This chapter describes some of the ways in which different types of land use affect emissions and sequestration of greenhouse gases and provides estimates of the scale of such emissions and sequestration. From a global warming perspective, the most important land use issues are those related to forest creation and destruction. Forests and forest soils remove and store large amounts of carbon from the atmosphere. In contrast, temperate-zone land-use-related emissions of methane and nitrous oxide are relatively small and contribute only minimally to global climate change. Because of these differences, more discussion in this chapter is devoted to forestland use than other types of land use.
Emissions of carbon dioxide from the combustion of fossil fuels are dwarfed by carbon dioxide emissions and absorption from natural processes. As noted in Chapter 1, natural processes in the oceans and biomass are responsible for most carbon dioxide absorption and emissions. This is also true of methane and nitrous oxide. Most methane and nitrous oxide are created by bacteria in soils and wetlands. Although the most important natural processes affecting greenhouse gas concentrations in the atmosphere are not subject to human control, modifications in land use can influence their concentrations in significant ways.
The magnitude of the influence is difficult to measure. Unlike service station pumps, trees and swamps do not come equipped with meters. Thus, analysts can only estimate emissions or sequestration on the basis of small sample surveys and extrapolate (with associated uncertainty) to much larger regions. A second, related problem is distinguishing between human-caused and natural phenomena. An electric power plant is clearly a human artifact. Trees growing back over abandoned farmland are a more ambiguous case.
To avoid confusion with other definitions in this report, land use is defined in this chapter to include only forestland, cropland, grassland, pasture and range, wetlands, and impervious surface areas such as urban areas, roads, and highways. Use and disposal of materials derived from these areas is not included, except in the case of wood products.
As noted in the Executive Summary of this report, of the trace gases, carbon dioxide is the largest single contributor to global warming, responsible for 85 percent of U.S. global warming potential (GWP)-weighted emissions of greenhouse gases. Thus, from a global warming perspective, the most important modifications in land use are those that significantly affect the carbon budget.
The most important changes in land use are those that increase or reduce forestland. U.S. forests removed a net 123 million metric tons of carbon in 1990, including the net 12 million metric tons sequestered in wood products and the net 15 million metric tons sequestered in landfilled wood product waste.(Note 1) This quantity would offset approximately 9 percent of the 1,430 million metric tons of carbon emitted by the United States in 1994 from the burning of fossil fuels (see Chapter 2).
In contrast, cropland, grassland, and pasture and range, once established, do not significantly affect greenhouse gas concentrations until they are changed to another land use. The carbon flux in cropland, grassland, and pasture and range is close to zero when considered over periods of a year or more at a time, not including the initial high carbon losses that are typical during the first few years after conversion. Carbon lost when plants are harvested or grazed or die back during the winter is normally recaptured the following year in new growth, resulting in no significant net change in the carbon budget.(Note 2) Such land does release nitrous oxide and methane to the atmosphere, but these gases are relatively unimportant contributors to U.S. GWP-weighted emissions of greenhouse gases, totaling only 13 percent in 1993 (the total share attributable to land uses is even less). Impervious surfaces, such as roads and large percentages of urban areas, once established, halt the process of carbon sequestration and emission entirely by preventing growth of vegetation and entombing soil carbon, thereby removing it from the cycle.
Of the land uses discussed in this chapter, changes in total forest area have the most important impact on U.S. anthropogenic contributions to greenhouse gas emissions. Forests sequester atmospheric carbon in biomass and soil. As noted above, on average all U.S. forests combined, and the wood products they produce, sequestered a net 123 million metric tons of carbon in 1994.
The mechanism driving forest carbon cycling is photosynthesis. Every green plant is, in effect, a solar-powered factory that extracts carbon dioxide from the atmosphere, separates the carbon atom from the oxygen atoms, returns oxygen to the atmosphere, and uses the carbon to make biomass in the form of roots, stems, and foliage. A fraction of dead vegetation accumulates as soil, and large quantities of roots annually die and slough off, sequestering additional carbon in the ground.
The opposite of photosynthesis is respiration, the release of carbon to the atmosphere as carbon is used for energy within plants. Trees add new cell layers each year. Old and new cells require energy for maintenance. Growth and increased maintenance cause respiration to increase. Eventually, because of limitations to total foliage area, the rate of photosynthesis cannot keep pace with respiration, and trees enter a stage of rough equilibrium between photosynthesis and respiration. Often they become net carbon emitters. Because trees must continually grow in diameter in order to survive, the imbalance leads to mortality, and the cycle shifts into reverse as a portion of the carbon is released through decay.
Accretion of carbon in living biomass is only one aspect of forest carbon sequestration. Leaf litter, fallen trees and branches, and other decaying biomass all add carbon to the soil as they decompose, even though most carbon is released to the atmosphere. More importantly, a large amount of carbon is added underground to the soil as roots die and slough off. The older a forest becomes without major perturbations (such as intense fires or conversion to cropland), the more carbon is stored in the soil. An estimate by Richard Birdsey of the U.S. Forest Service places the percentage of carbon in mature forests stored in the soil at 59 percent--only 31 percent is stored in live roots, stems, branches, and foliage. About 9 percent of all carbon is stored in litter, humus, and coarse woody debris on the forest floor, and about 1 percent is found in understory vegetation. The estimated amount of carbon stored in the soil is 42 metric tons per acre, with a total of 30 billion metric tons stored in the soils of all U.S. forests.(Note 3) In con trast, agricultural fields can contain 10 or fewer metric tons of carbon per acre.(Note 4)
Even without the gradual sequestration of carbon in the soil, the current inventory of biomass in the form of forests and natural vegetation contains enormous amounts of carbon. Birdsey's study estimated that U.S. forest ecosystems contained 52.5 billion metric tons of carbon in 1987--the equivalent of nearly 40 years of U.S. carbon emissions from fossil fuel consumption.
In the tropics, the continuing destruction of forests is releasing large volumes of carbon dioxide into the atmosphere. In the United States, however, the process is stable or moving slightly in reverse. Many forests in North America, especially in the East, were cleared for farming during the 19th century. As eastern farms have been abandoned for more productive lands in the Midwest, much of the land has reverted to forests that will continue to grow for decades to come. In addition, more than 4 million acres of marginal cropland have been reforested since 1974 under such Federal programs as the Conservation Reserve Program, Agricultural Conservation Program, and Forestry Incentives Program.(Note 5)
As noted above, U.S. forests and wood products absorbed an estimated net 123 million metric tons of carbon in 1990, equivalent to about 9 percent of energy-related carbon emissions in 1994. An estimate made in a report by the U.S. Environmental Protection Agency (EPA), placed net absorption from forested areas in the coterminous United States at 114 million metric tons for the year 1990.(Note 6) That report also estimated total carbon storage in U.S. forestland at 38.5 billion metric tons, 14 billion metric tons less than Birdsey's estimate, which illustrates some of the uncertainties associated with calculations of this sort. Birdsey estimated that U.S. forests absorbed 460 million metric tons of carbon between 1982 and 1987--equivalent to an annual growth rate in sequestered carbon of 0.8 percent.
While trees on existing acreage continue to absorb carbon, the acreage planted in forests can also expand or contract. Farm and pastureland can be converted to forestland, and forestland can be cleared and converted to other uses. Clearing forestland sets the stage for large-scale losses of carbon to the atmosphere, while allowing crop or pastureland to grow trees sets the stage for large-scale absorption of carbon.
Other types of land use conversion produce more ambiguous results. In general, converting pastureland or grassland into cropland typically produces emissions of carbon through the destruction of biomass and loss of soil carbon through tillage and crop harvest. Typical estimates of the amount of soil carbon lost are approximately 30 percent of the amount in place at the time of conversion.(Note 7) These losses can be expected to take place over a period of 20 years, or longer, following conversion. Similarly, abandoning cropland or converting it to forest normally leads to net carbon sequestration through the creation and long-term growth of biomass and corresponding additions to soil carbon.
Table 39 shows U.S. Department of Agriculture (USDA) estimates of the major uses of land in the United States, developed through a periodic sample survey. The survey is conducted every 5 years, most recently for the year 1992. The USDA has not yet published all the 1992 survey results: in particular, the publication that gives the comprehensive view of land use, Major Uses of Land in the United States, has not yet been published. However, draft estimates for the basic information contained in Table 39 were obtained, and published material indicates that between 1987 and 1992 there was a net increase of 6 million acres of forestland.(Note 8)
Table 39. Major Uses of Land in the United States
(Million Acres)
Sources: A. Daugherty, Major Uses of Land in the United States: 1987, Economic Research Service Report 643 (Washington, DC: U.S. Department of Agriculture, January 1991), p. 4; and A. Daugherty, Major Uses of Land in the United States: 1992, unpublished review draft (1995).
Land Use 1978 1982 1987 1992
Cropland 471 469 464 461 Used for Crops 369 383 331 338 Idle Cropland 26 21 68 56 Pasture 76 65 65 67 Grassland Pasture and Range 587 597 591 591 Forest-Use Land 737 721 731 737 Grazed Land 172 158 155 145 Special Use 34 66 83 89 Other Use 531 497 493 503 Special Use Areas 124 204 196 281 Miscellaneous Other Land 345 274 283 283 Total Land Area 2,264 2,265 2,265 2,263
Between 1982 and 1992 there was a large net reduction in the amount of active cropland. The amount of land used for crops declined by 45 million acres, while the amount of idle cropland increased by 35 million acres. The total amount of non-Federal rural cropland (excluding Alaska) declined by 24 million acres.(Note 9) While the total amount of cropland declined, the amount of idle cropland increased by 1 million acres. There were much smaller declines in private pastureland and rangeland (1.6 million acres and 3.7 million acres, respectively). The shift from cultivated to idle cropland--or shifts from cropland to range or pasture--should, in principle, lead to small increases in net carbon storage. Shifts from any of the above to urban land should lead to stable or slightly reduced storage.
It is difficult to be specific about how much carbon might be gained or lost through transformations of grasslands, pasturelands, or croplands. Although the amount of carbon in a square meter of forest might be on the order of 9 to 26 kilograms, depending on the condition of the forest and the age and type of trees growing, typical estimates of carbon storage in cultivated lands range from 1 to 8 kilograms per square meter, and estimates for uncultivated (but cultivatable) lands range from 2 to 10 kilograms per square meter.(Note 10) Thus, there is less carbon to be gained or lost, and the range of possible outcomes per unit of land is consequently smaller.
A recent study commissioned by the EPA estimated a current average soil carbon content for an area of 272 million acres of farmland in the United States at 4.8 to 7.9 kilograms per square meter.(Note 11) The study estimated that 1.0 billion to 1.6 billion metric tons of soil carbon had been lost from the farmland since it had been placed in cultivation, equivalent to 16 percent of the estimated original carbon content of the soil. The study also noted, however, that land with a soil carbon content of less than 4 kilograms per square meter was generally not being cultivated at the time of the study.
It would not be surprising if the least fertile farmland were the most likely to be removed from cultivation. Therefore, assuming that no trees are planted or naturally regenerate, the carbon gains from idling cropland are likely to be small: if the gains were commensurate with the original losses, they would be on the order of 0.6 kilograms per square meter (2.6 metric tons per acre), accrued over 20 to 50 years. However, the small gain would be distributed over a large acreage, since, as noted above, some 45 million acres were withdrawn from cultivation between 1982 and 1987. This implies eventual carbon storage of nearly 114 million metric tons, accrued at a rate of 2 million to 5 million metric tons per year. An estimate made for the year 1980 suggests that carbon sequestration from the abandonment of croplands in the United States and Canada was 3 million metric tons.(Note 12) These estimates would be considerably higher if all idled cropland had been afforested.
Converting land to forest should produce carbon gains, both through the addition of biomass (i.e., carbon stored in trees) and through the accretion of carbon into the soil, as dead limbs, trees, and roots gradually decay above and below ground. On average, the amount of carbon stored in U.S. forests is 17.7 kilograms per square meter of forestland (using the Birdsey/USDA estimate for storage), or 13 kilograms per square meter (using the EPA estimate). The range in forest storage across States is very large: from 9 kilograms per square meter in Nevada to 26 kilograms per square meter in Alaska, according to Birdsey.
As a crude numerical example, converting former cropland to forest might sequester approximately 10 kilograms per square meter over a 70-year period. (The actual amount would depend on the soil fertility and original carbon content of the land, the type of trees planted, and other factors specific to a particular plot of land.) This would produce 40.5 metric tons per acre over a 70-year period, or an average of 0.58 metric tons per acre per year. If the actual characteristics of the 6 million acres converted over the past 5 years matched this example, the conversion would ultimately store about 220 million metric tons of carbon over the next 70 years, at an average rate of 3.5 million metric tons per year. The actual annual rate could vary considerably from the average, since some species of trees grow much more rapidly in their early years than do others.
The examples above highlight two methodological problems common to adding carbon sources and sinks derived from land use data to more conventional greenhouse gas emissions:
There are parts of the United States (some areas in national parks, or parts of the interior of Alaska, for example) that remain close to an undisturbed state and continue to add biomass without human intervention. It is less clear that this carbon sequestration should "count" as anthropogenic. The broadest definition of anthropogenic would take the view that because humans control all land use in the United States, all land use decisions, whether of omission or commission, are anthropogenic acts. This argument might lead to the conclusion that by not cutting down and burning all its forests, the United States has saved 50 billion metric tons of carbon emissions in each year in which the forests were not cut down. Alternatively, too narrow a definition of anthropogenic could exclude unambiguous reforestation activities. There is no single universally acceptable definition of anthropogenic for the purpose of making an emissions inventory, and any decision that is made will inevitably be arbitrary to some degree.
As discussed in Chapter 1, there are numerous natural sources of methane. Anthropogenic land use changes inevitably affect those natural sources. One such natural source is wetlands. However, the stock of natural wetlands in the United States has diminished considerably over the past 2 centuries, which should, in principle, have reduced biogenic methane emissions. A recent study of wetland losses concluded that the United States had lost approximately 30 percent of its wetlands between colonial times and the mid-1980s. Almost all of this loss has occurred in the lower 48 States, which have lost 53 percent of their original wetlands.(Note 13) Ten States--Arkansas, California, Connecticut, Illinois, Indiana, Iowa, Kentucky, Maryland, Missouri, and Ohio--have lost 70 percent or more of their original wetland acreage. Nationally, remaining wetlands totaled approximately 274 million acres in 1985; wetlands lost totaled 119 million acres by the mid-1980s.
An update of the wetlands study indicates that 654,000 acres were converted from wetlands to other uses between 1982 and 1987, and that an additional 431,000 acres were converted between 1987 and 1991.(Note 14) Extra polating from these data, it is estimated that wetlands in the United States are currently destroyed at a rate of approximately 86,000 acres per year. It is difficult to find information on the conversion of other land categories to wetlands. It is assumed that the number and extent of wetland creations is small enough to leave the above loss estimates essentially unchanged.
The range of observed methane fluxes from U.S. wetlands is enormous. One survey of experiments conducted in the United States found estimates ranging from a negative flux (methane absorption) to a flux of 213 grams of methane per square meter per year, largely dependent on habitat type.(Note 15) Thus, it is difficult to extrapolate from experimental data to large-scale emissions estimates.
Estimates of global methane fluxes from wetlands tend to indicate that methane emissions from temperate-zone wetlands are minimal--typically between 5 and 10 million metric tons of methane per year for worldwide temperate-zone wetlands (which include U.S. wetlands)--when compared with estimated global wetlands emissions of 110 million metric tons.(Note 16) The U.S. share of all temperate-zone wetlands is about 57 percent, and temperate-zone wetlands lost during the 1980s accounted for about 0.5 percent of U.S. wetlands at the beginning of the period. Consequently, the reduction in natural methane emissions from wetlands lost might be on the order of 0.57 0.005 5 to 10 million metric tons of methane, or from 10,000 to 20,000 metric tons of methane annually over the decade.
The scientific literature suggests that grasslands and forestlands are both weak natural sinks for methane and weak natural sources for nitrous oxide. Natural soils apparently serve as methane sinks: well-aerated soils contain a class of bacteria called "methanotrophs," which use methane as food and oxidize it into carbon dioxide. Experiments indicate that cultivation reduces methane uptake by soils and increases nitrous oxide emissions.(Note 17)
Exactly how much methane is absorbed by natural soils, and how much nitrous oxide is emitted, is more difficult to estimate, although total amounts are very small. One report, based on experimental data, indicates that methane uptake in temperate evergreen and deciduous forests in the United States ranges from 0.19 to 3.17 milligrams (measured in carbon units) per square meter per day, equivalent to the uptake of 36.8 to 624.4 metric tons of methane per million acres per year.(Note 18) Thus (assuming no methane uptake at all from previous use), adding 6 million acres of forestland from 1987 to 1992 should increase methane absorption by 221 to 3,746 metric tons per year. Using the same numbers, all forestland in the United States may remove from 27,122 to 460,183 metric tons of methane per year. Comparing this figure with U.S. anthropogenic methane emissions of about 27 million metric tons per year, as estimated in this report, indicates that the magnitude of methane uptake by natural soils is not great.
Another report in the scientific literature indicates that some sample plots of pastureland in the United States have methane uptake rates of 4.1 milligrams (measured in carbon units) per square meter per day (for fertilized pasture) and 6.3 milligrams per square meter per day (for unfertilized pasture), with uptake from fertilized wheat and maize fields ranging from 0.2 to 0.9 milligrams per square meter.(Note 19) This study implies that an additional 0.6 to 6.1 milligrams of methane per square meter per day would be absorbed by abandoned farmland, equivalent to 118.1 to 1,201.1 grams per acre per year. Applying these figures to the 35 million acres of cropland taken out of production between 1982 and 1992 implies an increase in methane uptake of 4,133 to 42,038 metric tons per year from this source.
If such estimates are to be applied to emissions inventories, the same problem of crediting the uptakes applies. Removing an acre of farmland from production in a particular year creates a permanent annual methane sink that will absorb small additional amounts of methane each year thereafter, or at least until the use of the land changes. The method that should be used to credit such permanent reductions to a particular year is not obvious.
Nitrous oxide, while a highly potent greenhouse gas molecule-for-molecule compared to carbon dioxide, is released to the atmosphere from anthropogenic and natural sources in the United States in such trace amounts that its contribution to global warming is minimal. As noted in the Executive Summary, emissions of nitrous oxide were only 2 percent of U.S. GWP-weighted emissions in 1993. Of those emissions, it is difficult to say what percentage resulted from land uses, primarily due to the wide range of estimates concerning the magnitude of nitrous oxide emissions from nitrogen fertilizer use (which in any case should be included in agriculture statistics instead of land use statistics--see Chapter 4), and the lack of research data on nitrous oxide emissions from forest and grassland soils.
It is known that conversion of forests and grasslands to cropland accelerates nitrogen cycling and increases nitrous oxide emissions from the soil. It is not known with certainty by how much.(Note 20) Some estimates have been made of the difference between fertilized and unfertilized soils. The same research paper that described methane uptake in fertilized and unfertilized plots also described changes in nitrous oxide emissions from fertilized soils.(Note 21) (The mechanism by which this occurs is discussed in Chapter 4.) That paper indicated that unfertilized soils had emissions of 0.25 to 0.35 milligrams (measured in nitrogen units) per square meter per day, while emissions from fertilized soils ranged from 0.6 to 1.65 milligrams per square meter per day. Thus, abandoning fertilization should reduce nitrous oxide emissions by 0.35 to 1.3 milligrams per square meter per day--the equivalent of 86 to 321 metric tons of nitrous oxide per million acres per year. Applying this figure to the 35 million acres of cropland taken out of production between 1982 and 1992 implies a reduction in nitrous oxide emissions ranging from 3,010 to 11,235 metric tons annually. In principle, however, about three-quarters of the reduction in emissions from this source should be captured by reduced application of nitrogen fertilizers; thus, reporting emissions reductions using this method would result in significant double counting of units already included in the agriculture statistics in Chapter 4.