United States
Environmental Protection
Agency
Air and Energy Engineering
Research Laboratory
Research Triangle Park, NC 27711
Research and Development
EPA/600/SR-92/239 March 1993
EPA Project Summary
Revised Emissions Estimation
Methodologies for Industrial,
Residential, and Electric Utility
Stationary Combustion Sources
Dan Bowman, Scott Lowe, David Winkler, and David Zimmerman
This report presents the development
of Improved and streamlined EPA emis-
sion estimation methods for stationary
combustion area sources by the Joint
Emissions Inventory Oversight Group
(JEIOG) research program. These
sources include categories traditionally
labeled "other stationary source com-
bustion": residential, commercial/insti-
tutional, industrial, and electric utility
fuel-burning sources that are not In-
ventoried as point sources because
they are too small or numerous to In-
ventory individually. The report con-
sists of 10 chapters: Introduction,
Industrial Fuel Combustion Methodolo-
gies for Existing Categories, Industrial
Wood Fuel Combustion Methodologies,
Crude Oil Combustion Emissions from
Field Operations, Residential Combus-
tion Source Methodologies, Sulfur Con-
tent, Ash Content, Electric Utility Area
Source Fuel Use, Error Handling Pro-
cedures, and Conclusions and Re-
search Recommendations
Each chapter summarizes existing
methodologies (if any exist), then pre-
sents proposed algorithms. Emission
factors and data sources required to
implement the new methodologies are
introduced In each chapter.
This Project Summary was devel-
oped by EPA's Air and Energy Engi-
neering Research Laboratory, Research
Triangle Park, NC, to announce key find-
Ings of the research project that Is fully
documented In a separate report of the
same title (see Project Report ordering
Information at back).
Introduction
This report presents the development
of improved and streamlined EPA emis-
sion estimation methods for stationary com-
bustion area sources by the Joint Emissions
Inventory Oversight Group (JEIOG) re-
search program. These sources include
categories traditionally labeled "other sta-
tionary source combustion1': residential,
commercial/institutional, industrial, and
electric utility fuel-burning sources that are
not inventoried as point sources because
they are too small or numerous to inven-
tory individually. This research supports
the National Air Data Branch's (NADB's)*
annual estimates of total suspended par-
ticulate matter (TSP), sulfur dioxide (SO2),
nitrogen oxides (NO,), reactive volatile or-
ganic compounds (VOCs) and carbon mon-
oxide (CO). It also supports state
implementation plan (SIP) emission inven-
tory methods for ozone precursors and
CO, as well as the potential SIPs for NOX,
particulate matter less than ten u,m in di-
ameter (PM-10) and other criteria pollut-
ants with important combustion emission
components.
The need for review and revision of the
stationary combustion source methodolo-
gies is emphasized by four factors:
Current methods are outdated. EPA's
National Emissions Data System/Area
and Mobile Source Subsystem
(NEDS/AMS) area source methods
were developed in the early 1970s,
relying on data sources dating back
' Part of the Technical Support Division of EPA's Office
of Air Quality Planning and Standards
Printed on Recycled Paper
-------
to the 1960 Census. These techniques
have never been validated, nor do
they account for drastic changes in
energy consumption patterns that
have taken place throughout the
1970s and 1980s. Many of the meth-
ods require that results be compared
and reconciled with other data
sources.
Current methods require a significant
level of effort to collect and process
data. However, most of these emis-
sion sources are relatively small in
magnitude.
No previous methodology develop-
ment has considered both national
annual Inventory needs and SIP in-
ventory requirements. If annual and
SIP inventories are to be comparable,
emission estimates of both types must
be comparable.
The Clean Air Act Amendments
(CAAA) of 1990 have introduced ad-
ditional requirements of current emis-
sion inventory techniques; e.g.,
additional SIP inventories and lower
point source emission magnitude cut-
offs.
EPA's Air and Energy Engineering Re-
search Laboratory (AEERL) has under-
taken a research program to develop and
revise area source emission methodolo-
gies for use in both national (annual) and
SIP inventories, coordinated under the
JEKX3 program. Goals of this methodol-
ogy development are to improve the reli-
ability of the stationary combustion area
source emission estimates and streamline
the estimation process where possible to
reduce the level of effort required to imple-
ment the national and SIP inventories.
This report presents alternative meth-
odologies for industrial, residential, electric
utilities, sulfur content calculations, and
ash content calculations. These were cho-
sen from a list of 36 categories initially
developed for this project and represent
those considered most important for meth-
odology development. A 37th category,
identified in previous research (crude oil
combustion), was added. The other cat-
egories are expected to be covered in the
future within this research program.
This report also presents the stationary
combustion area source methods. Each
chapter includes a narrative description of
the existing NEDS/AMS and SIP methods
and possible alternatives. Wherever rea-
sonable, at least two alternatives to the
existing method are described. Following
the narrative phase, algorithms to com-
pute fuel consumption for the alternative
methods are presented with the emission
factors that would be employed. Data
sources for the alternative methods are
documented for the user. Finally, the con-
clusions derived from this research and
the opportunity for reliability and streamlin-
ing that the new methodologies represent
to the emission inventory community are
summarized.
Industrial Fuel Combustion
Methodologies for Existing
Categories
Industrial fuel combustion categories
include 11 area sources of air emissions.
The following sources are listed in the
order of their relative importance to the
emission inventory community: Natural Gas
(Boilers) Combustion, Natural Gas (1C En-
gines) Combustion, Bituminous Coal Com-
bustion, Residual Oil Combustion, Distillate
Oil Combustion, Anthracite Coal Combus-
tion, Liquefied Petroleum Gas (LPG) Com-
bustion, Wood Combustion, Coke
Combustion, Process Gas Combustion,
and Crude Oil Combustion
Methodologies were developed for all
of these categories except coke and pro-
cess gas combustion, which were elimi-
nated from methodology development
based on the low priority given to these
categories, their relatively low estimated
emissions levels, and the limited availabil-
ity of data to predict area source consump-
tion of coke and process gas based on a
review of U.S. Department of Energy (DOE)
and other data. Wood and crude oil con-
sumption are treated separately from the
remaining fuels because there is no exist-
ing methodology for either category.
Three closely related alternatives are
proposed for most industrial fuel catego-
ries. The first alternative eliminates the
current AMS normalization process through
the use of an alternate data source and
ensures that county industrial area source
bituminous coal consumption estimates
sum to state industrial area source con-
sumption by industry. However, this ap-
proach relies on county area source
employment estimates, so while it may be
a better approach (i.e., more consistent in
its use of available data), it is still some-
what data-intensive.
The second alternative makes use of
county-level fuel intensities for each indus-
try and also yields estimates of consump-
tion for each county and industry. The
industrial estimates will use planned DOE
reports of state totals. This approach is
more data-intensive than the first because
it employs area source employment esti-
mates and industry-specific fuel intensity
calculations.
The third methodology does not use
area source employment estimates. The
consumption estimates derived from the
third strategy may not be as accurate as
the estimates from the first two approaches,
but it is the least data-intensive of the
estimation procedures. This third technique
uses the number of establishments in each
county and industry to apportion reported
industrial area source fuel consumption!
among counties.
These methodologies represent a de-|
parture from the underlying Census and
DOE data sources which underpin the cur-
rent AMS fuel consumption estimates. Al-
though the current AMS methods use the
best DOE fuel use data currently available,
DOE is planning to release data that will
simplify the existing methodologies. The
following methodologies presuppose the
release of these data. (Proposed release
dates for these data are being investi-
gated.) the current methods require data
from at least four DOE data sources, while
the proposed method uses only two DOE|
data sources.
Industrial Wood Fuel
Combustion Methodologies
Due to the limited number of data I
sources on wood fuel usage, wood con-1
sumptiop data must be extrapolated from
the census region level. To simplify the
methodology and use available data
sources, an industrial wood-burning sub-
set was1 sought. Approximately 97% of
wood burning for fuel in industrial settings
occurs Within two Standard Industrial Clas-
sifications (SICs): SIC 24, the paper and
allied products industry (79%), and SIC
26, the lumber and wood products industry
(18%). Ten other industrial SICs consume
the remaining 3% of fuel wood.
The proposed methodology for estimat-
ing emissions from industrial wood fuel is:
(1) distribute regional wood consumption
to the state level, (2) allocate state level to
the county level based on employment in
S!Cs 24: and 26, and (3) eliminate point
sources from the county estimate of wood
consumption.
Crude Oil Combustion
Emissions from Field
Operations
There is no methodology for estimating
crude oil: consumption and emissions from
oil field operations in the NEDS, current
AMS, ori SIP area source methodologies.
During the compilation of the 1985 NEDS
for the National Acid Precipitation Assess-
ment Program (NAPAP) inventory, NADB
discovered that California reported crude
consumption data and emissions in the
residual oil category. This reporting in turn
affected comparisons made on residual oil
consumption between NEDS and DOE to
-------
determine area source residual oil con-
sumption.
The two proposed methods for estimat-
ing emissions from crude oil field opera-
tions rely on available DOE data sources
to estimate and allocate crude oil con-
sumption. Census data from the U.S. De-
partment of Commerce's Bureau of Census
are used to allocate total crude oil con-
sumption for combustion to the county level.
Both of the proposed methods consist of
four steps:
(1) Estimate crude oil consumption
for combustion at oil and gas pro-
duction fields.
(2) Allocate consumption estimate to
the state level.
(3) Allocate consumption estimates
to the county level.
(4) Adjust for point source consump-
tion.
Residential Combustion Source
Methodologies
Historically, a significant level of effort
has been required to collect and process
data used to generate emission estimates
for residential combustion sources on both
the national and local levels. Most of the
current methods and underlying data
sources for the fuel categories have been
unaltered since the early 1970s. There-
fore, there is a need for methodologies
which are simplified from the tedious and
sometimes lengthy methods presently used
for consumption estimates.
Four streamlined methods are proposed
county-level fuel consumption estimates.
These methodologies estimate county-level
fuel consumption based on population, the
number of homes using a specific fuel for
heating, the median number of rooms per
heated dwelling, or heating-degree days
(HDDs). The new simplified methodolo-
gies will enable the user to interchange the
residential fuel types in the same algo-
rithms and calculate the residential fuel
consumption of those fuels without adapt-
ing to numerous algorithms for each fuel
type. These streamlined methodologies use
the most complete, updated data sources
available to ensure that the fuel consump-
tion estimates are as accurate as possible
for the study year.
The methodologies allocate DOE-re-
ported state-level residential fuel consump-
tion to the county based on available
county-level surrogate data. The surrogates
used for allocation in the four proposed
methodologies are: (1) county-level popu-
lation, (2) county-level number of homes,
(3) number of rooms in county homes, and
(4) county population and HDD data.
Sulfur Content
After reviewing current methodologies,
it was concluded that most of the method-
ologies should remain intact because of
the lack of documented sulfur content from
public documents or agencies. Revisions
to the current methodologies are limited to
the identification of up-to-date data sources
which include sulfur content.
The current methodologies for estimat-
ing sulfur content from anthracite coal are
sufficient and do not need revision or sim-
plification. Since anthracite coal is mined
and used in a small region of the U.S.,
primarily in Pennsylvania and the surround-
ing states, the amount of sulfur content on
the national level reported by DOE is rep-
resentative of the county-level sulfur con-
tent. The current methodology does not
require computation or manipulation of any
data. The method implies that the county-
level sulfur content of anthracite coal is
equal to the national-level sulfur content
for all consumer categories.
The current methodologies used for bi-
tuminous coal are separated into two
groups: sulfur content of bituminous coal
which is used in the residential and com-
mercial sectors; and coal which is used in
the industrial sector. Revising the current
methodologies will eliminate the lengthy
algorithm currently used to estimate sulfur
content.
The proposed methodologies are
equalities based on the assumption that
the state-level sulfur content of coal is
distributed equally among the counties in
the study state. DOE publishes the source
for sulfur content data for all states which
use bituminous coal. The proposed meth-
odology assumes that the sulfur content of
bituminous coal is consistent throughout a
given state.
The current methodologies for the esti-
mation of sulfur content in distillate and
residual oils are not a computation or ma-
nipulation of data. Revisions of the current
methodologies will present better estimates
of sulfur content on the state and county
levels. Sulfur content data are available for
distillate and residual oils at a regional
level from the National Institute for Petro-
leum and Energy Research (NIPER). The
proposed methodology assumes that the
sulfur content of distillate fuel is consistent
throughout the regions reported on by
NIPER.
Ash Content
As was the case for sulfur content,
review of the current methodologies and
the available data sources reporting ash
content indicated that the current method-
ologies would essentially remain as they
are and would not undergo major revi-
sions. Primarily, the lack of available data
sources for ash content limits the revision
process for the methodologies. The cur-
rent methodologies do not recommend us-
ing a regularly updated data source. The
revised methodologies for estimating ash
content identify data sources that regularly
report and update ash content for coal.
Methodologies for estimating ash con-
tent in anthracite and bituminous coals are
presented in the report. The methodolo-
gies assume that ash content is consistent
throughout the U.S., and use published
national statistics for ash content at the
county level.
Electric Utility Area Source Fuel
Use
No AMS methodologies exist for calcu-
lating area source emissions from fuel com-
bustion at electric utilities. The assumption
implicit in this omission is that the point
source inventory adequately covers emis-
sions from electric utility fuel consumption.
However, EPA has discovered that a sig-
nificant number of small electric utilities
are not included in the point source inven-
tory.
A single methodology was developed
for estimating area source emissions from
electric utility fuel combustion. The data
used for the methodology are regularly
updated and publicly available. The meth-
odology is essentially a three-step alloca-
tion process that ensures that total DOE
electric utility fuel is counted: (1) generate
a fuel-specific inventory of all power plants
in the study county, (2) eliminate point
source power plants from the inventory,
and (3) determine the amount of fuel used
by the power plants in the remaining in-
ventory
Error Handling Procedures
Several of the alternative algorithms
proposed in this report rely on Aerometric
Information Retrieval System (AIRS) Facil-
ity Subsystem (AMS) point source through-
put or employment data to derive state-level
area source activity or county-level fuel
intensity data. The equations given pro-
vide an error-handling capacity that en-
ables the methodologies to proceed if there
are erroneous data or data gaps. The fol-
lowing algorithms and equations are af-
fected:
Industrial Fuel Consumption (Chapter
2, Algorithm 1, Equation 1), where
state-level reported point source fuel
consumption exceeds total estimated
-------
state-level fuel consumption from the
Manufacturing Energy Consumption
Survey (MEGS) for one or more SICs.
* Industrial Fuel Consumption (Chapter
2, Algorithm 1, Equation 7), where
county-level reported point source em-
ployment exceeds total estimated
county-level fuel consumption from the
County Business Patterns for one or
more SICs.
Industrial Fuel Consumption .(Chapter
2, Algorithm 3, Equation 16), where
state-level reported point source fuel
consumption exceeds total estimated
state-level fuel consumption from
MECS for one or more SICs.
Industrial Wood Consumption (Chap-
ter 3, Algorithm 1, Equation 3), where
state-level reported point source wood
consumption exceeds total estimated
state-level wood consumption from
Estimates of U.S. Biofuels.
Industrial Crude Oil Consumption
(Chapter 4. Algorithms 1 and 2, Equa-
tions 1 and 4), where county-level
reported point source crude oil con-
sumption exceeds total estimated
county-level crude oil consumption
based on Petroleum Administration
Defense District (PADD) statistics.
Conclusions and Research
Recommendations
Two objectives guided the emissions
estimation methodology development for
stationary area source categories: (1)
streamlining the methodologies to reduce
the data requirements and complexity of
the calculations, and (2) increasing the
reliability of the emission estimate. This
project has concentrated on development
of new methodologies to estimate fuel con-
sumption by county for these categories;
future work will Investigate emission factor
revisions. Streamlining has been accom-
plished by using readily available data
sources and offering straightforward meth-
ods that can easily be implemented by
state or local Inventory staff to meet an-
nual or SIP requirements. The alternatives
presented here attempt to meet the sec-
ond criterion (reliability) by employing ex-
isting DOE data sources that are updated
regularly.
Development of these alternatives has
revealed both weaknesses and strengths
in the pursuit of options to the existing
AMS methodologies. The drawbacks to
many of the methodologies proposed here
are consequent to the stated objectives:
The proposed methods center on al-
location of state-level fuel consump-
tion.
The proposed methods are
invalidated.
The availability and types of fuel data
limit the range of methods that can
be developed.
The proposed methods rely on DOE
data sources that represent fuel con-
sumption as equivalent to distribution.
The industrial methodologies rely on
data not yet published by DOE.
The lack of reported data for ash and
sulfur content in the fuel categories
has limited the amount of revision
possible to the AMS methods.
The alternatives do represent an im-
provement to existing methods, especially
for application to SIP inventories, because
the methods no longer rely on an entire
state area source inventory. The improve-
ments also provide a marked simplification
in some methods. The following character-
istics are seen as the significant advances
from the AMS methods:
The industrial methods have been
streamlined.
New methodologies have been cre-
ated for industrial wood and crude oil
consumption.
The industrial LPG methodologies ac-
count for feedstock consumption.
The proposed residential methodolo-
gies are capable of use for all cat-
egories, but they are consistent with
existing results.
A new method is included to cover
electric utility fuel consumption and
emissions that escape the point
source inventory.
This stage in the program to provide!
stationary combustion area source emis-l
sions estimation methods provides the first I
new methods for these sources since the!
inception1 of area sources about 15 years!
ago. It also represents the use of earlier!
stages which reviewed existing methods,!
detected problems, sought user needs, and J
identified, new data sources. Further work!
under this research program should pro-1
duce methodological alternatives to the!
categories not covered here: commercial, I
waste disposal, and miscellaneous sta-J
tionary combustion.
The development of consumption esti-
mation methodologies should lead to revi-1
sion of emission factors for these]
categories. For example, residential fur-1
nace and commercial boiler emission fac-j
tors are in reality based on larger!
commercial and industrial boiler emission j
factors with different design, operation, and
maintenance characteristics. The research
plan calls for parallel development of new |
emission factors.
Finally, work on improving the baseline I
consumption data and validating consump-
tion allocation should be discussed. It is
evident from the review of data sources
that DOE has access to much more de-
tailed data than are compiled or published.
The involvement of DOE representatives
in the identification or compilation of data
could improve the reliability of these meth-
ods and reduce the effort required through
the availability of electronically formatted
data sets. Data validation has not been
undertaken to date for area sources. How-1
ever, validation of consumption estimates,
if not emissions, is possible for some cat-1
egories. For example, residential natural
gas consumption is accurately tracked by
natural gas companies. Accessing, com-1
piling, and matching these data to the
county-level estimates derived from the
equations represent a significant task, but |
some pilot studies could be designed and
implemented to gauge the probable reli-
ability of the methods proposed in this
report.
U,S.Gwwnnปnt Printing Office: 1993750-071/60202
-------
-------
D-Bowman, S. Lowe, DWinkler, and D.Zimmerman arewith Alliance Technobgies
Corp., Chapel Hill, NC27514.
ฃ Sue KImbrough Is the EPA Project Officer (see betow).
The complete report, entitled 'Revised Emissions Estimation Methodologies for
Industrial, Residential, and Electric Utility Stationary Combustion Sources"
(Order No. PB93-135663/AS; Cost: $27.00; subject to change) will be available
only from
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at
Air and Energy Engineering Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
United States
Environmental Protection Agency
Center for Environmental Research Information
Cincinnati, OH 45268
Official Business
Penalty for Private Use
$300
BULK RATE
POSTAGE & FEES PAID
EPA
PERMIT No. G-35
EPA/600/SR-92/239
------- |