development of AP-42 emission
factors. In other instances the emission factors differed because the SCAQMD
often based their emission factors on their own test data in preference to
using AP-42 values.
It was recognized that the emission factors presented in AP-42 were
never intended to be totally representative of all such units within a specific
source type. Their widespread use, however, has generated several investiga-
tions by the EPA and others to broaden their data base and thereby increase
their usefulness and accuracy.
P-193
20-22
-------
A specific objective of this study was to examine the point source
emission factors used by the SCAQMD and AP-42. This was done for three
reasons. First, much of the data used to generate emission factors for
specific source types such as petroleum operations and the combustion of
fuels stemmed from studies conducted as far back as the 1950's. Considerable
debate has been raised about their continued applicability in view of
improved technology, sampling procedures, etc. Second, certain emission
factors listed in AP-42 intended for use nationally may not necessarily
represent conditions in the Basin. Finally, it was necessary to generate
entirely new emission factors where none had existed previously.
Field tests were conducted to provide data to assist in emission
factor evaluation and development. In addition, data from several related
projects specifically oriented to improving AP-42 emission factors have been
incorporated into this analysis. In most cases, these studies had been
directed at conditions within the Basin making them directly applicable to
the current study.
Comparisons have been made between the emission factors used by
the SCAQMD, those contained in AP-42 and those generated in this and related
studies. Where the SCAQMD emission factors have been shown to be in error,
correction factors have been applied to the emission rates listed in the
EIS data system to update these emission estimates. The intent was to
have the EIS data system reflect the best and most recent information avail-
able. This was a vital part of the improvements incorporated into the final
data base.
Since it was the intent of this paper to present general methodology,
the analysis used to investigate emission factors for the combustion of fuels
and establish correction factors will be discussed in detail as an example
of the approach. A similar analysis for petroleum operations including
marketing, storage and fugitive emissions will be incorporated into the final
report.
P-193
20-23
-------
It was appropriate to investigate the organic compound emission
factors for the combustion of fuels used in the Basin. Emission factors
used by the SCAQMD stem from data generated in the 1950's (Ref. 6) and
were currently under revision during the study using more up-to-date test
data.
AP-42 emission factors represented data accumulated over the last
several years and generally had an emission factor rating of A. These
emissions factors, however, are still subject to revisions for specific
sources.
Emission factors developed during the current program were themselves
subject to error due to the broad nature of the test program and the limited
number of samples that could be obtained for any one source type. Sources
tested were therefore carefully selected so that the tests were as representa-
tive as possible of sources of that general type. In addition, a thorough
evaluation of the test data was made to assure its accuracy.
The approach for this fuel combustion analysis was, therefore, to
compare the emission factors from the three sources (AP-42, SCAQMD and KVB)
and thereby ascertain the "best" emission factor. In cases where this
emission factor differed from that employed by the SCAQMD, an appropriate
correction factor was formulated and applied to the EIS data tape.
Table 4 presents a summary of the sources and emission factors for
the combustion of fuels that were evaluated. The table provides a descrip-
tion of the source, the appropriate SCC numbers, the units employed, emis-
sion factors listed in AP-42, those used by the SCAQMD, those resulting
from field tests conducted during this program and the emission factor used
in the final data processing. Also included is the correction factor used
to update the EIS data files as described above.
The analysis of the data presented in Table 4 resulted in some
interesting conclusions discussed in the following paragraphs.
P-193
20-24
-------
TABLE 4. ANALYSIS OF EMISSION FACTORS
Combustion of Fuels
in
Emission Factor
EIS
Applicable Emission Factors Used in the Correction
Description
Residual Oil
Combustion
Power Plants
Natural Gas
Combustion
Power Plants
Refinery Gas
Combustion
Natural Gas
Industrial
CO Boiler
Natural Gas
1C Engines
*Added to the
SCC Codes Units AP-42 SCAQMD KVB/ARB Current Study Factor
1-01-004-XX Ib THC/103 Gal. 1.0 2.6 0.7 1.0 0.40
1-01-006-XX Ib THC/106 ft3 1.0 8.8 - 8.8 1.00
3-06-001-02 Ib THC/106 ft3 30.0 21.9 20.0 21.9 1.00
3-06-001-04
1-02-006-XX Ib THC/106 ft3 3.0 7.0 12 7.0 1.00
3-06-002-01 lb/103 Bbl Feed 200 1.6 1.1 1.6 1.00
2-02-002-02 Ib THC/106 ft3 1400 - 1850- 1400 NA*
11600
data base by KVB
\ GscxyO) «= ^
.
P-193
-------
Among the most important of the emission factors investigated was
that from the combustion of residual oil in utility boilers. This represented
one of the largest uses of fossil fuels in the Basin. The emission factor
used by the SCAQMD was approximately 2.5 times that listed in AP-42. For
\
this source type, the results of the KVB test program conducted on a utility
boiler firing low sulfur residual oil tended to support the lower figure.
Consequently a correction factor was incorporated into the data management
program.
Similarly, the SCAQMD emission factor for natural gas combustion
in utility boilers was nine times that of AP-42. However, in this case,
since the quantity of natural gas used by utilities has decreased dramat-
ically in recent years, a decision was made to forego emissions testing of
this source type. Since this represented a relatively insignificant source
of organic compounds, the SCAQMD emission factor was not changed.
Refinery gas combustion, on the other hand, represents an important
industrial source of organic compounds. In this case, good agreement
between the three emission factor sources was obtained and no correction
factor was necessary.
Industrial natural gas consumption also represents an extremely
large energy use in the Basin. As shown in Table 4, the emission factor
used by the SCAQMD was between the AP-42 and KVB emission factors so that
again no correction factor appeared warranted.
Emissions from CO boilers treating exhaust gases downstream of an
FCC unit were also evaluated. Since the SCAQMD used test results from
these units for emissions estimates rather than emission factors, the
comparison between SCAQMD values and those obtained in this program has
been made for a particular unit. Again, the SCAQMD value appeared to be
reasonably close to that obtained during this study and no correction
factor was necessary.
P-193
20-26
-------
Natural gas combustion in 1C engines represented another somewhat
unique case. At the time of the study, sources of this type were not
included in the EIS data file although a preliminary inventory of 1C
engines in the Basin had been made. This inventory was incorporated into
the area source data base for this study. A decision on an appropriate
emission factor was somewhat difficult to make. Data presented in AP-42
represented emission factors of 800 hp units much larger than those typically
found in the Basin. The results of the tests conducted by KVB on 1C engines
also resulted in large emission factors however insufficient data were
obtained to generalize an emission factor. The AP-42 value was used as it
was somewhat conservative, although it was fully recognized that the emis-
sion rates from these sources may be higher.
In conclusion, the emission factors for the SCAQMD appeared
reasonably good, requiring only one correction factor be applied for the
sources listed in the EIS data file. Emission factors employed for natural
gas and refinery gas combustion appeared to be adequate for this program.
Emission estimates for CO boilers made by the SCAQMD also were verified.
There seemed to be additional data required to improve the emission factors
used for natural gas 1C engines.
4.2 AREA SOURCES
An important aspect of the KVB organic compound inventory was the
identification of sources of organic compound emissions not under permit
and generally not included or adequately characterized in previous inven-
tories. These sources were grouped as waste disposal, petroleum operations
other than refining, domestic and agricultural sources, geogenic sources
and natural emissions. Because these were diffuse sources, not concentrated
like industrial point sources, they were referred to as area sources.
Emission factors for these sources were therefore based on land area,
population, land use or other criteria characteristic of the area source.
Table 5 presents a summary of total organic emission estimates from
each of the area sources considered. Precise emission rates were difficult
to estimate due to the complex nature of each source type. A discussion of
inventory and emission factor development for some of the larger area sources
is included in subsequent sections.
P-193
20-27
-------
TABLE 5. EMISSION ESTIMATES FOR AREA SOURCES (TOTAL ORGANICS)
Waste Disposal
Agricultural Burning
Sanitary Landfills
Petroleum Operations
Production Operations
Marine Terminals
Gasoline Marketing
Natural Gas Transmission
Domestic and Commercial Sources
Architectural Surface Coatings
Solvent Use
Fuel Consumption
Agricultural
Natural Emissions
Orchard Heaters
Animal Wastes
Geogenic Sources
Natural Seeps
Forest Emissions
Natural Emissions
Forest Fires
Other Sources
Dry Cleaning
Asphalt Paving Operations
Tons/Day
2
930
60
3
83
83
93
31
5
14
3
77
11
1150
38
26
1
P-193
20-28
-------
It must be pointed out that the emission factors and inventories
presented were developed for use in this study of the California South
Coast Air Basin and caution must be exercised in their application to other
study areas.
4.2.1 Sanitary Landfills
Over 15 million tons of liquid and solid wastes were deposed annually
in the 45 major landfill sites within the Basin. Several studies (Refs. 7,
8f 9) indicated that appreciable amounts of methane rich gas were generated
due to the biological anerobic decomposition of these wastes. These gases
represented not only a potential source of useful energy but a large,
currently uncontrolled source of organic compounds to the atmosphere.
No precise estimate of the emissions from landfill operations for
the study area existed. Results from the above mentioned references and
field tests conducted as part of this program were used to estimate the
emissions from these sources.
The approach used was to estimate the rate of carbon escape over the
"life" of the fill as presented in a study by the California State Water
Quality Control Board (Ref. 10). In this study, it was found that
177
3.75 + 1.95t
where r = rate of carbon escape (Ib/ton refuse-yr)
t = age of refuse (years)
Note that it was assumed that carbon is released as both methane and carbon
dioxide gas. Using this relation and a gross estimate of the total quality
of wastes presently "alive" in the Basin, an estimate of the current organic
compound emission rate of 930 tons per day was made.
4.2.2 Petroleum Production Operations
Extensive petroleum production operations were underway in the Basin.
Nearly 150 million barrels of crude oil and 115 billion cubic feet of natural
gas were produced in 1975. For this inventory only onshore production opera-
tions were considered.
-*
P-193
20-29
-------
Prior to the current study, the magnitude and composition of organic
compound emissions fron production operations other than tank storage was
essentially unknown. Tests were conducted at two locations recommended'by
the Western Oil and Gas Association as typical of such operations in the
Basin. Since only brief test programs were possible during the current pro-
gram, the emission factors developed should be considered as representative
and useful for estimating purposes only.
A summary of the emission factors and inventories used in the current
study are given in Table 6. These inventories were made based on data from
the California Department of Oil and Gas (Ref. 12), data from local control
agencies (Ref. 13) and numerous discussions with representatives of the major
petroleum production companies operating in the Basin. It must be reemphasized
that the data represented the study area only and conclusions should not be ap-
plied arbitrarily to any other situation. Confidence factors on a scale of A
to E (A-high, E-poor) were also assigned to assist in the evaluation .of these.data.
As shown in Table 5, petroleum production operations represented
approximately 67 tons per day of which 50% were from storage tanks. The
balance were primarily fugitive emissions from leaking valves and metal con-
nections and evaporation from standing oil.
4.2.3 Gasoline Marketing
Orgariic compound emissions from the transfer of gasoline to automobile
tanks has been recognized as a major source of emissions to the atmosphere and
has been studied extensively. Control measures were implemented in counties
in Southern California to reduce these emissions through vapor recovery
techniques at both the tanker truck to storage tank transfer (Phase I) and the
nozzle to vehicle tank transfer operations (Phase II). Phase I had been
essentially completed within the Basin. Phase II had just been initiated.
Emission factors have been recently revised (Ref. 14) to reflect
additional test data on gasoline marketing operations and the effectiveness
of these control measures. The emission factors used in this study (given
in Table 7) represent not only information from Reference 14 but also
P-193
20-30
-------
TABLE 6. PETROLEUM PRODUCTION EMISSION FACTORS AND INVENTORIES
Source
Units
Emission
Emission Inventory Estimate(2) Confidence
Factor No. in Basin (tons per day) Level (3)
N>
Crude Oil Storage
Tanks (Fired Roof)
(1)
Process Drains Ib THC/rod pump well-day 2.0
Oil/Water
Separators
Fugitive Leaks
from Valves
Pump Engine
Exhausts
Heaters and
Boiler Exhausts
TOTAL
Ib THC/ft -day
Ib THC/valve-day
Ib THC/106 ft3
Ib THC/106 ft3
1400
30
1650 tanks
8000 rod pump
wells
34
B
0.1 184,000 ft
0.10 150,000 valves
4250xl06 ft3/yr
SlOOxlO6 ft3/yr
8
B
60
(1) Emissions adjusted to 60% of API Bulletin 2518 values based on the Engineering
Sciences report, (Ref. 11).
(2) Emissions have been rounded to the nearest ton per day.
(3) Confidence Levels; A - high, E - poor
P-193
-------
TABLE 7. GASOLINE MARKETING EMISSION FACTOR
Emission Source lb/103 Gal Throughput^
Submerged Filling of Underground 0.7
Tank (Controlled)
Underground Tank Breathing 1.0
Vehicle Refueling Displacement 9.0
Loss
Vehicle Refueling Spillage Loss 0.7
discussions with representatives of the EPA and the San Diego County APCD,
one of the most active of the local control agencies in assessing the validity
of these emission rates and effectiveness of Phases I and II control measures.
These emission factors have been incorporated with an estimated
gasoline sales of 5.3 billion gallons per year (1975) resulting in an average daily
emission rate of organic compounds from this source of 83 tons per day for the
Basin.
4.2.4 Architectural Surface Coatings
Estimates of the total volume of coatings applied to the surfaces of
stationary structure and marketed within the Basin was difficult to make due
to the large number of manufacturers and suppliers involved. The most effective
approach was to use marketing questionnaires. This approach proved to be a
very time-consuming and costly operation without the legal authority to require
responses by those questioned.
Surveys that were performed by the local control agencies in
California showed that emissions of organic compounds from architectural
surface coating applications were from 3.4 to 3.7 tons/1000 people/year.
The ARE estimated that 93 tons per day of emissions result from architectural
coating within the Basin (Ref. 15) for annual emission factor of 3.3 tons per
P-193
20-32
-------
1000 people. In the same study it was estimated that total emissions from
architectural coating for the entire state amounted to 186 tons per day
or an annual emission factor for the state of 3.3 tons per 1000 people con-
sistent with the estimate for the South Coast Air Basin. Thus 3.3 tons/1000
people/year was used in this inventory.
4.2.5 Natural Forest Emissions
Studies (Refs. 16, 17) had shown that there was considerable
organic interaction between plant life common to the forest areas in Southern
California and the surrounding atmosphere. These emissions are in the form
of a-pinenes, 3-pinenes and isoprenes generally termed terpenes.
Estimates of the emission rate from typical forest plant life groups
had been made by Zimmerman (Ref. 17) . A summary of forest group emission
rates are presented in Table 8.
Inventories of both the total average and the geographical distribu-
tion of forest vegetation were obtained from the National Forest Service
(Ref. 18). TO these were added estimated acreage of private lands assumed
to have similar forest vegetation compositions as in the National Forests.
Table 8 also presents a summary of the total acres of each forest group within
the study area.
Applying the emission factors and inventory results presented 'in
Table 8 resulted in a total emission rate from these sources of approximately
1150 tons per day. This amount was at least equal to that from all anthro-
pogenic sources within the study area. These results were not surprising
since nationally it was estimated that the natural emissions from such vege-
tations are 3-4 times the man-made emissions (Ref. 17).
P-193
20-33
-------
TABLE 8. SUMMARY OF NATURAL FOREST EMISSIONS
Forest Type
Total Acreage
in Basin
(1Q3 Acres)
Annual Emission
Factor
Tons/AcreYr
Emissions
Tons
Hardwoods
Douglas Fir
Mixed Conifer
Pines
Pinjon Juniper
Brush
Total
671
372
107
130
166
2147
3593
0.0211
0.0186
0.0782
0.0963
0.0353
0.172
38.8
19.0
22.9
34.3
16.1
1011.7
1142.8
P-193
20-34
-------
SECTION 5.0
FIELD TESTING
Field testing was conducted to determine the rate and chemical com-
position of organic emissions of representative sources in the Basin. Over
600 samples were collected and analyzed from various equipment in the
following locations:
Adhesives Mfg. Plant
Aircraft Plant (2)
Appliance Plant (2)
Asphalt Plant
Auto Body Shop (2)
Automobile Plant (2)
Chemical Plant (2)
Dry Cleaning Plant
Equipment Mfg. (2)
Gas Compressor Plant
Gas Pumping Station
Gasoline Station
Equipment tested included:
Adhesive Spray Booth
API Separator (6)
Asphalt Paving
Basic Oxygen Furnace
Blast Furnace
Charcoal Adsorbers (4)
Chemical Mill
Chemical Process
Chemical Transfer
Coke Oven
Compressors (28)
Cooling Tower (2)
Degrease Tank (11)
Dip Tank
Dry Clean Tumbler
Drying Ovens (8)
Fiberglass Impregnation (2)
Filling Rack
Flow Coater (2)
Gravure Press (5)
Landfill
Magnetic Tape Plant
Oil Field (2)
Oil Refinery (3)
Packaging Mfg. Plant
Printing Plant (2)
Roofing Kettle
Rubber Mfg.
Solvent Mfg. Plant
Steel Mill
Utility Boiler
Utility Gas Turbine
Heater Treater
I.C. Engines (6)
Incinerator (10)
Lithograph (3)
Open Hearth Furnace
Paint Booth (32)
Precip. Outlets
Printed Circuit Board Proc.
Process Heater (3)
Pumps (200)
Rubber Process (3)
Sintering Plant
Sludge Incinerator
Storage Tank (5) (Species only)
Sumps (6)
Valves (24,000)
Vapor Recovery Tank to Car (8)
Vapor Recovery Truck to Tank
Well heads (5)
P-193
20-35
-------
5.1 FIELD MEASUREMENTS AND SAMPLING
In the field,organic emission rates from stacks and from fugitive
emission sources like leaking seals, open ponds or spills were determined.
The general approach was either to measure the emission rate or to determine
it by calculations from process data or by experiment. From ducted sources,
such as stacks, emissions were determined by conventional velocity determi-
nations. Where information was available on the amount of product lost from
a process, this was used to determine emissions. Where fugitive emissions
due to leaks or spills were involved they were either measured or estimated.
In some instances special experiments were conducted to obtain estimates of
emission rates. An example of the type of experiments that were conducted
involved the determination of the amount of solvent which was emitted from an
architectural coating as it was drying or curing. Investigation in this area
has revealed that in some instances as much as 30 or 40 percent of the solvent
is actually retained in the paint after it is cured, and is not emitted.
For analytical purposes, samples of emission gases were collected in
one or more of the following type of containers:
tubes filled with activated charcoal
borsilicate glass bottles
Tedlar bags
glass bulb containing 1% sodium bisulphite solution
(aldehyde determinations).
The charcoal sorbent tubes were used to collect aliphatic organic
compounds with carbon numbers of six or greater and all other compounds from
C - up. The gas collection jars and bags were used to collect aliphatic
compounds with carbon numbers less than six. On most major sources, a combina-
tion of sorbent tubes and either bags or bottles were used. Bags or bottles
were used for the entire compound range when used for grab sampling.
P-193
20-36
-------
5.1.1 Sampling Equipment
Two identical portable sampling units were designed and built with
the following capabilities:
measuring stack temperature and velocity
filtering out particulates larger than 2 microns
collecting samples in sorbent tubes, glass jars or polybags.
Figure 1 illustrates the assembled sampling trains. Materials of
construction are as follows:
all metal components are stainless steel
seals are Viton or Teflon
containers are borosilica glass
flexible connections are latex rubber of minimal length.
The general flow diagram illustrated in Figure 2 illustrates all components
of the assembly which are available to be switched into several sampling
modes to conform to requirements dictated by the source to be tested. The
components are:
1. a sample nozzle
2. a filter holder with 2.5 micron pore size glass fiber filter
3. a filter and line heater and thermostatic control
4. an impinger train containing LiOH crystals
5. a borosilicate (Pyrex) gas collection bottle
6. a charcoal tube train with thermometer and vacuum gauge
7. a Brooks flowmeter with needle valve flow control
8. various interior and exterior valves and connectors as
indicated in Figure 2
9. a meter connection to PD gas meter
10. a Magnehelic velocity gauge and pyrometer for use with a
pitot tube.
The lithium hydroxide in the dry impinger train was used only on
combustion sources and was selected for use based on experience gained on
the Apollo space capsule. Initially an ice water impinger was considered
for moisture,' NOx, SOx, and CO removal. The problem with this approach was
that it was felt that the alcohols and some other oxygenates would form
P-193
20-37
-------
KVB
Figure 1. KVB hydrocarbon sampling trains.
20-38
-------
to
Heater
Gas Collection
Bottle
cxo-C
KX3 ,
Activated
Charcoal *
Sorbent Tubes
Gas Stream
Splitting Valve
Plow
Regulator
Sorbent
Gas Temp.
Flow
Meter
Sorbent
Pressure
Drop
Total Hydrocarbon
Analyzer
Aldehyde
Bulb
Sampler
System Flow
Throttle
Gas Stream System
Pressure Drop
(Meter Vacuum)
Vacuum
Source
Gas Meter
Figure 2. Complete hydrocarbon sampling train as set up for a hot combustion source
(> 180 °F).
P-193
-------
azeotropes with water and would not be easily separated for analysis. (The
impinger solution was analyzed for hydrocarbons.) LiOH was used in the
Apollo life support system to absorb primarily CO. In the sampling train it
also neutralized NOx and SOx which would react with the hydrocarbons. Also
the LiOH is hygroscopic and would absorb most of the condensed moisture.
Furthermore the LiOH would not adsorb hydrocarbons according to Apollo data.
CS extractions and hydrocarbon analyses were made on the impinger contents
and they were found to contain no hydrocarbons. On non-combustion sources
where moisture was no problem the dry impinger was not employed.
The suitability of several different types of sorbent materials
was investigated. The materials tested included: Tenax Gc, Carbosieve
B, activated charcoal, and XAD-2 resin. The criteria observed in the
selection of the sorbent included quantitative retention and recoverability
of every analyte possible. These qualities are dimensionalized by measure-
ment of breakthrough volumes and recovery efficiencies. Table 9 presents
the breakthrough volumes of the sorbents (25 °C) for hexane and benzene.
These analytes were considered to represent about the upper limit of
materials that can be analyzed in gas grab samples. Carbosieve B and
activated charcoal showed particularly high retention capacities.
Another important parameter in sorbent selection was the analyte
recovery efficiency. High temperature thermal stripping (with a purge gas
or in a vacuum) of the adsorbed components on Tenax, Carbosieve B and XAD-2
was considered but later rejected because the entire sample must be committed
in a single determination. Recovery efficiencies using the thermal/purge-gas
techniques also showed high molecular weight discrimination (see Table 10 ).
TABLE 9* RETENTION EFFICIENCIES OF VARIOUS SORBENTS
Breakthrough Volumes,* 1/g sorbent
Benzene Hexane
Carbosieve B 47 65
Tenax GC 3 4.4
XAD-2 Resin 12 20
Activated Charcoal 30 43
*Measured as the volume of gas/grams of sorbent in cartridge to give a 0.1%
FID response to gas stream containing 50 ppm of test component.
P-193
20-40
-------
TABLE 10. RECOVERY EFFICIENCY OF PURGE-THERMAL
STRIPPING OF SELECTED ANALYTES
Benzene
n-C?H16
n-C8H18
n-C9H20
n-C10H22
n~CHH24
n"C12H26
n-C13H28
n"C14H30
n"C15H32
n-C17H34
TENAX Carbosieve B
% Recovery % Recovery
105
100 11
99 <1
94 <1
72 <1
67 <1
67 <1
58 <1
56 <1
61 <1
46 <1
XAD-2
% Recovery
62
60
;
--
Solvent stripping for analyte elution prior to chromatographic
analysis was investigated. Carbon disulfide was found to be an attractive
solvent because of its excellent solvent properties. Many of the other
common solvents, such as methylene chloride, chloroform, hexane, benzene,
etc., tend to swamp the chromatogram, obliterating any signals of components
that have boiling points even decades higher.
Unfortunately, it was found that Tenax GC is soluble in CS as well
as in CH Cl . Carbosieve B showed poor recoveries with solvents. Testing
was therefore primarily focused on solvent extraction of- activated charcoal
with CS and XAD-2 resin extraction with CH Cl (CS also dissolved XAD-2).
^ £ £ £
Table 3-4 presents the results. Mueller and co-workers (Ref. 19 ) have
reported similar efficiencies for halogenated and oxygenated hydrocarbons
using charcoal adsorption followed by CS elution. Based on the data they
presented and the precedent set by the National Institute for Occupational
Safety and Health (NIOSH) in the selection and published (Refs. 20 to 22)
characterization of the charcoal/CS analysis scheme, the' use of coconut-
derived activated charcoal as supplied i>y Mine Safety Appliances or SKC, Inc.
was selected as the material of choice for source sampling.
20-41
P-193
-------
The total hydrocarbon analyzer shown in Figure 2 is the Bacharach
TLV Sniffer which was used to provide an indication of hydrocarbon concentra-
tion levels and process variations. At under $1000 in price it served its
purpose well.
5.1.2 Sampling Method
The sampling train shown in Figure 2 was used or modified for each
sampling. It was already mentioned that the LiOH impingers were removed for
other than combustion tests. On some simple one solvent sources even the
charcoal was omitted and only a gas bottle sample was obtained.
On fugitive sources the approach taken is shown in Figures 3 through
6. Figure 3 shows the setup for measuring leak rates when the rate is so
great that it will drive the meter itself. Most large pipeline leaks are of
this type. Figure 4 shows the setup if the leak will not drive the meter.
A pump is added to draw filtered air through the bag. The measured flow rate
times the hydrocarbon concentration measured by sampling the gas provided
a measure of the hydrocarbon emission rate.
Figures 5 and 6 illustrate test setup for sampling a high
temperature fugitive emission source. In Figure 5 aluminum foil was
substituted for polyfilm and rates were measured as Figure 3 or 4. When
the source was too hot, the foil could not be attached and the setup in
Figure 6 was used. The temperature of the source was measured, a grab
sample was obtained in a gas collection bottle, and the concentration of
total hydrocarbons was measured. The leak rate was then obtained by applying
engineering judgments.
5.2 LABORATORY ANALYSIS
Most ambient pressure gas samples were analyzed within 2-3 days
following receipt, except for a small number that were processed as long
as two weeks later. Several tests were made with synthetic samples to
evaluate storage effects on the contents of capped charcoal sampling tubes.
Recoveries did not change, within experimental error, between 24 hours and
P-193
20-42
-------
4 nil »»ty«thyl«m
<*«n top
Surga Boetia
J L
Cinooj- tMt«d) V.lv.
O O
Tout Hydroeubo*
SugiU
ottlt
Figure 3. Leak rate and concentration measurement of ambient temperature
fittings. High leak rates.
Mil
(*«OOS* H.t.dl V.lve
fucp
Open Top
tottl*
Figure 5. Leak rate measurement and concentration measurement of high
temperature fitting.
Figure 4. Leak rate by dilution sweep and sampling of ambient hydrocarbon
fitting. Low leak rates.
Figure 6. Hydrocarbon sampling from hot oil or solvent transfer.
P-193
-------
30 days. Therefore, the charcoal samples could stand for longer periods
without fear of losses, and were not usually analyzed until after the gas
samples in the same sets had been analyzed. The charcoal eluates were
usually run within an hour after the carbon disulfide was added to displace
the sample components.
Initial analysis of all samples was conducted using a gas chromato-
graph (GC). Lower boiling component identifications were based on retention
times established by repeated analyses of standards. If there were questions
as to the positive identity of a GC peak, the sample was rerun using GC/MS
methods for the identification. This approach was often necessary because
a number of chromatographic peaks contained at least two and sometimes three
components. The mass spectra also provided a basis for determining ratios
of the components in the GC peak being examined. These data were then used
in making quantitative measurements of the contents of chromatographically
unresolved but computer-integratable peaks.
A Beckman Model GC-55 equipped with a precision temperature programmed
column oven and a flame ionization detector (FID) was used for most of the
GC work performed on the program. The column used was 1/8" O.D. by 6 ft.
long stainless steel tubing containing a stationary phase of 100-200 mesh
Poropak Q. Using the analtyical conditions described below, this, column
furnished good resolution of the lowest boiling materials encountered while
still eluting with good results the higher boiling hydrocarbons representing.
the top of the range of interest.
Analyses were performed using helium as the carrier gas at a flow
rate of 30 cc/min. Detector gas flows were: H - 40 cc/min; air - 300
cc/min. The following conditions were used for GC analyses: 6 min. at 40 °C
followed by temperature programming at 10 °C/min to 190 °C and isothermally
held at 190 °C for approximately one hour.
The GC column effluent of the Beckman GC-55 gas chromatograph was
split into two streams. One stream was directed to the FID of the GC, the
other to a heated transfer line which carried the stream to a Finigan Jet
Separator and into the mass spectrometer. The separator provided a twenty
fold concentration of the material of interest in the gas stream.
20-44
-------
The mass spectrometer used on this program was a Consolidated
Electrodynamics Corporation (CEC) Model 21-104. This was a 180 degree
magnetic sector instrument having an electron impact ion source and an
electron multiplier detector system which permitted moderately high-speed
mass scanning.
Multiple MS scans were taken when a GC signal was observed on the
strip chart recorder. Multiple scan studies indicated that approximately
2 seconds were required for the maxima to be observed by the MS. Multiple
scans were required to insure representative ion pair formation.
Mass spectra were interpreted manually using such reference works
as:
"Compilation of Mass Spectral Data," Cornu, A. and R. Massot,
Heyden & Son, Ltd., London, England, 1966.
"Index of Mass Spectral Data," AMD II, American Soc. for
Testing and Materials, Philadelphia, 1969.
"Eight Peak Index of Mass Spectra," Atomic Weapons Research
Establishment, Aldermaston, England, 1970.
"Atlas of Mass Spectra Data," Stenhagen, E., et al., Inter-
science, New York, NY, 1969.
"API Project 44 Selected Mass Spectra Data," Thermodynamics
Research Center, Texas ASM University.
When an unknown peak could not be positively identified by this
means, the spectrum was compared with the mass spectra of some 27,000
different compounds in the library of the Cyphernetics Corp. Mass Spectral
Search System. This computerized search system is directly accessible on
a time-shared basis. It was successfully used to verify assignments made
during the earlier work on this program.
A spectrophotometric method similar to that specified by the NIOSH
was used for the determination of aldehydes. The total volume of liquid in
the aldehyde sample bulbs was measured, and an aliquot taken for the de-
termination. The sample was allowed to react with a modified Schiff's
reagent prepared from rosanaline hydrochloride and sodium bisulfite. After
a suitable development time, the adsorbance was read at 580 my against a
P-193
20-45
-------
reagent blank on a UV-vis spectrophotometer. Concentration was read from
a calibration curve. The same determination was performed on a sample of
the sodium bisulfite used for collecting/stabilizing the aldehydes and a
1 yg/ml formaldehyde standard. Results were calculated and reported as total
micrograms of formaldehyde equivalent in the sample. The minimum amounts
of aldehydes that can be detected by this method are typically 1-3 yg total
(as formaldehyde).
5.3 QUALITY CONTROL
Despite a program budget which limited the special tests designed
for quality assessment every effort was taken in the field and laboratory to
obtain as accurate and precise data as possible. A consulting firm,
EcoScience Systems, Inc. of Riverside, California* was retained to supervise
the quality program, i.e. specify tests to be conducted and procedures
to be followed and determine the overall error in data. Calibration gas
mixtures, round robin testing, unknown blank samples, redundant samples
were techniques used to determine the measurement error which was assessed to be
as follows:
the calculated total hydrocarbon emissions were good to
within +_ 25%
values for the emissions of individual hydrocarbons, however,
were less certain than that for total hydrocarbons
the sum of the errors in sampling and analyses for individual
alkanes probably was in the range of 25-50%, and
the concentrations of oxygenates, aromatics and halogenates
must be considered lower limits only with the possible error
being a factor of three or more.
*Principals included: Dr. James N. Pitts, Jr., Dr. Daniel Grosjean and
Dr. Barbara Finlayson-Pitts
P-193
20-46
-------
REFERENCES
1. Environmental Protection Agency, "Comprehensive Data Handling Systems,
Emissions Inventory/Permits and Registration Subsystem (EIS/P&R) Program
Documentation and Users Guide," February 1975.
2. Grisinger, J. E., "Development of Coordinate System Transformation
Equations Required for Air Quality Modelling in the SCAB," GARB Staff
Report, July 1977.
3. Environmental Protection Agency, "Compilation of Air Pollution Emission
Factors," Supplements 1-7, Publication AP-42, April 1977.
4. Trijonis, J. C. and Arledge, K. W., "Impact of Reactivity Criteria
on Organic Emission Control Strategies in the Metrolopitan Los Angeles
AQCR," Report for EPA Contract 68-02-1735, July 1975.
5. Environmental Protection Agency, "Compilation of Air Pollution
Emission Factors," Supplements 1-7, Publication AP-42, April 1977.
6. Personal communication with Wayne Zwiacher, SCAQMD.
7. Mery, R. C. and R. Stone, "Sanitary Landfill Behavior in an Aerobic
Environment," Public Works, January 1966.
8. MacFarlane, I. C., "Gas Explosion Hazards in Sanitary Landfills,"
Public Works, May 1970.
9. Dair, F. R. and R. E. Schwegler, "Energy Recovery From Landfills,"
Waste Age, March/April 1974.
10. California State Water Quality Control Board, "In-Situ Investigation
of Movements of Gases Produced From Decomposing Refuse," Publication
No. 31, 1965.
11. Engineering Science, Inc., "Hydrocarbon Emissions from Fixed-Roof
Petroleum Tanks," Western Oil & Gas Association, July 1977.
12. California Division of Oil and Gas, "Sixty-first Annual Report of
the State Oil and Gas Supervisor," Report Number PR06, 1975.
13. Personal communications with Robert Murray (SCAQMD) and Greg Barbaric,
(VCAPCD) and John Laird (SBAPCD).
P-193
20-47
-------
14. Burklin, C. E. and Honnckamp, R. L. , "Revision of Hydrocarbon
Evaporative Emission Factors," EPA 450/3-76-039, August 1976.
15. Air Resources Board, "Consideration of Model Organic Solvent Rule
Applicable to Architectural Coatings," Staff Report, June 1977.
16. Rasmussen, R. A., "What Do Hydrocarbons from Trees Contribute to
Air Pollution," Journal of the APCA, Vol. 22, No. 7, July 1972.
17. Personal Communications with Pat Zimmerman, Washington State
University.
18. Personal Communications with Mike Welsh, San Bernardino National
Forest, National Forest Service.
19. Mueller, F. X. and Miller, J. A., "Determination of Organic Vapors
in Industrial Atmospheres," Amer. Lab., 49-61, May 1974.
20. Levache, B.'and MacAskill, S. M., "Analysis of Organic Solvents
Taken on Charcoal Samplers," Anal. Chem., 48, (1), 76-78, 1976.
21. Nelson, G. O., et al., "Respiratory Cartridge Efficiency Statistics;
VII. Effect of Relative Humidity and Temperature," Amer. Ind. Hyg.
Assoc. J., 37, (5), 280-288, 1976.
22. Parkes, D. G., et al., "A Simple Gas Chromatographic Method for the
Analysis of Trace Organics in Ambient Air," Amer. Ind. Hyg. Assoc. J.
37, (3), 165-173, 1976.
P-193
20-48
-------
QUESTION:
ANSWER:
CONDENSED DISCUSSION
What provisions do you have for updating the
i nventory?
KVB does not have any contract or any respon-
sibility for updating it, per se. We have
turned over a data system in EIS format on
tape. If one wanted to update and change
emission factors, the program is written in
this way. If someone wants floating roof
tanks, for example, reduced by 60%, we can
do that. The program is set up that it can
be done. We will give to ARB a complete set
of documentations so they will know how to do
it. Not being a programmer myself, I don't
really know how easy that is done. Also
emission profiles can be changed by source
classification code.
20-49
-------
"HIGHWAY MOTOR VEHICLE EMISSION FACTORS"
Presented at the 1977
Environmental Protection Agency
Emission Inventory/Factor Workshop
Raleigh, North Carolina - September 13-15, 1977
By
Motor Vehicle Manufacturers Association
of the United States, Inc.
For Presentation at EPA Emission
Inventory/Factor Workshop
300 New Center Building
Detroit, Michigan 48202
September 13-15, 1977
Presented by
Walter S. Fagley, Jr.
Chrysler Corporation
P. 0. Box 1118
Detroit, MI 48288
21-1
-------
Abstract
The MVMA commends the holding of this workshop
intended to bring together information on the current
status, uses and needed improvements relating to
emission factors and inventories. The emission factors
and inventory data are a necessary part of information
to relate the complex phenomena of emissions and air
quality.
The purpose of this paper is to provide MVMA's
comments on the June, 1977, EPA Interim Document
entitled "Mobile Source Emission Factors" eventually
to be published as a supplement to AP-42.
The EPA document represents a monumental effort
and consequently these comments are not intended
to be exhaustive. While the Interim Document provides
substantially improved discussions of the methodologies
used to derive the emission factors, it lacks complete
and detailed documentation of the experimental data
base and the computational and judgmental methods
used for translating those data into emission factors.
The major thrust of this paper is on passenger
car exhaust and evaporative emission factors. The
paper reviews data from past emission surveillance
programs and compares the results to the emission
21-2
-------
factors contained in the Interim Document. It is
concluded that emission levels from uncontrolled cars
varied with model years. However, a single set of
"benchmarks" from which to judge progress is needed
and the 1960 model year was selected for this purpose.
Analysis shows that the 1960 model year baselines are
19 g/mi HC (10.6 exhaust, 4.3 evap., and 4.1 crankcase),
84 g/mi CO and 4.1 g/mi NOx. Exhaust emission values
are based on the 1975 Federal Test Procedure.
In estimating the contributions of vehicle
emissions to overall air quality, the emission
factors for controlled carspast, present, and
futureare more significant than emissions from
uncontrolled cars. The EPA Interim Document appears
to substantially overstate emissions from controlled
cars. The overstatement of expected emission rates
for future cars is very large, if one concedes even
moderate success by both manufacturers and regulators
to improve emission performance of in-use vehicles.
21-3
-------
Introduction
The purpose of this paper is to comment on the
motor vehicle emission factors contained in the June,
1977 EPA Interim Document entitled "Mobile Source
Emission Factors" (Reference 1). This EPA Interim
Document has been proposed to eventually revise the
previous mobile source emission factors that are
contained in the October, 1975 Supplement No. 5 to
AP-42. In commenting on the proposed revision to
Supplement 5, it should be pointed out that the basic
purpose of AP-42 is to provide a compilation of
emission rate data without attempting to relate the
complex phenomena of emissions and ambient air quality.
Because air quality models are continually being
revised and improved, it is entirely proper to place
the responsibility of correctly applying the emission
factors with the user of AP-42.
However, we believe the users of AP-42 should be
aware of the uncertainties involved in the emission
factors. Data can only be intelligently applied if
the user is fully aware of the associated questions
*
and qualifications. It is to the credit of EPA that
this workshop allows airing of any controversy related
to its published emission factors. In this same
21-4
-------
spirit, we appeal throughout this paper for more
complete documentation of the bases for emission
factors to be published in the future. Such
documentation would afford all future users some of
the advantages which will accrue to the attendees of
this workshop.
This paper deals with mobile source emission
factors and discusses some of the problems and
uncertainties in the application of emission factors.
Most of the discussion centers on data related to the
1975 FTP certification driving cycle and procedures
developed to simulate urban driving, low average
speed, under conditions which may contribute to peak
oxidant pollution. There are, however, other urban
driving cycles and procedures characterized by low
average speed (under 20 mph), such as the US 7-mode,
the European ECE Regulation 15 and the Japanese hot
10-mode and cold 11-mode cycles.
Although the low speed urban FTP cycle is the
basis for U.S. certification, some effort has gone
into characterizing emissions under other conditions.
Emission data are sometimes available on special
purpose cycles such as the EPA highway and sulfate
cycles; the SAE urban, suburban, and interstate
21-5
-------
driving cycles; and on various short test cycles that
are under investigation to find a less expensive way of
measuring emissions. Since emission rates can be quite
different on each of these cycles, care should be
exercised in deciding which cycle and what adjustments
are necessary for the particular application at hand.
In practice, one usually selects the cycle for which
there are data and then may make appropriate adjust-
ments if needed. Suggested procedures for adjustments
are included in AP-42 related to the 1975 FTP
certification cycle which is usually the cycle
selected for current applications in the United
States. The validity of such adjustments is still
under investigation.
As one example, EPA is currently studying the
possibility that the emission factors generated using
the 1975 FTP driving cycle overestimate the relevant
CO emissions. This would be due to the fact that the
1975 FTP cycle has a cold start causing FTP CO
emission factors to be relatively high compared to hot
vehicle operation, particularly for catalyst cars.
Since most vehicles are usually warmed up by the time
they reach (and have influence on) a CO "hot-spot"
in a city, the 1975 FTP cycle would not be appropriate
21-6
-------
and should be adjusted for such cases. Since traffic
flow varies during the day and during the year, speed
correction factors and other adjustments should be
based on the conditions leading to high ambient CO
concentrations rather than blindly using only the FTP
cycle emission factors.
When applying emission factors, one has to deal
with the definition of "ambient air" which is not
precisely defined. At the tailpipe, CO emissions
are concentrated. Somewhere between the tailpipe
and people the air quality standard is intended to
be achieved. Exactly where this point occurs depends
primarily on judgment. For example, New York City is
reported to have the worst carbon monoxide pollution
problem in the United States, based on EPA and New
York City measurements. However, part of the reason
for the relatively high CO measurements is due to
differences in the location of the measurement
equipment. In New York City the sampling probe is
closer to automotive tailpipes than sampling probes
in other cities allowing less time for dilution of
carbon monoxide. According to the President's Council
on Environmental Quality (Reference 2):
21-7
-------
"The New York City location of the
sampling probe was at 8 feet above the
curb; all other sites measured CO at
heights of 8 to 18 feet, but at the
building face several feet back from
the curb. One would expect that if the
New York measurements were made at the
building face, as in other cities, the
measured CO levels and their frequencies
of occurrence would be somewhat lower.
Mobile monitoring data taken by the
local New York agency, support this
conclusion."
It would be extremely unfortunate if trans-
portation control plans were established from
measurements based on the shortest distance between
sampling probe and tailpipe without considering
public health impacts.
It has been assumed in the past that the exhaust
emissions from cars built prior to first exhaust
control in 1968 (1966 in California) could be
represented by a single set of emission rates
i
irrespective of model year. Such values have been
21-8
-------
important not only as emission factors used in
calculation of emission inventories, but as
"benchmarks" from which progress in emission control
could be based. Analyses developed later in this
paper lead us to conclude that this assumption is
incorrect.
As time progresses, the fraction of total
vehicle miles traveled (VMT) represented by pre-
controlled cars becomes smaller and smaller.
Furthermore, the proliferation of air monitoring
equipment has substantially decreased the possibility
of interest in selecting "base years" for air quality
projection calculations which would involve a dominant
fraction of total VMT accounted for, by uncontrolled
cars. Therefore it might be argued that concern over
the emission levels from such cars is academic.
While this may be largely true for the emission
factors use of uncontrolled car emission levels,
there remains a need for a set of benchmarks from
which to judge progress. It has little meaning to
the layman to state that today's cars are controlled
to carbon monoxide emission levels of 15 g/mi.
However, one can make a reasonable judgment from the
statement that passenger car CO emissions are
21-9
-------
controlled to 82% of the level from uncontrolled cars.
We therefore believe it is desirable to have the best
possible assessment of emission levels from uncon-
trolled cars from available information.
However, since the best information argues
against a single set of uncontrolled car emission
rates, it is necessary to select a model year
definition of the desired benchmark. We have chosen
to define this benchmark as a 1960 model year
passenger car. The reason for this choice is that
1960 was the most recent model year for which there
were no emission controls. The positive crankcase
ventilation system (PCV) to control blow-by crankcase
emissions was introduced in 1961 in California.
Nationwide installation began in 1963. To complete
the history, exhaust emission control was first
established in California for 1966 model year vehicles
Federal exhaust emission control was first applied
for model year 1968 vehicles. In 1970, evaporative
emission control was added to California vehicles
and nationwide in 1971. Therefore, an uncontrolled
evaporative vehicle could be defined as a 1969 or
older vehicle.
21-10
-------
Due to the different dates for implementing
emission controls, a number of possible definitions of
uncontrolled vehicles could be made. Although model
year 1960 was selected to represent baseline emissions
for completely uncontrolled passenger cars, data from
other model years are also presented for comparison
should a different definition be desired.
Uncontrolled Crankcase Emissions
Uncontrolled crankcase emissions were listed
as 4.1 g/mi hydrocarbon in EPA's Interim Document.
It appears that the 4.1 g/mi crankcase emission rate
for 1959-1962 cars was obtained from the work of
Kramer and Cernansky (Reference 3). This value was
estimated by assuming an average blow-by volume of
1.1 cfm and an average hydrocarbon concentration of
15,000 ppm (as hexane) and an average urban area
speed of 25 miles per hour. Blow-by flow depends
not only upon vehicle speed but also upon fuel
tetraethyl lead concentration and mileage
(Reference 4). We are aware of no data which would
indicate the 4.1 g/mi rate is not a good estimate;
therefore, we endorse that value.
21-11
-------
Uncontrolled Exhaust Emissions
Sources of data on in-use exhaust emissions from
1965 and older vehicles are listed in Table 1. Since
the Federal Test Procedure for exhaust emissions has
changed several times over the past, some of the
earlier data must be converted in order to compare
to more recent data using the 1975 FTP (see the
Appendix for a discussion of conversion factors).
There are uncertainties in this conversion process.
However, since the largest body of data on uncontrolled
passenger cars (over 1,200 vehicles) was taken using
tailpipe concentration measurements, this estimation
of 1975 FTP emissions was considered worthwhile.
The conversion was found to compare favorably with
measurements on uncontrolled vehicles using the 1975
FTP procedures directly.
Data on uncontrolled exhaust levels are contained
in Table 2 and are plotted in Figures 1, 2 and 3.
The "mean line" in each Figure was plotted based
on a 2nd degree regression analysis according to the
methodology outlined in the Appendix. It is
interesting to note that uncontrolled exhaust emissions
appear to have changed over time. Hydrocarbon and
carbon monoxide emissions appear to have decreased
21-12
-------
from 1947 to 1965 while NOx increased. One reason
for this change can be attributed to a shift in model
mix incorporating greater numbers of automatic
transmissions. Other changes in engine design
(e.g., increasing compression ratios and the shift
from L-head to overhead valve engines) also had their
effect.
Based on the "mean line" in Figures 1, 2 and 3
the uncontrolled (model year 1960) baseline exhaust
emission level on the 1975 FTP is 10.6 g/mi HC, 84
g/mi CO, and 4.1 g/mi NOx. Emission levels in EPA's
Interim Document for model year 1960 cars driven
in calendar year 1970 were listed as 9.9 g/mi HC,
124.8 g/mi CO, and 3.4 g/mi NOx. The basis for EPA's
numbers is not clear. Carbon monoxide emissions
appear excessively high. There appears to be an
obvious need for better documentation of the basis
for that value.
Uncontrolled Evaporative Emissions
Table 2-13 in EPA's Interim Document includes
evaporative emission rates in grams per mile for light
duty gasoline powered vehicles. A value of 2.53 g/mi
was used for cars without evaporative emission
controls through and including model year 1970.
21-13
-------
There is a preponderance of evidence demonstrating
that this emission rate for cars without controls
is incorrect. This rate errs on the low side as
a result of neglecting fuel tank running losses.
In addition, there is a less serious but still
important controversy over conversion of the regulatory
gram per test basis for expressing evaporative
emissions to the more useful g/mi basis. The first
apparent publication of such a conversion was made
by Professor John B. Heywood of MIT (Reference 5).
Professor Heywood proposed the following formula:
EVAP g/mi = PS g/day HS g/test
27 mi/day + 7.5 mi/test
Where DS = Diurnal Soak Emissions
Where HS = Hot Soak Emissions
The EPA Interim Document, on the other hand, used
the following formula:
EVAP g/mi = DS g/day + 3.3 trips/day x HS g/test
29.4 mi/day
21-14
-------
The 3.3 trips per day and 29.4 miles per day values
are indicated to be national averages.
To illustrate the magnitude of the difference
between these computational methods, the EPA Interim
Document uncontrolled value of 2.53 g/mi would become
2.92 g/mi under the MIT method. While there is a
difference of approximately 2-1/2 miles per day
between these two methods, the more important
discrepancy is the effective miles per trip difference,
The Interim Document 29.4 mi/day with 3.3 trips
per day amounts to 8.9 mi/trip compared to the 7.5
mi/trip of the MIT method. While numerically smaller
than the mi/day discrepancy, this is fractionally a
larger difference. In the case of controlled cars,
the daily mileage divisor generally operates on a
substantially smaller g/test value, thus it makes
little difference in the diurnal soak contribution
to the g/mi value, whether one assumes a daily total
of 27 or 29.4 miles.
Since urban emissions are of concern, it would
appear that urban statistics rather than national
average statistics should be the basis of emission
computations. Urban statistics are presumably the
basis of the exhaust emission 7.5 mile "test trip"
21-15
-------
and since this is the actual trip which "drives"
the measured hot soak emissions, it seems more
reasonable to us to select the 7.5 mile per trip
value.
The Interim Document emission factor of 2.53
g/mi is based on data from EPA's Emission Factors
Surveillance Programs, primarily the FY 71 program
(Reference 9). The FY 71 program unexplainably
neglected to include measurement of tank running
losses on uncontrolled cars. Virtually all gasoline
powered passenger cars built in the last 40 years
used vented fuel tanks (prior to evap. control).
Since automobile fuel tanks are heated during operating
of vehicle as a result of engine heat rejection, the
tank vapor space will necessarily exhale HC vapors
(and air) through the tank vent during driving periods.
However, for the purpose of emission factors, 10 g/test
is a representative value (Reference 8). Therefore,
we recommend the following formula be used:
EVAP g/mi = PS g/day + HS + RL g/test
27 mi/day 7.5 ml/test
21-16
-------
Published values of running losses from
automobile fuel tanks (References 6 and 7) vary from
8 to 52 g. If this latter value is added to the 14.7
g/test hot soak value of EPA's Interim Document for
the uncontrolled car and the MIT method of conversion
is applied, a "correct" value for the uncontrolled
car of 4.3 g/mi is obtained.
It is recommended that first the regulatory
test-to-g/mi emission rate conversion be based on the
MIT method; secondly, that the emission factor for
cars without evaporative emission controls be
corrected to include tank running losses, specifically
a value of 10 grams for the LA 4 test. This would
result in an uncontrolled emission rate of 50.6 grams
per test or 4.3 g/mi.
Uncontrolled Baseline Summary
Based on the above analysis, the following base-
line was determined (1960 model year):
HC CO NOx
Exhaust 10.6 g/mi 84 g/mi 4.1 g/mi
Evap 4.3 -
Crankcase 4.1 -
TOTAL 19.0 g/mi 84 g/mi 4.1 g/mi
21-17
-------
Emissions from Controlled Vehicles
Controlled Crankcase Emissions
It is indicated in the Interim
Document that "...crankcase hydrocarbon
emissions from post-63 vehicles are
negligible." The values in Table 2-13
of that document, on the other hand,
imply a crankcase emission rate for
MY 1963-1967 of 0.8 g/mi. This is
consistent with the information in
Reference 3 that the so-called open
crankcase ventilation system applied
during those model years was 80%
effective in controlling crankcase
hydrocarbon emissions.
We believe the implied 0.8 g/mi rate
for MY 1963-67 is a reasonable estimate
and suggest the text be altered to
correct the associated error. A sep-
aration of the tabular values for crank-
case and evaporative emission factors
in the Interim Document would further
clarify the situation.
21-18
-------
We concur with the post-1967 crankcase
emission factor of zero, as given in the
Interim Document. The "closed positive
crankcase ventilation system" used since
that time vents both the crankcase air
inlet and the ventilation outlet to the
engine induction system. Since crankcase
blow-by emissions can only occur with the
engine operating, and the engine induc-
tion system is under negative pressure
when the engine is operating, it is
not possible for emissions to escape
from the crankcase directly to the
atmosphere.
Controlled Exhaust Emissions
Sources of data on in-use exhaust emissions
from controlled passenger cars are listed
in Table 3. Most of these data come from
EPA's in-use surveillance programs. Analysis
of these data will be divided into 1968 to
1974, and post-1974 model year categories.
21-19
-------
- Post 1974 Cars
Tables 2-1 and 2-2 in EPA's Interim
Document list a common set of HC
and CO exhaust emission factors
Cover the life of the car) for model
years 1975 through 1978. The tabu-
lation on page II-9, on the other hand,
lists equations for computing emission
factors as a function of chronological
age for several standards scenarios
which might apply to post-1978 cars.
These emission factors reflect extremely
high additive "deterioration rates",
particularly with respect to the appli-
cable emission standards. The accom-
panying text (pages II-6 and II-7)
indicates that these emission factors
were obtained from mileage regressions
of FY 1974 program data of 1975 cars
with some adjustments made (in the case
of CO) to compensate for "... bias due
to the rapid deterioration of the mal-
adjusted vehicles..."
21-20
-------
On Figures 4 and 5 the HC and CO
emission factors are plotted from
EPA's Tables 2-1 and 2-2 as a
function of miles based on the Interim
Document's indicated assumption of a
constant 10,000 miles per year. The
figures also contain linear regressions
of the individual 1975 car data from
the FY 1974 program (Reference 10).
In the case of exhaust HC, there is
virtually a perfect match of the least
squares straight line fit of the data
with the Interim Document emission
factors of Table 2-1. In the case of
the CO data (Figure 5) the change
stated in the Interim Document made
to the slope of the linear fit for
the "maladjustment bias" is evident.
The actual mileage extent of the test
data in Figures 4 and 5 was only
slightly over 40,000 miles. Thus,
the extension of those lines to the
entire life of the car represents
21-21
-------
substantial extrapolation and
possible large errors. Furthermore,
the data are assumed to be represented
by a straight line. The corre-
lation coefficients of .195 and
.127 for the HC and CO data respec-
tively suggest that this is an
arbitrary assumption.
Also shown in Figures 4 and 5 are
the plot of a least squares fit of a
logarithmic mathematical model,
perhaps as equally arbitrary as a
linear model but, on the other hand,
just as rational (correlation
coefficients with the log model are
also low at .164 and .133 for HC and
CO respectively). The logarithmic
model of these alternate curves
is similar to that employed by the
California Air Resources Board for
a number of years in analyzing its
surveillance data (Reference 11
for example). The logarithmic
21-22
-------
model suggests substantially lower
emission factors, particularly
at the higher mileage associated
with vehicle ages beyond four or
five years. The logarithmic model
assumes that the rate of deterioration
of exhaust emissions will be a con-
tinually decreasing function of miles.
It is suggested that this is con-
sistent with the assumption made in
the Interim Document that emission
levels cease to deteriorate after
10 years or 100,000 miles. At
some point emissions would cease
to increase with age.
Data on NOx are plotted in Figure 6.
The linear regression line through
the FY 74 data has a significantly
lower slope than that of the Interim
Document emission factors. Further,
the logarithmic fit of the FY 74 data
shows very low deterioration beyond
the initial 20,000 miles. It is
21-23
-------
believed the latter is more consistent
with actual observations of high
mileage NOx emissions. It is a
common observation that if CO emissions
increase with mileage as a result of
air/fuel ratio richening, NOx emissions
are expected to decrease concurrently.
The brief analysis represented by
Figures 4 to 6 strongly suggests
that the very large deteriorations
assumed by the emission factors
even for 1975 cars are unrealistic.
Furthermore, the assumption that such
levels of deterioration will continue
indefinitely for future model years is
even more unrealistic. While any
assessment of the FY 74 data indicates
that CO emission levels from 1975
cars are unfavorable relative to the
applicable standard, it is reasonable
to expect that situation to improve
as a result of the normal "learning
curve" process.
21-24
-------
Moreover, there is activity underway
both by the manufacturers and within
the regulatory community to reduce
the probability and magnitude of
field maladjustment which appears
to be contributing so substantially
to this discrepancy between in-use
car emissions and their certification
levels (References 12, 13, 14).
Given the present statutory and
regulatory climate for air pollution
control, the consequences of grossly
overstating probable emissions from
future passenger cars is profound.
MVMA, therefore, strongly urges that
the analytical bases and assumptions
used to generate the post-1974 model
year exhaust emission factors of the
Interim Document be re-examined.
21-25
-------
- 1968 through 1974 Cars
It is disconcerting that the emission
factors for these model years listed
in Tables 2-1 and 2-2 of the Interim
Document show the same levels of
exhaust hydrocarbon and carbon monoxide
control throughout. The HC and CO
standards, on the other hand, represent
three levels of control with stringency
of the original 1968 standards (1966
in California) having been increased
in 1970 and again in 1972. The oxides
of nitrogen emission factors of Table
2-3 appear to be at least directionally
consistent with the certification
standards throughout these model years.
These emission factors are indicated
by the Interim Document to be based
on the data from EPA's Emission Factors
Surveillance Programs of Fiscal Years
1971, 72, 73 and 74. The data for
those programs prior to the FY 74
21-26
-------
program have not been available to
us in the form which facilitated the
analysis of Figures 4 through 6 for
the 1975 cars data for which were
included only in the FY 74 program.
However, test data for some vehicles
of the 1968 through 1974 model years
were included in the FY 74 program.
We have consequently performed
similar analyses of those data com-
paring the logarithmic with the
linear model.
Figures 7 through 9 depict those
analyses for the 1974 model data
illustrating that the relationship
of the logarithmic to the linear
models are generally similar to
that shown in Figures 4 through 6
for the 1975 model year again
suggesting that the linear model
overstates the emission rates at
relatively high mileages.
21-27
-------
An interesting relationship is
depicted by Figure 10 which shows
the analysis for exhaust hydrocarbon
emissions from 1973 model year cars.
Here the linear model shows a
decreasing emission rate trend with
increasing miles, whereas the
logarithmic model shows the normally
expected, albeit very modest, increase
in HC emission rate with miles.
As with the prior discussion of the
1975 model year data analysis, the
correlation of the emission data with
miles is very poor whether the linear
or the logarithmic model is chosen.
Correlation coefficients for these 21
data sets vary from .006 to .327.
In some cases, the logarithmic model
displays slightly better correlation
coefficients whereas in other cases
the linear model appears to be a
better fit. While MVMA does not
21-28
-------
consider the "score" significant,
the logarithmic model shows a better
correlation for 12 out of these 21
data sets.
We believe these limited analyses
of a portion of the data, which is
indicated to support the emission
factors for the 1968 through 1974
model years, strongly suggest that
those emission factors overstate
the emission rates, particulary
for HC and CO at high mileages.
We therefore urge the EPA to re-
analyze the total body of Emission
Factor Surveillance data considering
alternates to the linear model which
could provide a significantly improved
expression of the emission factors
as a function of miles.
Controlled Evaporative Emissions
Table 2-12 of EPA's Interim Document indicates
evaporative emission rates for controlled light duty
21-29
-------
gasoline powered vehicles to be 1.76 g/mi for 1971
through 1977 cars, and .60 g/mi for 1978 and 1979
vehicles. The value stated for the first generation
of evaporative controls (though 1977) appears to
be incorrect. That rate errs on the high side
because of reliance on a single data body which
apparently involves erroneous measurements.
In addition, there is the less serious but still
important controversy over conversion of the
regulatory gram per test basis for expressing
evaporative emissions to the g/mi basis as dis-
cussed earlier. The 1.76 g/mi value would
become 2.05 g/mi under the MIT method (Reference 5).
The Interim Document emission factors are based
on data from EPA's Emission Factors Surveillance
Programs, primarily the FY 71 program (Reference 9).
The indicated emission levels from first generation
controlled cars (1971 through present) appear exces-
sively high. Table 5 compares similar data from
three separate independent programs with the FY 71
data. The independent data referenced in that
Table indicate that the EPA Emission Factor
Surveillance Program measurements of 24-27 grams per
test are high by more than a factor of 2. The
21-30
-------
California Air Resources Board (GARB) program involves
several times more in-use cars than did the EPA
program. Furthermore, the GARB program was begun in
calendar year 1975 and benefitted by the mistakes
made in the earlier EPA programs.
These high values obtained in the EPA surveillance
programs have not been explained. However, from
available information it appears probable that
installing the fuel tank thermocouple through the
fuel tank cap may have resulted in vapor leaks. This
assumption may explain both the larger discrepancy of
the diurnal soak values (diurnal soak emissions are
virtually 100% tank emissions) and the agreement
of the FY 71 results from uncontrolled cars with the
previously accepted values. Such a vapor leak would
not affect emissions from uncontrolled vehicle tanks
which are already vented to atmosphere.
Another possible explanation for the high
emissions from the FY 71 program is the difficulty
of proper placement of the tank fuel thermocouple
with the lead-through-the-cap approach. With that
approach the thermocouple junction can easily lodge
in the fill pipe or the tank vapor space. Attempts
to drive the resulting vapor temperature, rather
21-31
-------
than the intended liquid temperature, through the
prescribed excursion by heating the tank liquid,
results in unrealistically high emissions.
A weighted average of the data from the first
three test programs of Table 5 yields 7.4 g/test
as representative of the first generation of emission
controls. Similar treatment of the modal values
yields .81 g/mi by the MIT method (i.e., 1.8 g diurnal
and 5.6 g hot soak). The 1978 6 g/test standard,
on the other hand, assuming a 1.5/4.5 gram diurnal
soak to hot soak distribution results in .65 g/mi.
Truck Exhaust Emission Factors
The subject of truck emission factors is so
complex that an attempt will not be made to provide a
detailed treatment in this paper. However, a few
brief comments are in order since there are serious
questions about the validity of the truck emission
factors contained in the Interim Document.
Actually, there are three discrete populations -
the light duty vehicle (less than 6000 Ibs. GVW),
the intermediate truck (6001-8500 GVW), and the heavy
duty truck (over 8500 Ibs. GVW).
Because of differences in driving patterns,
load factors, N/V ratio, weight/power ratio and other
21-32
-------
factors, the use of passenger car deterioration rates
to the intermediate class of vehicles cannot be
justified. The intermediate vehicle will not be
subjected to a chassis dynamometer certification test
until the 1979 model year, so there is almost a
total absence of vehicle data to provide an
uncontrolled baseline.
The heavy duty truck is not subjected to
certification testing. In this category, the engine
is tested and certified. A separate MVMA panel is
currently engaged in a project to develop factual
heavy duty engine baselines. As with the inter-
mediate vehicle, the extensive data and deterioration
rates determined for passenger cars and intermediate
trucks to the heavy truck category cannot be justified.
Data are simply not interchangeable between the
three populations.
As an example, our concern over the validity
of the truck data, Table 3-1 of the Interim
Document shows higher than precontrol carbon monoxide
emissions for the 1970-78 model year trucks C6001 to
8500 Ibs. GW) . By contrast, an analysis of the test
results from the same program, as shown in Figure 11
(Reference 15), indicates more than a 60% control of
21-33
-------
CO from 1972-73 trucks. This discrepancy between
two analyses is not explained. Figure 11 also raises
a similar question on exhaust hydrocarbon emissions.
In the case of hydrocarbons, the new vehicle emission
rates of Figure 3-1 indicate a control of about 33%,
whereas, the data in Figure 11 suggests control
of more than 50%.
It. is further indicated in the Interim Document
that the average age of the uncontrolled 6001-8500
Ibs. GVW trucks used in the program was seven years
and that the new truck emission factor was
extrapolated back from the deterioration rate of
precontrolled passenger cars. In view of our earlier
comments on passenger car deterioration rates, the
application of these deterioration rates to a
completely different class of vehicles is unjustified.
21-34
-------
Summary
The following summarizes the major conclusions of
this paper:
. Emission factor data in AP-42 are based on
the 1975 FTP cycle used for U.S. certification.
In applying such data the user should first
check to see if this cycle is applicable
to his analysis. If not, appropriate
adjustments should be made.
Emission levels from uncontrolled cars was
found to vary with model year. Model year
1960 was selected as the uncontrolled
"benchmark". Uncontrolled emissions from
1960 model year passenger cars were found
to be 19 g/mi HC (10.6 exhaust, 4.3 evaporative
and 4.1 crankcase), 84 g/mi CO and 4.1 g/mi NOx.
The AP-42 uncontrolled CO emission rate
of 124.8 g/mi for 1960 cars driven in calendar
year 1970 appears to be excessively high.
Better documentation of the basis for AP-42
emission numbers is needed.
The AP-42 evaporative emission rate for
uncontrolled cars errs on the low side as a
result of neglecting fuel tank running losses.
21-35
-------
It is recommended that the regulatory
evaporative test-to-g/mi emission rate
conversion be based on the MIT method
(Reference 5).
Emission factors in AP-42 for controlled
vehicle exhaust emissions reflect extremely
high additive deterioration rates extrapolated
beyond the set of data used to estimate these
rates. It is recommended that some credit
be given to the activity by both manufacturer
and regulator to reduce the probability and
magnitude for high deterioration in future
vehicles.
. Controlled evaporative emissions for the
first generation of emission controls are
estimated to be 0.81 g/mi. The AP-42 appears
to be excessively high (1.76 g/mi).
Trucks of different Gross Vehicle Weights
have been categorized as light, medium and
heavy duty. All of these categories have
emission related characteristics different
than passenger cars. The application by EPA
of passenger car emission deterioration rates
21-36
-------
and other characteristics to these vehicles
is therefore inapproriate and appears to have
resulted in substantial understatement of
current levels of control.
21-37
-------
REFERENCES
1. "Mobile Source Emission Factors," EPA Interim
Document, OTLUP, Washington, D.C., June, 1977.
2. Seventh Annual Report to the President's Council
on Environmental Quality, 1976.
3. R. L. Kramer and H. P. Cernansky, "Motor Vehicle
Emission Rates," U.S. Department of Health,
Education and Welfare, Durham, N.C., August 15,
1970.
4. J.C. Gagliardi and F.E. Ghannam, "Effects of
Tetraethyl Lead Concentration on Exhaust
Emissions in Customer Type Vehicle Operation,"
SAE 690015, January, 1969.
5. John B. Heywood, "Statement to the Subcommittee
on Public Health and Environment," Interstate
and Foreign Commerce Committee, U.S. House of
Representatives, Washington, D.C.
December 4, 1973.
6. D.T. Wade, "Factors Influencing Vehicle
Evaporative Emissions," Society of Automotive
Engineers Paper No. 670126, January, 1967.
7. "Fuel System Evaporative Losses," issued by the
Automobile Manufacturers Association, AMA
Engineering Notes 616, September, 1961.
8. S.W. Martens and E.E. Nelson, "Current Status
of Vehicle Evaporative Emission Control,"
presented at Fourth Annual North American
Motor Vehicle Emissions Control Conference,
Anaheim, California, November 5, 1975.
9. CALSPAN Corporation, "Automobile Exhaust Emission
Surveillance - A Summary," Document No. APTD-1544,
prepared for Environmental Protection Agency,
Office of Air and Water Programs, Office of
Mobile Source Air Pollution Control, Ann Arbor,
Michigan, May, 1973.
21-38
-------
10. "Automobile Exhaust Emission Surveillance
Analysis of the FY 1974 Program," EPA-460/
3-76-019.
11. "Exhaust Emissions from Privately Owned 1966-
1973 California Automobiles, A Statistical
Evaluation of Surveillance Data," Supplement
of Progress Report No. 34, Surveillance of Motor
Vehicle Emissions in California, California Air
Resources Laboratory, El Monte, California,
May 1974.
12. Statement of General Motors Corporation to EPA -
Waiver Hearing on Adjustability of Idle Mixture
Mechanism, San Francisco, California, May 18, 1977,
13. State of California Air Resources Board
"California Exhaust Emission Standards and Test
Procedures for 1980 and Subsequent Model
Passenger Cars, Light Duty Trucks, and Medium
Duty Vehicles" - Adopted November 23, 1976,
amended June 22, 1977.
14. "New Motor Vehicle Certification, Intent to
Develop Rulemaking," EPA, Federal Register,
Volume 42, No. 104, May 31, 1977.
15. "Historical Development of Heavy Duty Gasoline
Engine Dynamometer Emissions Test Cycle and
Emissions Standards," Motor Vehicle Manufacturers
Association, Detroit, Michigan.
16. "Los Angeles Auto Exhaust Test Station Project -
A Joint Agency Report, 1961-1963", Air Pollution
Control District, County of Los Angeles,
California, Automobile Club of Southern
California, Automobile Manufacturers Association,
Detroit, Michigan California Department of
Public Health, California Motor Vehicle Pollution
Control Board, California Highway Patrol, United
States Public Health Service.
17. "Baseline Reactivity Survey," California Air
Resources Board, Los Angeles, California,
February, 1968.
21-39
-------
18. "Surveillance of Motor Vehicle Emissions in
California, Quarterly Progress Report No. 15,
January-March, 1969", California Air Resources
Board, Los Angeles, California.
19. "A Study of Mandatory Engine Maintenance for
Reducing Vehicle Exhaust Emissions," CRC-APRAC
Project CAPE-13-68 and EPA by TRW and Scott
Research Laboratories, July, 1973.
20. D.W. Houser, R.F. Irwin, L.J. Painter, G.H.
Amberg, "Field Tests Show Gasoline Deposit
Control Additives Effective in Emission
Reduction," APCA Meeting June, 1971.
21. The Great Plains Surveillance Program, FY 69.
22. The National Surveillance Program Phase I,
FY 70.
23. The National Surveillance Program Phase II,
FY 70.
24. Marcia E. Williams, John T. White, Lois A.
Platte, Charles J. Domke, "Automotive Exhaust
Emission Surveillance - Analysis of the FY 72
Program," EPA - 460/2-74-001, U.S. Environmental
Protection Agency, Ann Arbor, Michigan,
February, 1974.
25. Jeffrey Bernard, Paul Donovan, H.T. McAdams,
"Automobile Exhaust Emission Surveillance
Analysis of the FY 73 Program," EPA -
460/3-75-007, Prepared by CALSPAN CORPORATION
for the U.S. Environmental Protection Agency,
Ann Arbor, Michigan, July, 1975.
26. H.A. Ashby, R.C. Stahman, B. H. Eccleston,
R.W. Hurn, "Vehicle Emissions - Summer to
Winter," Society of Automotive Engineers
Paper No. 741053, Presented at Toronto,
Ontario, Canada, October, 1974.
21-40
-------
27. Study of Emissions from Light-Duty Vehicles
in Seven Cities, FY 75 (unpublished).
28. P.J. Clarke, "Investigation and Assessment
of Light Duty Vehicle Evaporative Emission
Sources and Control," Document No. EPA-
460/3-76-014, prepared for Environmental
Protection Agency, Office of Air and Waste
Management, Office of Mobile Source Air Pollution
Control, Ann Arbor, Michigan, June, 1976.
29. T.M. Fisher, Letter to Eric O. Stork of the U.S.
Environmental Protection Agency, General Motors
Corporation Warren, Michigan, May 14, 1976.
30. "Surveillance of Evaporative Emissions from 1970
to 1976 Year Model Vehicles by the SHED Method,"
Project E. Progress Report No. 5, Air Resources
Board, 9528 Telstar Avenue, El Monte, California,
March, 1976.
31. California Department of Public Health, Progress
Report - Motor Vehicle Emissions and Proposed
Standards, January 8, 1964.
32. Thomas A. Huls, "Evolution of Federal Light
Duty Mass Emission Regulations," Society of
Automotive Engineers paper 730554, presented
at the Automobile Engineering Meeting, Detroit,
Michigan, May 1973.
33. S.H. Mick and J.B. Clark, Jr., "Weighing
Automotive Exhaust Emissions," Society of
Automotive Engineers paper 690523, May 1969.
34. "California Test Procedure and Criteria for
Motor Vehicle Exhaust Emission Control,"
State of California, Motor Vehicle Pollution
Control Board, January 23, 1964.
35. "1975 Emission Standards for Hydrocarbons and
Carbon Monoxide Applicable to Light Duty
Vehicles," Notice of Proposed Rulemaking,
U.S. EPA, Federal Register, Vol. 36, No. 39,
February 26, 1971.
21-41
-------
36. "Exhaust Emission Standards and Test Procedures",
U.S. EPA, Federal Register, Vol. 36, No. 128,
July 2, 1972.
37- A.J. Hooker, "Summary of Oxides of Nitrogen
Test Methods and Test Data," California Air
Resources Board, Los Angeles, California,
May 26, 1969.
21-42
-------
TABLE I
Exhaust Emission Data on 1965 and Older Vehicles
Number of Vehicles
Source
1962 California
Program
Baseline
Reactivity
Survey
CARS Study
Rose Bowl Study
Year
Completed
1962
1968
1969
1965
and Older
1000
212
12*
Driving
Cycle
Hot 8-
Mode
Hot 7-
Mode
Hot 7-
Node
Reference
Dumber
16
17
18
1971
EPA FY 71
Program by AESI** 1973
CRC - Maintenance
Study
161
337
1973
7-Mode
1975 FTP
1972 FTP
20
19
* Includes HC and CO data only for a substantially
larger number of 1965 cars.
** FY 71 program included uncontrolled 1966/67 49 state cars.
21-43
-------
TABLE 2
Model
Year
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
otive
HC
16.6
15.9
15.3
14.7
14.1
13.6
13.1
12.6
12.2
11.8
11.4
11.1
10.8
10.6
10.4
10.2
10,1
10.0
9.9
Exhaust Emissions
Mean Line
Exhaust Emissions g/mL
(1975 FTP)
CO
108
105
102
100
97
95
93
91
90
88
87
86
85
84
84
83
83
83
84
NOV
i
2.8
2.9
3.1
3.3
3.4
3.5
3.7
3.8
3.9
3.9
4.0
4.1
4.1
4.1
4.1
4.1
4.1
4.1
4.1
21-44
-------
TABLE 3
In-Use Exhaust Emission Data on Controlled Vehicles
Source
CRC Vehicle
Emission
Tests
EPA FY 69
EPA FY 70-1
EPA FY 70-11
EPA FY 71
CRC Main-
tenance
Study
EPA FY 72
EPA FY 73
EPA FY 74
EPA FY 75
Model Number of Driving Evap.
Years Vehicles Cycle Emissions
1967-1968
1968-1969
1970
1971
1957-1971
1960-1971
1966-1972
1967-1974
1967-1975
1965-1975
1968-1976
188
2029
2101
369
1020
486
1020
1080
26
1968
2220
7-mode
7-mode No
7-mode No
7-mode No
1972/1975 FTP Yes
1972 FTP,
loaded mode-
idle
1972/1975 FTP Yes
1972/1975 FTP, Yes
short tests,
and modal
1975 FTP No
1974 FTP, HWFET, No
short tests, and
modal
1975 FTP, HWFET,
short tests, and
modal
Reference
Number
*
21
22
23
8
18
24
25
26
9
27
* Unpublished data generated in the program summarized
in Reference 19
21-45
-------
TABLE 4
Evaporative Emissions from In-Use Cars
Measured by the SHED Techniques
Model
Laboratory Year
Exxon 1973-75
GM 1970-72
CARB 1970-76
AESI(PY 71) 1970-71
Number
of
Cars
20
20
109
31
Mean
Emission
Total
g/test
8.99*
8.7
6.9
27.2
Diurnal
Soak
9
1.6*
3.37
1.5
16.28
Hot
Soak
g Reference
7.9* 28
5.36 29
5.4 30
10.92 8
* Exxon data uncorrected for non-fuel background emissions,
modal average based on 18 of the 20 test care.
21-46
-------
FIGURE I
UNCONTROLLED AUTOMOTIVE HC
ro
Y->
IS
HC
g/ml by
1975 FTP
10
0
ToI Iplp* Dolo
MEAN LINE
VFY71 Dot*
i
1945
1950
I8S5 I960
MODEL YEAR
tees
1970
-------
FIGURE 2
UNCONTROLLED AUTOMOTIVE CO
N>
e*
s
126
100
CO
Q/nl by 75
1875 FTP
S0
25
MEAN UZNB
liiiitii iJt it tit ill 1 I t i . i _. I
104S I960 1866 IO60 1966
HOOEL YEAR
IQ7«
-------
FIGURE 3
UNCONTROLLED AUTOMOTIVE NOx
NJ
h-
4N
VO
NOx
8/ml by
1975 FTP
e
**
/ Tollplp* Doto
Doko
1
I I I I I I I I I I I I I I I I I I I i lilt
1948
I860
1856 1868
MODEL YEAR
1866
1878
-------
HC
3
2
f irmnni
FIGURE 4
LOV EMISSION KATES
MODEL YEAR 1975
EXHAUST HYDROCARBONS
UiwarFtt
0f FY74 Data
I I I I I I II II
8.8 8.1 8,2 9.3 8.4 8.5 8.6 8.7 8,8 8.9 1.8
MILES/10s
21-50
-------
80
60
CO
/ml
40
20 H
0
FIGURE 5
LDV EMISSION RATES
MODEL YEAR 1975
CARBON MONOXIDE
Ufwar Fit
X DMlonatw toimlnal mlUag* of FY74 Data
T T T
0.0 0.1 0.2 0.3
I I I II I I
B..4 0.5 0.6 0.7 0.8 0.9 1.0
MILES/105
21-51
-------
3
NOv
a/ml
2
FIGURE 6
LDV EMISSION RATES
MODEL YEAR 1975
OXIDES OF NITROGEN
Emiulon Focton
Logarithmic Fit
X DMignatw Terminal MiUaa* of FY74 Data
I I I I I 1 I I I I
0.8 e.l 0.2 8.3 0.4 0.5 0.6 8.7 8.8 8.8 1.8
MILES/105
21-52
-------
HC
g/mi
FIGURE 7
Automotive Exhaust HC Emission Levels
Fits of FY 74 Data
1974 Models
Logarithmic Model
I I I I I I I I I
e.e e.i 9.2 0.3 e.4 e.s e.o 8.7 e.e a.9 t.e
HC74/HC74A
21-53
-------
FIGURE 8
Automotive CO Emission Levels
Fits of FY 74 Data
1974 Models
681
48
CO
g/ai
28-
8-
8.8
I
8.1
8.2 8.3
I
8.4
I
8.5
Miles/(8S
I I I I I
1.6 8.7 8.8 8.8 1.8
C074/C074A
21-54
-------
FIGURE 9
Automotive NOX Emission Levels
Fits of FY 74 Data
1974 Models
5
3-
1
I I I I I I I I I I
0.0 0.1 8.2 0.3 0.4 0.5 0.0 0.7 0.8 0.9 1.0
Miles/18~
NOX74/NOX74A
21-55
-------
41
3
2
(Jl
FIGURE 10
Automotive Exhauat HC Emission Levels
fit* of r* 74 Bata
It?3 Model*
U)q«rithrai-c Model
Una«r
i i i i t i i i r I
8.8 0-t 8,2 8.3 8.4 8,5 8.8 8.7 8.8 8.9 1.8
»6
X18*
HC73/HC73A
-------
§
M
Bh
II
So
* i
£83 '
tan a
M T>
24
22.
20
18
16
U
12
10
8
6
4
2
0
MC
1963 - 69 Unoootrolled
1969 Calif. 1970 - 71 Fed
HC std - 275
1972 - 73 HC St
-------
APPENDIX
Uncontrolled Car
Exhaust Emission Data
and
Their Adjustment
21-58
-------
Huls formally published in 1973 (Reference 32}
the EPA development of uncontrolled car baseline
exhaust HC and CO values for the 1972 Federal Test
Procedure (FTP). The 1972 FTP and its successor,
the 1975 FTP, obtain a direct mass rate of emissions
by diluting the vehicle exhaust stream (with air)
to a constant fixed flow rate and measuring the
pollutant concentrations in that streamhence the
term Constant Volume Sampler (CVS); see for example
Reference 33. Huls1 values, 16.8 g/mi HC and 126
g/mi CO were obtained by correcting a large bank of
7-mode (Reference 34) and similar tailpipe test data
to the CVS procedure via empirical factors
developed from tests of 30 uncontrolled cars. These
values further adjusted to 15 g/mi HC and 90 g/mi
CO when a correction was made for the "cold/hot
weighted" 1975 FTPReferences 35 and 36 list
"statutory" HC/CO standards of .46/4.7 and .41/3.4
based on the 1972 FTP and the 1975 FTP respectively.
The HC and CO 1975 FTP/1972 FTP corrections are
therefore .891 and .723.
At about the same time as the Huls' publication,
EPA released results of a field survey in which
21-59
-------
customer cars were actually directly tested using the
CVS method and included a substantial number of
uncontrolled cars (Reference 9). Data from those
uncontrolled cars averaged 8.7 g/mi exhaust HC and 87
g/mi CO over the 1975 FTP. This latter exhaust HC
level represents a very large discrepancy compared
to the above 15 g/mi value. We are aware of no
previous attempt to rationalize the difference between
these two sets of "baseline" values.
It seems apparent to us from the plot of the
two bodies of data in Figure 1 that the major
difference in the two "baselines" is the model year
make up of the two car populations involved.
Although to our knowledge EPA has never formally
published the sources of the data body which made up
the 7-mode cycle baseline values, we believe that
the HC data we have assembled from the literature
represents essentially all that is available and
therefore must be roughly equivalent to the data body
used by Huls.
Note in Figure 1 that the so-called tailpipe
data (primarily 7~mode cycle data) have been
corrected to the 1975 FTP basis. Table A-l details
the actual tailpipe data from the various surveys
21-60
-------
together with the corresponding 1975 FTP value.
In addition, it lists the weighted average of the
latter values which were used to form the "tailpipe
data" plot of Figure 1.
Huls showed that the average car over the 7-mode
cycle, at the 2.2 g/mi HC standard of 1970, would have
a composite concentration value of 180 ppm. Conse-
quently, the correction factor from composite tailpipe
concentration values to 1970 FTP g/mi values is .0122.
This value is then adjusted to the 1972 FTP by a
multiplying correction factor according to Huls of
1.38, and in turn by a factor of 0.891 as indicated
above to correct to the 1975 FTP. It is shown in
Reference 31 that the average car tested on the cold
start 7-mode cycle (1970 FTP) produced HC values
1.101 times that of the hot7-mode cycle. All of the
concentration data as indicated on Table A-l were
based on test from a hot start because of the
practical problems of testing very large numbers of
cars. In addition, Reference 30 details the basis
for an adjustment factor of 1.212 to the 8-mode cycle
hot start HC concentrations of the Los Angeles Test
Station Project (Reference 16). Thus, the latter
data have been corrected to the 1975 FTP basis through
21-61
-------
multiplication by correction factors of 1.212
x .0122 x 1.38 x .891. The Baseline Reactivity
survey (Reference 17) and the 194 Car Survey data
were corrected to 1975 FTP by a multiplicitive
correction factor of 1.101 x .0122 x 1.38 x .891.
Finally the Rose Bowl survey performed on the 1970
FTP basis was corrected simply by multiplication
of 1.101 x 1.38 x .891.
While it is obvious from Figure 1 that the
direct 1975 FTP measurements of the FY 1971 program
are similar to the older tailpipe data for the same
model years, there does remain a small discrepancy.
Since we have no basis to choose one set of values
over the other, we have assumed in Table A-2, whose
values generated the mean line of Figure 1, that the
true value for the average emission rate of the
"overlapping" model years is midway between the
corrected tailpipe values and the FY 71 values. To
be specific, the corrected data of Table A-2 were
obtained by averaging the 1957 through 1965 values
from both data bodies, and adjusting the tailpipe
data downward by half the difference and the FY 71
data upward by half the difference. The mean line
of Figure 1 then is a least-^s^uaresr fit ef a second
21-62
-------
degree polynomial to that corrected data.
The carbon monoxide tailpipe data of Table A-3
were similarly obtained from the correction factors of
Huls*~-23.0 from tailpipe concentrations to 1970 FTP
g/rai and 1.58 from 1970 FTP to 1972 FTP and from
References 35 and 36 the adjustment from 1972 FTP
to 1975 FTP for carbon monoxide is .723. Reference
31 indicates that the tailpipe carbon monoxide
values for uncontrolled cars from hot start compared
with cold start were not significantly different
and consequently no hot start to cold corrections
factor is required. Similarly, Reference 31
further states that the test cycle and other
corrections made to the Los Angeles Test Station
HC data are unnecessary for carbon monoxide.
Consequently, the tailpipe concentration CO data
for the Los Angeles Test Station Project, the Baseline
Reactivity Survey and the 194 Car Survey have been
corrected to the 1975 FTP by multiplication of 23.0
x 1.58 x .723. The Rose Bowl Survey data, on the
other hand, were simply corrected by adjustment from
the 1970 FTP to the 1975 FTP, i.e., by a multiplication
of 1.58 x .723.
21-63
-------
Analogous to Table A-2, the data detailed in
Table A-4 show the adjustment made to the tailpipe
CO data and the FY 71 1975 FTP data to a common
1957-1965 average value. The latter finally resulted
in the mean line of Figure 2 which is, similar to the
HC mean line of Figure 1, a least-squares second
degree polynomial fit of the Table A-4 data.
Published documentation for conversion of
tailpipe concentration NOx data from one test type
to another and from the tailpipe to the CVS basis
is less complete than the foregoing conversion
procedures for exhaust HC and CO. As far as we
are aware, there has been no publication of back-
to-back 1970 FTP measurements versus 1975 FTP
measurements for NOxof the type described by Huls
for HC and CO. However, the conversions used for
Table A-5 require such relationship.
An experimental program was conducted by
General Motors using 27 pre-1968 cars. That test
program, which involves back-to-back tests on the
1970 FTP and the 1972 (CVS) FTP, is summarized in
Table A~7. The indicated weighting of those
"equivalence ratios" for a realistic distribution
of the cars in the population of the early 1960's
21-64
-------
(by manufacturer) is indicated to produce a correction
factor of 1.082. As indicated on Table A-7 References
35 and 36 indicate that this conversion factor should
be in turn multiplied by the ratio 3.1/3.0 to result
in the final correction factor of 1.118 to obtain
1975 FTP values from 1970 FTP data.
Analogous to the previously described relationship
developed by Huls for converting tailpipe concentration
values to 1970 FTP g/mile values for HC and CO, the
NOx tailpipe concentration from the average car should
be multiplied by the factor .0037 to obtain the 1970
FTP value. Reference 31 indicates that there is not
an appropriate correction factor for converting the
8-mode cycle tailpipe concentration values of the
Test Station Project, nor the hot start test to the
1970 FTP basis. Consequently, these values have been
assumed to be related to 1970 FTP values by a factor
of 1.
The Test Station Project tailpipe NOx concentration
values have consequently been multiplied by an overall
correction factor of .0037 x 1.118 x 3.1/3.0 to
convert to the 1975 FTP g/mile basis. It was necessary,
in addition, to divide the resulting values for the
Baseline Reactivity Survey data by 1.5 since those NOx
21-65
-------
data were obtained by "bagging" the entire exhaust of
the 7-mode cycle which includes some normally unsampled
heavy load operation (Reference 37). The Rose Bowl
survey, on the other hand, which is originally given
in 1970 FTP g/mile, has simply been multiplied by
1.118 x 3.1/3.0.
As a final note, it is well to acknowledge that
the tailpipe data used in these analyses does not
include the "GARB Study" (Reference 18) included in
Table I. As indicated in Table I, only 12 of the 1965
test cars involved NOx measurements. Furthermore,
»-
the only other model years accounted for by that
"study" were 1966 and 1967. Exhaust controls were
required for both those model years in California.
We understand that the 1966/67 cars considered
"uncontrolled" in that program were "imports" from
outside the state. Nonetheless, their inclusion
raise questions for which satisfactory documentation
is unavailable. Consideration of those questions,
and the fact that a single model year from one test
program could unduly influence the "shape" of the
final emissions--vs.model year relationship, led us
to reject the "CARS Study" data for this analysis.
21-66
-------
to
H*
Table A-l
Exhaust HC Tailpipe Measurements from Uncontrolled Cars
Model
Year
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
Test
Los Angeles ,*
Station Project16
Reactivity Survey
Cone.. 1975 FTP,
N
4
6
7
23
25
16
39
50
89
100
112
72
109
109
108
151
13
»
.
PUB
862
1148
918
896
932
900
832
805
780
732
629
625
657
646
731
635
563
»
.
g/ml
15.7
20.9
16.8
16.1
16.9
16.5
15.2
14.7
14.2
13.4
11.4
11.3
11.9
11.8
13.4
11.5
10.3
.
-
N
.
»
1
..
»
2
2
1
1
6
5
6
9
16
21
51
45
46
1974 Car Survey31
Cone., 1975 FTP,
ppm
M
795
.
.
M
987
841
617
509
834
912
635
1987
942
711
628
937
763
3/ml _
.
13.1
_
.
.
16.3
13.9
10.2
8.4
13.8
15.1
10.5
32.8
15.6
11.7
10.4
15.5
12.6
N
.
2
5
8
13
9
12
13
26
22
15
n
17
16
19
.
.
-
Cone., 1975 FTP.
ppm
957
854
1028
951
1083
914
900
774
1008
1005
732
718
757
751
.
.
-
g/ml
15.8
14.1
17.0
15.7
17.9
15.1
14.9
12.8
16.7
16.6
12.1
11.9
12.5
12.4
_
.
-
Rose BOM! Survey
1970 FTP,
N
.
5
3
12
14
24
30
38
35
o/m1
.
7.00
6.52
6.94
5.93
6.06
8.15
7.14
7.21
1975 FTP,
a/ml
-
9.5
8.8
9.4
8.0
8.2
11.0
9.7
9.7
Weighted
Average
1975 FTP,
g/m1
15.7
. 19.6
15.5
16.3
16.5
17.0
15.2
14.7
13.9
13.9
12.1
11.5
11.6
13.0
13.0
11.1
10.6
12.8
11.3
Notes o Superscripts on each survey title Indicate associated reference number.
ON- Number of cars tested.
o LA Test Station data were obtained using an 8-mode cyclefrom hot start.
o Reactivity Survey data were obtained using the 7-mode cyclefrom hot start.
o 194 Car Survey data Mere obtained via weighted 7-mode/11 -mode cyclefrom hot start. (Listed
reference Includes only summary data, details obtained from raw data.)
o Rose Bowl Survey data were obtained using 1970 FTPexcept tests run from hot start.
-------
Table A-2
Exhaust HC Measurements from Uncontrolled Cars
Adjusted Data
All data are 1975 FTP g/ra1
Data Corrected* to
Htd. Avg. FY 71 Common 1957-1965 Mean
Model Tailpipe Data CVS Data Tailpipe
Year From Table A-l From Ref. 9 Data
1947 15.7 14.2
1948 19.6 18.1
1949 15.5 14.0
1950 16.3 14.8
1951 16.5 15.0
1952 17.0 15.5
1953 15.2 13.7
1954 14.7 13.2
1955 13.9 12.4
1956 13.9 12.4
1957 12.1 6.63 10.6 8.2
1958 11.5 10.04 10.0 11.6
195-9 11.8 10.80 10.3 12.4
1960 13.0 8.80 11.5 10.4
1961 13.0 5.94 11.5 7.5
1962 11.1 8.88 9.6 10.4
1963 10.6 9.44 9.1 11.0
1964 12.8 7.28 11.3 8.8
1965 11.3 11.18 9.8 12.7
1966 - 8.26 - 9.8
1967 - 7.38 - 8.9
Mean 1957-1965 11.9 8.8
*Correction Factor (11.9-8.8)/2; subtracted from tailpipe data, added to
FY 71 data.
"Excludes Denver Data and 1966/67 Los Angeles Data.
21-68
-------
ro
h-
0\
vo
Table A-3
CO Tailpipe Measurements from Uncontrolled Cars
Model
Year
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
Test
N
4
6
7
23
25
16
39
50
89
100
112
72
109
109
108
151
13
_
.
Los Angeles IK
Station Project16
Cone.,
%
3.68
3.55
4.36
3.22
3.78
4.40
3.95
4.26
3.44
3.63
3.54
3.46
3.06
2.75
3.43
3.23
3.19
.
-
1975 FTP,
g/ml
96.7
93.3
114.6
84.6
99.4
115.7
103.8
112.0
90.4
95.4
93.1
91.0
80.4
72.3
90.2
84.9
83.1
.
-
Reactivity
N
.
.
1
.
.
.
2
2
1
1
6
5
6
9
16
21
51
45
46
Cone. ,
X
.
.
6.62
.
.
.
3.49
2.86
2.27
1.55
3.75
2.23
2.87
3.98
2.92
2.64
2.47
3.30
2.89
Survey17
1975 FTP,
g/ml
.
.
173.0
.
.
.
91.5
75.0
59.4
40.6
98.1
58.4
75.1
104.0
76.4
69.1
64.6
86.3
75.6
1974 Car Survey31
N
2
5
8
13
9
12
13
26
22
15
11
17
16
19
.
.
.
.
Cone . ,
X
.
3.25
4.27
3.46
3.89
4.17
2.88
3.42
3.57
2.99
2.90
2.60
2.81
2.66
2.81
.
_
.
.
1975 FTP.
g/m1
m
85.4
112.3
91.0
102.3
109.6
75.7
89.9
93.9
78.6
76.2
68.4
73.9
69.9
73.9
_
_
.
.
20
Rose Bowl Survey
N
.
.
.
.
.
.
.
.
_
.
5
3
12
14
24
30
38
35
1970 FTP,
a/ml
m
_
.
.
.
.
.
.
_
.
88.15
33.56
52.71
71.38
49.57
75.51
70.13
69.97
1975 FTP,
a/Hi
.
.
-
.
-
-
.
-
.
.
.
100.8
38.4
60.2
81.6
56.7
86.3
80.2
80.0
Weighted
Average
M975 FTP,
g/ml
96.7
91.3
118.2
86.3
100.4
113.5
97.0
106.4
90.9
91.9
91.4
87.1
78.4
73.0
86.1
79.8
74.1
83.5
77.5
See Notes on Table A-l
-------
Table A-4
CO Measurements from Uncontrolled Cars
Adjusted Data
All data are 1975 FTP g/m1
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
Mtd. Avg.
Tailpipe Data
From Table A-3
96.7
91.3
118.2
86.
100.
113.5
97.0
106.
90.
91.
91.
.3
.4
.4
.9
.9
.4
87.1
78.4
73.0
86.1
79.8
74.1
83.5
77.5
FY 71
CVS Data
From Ref. 9
81.4
78.2
77.
81,
79.
78.0
96.
81.
87.
91.
93.
Data Corrected* to
Common 1957-1965 Mean
Tailpipe
Data
97.4
92.0
118.5
87.0
101.1
114.2
97.7
107.1
91.6
92.6
92.1
87.8
79.1
73.7
86.8
80.5
74.8
84.2
78.2
FY 71
Data**
_
-
_
_
-
-
-
-
-
-
80.8
77.5
77.6
81.0
79.1
77.3
95.8
81.1
87.3
90.4
93.0
Mean 1957-1965 81.2
82.5
^Correction Factor (82.5-81.2)/2 * 0.65; added to tailpipe data,
subtracted from FY 71 data.
"Excludes Denver Data and 1966/67 Los Angeles Data.
21-70
-------
Table A-5
NOx Tailpipe Measurements from Uncontrolled Cars
Los Angeles ^ 17 «- Weighted
Rose Bowl Survey Average
1970 FTP, 1970 FTP, 1975 FTP,
_N g/m1 g/m1 g/nrl
3.40
3.53
2.83
4.38
4.18
3.68
4.13
3.28
4.60
4.34
4.41
5 4.75 5.31 4.17
3 5.31 5.94 4.79
12 4.87 5.44 4.95
14 3.91 4.37 4.04
22 5.64 6.31 4.63
30 4.49 5.02 4.93
38 4.39 4.91 4.56
35 4.29 4.80 4.85
See Notes on Table A-l
Test Station Project10
Model
Year
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
N
4
6
7
20
24
11
31
39
77
85
89
- 47
76
78
70
122
10
_
-
Cone . ,
ppm
802
831
738
1032
985
865
959
742
1079
1019
1045
947
1134
1194
'900
997
1033
-
-
1975 FTP,
g/m1
3.40
3.53
3.13
4.38
4.18
3.68
4.07
3.15
4.58
4.33
4.44
4.02
4.82
5.08
3.82
4.24
4.39
-
-
Reactivity Survey
N
-
1
-
-
.
2
2
1
1
6
5
6
9
16
21
51
45
46
Cone . ,
ppm
-
266
.
-
_
1774
2019
2270
1696
1413
1558
1327
1106
1665
1802
1754
1504
1722
1975 FTP,
g/ml
-
0.76
.
-
-
5.04
5.73
6.45
4.82
4.01
4.42
3.77
3.14
4.73
5.12
4.98
4.27
4.89
-------
Table A-6
NOx Measurements from Uncontrolled Cars
Adjusted Data
All data are 1975 FTP g/m1
Model
Year
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
Htd. Avg.
Tailpipe Data
From Table A-S
3.40
3.53
2.83
4.38
4.18
3.68
4.13
3.28
4.60
4.34
4.41
4.17
79
95
04
63
93
56
4.85
FY 71
CVS Data
From Ref. 9
Data Corrected* 'to
Common 1957-1965 Mean
Mean 1957-1965 4.59
3.84
3.62
4.50
3.94
3.07
3.33
3.64
3.67
3.37
3.57
3.28
3.66
Tailpipe
Data
2.94
3.07
2.37
3.92
3.72
3.22
3.67
2.82
4.14
3.88
3.95
3.71
4.33
4.49
3.58
4.17
4.47
4.10
4.39
FYTI
Data**
4.31
4.09
4.97
4.41
3.54
3.80
4.11
4.14
3.84
4.04
3.75
Correction Factor (4.59-3.66)72 - .465; subtracted from tailpipe data.
added to FY 71 data.
"Excludes Denver Data and 1966/67 Los Angeles Data.
21-72
-------
Table A-7
NOx Test-to-Test Conversions for Uncontrolled (pre 1968) Cars
General Motors Test Program Results
Car Number 1970 FTP Average Equivalence Ratio
Manufacturer of Cars g/m1 1972 FTP/1970 FTP
GM 15 4.06 1.20
Ford 5 3.98 0.98
Chrysler 4 4.24 0.97
AMC 1 2.96 1.08
Foreign 2 1.59 0.93
Mean 3.86 1.10
Overall average Equivalence Ratio should be weighted by car population.
Reference 31 provide car population weighting factors (for the early
1960's) of:
GM .410
Ford .281
Chrysler .160
Other .088
Applying these weighting factors to the above GM results gives an over-
all equivalence ratio (1972 FTP/1970 FTP) of 1.082. Correcting this
ratio to the 1975 FTP via the ratio of the 1975 to the 1974 standard
gives: 1.082 (3.1/3.0) = 1.118 1975 FTP/1970 FTP
21-73
-------
The following written comments on this paper were received
from Marcia Williams, EPA, Office of Motor Source Air Pollution
Control, Ann Arbor, Michigan and are included for information.
This constitutes the only written response received to the papers
presented at the workshop.
21-74
-------
Comments on "Highway Motor Vehicle Emission Factors"
paper by MVMA, presented at EPA Emission Inventory/
Factor Workshop (Sept, 1977)
The following comments are submitted after a careful review of the MVMA
paper entitled "Highway Motor Vehicle Emission Factors". The MVMA paper
discussed the recently released draft version of the EPA Revised Emission
Factor Document. That draft is currently undergoing changes to reflect
the recent 1977 Clean Air Act Amendments, some new data, and new information
on California standards. The comments reflected here will pertain to
the draft document used by MVMA, not the revised version of that document.
MVMA has commented that the EPA document is still deficient on detailed
discussions regarding the experimental data base and computational
methods. This information has been carefully compiled and will be
released in a separate document concurrently with the final Emission
Factor Document.
Introduction
1. EPA has not concluded that CO emissions are being overestimated in
the current test procedure as is implied in the MVMA paper. EPA is
currently performing a study to characterize the types of driving
conditions which occur in CO hot spot areas. These results will be
compared with the driving patterns and test conditions (vehicle
temperature, ambient conditions) which are currently reflected in
the FTP. If necessary, the FTP will be modified to ensure that
emission reductions achieved over the test procedure will equate to
emission reductions on the road in CO hot spots. The study will
also quantify the fraction of the CO problem which is due to local
sources and that which is due to regional sources since typical
local/regional driving conditions are not necessarily the same.
2. The MVMA paper discusses the sensitivity of ambient air quality
measurements to the location of ambient monitors. The discrete
nature of the monitoring system makes it impossible to properly
quantify the air at every point. Likewise, since the main CO
standard of concern is an 8 hour average standard, one ideally
wants to average the air quality levels that a person can be expected
to be exposed to under typical worst case conditions. Such conditions
are likely to be those that affect policeman, airport porters, road
work crews, etc. These people are breathing air at curb level and
at approximately 8 feet, not at 18 feet and at the building face.
Thus, the CO levels measured in NYC may be more realistic of typical
-worst case conditions that those measured elsewhere. The MVMA
statement that "It would be extremely unfortunate if transportation
control plans were established from measurements based on the
shortest distance between sampling probe and tailpipe without
considering public health aspects" does not appear applicable to
the NYC situation as implied in the paper.
21-75
-------
3. The MVMA paper assumes that since it is desirable to have a bench-
mark emission rate, it is necessary to select a model year definition
for the desired benchmark. MVMA has chosen 1960 but has not specified
the age or mileage of the vehicles at the time their emissions were
computed. If one infers that the emission levels are measured in
1970, there is no reason for using these numbers as a benchmark.
For baseline and percent reduction work, one wants to compare pre-
controlled vehicles at a similar age to controlled vehicles.
EPA has determined that the exhaust emissions of pre-controlled
vehicles deteriorate with mileage. The appropriate benchmark is
considered to be the 50,000 mile point. Since 50,000 miles is
defined as the useful life, vehicles are expected to just meet
exhaust emission standards at the 50K mileage point. Thus, the 50K
exhaust emission levels of pre-controlled vehicles (of all model
years.) can be compared with later emission standards or 50K in-use
emission levels to determine the appropriate emission reductions.
If desired, the evaporative and crankcase emissions, which are not
assumed to deteriorate with mileage, can be added to the exhaust
emissions before reductions are compiled.
Uncontrolled Exhaust Emissions
4. The recent version of the Emission Factor Document contained data
on approximately 1000 precontrolled vehicles, all of which were
tested over the 1975 FTP. Table I in the MVMA paper contained only
a fraction of the total EPA data base. EPA considers the use of
the MVMA concentration data base to be highly questionable. The
ability to accurately convert concentration data to mass data is
difficult if not impossible. (The MVMA Appendix to the paper
supports this fact.) Both EPA and GARB have stated that since such
conversions are extremely vehicle dependent, they should not be
performed on in-use emission factor data bases.
The incorrect use of a single conversion factor over many model
years and vehicle mixes could account for the downward trend in
emissions shown in Figures 1, 2, and 3. The other factor which
could account for an artificial downward trend is the testing of
different model year vehicles at different mileage/age points. EPA
FTP data do not support a downward emission trend from 1957 to 1967
although within each model year, a trend with mileage/age is apparent.
Uncontrolled Evaporative Emissions
5. MVMA suggests that the MIT method be used to convert from grams/test
to grams/mile. The EPA procedure is based on the most recent
national average trip statistics while the origin of the MIT
statistics are unclear (the 7.5 miles/trip figure is based on over
10 year old data from Los Angeles; the basis for the 27 mile/day
figure is not known).
21-76
-------
The inclusion of running losses is not possible at this time due to
the absence of a defendable procedure over which to measure the
losses. In order for vehicle operation to cause evaporative
emissions in excess of those produced during the diurnal phase, the
vehicle operation must result in a temperature rise of the fuel
tank which is greater than that which occurs during the diurnal.
At this time, the extent of these losses in real world operation
have not been adequately quantified in order to include them in the
evaporative emission factor.
Controlled Exhaust Emissions
6. Post-1974 vehicles - The EPA regression equations for 1975 model
year vehicles were performed by dividing all vehicles into two
distinct groups: within specification vehicles and out of specification
vehicles. Regressions were performed on each group and the results
were weighted to get a final deterioration equation. It is this
difference in analysis methods that accounts for the difference in
the MVMA and EPA linear regression estimates. The correlation
coefficients for the EPA data are higher than the MVMA values.
Moreover, data from the recently completed FY75 Emission Factor
Program support the earlier regression equations and linearity
assumption. The logarithmic model used by MVMA has been discredited
for pre-1975 models as more data have been collected; GARB no
longer uses the model.
While an increase of CO normally is associated with a decrease in
NOx for a given vehicle, the correlation is not excessively strong
for newer model vehicles. More importantly, as vehicles accumulate
age, the types of maladjustments that cause high CO (and lower NOx)
can be totally offset by concurrent maladjustments to the EGR
system.
EPA estimates of future model year emission levels are based on a
technology evaluation model. The model projects the percentage of
vehicles that are properly tuned, maladjusted, and have emission
control systems inoperable as a function of vehicle age. Using
manufacturer supplied data on the emission behavior of future
control systems under various malfunction conditions, the composite
emission rates as a function of mileage can be computed. These
regressions assume that some type of EPA regulation regarding
limited vehicle parametric adjustability will be in effect by the
1980 model year.
7. 1968-1974 Cars - The straight linear regression approach used by
MVMA is not an optimal regression approach due to the large scatter
in the data. The EPA has applied a regression analysis using mean
emissions and mean mileage inputs for each model year vehicle group
in each test program. The result of this type of analysis is to
21-77
-------
predict mean emissions of groups of vehicles (not individual vehicles)
as a function of mileage/age. The correlation coefficients for a
linear model are excellent (greater than .8) for a range of data
that goes from low mileage to approximately 80,000 miles. The
regression equations were used to predict the results in the most
recent test program and predictions were correct within 10 percent.
Controlled Evaporative Emissions
8. The EPA emission factor data are based on three separate EPA test
programs. All three programs gave similar results. The possibility
of gas tank cap leaks was carefully checked and EPA is convinced
that significant leakage did not occur.
Regarding the test data presented in Table 4 (referred to in the
text as Table 5), the vehicles in the Exxon Study were fitted with
new fuel tanks prior to the test so that the vehicles would not be
typical of in-use vehicles.
Truck Exhaust Emission Factors
9. Although the heavy-duty trucks are not subject to a chassis dynamometer
(grams/mile) certification test, EPA has collected emission data
from these vehicles over chassis dynamometer transient tests and
over a real road route in Texas.
10. Figure 11 is not clearly labeled but there are believed to be two
discrepancies in the MVMA comparisons; pre-controlled vehicles had
much higher mileages at time of test than 1972-73 models (emission
deterioration was not adjusted for); 1972-73 trucks were tested in
a tuned condition and pre-controlled trucks were tested in an as-
received condition.
21-78
-------
CONDENSED DISCUSSION
COMMENT:
FAGLEY:
I have one comment on your procedure you
recommend for evaporative loss emission
factors. The intent that we (EPA) had in
putting that equation in there is that
area specific data be used. In other words
your origin-destination data for a number
of trips per day per vehicle and average
trip lengths - you put that into that
equation. As I see what you are doing, you
are putting in a different set of default
values, and it was not our intent. At least
in Supplement 5, that equation to recommend
default values was only to make it so that
people could put in individual urban data.
The point we were trying to make in the
paper is that the numbers for evaporative
emissions come out of the test cycle.
That has a mileage associated with it. You
know that might not be appropriate to par-
ticular urban areas being addressed, but
that was the basis the actual measure of
21-79
-------
emissions. So they made the hot soak
«
emissions associate with 7.5 miles. Not some
other value like ten or something elese.
COMMENT: Again, I think that you could use any default
value. You chose one a little bit lower
than we did. The difference is not very
significant. I do think we should look into
running losses. But that is just a personal
observation.
21-80
-------
FTP Emission Factor Development:
Correction for non-FTP Conditions
Presented at the 1977
Environmental Protection Agency
Emission Inventory/Factor Workshop
Raleigh, North Carolina - September 13-15, 1977
By
Janet Becker
Mathematical Statistician
Marcia Williams
Chief
U.S. Environmental Protection Agency
Office of Mobile Source Air Pollution Control
Emission Control Technology Division
Characterization and Applications Branch
22-1
-------
I. Introduction
The problem of estimating exhaust emission levels from in-use
light duty vehicles has been of interest for many years,
dating back to the early 1950's in California. However, this
problem did not receive a federal mandate for action until the
passage of the Clean Air Act of 1963 and the Clean Air Act
Amendments of 1970, which called for a 90% reduction in HC and
CO from allowable 1970 levels, and for a 90% reduction in NOx
from average NOx emission levels actually measured from light
duty vehicles manufactured during model year 1971. These
reductions were originally to be achieved by model year 1976,
but a series of one year extensions and the 1977 Clean Air Act
has revised final implementation dates until at least 1981.
The federal government has implemented programs designed to
ensure that properly maintained and used vehicles meet emission
standards in the field. These programs include certification,
assembly line testing, and recall. However, the states and/or
air quality control regions (AQCRs) were given the basic
responsibility for preparing plans under which the progressively
more stringent national ambient air quality standards could be
achieved. Of necessity, a part of any such plan includes a
statement regarding the impact of motor vehicle emissions on
air quality, and, if appropriate, methods for controlling the
exhaust emissions from in-use vehicles.
To help states and AQCRs in the development of such plans and
to help estimate the deterioration of emissions in a real
world environment, the EPA instituted the Emission Factor
Program in FY 1971.
This program is designed to estimate areawide urban emission
levels for in-use vehicles on a. nationwide basis. Although
testing of other than light duty vehicles has occurred in
recent years, the emphasis of the program has been on estimating
passenger car emission levels. Testing of vehicles occurs in
cities representing various climatic and geographic conditions.
Light duty vehicles are currently tested according to the 1975
Federal Test Procedure (FTP) as stipulated in the Federal
Register . Illustration 1 summarizes the conditions and
assumptions which occur with the FTP. The FTP driving cycle
1 Federal Register, Vol. 137, No. 211, Nov. 15, 1972.
22-2
-------
has been determined to be a reasonably representative urban
stop-and-go commute on the basis of data collected in Los
Angeles in the 1960's. The ambient conditions are not inherent
in the driving cycle, but were established as part of the FTP
to ensure comparability of FTP test results. The hot/cold
weighting factor used in the FTP is not inherent in the driving
cycle, but was determined to be representative of the percent
of hot start vs. cold start miles on the basis of data collected
in Los Angeles.
Thus, the estimates of exhaust emission levels derived from
the annual Emission Factor Programs are representative of
emission levels in areawide urban scenarios which conform to
the assumptions and conditions specified in Illustration 1.
These estimates are updated annually, and are published in AP-
42, Compilation of_ Air Pollution Emission Factors. For scenarios
which do not conform to the FTP conditions and assumptions,
AP-42 estimates are more or less inappropriate, depending on
the extent of the nonconformity. Therefore, in recent updates
to AP-42, starting with Supplement 5, so-called "correction
factors" have been provided and recommended for use in incorporating
scenario-specific considerations into emission estimates. The
next supplement on mobile source emission factors was originally
scheduled for promulgation as AP-42, Supplement 8, and will
contain light duty vehicle correction factors for ambient
temperature, average speed, % hot start/% cold start operation,
humidity, a,iT conditioning, and extra vehicle loading. This
supplement number is no longer firm; however, the Supplement
8 designation will be used throughout this document. This
paper will focus on the methodology used to correct FTP exhaust
emission estimates for ambient temperature, average speed, and
% hot start/% cold start operation (hot/cold).
II. Data Bases
In order to correct for non-FTP conditions and assumptions, it
is necessary to have data taken under non-FTP conditions to
co.mpa.re with AP-42 numbers. There were two data bases available
to EPA, representing FTP testing which occurred outside the FTP
ambient temperature range of 68 to 86°F. One set of data was
provided by the study Ambient Temperature and Vehicle Emissions
CEPA 460/3-74-028, October 1974), in which FTP emission tests
were performed on 26 1967-1974, and 1975 prototype model year
non-California light duty cars at ambient temperatures of 20°
to 11Q°F, Moat cars underwent four tests at ambient temperatures
of 20% 50% 75% and 100QF respectively. The other set of data
22-3
-------
was provided by the Environmental Protection Service, Ottawa,
Ontario, and represents FTP emission tests on 12 1975 model
year light duty vehicles, 7 of which were designed to meet
which federal emission standards, and 5 of which were designed
to meet the more stringent California emission standards.
Multiple tests were performed on each vehicle at ambients of 0°
to 80°F. For both data sets, the NOx emissions were normalized
to 25 grains of water/lb. of dry air, as stipulated by the
FTP . Also, all cars were in a tuned-up condition immediately
prior to testing. A description of each set of vehicles is
provided in Illustrations 2 and 3 respectively. Illustration 4
summarizes some of the characteristics of these two fleets.
III. Methodology of Correction Factor Development
The general approach to developing correction factors in both
Supplement 5 and Supplement 8 has been to calculate multiplicative
factors which are applied to the AP-42 estimates of HC, CO, and
NOx. The final forms of the Supplement 5 and Supplement 8
correction factors are given in Illustration 5, and are viewed
as part of the basic emission factor calculation. In Supplement
5 three separate factors were used to correct for temperature,
average speed, and hot/cold operation, respectively. Due to
lack of data, the possibility of interaction among these three
factors was by and large disregarded. In the process of developing
Supplement 8 correction factors it was decided on the basis of
the newly available Canadian data that interaction between
ambient temperature and hot/cold operation was significant
enough to warrant the use of a single, more complex correction
factor which would incorporate the interdependence of the
temperature and hot/cold variants. Future correction factor
work will probably be oriented towards the development of one
correction factor which incorporates the interdependency among
a.11 factors affecting exhaust emission levels. Although joint
data on average speed and hot/cold operation were not available,
it was felt that the modeling of an interaction between these
factors- was appropriate. Certain assumptions, which will be
discussed later, had to be made in order to Incorporate this
ititerdepeadency. One correction factor, the R1 correction
fa,ctor, results in Supplement 8 and consists of temperature,
hot/cold operation, and average speed components.
2 Federal Register. Vol. 28, No. 151, August 7, 1973.
22-4
-------
The approach to calculating the R1 factor was to first develop
the temperature-hot/cold portion by finding an expression for
CF which solves the following equation:
t ,w,x
Equation Is FTP^^ = (CF^^) (FTP75%
where t = ambient temperature (°F) ,
w = fraction of total miles driven which are driven
in cold start operation (representative of
emissions which occur on and shortly after
start-up following a long engine-of f period) ,
x = fraction of total miles driven which are driven
in hot start condition (representative of emissions
which occur on and shortly after start-up following
a short engine-of f period) ,
FTP = emissions (gm/mi) as measured over the FTP
>W)X driving cycle at t°F, with a % hot start/% cold
start operation ratio of x/w.
The situation where t is 75eF, w is .2058, and x is .2728
defines the FTP conditions and assumptions for ambient temperature
and hot/cold operation. (Although the FTP requirement is only
that vehicles be tested in an ambient temperature range of 68-
86°F, 75" was taken as representative of the temperature range
used for FTP testing).. As discussed previously, the FTP conditions
are reflected in AP-42 emission estimates. The above Equation
1 is equivalent to Equation 2:
FTP
Equation 2: CFtjW>x = FTP £i* .2058> .2728
The ambient temperature - hot/cold portion of the R' correction
factor is based on Equation 2. Thus, in conjunction with this
formulation of correction factors an expression for FTP
was fundamental. t,w,x
An equation for FTP,. can be formulated as follows:
t,w,x
Equation 3: FTP = w(Bag 1) + (1-w-x) (Bag 2) + x(Bag 3) ,
t,w,x t t t
where (Bag i) = emissions (gm/mi) in Bag i, i=l,2,3, at
temperature t.
22-5
-------
The next step was to determine Bag 1,2, and 3 emissions as an
appropriate function of ambient temperature.
The data provided by the Environmental Protection Service on
1975 model year cars included extensive replicate testing on
each vehicle at ambient temperatures of approximately 0 to
80°F. It was felt that these data would provide the best
estimate of the appropriate functional form for each bag's
gm/mi as a function of temperature, since each vehicle's data
represented a unit for analysis in and of itself. Data on
vehicles which were designed to meet California emission
standards were analyzed separately from data on vehicles which
were designed to meet the federal emission standards.
3
For each pollutant (HC, CO, and NOx), bag 1 , bag 2, and bag 3
emissions (gm/mi). were analyzed to determine the effect, if
any, of ambient temperature on emission levels. Based on an
examination of scatter plots of emissions vs. temperature and
an examination of residuals from the linear and loglinear
regression models, the loglinear model was selected. The
regression model was as follows:
Equation 4: ln(y) = a + a- t,
where In = natural logarithmic function,
y = emissions (gm/mi),
t = ambient temperature ("I?).
Due to the confounding effects of car-to-car variability which
arise when analyzing emission data on several different vehicles,
it was decided that covariance analysis could be useful in the
determination of an appropriate slope for the regression line.
Large experimental errors due, in this case, to car-to-car
variability would be encountered in a regression analysis which
does not stratify on vehicle, and would possibly result in an
inaccurate regression slope. By stratifying on vehicle for the
Bag 1 represents emissions which occur on and shortly after
start-up, following a long engine-off period (cold start),
bag 2 represents emissions which occur when a vehicle's
engine and/or emission controls are in a warmed-up condition
(stabilized), and bag 3 represents emissions which occur on
and shortly after start-up, following a short engine-off
period (hot start).
22-6
-------
analysis of covariance, smaller experimental errors result.
Thus, the relationship between emissions and temperature
becomes clearer.
Only bag 1 HC and bag 1 CO emission levels appeared to change
significantly with temperature. The expressions for these bag's
emissions in terms of temperature, derived from the best-
estimate slope determined by the analysis of covariance and the
average of the individual vehicle regression intercepts, are
presented in Illustration 6. It is emphasized that these
prediction equations reflect emission-temperature relationships
for 1975 model year vehicles in a tuned-up condition. As a
check on the robustness of these equations, a visual comparison
of the observed average bag 1 HC and bag 1 CO emissions for the
4 1975 model year prototype cars tested in the EPA study Ambient
Temperature and Vehicle Emissions with the predicted values was
made. The emissions predicted from the prediction equations
indicated reasonable predictability of bag 1 HC and bag 1 CO
emissions for these vehicles. The observed and predicted
emissions are presented in Illustration 7.
Also on the basis of these visual comparisons, it was decided
that extrapolation of the prediction equations to 110°F was
reasonable. The expressions presented in Illustration 6 are
therefore considered appropriate for temperatures in the range
0 to 11Q°F. Although the expressions are based on data for
1975 model year cars, they are assumed applicable to 1975-79
cars, since technology over these years has not changed and
probably will not change substantially.
The only source of ambient temperature-emissions data on pre-
1975 model years cars available to EPA was the study Ambient
Temperature and Vehicle Emissions. Since at most four FTP
tests were performed on a single vehicle, the reliability of a
functional form derived on the basis of these data alone was
considered poor. The general functional form for the pre-1975
model year cars' temperature-emissions relationships were
therefore taken to be loglinear, this form having been determined
acceptable for 1975 model year cars. Using this regression
model, an analysis of covariance was performed on all pre-1975
model year cars' data to determine expressions for bag emissions
(gm/mi) in terms of ambient temperature. As was the case for
the 1975 model year data, only bag 1 HC and bag 1 CO emissions
appeared to be significantly affected by ambient temperature
for pre-1975 cars. The expressions for bag 1 HC and bag 1 CO
emissions for pre-1975 model year cars are displayed in Illustra-
22-7
-------
tion 8. (A separate analysis of pre-controlled cars' data
yielded relationships similar to those for 1968-74 model year
cars, so all pre-1975 data were combined).
Since states and Air Quality Control Regions are interested in
estimating the emission levels of in-use, as opposed to tuned-
up vehicles, the expressions for bag 1 HC and bag 1 CO emissions
which are given in Illustrations 6 and 8 were modified by
adding a constant equal to the difference between low mileage
0-0,000 miles) AP-42 bag 1 FTP emission levels, and bag 1
emission levels at 75° as predicted by the expressions for
tuned-up vehicles. At this point in the development, the
temperature - hot/cold correction factors took the following
form:
CF ...
preliminary
w [exp CaQ + a- t) + c] + (1-w-x) (d) + x (e)
.2085 lexp CaQ + 75 a^ + c] + .5213 d + .2728 e,
c = difference between AP-42 low mileage bag 1
emissions Cgm/mi) and bag 1 emissions at 75° as
predicted by the expressions for tuned-up
vehicles,
d = AP-42 low mileage bag 2 emissions (gm/mi),
e = AP-42 low mileage bag 3 emissions (gm/mi),
w = fraction of total miles driven which are driven
in cold start condition,
x = fraction of total miles driven which are driven
in hot start condition.
For the above equation, FY72 and FY74 emission factor data on
pre-1975 and 1975 model year cars, respectively, were used to
calculate the percentages of the FTP which were in bags 1, 2,
and 3, respectively, for precontrolled cars, for 1968-74 model
year cars, and for 1975 model year cars. These percentages
were applied to the AP-42 low mileage FTP levels for each model
year group to obtain the values for c, d, and e. The denominator
in the above expression for CF preliminary is a low mileage
normalizing value, equal to the low mileage FTP level under the
FTP assumptions.
22-8
-------
A review of these preliminary temperature - hot/cold correction
factors resulted in a decision that a mileage or age consideration
should be incorporated. This decision was based on the fact
that these factors were extremely sensitive to the FTP levels
at 75° which were used as normalizing factors. For example,
the preliminary correction factors described above, which were
based on low mileage normalizing factors, were often substantially
different from those which were based on 100,000 mile normalizing
factors. The underlying assumption in applying a set of
correction factors which are normalized to a fixed mileage is
that the predicted steepness of the effect of low temperature
on emissions depends on mileage accumulation, as illustrated
below. However, available data from the EPA study suggest that
the assumption of such a dependency is unwarranted. Thus, the use
of correction factors which are normalized to a fixed mileage is
considered inappropriate.
Prediction of HC Bag 1 Emissions (gm/mi)
Fixed-mileage Normalized Temperature -
Hot/Cold Correction Factors
1975 Model Year Federal Light Duty Vehicles
1 Ht
I I
10 30 tO 50 frO 10
All Cold Start Driving CFs*
Normalized to Normalized to
10,000 miles 100,000 miles
20°
35°
50°
75°
5.64
4.10
3.03
1.89
2.67
2.14
1.76
1.36
* Curves A and C are based on
the application of the 10,000
mile normalized CFs to the
100,000 mile AP-42 FTP (3.94)
and to the 10,000 AP-42 HC FTP
(1.38), respectively. Curves
B and D are based on the appli-
cation of the 100,000 mile-
normalized CFs to the FTP levels
3.94 and 1.38, respectively.
22-9
-------
A consideration of vehicle age (mileage) was incorporated into
the expression for the temperature - hot/cold correction
factors in the following way:
CF =
w[exp (aQ+ a-j^
.2058[exp (aQ
where
t) + c + A df-]
+75a- ) + c + A
+ (1-w-x) (d + A df
df^] + .5213(d+A df
) + x(e + A df3)
2) + .2728(e + A df3)
df. = deterioration factor (gm/mi/year) in bag j, j = 1,2,3,
A = (vehicle age-1) , in years; A = 0 to 9. (One year
is assumed to be equivalent to 10,000 miles).
The above expression for CF is equivalent to the following:
CF -
w[exp(aQ+a1t) + c + A df^ + (1-w-x) (d + A df^ 4- x(e + A dfj)
v + ^
where
v = low mileage AP-42 FTP emission factor at 75°,
r = AP-42 FTP deterioration factor (gm/mi/yr) at 75°.
With regard to the above expression, for pre-1975s the assumption was
made that, for a given pollutant, percent deterioration per
year occurring in bag i (based on low mileage AP-42 emission
factors, by bag), which is fairly constant for all bags at
75°*, is constant for all temperatures. (The percent deterioration
in each bag is assumed to equal the overall percent deterioration,
r/v). For 1975-77 model year cars, the percent deterioration
which occurs in bags 2 and 3 has been shown to differ from
that in bag 1 for all pollutants.* For HC and CO, percent
deterioration is greater in bags 2 and 3 than in bag 1. For
NOx, deterioration is less in bags 2 and 3 than in bag 1. The
bag specific gm/mi/yr deterioration factors were obtained by
multiplying the percent deterioration by the AP-42 low mileage
bag 1, bag 2, and bag 3 emission levels. The basic input
July 20, 1976 internal EPA memo.
22-10
-------
values for the terms in the above expression for CF are given
in Illustration 9.
Incorporation of Speed Correction Factors; So far the discussion
of ambient temperature, average speed, and hot/cold correction
factors has centered around considerations of temperature and
hot/cold operation. Separate consideration of the speed portion
was necessary due to the lack of data covering all three
factors at once. Despite the lack of data, bag-specific speed
correction factors were incorporated into the correction
factors for temperature and hot/cold weighting as opposed to
being given as separate factors. Although the available bag-
specific data on speed effects were insufficient for analysis
purposes, engineering judgment supports the assumption of a
hot/cold-speed interdependency.
Until further data on speed-temperature interdependencies
become available, speed and temperature effects on emissions
will be assumed to be independent. This assumption will be
checked as data become available. The result of incorporating
speed correction factors into the temperature - hot/cold
correction factors given above is more accurate predicted
emission levels, at the expense of quite complex correction
factor calculations, if carried out by hand. Due to this
complexity, EPA is currently in the process of computerizing
the calculation of these factors. The derivation of the speed
correction factors and the final temperature, ambient temperature,
hot/cold correction factors is discussed below.
Derivation of Speed Correction Factors; Speed correction
factors for warmed-up vehicle operation were developed using a
four stage process. First, large quantities of second-by-
second speed-time data were collected, divided into acceleration,
deceleration, and steady state modes, and transcribed into
transition probability matrices. A transition probability
matrix gives the probability that a vehicle moves to speed x
given that it is currently at speed y. The data used to
develop the transition probability matrix were collected in
the 1970 Vehicle Operations Survey and the 1974 GM Chase Car
Survey, and covered 1957-1975 model year vehicles.
The second stage of the procedure generated second-by-second
speed-time cycles at average speeds between 5 and 60 mph.
Monte Carlo simulation techniques were applied to generate
cycles based on the transition probability matrix, and cycles
were screened so as to have the appropriate amount of idle
22-11
-------
time, acceleration time, deceleration time, cruise time, and
average speed, as determined from the observed average driving
cycle. Multiple cycles were generated for each average speed
point. All cycles were transient; no steady state only
cycles were generated.
Each of the transient speed-time cycles was fed into the EPA
modal emission model (EPA Report No. 460-3-74-065). The modal
emission model is a mathematical regression model which predicts
warmed-up vehicle emissions in gm/mi over arbitrary transient
driving sequences from speed-time data. The model is based on
modal input data from approximately 2000 1957-1975 model year
vehicles. For each model year vehicle, emission estimates
were obtained for each of the generated speed-time sequences.
Plots were made of predicted emissions vs. average speed and
regressions were performed. In all cases, the r value was
above .95. The regressions were then normalized to the average
speed of 19.6; that is, an average speed of 19.6 will give a
correction factor of 1.0. Illustration 10 lists the model
year/city groupings for which speed correction factors for
warmed-up vehicle operation were formulated, and Illustration
11 lists the formulas for the speed correction factors (normalized
to 19.6 mph) for stabilized and hot start driving conditions.
Illustration 12 provides the stabilized and hot start speed
correction factors in 5 mph increments. All HC and CO regressions
are exponential fifth order polynomials. All NOx regressions
are fourth order polynomials. This work was performed under
an EPA contract to Olson Laboratories. More detailed information
on the development of the speed correction factors can be
found in the contract final report which should be available
by late 1977.
These speed correction factors for stabilized and hot start
operation were used to derive bag-specific correction factors.
The need for bag-specific correction factors is based on two
considerations. First, the fact that the cold and hot bags of
the FTP have a different average speed from the stabilized bag
indicates a need for bag-specific speed factors. That is, if
the percentage operation in each of the bags is altered, the
average speed will also be altered. The second reason is more
complex. The speed correction factors discussed up until now
were developed to predict changes in warmed-up vehicle emissions.
It is not reasonable, from an engineering standpoint, to
expect these same emission changes to occur during cold operation
because of the operation of the choke: Pre-controlled vehicles
22-12
-------
were calibrated with rich air-fuel mixtures and as a result,
the difference in cold emissions and hot emissions at a given
average speed is not as great as for later model year vehicles
in which the air-fuel mixtures are leaner. During cold
operation of controlled vehicles, the operation of the choke
can be expected to result in a situation where emissions are
relatively less sensitive to changes in average speed than to
changes due to choke operation. Lacking conclusive data, the
emission dependency on speed during cold operation is assumed
to be similar for all model year vehicles and to be equal to
the dependency of pre-controlled vehicles during warmed-up
operation.
The calculations of the bag-specific speed correction factors
for a given pollutant were made according to the following
equations:
Bag 1 speed correction factor = v_ /v0 0,
2'S1 2»26
Bag 2 speed correction factor = v /v .fi
Bag 3 speed correction factor = v /v ~,
& Q O ft\
B> **o 6» *"»
where s = average speed in bag i, i=l,2,3
v, v - group-specific speed correction factor for
^'Si warmed-up operation normalized to 19.6 mph
Cg=2 designates pre-controlled low altitute
vehicles).
Due to lack of data, a consideration of modal effects in a
cycle of interest, such as the cycle's percent time or percent
miles spent in acceleration, deceleration, and idle, or the
sequence of modes in the cycle has not been included in the
development of the speed correction factors. Although EPA has
not investigated this question thoroughly, it appears that
for HC and CO as measured over various transient cycles, the
average speed difference accounts for the majority of the
effect on emission. NOx emission seems to be influenced a
little more by the cycle than is HC or CO emission, although
speed still has the major impact. It is expected that cycles
with higher percentages or rates of acceleration will have
higher NOx emissions for the same average speed.
22-13
-------
The bag-specific speed correction factors were incorporated
into the temperature-hot/cold factors, resulting in the general
formulas presented in Illustrations 13 to 15.
Applicability of Correction Factors; The general correction
factors are applicable to ambient temperatures of 0-110°F,
speeds of 5 to 60 mph, and all combinations of hot/cold driving.
Also, the factors should only be applied to transient driving
situations. To predict the emissions of a steady state driving
sequence such as constant 20 mph operation, the modal model
should be applied to the specific speed of interest and/or EPA
surveillance reports should be referenced. The difference
between emissions as measured over steady state vs. transient
cycles is considerable at low average speeds (greater than 20
percent) and becomes negligible at speeds of around 45 mph.
22-14
-------
Illustration 1
1975 Federal Test Procedure Conditions
for Light Duty Vehicles
1. Ambient temperature = 68° - 86°F (An average temperature of 75°F
is assumed to be representative of all tabled emission factor
values).
2. Ambient humidity = 75 grains
3. Average speed for entire FTP driving cycle =19.6 mph, 18% idle
operation.
a. Average speed for bags 1 and 3 (cold and hot start
bags) = 26 mph
b. Average speed for bag 2 (stabilized bag) = 16 mph
4. Average % cold start VMT operation = 20.58%
5. Average % hot start VMT operation = 27.28%
6. Average % stablilized VMT operation = 52.13%
7. Air-conditioning does not pull large power load
8. Car contains driver and fuel only, no passengers, luggage, etc.
9. Car is not pulling a trailer
22-15
-------
Illustration 2
Environmental Protection Agency Study Fleet
Model Year
Make & Model
2 3
Engine Size Transmission Emission Control
1967
1967
1967
1969
1969
1969
1969
1970
1970
1971
1971
1971
1971
1971
1972
1973
1973
1973
1974
1974
1974
1974
Prototype 1975
Prototype 1975
Prototype 1975
Prototype 1975
Ford Galaxie 289
Chevrolet Impala 283
Plymouth Fury 318
Chevrolet Malibu 307
Ford Galaxie 302
AMC Ambassador 290
Mercury Montery 390
Oldsmobile Cutlass 350
Chrysler Newport 383
Ford Galaxie 351
Chevrolet Impala 350
Dodge Coronet 318
Buick Electra 455
Chevrolet Impala 400
Ford Torino-C 351
Volvo 142 121
Chevrolet Laguna 350
Ford LTD 351
Ford Torino-C 351
Plymouth Fury III 360
Chevrolet Chevelle 350
Ford Torino 351-W
Ford Pinto 140
Plymouth Satellite 318
Ford LTD 400
Chevrolet Belair 350
A3
A2
A3
A2
A3
A3
A3
A3
A3
A3
A3
A3
A3
A3
A3
M4
A3
A3
A3
A3
A3
A3
A3
A3
A3
A3
PCV
PCV
PCV
EM, PCV
EM, PCV
EM, PCV
EM, PCV
EM, PCV
EM, PCV
EEC, EM, PCV
EEC, EM, PCV
EEC, EM, PCV
EEC, EM, PCV
EEC, EM, PCV
EEC, EM, PCV
EGR, EFI
EGR
EGR
EGR, MAI
EGR
EGR, MAI
EGR
EGR, MAI, Ox. Cat.
EGR, MAI, Ox. Cat.
EGR, MAI, Ox. Cat.
EGR, MAI, Ox. Cat.
Ambient Temperature and Vehicle Emissions, EPA Report 460/3-74-028, October 1974.
Transmission Code
A2 - automatic 2-speed
A3 - automatic 3-speed
M4 - manual 4-speed
Emission Control Code
PCV - positive crankcase ventilla.tion
EM - engine modifications
EEC - evaporative emission control
EGR - exhaust gas recirculation
MAI - manifold air injection
Ox. Cat. - oxidizing catalytic converter
22-16
-------
Illustration 3
Environmental Protection Service Fleet: 1975-1976
1975 Model Year Cars
Make & Model
Engine Size Transmission Emission Control'
Chevrolet Impala 350
Chevrolet Biscayne (Calif.) 350
Honda CIVIC (Calif.) 90.8
Honda CIVIC (Calif.) 90.8
Chevrolet Monza (Calif.) 140
Dodge Dart (Calif.) 318
Ford Maverick 250
Ford Custom 500 351
Dodge Dart 225
Chevrolet Nova 250
AMC Hornet 258
Dodge S/W Monaco 440
A3
A3
M4
M4
A3
A3
A3
A3
A3
A3
A3
A3
EGR, Ox. Cat., MAI
EGR
CVCC^
CVCC
EGR, Ox. Cat., MAI
EGR, Ox. Cat., MAI
EGR, MAI
EGR, MAI
EGR
EGR, Ox. Cat.
EGR
EGR
Transmission Code
A3 - automatic 3-speed
M4 - manual 4-speed
Emission Control Code
EGR - exhaust gas recirculation
MAI - manifold air injection
Ox. Cat. - oxidizing catalytic converter
Compound Vortex Controlled Combustion - CVCC is a stratified charge
system engine design which is used in place of auxiliary emission
control devices.
22-17
-------
Illustration 4
Ambient Temperature Data Bases
1. Ambient Temperature and Vehicle Emissions (EPA 460/3-74-028,
October 1974)
2. Environmental Protection Service, Ottawa, Ontario: 1975-76 Cold
Weather Fleet.
Data Base Characteristics
No. Tests
per Vehicle Model Years Ambient Tempera-
Source No. Vehicles (Range) Total Tests Represented ture Range
1. 26 3-4 96 1967-1975 20 to 110°F
2. 12 18-59 479 1975 Federal 0 to 80°F
and California
22-18
-------
Illustration 5
Basic Emission Factor Equation - Light Duty Vehicles
Supplement 5
n
npstwx =_ CiPn in
Supplement 8
n
Enpstwx = l ° ipn Min R'ipstwx Aip Lip Uipw Hip
i = model year
n = calendar year
s = speed
t = ambient temperature
w = fraction cold operation
x = fraction hot start operation
p = pollutant
c, c" = emission factor in AP-42 tables
M = fraction of total mileage
V = average speed cf
Z = temperature cf
R = hot/cold cf
R' = temperature, average speed, and hot/cold cf
A = air-conditioning cf
L = vehicle load cf
U = trailer towing cf
H ^ = humidity cf
22-19
-------
Illustration 6
Prediction of Bag 1 HC and Bag 1 CO Emissions from Ambient Temperature
1975 Model Year Vehicles in a Tuned-up Condition
1975 Model Year Bag 1 HC: y = exp (2.4339 - ,023591t)
Federal Cars: Bag 1 CO: y = exp (5.5460 - .028945t)
1975 Model Year Bag 1 HC: y = exp (1.9934 - .022269t)
California Cars: Bag 1 CO: y = exp (4.2391 - .017522t),
where y = emissions (gm/mi), exp = natural exponential function,
t = ambient temperature (°F). The equations are based
on data for 12 1975 model year vehicles tested at ambient
temperatures of 0° to 80°F. A total of 479 tests were
performed.
22-20
-------
Illustration 7
Observed* vs. Predicted** Emissions (gm/mi)
for Four 1975 Prototype Vehicles
Bag 1 HC
Ambient Temperature (°F)
20° 50° 75° 110°
Observed Level 5.2 2.6 1.3 1.0
E.P.S. Federal Predicted Level 7.1 3.5 1.9 .85***
E.P.S. California Predicted Level 4.7 2.4 1.4 .63***
Bag 1 CO
Ambient Temperature (°F)
20° 50° 75° 110°
Observed Level 126.3 67.5 17.0 10.1
E.P.S. Federal Predicted Level 143.6 60.3 29.2 10.6***
E.P.S. California Predicted Level 48.9 28.9 18.6 10.1***
* Observed levels were taken from EPA study 460/3-74-028, Ambient
Temperature and Vehicle Emissions.
** Predictions were based on the results of covariance analyses per-
formed on the Environmental Protection Service data on federal
and California cars, respectively.
*** Extrapolated values.
22-21
-------
Illustration 8
Prediction of Bag 1 HC and Bag 1 CO Emissions from Ambient Temperature
Pre-1975 Model Year Vehicles in a Tuned-up Condition
Bag 1 HC: y = exp (2.9310 - .014779t)
Bag 1 CO: y = exp (5.6548 - .015965t),
where y = emissions (gm/mi), exp = natural exponential function,
t = ambient temperature (°F). Equations are based on data
for 22 1967-74 model year cars tested at ambient temperatures
of 20° to 110°F from EPA study 460/3-74-028, October 1974. A
total of 83 tests were performed.
22-22
-------
to
r
NJ
Illustration 9
Numerical Inputs to Correction Factor Expressions
Pre-1968
1968-74
1975 Federal
1975 Calif
HC
CO
NOx
HC
CO
NOx
HC
CO
NOx
HC
CO
NOx
al
-0.014779
-0.015965
0.
-0.014779
-0.015965
0.
-0.023591
-0.028945
0.
-0.022269
-0.017522
0.
The values
ao
2.9310
5.6548
0.
2.9310
5.6548
0.
2.4339
5.5460
0.
1.9934
4.2391
0.
for d^,
c e
.673 4.746
-14.74 42.84
2.876 4.25
-2.41 2.43
-33.89 25.26
4.44 5.92
0.623 1.11
11.29 15.855
2.26 2.989
-.032 .497
-0.20 4.12
2.05 2.88
d
5.685
57.57
2.768
2.61
35.9
3.77
1.053
21.167
1.887
.243
3.96
2.009
df2, and df, were calculated from
df1 - (rc/v)*(EXP(75
df2 = (rh/v)*d
df, = (rh/v)*e.
.*an+a )+c)
1 o
V
5.67
56.43
3.40
2.8
36.4
4.7
1.38
23.7
2.47
0.54
6.98
2.46
r
0.47
7.59
0.
0.64
6.79
0.
0.28
3.14
0.18
0.28
3.14
0.18
the following
re
0.47
7.59
0.
0.64
6.79
0.
.162
2.48
.2540
.178
2.645
.2563
formulas :
rn
0.47
7.59
0.
0.64
6.79
0.
.3532
3.50
.1524
.388
3.729
.1538
-------
Group Number
Group 1
Group 2
Group 3
Group 4
Group 5
Group 6
Group 7
Group 8
Group 9
Group 10
Group 11
Group 12
Group 13
Group 14
Group 15
Group 16
Group 17
Group 18
Illustration 10
Group Definitions for Speed Correction Factors
Group Definition
Denver pre-controlled
All low altitude cities
pre-controlled
1966-1967 California
1968 low altitude cities
1969 low altitude cities
1970 low altitude cities
1971 low altitude cities
1968 Denver
1969 Denver
1970 Denver
1971 Denver
1972 Denver
1972 Los Angeles
1972 low altitude cities
1973-1974 Denver
1973-1974 Los Angeles
1973-1974 low altitude cities
1975 low altitude cities
22-24
-------
Illustration jtl
Speed Correction Factor Formulas
Normalized Equations
In HC - AQ + A1 s + A2 s2 + Aj s3 + A/, s4 +
in CO -A0+Al8 + A2B2 + A3a3+AA84 +
NOx - A
a + A2.82 + A, s3 + A,
GROUP* i
HC* 0.22461E*01
C0= 0.1SJ98E*01
0.24442E+01
0.29097E*00
0.25466E*00
0.25011E*00
0.158S9E-01
0.15235E-01
0.13829E-01
0.47249E-03
0.48740E-03
0.28703E-03
0.69408E-05
0.75821E-05
0.20758E-05
0.39280E-07
0.44951E-07
0.0 '
1
Ul
GROUPS
HC=
CO*
NOX=
2
0.«!3103E*01
0.23399E*01
0.16863E01
. -0,28957E*00
-0.29698E*00
-0.11330E*00
0.15299E-01
0.16007E-01
0.65497E-02
-0.44669E-03
-0.47740E-03
-0.13714E-03
0.64818E-05
0.70675E-05
0.10085E-05
-0.36346E-07
-0.40398E-07
0.0
GROUP= 3
COs
NOXa
0.21656E-+01
0»2A415E*01
0.il265E*01
0,26999E*OQ
0.29147E*00
0.3934QE-01
0.14422E-01
0.14295E-01
0.26864E-02
0.43364E-03
0.38785E-03
0.60802E-04
0.65074E-05
0.52978E-05
0.47729E-06
0.37810E-07
0.28244E-07
0*0
-------
ro
N»
ON
GROUP"
HCs
C0=
NOXs
QROUPa
HO
CO*
NOXs
GROUP*
HCs
CO*
NOX=
GROUP*
HCs
CO*
NOXs
GROUPS
HCs
co«
NOXs
GROUPs
HCs
CO*
NOXs
(Illustration
4
0.23973F*01
0.24655E+01
0.1226flf>01
5
0.240B7E*Ol
0.27780E+01
U.10174E+01
6
0,.22322E*01
0.27890E*01
0.98760E+00
7
0.22522F>01
0.27074E+Q1
0.11592E+01
8
0.20278E*01
0.18692E+01
0.18866E+01
9
0.21506E+01
0.18213E+01
0.15578E+01
11 cont'd)
-0.29998E*00
-0.30502E*00
-0.44498E-01
-0.30819E*00
-0.31913E*00
-0.11R96E-01
-0.28499E+00
-0,32711E*00
-0.19567E-01
-0.28773E*00
-0.33131E*00
-0.44454E-01
-0.27305E+00
-0.27668E*00
-0.16129E*00
-0.28362E*00
-0.27205E*00
-0.11303E*00
*
0.16M5E-01
0.16050E-01
Ot262^8E-02
\
Otl6817E-01
0.15318E-01
0.91437E-03
0.1S383E-01
0.16294E-01
0.16964E-02
0.156R2E-01
0.17618E-01
0.29643E-02
0.1S358E-01
0.17233E-01
0.90499E-02
0.15384E-01.
0.17030E-01
0.67183E-02
-0.48749E-03
-0*47397E-03
-0.56715E-04
-O.S0684P-03
-0.42233fT-03
.-0.21574F-04
-0.45674F-03
-0.46757E-03
-0.40400E-04
-0.47318E-03
-0.53858E-03
-0.66899E-04
-0,46030E-03-
*.O.S58?8E-03 .
-0.18S61E-03
-0.44214E-03
-0.55202E-03
-0.14341E-03
0.72909E-05
0.69908E-05
0.43429E-06
0.75385E-05
0.5fl495E-05
0.1fl230E-06
0.67349E-05
0.67191E-05
0.32800E-06
0.70795E-05
0.81740E-05
0.52236E-06
0.67853E-05
0.87168E-05
0.13256E-05
0.62873E-05
0.86254E-05
0.10608E-05
-0
-0,
0
-0,
-0,
0,
-0,
-0,
0(
-0,
-0,
0.
-0,
-0,
0.
-0,
-0«
0.
0.41977E-07
0.39976E-07
0.43160E-07
0.31497E-07
0.38380E-07
0.37440E-07
.0
0.40846E-07
0.47780E-07
0.38488E-07
0.51698E-07
GROUP"10
HCs 0,22302fT*01
C0s 0.20342E*01
NOXs 0,20452t-+01
0.29365E+00
0.29519E+00
0.19401E+00
0.16236E-01
0.18635E-01
0.1l074ErOl
0.48415E-03
0.62161E-03
0.23175E-03
0.71159E-05
0.99366E-05
0.16837E-05
0.34631E-07
.51144E-07
.0
0.40286E-07
0.59978E-07
0.0
-------
(Illustration 11 cont'd)
10
N>
N>
GROUP* 11
HC= 0.21223?*01
C0= 0.20453f:*01
NOXs 0.101
FE= 0.35076E-01
GROUP«12
, HC= 0.21536E*01
C0= 0.23187E+01
NOX= 0.14482E*01
GROUP«13
HC= 0.20735E+01
C0= 0.25752fi*01
NOX= 0.24597E+00
GPOUPsl4
HC= 0.23495K*01
C0= 0.26B45E+01
NOX« 0,12817E*01
GROUP* 15
HC= 0.21134E*01
C0= 0.2154QE*01
NOXs 0.15345E*01
GROUP=16
HCs 0.21194E*01
CO* 0.25456E*01
NOX= 0.70481E*QO
6ROUP=17
HC« 0.2fi838F:*Ol
C0= 0.2S393E+01
NOXs 0.78384E*00
GROUP= 18
HO 0.23954fT*01
C0= 0.24875E*01
NOXa 0.94213E*00
-0.29107E*00
-0.31062E*00
-0.12186E*00
0.87843E-01
-0.28345E+00
-0.34115E*00
-0.12?64E+00
-0.28935E*00
-0.32889E*00
0.84195E-01
-0.30^96E*00
-0.33282E*00
-0.80487E-01
-0.28568E+00
-0.32912E*00
-0.12567E+00
-0.29863E*00
-0,36295E*00
0.38153E-01
-0,34463E+00
-0.36876E+00
0.32855E-03
-0.33578E*00
-0.39156E*00
-0.42324E-01
0.16909E-01.
0.20485E-01
0.70302E.-02
-0. 27727^-02
0.15695E-01
0.20945E-01
0.79502F--02
d,l73Q4E-01
0.18975E-01
-0.34Q84E-02
0.16842E-01
0.17628E-OI
0.53574E-02
0.16318E-01
0.210UE-01
0.78592E-02
0.18447E-01
0.23277E-01
-0.17391E-02
' *. ;
0.19542E-01
0.21078E-01
0.10603E-02
0.21161E-01
0.27072E-01
0.38625E-02
-0.52615E-03
-0.70853E-03
-0.14629E-03
0.47466E-04
-0.46976E-03
-0.66S89F.-03
-0.17108E-03
-0.55471E-03
-0.62826E-03
0.62988E-04
-0.50962E-03
-0.52412E-03
-0.11889E-03
-O.S0079E-03
-0.68906F-03
-0.16943E-03
-0.61654E-03
-0.81504E-03
0.32614E-04
-0.62572E-03
-0.67644F-03
-0.31935E-04
-0.73155E-03
-0.97618E-03
-0.93985E-04
0.80271E-OS
0.11621E-04
0.10614E-05
,.-0.33220E-06
0.69383E-05
0.10223E-04
0.12578E-05
1
0.86420E-05
0.10092E-04
-0.41397E-06
0.75952E-05
0.77222E-05
0.90106E-06
0.75507E-05
0.10839E-04
0.12549E-05
0.99206E-05
0.13623E-04
-0.20385E-06
0.97844E-05
0.10627E-04
0.29039E-06
0.12072E-04
0.16527E-04
0.75388E-06
-0.47012E-07
-0.71569E-07
0.0
0.0
-0.39471E-07
-0.59827E-07
0.0
-0.51311E-07
-0.61273E-07
< 0.0
-0.43496E-07
-0.43702E-07
0.0
-0.43719E-07
-0.64712E-07
0.0
-0.60402E-07
-0.85591E-07
0.0
-0.58337E-07
-0.63641E-07
0.0
*
-0.74857E-07
-0.10432E-06
0*0
-------
Illustration 12
Selected Speed Correction Factors - Warm Operation
N>
K>
00
GROUP
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
- Hydrocarbon
Average Speed
5.000
3.107
3.297
3.083
3.470
3.419
3.123
3.160
2.700
2.903
3.039
2.798
2.928
2.705
3.276
2.815
2.763
3.963
3.194
10.
1.679
1.749
1,708
1.808
1.773
1.695
1.709
1.548
1.600
1.649
1.571
1.623
1.548
1.726
1.582
1.575
1.931
1.708
000 15.
1.201
1.224
1.219
1.2*6
1.231
1.208
1.215
1.160
1.170
1.190
1.168
. 1.184
1.165.
1.216
1.172
l.lfll
1.285
1.228
000 20.
0.987
0.986
0.986
0.984
0.985
0.987
0.986
0.990
0.990
0.988
0.989
0.938
0.989
0.986
. 0.989
0.987
0.981
0.984
000 25.
0.858
0.844
0.841
0.821
0.834
0.853
0.845
0.85-9
0.891
0.867
0.877
0.871
0.871
0.844
0.876
0.848
0.784
0.803
000 30.
0,761
0.740
0.733
0.700
0.720
0.754
0.740
0.811
0.819
0.775
0.788
0.781
0.773
0.736
0.788
0.729
0.635
0.653
000 35.
0.684
0.659
0.650
0.606
0.630
0.677
0.658
0.748
0.762
0.702
0.716
0.709
0.694
0.650
0.717
0.635
0.523
0.540
000 40.
0.629
0.600
0.592
0.538
0.565
0.622
0.600
0.703
0.720
0.649
0.667
0.659
0.640
0.589
0.669
0.573
0.446
0.468
000 45.
0.597
0.565
0.556
0.497
0.526
0.591
0.567
0.681
0.699
0.619
0.644
0.632
0.616
0.554
0.646
0.543
0.401
0.432
000 50.
0.585
0.547
0.534
0.472
0.504
0.576
0.551
0.678
0.694
0.609
0.638
0.624
0.610
0.538
0.641
0.531
0.373
0.414
000 55.
0.571
0.530
0.503
' 0.445
0.479
0.557
0.529
0.672
0.691
0.596
0.617
0.612
0.589
0.519
0.628
0.497
0.337
0.373
000 60.000
0.516
0.482
0.429
0.381
0.414
0.495
0.460
0.616
0.649
0.538
0.525
0.553
0.493
0.456
0.553
0.383
0.258
0.260
-------
(Illustration 12 -oont'd)
N>
VO
GROUP
1
2
3
4
5
6
7
8
9 *
10
11
12
13
14
15
16
17
13
GROUP
I
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Carbon Monoxide
Average Speed
5.
2.389
3.319
3.656
3.621
.4.554
4.511
4.174
2.345
2.277
2.541
2.516
2.885
3.791
4.056
2.599
3.384
4.239
2.988
000 10.
1.463
1.751
1.856
1.844
2.120
2.103
2.003
1.418
1.395
1.488
1.474
1.540
1.916
1.950
1.459
1.744
1.980
1.580
000 15.
1.142
1.225
1.251
1.253
1.329
1.326
1.299
1.121
1.113
1.149
1.148
1.149
1.291
1.281
1.127
1.237
1.293
1.183
000 20.000 25.000
0.991
0.986
0.985
0.984
0.979
0.979
0.975
0,992
0.993
0,990
0.989
0.991
0.980
0.982
0.992
0.983
0.981
0.986
0.889
0.841
0.837
0.823
0.781
0.778
0.776
0.905
0.913
0.873 .
0.863
0.891- .
0.771
0.804
0.900
0.795
0.782
0.821
30.000
0.803
0,734
0.738
0.707
0.644
0.637
0.633
0.827-
0.841
0.770
0.746
O.S04
0.612
0.675
0.814
0.641
0.634 '
0.671
35.000 40.000 45.000 50.000 55.000 60.000
0.733
0.650
0.663
0.619
0.543
0.533
0.526
0.760
0.780
0.684
0.651
0.729
0.496
0.577
0,739
0.527
0.525
0.557
0.686
0.592
0.608
0.556
0.469
0.457
0.453
0.717
0.743
0.629
0.591
0.677
0.420
0.510
0.691
0.458
0,454
0.493
0.665
0.557
0.572
0.517
0.418
0.407
0.406
0.703
0.737
... 0.608
0.568
0.657
0.378
0.470
0.67S
0.427
0.415
0.475
0.663
0.539
0.554
0.493
0.384
0.374
0.375
0.711
0.755
0.608
0.566
0.658
0.355
0.451
0.681
0.410
0.395
0.478
0.647
0.518
0.542
0,465
0.358
0.345
0.341
0.701
0.756
0.589
0.534
0.643
0.324
0.436
0.661
0.360
0.364
0.434
0.558
0.454
0.515
0.399
0.323
0.299
0.273
0.600
0.662
0.478
0.397
0.538
0.246
0.391
0.530
0.227
0.278
0.265
Nitric Oxide
5.
1.505
1.242
0.990
1.063
0.978
0.927
1.003
1.284
1.143
1,324
1.181
1.014
0.589
0.999
1.082
0.856
0.808
0.816
000 10.
1.060
1.031
0.946
0.992
0.970
0.924
0.949
1.006
0.966
0.997
0.981
0.860
0.806
0.903
0.907
0.943
0.864
0.819
000 15.
0.941
0.974
0.960
0.980
0.9B1
0.956
0,960
0,944
0.944
0.930
0.946
0.887
0.934
0.924
0.909
0.986
0.934
0.897
000 20.
1.010
1*004
1*004
1.002
1.002
1.004
1,004
1,008
1,007
1.010
1.007
1.012
1.004
1*008
,. 1*010
1.001
1.006
1.009
Average
Speed
000 25.000 30.000 35,
1.161
1.074
1.058
1.038
1.026
1.056
1.059
' 1.128
1.105
1.152
1.109
1.174
1.043
1.112
1.148
1.002
1.069
1.124
1.319
1.146
1,109
1.075 ,
1.049
1.102
1.110
1.255
1.200
1.297
1.214
1.330
1.070
J.208
1.280
1.000
1.121
1.222
1.440
1.203
.150
.105
.070
.141
,150
.359
1.275
1,410
1.300
1.454
1.097
1.282
1.382
1.002
1.161
1.294
000 40.
1.511
1.239
1.182
1.129
1.091
1.173
1.180
1.429
1.323
1,480
1.361
1.542
1.132
1.332
1.452
1.014
1.193
1.344
000 45.
1.551
1.26S
1.213
1.152
1.115
1.206
1.207
1.477
1.358
1.524
1.407
1.606
1.175
1.369
1.501
1.036
1.226
1.386
000 50.
1,608
1.306
1.258
1.189
1,151
1.250
1.250
1.531
1,406
1.582
1.462
1.678
1.221
1.421
1.564
1,067
1.274
1.446
000 55.
1.764
1.404
1.340
1.257
1.208
1.323
1.331
.. 1.641
1.511
1.721
1.570
1.B10
1.258
1.526
1.691
1.103
1.353
1.560
000 60.000
2.129
1.615
1.489
1.384
1.298
1.445
1.483
1.877
1.733
2.031
1*786
2.070
1.268
1.736
1.955
1.136
1.486
1.777
-------
(Illustration 12 cont'd)
K>
Ni
CROUP
1
2
6
7
8
9
10
II
12
II
14
15
16
1"
18
GROUP
1
2
7
8
9
10
11
1?
15
16
.17
IS
Hydrocarbon
Average Speed
6.000
2.662
2.816
2.663
2.951
2.901
,2.680
2.710
2.343
2.497
2.604
2.413
2.522
2. 343
2.788
2.430
2.389
3.309
2.717
11.000
.544
.601
.573 :
.650
. 620
..558
.571
.438
.477
.519
.457
.499 1
.440 1
.582 1
. 466 1
.464 1
.748 1
. 573 1
16.000 21. 000 "'26. 000" 31.000 36.000 41.000 46.000 51.000 56.000 61.000
1.146 0.957' 6.837. 0.744 0.671 0.621. 0.594 0.583 0.565 0.495
1.163 0.953 0.821 0.722 0.645 0.591 0.560 0.544 0.525 0.464
1.160 0.952 0.817 0.715 0.637 0.583 0.551 0.529 0.493 0.406
1.180 0.946 0.794 0.679 0.590 0.528 0.491 0.468 0.437 0.361
.168 0.950 0.809 0.700 0.615 0.555 0.520 i 0.500 0.471 0.793
.151 0.956 0.832 0.737 0.664 0.614 0.587' 0.5"3 O.f.50 O.H/4
.156 0.954 0.822 0.7£? 0.644 0.592 0.563 0.548 0.520 0.437
.116 0.967 0.872 0.797 0.737 0.697 0.680; 0.679 0.667 0.594
.123 0.966 0.875 0.807 0.752 0.714 0.697 0.694 0.687 0.631
.138 0.960 0.847 0.759 0.689 0.641 0.6161 0.607 0.589 0.51*
.122 0.964 0.858 0.772- 0.704 0.661 0.642' 0.636 0.606 O.H'f'l
.134 0.961 0.851 0.765 0.698 0.652 0.629 0.623 O.tOt 0.531
.121 0.963 0.850 0.756 0.681 0.633 0.614 0.609 0.578 O.HsO
.158 0.953 .0/821 0.717 0.636 0.580 0.550 0.535 0.511" 0.434
.125 0.963 0.857 0.772 0.705 0.662 0.644 0.640 0.620 0.526
.133 0.957 0.823 0.708 0.620 0.564 0.540 0.527 0.482 0.3H8
.209 0.936 0.751 0.610 0.505 0.435 0.395 0.368 0.325 0.3-75
.169 0.945 0.770 0.627 0.522 0.458 " 0.428 " O.'UQ O.J^ O.t**'
Carbon Monoxide
Average Speed
6.000 11.000 16.
2.105 1.373 1.104
2.829 .602 1.164
3.096
3.063
3.784
3.745
3.482
2.054 1
2.001 1
2.208
2.183
2.447
3. 196
3.383
2.228
2.848
3.500
.685 1.181
.678 1.184
.896 1.239
.882 1.237
.804 1.217
.333 1.088
.314 1.082
.392 1
.381 1
.425
.742
.759
.361
.601
.783. ... 1
"2.313 .465 1
. 110
.110
. 107
.215
.204
.092
.176
.214
.137
000 21.
0.968
0.952
0.949
0.946
0.932
0.932
0 . 929
0.973
0 . 975
0 . 965
0.963
0.969
0 . 933
0.940
0.972
0.942
0.935
0 . 952
000 26.
0.871
0.817
,0'.815
0.797
0. 750
0.746
0.744
0 . 889
0.898
0 . 852
.0.839
0 . 873
0.735
0.775
0.882
0.762
0.749
0.789
000 31.
0.787
0.715
0.721
0.687
0.622
0.613
0.609
0.812
0 . 827
0.751
0.725
0.788
0.585
0 . 653
0.797
0.615
0.609
"0.644
000 36.
0.721
0.637
0.650
0.604
0.526
0.515
0.509
0.749
0.770
0.670
0.636
0.716
0.477
0 . 562
0.727
0.510
0.508
0.540
000 41.
0.679
0.583
0.599
0.546
0.457
0.4H6
O.HHI
0.712
0.739
0 . 622
0.583
0.671
0.409
0.500
0.685
0.449
O.H44
0.486
000 46.
0.664
0.552
0.567
0.511
0 . H 1 0
0 . 399
0.399
0.704
0.739
0.607
0.567
0.656
0.373
0.465
Ci . 676
0.423
O.H10
0.475
000 51.000
0.663
0.536
0.551
0.488
0 . 3~9
0.369
0.370
0.712
0 . "58
0.608
0.564
0.658
0.351
0.449
0 . 682
0.405
0.391..
0.476
56.000
0.637
0.510
0.539
0.456
OT«s"T
«..;
0 . 338
0.331
0.691
0.748
0.577
0.517
0.632
0.313
0.431
0.646
0.341
0.352
0.411
61.000
0.?56
0.473
0.504
0 . 7~8
0.313
O.S85
0.253
0 . 563
0.624
0.440
0.355
0.501
0.223
0.375
0.486
0.193
0.253
0.221
-------
(Illustration & cont'd)
NJ
r
CO
GROUP
1
5
7
4
I?
6
7
8
9
10
II
15
13
14
15
16
17
18
Nitric Oxide
Average Speed
6.000 11.000 16.000
1.385 1.015 0.943
1.184- 1.010 0.975
0 . 975 0 . 945 0 . 967
1.043 0.986 0.983
0.975 0.971 0.985
0 .953 0 . 959 0 . 965
0 . 985 0 . 947 0 . 967
1.506 0.980 0.949
1.095 0.955 0.951
1 i 535 0 . 967 0 . 937
1 . 1 54 0 . 964 0 . 953
0.964 0.854 0.906
0.645 0.837 0.951
0.967 0.900 0.937
1.058 0.896 0.954
0 . 878 0 . 954 0 . 990
0.817 0.877 0.949
0.808 0.830 0.918.
51.000
.036
.017
.015
.009
.007
.014
.015
.059
.055
.036
.055
.043
.014
.058
.036
.001
.019
.033
<
I/
l.l
.1
.1
1.1
1.1
.1
.
.
.
,
.
.
.
.
.0
.c
.1
6.000
194
389
369
346
330
365
370
54
55
83
30
-07
49
33
75
01
81
45
. 31.000
1347
. 159
. 118
.081
.053
.111
. 119
.578
.517
.353
.533
.358
.075
.555
.303
.000
. 130 1
.538 1
36.000
.458
.511
.157
.110
.074
. 148
.157
.376
.586
.457
.314
.475
.103
.594
.399
.004
. 168
.306
41.000
.551
.544'
.187
.133
.095
.180'
.185
.440
.331
.490
.371
.556
. 140
.340
. 465
.017
. 199
.355
46.000
.559
.571
.550
.158
.151 ;
.513 ;
.514
.486 .
.365 1
>?77 |
«P ' <^
.416
.618
. 1 84
.378 .
.511
.045
.534
.396
51.000
.658
.350
.571
. 199
. 160
. 565
.565
.546
.451
.601
.. 478
.698
. 559
.436
. 585
.074
. 587
.463
56.000 61.000
.816 5.539
.436 1 . 677
.364 1 . 530
.577 1.419
.553 1.351
.343 1 . 478
.354 1 . 555
.675 1.946
.544 1 . 798
.766 5.154
.605 1.849
.849 5.145
. 563 1 . 564
.557 1.797
.731 5.031
.110 1.141
.375 1.551
.594 1 . 838
-------
Illustration 13
Light Duty Vehicle Hydrocarbon Correction Factors for Ambient Temperature and/or Average Speed
and/or Z Hot Start/Z Cold Start Driving which differ from the FTP Assumptions*
Model Year
pre-1968
1968-74
[w(exp(2.9310-.014779t) + . 673+. 569A) (v.
Correction Factor**
/.821) + x(4.75 +.393A)(v
/VB ,,) + (l-w-x)(5.69 +.471A)(vB 0 /VB .,)]/(5.67+.47A)
[w(exp(2.9310-.014779t) - 2.41 +.863A)(v, /.821) + x(2.43 +.555A)(v /v ,,) + (l-w-x)(2.61 +.597A)(v /VB -,)]/(2.8 +.64A)
*»»j_ 8t»3 8»*o 8»»^ at*"
K>
1975 Federal [w(exp(2.4339-.023591t) + .623+.301A)(v2
1975 California [w(exp(1.9934-.022269t) - .032+.445A)(v2
* FTP assumptions are outlined in Table I.I.
/.821) -f x(l.ll +.284A)(v,
x( .497+.357A)(v
/va ,,)
(l-w-x)(1.05 +.270A)(v
,243+.175A)(v
Jv
)]/(!. 38+. 28A)
.54+.28A)
**
v < fraction of total miles which are driven in cold start condition,
x fraction of total miles which are driven in hot start condition,
t - ambient temperature (*F),
A vehicle age minus 1, in years,
g index for model year-city groups for which average speed data wsre available (g 1, 2,... ,18),
s. average speed (miles per hour) in bag 1, 1 » 1,2,3,
v( )^v( ) " ^)a8~8Peci^ic speed correction factor (.821 equals v. ,g)»
exp - exponential function, base e.
These 'correction factors are applicable to ambient temperatures of 0* to 110*F, vehicle ages of 1 to 10 years, average vehicle
speeds of 5 to 60 mph, and to all combinations of Z hot start/% cold start driving. The incorporation of the appropriate v
terms (calculated from Table 11.13) in the above equations would yield a correction factor for each of the 19 model year-clE^ groups
for which data were analyzed. Table II. 4 provides the method for combining the basic equations provided in the above table with
the appropriate v terms In order to perform calculations without the aid of a computer.
-------
Illustration 14
Model Year
Light Duty Vehicle Carbon Monoxide Correction Factors for Ambient Temperature and/or Average Speed
and/or Z Hot Start/Z Cold Start Driving which differ from the FTP Assumptions*
Correction Factor**
to
u>
pre-1968 [w(exp(5.6548-.015965t) - 14.74+9.62A)(v2 fl /.817) 4- x (42.844-5.76A)(v g /v 2fi) 4- (1-w-x)(57.57f7.74A)(v g /v 16)J/(56.434-7.59A)
1968-74 [w(exp(5.6548-.015965t) - 33.89+9.77A)(v, /.817) 4-x(25.2644.71A)(v /v ,,) 4- (l-w-x)(35.9 +6.70A)(v /v 1R)]/(36.4 +6.79A)
f**i oft .1 o§ " & ' A o v *»
1975 Federal (w(exp(5.5460-.028945t) 4- 11.294-4.24A)(v2|8 /.817) 4- x(15.854-2.34A)(vgjB /vg>26) 4- (1-w-x)(21.174-3.13A)(vg g /v 16)]/(23.7 4-3.14A)
1975 California [w(exp(4.2391-.017522t) - .20«.99A)(v2jB /.817) 4- x( 4.124-2.20A)(vgj(j /vg>26) 4- (1-w-x)( 3.96«.12A)(v g /v lfi)]/( 6.984-3.14A)
* FTP assumptions are outlined in Table Z.I.
** w fraction of total miles which are driven in cold start condition,
x " fraction of total miles which are driven in hot start condition,
t * ambient temperature (*F),
A vehicle age minus 1, in years,
g « index for model year-city groups for which average speed data were available (g 1, 2,...,18),
s. average speed (miles per hour) in bag i, i - 1,2,3,
v< »/v, . bag-specific speed correction factor (.817 equals v. .g)»
exp - exponential function, base e.
These correction factors are applicable to ambient temperatures of 0* to 110*F, vehicle ages of 1 to 10 years, average vehicle
speeds of 5 to 60 mph, and to all combinations of % hot start/Z cold start driving. The incorporation of the appropriate v
terms (calculated from Table 11.13) in the above equations would yield a correction factor for each of the 19 model year-cifyagroupa
for which data were analyzed. Table II.4 provides the method for combining the basic equations provided in the above table with
the appropriate v terms in order to perform calculations without the aid of a computer.
-------
_
r*
l
Illustration 15
Light Duty Vehicle Oxides of Nitrogen Correction Factors for Average Speed and/or
. X Hot Start/Z Cold Start Driving which differ from the FTP Assumptions*
Model Year
pre-1968
1968-74
1975 Federal
1975 California
[v(3.26 » .335A)<2.01
/v
/v
.116A)(vB /VB
*' ft. 8
.126A)
)l/(2.47 + .ISA)
l)]/(2.46 + .18A)
These correction factors are applicable to ambient temperatures of 0* to 110*?, vehicle ages of 1 to 10 years, average vehicle
speeds of 5 to 60 aph, and to all combinations of Z hot start/% cold start driving. The Incorporation of the appropriate v
terms (calculated from Table 11.13) in the above equations would yield a correction factor for each of the 19 model year-cify groups
for which data were analysed. Table II. 4 provides the method for combining the basic equations provided in the above table with
the appropriate vn a terms in order to perform calculations without the aid of a computer.
B»a
-------
LAND USE BASED EMISSIONS FACTORS
Presented at the 1977
Environmental Protection Agency
Emission Inventory/Factor Workshop
Raleigh, North Carolina - September 13-15, 1977
By
Frank Benesh Thomas McCurdy
Environmental Planner Community Planner
GCA/Technology Division Land Use Planning Office
Bedford, Massachusetts Strategies and Air Standards
Division
U.S. Environmental Protection
Agency
23-1
-------
ABSTRACT OF "LAND USE BASED EMISSION FACTORS"
A set of land use based emission factors were developed for
five criteria pollutants and thirteen land use types. The pollu-
tants are particulates, sulfur oxides, carbon monoxide, total hydro-
carbons, and nitrogen oxides; the land uses investigated were five
classes of residential (detached single family, attached single
family, mobile homes, lowrise multifamily, and highrise multifamily),
retail commercial, wholesale commercial, transient lodging, offices,
hospitals, cultural buildings, churches, and schools. Units of the
emission factors for heating degree day or compressor operating
hour. Non-residential emission factor units are given in pounds
of pollutant per square feet of floor area per heating or cooling
degree day. Emission factors are also presented in SI units.
Methods used to develop the emission factors are discussed in
the paper. The factors are applied to one land use type to deter-
mine what project size (in terms of number of dwelling units) would
be a "major source" under EPA's 1976 Interpretative Ruling on the
location of new sources within non-attainment areas. This appli-
cation indicates that land use based emission factors are easy-to-
use.
23-2
-------
INTRODUCTION
Any air resource management approach that includes evaluation
of community land use plans for their ambient air quality impacts
requires that estimates of emissions be devised for the time period
of analysis. Normally this is a time consuming and laborious
process as evidenced by the amount of work that went into the
emission inventories discussed at this conference. It is a process
well beyond the limited expertise and budget of most planning
agencies in this country. Past EPA attempts to standardize and
simplify the procedure, such at its Volume 7 and 13 of the Air
Quality Maintenance Planning Analysis guidelines (1), are still
very complex and tedious. (In fact, an effort is currently under-
way to determine if the procedure should be computerized or not
(2).) If routine evaluation of land use plans for air quality
impacts is to become a reality, a method has to be devised to reduce
the complexity and labor involved especially if urban planners
are the ones who will be doing the evaluation (3).
The Land Use Planning Office of EPA's Strategies and Air
Standards Division has long been interested in developing easy-to-
use land use oriented emission factors. The Office sponsored two
projects in 1972-1973 that developed and used land use based emission
factors: the "Air Pollution-Land Use Planning Project" with Argonne
National Laboratories (4) and the "Hackensack Meadowlands Air Pol-
23-3
-------
lution Study" with a private contractor (5).*
These two efforts treated emissions as a function of land area, viz:
emissions
emission factor = unit land area
Because space heating, process emissions, and the like occur indoors
and have no known consistent relationship with land area per se,
the above formulated emission factor is incomplete because another
piece of information is needed to use it -- density of development
(building area per lot area). To avoid this problem, land use
based emission factors based on building area were developed as
part of the "Growth Effects of Major Land Use Projects" (GEMLUP)
effort (9). These building area emission factors are the subject
of this paper.**
RELATIONSHIP WITH CONVENTIONAL EMISSION FACTORS
Reference 4 has shown the lack of ability of land use based
emission factors to adequately predict emissions of large industrial
sources, i.e., point sources. (It is also shown that other economic
and planning indicators such as employment and output are also
*A critical review of land use emission factors developed in these
projects is found in references 6-7. An elaboration of Hackensack
Meadow!and factors is found in reference 8. How land use data can
be used to improve area source emission inventories is contained
in reference 27.
**A land area emission factor can easily be derived from a building
area factor by multiplying the latter by "floor area ratio," or FAR,
which is a density measure with units of building floor area per
unit lot area. This ratio has been standardized for zoning purposes
by the Federal Housing Administration, and is part of a larger scheme
to relate building areas with open space and parking. See reference
10 for additional information.
23-4
-------
incapable of predicting industrial point source emissions. This is
due, in part, to the variability of emission rates within even fine
industrial categories such as four digit SIC codes). Thus, large
industrial sources must be treated individually when using land
use based emission factors, as is the case when conventional emission
factors are utilized in an emission inventory.
Land use based emission factors can be expected to handle
area source emissions adequately. In the special (and hypothet-
ical) case where the land use categories are identical to the
area source emission categories in a conventional emission inventory,
an emission inventory prepared with land use based emission factors
will be exactly equivalent to a conventially prepared inventory.
In fact, one method of estimating locally specific land use based
emission factors is through a regression analysis of a convention-
ally prepared inventory, such as was done by the California Air
Resource Board (30). Depending on the data available in a specific
application, land use based emission factors may be able to provide
a more accurate emission inventory (both in terms of aggregate
emissions and their spatial distribution) then the conventional
approach. The classification scheme employed in the land use inven-
tory and the availability of accurate fuel consumption data for the
region of interest are principal determinants of which approach will
be more accurate. The utilization of land use based emission factors
would appear to be less time consuming in all instances.
23-5
-------
GEMLUP APPROACH TO DEVELOPING LAND USE BASED EMISSION FACTORS
Land use based emission factors are usually disaggregated into
two components, an activity factor (i.e., fuel throughput, etc.,
per unit floor area), and the standard emission factor (i.e.,
emissions per unit fuel). For example, in the case of fuel oil
space heating consumption, this would be:
emissions (gr) _ oil consump. (gals) * emissions (gr)
floor" area ( 1 03 f tz ) ~ floor area (TCP ft*) oil consump, (gals)
Given this approach, a complete set of land use based emission fact-
ors would consist of an n-dimensional array with specific values
given for a pollutant species, fuel or process type, building cate-
gory, and, in some cases, energy requirements (e.g., region of the
country).
Ignoring solvent evaporation, solid waste disposal, and other
miscellaneous emissions*, an energy consumption related emission
factor can be generalized as follows:
emissions.- ^ , Btu. Btu_. Btu.
T »J»K = TI . i ' '
uz
_ . _ _ _ _^ . .. .
ftz * year uftz * year ftz * ht.d.d. ftz * cl.d.d.;
1 _ ^ 1 * emissions -i
heat content. seasonal ef f i ciency unit fuel^. k
*Emissions from these sources were not considered in the GEMLUP
project, because (1) there is limited information about their
characteristics, and (2) they may be expected to display a lot of
variation in per unit floor area emissions among parts of the
country. However, the emission factor structure discussed above
is amenable to their inclusion.
23-6
-------
Where:
ht.d.d. = heating degree days per year
cl.d.d. = cooling degree days per year
and for a particular fuel type i, pollutant species j
and building category k.
The fourth term in this equation, emissions per unit fuel,
is the commonly used values determined directly from EPA's
Compilation of Air Pollutant Emission Factors (11). Hence, focus
of the GEMLUP project was on generating the first three terms
(i.e., the activity factor).
The activity factor identifies fuel consumption per building
floor area given a number of heating and cooling degree days. The
heat content of fuel in British thermal units is approximately
constant and is well known (12). It does display some variation
for every fuel, especially for natural gas in different regions of
the country (13).
Efficiency values for building types and fuels are less well
known. Efficiency can be defined in a variety of ways, and is used
to account for differences in the amount of energy consumed by a
building depending on the fuel type selected to provide that energy.
This is not heating unit efficiency, which is measured at full load/
steady state operation. Thus, it does not account for rapid on and
off cycling associated with the typical oversized furnace. Nor (in
the case of gas furnaces) does it measure the pilot light fuel
23-7
-------
consumption when the furnace is off.
The desired efficiency measure for GEMLUP purposes was the ratio
of heat loss from a structure to the energy input to the structure.
This is variously defined as efficiency of utilization or seasonal
efficiency. However, even with this definition of efficiency, there
is some disagreement in the literature over what are appropriate
values to use.
The term in brackets, the energy requirement per square foot
and per square foot degree day, represents building energy require-
ments. It is divided into three components:
1. Process use of energy that is not related to climate, such
as:
Lighting Water heating equipment
Elevators Cooking equipment
Refrigeration Ventilation
2 Energy requirements for space heating as a function of heat-
ing degree days.
3. Energy requirements for air conditioning as a function of
cooling degree days.
The energy requirement factor, the efficiency of utilization,
and the standard emission factors are all estimates of the mean of
population values and can be expected to display a large variation.
In general, these factors are not precise indicators of energy
requirement, efficiency, or emissions of a single source. They are
more valid when applied to a large number of sources. Sources of
variation in the energy requirement and efficiency factors are
discussed in the GEMLUP report (9) and will not be discussed here.
23-8
-------
The critical element in the development of land use based emis-
sion factors is estimating energy requirements per square foot for
various building types; the rest of the information needed to develop
the factors is generally available.
Much of the existing literature on energy consumption in build-
ings is not applicable to the development of energy requirement fact-
ors. Most of it is devoted to predicting energy consumption of a
single structure. The literature that was applicable to the GEMLUP
study fell into two classes: (1) typical energy consumption data
for building categories based on engineering estimates, and (2)
average energy consumption information from a sample of structures
in a particular building category. Both classes of literature
were used in GEMLUP analyses with slightly more emphasis given to
the latter category, particularly utility company surveys of customer
energy utilization. Where conflicting data were presented, a compro-
mise value was chosen that seemed to represent an "average" situation.
Not all of the data come from secondary sources. Regression
analyses of office building energy usage were undertaken as part of
the GEMLUP project. Input data for the analyses came from a Building
Owners and Managers Association annual survey of energy use (24).
The regressions yielded very low coefficients of determination, even
though various combinations of independent variables were used (9).
LAND USE BASED EMISSION FACTORS
Land use based emission factors for uncontrolled single family
residences are presented in Table 1. Similar factors for all other
23-9
-------
TABLE 1
SINGLE FAMILY RESIDENTIAL LAND USE BASED EMISSION FACTORS
Pound of pollutant (or kilowatt-hours) per measure
PM SOV CO HC NOV kWh Measure
X A
Space Heating
Electricity - - -
Gas 2.6 x IO"4 1.5 x IO"5 5.1 x IO"4
011 2.2 x IO"3 3.2 x 10"2S 1.1 x 10"3
A1r Conditioning
Central
Electricity - - -
N» Gas 1.8 x IO"4 1.1 x IO"5 3.5 x 10"4
,L Room
0 Electricity - - -
Process
Hot Water
Electricity -
Gas 3.0 x 10"1 1.8 x IO"2 6.0 x IO"1
011 2.5 3.7 x lO^S 1.2
Cooking
Electricity -
Gas 1.1 x IO"1 6.6 x IO"3 2.2 x 10"1
Miscellaneous - - -
3.8 dwelling
2.0 x IO"4 2.6 x TO"3 - dwelling
6.6 x IO"4 2.6 x 10"3 - dwelling
4.7 dwelling
1.4 x 10"4 1.8 x IO"3 - dwelling
unlfht.d.d.
unlfht.d.d.
unlfht.d.d.
uh1fop.hr.
un1fop.hr.
5.1x10 a.c. un1t°operat1ng
hour
1.4xlO+4 dwelling
2.4 x IO"1 3.0 - dwelling
7.5 x IO"1 3.0 - dwelling
3.5xlO+3 dwelling
8.8 x IO"2 1.1 dwelling
7.9xlO+3 dwelling
un1 fyear
unifyear
un1 fyear
unifyear
un1 fyear
uni fyear
Note: A 1600 square foot dwelling unit Is assumed.
S' represents the sulfur percentage of oil, by weight.
-------
land uses investigated in the GEMLUP project are contained in Refer-
ence 9. The factors are presented in units of pounds of pollutant
emitted per "measure" for oil and gas combustion. For electricity
consumption, the factors are in terms of kilowatt-hours per "measure".
The measure, depending on the activity involved, may be square foot
of building floor area per heating degree day, dwelling unit per year,
air conditioner operating hour, etc.
The quantity of secondary, i.e., offsite, emissions occuring
due to electricity consumption depends on the nature of the local
electric utility generating station. In developing an estimate of
secondary emissions caused by a land use type, the local utility
should be contacted to determine the proper emission rate. Default
values for these emissions have been developed, however, and appear
as Table 2. They come from data in references 11,25, and 26.
None of the tables present process emissions for industrial
sources. The estimation and use of land use based emission factors
for industry presents severe problems. The variation in emissions
per square feet of industrial land area, and probably in industrial
building area also, is great (4,7). Therefore, GEMLUP industrial
emission factors only includes fuel combustion and treats these
emissions like area source emissions in a typical emissions inven-
tory. Process emissions have to be obtained separately and added
to fuel use emissions.
23-11
-------
TABLE 2
TYPICAL EMISSION FACTORS FOR ELECTRIC UTILITIES
Pounds of
PM
N>
V
l-» :
K>
coal
oil
gas
5.
6.
1.
23 x
34 x
19 x
ID'3
io-4
ID"4
Kilograms of
coal
oil
gas
6.
7.
1.
PM
59 x
99 x
50 x
io-10
lo-11
ID'11
pollutant emissions
SO
x
1.53 x 10"2S
1.26 x 10"2S
7.13 x 10"6
4
2
2
pollutant emissions
sov
x
1.93 x 10"9S
1.59 x 10"9S
8.98 x 10"13
5
3
2
per kilowatt hour
CO HC
.03 x
.38 x
.02 x
ID'4
ID'4
ID'4
per joule
CO
.08 x
.00 x
.55 x
ID'11
io-11
ID'11
1.21 x
1.58 x
1.19 x
sold to
HC
1.52 x
1.99 x
1.50 x
sold to customer
NO..
io-4
io-4
ID'5
customer
io-11
ID'11
ID'12
X
2.21 x IO"2
8.32 x 10"3
8.32 x IO"3
(SI Units)
NOV
x
2.78 x 10"9
1.05 x IO"9
1.05 x IO"9
Note: A 33.3% overall plant efficiency is assumed for coal fired plants [34].
A 31.6% overall plant efficiency is assumed for oil and gas fired plants [34].
A 10% transmission loss is assumed [35].
'S' and 'A1 represent, respectively, the sulfur and ash percentage of fuel by weight.
-------
APPLICATION OF LAND USE BASED EMISSION FACTORS: DETERMINING " MAJOR
SOURCES" IN NON-ATTAINMENT AREAS
To illustrate how land use based emission factors can be used,
a "real-life" example will be presented. On December 21, 1976
EPA issued "Requirements" (28) and an "Interpretative Ruling" (29)
on the location of new sources within non-attainment areas. The
Ruling states that a major new source may Icoate in an area with
air quality problems only if stringent conditions can be met. The
conditions are: emissions are controlled to the "greatest degree
possible, "equivalent emission reductions will be obtained from
existing sources, and progress toward attaining air quality standards
will be achieved (41 FR_ 55525). A preconstruction review process of
all sources emitting 100 tons of pollutant* .per year or more (1,000
tons for carbon monoxide) is mandated to implement the ruling. If
these "major" sources do not meet the three conditions mentioned
above, a State is not allowed to issue a construction permit under
40 CFR 51.18 (41 FR 55525).
In the other Federal Register notice, EPA proposes to lower
the "major source" definition down to 50 tons of pollutant* per year
(500 tons for carbon monoxide) (41 FR_ 55559). In other words,
facilities emitting 50 tons of most criteria pollutants will be
required to undergo new source review and meet the stringent require-
ments.
*The pollutants of interest are particulate matter, sulfur oxides,
nitrogen oxides, and non-methane hydrocarbons.
23-13
-------
For a private developer 6f land, a question immediately arises:
what size of project will trigger a new source review? Phrased
another way, what is the biggest shopping center (or apartment house,
etc.) a developer can build without emitting 50 tons per year of
SOV or TSP? Land use based emission factors can be used to answer
A
these questions, and doing so constitutes the example to be shown in
the remainder of this paper.
Basically, the answer is obtained by using Table 1 to solve the
following simple equation (or ones similar to it):
sq ft = 50 tons sq ft ht.d.d. yr.
yr tons ht.d.d.
land use based heating
emission factor degree
days.
Additional information needed, or needed to be assumed, to solve
the equation includes: sulfur content of oil, location of the new
source (because of differences in heating/cooling degree days), air
conditioner operating hours, and so forth.
Maps of heating and cooling degree days and air conditioner
compressor operation hours were consulted (9) to choose representa-
tive areas of the country for analysis. A list of a sample of cities
appears as Table 3. Five cities will be used for the analysis:
Atlanta, Bismarck, Boston, Miami, and San Francisco.
The sulfur percentage of #2 fuel oil used for heating and the
coal and oil used in power plants has to be assumed in order to
23-14
-------
TABLE 3
LIST OF REPRESENTATIVE CITIES AND HEATING/COOLING PARAMETERS
ro
V
!-
Ui
»
Cities Heating Degree Day
Albuquerque
Atlanta*
Bismarck*
Boston*
Chicago
Denver
Lincoln
Los Angeles
Miami*
Missoula
New Orleans
New York
Okla. City
Philadelphia
Portland
Raleigh
Reno
Salt Lake
San Antonia
San Diego
San Francisco*
Seattle
St. Louis
Washington
5000
3000
9000
6000
6500
6500
6000
2000
100
9000
1500
5500
3500
5000
7500
3000
6500
6000
1500
1500
3000
5800
4500
4500
Compressor Operating Hour Coolii
600
1600
300
400
600
200
1000
500
2500
100
2000
700
1700
850
300
1300
500
100
2500
500
100
100
1200
1000
1000
1400
500
500
700
500
1000
1000
4000
500
2700
1000
2000
1000
500
1500
500
1000
3000
1000
500
500
1400
1000
*Cities chosen for further analysis
-------
estimate sulfur oxide emissions. The percentages used for the three
items listed in the previous sentence are 0.5%, 2.0%, and 1.0%,
respectively.
Using the above assumptions and locations, emissions from land
uses can be derived for the eleven land use types listed earlier.
Presenting data for all these types would require a lot of space,
so only one example will be given. It focuses on calculating SO
X
emissions for "single family residences" (1600 square feet in area)
having three different energy supply combinations: all gas, all
electric, and part oil and part electric.* The results are presented
in Table 4.
Both "Direct" and "Indirect" emissions are given in Table 1.
Direct emissions are those at the residence itself, while indirect
emissions are those at the power plant supplying electricity to the
house. Indirect emissions are further divided into emissions from
oil burning power plants (Oil) and coal burning power plants (Coal).
Indirect emission factors are found in Table 2.
Multiplying emission factors times the appropriate "measure"
results in an estimate of total yearly SO emissions. For the five
A
*An all gas residence uses natural or synthetic gas for space heat-
ing and cooling, hot water, and cooking. The last two uses are con-
sidered to be "processes," resulting in process emissions. See the
emission factors tables. (All other process emissions are called
miscellaneous emissions, and these are utility plant emissions ass-
ociated with the residence's demand for electrical energy to run
appliances, lights, etc.) An all electric house uses electrical
energy for space heating and cooling, cooking, hot water, and pro-
cesses. Emissions from an all electric residence are all indirect.
A part oil-part electric residence uses #2 fuel oil for space
heating and electricity for everything else.
23-16
-------
TABLE 4
SOX EMISSIONS FROM SINGLE FAMILY RESIDENCES
(Pounds of Sulfur Oxide Emissions Per Dwelling Unit Per Measure)
All Gas
All Electric
Part Oil/Part Electric
Measure
Space Heating
Direct
Indirect:
Oil
Coal
.00002
.04788
.11628
.016
* »
Heating Degree Day
Heating Degree Day
Heating Degree Day
Space Cooling
Direct
Indirect:
Process
Di rect
Indirect:
Oil
Coal
Oil
Coal
.00001
.41
99.54
241.74
__
,05922
.14382
320.04
777.24
.05922
.14382
«» w
320.04
777. 24
Compressor Operating Hours
Compressor Operating Hours
Compressor Operating Hours
Yea-r
Year
Year
Note: Total Yearly Emissions = (Space Heating Emissions)(Heating Degree Day) + (Space Cooling Emissions)x
(Compressor Operating Hours) + (Process Emissions)(Year).
-------
cities previously chosen, estimated total direct yearly SOY emissions
f\
for a single family residence are presented in Table 5. Also given
in the Table is the number of dwelling units needed before a 50 ton
new source review is triggered; the number should be doubled for a
100 ton new source review.
If both direct and indirect emissions of a residence are con-
sidered, then the New Source Review triggering point reduces dramat-
ically, as shown in Table 6. The seemingly anomolous results in the
Table, where SO emissions for a single family residence are higher
A
in warm areas than in (moderately) cold ones, is due to the very
high indirect emissions rates caused by operating air conditioner
compressors. Because this rate is 2.5 times higher for coal fired
power plants than oil fired plants, a high number of compressor hours
(such as in Miami and Atlanta) causes some changes in emissions
ranking among areas depending on type of fuel burned in the plant.
It should be reemphasized that the emission estimated are for
uncontrolled emissions. Because most power plants control SO
/\
emissions to some extent, the emission estimates and triggering
level of New Source Review have to be changed to reflect the amount
of emissions control realized in an area.
SUMMARY & CONCLUSIONS
Land use based emission factors were presented for single
family residences (area sources). The factors were applied to a
problem associated with EPA's evolving New Source Review policy,
23-18
-------
TABLE 5
ESTIMATED DIRECT TOTAL YEARLY SOX EMISSIONS FOR A SINGLE FAMILY
RESIDENCE IN FIVE CITIES AROUND THE COUNTRY
City
Atlanta
All Gas
All Electric
Part Oil, Electric
Bismarck
All Gas
All Electric
Part Oil, Electric
Boston
All Gas
All Electric
Part Oil, Electric
Miami
All Gas
All Electric
Part Oil, Electric
San Francisco
All Gas
All Electric
Part Oil, Electric
SOX Emissions
(Pounds Year"1)
0.49
48.00
0.59
144.00
0.53
96.00
0.44
1.60
0.47
48.00
Number of DUs
Needed Before a 50 Ton
NS Review is Triggered
204,081
N.A.
2,083
169,491
N.A.
694
188,679
N.A.
62,500
227,272
N.A.
62,500
212,765
N.A.
2,083
N.A. = New Source review is not applicable (an infinite trigger).
DUs = Dwelling Units.
23-19
-------
TABLE 6
ESTIMATED DIRECT AND INDIRECT TOTAL YEARLY SQy
EMISSIONS FOR A SINGLE FAMILY RESIDENCE USING OIL HEAT
IN FIVE CITIES AROUND THE COUNTRY
City
SOX Emissions
(Pounds Year "1)
Number of DUs
Needed Before a 50 Ton
NA Review is Required
Atlanta
Oil PP
Coal PP
Bismarck
Oil PP
Coal PP
Boston
Oil PP
Coal PP
Miami
Oil PP
Coal PP
San Francisco
Oil PP
Coal PP
462.79
1,055.35
439.73
930.77
469.69
1,138.39
373.96
839.62
216
94
207
103
227
107
212
87
267
119
Note: PP stands for power plant.
DUs = Dwelling Units
Total Yearly Direct and Indirect Emissions = Direct Emissions
+ Indirect Emissions + (Compressor Operating Hours) (Indirect
Space Cooling Emission Factor).
23-20
-------
and were found to be both tedious and easy to use in practice.
The tedium conies from the numerous steps involved in computing
emissions for a particular location for particular assumptions.
There may be a way to simplify the process some, but it is felt that
combining components or otherwise reducing the steps may result
in too large an error in estimate. The amount of labor involved
in using land use emission factors is still less than that needed
to develop a typical area source emission inventory. Also, the units
and organization of the factors are familiar to urban planners,
which means that they probably will use the factors more readily
than, say, fuel use factors.
While the factors have not been compared to other emissions
estimating procedures, the extensive use of disaggregated engineering
estimates and actual samples of energy consumption gives the authors
a lot of confidence in the factors. However, it is recognized that
a validation of land use based emission factors is in order.
The application of land use based emission factors to a new
source review illustrate their convenience and applicability with
project level data especially with planned projects whose charac-
teristics are still vague. They are also expected to be useful in
reviewing community or municipal scale plans and programs situa-
tions where convential emission inventory techniques are .awkward
or impractical to use.
The land use based emission factors presented in this paper have
23-21
-------
been applied successfully in the following situations.
o To estimate the stationary source CO emissions in an indirect
source air quality evaluation.
o To estimate the sulfur oxide and particulate emissions from
development in the corridor associated with a relocated major
urban arterial highway.
o To evaluate energy consumption associated with several alter-
nate plans for a two city block urban renewal project.
o To estimate residential emissions in an environmental impact
statement for a residential development.
23-22
-------
REFERENCES
1. Office of Air Quality Planning and Standards. Guidelines for Air
Quality Maintenance Planning and Analysis. Research Triangle Park,
North Carolina: Environmental Protection Agency. Volume 7:
Projecting County Emissions (1975), EPA-450/4-74-008; Volume 13:
Allocating Projected Emissions to Sub-County Areas (1974), EPA-450/
4-74-014; "Accounting for New Source Performance Standards in
Projecting and Allocating EmissionsHypothetical Example" (1975),
a supplement to Volume 13; and Volume 13: Appendices A and B (1975);
EPA-450/4-74-014a.
2. Richard R. Cirillo and Michael J. Senew. Development of Computer-
ized Emmissions Projection and Allocation SystemPhase I: Prelim-
inary Feasibility Study. Research Triangle Park, North Carolina:
Environmental Protection Agency, (EPA-450/3-77-001).
3. Peter C. Cosier, IV. "Land Use Based Emission Stretegies: Their
Promise and Problems", Planning Comment (November 1976) 12:31-48.
4. A. S. Kennedy, et al. Air Pollution Land Use Planning Project;
Volume II: Methods for Predicting Air Pollution Concentrations
from Land Use. Research Triangle Park, North Carolina: Environ-
mental Protection Agency, 1973. (EPA-450/3-74-028b)
5. John C. Goodrich. Hackensack Meadow!ands Air Pollution Study-
Emission Projection Methodology. Research Triangle Park, North
Carolina: Environmental Protection Agency, 1973. (EPA-450/3-74-
56b)
6. Bill Swindaman. "Notes on Land Use Based Emission Factors",
Unpublished paper available at the Land Use Planning Office,
Strategies and Air Standards Division, U.S. Environmental Pro-
tection Agency, June 1972.
7. Thomas E. Baldwin and Allen S. Kennedy. "The Feasibility of
Predicting Point Source Emissions Using Industrial Land Use
Variables: A Path Analysis", presented at the 67th Annual Meet-
ing of the Air Pollution Control Association, 1974. (APCA paper
#74-145)
8. John C. Goodrich and Byron H. Willis. "A Methodology for Deter-
mining Emissions from Land Use Planning Data", presented at the
65th Annual Meeting of the Air Pollution Control Association, 1972.
CAPCA paper #72-122)
23-23
-------
9. Frank Benesh. Growth Effects of Major Land Use Projects; Volume
IICompilation of Land Use Based Emission Factors. Research
Triangle Park, North Carolina: Environmental Protection Agency,
1976. (EPA-450/3-76-012b)
10. J. Ross McKeever (ed.). The Community Builders Handbook. Washing-
ton, D.C.: Urban Land Institute, 1968.
11. Office of Air Quality Planning and Standards. Compilation of Air
Pollutant Emission Factors. 2nd Edition. Research Triangle Park,
North Carolina: U.S. Environmental Protection Agency, April 1973,
and supplements. (AP-42)
12. See, for example, the Keystone Coal Industry Manual. New York:
Mining Information Services of McGraw Hill, 1969.
13. James Couillard. Browns Directory of North American Gas Companies,
Duluth, Minnesota: Harcourt Brach Jovanovich, 1973.
14. Arthur D. Little, Inc. Project Independence Blueprint: Volume I.
Washington, D.C.: Federal Energy Administration, 1974.
15. Marketing Division. Residential Appliance Gas Consumption, Phase
4^ Lincoln, Nebraska: Northern Natural Gas Company, 1973.
16. R. Anderson. Residential Energy Consumption: Single Family
Housing Final Report. Washington, D.C.: U.S. Department of
Housing and Urban Development, 1973.
17. American Gas Association. Gas House Heating Survey (Annual).
Arlington, Virginia: American Gas Association, various years.
18. American Gas Association. Info Data Sheet: Use of Gas by Resi-
dential Appliances. Arlington, Virginia: American Gas Associ-
ation, 1971.
19. Hittman Associates. Residential Energy ConsumptionMultifamily
Housing. Washington, D.C.: U.S. Department of Housing and Urban
Development, 1974.
20. Electric Heating Association. EHA Case History. New York: Elec-
tric Heating Association, various cases.
21. Gordian Associates. Environmental Impact of Electric vs. Fossil
Fuel Space Heating for the Welfare Island Development Project.
New York: New York State Urban Development Corporati-on, 1972.
23-24
-------
22. H. Hansteen and J. Kirkwama. "The Fossil Electric Ratio", paper
presented at the Annual Meeting for the American Society of
Mechanical Engineers, 1968. (ASME paper #68-WA/PEM-3)
23. American Society of Heating, Refrigerating, and Air Conditioning
Engineers. ASHRAE Heating. Ventilating, and Air Conditioning
Guide. New York: ASHRAE, 1958.
24. Building Owners and Managers Association International. 1975
Office Building Experience Exchange Report for the Calendar Year
1974. Chicago: BOMA, 1975.
25. National Coal Association. Steam Electric Plant Factors. Wash-
ington, D.C.: NCA, 1973.
26. Edison Electric Institute. Statistical Yearbook of the Electric
Utility Industry, New York: EEI, 1972.
27. Kenneth A. Hagg, et al. "Maintenance Planning in Massachusetts:
Use of Land Cover Factors for Apportionment and Projection of
Areas Source Emissions", paper presented at the 68th Annual Meet-
ing of the Air Pollution Control Association, 1975. (APCA paper
#75-22.8).
28. 41 Federal Register 246: 55558 (December 21, 1976). The notice
is entitled "Review of New Sources and Nodifications" and is part
of 40 CFR 51.
29. 41 Federal Register 246; 55524 (December 21, 1976). The notice
is entitled "Air Quality Standards: Interpretative Ruling" and is
related to 40 CFR 51.
30. Evaluation and Planning Division, California Air Resources
Board. The Land Use/Oxidant Precursors Emissions Study( DRAFT).
Sacramento, CA.: A.R.B., 1976.
23-25
-------
QUESTION:
MC CURDY:
CONDENSED DISCUSSION
On these housing developments£ what new source
review are you looking at from a standpoint
that a new housing development will bring in
new cars and put out new hydrocarbons and
whether they should be subject to the new
source review?
That's an interesting question because under
indirect source review you were supposed to
have done that; but indirect source review
as you probably know is dead for all types
of land use except for highways and airports.
This growth effects the major land use projects.
It's specific aim was to develop a procedure
where we can estimate the amount of secon-
dary development caused by the construction
and operation of major land use types. The
first two types we looked at were residential
development and industrial parks. You put in
an industrial park and what happens to the
urban structure around it? What happens to
VMT, etc.? We have a simple predictive
methodology which is based on multiple re-
2
gression again, and we are quite high r 's.
23-26
-------
2
I was appalled at the r 's that were mention-
ed earlier. Our's are .75 - .54. At any
rate our office is looking Into that. Right
now we have a project to do exactly the same
thing for sewage treatment plants. If you
put in a sewage treatment plant, what is
going to happen next? We are developing
easy to use procedures so that if you knew
what size treatment plant you had or what
size inceptor you had, we could estimate
the amount residential development induced
by and associated with this major project.
I don't think EPA is going to have a policy
saying that you have to do this for all
development types.
23-27
-------
EMISSION RATES FOR BIOGENIC NO
A
Presented at the 1977
Environmental Protection Agency
Emission Inventory/Factor Workshop
Raleigh, North Carolina - September 13-15, 1977
By
Hilman C. Ratsch and David T. Tingey
Terrestrial Ecology Branch
Corvallis Environmental Research Laboratory
U. S. Environmental Protection Agency
Corvallis, Oregon 97330
24-1
-------
Abstract
A literature review of biogenic sources of NO was conducted to
A
determine their emission rates into the atmosphere. NO are some of
X
the products of microbial denitrification, chemical decomposition of
nitrites and the oxidation of organic nitrogen compounds. There
appears to be no significant emission of NO from either oceans or
A
freshwaters. Biogenic emission rates for NO and N02 from soil range
from 0.015 to 0.02 kg NO km-2 hr-1 and 0.01 to 0.2 kg N02 km-2 hr-1.
Submerged soils, sediments, marshes and swamps could be sources of NO
but emission data are not available. There is no significant evidence
of NO emission from living vegetation. During decay, decomposition
A
and ensiling of vegetation, NO can be formed. Although the emission
a
rates are not known, they are probably not significant.
The estimates of NO emission rates from the above biogenic (soil)
sources were computed for a global basis and then compared closely to
previously estimated natural global emissions. Also, the background
atmospheric concentrations of NO are similar to those levels
predicted from biogenic emission rates.
24-2
-------
Introduction
Several authors have Estimated global biogenic NO emissions.
A-
Robinson and Robbins (1970) estimated global natural emissions of NO
A
at 768 x 109 kg N02 which was approximately 15 times the anthropogenic
emissions (53 x 109 kg N02). McConnell (1973) suggested that the
natural sources of NO were 4 times the anthropogenic sources.
A
Galbally (1975) estimated biogenic NO emissions for the northern
A
hemisphere at 1 x 109 kg N02 yr-1 and the anthropogenic emissions at
0.5 x lo9 kg N02 yr-1. Estimates of biogenic emissions for NO for
A
Ohio and surrounding states, ranged from 1.7 to 4.1 x 108 kg N02 yr-1
and the anthropogenic emissions ranged from 2.5 to 33.3 x 10s kg N02
yr-1 (RTI, 1974). These calculated NO emissions were based on
atmospheric concentrations and a theoretical balance of nitrogen
compounds. At present there is no general agreement on the relative
contributions of the biogenic and anthropogenic emissions. Most
biogenic emission rates were derived from the amounts of NO needed to
A
balance nitrogen cycles or were deduced theoretically.
The objectives of this literature review were to gather infor-
mation on possible biogenic sources of NO , to determine biogenic
A
emission rates and to discuss the factors affecting NO emissions. The
A
nitrogen cycle is discussed to suggest possible biogenic sources of
24-3
-------
nitrogen oxides. The biogenic sources and emission rates of NO are
divided into the following categories: 1) water: ocean and
freshwaters; 2) soil; 3) flooded soil, sediments, swamp and marsh; and
4) vegetation. The biogenic emission rates determined from the
literature were compared to other emission estimates and atmospheric
concentrations.
24-4
-------
Nitrogen Cycle
Nitrogen is found in five major sinks in the biosphere: primary
rocks, sedimentary rocks, the deep-sea sediment, the atmosphere, and
the soil-water pool. Approximately 98% of the earth's nitrogen is in
primary and sedimentary rocks, while the atmosphere contains about 2%;
the deep-sea sediment and soil-water pool together contain less than a
percent of the global nitrogen (Burns and Hardy, 1975).
In the atmosphere, N2 is the major nitrogen constituent while
nitrogen oxides (N20, NO, N02), NOs, N02, NHg, and NH4 are present in
the ppm range or less. The soil-water pool can contain a large amount
of dissolved N2. N20, NO, and N02 can also occur in soils for short
time periods under specific environmental conditions.
A simplified nitrogen cycle showing nitrogen transformations in
the soil-water pool and transfers between the soil-water pool and the
.atmosphere is shown in Figure 1. Nitrogen enters the soil-water pool
through biological nitrogen fixation, industrial fixation, precipi-
tation and application of fertilizers. Biological fixation reactions
occur either in free-living organisms (i.e. Azotobacter) or in
symbiotic plant-microbial associations (i.e. Rhizobium) . Plants and
microorganisms utilize NH4, NOs, and N02 from the soil-water pool and
nitrogen undergoes a multitude of chemical and biological trans-
formations.
24-5
-------
NH,
Precipitation
(105)
Biological
Nitrogen
Fixation ,
(135) I Industrial
Fixation
(30)
NO; NO;
Fertilizer*
(40)
Soil Organic
Material
N,
Outgasslng)
Volatili-
zation
(130)
N20
NO
Denitrlfication
(140)
NO N02
NH,
Combustion
(20)
N02
Nitrification
Leaching
^ (15)
Fig 1. Terrestrial nitrogen cycle
transfer and transformations.
showing. nitrogen
Values given are
metric tons (N) x 106 yr-1 (Burns and Hardy,
1975; Hardy and Havelka, 1975).
24-6
-------
In nitrification (Figure 1), an aerobic process, ammonium is
oxidized to nitrate.
Oxidation occurs mainly by the autotrophic bacteria of the Nitro-
bacteriacea. The genus Nitrosomonas oxidizes ammonium to nitrite and
the genus Nitrobacter oxidizes nitrite to nitrate. Other micro-
organisms , including certain bacteria , molds and fungi , are capable of
limited oxidation of nitrogen, but their contribution to nitrification
is limited. The rapid nitrification of ammonium is important
agriculturally, since fertilizer ammonium can be rapidly oxidized to
nitrate and then lost from the soil by denitrification, leaching and
chemical decomposition (Hauck, 1971).
In denitrification (Figure 1), an anaerobic metabolic process,
nitrate is reduced sequentially to NO, N20 or N£.
NOg -> N02 -» NO -* N20, N2
The reduction is carried out by a diverse group of bacteria but the non
spore formers such as Pseudomonoas , Micrococcus, and Achromobacter and
spore-forming species of Bacillus are the principal denitrifiers.
24-7
-------
Biogenic Sources of^ Nitrogen Oxides
Water
Oceans
The nitrogen cycle in the oceans is complex in respect to the
large geographical translocations of nitrogen and in species
composition of micro- and macro-organisms (Dudgale, 1969).
Theoretical models of nitrogen circulation in oceans and mass balances
for marine nitrogen cycles indicate that inorganic nitrogen is removed
or lost from the oceans (Yoshinari, 1976; Dugdale, 1969).
Denitrification is a significant factor in loss of nitrogen from the
marine environment, especially in oxygen-deficient waters. Goering
and Cline (1970) determined that denitrification in raw seawater
occurred in two stages. First, nitrate was reduced to nitrite, and
second, the nitrite was further reduced, presumably to N£. Nitrate15-N
added to water samples from the oxygen-deficient layer in the tropical
Pacific resulted in the production of nitrite and N£- Denitrification
rates varied from 7-150 |Jg N/l/day depending on sample depth (Goering,
1968). Estimated denitrification losses in oxygen-deficient waters of
the Black Sea (7 x 109 - 2 x 1011 g N yr-1) and Cariaco Trench (1 x 1010
g N yr-1) were not significant compared to those in the eastern
tropical North Pacific (1 x 1013 g N yr-1) (Goering et al., 1973).
Patriquin and Knowles (1974) examined shallow-water marine sedi-
ments from several locations for denitrifiers and found that most of
the bacteria that reduced nitrate to nitrite also reduced nitrite to
24-8
-------
gaseous nitrogen. Bacterial isolates varied considerably in the rate
of N£ production and N£0 reduction, and in accumulation of ^0.
Barbaree and Payne (1967) demonstrated that N20 and very small
quantities of NO were present transiently in the atmosphere over
reaction mixtures containing cell-free extracts of Pseudomonas
perfectomarinus cells.
The above data support the conclusion that denitrification is an
active process in the oceans and that both N£ and ^0 are products but
there is no evidence of NO production in oceans.
Freshwater
As in the oceans, denitrification in freshwaters is a significant
factor in nitrogen loss. The nitrogen cycle and the fate of nitrogen
in freshwaters has been reviewed recently (Keeney, 1973; Brezonik,
1973).
In lakes, denitrification occurs mainly in the oxygen depleted
hypolimnic layers where inorganic nitrogen levels are at a minimum
(Vollenweider, 1968). Denitrification rates in Smith Lake, Alaska, in
water one meter below the ice were measured at 15 (Jg N/liter/day.
Molecular nitrogen appeared to be the only significant product and NO
and N20 were not detected (Goering and Dugdale, 1966). Brezonik (1973)
concluded that denitrification did not appear to be significant in
Florida lakes. Of the 55 north central lakes sampled in Florida, only
four developed anoxic conditions at the bottom and the nitrate
24-9
-------
concentrations in Florida lakes are typically low. Nitrogen is the
only product of denitrification released in significant quantity, and
seldom is NO or N20 detected (Payne, 1973).
In addition to denitrification, there is the possibility of
chemical decomposition of nitrite to produce gaseous nitrogen prod-
ucts. In some lakes, concentrations of polyphenolic substances
(tannins, lignins, humic acid) are high and pH of water is acidic.
Under these conditions the reaction of nitrous acid with organics could
be important as a source of nitrogen oxides (Brezonik, 1973).
Soil
The soil is an open system from which various nitrogen forms
volatilize into the atmosphere. As early as 1944, it was discovered
that nitrous oxide was one of the constituents of soil air (Kriegel,
1944). Adel (1946, 1951) suggested that N20 was produced in the soil
and was the source of atmospheric N20. Arnold (1954) confirmed N20
losses from the soil and Wijler and Delwiche (1954) found that N20 was
the major initial product of denitrification under some soil
conditions, but also identified NO as an additional product. More
recently, in addition to losses from denitrification, investigators
have found that NO and N02 are produced by chemical reactions in soil
involving HN02 and nitrite. (Reuss and Smith, 1965; Nelson and
Bremner, 1970; Bollag, etal., 1973).
24-10
-------
Denitrification
The terra "denitrification" implies the gaseous loss of nitrogen
to the atmosphere, usually as N2, N20 or NO, through some biological
agency and it is the subject of two recent reviews (Payne, 1973;
Delwiche and Bryan, 1976). In the absence of oxygen, microorganisms
use nitrate or its reduction products as electron acceptors for the
oxidation of some organic compound as an energy-yielding reaction. The
intermediates and products of denitrification have been studied ia
cultures and in soil experiments, but the pathways are varied and not
presently elucidated. One possible pathway is as follows:
NOs -» N02 -» NO » N20 » N
Many genera of bacteria are able to reduce nitrate to nitrite and
of those a limited number are able to further reduce nitrite to N20 or
N2. Studies have determined that NO is a specific product of nitrite
reduction; nitrous oxide (N20) results from NO reduction and is the
terminal denitrification product of several bacterial strains (Renner
«
and Becker, 1970; Payne, Riley and Cox, 1971).
Factors Affecting Denitrification. Denitrification is influenced
by factors such as soil oxygen concentration, redox potential, pH, soil
organic matter content and temperature. In general, conditions that
directly or indirectly decrease the soil oxygen content or increase
microbial activity increase denitrification. When soil oxygen level
decreased from 8.5% to 1%, N20 production increased, but as the oxygen
level decreased to zero the N20 was reduced to N2 (Cady and
24-11
-------
Bartholomew, 1961). Although it was not measured, a similar response
would be expected for NO. In a closed soil-plant system (Stefanson,
1972) N2 was the major component of denitrification but in the absence
of plants the main product was N20. Over the temperature range of 15
to 35°C, increasing the temperature 10°C doubled the rate of
denitrification (Stanford et al., 1975). Bailey (1976) reported that
as soil temperature increased the rate of N2 production increased and
NO production decreased. Denitrification capacities of 17 surface
soils were significantly correlated with total organic carbon content
and very highly correlated with water-soluble organic carbon or
mineralizable carbon (Burford and Bremner, 1975). Nitric oxide was
detected in the atmosphere of several soils incubated at 20°C for 7
days, but the amount represented not more than 0.1% of the nitrate
lost. As the soil pH increased from 5 to above 7, the ratio of N20 to
*.
N2 decreased (Burford and Bremner, 1975); below pH 7.0, N20 was the
main product (Wijler and Delwiche, 1954). The pH dependency of NO
Ju
production should be similar to N20. In general, low temperature, low
pH, and marginal anaerobic conditions favor the production of N20
relative to N2 (CAST, 1976).
Chemodenitrification or Nonbiological Chemical Decomposition
In addition to microbial denitrification there is increasing
evidence that non biological reactions produce nitrogen gas or
nitrogen oxides under some circumstances (Delwiche and Bryan, 1976;
24-12
-------
Porter, 1971). Non biological loss of nitrogen may result from "side-
tracking" during nitrification and denitrification processes. This
can occur when an intermediate in the process (i.e. N02) is produced
more rapidly than it can be oxidized or reduced biologically and
undergoes chemical decomposition (Lance, 1972). There are several non
biological reactions that could release NO into the atmosphere.
Nitrosation denotes the addition of the nitroso group (-N = 0) to
an organic molecule, and is brought about by HN02 and other compounds
to form an organic complex (0 = N = X). The nitroso groups formed are
labile and react further with the nitrosating agent to form gaseous
products.
HN02 * NO- + OH
NO-+R-»N = 0-R
N + 0 - R +"A -> NO + A - R
Stevenson, et al. (1970) showed evidence that NO, N20, and N2 could be
produced by nitrosation of humic and fulvic acids, lignins, and
*
aromatic substances at pH 6.0 and 7.0 in the absence of oxygen. Steen
and Stojanovic (1971) found that NO was volatilized from a calcareous
soil when high concentrations of urea were nitrified with concurrent
accumulation of nitrite and assumed that nitrosation between nitrous
acid and organic matter was the main pathway by which NO was formed.
i
Wullstein and Gilmour (1964, 1966) reported that nitrite reacted
with certain reduced transition metals in sterile, moderately acid
systems to yield NO as a primary gaseous product. In their proposed
24-13
-------
scheme, N02 and NO reacted with metallic ions to form complexes which
were either stable or reactive. The metals could also react directly
with N02 or NO without forming complexes or intermediates to form NO or
N2. They concluded that copper and manganese ions in the soil were
primarily responsible in reacting with nitrite. Nelson and Bremner
(1970) found that Cu+, Sn2+ and Fe2* were the only metallic cations
that promoted nitrite decomposition and that soils normally do not
contain sufficient amounts of these cations under conditions suitable
for chemodenitrification to be significant in nitrite decomposition.
Reuss and Smith (1965) showed that N2 and N20 are formed by
chemical decomposition of nitrite in soil and also showed that N©2 was
evolved when nitrite was added to acid soils. Bremner and Nelson
(1968) found that N2 and N02 and small amounts of N20 were formed when
nitrite was added to neutral"and acid soils. They suggested that the
reactions between soil organic constituents and nitrite were respon-
sible for the formation of N2 and N20, while self-decomposition of HN02
*
was responsible for the formation of NO and N02- In the steam-
sterilized raw humus samples incubated with nitrite, NO was the
predominant gaseous reaction product (Nommick and Thorin, 1971).
Nelson and Bremner (1970) found that the formation of N02 by
decomposition of nitrite in pH 5.0 solution was not promoted by organic
)
and inorganic soil constituents, and concluded that most of the N02
evolved was formed by self-decomposition of HN02- The amount of N02
formed was inversely related to soil pH, but pH had little effect on
24-14
-------
the amount of nitrite converted to nitrate. These findings led to the
conclusion that the self-decomposition reaction of HN02 was best
represented by the equation:
2HN02 -> NO + N02 + H20
Measurements of NO and N02 emissions from soil and biological and
non biological reactions are difficult and only limited data are
available (Table 1). Soil emission rates were determined by directly
trapping liberated gases at the soil surface. Based on the limited
number of studies, NO is produced at lower rates (0.004 to 0.02 kg NO
km-2 hr-1) than N02 (0.01 to 0.2 kg N02 km-2 hr-1). These rates are
temperature dependent and would be expected to increase as the
temperature increases (Bailey, 1976).
Flooded Soil, Sediments, Swamp and Marsh
Flooded soil or sediments have characteristics that are unique
and separate them from arable soil, such as the interruption of gaseous
exchange between air and soil (Patrick and dikkelsen, 1971). The
restriction of oxygen diffusion results in an oxidized layer of up to
one cm. in the soil-water interface, but below this layer the oxygen
content declines rapidly. A second characteristic of flooded soil is a
)
change .in microbial forms from aerobic to facultative anaerobic
organisms to anaerobic bacteria. Retarded metabolic processes are
reflected in reduced organic matter decomposition and a lowered
nitrogen requirement for decomposition. In terms of physiochemical
24-15
-------
changes, the pH of flooded soils tends to change toward neutrality,
redox potentials are low (-300 mv) and there is an increase in the
amount of ions in the soil solution (Ponnamperuma, 1972).
In submerged soil, inorganic nitrogen is present almost
exclusively as NH4 because the lack of oxygen prevents the
nitrification of NH4 to N03. However, the NOs formed in the aerobic
layer at the sediment-water interface diffuses downward to the
anaerobic layer where denitrification occurs. As in soils,
denitrification (biological and non-biological) is the major source of
nitrogen oxides from flooded soils, sediments, swamps and marshes.
^
Oxygen content, pH, redox potential, temperature, nitrate content and
organic matter content of submerged soils are factors that affect the
amount and products of denitrification.
Denitrification as measured by nitrate and nitrite reduction
rates and Nj formation is significant in submerged soil and sediments
(Chen, Keeney, Konrad, Holding, Graetz, 1972; Chen, Keeney, Graetz and
Holding, 1972; Reddy and Patrick, 1975; Engler and Patrick, 1974;
Goering and Dugdale, 1966). The disappearance of nitrate in sediments
and submerged soils is usually accompanied by the production of KT2 but
recently several workers have shown other denitrification products.
Lake sediments incubated with nitrate and nitrite produced N20 in
addition to N2 (Macgregor and Keeney, 1973; Chen, Keeney^ Konrad,
Holding, Graetz, 1972). In the decomposition of nitrite in flooded
soils, N£ was produced at all pH's but NO and ^0 were produced only at
24-16
-------
pH 6.0 and below (Van Cleemput, et al. 1976). The addition of a soil
sterilant (HgCl2) increased the rate of NO production. The important
conclusions from this study were that under acid conditions
significant amounts of N2, ^0 and NO were formed. The production of
NO under acid conditions with and without a sterilant suggested the
self-decomposition of nitrous acid similar to what occurs in arable
soils. However, the data are not adequate to calculate emission
factors.
Vegetation
Plant leaves and roots can absorb both reduced or oxidized forms
of nitrogen from the environment and a relatively large concentration
of nitrogen compounds are found in plants. However, there is no direct
evidence that any of this nitrogen is emitted into the atmosphere as
NO . In contrast during decomposition NO can be emitted into the
A A
atmosphere but its significance is not known. During the ensiling
process high concentrations of nitrogen oxides can be emitted. In an
unventilated silo or enclosure, NO and N0£ can reach hazardous levels
and such levels have accounted for several fatalities (Commins, et al.,
1971; Scaletti, et al., 1960). In decomposition of plant products only
a small fraction of total nitrogen losses from the system are
attributed to denitrification. In an eastern mature hardwood forest,
an estimated 3.6% of the total nitrogen flux was lost through
denitrification processes (Mitchell, et al., 1975). Estimated
24-17
-------
denitrification rates of 0.17, 1.61 and 0.08 kg N ha-1 yr-1 were
measured for branches, logs and litter (H layer), respectively (Todd,
et al., 1975). It is conceivable that under some conditions small
amounts of nitrogen oxides are emitted during litter decomposition.
However, data on the emission rates from decomposition are lacking.
Estimates of Global NO Emission
The source of NO in the atmosphere is both anthropogenic and
A
biogenic. Estimates of anthropogenic emissions of NO on a global
A
basis are in good agreement as shown in Table 2. But the biogenic
emissions are varied and are much more difficult to identify, measure,
or estimate.
The biological and chemical transformations of nitrogen compounds
in the soil appear to be the major source of NO . The estimates of NO
X A
emission rates from soil (Table 1) range from 0.01 - 0.2 kg N k
This rate computed on a global scale gives an emission from 3 to 58 x
109 kg N yr-1 and compares closely to the estimated global NO
A
emissions by Galbally (1975) and Soderlund and Svensson (1976).
The background concentrations of NO and N02 reported by
Rasmussen, et al. (1975) is about 0.3 - 2.5 |Jg m-3 and 2 - 2.5 M8 m~3>
respectively. Soderlund and Svensson (1976) reported background
concentrations of NO to range from 0.5 - 7.5 M8 ffl3 (as ^2) depending
A
on geographical location. If it is assumed that background NO is in a
A
steady state condition, then NO deposition must be balanced by NO
A A
24-18
-------
emission. The total wet and dry deposition of NO for the terrestrial
system was estimated at 32 to 83 x 109 kg N yr-1 and for the aquatic
system at 11 to 33 x lo9 kg N yr-1 (Soderlund and Svensson, 1976).
Peterson (1977) estimated that NO is removed by wet and dry deposition
A
processes at about 0.04 kg N km^hr-1. This value is similar to soil
NO emission rates given in Table 1. Therefore, background
A
concentrations of NO can be explained in large part by biogenic NO
X X
emissions.
24-19
-------
Table 1. Nitrogen Oxide Flux From Soil
Soil or
Medium
N Type N
Addition Oxide
kg/km2/hr. Reference
Emission Rate
low humic NH4 N03
Ordinary Chernozem urea
Chernozem
Podzol
NH4 N03
100 Kg.
N/ha
top soil from pine, none
oak, sod stand
(sandy loam)
NO
NO
N02
.015-.02
.006
.004
Getmanets, 1972
Borisova, et al.,
1972
Sod- Podzolic
NH4 N03 N02
Sod - Podzolic
medium clay loam
urea N02
240 Kg/ha
pine 0.125
oak 0.07
sod 0.11
0.01-0.2
comment: vari-
ation during
the growing
season
0.03-0.05
Kim, 1973
Makarov, 1969
* emission rate calculated as N02
Makarov and
Ignatova, 1964
raw humus
in spruce
stand
(feather
moss)
raw humus
(sphagnum
moss)
KN03 *(NO+N02)
NH4 N03
Ca (N03)2
Al (N03)3
NH4 N03 *(NO+N02)
Ca (N03)2
Al (N03)3
.018
.029
.035
.104
.012
.018
.035
Mahendrappa ,
1974
Mahendrappa ,
1974
24-20
-------
Table 2. Estimates of Global NO Emission
A
Sources of NO (109
A.
Anthropogenic
16
18
15
19
kg N yr-1)
Biogenic
234
* 72
20
8-25
3-58
References
Robinson and Robbins, 1970
McConnell, 1973
Galbally, 1975
Soderlund and Svensson, 1976
estimated from data in Table 1.
* natural sources estimated at 4 or more times the anthropogenic
sources.
24-21
-------
REFERENCES
Adel, A. 1946. A possible source of atmospheric N20. SCIENCE 103,
280.
Adel, A. 1951. Atmospheric nitrous oxide and the nitrogen cycle.
SCIENCE 113, 624-625.
Arnold, P. W. 1954. Losses of nitrous oxide from soil. J. SOIL SCI.
5, 116-128.
Bailey, L. D. 1976. Effects of temperature and root on
denitrification in a soil. CAN. J. SOIL SCI. 56, 79-87.
Barbaree, J. M. and Payne, W. J. 1967- Products of denitrification by
a marine bacterium as revealed by gas chromotography. MAR. BIOL.
1, 136-139.
Bollag, J. M., Drzymala, S. and Kardos, L. T. 1973. Biological versus
chemical nitrite decomposition in soil. SOIL SCI. 116, 44-50.
Borisova, N. N., Burtseva, S. N., Rodionov, V. N. and Kirpaneva, 0. L.
1972. Determination of nitrogen losses from soil in the form of
different oxides and ammonia under field conditions. SOV. SOIL
SCI. (4), 540-546.
Bremner, J. M. and Nelson, D. W. 1968. Chemical decomposition of
nitrite in soils. 9th TRANS. INT. CONGR. SOIL SCI. 2, 495-503.
Brezonik, P. L. 1973. Nitrogen sources and cycling in natural waters.
EPA-660/3-73-002 Ecological Research Series, U. S. Environmental
Protection Agency, p. 7-19.
Burford, J. R. and Bremner, J. M. 1975. Relationships between the
denitrification capacities of soils and total, water-soluble and
readily decomposable soil organic matter. SOIL BIOL. BIOCHEM. 7,
389-394.
Burns, R. C. and Hardy, R. W. F. 1975. Nitrogen fixation in bacteria
and higher plants. Springer - Verlag, New York, 189 pp.
24-22
-------
Cady, F. B. and Bartholomew, W. V. 1961. Influence of low p(>2 on
denitrification processes and products. SOIL SCI. SOC. AM PROC.
25, 362-365.
CAST, 1976. Effect of increased nitrogen fixation on stratospheric
ozone. Rep. 53. Iowa: Council for Agricultural Science and
Technology, 33 pp.
Chen, R. I., Keeney, D. R., Konrad, J. G., Holding, A. J., and Graetz,
D. H. 1972. Gas production in sediments of Lake Mendota, Wise.
J. ENVIRON. QUAL. 1(2), 155-158.
Chen, R. L., Keeney, D. R., Graetz, D. A. and Holding, A. J. 1972.
Denitrification and nitrate reduction in Wisconsin lake
sediments. J. ENVIRON. QUAL. 1(2), 158-162.
Commins, B. T., Raveney, F. J. and Jesson, M W. 1971. Toxic gases in
tower silos. ANN. OCCUP. HYG. 14, 275-283.
Delwiche, C. C. and Bryan, B. A. 1976. Denitrification. ANN. REV.
MICROBIOL. 30, 241-262.
Dugdale, R. C. 1969. The nitrogen cycle in the sea. BIOLOGY AND
ECOLOGY OF NITROGEN. Conf. Proc. University of California, Davis,
NAS, Washington, D. C. p. 16-18.
Engler, R. M. and Patrick, W. H. 1974. Nitrate removal from
floodwater overlying flooded soils and sediments. J. ENVIRON.
QUAL. 3(4), 409-413.
Galbally, I. E. 1975. Emission of oxides of nitrogen (NO ) and
ammonia from the earth's surface. TELLUS 27, 67-70. X
Getmanets, A. Y. 1972. Loss of mineral fertilizer nitrogen from soil
in gaseous form. SOVIET SOIL SCI. 4, 172-175.
Goering, J. J., Richards, F. A., Godispoti, L. A. and Dugdale, R. C.
1973. Fixation and denitrification in the ocean. In E. Ingerson
(Ed.) Proc. Intern. Symp. Hydrogeochim. Biogeochem. Vol. II.
Clarke, Co, Washington, D. C. pp 12-27.
Goering, J. J. 1968. Denitrification in the oxygen minimum layer of
the eastern tropical Pacific ocean. DEEP SEA RES. 15, 157-164.
Goering, J. J. and Cline, J. D. 1970. A note on denitrification in
seawater. LIMNOL OCEANOGR. 15, 306-309.
24-23
-------
Goering, J. J. and Dudgale, V. A. 1966. Estimates of the rates of
denitrification in a subarctic lake. LIMNOL. OCEANOGR. 11, 113-
117.
Hardy, R. W. F. and Havelka, V. P. 1975. Nitrogen fixation research:
a key to world food? SCIENCE L88, 633-643.
Hauck, R. D. 1971. Quantitative estimates of nitrogen-cycle
processes. Nitrogen-15 in soil plant studies, (Int. Atomic
Energy Agency, Vienna)pp. 65-80.
Keeney, D. R. 1973. The nitrogen cycle in sediment-water systems. J.
ENVIRON. QUAL. 2(1), 15-29.
Kim, C. M. 1973. Influence of vegetation types on the intensity of
ammonia and nitrogen dioxide liberation from the soil. SOIL BIOL.
BIOCHEM. 5, 163-166.
Kriegel, M. W. 1944. Analysis for hydrocarbons in the presence of
nitrous oxide. GEOPHYSICS 9, 447-462.
Lance, J. C. 1972. Nitrogen removal by soil mechanisms. J. WAT.
POLLUT. CONT. FED. 44, 1352-1361.
Mahendrappa, M. K. 1974. Volatilization of oxides of nitrogen from
nitrate-treated Black Spruce raw humus. SOIL SCI. SOC. AM. PROC.
38, 522-524.
Makarov, B. N. 1969. Liberation of nitrogen dioxide from soil.
SOVIET SOIL SCI. 1, 20-25.
Makarov, B. N. and Ignatova, V. P. 1964. Losses of nitrogen from the
soil in the gaseous form. SOVIET SOIL SCI. (4), 407-413.
McConnell, J. C. 1973. Atmospheric ammonia. J. GEOPHYS. RES. 75,
7812-7821.
Macgregor, A. N. and Keeney, D. R. 1973. Denitrification in lake
sediments. ENVIRONMENTAL LETTERS 5(3), 175-181.
Mitchell, J. E., Waide, J. B. and Todd, R. L. 1975. A preliminary
compartment model of the nitrogen cycle in a deciduous forest
ecosystem. Mineral Cycling in S. E. Ecosystems. ERDA symposium
series CONF-740513. p. 41-57-
Nelson, D. W. and Bremner, J. M. 1970. Role of soil minerals and
metallic cations in nitrite decomposition and
chemodenitrification in soils. SOIL. BIOL. BIOCHEM. 2, 1-8.
24-24
-------
Nelson, D. W. and Bremner, J. M. 1970. Gaseous products of nitrite
decomposition in soils. SOIL BIOL. BIOCHEM. 2, 203-215.
Nommick, H. and Thorin, J. 1971. Transformation of N-15 labelled
nitrite and nitrate in forest raw humus during anaerobic
incubation. Nitrogen-15 in soil plant studies, (Int. Atomic
Energy Agency, Vienna) pp 369-381.
Patrick, W. H. and Mikkelsen, D. S. 1971. Plant nutrient behaviour in
flooded soil. In Rinauer, R. C. (Ed.), Fertilizer technology and
use, Second Edition. SOIL SCI. SOC. of AMER. Madison, Wise. 187-
215.
Patriquin, D. G. and Knowles, R. 1974. Denitrifying bacteria in some
shallow-water marine sediments, enumeration and gas production.
CAN. J. MICROBIOL. 20(7), 1037-1041.
Payne, W. J. 1973. Reduction of nitrogenous oxides by
microorganisms. BACTERIOL. REV. 37, 409-452.
Payne, W. J., and Riley, P. S. and Cox, C. D. 1971. Separate nitrite,
nitric oxide, and nitrous oxide reducing fraction from
Pseudomonas perfectomarinus. J. BACTERIOL. 106, 356-361.
Peterson, E. W. 1977. Wet and dry deposition: A synopsis containing
estimates of deposition velocities for some pollutant and trace
gases in the atmosphere. EPA. CERL-037. Corvallis Environmental
Research Laboratory, Oregon.
Ponnamperuma, F. M. 1972. The chemistry of submerged soils. ADVAN.
AGRON. 24, 65-71.
Porter, L. K. 1971. Techniques for the measurement of soil
nitrogenous pollutants. Int. symposium on identification and
measurement of environmental pollutants, Ottawa, Ontario, Canada.
pp. 84-91.
Rasmussen, K. H., Taheri, M. and Kabel, R. L. 1975. Global emissions
and natural processes for removal of gaseous pollutants. WATER,
AIR and SOIL POLLUTION 4, 33-64.
Reddy, K. R. and Patrick, W. H. 1975. Effect of alternate aerobic and
anaerobic conditions on redox potential, organic matter
decomposition and nitrogen loss in a flooded soil. SOIL BIOL.
BIOCHEM. 7, 87-94.
24-25
-------
Renner, E. D. and Becker, G. E. 1970. Production of nitric oxide and
nitrous oxide during denitrification by Corynebacterium
nephridii. J. BACTERIOL. 101, 821-826.
Research Triangle Institute, 1974. Natural emissions of gaseous
organic compounds and oxides of nitrogen in Ohio and surrounding
states. EPA Contract 68-02-1096, Task 12, 13.
Reuss, J. 0. and Smith, R. L. 1965. Chemical reactions of nitrites in
acid soils. SOIL SCI. SOC. AM. PROC. 29, 267-270.
Robinson, E. and Robbins, R. C. 1970. Gaseous nitrogen compound
pollutants from urban and natural sources. J. AIR POLL. CONTROL
ASSOC. 20, 303-306.
Scaletti, J. V., Gates, C. E. Briggs, R. A. and Schuman, L. M. 1960.
Nitrogen dioxide production from silage. I. Field survey. AGRON.
J. 52, 369-372.
Stanford, G., Legg, J. D., Dziena, S. and Simpson, E. C. 1975.
Denitrification and associated nitrogen transformations in soils.
SOIL SCI. 120, 147-152.
Soderlund, R. and Svensson, B. H. 1976. Nitrogen, Phosphorus and
Sulfur Global Cycles. SCOPE Report 7. Ecol. Bull. (Stockholm)
22, 23-73.
Stefanson, R. C. 1972. Soil denitrification in sealed soil-plant
systems. I. Effect of plants, soil water content and soil organic
matter content. PLANT and SOIL 37, 113-127.
Steen, W. C. and Stojanovic, B. J. 1971. Nitric oxide volatilization
from a calcareous soil and models aqueous solutions. SOIL SCI.
SOC. AM. PROD. 35, 277-282.
Stevenson, F. J. , Harrison, R. M., Wetselaar, R., and Leeper, R. A.
1970. Nitrosation of soil organic matter. III. Nature of gases
produced by reaction of nitrite with lignins, humic substances,
and phenolic constituents under neutral and slightly acid
conditions. SOIL SCI. SOC. AM. PROC. 34, 430-435.
Todd, R. L., Waide, J. B. and Cornaby, B. W. 1975. Significance of
biological nitrogen fixation and denitrification in a deciduous
forest ecosystem. Mineral Cycling in S. E. Ecosystems. ERDA
symposium series CONF-740513. p. 729-735.
24-26
-------
Van Cleemput, 0., Patrick, W. H. and Mcllhenny. 1976. Nitrite
decomposition in flooded soil under different pH and redox
potential conditions. SOIL SCI. SOC. AN. J. 40, 55-60.
Vollenweider, R. A. 1968. Scientific fundamentals of the
eutrophication of lakes and flowing waters, with particular
reference to nitrogen and phosphorus as factors in
eutrophication, O.C.E.D. Report, DAS/CSI 68.27
Wijler, J. and Delwiche, C. C. 1954. Investigations on the
denitrifying process in soil. PLANT SOIL 5, 155-169.
Wullstein, L. J. and Gilmour, C. M. 1966. Non- enzymatic formation
of nitrogen gas. NATURE 210, 1150-1151.
Wullstein, L. H. and Gil mour, C. M. 1964. Non-enzymatic gaseous loss
of nitrite from clay and soil systems. SOIL SCI. 97, 428-430.
Yoshinari, T. 1976. Nitrous oxide in the sea. MARINE CHEM. 4, 189-
202.
24-27
-------
CONDENSED DISCUSSION
QUESTION: Which data did you think was most accurate
of that in the table?
TINSEY: Well I think our own 1s, but if you look
honestly at them I think in reality they are
all relatively suspect.
24-28
-------
PROCEDURES FOR CONDUCTING HYDROCARBON EMISSION INVENTORIES OF
BIOGENIC SOURCES AND SOME RESULTS OF RECENT INVESTIGATIONS
Presented at the 1977
Environmental Protection Agency
Emission Inventory/Factor Workshop
Raleigh, North Carolina - September 13-15, 1977
By
Pat Zimmerman
Assistant Environmental Scientist
Washingtyn State University
Air Pollution Research Section
Pullman, Washington 99164
25-1
-------
Introduction
All of us have enjoyed the smells of a thick pine forest or
a blossoming orange grove. Few of us can forget majestic views
of distant blue mountains.
Pollution from plants? Not likely. Plants manufacture
oxygen and we all know how important oxygen is.
Rasmussen and Went were among the first to attempt to measure
vegetation emissions (Rasmussen and Went 1965, Rasmussen 1972).
Essentially all previous predictions of natural hydrocarbon
emission rates have either been made by Rasmussen or based upon
his early work. That early work established that natural sources
can contribute hydrocarbons to ambient air and it postulated that
the amounts of emissions may be quite large. Smog chamber
studies showed that ozone could be generated when car exhaust was
irradiated by sunlight. It was then found that a mixture of
"natural" hydrocarbons, NO and sunlight could also produce ozone
X
(Ripperton et al. 1967). During the same time period the technol-
ogy to accurately monitor ozone levels on a routine continuous
basis was developed. As monitoring systems expanded, agencies
began reporting high ozone levels in rural areas well away from
local anthropogenic sources of ozone precursors. Local, State
and Federal officials in charge of maintaining specified oxidant
levels were faced with the monstrous task of designing comprehensive
and often very expensive strategies to control the emission of
25-2
-------
oxidant precursors. Detailed inventories of anthropogenic point
and area sources were conducted so that the most effective control
techniques could be employed. Yet the natural component - though
known to exist and known to have a potential for ozone development
was ignored. No standardized technique for the evaluation of
natural sources existed.
In 1976 the National Air Data Branch of the Environmental
Protection Agency funded the Washington State University Air
Pollution Research Section to develop a technique for the evalu-
ation of natural organic emissions. Specifically the project
objectives were:
1. To Develop a standardized sampling and analytical
methodology
2. To develop emission factors for a limited number of
species
3. To develop standardized emission inventory techniques
4. To develop preliminary nationwide emission estimates
Sample Methodology
A static/dynamic enclosure system was developed which satis-
fied the project objectives. In addition the technique is rela-
tively simple, inexpensive and easily reproducible by other
laboratories.
Basically, the method involves enclosing a portion of the
vegetation of interest in a Teflon bag, sucking most of the
25-3
-------
ambient air out of the bag and thus collapsing the bag around the
branch. A "background" sample of the air from the collapsed bag
is then collected. The bag is then quickly filled with a known
volume (60«, @ lOfc/min.) of air which is free of all hydrocarbons
and has a controlled level of C02- After the bag is partially
inflated, an "emission rate" sample is collected while zero air
continues to flow into the enclosure at a rate slightly higher
than the sample rate (- ZVmin.). The total enclosure time is
approximately 15 minutes. The emission rate is equal to the
difference in the hydrocarbon content of the bag as measured in
the emission rate sample minus the hydrocarbon content of the bag
as measured in the background sample collected prior to inflation
with zero air. This emission rate is then divided by some unit
of foliage measurement such as by leaf biomass. The result is a
raw emission rate of the units: micrograms emission per gram leaf
biomass per hour (vg/g.hr).
To collect emission samples from vegetation the following
equipment is needed:
1. A portable source of pure air which is hydrocarbon free
and can be regulated to give a precise flow rate.
2. A vegetation enclosure which does not add to or take
away from organic vegetation emissions.
3. A method to collect an air sample from the enclosure and
to move it to a laboratory.
25-4
-------
4. A temperature sensor to measure bag temperature and
ambient air temperature.
The WSU sample train included an open-ended 110 x 150 cm Teflon
bag (capacity - 120£) as an enclosure. A sample manifold (Fig.
1) provided flow control for the pure air entering the bag and
sample leaving the bag. A metal bellows pump was used for the
transfer of samples from the enclosure into 5.5I electropolished
stainless steel cannisters. Clean 1/8" (od) copper tubing or
1/8" od Teflon tubing was used for pure air and sample transfer
lines. Figure 2 is a schematic drawing which illustrates the
experimental design. The pure air was obtained by cryogenically
compressing air purified by an Aadco pure air generator into
empty medical oxygen cylinders. Other sources of clean air such
as zero-grade air commercially available in high pressure cylinders,
could probably be used. However, it would be important to conduct
a blank analysis of each cylinder to determine the COo concentration
and hydrocarbon species present. An indoor outdoor thermometer
was used to obtain the ambient air temperature and the enclosed
air temperature simultaneously.
The emissions from the surface of soil/leaf litter can also be
sampled with this basic setup plus some additional equipment. The
additional equipment consists of a sealing ring and a stainless-
steel bag collar (Fig. 3). The sealing ring is driven into the soil;
the bag collar is then placed in the center of the sealing ring.
25-5
-------
FIGURE 1
PORTABLE SAMPLE MANIFOLD
0-AIROUT
PRESSURE
GAUGE
SAMPLE' PRESSURE
REGULATOR
0-AIR IN
REGULATOR
EVAC.
SAMPLE 0-AIR
FLOW FLOW
0-AIR OUT
PRESSURE
GAUGE
NEEDLE VALVES
25-6
-------
FIGURE 2
VEGETATION EMISSION SAMPLE COLLECTION SYSTEM
ts»
PORTABLE SAMPLE MANIFOLD
-THERMOMETER
STAINLESS STEEL
CANNISTER
METAL
BELLOWS PUMP
-------
FIGURE 3
SOIL LEAF-LITTER SAMPLING SYSTEM
EVAC.
COLLAPSIBLE
TEFLON BAG
[r
BAG COLLAR
SAMPLE
ZERO AIR INLET
MOIST SOIL SEAL
SEALING RING
(2)J4"SWAGLOCK BULKHEAD
SOIL
SHARP CUTTING
EDGE
SEALING RING
BAG COLLAR
*all dimensions in centimeters
25-8
-------
Next moist dirt is used as a filler between the sealing ring and
the bag collar. After the collar and ring are in place, the
Teflon bag is attached by means of a wide elastic strap to the
bag collar. The sample collection procedure is then identical to
that for vegetation.
Surface waters of bays, estuaries and marshes can be sampled
by the addition of a floatation ring to the bag collar. The WSU
laboratory used two closed cell polyurethane filled water-ski
belts sewn together and strapped around the bag collar as a
floatation ring.
Analysis
Most samples were returned to the laboratory for f.i.d. gas
chromatographic analysis within 24 hours. Columns and conditions
are shown in Table 1. Three separate analysis were performed on
each sample. One instrument was required to quantify methane,
ethane, ethylene, and acetylene. Another instrument was equippped
to quantify ^ - Cg hydrocarbons. A third instrument was committed
to the analysis of C^ - C^ hydrocarbons.
The overlap in analysis served as a check on GC performance.
GC response was calibrated with respect to a standard concentration
of neo-hexane.
In order to achieve sufficient GC response a sample concen-
tration step is required. Prior to analysis, 500 ml of sample is
drawn through a 3.2 mm (1/8") o.d. stainless steel sample loop
25-9
-------
TABLE I
Hydrocarbon Analysis Conditions
Compound
Instrument
Operating Conditions
ro
»
o
Ethylene
Ethane
Acetylene
Methane
Light
Hydrocarbon
crce
P.E. 3920 Isothermal
FID
P.E. 3920 Temp. Prog.
FID
Heavy Hydrocarbon
and Oxygenates
Column: 10' x 1/8" OD Porapak 0
Carrier: He 80 psi, 7 ml/min
Hydrogen: 22 psig
Compressed Air: 50 psig
Column: 20' x 1/16" OD Durapak
N-Octane
Carrier: He 90 psig, 6 ml/min
Hydrogen: 40 psig
Compressed Air: 50 psig
Oven: Initial Temp. -70°C
final 65"C
Delay time: 4 min.
Program rate: 16°/min
Total Run Time: 40 min.
Column: 10' x 1/8" Durapak Low K,
Carbowax 400
Carrier: He 90 psig, 8 ml/min
Hydrogen: 40 psig
Compressed Air: 50 psig
Oven: -20 to 100°C
Delay Time: 2 min.
Program Rate: 8°/min
Total Run Time: 20 min.
-------
TABLE 1 (cont'd)
Hydrocarbon Analysis Conditions
Heavy P.E. 3920 Temp. Prog. 4. Column: 200' SCOTOV-101,
Hydrocarbon FID GC 10' x 1/16" OD
C.-C1? Durapak low-K, Carbowax
^ ' 400 precolumn
Carrier: He 90 psig, 5 ml/min
Hydroqen: 40 psig
Compressed Air: 50 psig
Oven: 0°C-100°C Temp. prog.
Delay Time: 6 min.
Program Rate: 6°/min
Total Run Time: 60 min.
5. Column: 30 m SE 30 Glass
Capillary Column
Carrier: He 90 psig, 1 ml/min
Oven: -30-80°C Temp. prog.
Delay Time: 8 min.
Program Rate: 4°/min
Total Run Time: 50 min.
-------
attached to a six port sample valve. The loop is filled with
60 - 80 mesh glass beads and immersed in liquid oxygen. The
sample valve is then switched on stream with the GC column and
the loop is immersed in boiling water to flush the organics
trapped on the glass beads onto the head of the column. Fig. 4 shows
the vacuum sampling system used to introduce known amounts of
sample into the sample loop. This system consists of a chamber
of known internal volume which has been evacuated. A vacuum
gauge is attached to the chamber and calibrated so that a specified
change in vacuum is equivalent to a specified sample volume.
Preparation of an Emission Inventory
After a specific area and time period of interest have been
designated for an emission-study, program there are five major
steps in developing an emission inventory.
1. Identify the major vegetation types and predominate
plant species.
2. Select the representative species to be sampled.
3. Quantify the biomass for each species selected.
4. Conduct a field program to collect emission samples
from each of the representative species.
5. Develop emission rate algorithums for seasonal and/or
daily weather variables (sunlight, temperature, moisture).
25-12
-------
FIGURE 4
VACUUM SYSTEM
for sample injection
OFF
SAMPLE
FRONT VIEW
VACUUM SAMPLE
INLET-
BACK VIEW
25-13
-------
These five steps can be classified into two basic components
necessary for inventory development: 1) a set of emission factors
representative of the vegetative species and conditions (season,
temperature, etc.) for the area to be inventoried and 2) biomass
factors, which are a measure of the quantity of vegetation or
litter present in the area. Inventorying vegetation is therefore
analogous to any other source category in that an appropriate
emission factor is multiplied by some source activity level in
order to estimate emissions.
The samples collected by WSU in Pullman in the fall of 1976
were used to construct a sample emission inventory for the
continental U.S. Figure 5 is a map of the major biotic regions
of the U.S. Raw emission factors were then determined for each
biotic region by approximating the vegetative mix with the average
emission rates of similar vegetation sampled in the Pullman area.
The raw emission factors were then corrected for the effect of
temperature. The temperature of the United States was approximated
by an average winter temperature of 7.5°C and an average summer
temperature of 23°C. The corrected raw emission factors were
then multiplied by the respective summer and winter leaf biomass
estimates for each biome. This resulted in an average summer
emission factor and in an average winter emission factor.
Multiplying these emission factors by the number of hours of
summer and winter respectively and adding the products results in
25-14
-------
Figure 5
MAJOR BIOTIC REGIONS OF THE U.S.
:XvXvXvXxX;:
jx[:|xjxv:jx|xjx
25-15
GKAMJI.AUD
SCt l.KOr'IIYU. '.iCI
TEMI'f-R/\TI: KA1N t'ORCST
DECIDUOUS I-OREST
CONIFEROUS FORCST
DESI-IRT
TUNDRA, ALPINi: FIELDS
vZA
-------
an estimation of the average annual emission rate from natural
sources (Fig. 6). As figure 6 illustrates, this total is equal
to aproximately four times the total emissions from anthropogenic
sources. This emission inventory will be updated this fall.
Recent sample programs conducted in Florida and North Carolina
indicate that the emission factor used for oaks may be five times
too small. Early samples were collected in a dark stainless
steel enclosure. The major emission from oaks is isoprene and
isoprene seems to be emitted almost exclusively in daylight
conditions. Similarly samples of pines collected in Pullman had
emission rates approximately one third lower than those of measured
in North Carolina and Florida. These emission rate differences
may partly be the result of species differences, but more likely
the differences are due to seasonal changes in emission rates as
reflected by the basic metabolic rate of the respective vegetation.
Some raw emission rates for selected vegetation types sampled in
Pullman, North Carolina and Florida are shown in Table 2. Figure
7 illustrates a preliminary emission inventory developed for the
Tampa, St. Petersburg area. This inventory is also only an
approximation. No area was subtracted for highways or developments
and no provision for temperature or light relationships was
included. Details of the inventory are explained in EPA Periodic
letter Report IIV, Interim Progress Report, prepared for Air
Programs Branch EPA Region IV, Contract -68-01-4432. Approximately
25-16
-------
Figure 6
TOTAL YEARLY EMISSION RATE FOR THE CONTINENTAL U.S.
BASED ON LEAF BIOMASS
Summer Emissions:
Average Vegetation Emission Rate
1.36 x 104T/hr x 708.7 hrs/mo x 5 mo = 4.82 x 107T
Litter Emission Rate
132 g/m2 hr x 1 x 106m2/km2 x 9.06 x 106 km2/U.S. = 1.2 x 103T/hr
1.2 x 103T/hr x 708.7 hrs/mo x 5 mo = 4.20 x 106T
Total Summer Emission 5.24 x 10 T
Winter Emissions:
Average Winter Emission Rate
7.04 x 103T/hr x 708.7 hrs x 7 mo = 3.49 x 107T
Total Yearly Emission Rate for Vegetation and Leaf-Litter 8.73 x 10 T/yr
Total Yearly Emission Rate from Anthropogenic Sources 2.12 x 10 T/yr
Natural Emissions as the Percent of the Total Emissions from All Sources
8.73 x 107T/yr = QQ%
2.12 x 10'T/yr + 8.73 x TO'T/yr
-------
TABLE 2
AVERAGE RAW EMISSION RATES FOR SOME SELECTED SAMPLE TYPES IN FLORIDA
NORTH CAROLINA AND WASHINGTON
Number X" Raw Emission
Sample Type of Samples Rate
Tampa, Florida April - May 1977
Selected Marine Samples ?
Mud Flat 8 119.8 yq/nyhr
Intertidal Samples 3 117.2 ya/nyhr
Decaying Vegetation 13 202.0 yg/nyhr
Marine Grass - 20 93.5 yg/nyhr
N> All Marine Samples 89 143.0 yq/m'hr
Ul 'o
I* Pastures 32 288.6 yg/m "hr
00
Oaks
Laurel Oak 10 11.2 yg/g'hr
Turkey Oak 3 26.1 yg/g'hr
Water Oak 3 27.2 ya/g'hr
Blue Jack Oak 6 16.5 ya/n'hr
All Species of Oak (Day) 25 21.14 yg/q'hr
All Species of Oak (Night) 3 1.20 yg/g'hr
Ci trus
Oranges 8 3.1 yg/q'hr
Grapefruit 3 2.8 yg/g'hr
Coni fers
Long Leaf Pine 16 3.3 yg/Vhr
Slash Pine 7 6.0 yg/n'hr
Sand Pine 2 7.3 yg/g'hr
All Conifers 26 5.1 yg/q-hr
-------
TABLE 2 (cont'd)
AVERAGE RAW EMISSION RATES FOR SOME SELECTED SAMPLE TYPES IN FLORIDA
NORTH CAROLINA AND WASHINGTON
Raleigh, North Carolina June 1977
Oaks 3 26.10 yg/g-hr
Conifers
Shortleaf Pine 2 16.3 yp/g-hr
Loblolly 2 4.85 yp/n'hr
E. Red Cedar 2 1.14 yg/g'hr
Virginia Pine 1 13.60 yg/o-hr
All Conifers 7 8.4 yg/g-hr
Pullman, Washington August - November 1976
Conifers
Ponderosa Pine 6 2.96 yp/g-hr
Mugo Pine 2 1.78 yg/g-hr
Douglas Fir 6 0.86 yg/p-hr
Juniper 4 3.25 yg/n-hr
Spruce 3 .2.26 yg/g-hr
All Conifers 21 2.26 yg/g-hr
Oak (Dark Chamber) 2 4.40 yg/n-hr
Pine Litter 1 132.00 yq/m2-hr
-------
Figure 7
PRELIMINARY NATURAL HYDROCARBON EMISSION INVENTORY
FOR THE TAMPA/ST. PETERSBURG AREA
N>
Y
S3
o
Vegetation Type
Col umn
Tampa Bay
Intertidal
Marine Grass
Improved Pasture
Unimproved Pasture
Citrus
Mangrove
Palmetto
Pine
Shrubs
Leaf
Biomass
Factor
g/m2
1
498
156
240
760
240
X Raw
Emissions
Factor
yg/tj-hr
2
3.02
1.77
9.74
5.13
9.74
X Raw
Emissions
Rate
pq/m2 hr
3
143
284
301
1,504
276
4,091
3,899
4,091
Percent
of area
covered
4
22
9
12
8
3
12
6
12
Est.
area
Km2
5
997
419
558
349
156
532
279
532
Total
Emission
Rate
(q)/hr
6
1.4X105
1.1X105
1.7X105
5.2X105
4.3X104
2.2X106
1.1X106
2.2X106
-------
Figure 7 (cont'd)
PRELIMINARY NATURAL HYDROCARBON EMISSION INVENTORY
FOR THE TAMPA/ST. PETERSBURG AREA
Oak-Hickory 498 14.87 7,405 8 349 2.6X106
Oak-Gum-Cypress 480 38.57 18,514 8 349 6.5X106
4,520 1.6X107q/hr
Total Average Hourly Emission Rate 16T/hr
V Study Area natural TNMHC Emission Rate: 16T/hr X 24hr/day 384T/day
to
!-
Hillsborough County Anthropogenic Hydrocarbon Emission Rate 132T/day
Pinellas County Anthropogenic Hydrocarbon Emission Rate 48T/day*
*No official estimates were available for Pinellas County.
This figure is a guess based upon the Tampa Bay (AQCR 052)
NEDS emission rate estimate of 539T/day.
Percentage of TNMHC Burden Attribritable to Natural Sources: 384/180+384=68%
-------
330 additional samples have been collected since tabulation of
the preliminary Florida emission inventory. Considerable time
has also been spent in better quantitation of the biomass in the
study area. The final emission inventory will be available after
March 1978 and will be included in an EPA computer modeling
program which is intended to predict oxidant levels for the
Tampa/St. Petersburg area.
Conclusion
Although all preliminary emission estimates predict that
natural sources of hydrocarbons contribute significantly to the
atmospheric burden, the actual concentration of natural hydrocarbons
in ambient air is dependent upon the dilution volume relative to
the hydrocarbon emission rate.
Westberg and Holdren (1976) reported on an analysis of
monoterpenes from a rural forested site in northern Idaho by a
gas chromatograph linked directly to a Mass Spectrometer. They
collected samples of ambient air as well as vegetation emission
samples using the Teflon bag enclosure technique described in
this paper. The emission rates for the terpene compounds were
then used to calculate expected ambient air terpene concentrations
within the forest canopy. Their calculations show that assuming
2
an average terpene emission rate of 3120 yp/m -hr (obtained for a
coniferous forest of this region) and a typical canopy height of
20 meters, 3120 yg of terpene compounds are emitted into a volume
25-22
-------
unit 1 meter x 1 meter x 20 meters (20 m ) in one hour. The time
(t) for the forest canopy to vertically exchange its volume can
be calculated utilizing the "random walk assumption" which states
that T 21 Z/2D, where Z is the mixing distance (20 m) and D is the
vertical diffusion coefficient in a forest canopy (assumed to be
10 cm /sec). Using these values the time for vertical exchange
was calculated to be 200 seconds. The expected ambient air
concentration within the canopy is then:
The calculated value of 1.6 ppb agrees relatively well with
the ambient air values measured during the course of the Westberg-
Holdren study. This calculation indicates that low levels of
natural hydrocarbons are expected in ambient air and that normal di-
lution processes can account for these low levels. Thus, although an
annual U.S. emission rate of natural hydrocarbons which is four
times the total of all anthropogenic sources seems quite large,
the actual concentration of natural hydrocarbons in ambient air
is expected to be quite low. These low levels can be attributed
to the diffuse nature of the source and natural dilution.
Another approach which can be used to put natural hydrocarbon
emissions into perspective is to compare emission densities from
vegetation and from anthropogenic sources in urban areas. Tom
Lahre, Engineer for the Air Management and Technology Branch of
25-23
-------
the Environmental Protection Agency has used this approach to
compare natural emission data generated by this study to data for
anthropogenic emissions gathered during the NEDS program (personal
communication).
Mr. Lahre's comparison shows that for the St. Louis urban
and suburban areas (a radius of 22 miles from the center of the
city) vegetative emissions comprise only 6% of the total of all
emissions in the developed part of the city. However the ratio
of the emission densities for Natural and Anthropogenic sources
reaches unity if a radius of 68 miles from the center of the city
is used. That is for the St. Louis area, assuming that the air
were evenly mixed within a 68 mile radius of the city, 50% of the
total hydrocarbon burden would be due to biogenic sources.
From this information it appears that the total of natural
hydrocarbon emissions is large with respect to anthropogenic
emissions, however anthropogenic sources tend to be much more
concentrated and that physical dilution of natural hydrocarbon
emissions can account for the low levels of natural hydrocarbons
observed in rural areas.
Laboratory research has demonstrated that the natural emission
products are very reactive both in photochemical processes and
ozonolysis reactions. However it has also been reported that
laboratory ozonolysis experiments typically required from 10 to
30 ppb of carbon from the terpene compounds to produce 1 ppb of
25-24
-------
ozone. In addition it has been predicted that many of the natural
hydrocarbons would react with ozone very quickly after entering
the atmosphere leaving only small quantities to exist long enough
to participate in photochemical processes (Westberg 1977).
Though the information generated by the research is important
a few researchers do not feel that it is conclusive. They point
out, for example that most previous estimates of the ozone producina
role of natural hydrocarbons are based upon a hydrocarbon mix of
terpenes. They point out that our research has shown that in
many areas isoprene is by far the largest natural emission product.
Isoprene is very reactive photochemically and is not as likely to
produce aerosols as the heavier terpenes. Also most bag irradiation
experiments in which large amounts of particulates were formed
were conducted at hydrocarbon levels much higher than those of
rural ambient air. They believe that this would probably cause a
shift in reaction kinetics as well as in the end products produced.
However a majority of previous research when coupled with
the biogenic emission rate data determined by WSU indicates that
biogenic emissions of natural hydrocarbons cannot in themselves
account for the existence of high ozone levels in rural areas.
Other mechanisms proposed to explain high ozone levels in rural
areas include the direct injection of stratospheric ozone into
the troposphere (Chatfield and Rasmussen, 1977), or the theory
that anthropogenic ozone precursors including NO and man-made
X
25-25
-------
hydrocarbons are transported downwind over long distances, reacting
to produce ozone. The relative importance of these mechanisms has
not been determined.
Although natural hydrocarbon emissions do not seem to be the
major cause of high rural ozone levels, it cannot be concluded that
these emissions do not play a significant role in the chemistry of
rural atmospheres. Recently much of the scientific literature has
been focused upon the OH budget of the atmosphere. OH radicals seem
to be very important in "scrubbing" trace gases such as S02, and
chlorofluorocarbons from rural air. OH radicals are postulated as the
reactive intermediates for many atmospheric reactions, including the
formation of ozone (Wofsy 1976). They therefore have an effect on
such far ranging problems as acid rain, stratospheric ozone depletion
and tropospheric ozone production. The key to understanding these
problems is a thorough understanding of the OH budget. One of the
basic inputs of OH budget models is the hydrocarbon emission rate
(including methane) into the atmosphere (Crutzen and Fishman 1977).
The interactions between stratospheric injection, downwind
transport, the production of natural hydrocarbons and the OH
budget are unknown. As a result the occurrance of high ozone levels
in rural areas is not understood. Though many questions remain to
be answered, determining the magnitude and types of biogenfc
emissions is an important step in understanding the air around
us.
25-26
-------
LITERATURE CITED
R. B. Chatfield and R. A. Rasmussen, "An Assessment of the
Continental Lower Tropospheric Ozone Budget," EPA Report No. 700-
lA-Jan., 1977.
R. J. Crutzen and J. Fishman, "Average Concentrations of OH in
the Northern Troposphere, and Budgets of CH,, CO and FL," NOAA,
Env. Res. Lab., Aeronomy Lab., Boulder, CO and NCAR, ATr Qual.
Div., Boulder, CO, 1977.
R. Rasmussen and F. W. Went, "Volatile Organic Material of Plant
Origin in the Atmosphere," Proc. Nat. Acad. Sci., 53:215 (1965).
R. Rasmussen, "What do the Hydrocarbons from Trees Contribute to
Air Pollution?" APCA Journal, 22, 537 (1972).
L. Ripperton, 0. White and H. Jeffries, "Gas Phase Ozone-Pinene
Reactions," Div. of Water, Air, and Waste Chemistry, 147th
National Meeting American Chemical Society, Chicago, IL, pp 54-56.
Sept. 1967.
H. Westberg, "The Issue of Natural Organic Emissions," prepared
for U.S. EPA Office of Res. and Dev., order No. DA-7-1290J,
1977.
H. Westberg and M. Holdren, Quarterly Technical Report for EPA
Grant No. 800670, J. Bufalini, Project OFficer, 1976.
S. C. Wofsy, "Interactions of CH, and CO in the Earth's Atmosphere,
Ann Rev. Earth and Planetary Sciences, 4, 1976, pp 441-469.
25-27
-------
Acknowlegements
This work was supported by the Following EPA Contracts:
Contract No. 68-02-2071 Title: Testing of Hydrocarbon
Emissions From Vegetation and Development of a
Methodology for Estimating Emission Rates from
Foliage ion Any Geographical Area in the United
States.
Contract No. Du-77-C063 Title: Additional Testing of
Organic Emissions from Vegetation and Leaf Litter/
Soil Surfaces.
Contract No. 68-01-4432 Title: Determination of
Emission Rates of Hydrocarbons From Indigenous
Species of Vegetation in the Tampa/St. Petersburg
Florida Area.
Don Sterns, Phil Sweany and Bob Watkins provided
significant contributions toward the successful
completion of the fieldwork.
Disclaimer
The mention of specific brand names in this paper is solely
for reference use, and does not constitute an endorsement by
Washington State University or the Environmental Protection
Agency.
25-28
-------
QUESTION:
ZIMMERMAN:
CONDENSED DISCUSSION
Could you estimate what percentage of total
is isoprene?
Isoprene has not been studied in bag studies.
Isoprene is a very active ozone former and
it doesn't tend to form particulates nearly
as readily as the heavy organics. Most bag
studies were also done at high concentrations
What that tends to do is favor the kinetics
to produce particulates etc. I would
say it depends on the forest. In a north-
west forest, the isoprene concentration is
lower. If you get back here where there are
oak trees and other isoprene emiters, it
would be higher. For an oak tree it is 99%
of the total emissions. So, if you had all
oak trees it would be 99%. But it is prob-
ably 60 or 70%, maybe as low as 50% in a
mixed deciduous oak-hickory, etc. forest.
Of course a lot of the oak-hickory areas that
were shown on the map are actually loblolly
pine now that the original forests have been
cut down and replanted.
25-29
-------
QUESTION:
ZIMMERMAN:
QUESTION:
ZIMMERMAN:
QUESTION:
ZIMMERMAN:
Will you report and try to break down the
emissions by compounds.
The report for Florida will. This one won't
as it is a survey report. We've got just
what I presented. The one for Florida will
have the emissions quantified by species for
each one by one grid area. On an hourly
basis throughout the day and throughout the
season the major emissions will be pointed out.
Could you measure emissions over a
diurnal cycle?
Yes, we did, day and night. Every four to
six hours we would resample the live oak trees
and pine trees and so forth.
Did you get no emissions from the Teflon
sample bags?
It was pretty low because it was a short 15
minute enclosure time. We did background
blanks. As long as we turn the bags inside
out and let them dry out good between
samples, background is low. It would also
have been taken into account in our tests
because we collected a background sample
which would have included the bag emissions.
25-30
-------
QUESTION:
ZIMMERMAN:
QUESTION:
ZIMMERMAN:
QUESTION:
How did you investigate the possibility that
by surrounding the branch with a bag you
were changing its natural emissions?
What we tried to do was measure different
parameters such as the bag temperature. We
found in the very worse cases, if you put
the bag in the sun it would heat up maybe 7
or 8 degrees F. We measured relative
humidity, etc. That was the best we could do,
We have also compared data with EPA people
that are doing flux estimates.
I went through some of their figures and
they came out pretty close. Anyway, it
seems that at lower temperatures at least,
there was a pretty good correlation.
Did you measure the COp or did you add COg to
the bag?
Yes. We added COp to the air that we put into
the bag. We did measure it. We put in 365
parts per million, which is equivalent to
ambient COp.
Did you check on the reduction of COp as the
reaction took place?
25-31
-------
ZIMMERMAN:
QUESTION:
ZIMMERMAN:
You mean did we measure it in the samples we
collected? Yes, we did periodically. We
didn't as a routine measure in Florida but
when we were developing the technique we
did. It didn't change a large amount, like
10% maybe. In the dark it would increase 10%.
You are talking about day and night sampling;
did you sample in direct sun light and
indirect sun light and were there any
differences?
We sampled in all different lights. We
recorded the cloud cover and so forth in the
notes and we haven't gone through all the
data yet so I can't make any strong corre-
lations. In other words, I don't know
where the isoprene emission peaks out with
sunlight intensity. We can get that infor-
mation from out data. We will determine a
pattern when we refine our data more and
more.
25-32
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing}
1. REPORT NO.
EPA-450/ 3-78-042 b_
4. TITLE AND SUBTITLE
Emission Inventory/Factor Workshop Volume II
3. RECIPIENT'S ACCESSION NO.
5. REPORT DATE
Mav 1978
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Various
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Air Pollution Training Institute
Air Management Technology Branch
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS
Same
13. TYPE OF REPORT AND PERIOD COVERED
Final Proceedings
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
Moderator: James Southerland
16. ABSTRACT
This report in two volumes presents the written form and summarized discussions
of "presentations" made at the Emission Inventory and Factor Workshop in Raleigh, N.C.
September 13-15, 1977. A total of twenty-five "papers" on emission inventory and
factor experiences and other information with emphasis on organics (hydrocarbons)
were presented. Authors represented EPA, state air pollution control agencies and
private industry.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
Emission inventory
Emission factors
Volatile organic compounds
Oxides of nitrogen
Natural organic emissions
State implementation plans
18. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (ThisReport)
Unrestricted
21. NO. OF PAGES
414
20. SECURITY CLASS (Thispage)
Unrestricted
22. PRICE
Form 2220-1 (Rev. 4-77) PREVIOUS EDITION is OBSOLETE
26-1
------- |