In preparing last year's edition of Emissions of Greenhouse Gases in the United States, the Energy Information Administration (EIA) developed new emissions coefficients for estimating carbon released from the combustion of fossil fuels in the United States. The EIA developed annualized emissions coefficients for coal, by rank, using 5,426 samples in EIA's coal analysis file. Coefficients for pipeline-quality and flared natural gas were calculated, based on 6,743 gas samples drawn from the U.S. gas distribution system. More than 20 petroleum products were assigned emissions coefficients based on their density, heat content, and carbon share, as revealed by product samples or underlying chemical composition. For a more detailed discussion of the methods employed in the development of these coefficients, please refer to Appendix A of last year's report.(Note 1)
Emissions coefficients specific for the United States were developed as an alternative to the more general coefficients recommended by the Intergovernmental Panel on Climate Change (IPCC).(Note 2) The IPCC coefficients are intended to be suitable for all nations, and to be used in conjunction with consumption data for various end-use products, as defined by the International Energy Agency (IEA). Using U.S. specific emissions coefficients allows EIA to capture the specific characteristics of fuel consumed in the United States, while taking advantage of the more detailed breakdown of product consumption in U.S. energy statistics.
With the exception of revised coefficients for jet fuel and liquefied petroleum gas (LPG), the emissions coefficients developed for last year's report were adopted for this year's edition. The emissions coefficient for LPG was also annualized, as were the coefficients for motor gasoline and crude oil. The composition of all petroleum products varies over time, as a result of economic changes (e.g., increases in the price of oil), regulatory changes (e.g., the Clean Air Act), or changes in refining technology. By annualizing or, in the absence of yearly data, periodically updating emissions coefficients, this variation is reflected, and the precision of carbon emissions estimates is increased. The new annual coefficients are listed in Table A1.
This Appendix describes the derivation of the revised and annualized emissions coefficients applied in this year's report. It also offers a brief discussion of the difficulties of defining a "typical" still gas and, thus, assigning an accurate emissions coefficient.
Motor gasoline consumption is the largest single source of anthropogenic greenhouse gas emissions in the United States, releasing nearly 270 million metric tons of carbon in 1993. As with all petroleum products, the principal determinants of an emissions coefficient for motor gasoline are its density and carbon content. Density of motor gasoline varies systematically between summer and winter grades of gasoline, and from low octane to high octane. Over the past decade, the density of all octane grades, across all seasons, has been increasing slowly. Table A2 shows that density has increased by as much as 3.5 degrees API (depending on season and octane) over the past 10 years, raising emissions coefficients over that period from 19.38 million metric tons per quadrillion Btu to 19.43 million metric tons per quadrillion Btu.
The trend toward increased density in motor gasoline can be traced, in part, to the gradual elimination of leaded gasoline. In 1984, more than 40 percent of all gasoline supplied was leaded; by 1991, the last year for which data on leaded gasoline production were collected, that share had fallen to 2 percent. As Table A2 shows, leaded gasoline was the least dense of all gasoline grades.
Table A1. Carbon Emissions Coefficients at Full Combustion, 1984-1994
(Million Metric Tons of Carbon per Quadrillion Btu)
*Composite coefficient based on final calculations, accounting for fraction not combusted and deductions for nonfuel use, using unpublished disaggregated energy data.
Fuel 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994
Petroleum Motor Gasoline 19.37 19.37 19.38 19.38 19.39 19.41 19.41 19.41 19.42 19.43 19.43 LPG 16.97 16.98 17.03 17.05 17.04 17.07 17.00 16.99 17.00 16.98 17.02 Jet Fuel 19.44 19.43 19.42 19.42 19.42 19.41 19.40 19.40 19.39 19.37 19.34 Distillate Fuel 19.95 19.95 19.95 19.95 19.95 19.95 19.95 19.95 19.95 19.95 19.95 Residual Fuel 21.49 21.49 21.49 21.49 21.49 21.49 21.49 21.49 21.49 21.49 21.49 Asphalt and Road Oil 20.62 20.62 20.62 20.62 20.62 20.62 20.62 20.62 20.62 20.62 20.62 Lubricants 20.24 20.24 20.24 20.24 20.24 20.24 20.24 20.24 20.24 20.24 20.24 Petrochemical Feed 19.37 19.37 19.37 19.37 19.37 19.37 19.37 19.37 19.37 19.37 19.37 Aviation Gas 18.87 18.87 18.87 18.87 18.87 18.87 18.87 18.87 18.87 18.87 18.87 Kerosene 19.72 19.72 19.72 19.72 19.72 19.72 19.72 19.72 19.72 19.72 19.72 Petroleum Coke 27.85 27.85 27.85 27.85 27.85 27.85 27.85 27.85 27.85 27.85 27.85 Special Naphtha 19.86 19.86 19.86 19.86 19.86 19.86 19.86 19.86 19.86 19.86 19.86 Waxes and Miscellaneous 19.81 19.81 19.81 19.81 19.81 19.81 19.81 19.81 19.81 19.81 19.81 Other* 20.00 20.19 20.03 20.06 20.04 19.93 19.99 20.16 20.19 20.24 20.26 Coal Residential/Commercial 25.92 25.89 25.88 25.90 25.87 25.94 25.92 26.00 26.13 25.97 25.97 Industrial Coking 25.43 25.43 25.41 25.39 25.40 25.40 25.51 25.51 25.51 25.51 25.51 Other Industrial 25.51 25.53 25.55 25.53 25.53 25.56 25.58 25.60 25.62 25.61 25.61 Electric Utility 25.63 25.64 25.63 25.64 25.67 25.67 25.68 25.69 25.69 25.71 25.71 Flare Gas 14.92 14.92 14.92 14.92 14.92 14.92 14.92 14.92 14.92 14.92 14.92 Natural Gas 14.47 14.47 14.47 14.47 14.47 14.47 14.47 14.47 14.47 14.47 14.47 Crude Oil 20.03 20.11 20.12 20.13 20.16 20.13 20.16 20.18 20.22 20.23 20.21
Source: Energy Information Administration, Emissions of Greenhouse Gases in the United States 1987-1992, DOE/EIA-0573 (Washington, DC, November 1994), pp. 78-92, and estimates presented in this appendix.
Table A2. Trends in Motor Gasoline Density, All Grades, 1984-1994
(Degrees API)
NA = not available.
Fuel Grade 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994
Winter Grade+ Leaded 63.4 63.3 62.9 62.6 62.2 62.0 62.0 * * * NA Low Octane 61.8 62.3 62.5 62.4 62.2 62.0 62.0 61.8 61.4 61.0 NA Mid Octane * * * * * 59.9 60.8 60.4 60.2 59.9 NA High Octane 60.1 59.7 59.7 60.3 59.6 58.8 59.0 59.3 59.0 58.7 NA Summer Grade Leaded 61.2 60.9 60.5 60.2 59.5 58.4 58.5 * * * NA Low Octane 59.4 59.3 59.5 59.2 58.9 58.2 58.2 58.0 57.4 56.1 NA Mid Octane * * * * * 57.1 57.4 57.1 56.4 55.5 NA High Octane 57.9 57.5 57.0 57.1 56.8 55.3 55.5 55.7 55.6 54.4 NA Average Emissions Coefficient (Million Metric Tons Carbon per Quadrillion Btu) 19.37 19.37 19.38 19.38 19.39 19.41 19.41 19.41 19.42 19.43 NA
*Data prior to 1989 do not include mid-grade octane. Data after 1990 do not include leaded gasoline.
+Winter grade densities are the average of winter months, including the end of one calendar year and the beginning of the next calendar year.
Source: National Institute of Petroleum and Energy Research, Motor Gasoline, Summer, and Motor Gasoline, Winter (1984-1994).
In order to create unleaded gasoline with the same octane rating and "anti-knock" quality as leaded gasoline, refiners must synthesize additional volumes of light aromatic hydrocarbons and blend them into paraffinic straight-run naphtha. Aromatic hydrocarbons have a much higher ratio of carbon to hydrogen than do other typical hydrocarbons. Because carbon is more dense than hydrogen, aromatics drive up the density of unleaded gasoline. Further, the increased ratio of carbon to hydrogen in aromatics also drives up the carbon share. The result is higher emissions coefficients. Unleaded premium gasoline achieves its higher octane rating with an even larger share of aromatics, and hence has higher density and carbon content than regular unleaded gasoline.
After density, the percentage of carbon in motor gasoline is the most important determinant of its carbon emissions coefficient. While the density of commercially supplied gasoline may vary by as much as 10 degrees API, the share of carbon in gasoline is unlikely to vary by more than plus or minus 2 percent. Carbon share is bounded by the ratios of carbon to hydrogen in the hydrocarbon compounds that compose gasoline. Thus, the uncertainty associated with estimating carbon content is limited.
The carbon contents used to develop emissions coefficients for this report are based on a study conducted by Mark DeLuchi.(Note 3) Using ultimate analyses of motor gasoline samples taken during 1981, DeLuchi estimated carbon contents of 87.1 percent for summer gasoline and 86.2 percent for winter gasoline. He also provided an estimate of 86.6 percent for the average carbon content for all grades of gasoline, based on samples taken during 1981, 1985, 1987, 1989, and 1990. The 86.6 percent average was adopted for this analysis, because the multi-year sample is more likely to incorporate the changing composition of gasoline.
To estimate annualized emissions coefficients for motor gasoline, data on density by season and octane grade were obtained from the National Institute of Petroleum and Energy Research (NIPER).(Note 4) These data were used in conjunction with the carbon share of 86.6 percent to develop an emissions coefficient for each grade of octane in each season. The coefficients were then weighted by the share of annual consumption each grade represents, to derive an annual emissions coefficient. Although gasoline grades are typically divided into only two seasons, winter and summer, each annual emissions coefficient represents a combination of three seasonal sets of coefficients: a winter season (January, February, and March), a summer season (April through September), and a second winter season (October, November, and December). The first and second set of winter coefficients used for each year may differ slightly due to variations in the density of motor gasoline over time, as reported to NIPER. For this report, it was assumed that "motor gas blending components" are identical to motor gasoline in their average composition and hence have the same emissions coefficients.
Text Box: Oxygenated and Reformulated Gasoline
EIA energy statistics distinguish between two types of jet fuel: "naphtha-based" and "kerosene-based." Naphtha-based jet fuels, used primarily by the military, represented about 10 percent of jet fuel consumption in 1993 but less than 3 percent in 1994. This rapid decline is believed to be the result of the Defense Department's conversion from naphtha-based JP-4 jet fuel to kerosene-based JP-8 jet fuel. Because of the diminishing importance of naphtha-based jet fuel, no additional work on its emissions coefficient was undertaken, and the factor developed for last year's report (19.95 million metric tons per quadrillion Btu) was retained.
Kerosene-based jet fuels include civil-grade Jet A, consumed by commercial jet airliners. Combustion of kerosene-based jet fuel is an important contributor to U.S. greenhouse gas emissions, accounting for emissions of 58 million metric tons of carbon in the United States during 1993. The emissions coefficient for kerosene-based jet fuel was revised from last year's report after additional data on the density and carbon share of commercially sold jet fuels were obtained. For last year's report, the density of Jet A was estimated at 42 degrees API and the carbon share at 86.3 percent. The density estimate was based on 38 fuel samples examined by NIPER.(Note 5) Carbon share was estimated on the basis of a hydrogen content of 13.6 percent found in fuel samples taken in 1959 and reported by Martel and Angello,(Note 6) and an assumed sulfur content of 0.1 percent.
Boeing, the leading commercial airline supplier in the United States, conducts its own tests of jet fuel properties, to ensure the safe and efficient operation of the airplanes it sells. Using data from 39 samples of Jet A analyzed by Boeing, the EIA now estimates the average density of Jet A at 44.5 degrees API and the average carbon share at 85.8 percent.(Note 7) Although the NIPER density data and the Boeing data are from contemporaneous samples, it is likely that the Boeing data are more representative of fuel being consumed by the commercial fleet. Because the work done by Martel and Angello used samples dating from 1959 to 1972 and the Boeing work used samples from 1989, the carbon share estimated from the Boeing data is presumed to be more reliable. A lower density is consistent with a slightly reduced carbon share, further supporting the use of the Boeing data for both density and carbon share.
The carbon emissions coefficient resulting from these density and carbon estimates may appear surprisingly low at 19.33 million metric tons per quadrillion Btu, down from 19.71 a year ago. This is a result of the continued use of EIA's heat content estimate of 5.67 million Btu per barrel for kerosene-based jet fuel. Because the estimated density of the fuel has decreased, the heat content (Btu per pound) has been artificially raised, reducing the overall emissions coefficient. If the heat content is estimated directly from product samples, the coefficient rises to 19.56 million metric tons per quadrillion Btu. In order to maintain consistency with EIA energy data, however, EIA's standard heat content was used.
For convenience of use, the coefficients for naphtha-based jet fuel and kerosene-based jet fuel were combined into a single coefficient. An annualized jet fuel coefficient was derived by weighting the two coefficients by consumption share. Annual coefficients range from a high of 19.42 million metric tons per quadrillion Btu in 1984 to a low of 19.34 in 1994 (Table A1).
Crude oil is rarely consumed directly in the United States, representing less than 0.1 percent of U.S. oil consumption. This report estimates emissions based on end-use consumption, limiting the importance of a crude oil coefficient. However, if emissions are estimated using a mass balance approach, the emissions coefficient for crude oil accounts for most of the carbon emissions from petroleum.
For last year's report, the EIA developed a regression equation for predicting the share of carbon in crude oil from its density and sulfur content. Density and sulfur content are used as measures of commercial value; thus, data for these independent variables are readily available. Using ultimate analyses of 182 crude oil samples, the following equation was developed:
Previously, a density of 31 degrees API and a sulfur content of 1 percent were adopted as representative of an "average" barrel of crude oil. However, the EIA collects and publishes annually a weighted average density and sulfur content for crude oil input at refineries.(Note 8) These data were assembled for the period 1980 through 1994 and entered into the regression equation outlined above. The result was a series of annual carbon shares for crude oil. When used in conjunction with the density data and EIA's standard heat content for crude oil of 5.8 million Btu per barrel, they produce annualized emissions coefficients that rise steadily from 19.94 million metric tons in 1980 to 20.23 in 1993, before dropping back slightly to 20.21 in 1994 (Table A1). The increase can be attributed directly to the increase in density of crude oil entering refineries. Crude oil density moved from 33.76 degrees API to 31.30 degrees API between 1980 and 1993.
LPG consists of light hydrocarbons extracted from natural gas and sold to end users. The EIA collects data on four categories of paraffinic hydrocarbons under the heading LPG: ethane, propane, isobutane, and n-butane. Each category represents a discrete, pure paraffinic compound.(Note 9) In last year's report, pentanes-plus, a fuel category including hydrocarbons with more than five carbon atoms, was erroneously included in the calculation of an emissions coefficient for LPG. Pentanes-plus are more accurately described as natural gas liquids (NGL). Pentanes-plus have a higher carbon content and, consequently, have a higher emissions coefficient than LPG. Thus, their removal from the overall weighted emissions coefficient for LPG lowers the factor by about 1 percent.
Because the densities and carbon contents of pure paraffinic hydrocarbons are well known, carbon emissions coefficients for the four LPG categories are easily derived. These coefficients were then weighted by yearly consumption to yield a series of annual overall LPG emissions coefficients. Propane accounts for the bulk of LPG consumption, between 53 and 63 percent of consumption in each of the past 10 years, and thus has the largest effect on the emissions coefficient. As shown in Table A3, the emissions coefficient for LPG has remained nearly stable over the past decade, varying by less than 1 percent from its low to its high.
Still gas, sometimes called refinery gas, is a byproduct of petroleum distillation and refining. It results from catalytic and thermal cracking of various crude fractions and has a highly variable composition. Some constituents of still gas, such as propane and methane, have economic value. Others, such as hydrogen sulfide, are considered waste. The gas itself is often sold or reused as a petrochemical feedstock, or purified and sold as pipeline-quality natural gas. On average, approximately 160 cubic feet of still gas are collected for every barrel of light crude oil processed.(Note 10)
Because different refinery operations result in different gaseous byproducts, it is difficult to determine what constitutes "typical" still gas. In last year's report, the emissions coefficient for still gas was based on a "typical" composition found in the Gas Engineers Hand book.(Note 11) For this report, the EIA obtained and examined three additional samples of still gas. Composition and carbon content varied significantly among the four samples, as do the resulting carbon emissions coefficients.
The size of an emissions coefficient for still gas is primarily a function of the ratio of heavier hydrocarbons, such as ethane and propane, to the share of unassociated hydrogen molecules within the sample. The share of unassociated hydrogen found in the four samples examined ranged from a low of 13 percent in sample 1 to a high of 72 percent in sample 3, while collectively ethane and propane had a combined share of 29 percent of sample 1 and only 14 percent of sample 3 (Table A4). The resulting emissions coefficients range from 10.23 million metric tons carbon per quadrillion Btu to a much higher 17.51 million metric tons carbon per quadrillion Btu.
In the absence of any concrete information on the composition of still gas, we are forced into the realm of supposition. Still gas samples with low emissions factors owe this characteristic to a very high proportion of free hydrogen in their composition. Gas streams with a large hydrogen content are likely to be used as refinery or chemical feedstocks, rather than being burned for energy. Therefore, "typical" still gas burned at refineries is likely to be composed largely of less valuable feedstocks, such as methane and carbon monoxide. Consequently, the EIA has decided to continue to use a carbon emissions coefficient of 17.51 million metric tons per quadrillion Btu for still gas.
Table A3. Consumption Shares for Liquid Petroleum Gases by Energy Content, 1984-1994
(Percent)
NA = not available.
Fuel (Emissions Coefficient) 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994
Ethane (16.25) 30.9 29.6 26.6 24.4 25.5 22.4 26.8 27.6 26.7 28.1 NA Propane (17.20) 57.3 59.6 58.6 60.8 59.2 62.5 63.2 62.5 63.1 62.5 NA Isobutane (17.75) 0.6 1.8 3.1 3.6 3.5 1.7 1.4 2.4 2.4 3.3 NA n-Butane (17.72) 11.2 9.0 11.7 11.3 11.8 13.5 8.6 7.5 7.8 6.0 NA Overall Emissions Coefficient (Million Metric Tons Carbon per Quadrillion Btu) 16.97 16.98 17.03 17.05 17.04 17.07 17.00 16.99 17.00 16.98 NA
Sources: Characteristics of Compounds: V.B. Guthrie (ed.), Petroleum Products Handbook (New York, NY: McGraw Hill, 1960), p. 3-3. Shares of U.S. Consumption: Energy Information Administration, Petroleum Supply Annual, DOE/EIA-0340 (Washington, DC, various years).
Table A4. Composition, Energy Content, and Emissions Coefficient for Four Samples of Still Gas
Sources: Sample one from J.F. Bell and H.R. Linden, in American Gas Association, Gas Engineers Handbook (New York, NY: Industrial Press, 1974), p. 3/71. Samples two, three, and four from C.R. Guerra, K. Kelton, and D.C. Nielsen, "Natural Gas Supplementation with Refinery Gases and Hydrogen," in Institute of Gas Technology, New Fuels and Advances in Combustion Technologies (Chicago, IL, June 1979).
Emissions Coefficient Composition (Percent) Energy Content (Million Metric Tons Carbon Sample Hydrogen Methane Ethane Propane (Btu per Cubic Foot) per Quadrillion Btu)
One 12.7 28.1 17.1 11.9 1,388 17.51 Two 34.7 20.5 20.5 6.7 1,143 14.33 Three 72.0 12.8 10.3 3.8 672 10.23 Four 17.0 31.0 16.2 2.4 1,100 15.99