3. Methane Emissions
3. Methane EmissionsEstimated U.S. anthropogenic methane emissions totaled approximately 27.2 million metric tons in 1992, virtually unchanged from 1990 and 1991 (Table 15). Increasing emissions from the agricultural sector have been offset by a decrease in emissions from coal mines and landfills. Emissions increased by an estimated 1.3 million metric tons between 1985 and 1990, largely because energy production and consumption increased during that period. About one- third of all anthropogenic methane emissions in the United States can be traced to energy production and use. Landfills are the largest single source of methane emissions in the United States, with more than 37 percent of all emissions in 1992. Methane emissions from landfills rose steadily between 1985 and 1990, because the amount of waste in place in landfills continued to rise. Between 1990 and 1992, however, increases in methane recovery for energy use and flaring outstripped the growth in gross emissions, leading to a net decline in overall methane emissions from landfills. The balance of U.S. emissions in 1992 was from industrial processes and agricultural sources, principally, the management of domesticated livestock. Emissions from enteric fermentation in domesticated animals and the solid waste of domesticated animals declined for much of the 1980s, as decreasing animal populations more than offset increases in average animal size. This trend was reversed between 1989 and 1992, with increasing animal populations and accelerating growth in average animal size leading to rising emissions from this source.
Methane emissions estimates are considerably more uncertain than carbon dioxide emissions estimates, largely because of limitations on the availability of data on which to base emissions estimates. Most carbon dioxide emissions are the result of the combustion of fossil fuels . Because fossil fuels are of economic value, the quantity consumed is carefully measured. For each fuel burned, a carbon content can be estimated, with an uncertainty band below 5 percent. For every unit consumed a known amount of carbon is released. In contrast, methane emissions are an unintended and often unrecognized side-effect of energy production and use, municipal waste disposal, and animal husbandry. Methane emissions are rarely metered, and when metering does occur, as in the case of ventilation emissions from coal mines and methane recovered for energy at landfills, data may be available for only a single or few years and must be extrapolated over time.
Emissions Trends. Methane emissions from coal mining are estimated at 4.2 million metric tons for 1992, down from just under 4.5 million metric tons in 1990 and slightly above the 4.1 million metric tons emitted in 1988 (Table 16). Emissions from coal mining represent about 16 percent of all methane emissions and 56 percent of emissions from energy production. These estimates have a large uncertainty bound. Of the four sources of coal mining emissions-underground mining ventilation systems, underground mining degasification systems, surface mining, and post-mining activities- actual data exist only for emissions from underground mining ventilation systems for some mines and only for selected years. Data have been compiled on ventilation emissions from approximately 200 of the gassiest underground coal mines in the United States, for the years 1985, 1988, and 1990. Emissions estimates for those years are likely to be most accurate, and they drive the estimates for other years. The growth of emissions from coal mining through 1990 and their subsequent decline after 1990 parallel changes in coal production over the same period (Table 16).
Methane Formation and Release. As a coalbed matures and coal is formed from organic matter, several gases, including carbon monoxide and methane, are produced. (29) The quantity of methane produced is largely a function of pressure and temperature. (30) Methane content in coalbeds is also loosely correlated with the depth of the coal seam. Of the methane produced during the formation of coal, up to 12 percent may be stored in the pores of the coal itself. (31) The remainder either migrates to the atmosphere through cracks in the coal seam or is stored in the surrounding strata of the coal seam. If the strata overlying the coal seam erode or are stripped away during coal mining, pressure in the coal seam drops, allowing methane to migrate to the surface. Coal mined from the surface is likely to release far less methane than coal mined underground, because it is subjected to lower pressures and methane in the seam will have had earlier opportunities to migrate to the surface through cracks and fissures. All methane remaining in the coal pores after mining will be emitted when the coal is transported and pulverized for combustion.
Estimating Emissions from Coal Production and Transportation . Emissions from underground coal mines accounted for 89 percent of all emissions from coal mines in 1992 (Table 16). Because methane in underground coal mines represents a potentially fatal hazard to coal miners, methane emissions from this source are better understood than those from surface mining or post-mining activities. Methane in atmospheric concentrations above 5 percent is explosive; consequently, the Mine Safety and Health Administration (MSHA) requires that mines be ventilated and sets standards for methane levels in underground coal mines. The MSHA conducts quarterly inspections of coal mines, measuring methane concentrations in the mines and methane emissions at fan exhaust points. The Department of Interior's Bureau of Mines (BOM) has collected these reports and developed a database for selected years (most recently, 1985, 1988, and 1990) of ventilation emissions for all mines emitting more than 100,000 cubic feet per day.
To estimate underground coal mining emissions, U.S. underground coal mines and coal mine production were divided into two groups:
Emissions factors for gassy mines in each coal basin were computed for 1985, 1988, and 1990 by dividing reported emissions from each basin by production from the gassy mines in each basin, as reported annually to the Energy Information Administration on Form EIA7A, "Coal Production Report." Emissions factors for intervening years were calculated by interpolating between the 1985, 1988, and 1990 factors. The 1990 emissions factor was assumed to remain unchanged for 1991 and 1992.
The BOM states that emissions from all other mines with measurable emissions amount to only 2 percent of total emissions. (32) Therefore, 1988 emissions from nongassy mines were estimated at 2 percent of emissions from gassy mines for each coal basin. Basin-level emissions factors for nongassy mines were calculated by dividing this figure by 1988 coal production from non- gassy mines. (33) This calculated set of emissions factors were then multiplied by production from nongassy mines for other years to estimate annual emissions from this source.
In mines where methane production is particularly high, the use of ventilation fans to remove sufficient quantities of methane to assure miner safety may be prohibitively expensive or require such high volumes of air movement as to make mining impractical. Instead, methane concentrations may be reduced through pre- mining degasification. Degasification may also occur where gas is present in sufficient volumes to make commercial sale feasible. To estimate emissions from degasification systems, 1988 emissions from those mines operating degasification systems, as identified by the EPA, were scaled to production from those mines. The resulting emissions factors were multiplied by annual production from mines with degasification systems. (34) Total emissions from degasification systems were estimated at 1.31 million metric tons in 1992, up from 1.25 million metric tons in 1988 (Table 16).
Gross emissions from underground mining were then reduced by coal mine methane captured for commercial recovery. According to the EPA, about 250,000 metric tons of methane was recovered in 1988 and sold to pipeline companies. (35)
Emissions from surface mines are much more difficult to measure. Because methane in surface mines does not represent a health hazard to miners, and because there is no practical means to capture it, these emissions are not directly measured. Some efforts to measure methane emissions at surface mines using Fourier Transform Infrared (FTIR) (36) spectroscopy have been undertaken, but the results are thus far limited and not yet generalizable. (37) However, it is known that emissions from surface mines should be far below those from underground mines because of the much lower in-situ methane contents of the coal in surface mines and opportunities for methane to migrate to the atmosphere in advance of mining. The Intergovernmental Panel on Climate Change (IPCC) recommends the use of an emissions factor for surface mines of 0.3 to 2.0 cubic meters per metric ton of coal mined. (38) That emissions factor was adopted for this report, resulting in a central emissions estimate of 0.42 million metric tons of methane from surface mines in 1992 (Table 16).
Post-mining emissions, like those from surface mines, are not measured. Because coal is pulverized before it is combusted, all methane remaining in coal pores that has not been released prior to or during coal mining will be emitted. The IPCC recommends emissions ranges for post-mining activities of 0.0 to 0.2 cubic meters per metric ton of coal mined for coal from surface mines and 0.9 to 4.0 cubic meters per metric ton of coal mined for coal from underground mines. (39) Those factors were applied here, leading to emissions estimates from post-mining activities of 0.04 million metric tons and 0.6 million metric tons in 1992, for surface-mined coal and underground-mined coal, respectively (Table 16).
The methane emissions estimates presented in Table 16 are 10 percent lower than those presented in last year's report because of changes in the estimation methods used. (40) Formerly, emissions from underground coal mines, surface coal mines, and post-mining activities were estimated as multiples of in-situ methane content. At an experts' workshop on coal mine methane emissions hosted by the EIA in April 1994, it was concluded that the use of ventilation data scaled to coal production is a more reliable estimation method for underground coal mine emissions than estimates based on in-situ methane contents.
In 1992, U.S. anthropogenic methane emissions from the oil and gas system amounted to 3.3 million metric tons, an increase of more than 15 percent since 1985, largely attributable to increased natural gas consumption. Emissions from the oil and gas system were 40 percent of all emissions from energy sources and 12 percent of total methane emissions (Tables 15 and 17).
Oil and natural gas are found trapped beneath impermeable cap rock in underground reservoirs. Natural gas may be found in a free gaseous state or as gas dissolved in oil or water. In general, natural gas marketed in the United States is between 90 and 95 percent methane. Most of the methane found in natural gas is produced through the thermal or microbial degradation of organic matter in sediments. More rarely, methane is produced through the thermal cracking of oil and asphalt . (41)
Oil and Gas Production and Processing . Most oil reservoirs also contain natural gas, often dissolved in the oil. Thus, when crude oil is extracted, associated natural gas is produced. In many nations, this associated gas would be vented or flared. In the United States, however, the gas is usually sold commercially. In 1992, 5.97 trillion cubic feet of associated natural gas was withdrawn from oil wells. The volume of gas withdrawn directly from gas wells was much larger, 16.16 trillion cubic feet in 1992. (42) Of the more than 22 trillion cubic feet of gas withdrawn from all sources in 1992, some was used for repressuring, some was vented and flared, and the remainder, 18.7 trillion cubic feet, was marketed.
After natural gas is extracted from either oil wells or gas wells, it is sent via gathering pipelines to a gas processing facility. As natural gas is moved from the wellhead through gathering pipelines, there is some leakage from valves, meters, and flanges. These fugitive emissions result in the release of methane to the atmosphere. In addition to fugitive emissions, there are maintenance- related emissions when pressure valves and gathering pipelines are emptied. At the processing facility, heavy hydrocarbons and contaminants are removed from the gas. There are both fugitive and maintenance-related emissions from the gas processing plant and equipment. There may also be system upsets or accidents, during which sudden increases in pressure require the release of gas as a safety measure, or a portion of the system ruptures, releasing large volumes of gas. Currently, the U.S. natural gas system is operating below capacity, reducing the frequency of such events.
The IPCC recommends the use of source-specific emissions factors based on examinations of model facilities or engineering studies to estimate methane emissions from oil and gas systems. (43) This report uses emissions factors modified from those developed by the EPA. (44) Emissions factors for oil and gas wells are based on studies of four model facilities. For gathering pipelines, emissions factors were developed from two model transmission line systems (discussed below). The emissions factors for gas processing plants developed by EPA are based on analyses of three model plants. Using EPA's emissions factors, the EIA developed aggregate emissions factors for natural gas wellheads, oil wells, gathering pipelines, and gas processing plants. Rather than scaling emissions from gas processing plants to gross gas withdrawals, they were scaled to gas throughput. An emissions factor for heaters, separators, and dehydrators was calculated by dividing the total emissions estimate for 1990 provided by the EPA for this source by gross gas withdrawals. (45)
Total methane emissions from oil and gas wells, gathering pipelines, and gas processing were estimated at 1.18 million metric tons for 1992, up slightly from 1.16 million metric tons in 1990 (Table 17). While the number of oil wells producing associated gas (46) and the miles of gathering pipeline have declined since 1990, (47) the decline has been more than offset by an increase in natural gas wellheads and throughput at gas processing facilities. (48)
Gas Venting. When the flow of associated gas from an oil well is too small or too intermittent to be of value, the gas does not contain sufficient heat content to be marketed, or there are no gas gathering and processing facilities attached to the site, the gas is vented or flared. The gas that is flared does not result in methane emissions, because the methane it contains is combusted and converted to carbon dioxide (estimates of carbon emitted from this source appear in Chapter 2). The methane contained in gas that is vented is released directly to the atmosphere. Although statistics are available for total gas vented or flared, there are no existing data on the share of each. The Department of Energy estimates that of gas vented or flared, 80 percent is flared and 20 percent is vented. (49) There is a great deal of uncertainty in this estimate. More than two-thirds of U.S. oil wells are "stripper wells " that produce less than 10 barrels of oil per day but collectively account for about 14 percent of U.S. oil production. Associated gas production at these wells is usually at low pressures and volumes, with little or no commercial value. This associated gas production is unlikely to be reported by well owners, and its magnitude is impossible to judge. Therefore, emissions from "stripper wells" are excluded from the emissions estimates in this report.
Assuming that 20 percent of associated gas reported as vented or flared is vented, methane emissions from this source are estimated at 640,000 metric tons for 1992, an increase of 60,000 metric tons from 1990 (Table 17). Gas vented and flared in Wyoming accounts for more than half of all gas vented or flared in the United States. Between 1990 and 1992, gas vented or flared in Wyoming grew from 63 billion cubic feet to nearly 90 billion cubic feet. (50) In the absence of this growth in Wyoming, the amount of gas reported vented or flared in 1992 would have fallen from the 1992 level, and consequently, estimated methane emissions would have also fallen.
Natural Gas Transmission and Distribution. Natural gas is transported from production fields and gas processing centers to local distribution companies and large-scale industrial users through high-pressure transmission pipelines. When gas reaches the local distribution network, its pressure is lowered in gate stations and it is sent out to customers via low-pressure pipeline. Methane may be emitted from the transmission system through a variety of sources. For example, there may be leaks from corroded pipeline or inadequately sealed valves, or emissions may result from compressor exhausts, pneumatic devices, and routine maintenance. The majority of emissions from gate stations and distribution pipeline are fugitive emissions from equipment leaks.
Emissions factors for transmission pipeline, gate stations, and distribution pipelines were also derived from the factors developed by EPA, (51) based on four model transmission systems. Emissions factors for gate stations were estimated from a study of 28 gate stations in 11 towns. Emissions factors for distribution pipeline were based on studies conducted by Pacific Gas and Electric (PG&E) and Southern California Gas Company on their distribution networks. The pipeline systems operated by PG&E and Southern California Gas may be newer and better maintained than the average natural gas system, thus biasing emissions estimates downward.
Methane emissions from the transmission system were estimated at 1.06 million metric tons for 1992, up from 1.04 million metric tons in 1990. Emissions from the remainder of the distribution system remained stable between 1990 and 1992 at approximately 330,000 metric tons (Table 18).
The increase in emissions from transmissions pipeline can be traced to a small increase in miles of pipeline. For this report, the emissions factors for gas transmission were scaled to pipeline mileage rather than broken down into individual system components. Pipeline mileage is the best nationally available time series proxy for the capital equipment of pipeline operators, but it is an imperfect tool for scaling emissions. Eberle has argued that as natural gas demand increases, producers and suppliers will upgrade the existing capital stock. (52) In particular, new pipeline is almost exclusively plastic pipeline, which has a much smaller emissions factor per mile than does nonplastic pipeline. Therefore, emissions may not grow proportionately with additional pipeline miles.
Oil Refining and Transportation. Methane is emitted during the transport of crude oil to the refinery and from refinery operations themselves. Emissions from the transportation phase occur when vapor is displaced during the loading and unloading of crude oil. Most crude oil produced in the United States, however, is transported via pipeline, which emits only minor amounts of fugitive methane. Accordingly, emissions from the transport of crude oil are small.
Methane is emitted during several phases of the refining process. When crude oil arrives at the refinery, it is loaded into storage tanks. Vapor displacement from the tanks results in methane emissions. During the refining process, methane is separated from the crude oil through vacuum or atmospheric distillation. Methane emissions occur during this stage from leaking equipment components. Methane that is not destroyed during refinery flaring operations is also released to the atmosphere. Altogether, these sources accounted for emissions of 81,000 metric tons of methane in 1992 (Table 19).
Estimates of methane emissions from oil refining and transport were derived from the Radian Corporation report, Global Emissions of Methane From Petroleum Sources. (53) Radian used two factors for crude oil transport emissions-one for tank trucks and rail cars and one for marine vessels. At present, information for determining the amount of crude that is transported by truck or rail to refineries indicates that the number is small. (54) Accordingly, this report uses an emissions factor for transport on marine vessels. Radian's coefficient of 2.55 x 106 tons of methane per barrel loaded was applied to the sum of annual crude oil imports, exports, and Alaskan production of crude. These categories should most accurately reflect the amount of crude subject to emissions from marine transport. It should be noted that Radian's methane emissions coefficient is based on volatile organic compound (VOC) emissions factors published in EPA's Compilation of Air Emission Factors (AP-42) and assumes that 15 percent of the VOC vapor is methane. (55) Radian's factor also assumes equal loading of ocean- going ships and in river barges. Estimation of methane emissions from oil refineries also relies on factors published by Radian.
Carbon dioxide is not the only major greenhouse gas released from combustion processes. Methane and nitrous oxide , as well as other radiatively important gases, such as carbon monoxide, nonmethane volatile organic compounds , and nitrogen oxides (see Chapter 6), are also emitted from both stationary and mobile combustion of fossil and renewable fuels. The most significant source of combustion-related methane emissions is woodburning in residential woodstoves. Together, stationary and mobile combustion account for a small share (2.8 percent) of all U.S. anthropogenic methane emissions (Table 15).
Estimates of methane emissions from stationary sources were calculated using EPA's Compilation of Air Pollutant Emissions Factors and the IPCC's Draft Guidelines for National Greenhouse Gas Inventories, which contain coefficients for each fuel type by end-use sector. (56) As Table 20 indicates, residential woodstoves account for the overwhelming majority of stationary source methane emissions, because of their particularly inefficient combustion process. However, there is a high degree of uncertainty associated with emissions from residential wood combustion. This report shows a much higher level of methane emissions from residential woodstoves than was shown in last year's report, because the EPA has revised its emissions factor for this source from 1 pound of methane per ton of wood burned to 26 pounds per ton. The EPA maintained a roughly constant emissions factor for all volatile gases (methane plus nonmethane volatile organic compounds ) but revised the methane share of emissions from about 5 percent of total gases to 50 percent. Using energy consumption data reported in EIA's State Energy Data Report 1992, Monthly Energy Review, and Estimates of U.S. Biomass Energy Consumption 1992, stationary source methane emissions are estimated at 505,000 metric tons for 1992.
If transportation fuels were completely combusted, the only products emitted would be carbon dioxide and water. In fact, not all of the fuel is fully burned and thus other gases, including methane , are created and released. Mobile source methane emissions are affected by a number of factors, including the amount of unburned hydrocarbons passing through the engine, the engine type and maintenance conditions, and any post-combustion control of hydrocarbon emissions, such as the use of catalytic converters. Emissions of methane are highest when the air-fuel mixture is "rich," that is, the amount of oxygen present is insufficient for complete combustion. This condition occurs especially in low speed and idle engine situations.
To develop estimates of mobile source methane emissions, this report uses emissions coefficients published by the IPCC, expressed in grams of methane per kilometer traveled. To use these coefficients, information is required on the types of fuels consumed in the transportation sector, the combustion technologies used, and the extent to which emission control measures are employed. For motor vehicles, it is necessary to know how many miles are traveled by the various vehicle types and models. The U.S. Department of Transportation's Federal Highway Administration (FHWA) reports annual estimates of vehicle miles traveled (VMT) by type of vehicle (cars, trucks, buses, motorcycles). (57) The EIA's Residential Transportation Energy Consumption Survey (RTECS) also reports VMT and energy consumption by households for personal transportation. (58) Surveys were conducted in 1983, 1985, 1988, and 1991.
For this report, a custom database sort from the RTECS was used to calculate VMT for household-sector passenger cars and trucks by model year for the years in which surveys were conducted. Data from the RTECS survey were used to compute a weighted average coefficient for each survey year, which was then applied to non-household-sector passenger cars and light trucks (business-owned vehicles, fleets, rental cars, etc.). Emissions for nonsurvey years were estimated by interpolating between the weighted average estimates for survey years. Emissions for 1992 were extrapolated using 1991 RTECS data and 1993 fleet age data as reported by the American Automobile Manufacturers Association. (59) As Table 21 illustrates, emissions from transportation have been declining steadily despite the annual increase in VMT, because newer model cars with more effective emissions control technologies account for an increasing share of the VMT each year.
Methane emissions from other forms of transportation were estimated by applying an IPCC emissions factor to the amount of energy consumed by each mode of transportation. Fuel consumption by domestic trade ships, locomotives, farm equipment, and construction and industrial equipment were taken from EIA's Fuel Oil and Kerosene Sales . Consumption of aviation gas and jet fuel are reported in EIA's State Energy Data Report and Petroleum Supply Annual. Emissions from recreational boats were estimated on the basis of fuel consumption estimates in the Transportation Energy Data Book. (60)
Landfills are the single largest source of anthropogenic methane emissions in the United States. Methane emissions from landfills remained nearly stable at just over 10 million metric tons between 1987 and 1992, declining slightly between 1990 and 1992 as a result of increased methane recovery for energy use (Table 22). Increasing amounts of landfilled waste in place have resulted in an increase in estimated gross emissions, from approximately 11 million metric tons in 1987 to nearly 12 million metric tons in 1992. The increase is offset, however, by significant increases in methane capture for energy use or for flaring as a safety measure. In 1992, it is estimated that as much as 1.7 million metric tons of methane was either flared or recovered for energy use. (61)
Methane is produced when organic material in landfilled waste decomposes under anaerobic conditions. When waste is generated, aerobic bacteria begin to decompose its organic material. These bacteria consume oxygen while converting organic substances to carbon dioxide, heat, and water. After the waste is landfilled, the process continues until the oxygen available for aerobic decomposition is depleted. This oxygen-free condition may occur rapidly if the waste is tightly compacted and covered with soil or other waste, or it may occur slowly if the waste is left exposed or lightly compacted at the landfill surface. The viability of methanogenic bacteria is dependent on temperature, acidity, and moisture content. In the absence of oxygen, anaerobic bacteria, including methanogens , begin digesting the organic material in the waste. Because landfilled waste is buried beneath soil covers, temperatures within the landfill are largely independent of ambient air temperatures and typically fall in a range conducive to methane generation. (62)
To prevent groundwater contamination and to comply with environmental regulations, sanitary landfills are designed to minimize the entry of moisture. Landfills are lined with an impermeable barrier such as clay or plastic, and after each day of operation freshly deposited waste is covered with 6 inches of dirt. (63) Thus, moisture content in landfills is largely a function of the moisture content of the waste, which varies both horizontally and vertically in a landfill. Accordingly, methane emissions rates are heterogeneous not only across different landfills but within landfills. Further, because moisture content in landfills is inhibited, decomposition rates of organic material are not maximized. This is demonstrated by the presence of newspapers and foodstuffs in landfills that have yet to degrade after 20 years or more. (64) These factors make the estimation of emissions from landfills highly uncertain.
Methods Used. Thorneloe et al. have demonstrated a relationship between landfilled waste in place and methane emissions. (65) Total amounts of waste in place were estimated using data from Franklin Associates and Biocycle magazine. (66) The data from Franklin Associates provide annual estimates of waste landfilled from 1960 to 1992. Biocycleprovides estimates of waste landfilled from 1988 to 1993. The estimates from Franklin are based on a model that estimates trash outputs based on production inputs. The Biocycle data are based on a survey of State agencies. At a methane emissions workshop hosted by the EIA in April 1994, experts suggested that the Franklin waste estimates may be as much as 33 percent too low. The Biocycle data are more likely to be representative of actual amounts of waste landfilled, but the method employed required the use of more historic data. Accordingly, an average ratio of Biocycle to Franklin estimates was calculated for the period 1988 to 1992 and applied to all years of Franklin statistics to develop an estimate of waste landfilled for the period 1960-1992.
To estimate emissions from landfills, municipal solid waste in the United States was divided into two groups:
Emissions from landfills with gas recovery systems were estimated at 2.08 million metric tons for 1992. (67)
A model based on the EMCON Methane Generation Model was used to estimate emissions from these landfills for years other than 1992. (68) The model divides landfilled waste into three categories: readily decomposable waste, moderately decomposable waste, and slowly decomposable waste. For each category, a time lag until methane generation begins is estimated, as well as a time constant during which the methane yield of the waste is realized. The methane yield represents the total amount of methane that a given amount of waste will produce over its lifetime. While the EMCON model represents emissions increasing linearly in the early years after waste has been landfilled, then decreasing exponentially over time, this report assumes that emissions occur at a constant rate over time. In the long run, this difference will have little impact on emissions estimates as a state of equilibrium between emissions from new waste and the expiration of emissions from older waste is reached.
The EMCON model was benchmarked to measured emissions for 1992. Methane yields were calculated that resulted in estimates equal to measured emissions for 1992. The resulting yields were 1.42 times larger than would be expected from the "default" time lags and methane yields from the EMCON model. It would not be surprising if landfills with methane recovery systems proved to be "gassier" than the general population of landfills, making recovery more economical. The share of waste in these landfills is assumed to have remained constant over time. This assumption is likely to bias older estimates somewhat, since it is known that the number of landfills is shrinking and that the disposal of waste is becoming more concentrated in fewer and fewer landfills.
To estimate methane emissions from the 87.5 percent of all waste for which no direct emissions measurements are available, the EMCON model was used again, but with the "default" time lags and methane yields. On this basis, methane emissions from all other landfills were estimated at 9.8 million metric tons in 1992.
The inconsistency of estimates of the amount of waste landfilled and the general lack of knowledge about the number of both open and closed landfills suggest significant uncertainties in estimates of waste in place. The estimates of waste in place used in this report may underrepresent actual waste in place, because they exclude waste landfilled prior to 1960. However, the excluded waste is likely to have only a negligible impact on methane emissions beyond 1980, because its organic material has already decomposed or been sequestered. Because of these uncertainties, the emissions estimates in this report are tied, to the extent possible, to the data on 105 landfills compiled by Thorneloe et al.
This report's estimate of emissions for 1990 is somewhat lower than the estimate contained in last year's inventory of greenhouse gas emissions. (69) The difference is due to a change in the estimation methods used. The former method, developed by Bingemer and Crutzen, (70) was based on the total amount of waste landfilled in a single year, the degradable organic content of the waste, and the fraction of waste actually converted to biogas . That method, which assumes that all methane is released as soon as the waste is landfilled, is likely to overestimate the share of organic material that is converted to methane. New methods were chosen because they include a component of measured emissions rather than theoretical estimates, and because they capture the time lag associated with emissions.
Total methane emissions from enteric fermentation in domesticated animals were 5.49 million metric tons in 1992, up from 5.25 million metric tons in 1990, an increase of almost 5 percent (Table 23). The increase is attributed to cattle populations that are growing in both total number and average size and productivity per head. Emissions from other animals have remained fairly stable since 1981.
Estimating Emissions. The breakdown of carbohydrates in the digestive tract of herbivores (including insects and humans) results in the production of methane. (71) The volume of methane produced by this process, known as "enteric fermentation," is greatest in ruminant animals, such as cattle, buffalo, sheep, goats, and camels. Ruminant animals possess a rumen or forestomach that allows them to digest plant material that monogastric (single stomach) species are unable to digest. (72) The rumen contains as many as 200 species of microorganisms. A small fraction (5 to 10 percent) (73) of these microorganisms are methanogenic bacteria. These methanogens are responsible for the removal of hydrogen from the rumen. Although nonruminant animals do produce methane from enteric fermentation in their large intestines, the production is small compared with that which occurs in the rumen of ruminant animals. The majority (some 90 percent) of the methane produced by ruminant animals is exhaled during respiration. The remainder is released during belching or as part of the flatus .
The digestion of food represents the transformation of food energy to heat to produce body work or tissue growth. The production of methane is a loss of energy. (74) The degree of this energy loss is a product of quantity and quality of feed intake, the growth rate of the animal, its productivity (reproduction and/or lactation), and its mobility. In order to calculate methane emissions from enteric fermentation in domesticated animals, the animals must be classified into distinct, relatively homogeneous groups. For a representative animal in each group, feed quality and quantity, as well as growth rate, productivity, and activity levels, are estimated. These variables are combined to derive a methane emissions factor for that animal. That factor is then applied to the total population of the animal group to calculate an overall emissions estimate.
Cattle. The feed energy intake and productivity of cattle have been extensively studied. Baldwin et al. developed a digestion model to predict the energy values provided by various types of feed. (75) The EPA combined this model with an animal growth model that predicts growth, pregnancy, and milk production. (76) The EPA defined 32 different diets for 5 different regions of the United States. Dairy cattle were divided into three population groups: replacement heifers 0-12 months old, replacement heifers 12-24 months old, and mature cows. Beef cattle were divided into six groups: replacements 0-12 months old, replacements 12-24 months old, mature cows, bulls, steers and heifers raised for slaughter under the weanling system, and steers and heifers raised for slaughter under the yearling system.
The EPA provides emissions factors for each class of cattle in each region and national-level emissions coefficients for each animal class. These national average emissions factors were adopted with modifications for this report. The emissions factors developed by EPA are representative of the cattle population in 1990. The average size of U.S. cattle has been increasing since 1981, and the milk production of dairy cows has increased rapidly over that time. To capture some of this variation, the EPA emissions factors were scaled to average slaughter weight as provided by the U.S. Department of Agriculture (USDA) National Agricultural Statistics Service, Livestock, Dairy and Poultry Branch, for the respective classes of cattle. The change in the average slaughter weight of calves was applied to replacements 0-12 months old. This is an imperfect adjustment factor for several reasons. The average slaughter weight may not reflect the average live weight of cattle classes, and it may not accurately characterize the milk production and/or the size of progeny. However, it does serve as an approximation of changes in the total energy intake of the cattle population. The result is emissions factors that increase steadily from 1980 to 1992, especially for younger cattle.
The size-adjusted emissions factors were applied to population data obtained from the USDA National Agricultural Statistics Service, Livestock, Dairy and Poultry Branch. Using this method, emissions from enteric fermentation in cattle were 5.19 million metric tons in 1992, up from 4.94 million metric tons in 1990 and slightly below 1981 emissions of 5.27 million metric tons (Table 23). Emissions declined between 1981 and 1989, because a shrinking cattle population (down about 10 percent in that period) overshadowed the effects of larger and more productive animals. Since 1989, cattle populations have been growing, and in combination with a more rapid increase in the size and productivity of the average animal, have resulted in growing emissions.
Other Animals. In addition to cattle, this report includes estimates of methane emissions from enteric fermentation in sheep, pigs, goats, and horses. While pigs and horses are not ruminants, their large populations and large sizes result in some methane emissions. The method for estimating emissions from these animals was drawn from the work of Crutzen et al. (77) This method estimates the gross energy intake of the animal and combines it with a methane yield to derive an emissions factor. An emissions factor of 13 kilograms methane per head per year was adopted for sheep, based on a methane yield of 6 percent and a gross energy intake of 35 megajoules per day for adult sheep and 22.5 megajoules per day for immature sheep. U.S. sheep population data were obtained from the USDA's National Agricultural and Statistics Service.
Methane emissions factors of 1.5 kilograms methane per head per year for pigs and 18 kilograms methane per head per year for horses were adopted from the work of Crutzen et al. (78) Crutzen provides an emissions factor only for goats from India. A goat emissions factor of 8 kilograms per head per year was adopted, based on an assumption that the relative energy intake of U.S. goats compared to Indian goats is the same as U.S. sheep compared to Indian sheep. (79) Pig population data were provided by the USDA. Horse and goat population data were obtained from the Census of Agriculture for 1982 and 1987. Data for 1983 to 1986 were estimated using straight- line interpolation, and data for 1988 to 1992 were calculated using straight-line extrapolation.
The estimates shown in Table 23 are 15 percent lower than those presented in last year's report. The change can be traced to the change of estimation methods used for calculating emissions from cattle. In the 1993 inventory, cattle were divided into three categories: milk cows, beef cattle, and cattle on range. Emissions factors for these categories were based on the work of Blaxter and Clapperton. (80) The IPCC has shown that these emissions factors may overestimate methane yield. (81) In addition, these factors did not take into account changes in the size and productivity of U.S. cattle.
Total methane emissions from the solid waste of domesticated livestock were estimated at 2.8 million metric tons for 1992, up from 2.6 million metric tons in 1990 (Table 24). The increase can be attributed to the increasing population and average size of U.S. cattle. While the average size of cattle has grown since 1980, cattle populations were declining until 1988, more than offsetting the increase in average size. Thus, current emissions are no higher than 1981 emissions.
Estimating Emissions. Methane is produced when methanogenic bacteria decompose the organic material in the solid waste of domesticated animals in an anaerobic environment. The amount of organic material that is susceptible to decomposition is described as the "volatile solids content." The volume of methane produced from a given amount of volatile solids under optimal anaerobic conditions is characterized as the maximum methane- producing capacity of the animal waste. Because conditions are rarely optimal, actual methane production usually is lower than the maximum possible.
The share of the maximum methane-producing potential of the waste that is realized is largely a function of the manner in which the waste is managed. Liquid-based waste management systems provide an anaerobic environment as well as the moisture required for methanogenic bacterial cell production and acidity stabilization. (82) In contrast, animal waste left to dry in the fields will decompose in the presence of oxygen, minimizing methane production. The share of the maximum methane produced using a given waste management technique is represented by its methane conversion factor.
Emissions estimates were based on an equation developed by Safley et al., which links emissions to the volume of solid waste produced by a given animal, the volatile solids in that waste, and the system in which the waste is handled. (83) The volume of waste produced by a given animal is a function of the animal's size, diet, and energy requirements. Emissions factors developed using this equation were then multiplied by population data for each animal group.
Methane emissions from solid waste were estimated for cattle, swine, poultry, sheep, goats, and horses. Cattle were divided into beef cattle on feed, other beef cattle, and dairy cattle. The categories of cattle were further subdivided by size and type. Swine were divided into market swine and breeding swine. Poultry were separated into broilers and layers.
Animal population data for cattle, poultry, swine, and sheep were obtained from the USDA's National Agricultural Statistics Service, Livestock, Dairy and Poultry Branch. Horse and goat population data were drawn from the U.S. Department of Commerce Agricultural Census for 1982 and 1987 and interpolated or extrapolated for other years. Cattle population estimates are an annual average based on the January 1 and July 1 inventories. Broiler chickens have a lifespan in the neighborhood of 7 weeks. To estimate an average broiler chicken population during a given year, the annual broiler chickens slaughtered estimate was multiplied by 0.1425, based on the recommendation of the USDA's Economic Research Service. (84)
Typical animal masses estimated by EPA were adopted with modifications. (85) The typical animal masses offered by EPA for U.S. cattle represent animal sizes in 1990. The size of U.S. cattle has been increasing over the past decade. For this report, typical animal masses for cattle were adjusted on the basis of changes in the slaughter weight of cattle from year to year. This is an imprecise scaling mechanism, but it does provide some sense of continuing growth in the typical size of animals in the U.S. cattle population. Typical animal mass for all other animals was assumed to be constant. The volatile solids produced per kilogram of animal weight, maximum methane-producing capacity of each animal's waste, and share of waste handled in each management system were also adopted from Safley et al. (86)
The methane conversion factors for each waste management system have been revised considerably since last year's greenhouse gas emissions report. Last year, the EIA adopted the methane conversion factors used in Safley's work. (87) Shortly before publication of the EIA report, Dr. Andrew Hashimoto of Oregon State University completed work on updated, substantially lower, methane conversion factors that varied by temperature. EIA published a set of alternative emissions estimates based on Dr. Hashimoto's methane conversion factors. Dr. Hashimoto's updated factors have since been reviewed and published by the IPCC. (88) Dr. Hashimoto provides coefficients for manure management systems in climates with average temperatures of 5 to 15 degrees Celsius, 15 to 25 degrees Celsius, and above 25 degrees Celsius. For this report, all cattle waste was assumed to be handled under ambient air temperatures between 15 and 25 degrees Celsius (59 to 77 degrees Celsius).
The effect of the change in methane conversion factors used for waste management systems can be seen by comparing the revised 1990 emissions estimate of 2.6 million metric tons of methane to the previous 1990 estimate of 3.5 million metric tons. In addition, the previous estimate showed cattle with nearly twice the emissions of swine, whereas the new method indicates that emissions from the solid waste of cattle and swine are almost equal.
On a global scale, rice paddy fields are one of the largest sources of anthropogenic methane emissions. However, the amount of land devoted to rice cultivation in the United States is relatively small and therefore represents only a minor source of methane emissions. There are three processes involved in the release of methane from rice paddies: methane production, oxidation, and transport. Methane is produced in flooded rice fields by the anaerobic decomposition of organic matter. Rice paddy soils, offering conditions of oxygen depletion, moisture, and high levels of organic matter, present ideal environments for the growth of methanogenic bacteria. A substantial amount of the methane that is produced, however, is oxidized by aerobic methanotrophic bacteria in the soil and floodwater of rice paddies. Experiments indicate that as much as 80 percent of the methane is internally oxidized before it can be released to the atmosphere. (89) The methane that is not oxidized is emitted to the atmosphere through gas bubbles, by diffusion through the floodwater, and most importantly, via plant-mediated transport.
A number of factors are believed to be significant in the processes of methane production and emission, including soil temperature, redox potential , and acidity; substrate and nutrient availability; the addition of chemical and/or organic fertilizers; the rate of methane oxidation; and the rice plant variety. Floodwater depth is also significant-dry, upland fields do not produce significant amounts of methane, nor do fields flooded to depths greater than 1 meter. Additionally, field studies have shown large seasonal variations in methane flux corresponding with the development of the rice plant.
Emissions of methane from rice paddies in the United States were estimated by applying a daily emissions rate range (0.165- 0.564 grams methane per square meter per day) to the daily cultivated rice area. The emissions range, based on field experiments in California, Louisiana, (91) and Texas, (92) was multiplied by the season length and the rice area harvested for each rice-growing State. According to the EPA, (93) the climate in southwest Louisiana and Texas allows these States to grow a second or "ratoon" rice crop. In calculating methane emissions, the estimated area harvested during the ratoon crop was added to the area for the primary crop, as reported by the USDA. (94) In 1992, there were approximately 1.4 million hectares of rice harvested, implying methane emissions of 0.44 million metric tons (Tables 15 and 25).
The production of certain petrochemicals requires the use of catalytic cracking. This process involves heating a feedstock to high temperatures (often by the addition of steam) in the presence of a catalyst. The molecular bonds of the feedstock are cracked and new bonds form. By controlling the temperature of the reaction, the feedstock, and the catalyst used, the production of specific petrochemicals is favored. However, a wide range of chemical reactions occur in the reaction vessel, leading to the generation of byproducts, including methane .
Coefficients have been developed to estimate methane emissions from the production of ethylene , ethylene dichloride , styrene , methanol , and carbon black . (95) These coefficients estimate the emissions of methane in grams per kilogram of product. Total U.S. production of the chemicals can be obtained from trade association statistics. (96) Applying the coefficients to the production figures results in national emissions estimates for each chemical (Table 26).
These emissions estimates are very uncertain. The emissions coefficients were developed for worldwide use, and it is possible that pollution control equipment and flares at U.S. plants may reduce net methane emissions. Further, materials balance equations for petrochemical processes suggest that the emissions coefficients used likely overestimate methane emissions. (97)
Ethylene and styrene are valuable building blocks in the manufacture of a wide variety of materials (in fact, ethylene is itself a building block for styrene). These chemicals are used to produce plastics, fibers, pharmaceuticals, synthetic rubber, and other materials. Consistent with their variety of uses, production of these chemicals has been growing. Ethylene dichloride is used primarily to produce vinyl chloride monomers, which when strung together become the widely used polyvinyl chloride (PVC). Use of this substance, and consequently production of ethylene dichloride, has also grown. Production of methanol and carbon black has remained relatively stable in recent years.
The manufacture of iron and steel includes three processes that give rise to methane emissions: the production of coke, sinter , and pig iron . Coke is produced by heating a carbon source (usually coal) in the absence of oxygen. Impurities in the coal are driven off, leaving nearly pure carbon coke. Coke ovens are required to have strict emissions controls, but seals in the coke oven itself may leak methane, which is one of the gases formed. Sinter contains iron ore, flux materials, and coke. The coke is ignited, and the heat generated causes the sinter to agglomerate into particles of a size appropriate to be used in the blast furnace for iron production. During this agglomeration stage, methane is produced as a byproduct gas. Iron (as ore, pellets, and sinter), coke, and flux materials are combined in the blast furnace, where hot gases are used to reduce them to iron, slag, and exhaust gases. The exhaust gases, which include carbon monoxide and methane, usually are collected to be used as fuel. However, some leakage occurs.
The IPCC has developed coefficients for methane emissions corresponding to each of the three iron and steel production inputs discussed above. (98) These coefficients were multiplied by the U.S. production totals for each respective input to obtain estimates of national emissions (Table 26).
Production of both coke and sinter has been declining in recent years, largely in response to environmental regulation and costs, with coke supply shifting from domestic production to imports. This production shift has led to a decrease in national emissions. Overall production of pig iron has remained relatively constant.
