Historic Emissions Of Sulfur and Nitrogen Oxides In The United States, 1900 To 1980

United States
Environmental Protection
Agency
Air and Energy Engineering
Research Laboratory
Research Triangle Park NC 27711
Research and Development
EPA/600/S7-85/009 May 1985
Project  Summary
Historic  Emissions  of Sulfur and
Nitrogen Oxides  in  the
United States from  1900 to 1980

Gerhard Gschwandtner, Karin C. Gschwandtner, and Kevin Eldridge
  Historic emissions of sulfur dioxide
(SO2) and nitrogen oxides (NOX) were
estimated for Task Group B, Manmade
Sources, of the National Acid Precipita-
tion Assessment Program for each state
of the conterminous U. S. The emissions
were estimated by  individual source
category on the state level from 1900 to
1980 for every fifth year and for 1978.
The source categories included power
plants, industrial boilers, industrial pro-
cesses,  commercial and residential
heaters, natural gas pipelines, highway
vehicles, off-highway diesel engines,
and all other anthropogenic sources.
These emissions were calculated from
salient statistics indicative of fuel con-
sumption or industrial output, estima-
tions of average statewide fuel proper-
ties, and estimations of emission factors
specific to each source category over
time. The emission estimates were then
aggregated to show the emission trends
by state, region and all states combined.
Total state emissions for each year were
then  estimated using an interpolation
procedure based on national annual fuel
consumption.


  This Project Summary was developed
by EPA's Air and Energy Engineering
Research Laboratory, Research Triangle
Park, NC, to announce key findings of
the research project that is fully docu-
mented in a separate report of the same
title (see Project Report ordering in-
formation at back).
Introduction
  Sulfur oxides (SO,) and nitrogen oxides
(NOX) are considered primary precursors
of acidic precipitation. The anthropogenic
emissions of these pollutants are sus-
pected  causes of  many biological and
chemical effects observed in recent years.
Understanding the historic emission
trends is important to understanding the
development of acid-precipitation-related
problems and causes of observed environ-
mental effects.
  Annual quantities of emissions of SO*
and NOx are presented for each of the
contiguous 48 states and the District of
Columbia. Emissions of each pollutant
were estimated for every fifth year from
1900 to 1980 and for 1978. The period
from 1900 through 1980 was selected to
allow study of early alkalinity measure-
ments and also to allow comparison with
the 1980 national emission  inventories
being developed under the National Acid
Precipitation Assessment Program
(NAPAP). Five-year intervals were selected
to provide an indication of the emission
trends sufficient for most effects studies
and to develop a methodology that could
be applied to all other years. The state
level was selected because it provides the
most complete and consistent body of
information on a historic basis and col-
lectively covers all geographic regions of
the country.
  For each state the estimates are based
on the apparent annual consumption rate
of fuels. The fuels include bituminous
coal, anthracite,  lignite,  residual and
distillate oils, natural gas, wood, gasoline,
diesel fuel, and kerosene. The consumers
of these fuels, which are also the emitters
I

-------
of SOx and N0«, are categorized as electric
utilities,  industrial boilers, commercial
and residential furnaces, pipelines, high-
way vehicles,  railroads,  coke plants,
smelters, vessels, or other major sources.
Emissions were also estimated for indus-
trial processes based on production rates,
wildfires,  and a  miscellaneous source
category. Collectively, these source cate-
gories account  for  all anthropogenic
emissions in each state.
Method
  Average emission rates for each study
year were calculated for individual source
categories for each state. The source
categories are listed in Table 1 according
to the type of fuel consumed. These
categories represent all types of boilers,
furnaces, engines, processes, and other
anthropogenic sources. The basic steps
involved  in calculating state emissions
are:

  1.  Obtain state level information on
     fuel use.
  2.  Allocate fuel quantity used by each
     source category.
  3.  Develop source category emission
     factors.
  4.  Determine fuel sulfur content by
     state for each category.
  5.  Calculate emissions, after emission
     controls.

  The actual procedure varied somewhat
depending on the usefulness and avail-
ability of information. It can generally be
described in more detail for two time
periods: (1) 1950 to 1980; and(2)1900to
1945.

Approach for 1950 to 1980
  Fuel Consumption—For electric utili-
ties, state consumption  rates  of fossil
fuels were derived by  individual power
plant from the Bureau of Census, the
American Petroleum Institute,  the  U.S.
Department of  Energy (DoE),  and the
National Coal Association. The consump-
tion rates were determined according to
boiler type. For all other categories, except
smelters and miscellaneous  sources,
annual fuel consumption  rates were
obtained  for the source category as a
whole from various publications. When
fuel consumption data were not available,
other salient statistics (e.g., fuel sales,
demand,  distribution,  shipments)  were
used. For the highway vehicles category,
vehicle miles traveled were used for 1970
Table 1.    Fuel Types and Emission Source
           Categories

Bituminous    Electric Utilities
  Coal:       Industrial Boilers and Space
               Heaters
             Commercial and Residential
               Uses
             Steam Railroads
             Coke Plants
Anthracite
  Coal:

Residual Oil:
All Uses

Electric Utilities
Industrial Boilers and Space
  Heaters
Commercial and Residential
  Uses
Vessels
Distillate Oil:   Electric Utilities
             Industrial Boilers and Space
               Heaters
             Commercial and Residential
               Heating
             Railroads
             Vessels

Natural Gas:   Electric Utilities
             Industrial Boilers and Space
               Heaters
             Pipeline Compression
               Stations
             Commercial and Residential
               Uses

Wood:        Electric Utilities
             Industrial Boilers and Space
               Heaters
             Commercial Heating
             Residential Wood Stoves and
               Fireplaces

Gasoline and  Highway Vehicles
  Diesel:      Off-Highway Vehicles
             Vessels

Other:        Wildfires
             Cement Plants
             Copper, Lead, and Zinc
               Smelters
             Miscellaneous Industrial
               Processes
             Miscellaneous Other Sources
to 1980 because these provided a better
estimate according to the mix of vehicle
types.  For earlier years,  gasoline con-
sumption by state was used.  For the
wildfire category, total forest area burned
in each state was used. For smelters and
miscellaneous industrial processes, esti-
mates were based on either individual
plant or state production rates. For 1950
to  1980, state-level  fuel  consumption
data were  available for  most source
categories.
  Emission Factors—The state-level data
were then multiplied by specially derived
emission factors to yield estimates  of
uncontrolled emissions. First, the most
recent emission factors were obtained for
stationary and mobile sources reported
by the EPA. These factors are based on
actual emission tests of each  type  of
combustion process or emission source
represented  in  each  source category.
They are most appropriate when applied
to a large number of sources (e.g., on the
state level). Periodically, they are revised
by the EPA to include  new, additional, or
improved test data.
  The  factors for each pollutant were
then adjusted to represent each source
category as  a  whole.  This  procedure
involved mathematically weighting each
factor according to the amount of fuel
consumed  by various types  of  boilers,
furnaces,  engines,  processes, or  other
emission sources comprising the cate-
gory. For highway vehicles, NO* emission
factors were state specific  and  were
weighted  according to the amount  of
urban and rural traffic, state elevation,
vehicle mix, and pollution controls in use.
These adjustments provided the most
representative factors for 1970 to 1980
for  which vehicle   miles traveled were
available. For earlier years, vehicle miles
traveled are not available; the factors
were based instead on gasoline consump-
tion and on estimated average miles per
gallon for both urban and rural traffic.
These factors yield generally the  same
results as those in  a  DoE report on the
trend in internal compression ratios  of
vehicle motors.
  SOa emission factors were also weight-
ed  according to fuel consumption  by
individual emission sources within each
category.  However, these factors  are
more dependent on fuel properties than
on combustion sources and include a fuel
sulfur content variable. The emission
factors account for the fraction of the fuel
sulfur that would be  emitted as uncon-
trolled  emissions  and the  remaining
fraction that would be captured in the
solid residue. These fractions are deter-
mined on the basis of source emission
tests and materials balance  analysis
involving various  coal ranks  that are
commonly  used by  each source. The
average statewide values of  sulfur con-
tent of coal on the  consumer level were
obtained from the  Bureau of Mines for
1965.  For  the  earlier years, average
statewide values were calculated from
fuel distribution reports and information
on fuel properties by  originating district.
The average values of sulfur content of

-------
fuel oils were obtained from information
published by the Bureau of Mines and the
U.S. DoE, Bartlesville Energy Technology
Center, for domestically produced fuel
oils.
Emission Controls—The amount of
emissions controlled by certain control
devices was then subtracted from each
source category. This step pertained to
S02 which is controlled by flue gas
desulfurization systems at power plants
and by-product sulfuric acid plants at
smelters. Controls applied to sources of
NOX emissions have generally had little
effect in reducing emissions through
1980.
Estimates were then compared with
EPA national emission estimates, with
the NAPAP emission inventory for 1980,
and with the estimates of E. H. Pechan, et
al. for electric utility emissions. This
provided an indication of the precision of
estimates for common years and a basis
for establishing the precision for earlier
years.

Approach for 1900 to 1945
For this period, state-level data on
actual fuel consumption by source cate-
gory were not always available, especially
for the earlier years. Also, the method for
collecting and reporting early data was
not always consistent with the method
for more recent years. Depending on the
type of information found, one of three
approaches was taken:

1. State-level data were used when
available.
2. National data were apportioned to
the states.
3. No estimates were made when
state and national data were un-
available and when the emissions
were so small as to be considered
negligible.
These approaches help account for most
of the early SO2 emissions which were
dominated by coal usage and for which
consumption data are available either on
the state or national level. Most NO,
emissions are also accounted for by this
approach, but in, terms of quantity are
comparatively less than SO2 because of
the low consumption rates of fuel oils and
natural gas in the early study years.

Assumptions
The same sulfur content values derived
for 1955 were assumed to apply to the
earlier years. This assumption was nec-
essarily made because no evidence was
found to suggest a general trend in sulfur
content. Available information indicates
that sulfur content of coal as mined did
not change significantly, and most coal
was consumed in or near the producing
states. Analysis of coal distribution pat-
terns also suggests little change com-
pared to the changes in the middle or
recent part of the century.
It was also assumed that the emission
factors used for 1955 applied to earlier
years. No evidence was found to suggest
a change in either the emission charac-
teristic of coal-fired sources or the popu-
lation mix of types of boilers and furnaces.
As research in historic emission patterns
and trends continues, this general as-
sumption may be replaced by specific
state-level data if such data can be
developed.
Aggregation of Emissions
Historic fuel consumption data were
tabulated for each state according to
source category and study year. Corre-
sponding fuel sulfur content values were
also tabulated for each source category
according to state and year. A third tabu-
lation contained the weighted emission
factors for S02 and NO* by source cate-
gory and in some cases by state. These
three matrices were multiplied to produce
two new matrices, one for S02 emissions
and one for NO, emissions. The emissions
of each state were then totaled by year to
provide an estimate of overall national
emission trends. They were also totaled
by fuel type and by source category to
show the effects of fuel switching and
changes in consumer sectors. State emis-
sions were also aggregated to show the
trends in broad geographic regions of the
country.
The national emissions were then
analyzed by season and also by stack
height ranges. For the seasonal analysis,
the percentage distribution of the emis-
sions by season was estimated for each
major source category based on engi-
neering judgment and known historic
characteristics of each source category to
give a general indication of the trend
since 1900. The total national emissions
of each source category were then multi-
plied by these percentages. As a result,
the estimated seasonal emissions reflect
both the trend in total emissions by source
category and the general change in the
seasonal distribution of emissions by
source category.
For the release height analysis, the
percentage distribution of national emis-
sions was estimated for each source
category according to four broad ranges
of stack heights. In the case of electric
utilities, individual power plant emissions
and stack height data were used to
determine the national distribution by
height from 1950 to 1980. For earlier
years and for other sources, the analysis
was based on the general trend in the
stack heights for the category as a whole.
Both the seasonal and stack height
analyses provide an approximate indica-
tion only of the trend on the national level.

Yearly Estimates
State total emissionsforthe intervening
years were interpolated from the state
emissions estimated for the study years
and the annual national energy consump-
tion reported by fuel type. The interpola-
tion was performed individually for each
major fuel category by state. For each
intervening year, the emissions of each
fuel category were then added.
Results for 1 900 to 1 980
National and regional trends of each
pollutant are presented here by fuel type
and source category.

Fuel Consumption,
Overall and for Categories
Figure 1 shows the total mineral fuel
consumption for the U.S. in terms of
energy consumed by major source as
obtained from the Bureau of Census.
Total coal consumption has remained
relatively constant over time since 1900
compared to the consumption of other
fuels. Since 1 960, coal consumption has
steadily increased on the national level by
30 percent. Crude petroleum and natural
gas consumption (the primary sources of
NO, emissions) has increased most rapid-
ly since the 1 930s. Wood and anthracite
accounted for a large portion of the total
energy consumption in the early part of
the century but not in the latter part. Until
1930, per cap
-------
Notes: 1 Btu = 1.055 kJ. The petroleum category
in the figure includes crude petroleum and
petroleum products as consumed, minus the
consumption of gasoline and diesel fuel which is
shown separately.
Gasoline
& Diesel
Petroleum
(see note)
Bituminous Coal
, |l.l ,, • II I ,. 1 1 II •...,•,, II 1 ,..,,.., nil ...... • II I 1 ,11111.1! I,.
7900 79/0 7920 7930 7940 7950 7950 7970 7950

Year

Figure 1. Total mineral fuel consumption of the United States by major source: 1900to 1980.
study years.) Overall, the SO2 trend
follows the general trend of coal con-
sumption except that total emissions
appear to have decreased by 10 percent
from a maximum around 1970. This
decrease is somewhat due to the general
decrease in sulfur content of fuels and
emission reductions brought about by
national and state environmental control
regulations. Sulfur content has decreased
to a large extent as a result of coal
cleaning and mixing eastern coal with
cleaner western coal, while in the early
years coal was mostly burned as received
from the nearest coal-producing district.
Figure 3 shows the overall trend by
source category. This plot reflects the
growth of major fuel-consuming sectors
and changes in fuel demand. For example,
electric utility emissions appear to have
increased sharply by the 1950s and 60s.
In contrast, 862 emissions from steam
locomotives almost completely disappear-
ed by 1950 with the advent of diesel-
powered engines.
Emissions,
Overall and for Categories
In contrast to SOa, total NO, emissions
appear to have increased constantly
throughout most of the study period as
shown in Figure 4. The total quantity of
emissions is plotted on the same scale as
SOz to allow a direct comparison. This
upward trend is primarily a result of
greater use of natural gas and petroleum
products and a conversion away from
coal. Figure 5 shows that the increase is
largely due to the growth in the number of
highway vehicles, natural-gas-fired
power plants, and many other sources
related to a large extent to a growth in
population and changes in technology
and lifestyles.
Analysis by Region
The total state emissions were aggre-
gated according to the Administrative
Regions of the USEPA shown in Figure 6.
These regions represent various broad
geographic regions of the country. It
should be noted that the regions vary in
size and in the number of states and that
these two factors will also affect the total
regional emissions. By selecting a differ-
ent combination of states other than these
Federal regions, different emission trends
may be shown. Recognizing this fact, the
regions were selected to provide only a
general indication of trends in various
regions of the country.
The overall emission trend of each
pollutant and the trend by category are
plotted in Figure 7 for each region. These
plots provide resolution of the national
trend and allow the historic emission
trends of each region to be compared. For
example, Regions 3, 4, and 5 appear to
have historically emitted more SO2 than
other regions in terms of total quantity.
The total S02 emissions of Regions 1 and
2 combined have historically remained
constant. In Region 6, NO* emissions
have increased more rapidly than in any
other region due to the growth in the
natural gas production industry and the
number of pipeline compression stations.
In all other regions, highway vehicles and
electric utilities together have accounted
for more than half the total NOX in the past
several decades.
In 1950, the regions east of the
Mississippi River emitted 75 percent of
the total national SOz emissions and 67
percent of the total NOX emissions. In
1980, the eastern regions emitted 77
percent of the total SO2 and 60 percent of
the total NOX. During this period, total
national SOa emissions increased 140
percent while total NOX emissions in-
creased 280 percent, or twice as much.
While most of the emissions have histor-
ically originated in the east, the western
regions have begun to emit a greater
share of the total national NOX in recent
years.

Analysis by Emission
Release (Stack) Height
Analysis of emissions by release height
(actual stack height) is important to
studying the potential for long-range
transport. Note that the potential for long-
range transport increases with each
higher range. This analysis does not
include stack exit velocities or atmos-
pheric mixing heights which are also
important considerations. The analysis in
-------
this study suggests that more S02 emis-
sions were released into the atmosphere
from stacks above 240 ft* than from
stacks below this height since about
1945. By 1980, approximately 30 percent
of the S02 emissions were emitted above
480 ft, for example, compared to only 5
percent above this height in 1950. Not
only have the percentages increased, but
total national S02emissionsalso increas-
ed and peaked around 1970. The percen-
tage of the total S02 emissions released
below 120 ft has generally decreased
over the study period. The distribution of
NOX emissions has historically remained
constant, although on the national level
the total emissions have steadily increas-
ed. Approximately 60 percent of the total
NO, emissions in 1980 were released
from ground level sources; predominantly
from transportation sources.
Analysis of the electric utility category
suggests that in the 1950s and 60s, most
of the S0a and NO, emissions from this
category were released below 480 ft*—
mostly between 240 and480 ft. By 1980,
about 50 percent of the total S02 emis-
sions and 40 percent of the NO, emissions
from this source category were released
above 480 ft as a result of the trend
toward taller stacks. Since the emissions
from electric utilities constitute a large
portion of the total national emissions in
recent years as shown in previous figures,
they have a significant effect on the
overall distribution of emissions by re-
lease height.

Conclusion
The current historic emissions data file
presents the best estimates available on
the state level. The emission trends of
each state vary over time reflecting
changes in a variety of economic and
technological factors. While the national
and regional scale emissions data provide
general indications of trends, it is recom-
mended that the reader refer to the state-
specific estimates presented in the full
report for studies of the historic relation-
ship between emissions and environmen-
tal effects. These emission estimates can
serve as the basis for future studies of the
relationship between emissions and en-
vironmental effects associated with acid
precipitation phenomena.
30000. Notes: j ton = go? kg. The bars in this figure are
interconnected only to highlight the overall trend.
25000.
S 20000.
Other

Residual Oil
Bituminous
Coal
0. 5. 10. 15. 20. 25. 30. 35. 40. 45. 50. 55. 60. 65. 70. 75. 80.

Year 78