RL30036: Global Climate Change:
Carbon Emissions and End-use Energy Demand
Senior Specialist in Science and Technology
Science, Technology, and Medicine
January 20, 1999
List of Tables
The United Nations (Kyoto) Protocol on greenhouse gas emission reductions sets a target
for the United States to achieve annual emissions of six greenhouse gases, as measured in
terms of their equivalency to carbon dioxide, over the 2008-2012 period that are 7% below
specified baseline years. The largest contributor to greenhouse gas emissions is the
combustion of fossil fuels for energy production to power a wide variety of end-uses such
as automobiles, space heating, and industrial process heat. This report presents an
analysis of the potential impacts of meeting the Kyoto Protocol targets on those end-uses.
Demand for each energy source is calculated using Energy Information Administration
(EIA) data and forecasts for 1996, 2008, and 2012 for 27 common end-uses making up the
four major sectors: residential, commercial, industrial, and transportation. Carbon
emissions are then determined using carbon-emission coefficients for each fossil fuel.
Finally, energy demand reduction requirements are calculated by applying the Kyoto
Protocol target of a 7% reduction in carbon emissions from 1990 levels to the 2008 carbon
emission levels for each end-use.
Based on EIA forecasts, by 2008, total carbon emissions from energy use by the 27
end-uses is calculated to be 1,721 million metric tons carbon-equivalent (MMTCE) compared
to 1,464 MMTCE in 1996. Five end-uses -- light duty vehicles (primarily automobiles, sport
utility vehicles, small trucks, and vans), freight trucks, residential miscellaneous
(small appliances and outdoor machinery), industrial machine drive, and miscellaneous
commercial (communications and information equipment) -- would comprise over 70% of the
1996-2008 increase in carbon emissions.
The Kyoto Protocol target would require average carbon emissions of 1,252 MMTCE from
the 27 end-uses over the 2008-2012 period if it were applied uniformly to all sources of
greenhouse gases. In this case, energy demand for each of the end-uses would have to
decline by about 28.7% below the levels now forecast for 2008. Further, resultant energy
demand for each of the end-uses would be about 20%, on average, below the actual 1996
values. Finally, on average, the required reductions from the current 2012 forecast would
be about 31%.
Reductions of that magnitude would require substantial increases in energy efficiency,
above those already forecast, and/or significant reductions in the services provided by a
given end-use. For example, if the target were to be met solely by increases in
efficiency, the average fuel economy of the light duty vehicle fleet would have to grow
from the current forecast value of about 20.3 miles per gallon (mpg) to over 29 mpg by
2008. If the reduction were to be met solely by driving less, annual passenger car
(passenger cars consume about 70% of the fuel used by light duty vehicles) travel would
have to drop from the current forecast of 14,500 miles per car to about 10,300 miles per
car. Such actions would be a substantial undertaking and consumers are likely to feel
significant effects from any strategy that is used to try to meet the Kyoto Protocol
The potential for global climate change from the buildup of greenhouse gases in the
Earth's atmosphere has elicited considerable concern by the world's nations.(1) A major source of that concern is the
contribution to that buildup by greenhouse gases resulting from human activity.
Historically the predominant greenhouse gas from human activity has been carbon dioxide
(CO2) although other gases have been gaining in importance in recent years.
Nevertheless, CO2 is expected to remain the largest single contributor to
greenhouse gas buildup from human activity for at least the next 60 years.
The principal source of such CO2 is energy use. It is a byproduct of
combustion of fossil fuels. In 1996, the Department of Energy (DOE) estimated that the
United States produced 1,462 million metric tons-carbon equivalent (MMTCE)(2) of CO2 from the combustion of 85.2 quadrillion
Btus (Quads) of fossil fuels.(3)
The growing concern about greenhouse gas-induced global climate change has prompted
major international efforts to limit the buildup. The most recent of these was the
December 1997 United Nations (Kyoto) Protocol on Global Climate Change, which established
greenhouse gas reduction targets for the developed nations signatory to the accord.(4) That agreement, which the United
States has signed but not ratified, would require the United States to achieve average
annual carbon-equivalent emissions of six greenhouse gases over the period 2008-2012 that
are 7% below specified baseline years. For CO2, the baseline year is 1990.
According to DOE, the United States produced 1,353 MMTCE in 1990 from energy use.
There have been several analyses of the implications of meeting the reduction targets
set by the Kyoto accord. Most recently, the Energy Information Administration (EIA) of DOE
reported on the impacts on the U.S. energy markets and economy of those reductions.(5) In addition, a DOE study carried out
by five of its national laboratories examined the potential for new energy supply and
demand technologies to help meet those targets with a minimum of economic dislocation.(6) Those studies carried out detailed
examinations of U.S. energy demand in analyzing the potential impacts of CO2
With some exceptions, however, the detail did not extend to the specific end-uses --
such as motor vehicles or air conditioning -- that consumers are familiar with. Although
the Interlaboratory (called the 5-lab) study looked at specific end-use technologies, it
did not provide data on end-use energy demand and CO2 emissions for all major
end-uses. While such information is not necessary to estimate potential economic impacts
of CO2 emission reduction, it can be helpful in gaining an understanding of the
potential consequences of such actions at the consumer level.
Report Purpose and Format
This report presents estimates of actual and forecast energy demand for all of the
common energy demand end-use categories for 1996, 2008, and 2012. It then calculates the
CO2 emissions for these estimates. With these values the reader should be able
to see which, of the several ways energy is used in this country, are the major sources of
CO2 emissions. A similar report was prepared in 1991, which presented energy
demand and CO2 emissions estimates for 1988 and 2000.(7)
This report goes further than the previous report by presenting an analysis of the CO2
emission reductions called for by the Kyoto accord. This analysis is based on a
spreadsheet model that estimates the CO2 emissions reductions for each of the
end-use categories for both 2008 and 2012, and calculates the energy demand reductions
needed to meet those emissions targets. In this way, the reader can see just how any of
the common energy demand end-uses would be affected by reaching the targets. The report
concludes with a discussion of implications of these reductions for representative
Energy Demand by End-use Category
The EIA divides U.S. energy demand into four sectors: industrial, residential,
commercial, and transportation. Each sector is characterized by sources of energy supply
and by specific end-uses. The sources of energy are the fossil fuels, renewables, and
electricity. Electricity, of course, requires primary energy sources, namely fossil fuels,
nuclear fuels, and renewables. End-uses are defined as specific functions or equipment
that perform services for consumers such as cooling, heating, rail transport, electric
motors, and lighting. Table 1 presents the energy sources and end uses considered in this
report. There are 27 end-uses in all. The term in parentheses after each end use is a
symbol used to denote the end-use in graphical representations to follow. The end-uses are
not presented in any special order. Also, the energy sources in the right column are not
meant to correspond to the end-uses in the left column. A given end-use usually requires
more than one source of energy.
While most of the end-uses are self-explanatory, some need further explanation. The
miscellaneous end-use in the residential sector includes motors and heating elements
commonly found on gardening equipment, machine tools, and small appliances. Appliances in
this sector include refrigeration, cooking, freezers, clothes washers and dryers, color
televisions, and personnel computers. In the commercial sector, miscellaneous includes
electronic office equipment, telecommunications, medical equipment, service station
equipment, and manufacturing performed in commercial buildings.
In the industrial sector, direct heat refers to manufacturing processes requiring the
direct application of heat such as metal treating and chemical production. Machine drive
refers primarily to the use of electric motors to control manufacturing processes. Steam
is used in manufacturing primarily as a heat source for chemical and metallurgical
processes, as a power source for metal-forming equipment, and to drive turbines.
Electrolysis is an electro-chemical process used primarily in aluminum production.
In addition to those manufacturing processes, the industrial sector contains
construction, agriculture, and mining. Each of these "subsectors" contains
several end-uses, but lack of good data prevents an accurate disaggregation. Therefore,
each is counted as a separate end-use. Finally, the industrial sector uses a significant
quantity of petroleum fuels for products such as asphalt, liquified petroleum gases,
petrochemical feedstocks, and lubricants. Most of the carbon in these nonenergy
applications is not released as CO2 emissions, and the nonenergy end-uses are
not included in the carbon emission analysis to follow.(8)
In the transportation sector, light duty vehicles include all vehicles weighing less
than 8,500 pounds. These include automobiles, minivans, sport utility vehicles, larger
passenger vans, and small trucks. Freight trucks include all commercial and freight trucks
weighing over 8,500 pounds.
In each sector, electricity is given as one of the energy sources. Electricity, of
course, requires primary energy sources for its generation. These sources are distillate
and residual fuel oil, natural gas, coal, and nuclear and renewable energy. In the
calculations to follow, it is assumed that for a given year, the mix of energy sources to
generate electricity is the same regardless of which sector uses the electricity. The mix
will change over time, however, and those changes are incorporated.
Energy demand in the United States for 1996 was adopted for the baseline because that
is the last year for which complete data are available for all sectors. For the
residential and commercial sectors, data for end-use energy demand by fuel are directly
available from the EIA.(9)
Consolidation of some of the end-uses provided by EIA was made by combining several
household appliances in a category called residential appliances, and combining all
commercial categories labeled other uses with commercial office equipment in a category
called commercial miscellaneous. These new categories are described above.
For the transportation sector, the EIA provides separate reporting of total energy use
by the various categories and by energy source.(10)
As a result, estimates have to be made of how much of a given energy source are used by
any end-use. This can be done by noting that there is a predominate energy source for a
given end-use and that most of the categories will use no more than two different kinds of
energy sources. For example, light duty vehicles will use primarily gasoline with a small
amount of distillate, air transport will use all of the jet fuel and a small amount of
gasoline, and rail is the primary user of electricity but also uses residual fuel oil. By
reconciling the total energy demand for each end-use with that for each energy source, an
accurate picture of energy demand by energy source for each end-use can be obtained.
For the industrial sector, the calculation is more complex. The sector is made up of
four subsectors: manufacturing, construction, agriculture, and mining. Detailed energy use
data by end-use exist only for the manufacturing sector. The EIA publishes a survey of
manufacturing energy use every three years, the most recent for 1994. The report provides
data on the major end-uses by energy source.(11)
A complication arises in that a large fraction of the totals reported by end-use and by
energy source are not specified. By going to the individual industry groups much of those
unspecified values can be estimated from noting the types of processes used by those
industries. For example, unspecified fuel in the paper and paper products industry is
likely to be wood used to produce steam. In petroleum refining, the unspecified fuel is
likely to be still gas, a product of the refining process, used for direct heat and steam.
Once energy demand by end-use and energy source is determined for 1994, the 1996 value can
be estimated by adjusting each energy source by the 1994-1996 growth rate. That method
assumes that there is no significant shift of the type of energy source used by the
end-uses over that period.
For mining, the census of mineral industries published by the U.S. Census Bureau of the
Department of Commerce provides data on energy use for 1992 by energy source for the
entire mining industry.(12) No data
are provided by end-use. For that reason, the entire mining industry is included as a
separate end-use. To obtain values for 1996, the value of each energy source is adjusted
by its the 1992-1996 growth rate. For construction and agriculture, no energy use data are
available. Total energy demand by energy source for the entire industrial sector, however,
is available.(13) By subtracting
total energy demand by energy source for the other two subsectors from the industry total,
and considering construction and agriculture as one end-use, its energy demand can be
estimated as the residual.
For the years 2008 and 2012, the EIA forecasts are used to estimate energy demand by
end-use and energy source.(14) The Annual
Energy Outlook, 1998 provides forecasts for the years 2005, 2010, and 2015, among
others.(15) To obtain forecast
estimates for 2008 and 2012, therefore, it is first necessary to determine the forecast
for energy demand by end-use for 2010.(16)
Energy demand forecasts for the residential and commercial sector end-uses are obtained
directly for 2010 just as described above for 1996. Similarly, estimated energy demand by
end-use for the transportation sector for 2010 can be obtained from the EIA data using the
same methods as for 1996.
For industry, a different method must be used because there are no forecasts of
manufacturing energy demand by end-use. The EIA, however, does provide forecasts of energy
demand by energy source for the entire industrial sector. By adjusting each energy source
for a given end-use by the 1996-2010 growth rate for that energy source for the entire
sector, an estimate of the energy demand for that end-use for 2010 can be made. This
method assumes that the relative mix of energy sources for a given end-use does not change
significantly over that period. That is a valid assumption because a major energy-source
mix change in one end-use would require an equal and opposite change in one or more of the
remaining end-uses in order to keep the totals for the sector unchanged. Such actions are
quite unlikely, particularly over a 14-year period.
Once the forecasts for 2010 are obtained, they can be modified to provide estimates of
the forecasts for 2008 and 2012. This modification is performed by first calculating the
annual growth rate between 2005-2010 and 2010-2015 for each energy source by sector. The
data for these calculations are obtained from the EIA Annual Energy Outlook, 1998.
With those growth rates, it is a straightforward matter to adjust each of the energy
sources for each of the end-uses by the appropriate annual growth rate. This method
assumes that the 2005-2010 and 2010-2015 growth rates for a given energy source are the
same for all end-uses in a sector. In the cases where that assumption can be checked --
the residential and commercial sectors -- it is not strictly correct. The error introduced
for those two sectors, however, will be small because the growth rates themselves are
small, 1% per year and less. For the other two sectors, it is unavoidable, given the data
available, but also quite small.
The results of those calculations are shown in Figure 1 (on next page) for all 27
end-uses. They are arranged in descending order of energy demand forecast in 2012. The
results present primary energy demand for each end-use; waste heat produced in the
generation of electricity is assigned to the electricity total for each end-use.(17) Tables presenting the actual data
are in the Appendix to this report.
Total energy demand represented by all 27 end-uses in 1996 -- 87.5 Quads -- is nearly
equal to the value of total energy demand reported by EIA -- 88.1 Quads -- indicating that
these 27 end-uses virtually capture the full range of U.S. energy demand. For 1996, EIA
also reports that non-energy uses, as discussed above, consumed 5.95 Quads of fossil
fuels. For 2008, energy demand for the 27 end-uses is 101.5 Quads and for 2012 it is 105.7
Quads. The EIA estimates for total energy demand including nonenergy uses of fossil fuels
for those two years are 109.6 Quads and 113.6 Quads, respectively. These values would
indicate that nonenergy use for 2008 and 2012 is 8.1 Quads and 7.9 Quads respectively.
Based on extrapolation of past trends those values are reasonable, although it is unlikely
that nonenergy, fossil fuel use would decline from 2008 to 2012. The value for 2008 is
probably a little high, while that for 2012 somewhat low. This means that the estimates
for energy demand determined from summing the end-uses is somewhat low for 2008 and
somewhat high for 2012. The discrepancy is small, however, and very likely to be within
the EIA forecast errors.
Of the individual end-uses, the one with the largest energy demand is light duty
vehicles. It is nearly twice the size of the next largest end-use, direct heat for
industrial processes. Furthermore, its energy demand is forecast to grow significantly
from 1996 to 2012 primarily as a result of a substantial increase in vehicle miles
traveled.(18) Other end-uses that are
expected to grow sharply between 1996 and 2012 are the miscellaneous categories in both
the residential and commercial sectors. A rapid expansion in office electronic and
telecommunication equipment, and in a variety of small residential appliances and outdoor
heating equipment and motors is expected to be responsible for that growth. Electricity is
by far the dominant energy source for those miscellaneous end-uses. Those three end-uses
-- light duty vehicles and the two miscellaneous categories -- constitute 48% of the
projected growth in energy demand between 1996 and 2012.
The first step in estimating CO2 emissions is to determine the carbon
emission coefficients for the fossil fuels used as energy sources. Each fossil fuel has a
characteristic CO2 production rate, or emission coefficient, determined by the
chemistry of that fuel. That rate is the amount of CO2 that will be produced
upon complete combustion of a specific quantity of fuel. Those rates are given in Table 2
in terms of millions of metric tons of carbon (MMTC) produced per Quad of energy used.(19) (20) The coefficients for a given fuel can change from
year-to-year depending on the quality of the fuel produced that year. The changes will be
small, however, and the EIA has reported no changes for petroleum products and natural gas
over the period 1986 to 1996.(21)
There have been slight changes for coal, but they are very small and possible future
changes will not be considered in this report. Therefore, the carbon emission coefficients
for fossil fuels are assumed to remain constant over the period covered by this report,
1996 to 2012.
For electricity, the coefficient is calculated from the coefficients of the fossil
fuels used in the generation mix, weighted by their contribution to that mix. This mix
will change over time. Using EIA forecasts for the generation mix in 2010, the carbon
emission coefficient for electricity for that year can be calculated. Note that the change
between 1996 and 2010 is small. While the value for 2008 and 2012 can be estimated by
using the 2005-2010 and 2010-2015 fossil fuel growth rates given by the EIA, the changes
from 2010 are likely to be insignificant. Therefore, the 2010 value is used in calculating
carbon emissions for 2008 and 2012.
Once the carbon emission coefficients are calculated, it is relatively straightforward
to calculate the total amount of carbon emitted by each end-use in the selected year. For
each end-use, the quantity of energy from a each energy source is multiplied by the
appropriate carbon emission coefficient and the results are summed. Using those
coefficients, of course, will yield carbon emissions not CO2 emissions.
Although it is simple to convert to CO2 emissions (see footnote 19), this will
not be done in order to be consistent with the way the EIA presents emissions data.
The results of this calculation are given in Figure 2. The detailed data are provided in
tables in the Appendix. The data in the figure are displayed by order of end-use emission
rankings forecast for 2012. Total carbon emissions in 1996, calculated by aggregating
end-uses, were 1,464 million metric tons carbon equivalent (MMTCE). The EIA reports that
total 1996 carbon emissions, including those from non-energy use of fossil fuel, were
1,463 MMTCE.(22) The CRS estimate
might be slightly high because no emissions from non-energy uses of fossil fuels were
included. As noted above, about 20% of the carbon in those fuels was released to the
atmosphere. This would amount to carbon emissions of about 15 to 20 MMTCE. Nevertheless,
the value calculated here is quite close to the EIA total indicating that the end-use
structure is valid.
For 2008 and 2012, aggregation of the end-use carbon emissions yields totals of 1,721
and 1,795 MMTCE respectively. The EIA forecasts for these two years are 1,757 and 1,837
MMTCE respectively. Considering the effect of non-energy uses of fossil fuels as discussed
in the previous paragraph, the results obtained from aggregating end-use emissions appear
quite consistent with the EIA forecasts.
The largest contributor to carbon emissions was, and is expected to continue to be,
light duty vehicles. That end-use is forecast to contribute almost twice as much carbon as
the next highest end-use, direct heat for industry. These observations, of course, are
consistent with those reported above in the discussion of energy demand. There is not,
however, a direct correlation between total energy demand and carbon emissions. Rather the
energy source mix can play an important role in determining the relationship between
energy demand ranking and carbon emission ranking. In particular, industrial steam and
residential space heat rank higher on the energy demand scale than the carbon emission
scale because both include a significant amount of biomass -- wood combustion -- in their
energy source mix. Wood combustion does not produce any net production of CO2
as long as the wood used for energy production was not previously counted as a sink for CO2.(23)
Nevertheless, there are a number of similarities in the two tables. Light duty
vehicles, and residential and commercial miscellaneous are projected to have the largest
increases in carbon emission between 1996 and 2008, and 1996 and 2012. Other end-uses that
indicate a large increase are freight trucks, air transport, and industrial machine drive.
These five end-uses make up 70% of the forecast increase of total carbon emissions from
energy use between 1996 and 2012.
For some end-uses, carbon emissions are forecast to change very little and, in some
cases, decline. All of the end-uses in the residential and commercial sector, except the
miscellaneous categories, are expected to be nearly the same in 2012 as in 1996. This
level behavior is due primarily to projected increases in energy efficiencies for those
end-uses. Most industry and transportation end-uses, however, are forecast for significant
growth. On average, the rate of forecast efficiency gains for the end-uses in those
sectors does not match their projected energy-demand growth resulting from continued
Kyoto Protocol Targets
In 1997, the United States joined with more than 160 nations to negotiate greenhouse
gas emission reductions targets. The result was the Kyoto Protocol that established such
targets for the Annex I countries.(24)
Those countries are among the ones that have ratified the 1992 United Nations Framework
Convention on Climate Change. According to the Protocol, the United States is to reduce
those emissions to a level 7% below 1990 levels. There are six greenhouse gases covered in
the Protocol and the targets are based on the carbon equivalent of each gas. These gases
are CO2, methane, nitrous oxide, hydrofluorocarbons, perfluorocarbons, and
sulfur hexafluoride. For last the three of the gases, there is an option that 1995 can be
adopted as the base year.(25)
In 1990, total emissions of the six greenhouse gases was 1,618 MMTCE .(26) If the 1995 base year is adopted for the three gases,
the total becomes 1,629 MMTCE. Of the total, 1,374 MMTCE is CO2, of which 1,346
MMTCE comes from fossil fuel combustion for energy use. The nonenergy CO2 is
from certain industrial processes, primarily cement production and limestone consumption.
There are considerable uncertainties involved in calculating the targets and reduction
quantities required.(27) One of the
principal ones is how the reduction will be allocated among the various gases. It might be
possible to reduce some gases well beyond the 7% target allowing smaller reductions in
others. In addition, the Protocol allows countries to trade emissions credits.(28) That process could allow a
loosening of the target level for the United States.
One of the purposes of this report is to show what would happen to energy demand levels
for the common end-uses if the targets were applied uniformly. Therefore, to determine the
required carbon emission targets, the 7% reduction will be taken from carbon emissions
resulting from energy production in 1990. This target level is 1,252 MMTCE. One other case
has been considered, that of a 3% reduction from the 1990 levels. This reduction level was
estimated by the U.S. Department of State and cited by the Council of Economic Advisors as
possible given the flexibility inherent in the Protocol.(29) This target level is 1,306 MMTCE. The difference between
the two target levels is small, however, and will not significantly affect the
implications of reaching these levels for any of the end-uses.
In addition to determining the target level, there is also the question of when the
reductions would start. The Protocol states that the average emissions over the five-year
period must equal the target. Therefore, if the 7% reduction is met in 2008, emissions can
remain flat throughout the period. If no reduction has taken place by 2008, the reduction
necessary by 2012 must be substantially greater. This behavior is shown in Figure 3 which
compares different trajectories to the target level. In the extreme case of waiting until
2008 to begin, and assuming a constant percentage decline each subsequent year, the 2012
target level becomes 756 MMTCE, 59% below the current EIA forecast for that year. Starting
in 2005, the case adopted by the EIA assessment,(30)
and assuming a constant percentage decline to the target of 1,252 MMTCE in 2008, would
require a 6.9% per year rate of decline. Starting in 1999 to the same target, would
require a 1.9% per year decline.
It is clear that the degree of difficulty in reaching the target levels would increase
dramatically as the year in which the reductions begins approaches 2008. For the purposes
of this paper, it is assumed that reduction would begin some time before 2008 and reach
the 7% (or 3%) level by 2008 so that carbon emissions are constant over the 2008-2012
To estimate the consequences of reaching the reduction targets of 7% below 1990 carbon
emissions on end-use energy demand in 2008, it is assumed that the reductions are applied
uniformly to all the end-uses. The 2008 carbon emissions of each end-use are multiplied by
the ratio of the target emissions level to the total carbon emissions from all energy use
in 2008 as forecast by the EIA.(31)
The results are the allowed carbon emissions for that end-use to meet the targets set by
the Kyoto Protocol. Once those target emission levels are determined, the energy demand
levels that would be required to produce those target emission levels can be calculated.
First, the average carbon emissions coefficient for each end-use is calculated using the
2008 values.(32) Then, the target
emission levels for a given end-use are divided by its average carbon emission
coefficient, giving the target energy demand level. Finally, the difference between the
forecast level and the target level gives the required energy demand reduction needed to
meet the Kyoto Protocol targets for each end-use.
The results of this calculation are shown in Figure 4 for 2008. The detailed data are in
the Appendix. Because carbon emissions would have to remain constant at 1,252 MMTCE over
that period to meet the Kyoto Protocol requirements, energy demand for each end-use would
also have to remain constant as long as each end-use is sharing the reduction
proportionately. The higher energy demand forecasts for 2012 as shown above mean that the
reductions for each end-use in 2012 would be correspondingly higher. These reductions are
also shown in Figure 4 and the detailed data are shown in the Appendix.
Figure 4 also presents the energy demand reductions from 1996 actual levels that would
result from meeting the Kyoto Protocol requirements. A negative value means that the
target energy demand level in the 2008-2012 period would be higher than the 1996 actual
level. Only two end-uses show such behavior, residential miscellaneous and air transport.
For both, EIA forecasts that energy demand is expected to grow well above the average of
all end-uses. Other end-uses that are forecast to grow rapidly, such as commercial
miscellaneous and light duty vehicles, show relatively small reductions from the 1996
actual levels. For nearly all of the end-uses, however, the target levels for 2008-2012
would be significantly below the 1996 recorded energy demand. Indeed, some end-uses show
larger reductions from 1996 levels than from the 2008/2012 forecasts. Those are primarily
in the residential and commercial sectors and are end-uses where substantial efficiency
gains are expected in coming years such as residential appliances and commercial space
heating. In percentage terms, the reductions that would be required from 1996 actual
levels range from about a negative 25% to a positive 49% with an average of about a
Under the method used in this report to apportion the reductions that would be required
by the Kyoto Protocol, each end-use contributes the same percentage reduction in carbon
emissions and energy demand for 2008. For that year, energy demand would be reduced by
about 28.7%. Because the energy demand target level remains fixed from 2008 to 2012 but
demand growth for each end-use is forecast to grow at different rates over that period,
the percentage reduction in energy demand for 2012 would not be the same for each end-use.
The percentage reductions for 2012 in end-use energy demand would range from about 29% to
41% with most around 31%.
Implications. Reaching the Kyoto Protocol target
for carbon emissions from energy use would result in energy demand in the year 2008 of
about 70.2 Quads.(33) This would be
about 28.3 Quads below the amount the EIA now forecasts for total energy use in 2008 and
about 14.8 Quads below the amount used in 1996. It would be the lowest total U.S. energy
demand since 1988. Taken from the 1996 level, to reach that target by 2008 would require a
decline of about 1.6% per year. If actions do not begin until 2005, as assumed in the EIA
study on impacts of the Kyoto Protocol, energy demand would have to decline by about 10.5%
per year to reach the target level. The longest stretch of declining energy demand in the
United States since 1949 occurred from 1979 to 1983 following the second Arab oil embargo
and oil price spike. Over that period, total energy demand dropped by 10.6% or about 2.8%
per year for the four-year period. While that rate is considerably greater than 1.6% per
year, it is much less than 10.5% per year, and the total percentage drop that took place
is about one-half that which would be required to meet the 2008 Kyoto Protocol target from
In discussing these results, three examples are considered: light duty vehicles,
residential space heating, and industrial direct heating. Those are large energy demand
categories in three different sectors. The assessment will consider changes in energy
efficiency needed to meet the targets as well as consequences of reducing energy demand
without any efficiency gains. Finally, a review of historical energy demand for each of
the three end-uses will be made to see whether there are precedents for reductions called
for to meet the Kyoto Protocol targets.
Light Duty Vehicles.The end-use with the
largest energy demand is light duty vehicles. Meeting the Kyoto target would require a
decline in energy use by these vehicles of 4.9 Quads from the level now forecast by the
EIA for 2008 and about 1.7 Quads from the amount used in 1996. This would amount to a drop
of about 880,000 barrels a day of gasoline consumption. To achieve that decline by 2008
would require an increase of the average fuel efficiency of the entire light duty vehicle
fleet from 20.3 miles per gallon (mpg) to 23.3 mpg assuming no increase in
vehicle miles traveled (VMT) between 1996 and 2008. If the EIA forecast of VMT for 2008 is
met, fuel efficiency would have to reach 29.6 mpg, an annual rate of increase of 3.2%.
Currently, the EIA forecasts that the light duty vehicle fleet will operate at 20.3 mpg in
While historical fuel efficiency data for light duty vehicles as a group do not exist,
data for passenger cars, which constitutes about 60% of all light duty vehicles, show that
fuel efficiency grew from 14.6 mpg in 1979 to 21.2 mpg in 1991. The annual rate of
increase for that 12 period was 3.2%.(34)
For all motor vehicles, however, the increase was considerably less over that period, 2.6%
per year. Because the latter includes freight trucks, the actual increase for light duty
vehicles was somewhere in between these two rates. Since 1989, however, fuel efficiency
for light duty vehicles has increased from 18.5 mpg to 20.2 mpg. This slow growth in fuel
efficiency plus the forecast that such growth will continue to be slow for the next two
decades at least, indicates that most of the effects of the fuel economy standards
established in the late 1970s might already have been achieved. Therefore, to achieve the
growth in fuel economy required to meet the Kyoto target, would require a return to rapid
growth in new vehicle fuel economy.
It should be noted, however, that since the mid-1980s, there has been little increase
in the Corporate Average Fuel Economy (CAFE) standards for new automobiles or for light
trucks such as sport utility vehicles, vans, and pickup trucks. The CAFE standards for new
automobiles has remained at 27.5 mpg since 1985, and the standard for new light trucks has
only increased from 20.5 mpg in 1987 to 20.7 mpg in 1996.(35) Furthermore, over the same period the automobile portion
of the light duty vehicle fleet has been decreasing. In 1979, over 79% of all vehicle
miles traveled by light duty vehicles were accounted for by automobiles. In 1996, that
percentage had dropped to just over 64%.(36)
The remainder of those miles were accounted for by light trucks, which are less fuel
efficient than automobiles. In 1996, the value for all automobiles was 21.5 mpg and for
all light trucks was 17.3 mpg.(37)
If no efficiency gains were to take place, a substantial decline in VMT would be
necessary. At a constant 20.2 mpg, vehicle miles traveled would have to decline about
12.4% from the 1996 actual level and about 28.8% from the level forecast for 2008. In
terms of automobiles which make up about 72% of the VMT, this reduction would force annual
miles traveled per automobile to drop to about 10,300 compared to the 1996 level of about
11,700 miles.(38) If miles traveled
per automobile per year increased at the same rate as VMT according to the EIA forecast,
the value in 2008 is now projected to be about 14,500 miles. Therefore, the VMT level
required to meet the Kyoto Protocol target without any efficiency gains would be
Residential Space Heating.To meet the
Kyoto targets for residential space heating energy demand, a decline would be required of
1.8 Quads, or 27.6%, from the 1996 level and about 2 Quads, or 28.7%, from the level now
forecast by EIA for 2008. The small difference between the 1996 and 2008 reductions is a
result of the continued efficiencies that the EIA forecasts will be forthcoming for
residential space heating.
To achieve those reductions would require large increases in the efficiencies of
residential building shell and/or heating systems, a substantial reduction in the
temperature levels maintained in a typical house, or some combination of both. For
example, if a typical residential building had an average building thermal barrier rated
at R = 16, it would have to increase to R = 22 to reduce its space heating energy
requirements for heating by 27.6%. It could also install new heating equipment that
operated at a higher efficiency. To achieve the necessary fuel reduction, an efficiency
gain of over 37% would be needed. For example, a furnace operating at 60% efficiency would
need to be replaced by one operating at about 82%. Finally, if neither of these two
efficiency improvements were possible, that residential building would need to decrease
its average indoor temperature. For example, if the building's normal indoor temperature
was 75 degrees, and the outdoor temperature averaged 40 degrees, the indoor temperature
would have to be lowered to about 65 degrees to reduce its heating requirements in line
with the Kyoto target.
A change in residential space heating energy demand of this magnitude has occurred
before, from 1978 to 1980. The large increase in energy prices in the late 1970s drove the
demand for space heating energy down by 22%, after correcting for changes in heat load,
over that two-year period.(39) After
that decline, however, residential space heating energy demand has stayed nearly constant,
increasing by less than 10% between 1980 and 1996 after correcting for changing heating
requirements. That behavior is due primarily to increasing building and heating system
efficiencies, which have kept pace with housing stock growth and declining real energy
DOE has assumed that those efficiency increases will continue in order to compensate
for future housing stock growth. DOE forecasts that space heating demand per household
will decline by 25% between 1996 and 2020.(40)
Therefore, the reduction in space heating energy demand per household needed to meet the
Kyoto Protocol levels, estimated at about 27% as described above, would have to come on
top of the gains already forecast by DOE.
Industrial Direct Heat. This end-use is
used for a variety of manufacturing processes. It is used principally for primary metal
production such as in steel mill blast furnaces, to drive chemical reactions in chemical
plants and petroleum refining, for glass and clay product manufacturing, and in food
processing. To meet the Kyoto Protocol levels, industrial direct heat use would have to
decline by about 1.5 Quads, or 18.7%, from the 1996 level, and 2.6 Quads, or 28.7%, from
the forecast 2008 level. The EIA expects industrial direct heat energy demand to grow
about 14.7% from now to 2008, because of continued growth in the manufacturing sector. By
2010, the EIA forecasts that manufacturing output will grow by about 42.5%.(41) The difference between the two
growth rates is, in part, a result of expected increases in industrial process efficiency
for those processes involving direct heat. In addition, manufacturing output in industries
that use little or no direct heat are expected to grow faster than those that use a great
To meet the Kyoto Protocol levels for industrial direct heat use, manufacturers would
have to increase the efficiency of process heat equipment, increase process productivity
for those goods requiring direct heat, reduce output, or undertake some combination of the
three steps. An average heater efficiency increase of 23% would be required to reduce 1996
direct heat energy demand to the Kyoto target level. To maintain that level to 2008, given
the current EIA forecast, would require a total heater efficiency increase of over 56% if
that was the only action taken. That increase would have to come on top of efficiency
increases already forecast by EIA.
To see the effect of meeting the Kyoto Protocol by reducing production, consider
petroleum refining. In 1996, the United States produced about 17 million barrels per day
(MMBD) of refined products. If the U.S. petroleum refining industry had to reduce its
direct heat energy use to contribute its share in meeting the Kyoto protocol level,
production would have to decline by about 18% or about 3 MMBD. In 2008, the EIA forecasts
that U.S. refineries will produce about 19.5 MMBD.(42)
In order to meet the Kyoto protocol by reducing production, U.S. refinery output would
have to drop by about 5.5 MMBD to a total of 14 MMBD. Similar analysis could be carried
out for other manufacturing areas requiring direct heat in their manufacturing process.
No historical data exist specifically on direct heat energy demand so it is not
possible to compare, directly, the requirements of the Kyoto Protocol, as discussed in the
preceding paragraph, with past trends. It is possible, however, to determine industrial
energy intensity by calculating the ratio of industrial energy demand to industrial output
given in dollars.(43) From 1977 to
1989, industrial energy intensity declined by over 30%. The decline in industrial direct
heat energy intensity needed to meet the Kyoto Protocol from the 1996 level is about 18%.
If one assumes that total industrial energy intensity and direct energy intensity track,
then it would seem that meeting the Kyoto protocol requirements is reasonable based on
historical precedent. It should be noted, however, that total industrial energy intensity
stopped its decline in 1989 and has risen slightly -- about 5% -- between 1989 and 1996.
Furthermore, the 18% decline required by the Kyoto Protocol would have to come on top of
efficiency gains already forecast by EIA. It projects a decline of 17% in industrial
energy intensity between 1996 and 2010.(44)
Again, assuming direct heat energy intensity and total energy intensity for industry
growth (or decline) at the same rate,(45)
the total decline in energy intensity required to meet the Kyoto requirements while
allowing industrial output to grow to levels now forecast by EIA to 2008 would be over
40%. This is well beyond any historical changes.
As for reducing output, the question is whether consumers can accommodate less
production, not whether industry can produce less because of lower energy demand. In the
case of oil refinery production, from 1978 to 1983, oil products supplied in the United
States declined by 3.6 MMBD. This exceeds the decline from 1996 supply needed for oil
refineries to meet the Kyoto Protocol if they were to choose the lower production path.
Accommodation to that production decline between 1978 and 1983 was a complicated
combination of fuel switching, reduced economic output and increased energy efficiency.
Since 1983, however, the volume of oil products supplied has increased steadily to a
current level near the 1978 peak. Furthermore, the EIA forecasts continued increases as
noted above. The reduction that would be needed for oil refinery output to meet the Kyoto
protocol from the 2008 forecast level -- 5.5 MMBD -- is considerably greater than past
Concluding Comments. The analysis presented here
provides a detailed view of how energy is used in the United States. It also provides a
clear picture of the contribution these end-uses make to the buildup of carbon dioxide in
the earth's atmosphere. Finally, it presents a way to analyze the contributions each
end-use would make to any strategy to reduce CO2 emissions, and the
implications of those strategies in terms of particular end-uses.
Obviously, end-use disaggregation could continue beyond that given in this report. For
example different types of light vehicles and different types of industrial direct heat
processes exist. In particular, it was seen above that the contribution of light trucks --
sport utility vehicles, vans, etc. -- to the light utility vehicle fleet is growing. Data
to carry out finer breakdowns, however, exist in only a few cases, and further
disaggregation would be less and less precise. While current and historical data for
different components of the light duty vehicle fleet exist, projections of that mix are
lacking. Furthermore, it is not clear that, with the possible exception of light duty
vehicles and the commercial and residential miscellaneous categories, more disaggregation
would add much to understanding how energy is used or to the analysis of the implications
of reduced energy demand.
It is clear from the examples given above that the reduction in energy demand needed to
reach the Kyoto Protocol targets under the assumptions made in this report would be
substantially greater than previous changes in U.S. energy demand. While reductions
required from 1996 levels, for the three examples considered above, appear to be
comparable to those taking place in the past, when growth in energy demand that is
expected between now and 2008 or 2012 is factored in, the changes required are nearly all
unprecedented. Similar observations would hold for other end-uses because of the
underlying sector growth now forecast.
Achieving the energy demand reductions by increases in efficiency to maintain the
growth in products and services supplied by each end-use appears to require substantial
gains in equipment efficiency. While for the three cases examined the gains do not appear
to be impossible, they are likely to be difficult to achieve in the 12-year period and
might be increasingly costly.(46)
Strategies to achieve the Kyoto Protocol levels would probably not involve efficiency
gains alone, but rather would also include fuel switching and product or service
substitution. The former involves substitution of energy sources that do not have any net
CO2 emissions, such as renewables or nuclear-generated electricity, for fossil
fuels. The latter involves using services or products that result in lower carbon
emissions than those currently used; for example, using less energy intensive materials or
modes of transportation. While such substitutions are possible, they would likely take
several years to implement on a scale that would contribute significantly to carbon
Nevertheless, substitution, particularly zero-emission energy sources, appear to be an
important consideration along with increased energy efficiency in any long-term strategy
to reduce carbon emissions. While it is beyond the scope of this report to consider such
substitutions in detail, an example is given here to show how that might work.(47) If one-half of the coal-fired
capacity projected for the nation's electricity supply for 2008 could somehow be replaced
by nuclear power and/or renewables, carbon emissions in 2008 would decline by about 15%
from the current forecast. That change would lower by about two-thirds the energy
reduction requirements that would be needed to meet the Kyoto Protocol levels.
Furthermore, all end-uses, even those that used negligible amounts of electricity, would
benefit if the burden of emission reduction were apportioned to all end-uses as is done in
this report. The long lead time needed to build new power plants combined with material
and personnel constraints, along with other environmental and regulatory issues, however,
would likely preclude a substitution of that magnitude within 10 years. Over a longer
period, such substitution is probably more feasible.
Carbon emission reduction to meet the Kyoto Protocol levels for the 2008-2012 period by
reducing energy demand for current end-uses appears to be a substantial undertaking as
seen in the above analysis. If apportioned to all end-uses, each would be affected
significantly by 2008 given the reduction requirements and the currently forecast growth
in that end-use. If all or a major portion of the energy demand reduction were a result of
a lower level of service from that end-use rather than greater energy efficiency,
consumers of those end-uses would likely be substantially affected.
Appendix: Detailed Data Tables
The following are the detailed data tables showing for each end-use the actual and
forecast energy demand and carbon emissions for 1996, 2008, and 2012, the carbon emission
levels and resultant energy demand levels needed to reach the Kyoto protocol levels of a
7% reduction from the 1990 levels, and the resultant changes from 1996, 2008, and 2012.
1. (back) Congressional Research
Service, Global Climate Change, by Wayne Morrissey and
John Justus, CRS Issue Brief 89005 (updated regularly).
2. (back) One ton-carbon equivalent
is equal to 3.67 tons of carbon dioxide.
3. (back) Energy Information
Administration, Department of Energy, Annual Energy Outlook, 1998: With Projections
Through 2020, DOE/EIA-0383(98) (December 1997), 100, 124.
4. (back) Congressional Research
Service, Global Climate Change Treaty: Summary of the Kyoto
Protocol, by Susan Fletcher, CRS Report 98-2, 22 December 1997.
5. (back) Energy Information
Administration, Department of Energy, Impacts of the Kyoto Protocol on U.S. Energy
Markets and Economic Activity, SR/OIAF/98-03, (October 1998).
6. (back) Office of Energy
Efficiency and Renewable Energy, Department of Energy, Scenarios of U.S. Carbon
Reductions: Potential Impacts of Energy Technologies by 2010 and Beyond, prepared by
the Interlaboratory Working Group on Energy-Efficient and Low-Carbon Technologies.
7. (back) Congressional Research
Service, Energy Demand and Carbon Dioxide Production, by Richard Rowberg, CRS
Report 90-204, 11 February 1991.
8. (back) A small fraction of the
carbon -- about 20% -- of these nonenergy uses, according to the EIA, does end up as CO2
in the atmosphere. The effect of this contribution will be discussed but will not be
included in the detailed results because it is very small. See: Energy Information
Administration, Department of Energy, Emissions of Greenhouse Gases in the United
States, 1996, DOE/EIA-0573(96), (October 1997), 70.
9. (back) Energy Information
Administration, Annual Energy Outlook 1998, 106-109.
10. (back) Ibid. 102, 111.
11. (back) Energy Information
Administration, Department of Energy, Manufacturing Consumption of Energy, 1994, DOE/EIA-0512(94),
(December 1997), 114.
12. (back) U.S. Census Bureau,
Department of Commerce, 1992 Census of Mineral Industries: Fuels and Electric Energy
Consumed, MIC92-S-2. See: http://www.census.gov/mcd/minecen/download/nc92feec.txt.
13. (back) Energy Information
Administration, Annual Energy Outlook 1998, 101.
14. (back) For this report, the
EIA reference case forecast is used. That forecast is based on a macroeconomic model that
calculates a series of indicators that are used to drive the energy demand model. Among
the indicators for the reference case are an annual growth of real GNP of 3.0% per year
between 1996 and 2020, and a decline in total energy intensity (1000 Btu/1992 dollar of
GDP) of 0.9% per year over the same period. See, Energy Information Administration, Annual
Energy Outlook, 1998, 125.
15. (back) Energy Information
Administration, Annual Energy Outlook, 1998, 101, 106-111.
16. (back) In all cases the EIA
reference case forecast is used.
17. (back) About two-thirds of the
energy used to produce electricity is lost as waste heat. Thus one Quad of electric energy
delivered to consumers represents about three Quads of primary energy used. Some of the
waste heat is used for low-level heat -- e.g., space heat -- in industry and commercial
buildings, and, as such, replaces other energy sources. This secondary use of "waste
heat" is particularly prevalent with electricity generated on-site by industry. This
process is called cogeneration.
18. (back) Energy Information
Administration, Annual Energy Outlook, 1998, 111. EIA forecasts only a small
increase in light duty vehicle energy efficiency over that period: 20.2 mpg in 1996 to
20.3 mpg in 2010.
19. (back) Energy Information
Administration, Emissions of Greenhouse Gases, 100.
20. (back) These coefficients are
presented in terms of the amount of carbon produced. To calculate the amount of CO2
produced it is necessary to multiply the coefficient by 3.67, the ratio of the molecular
weight of CO2 to that of carbon.
21. (back) Energy Information
Administration, Emissions of Greenhouse Gases, 100.
22. (back) Energy Information
Administration, Annual Energy Outlook, 1998, 124.
23. (back) During tree growth,
carbon is sequestered as a result of the photosynthesis process whereby plants consume CO2
and give off oxygen. Because this growth takes place within years of the time when the
combustion of the wood takes place, there is no net production of CO2 over the
time frame of concern for possible greenhouse gas induced climate change.
24. (back) Congressional Research
Service, Global Climate Change Treaty: Summary of the Kyoto
Protocol; and Congressional Research Service, Global
25. (back) Energy Information
Administration, Impacts of the Kyoto Protocol on U.S. energy Markets and Economic
26. (back) Energy Information
Administration, Emissions of Greenhouse Gases, x, 18.
27. (back) For an extensive
discussion of the uncertainties, see, Congressional Research Service, Global Climate Change: Reducing Greenhouse Gases -- How Much from
What Baseline? by Larry Parker and John Blodgett, CRS Report 98-235 ENR, 11 March
28. (back) Congressional Research
Service, Global Climate Change.
29. (back) Energy Information
Administration, Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic
30. (back) Ibid.
31. (back) The EIA forecast is
used instead of the aggregate from all of the end-uses calculated by CRS because the
former accounts for the nonenergy contribution as described above, and such contributions
are included in the target level. Therefore, both parts of the ratio will be comparable.
32. (back) This is just the ratio
of total carbon emissions for that end-use to the total energy demand for that end-use
from all energy sources.
33. (back) In addition, there
would be about 7.5 Quads of fossil fuel use for nonenergy purposes and about 3.6 Quads of
biomass making a total of about 81.3 Quads.
34. (back) Energy Information
Administration, Department of Energy, Annual Energy Review 1997, DOE/EIA-0384(97)
(July 1997), 53.
35. (back) National Highway
Transportation Safety Administration, Department of Transportation, Automobile Fuel
Economy Program: Twenty-second Annual Report to Congress, Calender Year 1997, 3. See;
36. (back) Oak Ridge National
Laboratory, Department of Energy, Transportation Energy Data Book: Edition 18,
ORNL-6941 (1998), 5-5. The report can also be found on http://www-cta.ornl.gov/data/tedb18/Index.html.
37. (back) Ibid., 2-16.
38. (back) Ibid., 5-6, 5-8.
39. (back) Heat load is measured
in terms of heating degree-days. Energy Information Administration, Annual Energy
Review 1997, 55, 23.
40. (back) Energy Information
Administration, Annual Energy Outlook, 1998, 41.
41. (back) Ibid., 125.
42. (back) Ibid., 116.
43. (back) Council of Economic
Advisors, Office of the President, Economic Report of the President, (February
44. (back) Energy Information
Administration, Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic
45. (back) To the degree that
future gains in industrial output are disproportionately accounted for by less
energy-intensive industry, this assumption becomes less valid. Nevertheless, the decline
in direct heat energy intensity to meet the Kyoto Protocol is still likely to be
substantial as long as total industrial output is not to suffer.
46. (back) For another view on
this issue see Office of Energy Efficiency and Renewable Energy, Department of Energy, Scenarios
of U.S. Carbon Reductions: Potential Impacts of Energy Technologies by 2010 and Beyond.
47. (back) The model built to
perform the analysis presented above can also be used to calculate the effects of fuel