ghgemis (f7revised) [ms]

1. INTRODUCTION

The present paper offers a structure for analyzing the greenhouse gas contribution of deforestation in Brazilian Amazonia (Figure 1). It is hoped that this structure will prove valuable beyond the short time that the series of numbers for greenhouse gas emissions presented here remains the current best estimate. As the rates and locations of deforestation activity change, and as better data become available on these and other important factors, the estimates can be continually updated. Deforestation rates declined over the period from 1987 to 1991, but this is largely explained by Brazil's deepening economic recession, and cannot be extrapolated into the future (Fearnside, 1993a).

(Figure 1 here)

A variety of measures exists for expressing greenhouse gas emissions from tropical deforestation (see Makundi et al., 1992). One of these is net committed emissions, which considers the emissions and uptake that will occur as the landscape approaches a new equilibrium condition in a given deforested area (here the 13.8 X 10³ km² of Brazil's Amazonian forest that was cut in 1990). The "prompt emissions" (emissions entering the atmosphere in the year of clearing) are considered along with the "delayed emissions" (emissions that will enter the atmosphere in future years), as well as the corresponding uptake as replacement vegetation regrows on the deforested sites. Not included are trace gas emissions from the burning and decomposition of secondary forest and pasture biomass in the replacement landscape, although both trace gas and carbon dioxide fluxes are included for emissions originating from remains of the original forest biomass, from loss of intact forest sources and sinks, and from soil carbon pools. Net committed emissions are calculated as the difference between the carbon stocks in the forest and the equilibrium replacement landscape, with trace gas fluxes estimated based on fractions of the biomass that burn or decompose following different pathways.

Net committed emissions are not the same as the annual balance of emissions from the region. An annual balance calculation would consider the entire region (not just the part deforested in a single year), and would consider the fluxes of gases entering and leaving the region both through prompt emissions in the newly deforested areas and through the "inherited" emissions and uptakes in the clearings of different ages throughout the landscape. Inherited emissions and uptakes are the fluxes occurring in the year in question that are the result of clearing activity in previous years, for example, from decomposition or reburning of the remaining biomass of the original forest. The annual balance also includes trace gases from secondary forest and pasture burning and decomposition. The Framework Convention on Climate Change, signed in Rio de Janeiro in June 1992 by 155 countries plus the European Union, requires national estimates of fluxes and stocks for use in calculating the annual balance of emissions for each country. Presumably, the annual balance will form the basis for assigning responsibility for global warming in any protocols that may be negotiated under the Framework Convention.

Net committed emissions has several advantages over the annual balance as a measure of the impact of tropical deforestation, for example, for purposes of comparing the impacts of deforestation to those of fossil fuel combustion. Burning fossil fuels releases all gases in a single pulse, whereas deforestation sets in motion a process that commits emissions for a number of years after the trees are cut. Comparisons limited to prompt emissions from the two sources therefore grossly understate the impact of deforestation. The inherited emissions included in the annual balance avoids much of this distortion, but changes in the deforestation rate alter its appropriateness as a measure of the importance of deforestation. When clearing rates are increasing, the annual balance will understate the impact of deforestation, and when clearing rates are decreasing, it will overstate it. Except for the relatively small contribution of trace gases from secondary forest and pasture biomass, the annual balance and net committed emissions would be equal if deforestation rates were constant over an extended period of years. Using net committed emissions as a measure of the importance of deforestation is a more appropriate measure than the annual balance for assessment of deforestation policy questions in terms of the global interest (as opposed to each country's national interest under the Framework Convention).

Although better than the annual balance, net committed emissions is not ideal as the measure of deforestation's impact. The time when emissions occur is important, as human societies place a higher value on avoiding short-term impacts than on impacts far in the future. Although this preference is reflected in the time horizon of the global warming potentials used to express different trace gases in terms of carbon dioxide equivalents, the computations only include time preference with respect to the atmospheric load originating from a single pulse of gas, not the timing of the fluxes themselves in the case of delayed emissions. To the extent that deforestation has remained constant, net committed emissions are the same as the annual balance and knowledge of the timing of releases becomes unnecessary, thus greatly simplifying both data requirements and computation. A better index than net committed emissions would be provided by what may be called "time preference weighted emissions"--a measure derived from a non-equilibrium calculation using explicit fluxes for each year in the cleared area, and weighting these using both a time horizon and a scheme for assigning time preference (as by discounting or other alternative procedures). Discount rates or other time preference adjustments appropriate for such calculations are unlikely to be the same as those employed in financial analyses (Fearnside, 1995a).

2. EXTENT AND RATE OF DEFORESTATION

The present paper uses estimates of the extent and rate of deforestation by state derived from LANDSAT imagery (Tables I and II). The cumulative area deforested through 1991 had reached 427 X 10³ km² (including old clearings and hydroelectric dams), or 10.7% of the 4 X 10⁶ km² originally forested portion of Brazil's 5 X 10⁶ km² Legal Amazon region. Forest loss from 1978 through 1988 proceeded at an average of 20.4 X 10³ km²/year (Fearnside, 1993b); this rate estimate is derived, with a variety of adjustments, from estimates of deforestation extent in 1978 (Skole and Tucker, 1993) and in 1988 (Fearnside et al., 1990, see Fearnside, 1993a). Deforestation rate fell to 19.2 X 10³ km²/year in 1989, 13.8 X 10³ km²/year in 1990, and 11.1 X 10³ km²/year in 1991 (Brazil, INPE, 1992, see Fearnside, 1993a). The 1991 rate of forest loss was 20% less than the 1990 rate on which the emissions calculations in this paper are based.

(Tables I and II here)

The above rates cover only loss of primary forest within the portion of the region that was originally forested. Rates of conversion of the nonforest area, mainly cerrado (central Brazilian dry scrub savanna) are far less certain, but fortunately have less impact on greenhouse gas calculations due to the much lower biomass of savanna vegetation. The cerrado clearing rate for 1990 is assumed to be 5 X 10³ km²/year, a value lower than the 10 X 10³ km²/year used previously for 1990 (Fearnside, 1992a), and much lower than the 18 X 10³ km²/year estimated for 1988 (Fearnside, 1990). Based on comparison of Brazil's 1970 and 1985 agricultural censuses, Klink et al. (1994) estimated that 20 X 10³ km²/year of cerrado were cleared over that period, including areas outside of the Legal Amazon. The cerrado loss rate used in my present emissions estimate is believed to be conservative.

The rate of deforestation, together with the biomass of forest being cleared, affects the current (as opposed to potential) contribution of deforestation to the greenhouse effect. The rate of clearing has been calculated for each state (Table II), and is apportioned among various forest types within each state by assuming that, within each state, each forest type is cleared in proportion to the area in which it occurs outside of protected areas.

3. BIOMASS OF AMAZONIAN FORESTS

The initial biomass of the vegetation is an important factor affecting the magnitude of greenhouse gas emissions from deforestation. The biomass estimate used in the present paper-- 434 metric tons per hectare (t/ha) pre-logging total biomass for forests cleared in 1990 [post-logging biomass is 407 t/ha]--is based on much more data than earlier estimates and is derived as a weighted average from estimates that are disaggregated by state and forest type. The estimate indicates a substantial increase in biomass per hectare estimated for locations currently the focus of deforestation activity in Amazonia. It more than doubles the 155.1 t/ha value for total biomass derived by Brown and Lugo (1984) from FAO forest volume surveys for "tropical American undisturbed productive broadleafed forests" that has been used in recent global carbon balance calculations (e.g., Detwiler and Hall, 1988). It is also much higher than the 169.8 t/ha above-ground estimate by Brown et al. (1989) used as total biomass by Houghton (1991) for carbon emission estimates. Hall and Uhlig (1991) used the volume and expansion factor from the Brown et al. (1989) study, plus a correction for below-ground biomass, to obtain an estimate of 98.2 t C/ha (196.4 t biomass/ha) for undisturbed forests in "Brazil." The estimate used in the present study is also higher than the 211 t/ha total biomass estimated for areas cleared in 1988 for emissions calculations (Fearnside, 1991); a major reason for the increase is better data (derived from forest inventories of Brazil, Projeto RADAMBRASIL, 1973-1983) for biomass in the southern portion of the region where deforestation activity is concentrated. The estimate also has improved information on wood density and on below-ground biomass.

The different types of vegetation present in the Legal Amazon have been measured (Fearnside and Ferraz, 1995) from a digitized version of the 1:5,000,000 scale vegetation map of Brazil published by the Brazilian Institute for Geography and Statistics (IBGE) and the Brazilian Institute for Forestry Development (IBDF--since incorporated into the Brazilian Institute for the Environment and Renewable Natural Resources - IBAMA) (Brazil, IBGE and IBDF, 1988). The IBGE/IBDF (IBAMA) map code used indicates 28 vegetation types within the Brazilian Legal Amazon, of which 19 are considered here to be forest (Table III). This is a liberal definition of forest, including all ecotones between a forest and a nonforest vegetation type such as cerrado. So defined, the area of forest present according to the map totals 3.7 X 10⁶ km², or 74% of the 5 X 10⁶ km² Legal Amazon. The area originally forested totals 4.3 X 10⁶ km² (Table IV). The areas that were originally forest and nonforest using this definition are mapped in Figure 2. The deforestation estimate used here has as its base a definition of originally forested area corresponding to approximately 4.0 X 10⁶ km², therefore implying some remaining inconsistency.^(*)

(Tables III and IV and Figure 2 here)

Because the Legal Amazon is so big, each of its nine states being the size of countries in many parts of the world, vegetation with the same map code in different states cannot be assumed to have the same biomass. Considering each vegetation type in each state as a separate unit, here designated "ecoregions," there are a total of 111 different ecoregions in the Legal Amazon, of which 78 are "forest."

In order to estimate the area of each forest type being cleared annually in 1990, it was assumed that forests within each state are cleared in proportion to the area of each type (outside of parks and other legally protected areas). Although protected areas are not immune to deforestation, the small amount of clearing activity currently taking place inside these areas is undoubtedly insignificant from the standpoint of greenhouse gas emissions.

All biomass values given here and elsewhere in this paper refer to oven dry weight biomass. Unless otherwise noted, values are for total biomass, including both above- and below-ground portions, and including dead vegetation (but not soil carbon). All biomass fractions are included (leaves, small trees, vines, understory, etc.). Values are expressed in terms of biomass, rather than carbon (carbon content of biomass is 50%).

Biomass loading (biomass per hectare) of the different forested ecoregions is estimated from forest volume inventories in surveys carried out by the RADAMBRASIL project in the 1970s (Brazil, Projeto RADAMBRASIL, 1973-1983) and by the Food and Agriculture Organization of the United Nations (FAO) in the 1950s (Glerum, 1960; Heinsdijk, 1957, 1958a,b,c). A total of 2954 ha of usable data has been extracted from these studies for vegetation types classified as forest. Almost all of the FAO and RADAMBRASIL data are from one-hectare sample plots.

The parameters used for deriving the biomass estimates from the forest volume data lead to estimated biomass values substantially higher than those derived by Brown and Lugo (1992) from the FAO data and a summary of a portion of the RADAMBRASIL dataset. The difference is largely because of biomass components omitted from the Brown and Lugo estimates, including palms, vines, trees smaller than 10-cm DBH, dead biomass and below-ground biomass (Fearnside, 1992b, 1993c). Adjustments for these components are made based on available direct measurements (Fearnside, nd-a).

The total biomass is derived for each sample and the average for each ecoregion is calculated. Of the 78 forested ecoregions appearing on the IBAMA (Brazil, IBGE and IBDF, 1988) map, 44 (56%) have forest volume data available in the RADAMBRASIL or FAO datasets, and 34 (44%) do not. Fortunately, most of the ecoregions without data are of relatively minor importance from the standpoint of current greenhouse gas emissions. Of estimated biomass cleared in 1990, they total only 21.5%. For the ecoregions with no forest volume measurements, the mean biomass for the areas sampled in the same forest type (in the other states) is used as a substitute. For 5 of the 19 forest types, no measurement exists for any state. Forest types with no sample in any state represent only 1.0% of the estimated biomass cleared in 1990.

Mean biomass per hectare in each of the 78 forest types, including the values substituted as described above, are presented in Table V. It is evident that significant variation exists among states and among forest types, with pre-logging biomass loading ranging from 336 to 613 t/ha.

(Table V here)

The biomass stock in each ecoregion can be calculated by multiplying the per-hectare biomass (Table V) by the area in hectares (values from Table IV multiplied by 100 ha/km²). Table VI gives the approximate biomass stock cleared in million t for each ecoregion in the Legal Amazon. For the region's forests as a whole, the mean biomass loading (t/ha) for pre-logging biomass (weighted by the area of each ecoregion present) is estimated at 464 t/ha (Table V). In Table VI the loading for pre-logging biomass of forests cleared in 1990 (weighted by the deforestation rate in each state) is calculated as 434 t/ha. The pre-logging biomass in the areas cleared in 1990 is 6.5% lower than the average in the region as a whole, a difference equivalent to 892 km² of forest clearing.

(Table VI here)

4. TRANSFORMATIONS OF GROSS CARBON STOCKS

4.1 Land Uses Replacing the Forest

Estimates of the impact of deforestation have usually assumed that all deforested land is converted to cattle pasture (the dominant land use in deforested areas in Brazilian Amazonia). Some have even assumed that the forest is replaced with bare ground. Pasture has been assumed to remain indefinitely as the replacement for forest in estimates of net greenhouse gas emissions (e.g., Fearnside, 1985, 1987), and in simulations of impact on the water cycle (e.g., Shukla et al., 1990) and of the less-threatening changes in surface albedo (Dickinson and Henderson-Sellers, 1988). The results of such calculations are useful in identifying potential consequences of continued deforestation, but are unrealistic as quantitative predictions of contributions to climatic changes because the landscape that really replaces forest is not only pasture. The principal reason for using cattle pasture as the replacement vegetation has been the lack of more realistic scenarios of the evolution of the landscape after its initial conversion from forest to pasture. Here a first approximation is made using a simple first-order Markov model of transition probabilities between land-use classes (Fearnside, in press-a).

Markov matrices carry the assumption that the transfer probabilities remain unaltered over time--something for which there is no guarantee in practice. However, in most agricultural systems the tendency has been for population pressure to increase, leading to increased use intensity over time and shorter periods in secondary forest, with resulting lower average biomass for the landscape (e.g., Vermeer, 1970; UNESCO/UNEP/FAO, 1978). The assumption of constant transfer probabilities therefore is conservative from the point of view of greenhouse gas emissions. The assumption of constant transition probabilities is also optimistic because degradation of soil under pasture, combined with rainfall changes expected if the scale of deforestation should greatly expand, are likely to make low-biomass dysclimaxes, including grassy formations, the dominant land cover in a deforested Amazonia (Shukla et al., 1990; Fearnside, 1995b).

Exponentiation of the present matrix of transfer probabilities (Table VII) yields a vector representing the proportion of land in each category after establishment of equilibrium (Jeffers, 1978: 92-97). Performing these calculations indicates that the equilibrium landscape (Table VIII) would contain 0.0% forest, 4.0% farmland, 43.8% productive pasture, 5.2% degraded pasture, 2.0% secondary forest derived from farmland and 44.9% secondary forest derived from pasture. In Table VIII a weighted average is calculated of the biomass of vegetation in this equilibrium landscape, resulting in a value of 28.5 t/ha.

(Tables VII and VIII here)

The above calculations refer only to land that is cleared for agriculture and ranching. Hydroelectric development also removes forest land.

4.2. Fate of Biomass Carbon Stocks

The carbon stocks will change over a period of years to approach those in the equilibrium landscape, with the quantities in each pool increasing or decreasing at a different pace. The initial burn releases carbon immediately, while subsequent burns do so over a period of about ten years. Bacterial decomposition and termite activity also occurs largely over the first decade. Soil carbon pools change relatively quickly at the surface, but adjustments take longer for deeper pools (only carbon to 20 cm is considered in the current calculation). Charcoal (char) in the soil is a very long-term pool, considered to be permanently sequestered in the analysis.

Charcoal formed in burning is one way that carbon can be transferred to a long-term pool from which it cannot enter the atmosphere. A burn of forest being converted to cattle pasture near Manaus resulted in 2.7% of above-ground carbon being converted to charcoal (Fearnside et al., 1993). Using the mean of this and additional studies at Altamira and Manaus (Fearnside et al., nd-a,c) gives a mean of 1.9% (Table IX). This is substantially lower than the 15-23% assumed by Seiler and Crutzen (1980: 236) when they identified charcoal formation as a potentially important carbon sink (more recent calculations have used 5-10% charcoal yield: Crutzen and Andreae, 1990: 1672). Using the observed lower rate of charcoal formation would make global carbon cycle models indicate a larger contribution of greenhouse gases from tropical deforestation than has been the case using the higher rates of carbon transfer to long-term pools (e.g., Goudriaan and Ketner, 1984).

The burning behavior of ranchers can alter the amount of carbon passing into a long-term pool as charcoal. Carbon budget calculations often assume that forest is burned only once, and that all unburned biomass subsequently decomposes (e.g., Bogdonoff et al., 1985). This is not the typical pattern in cattle pastures that dominate land use in deforested areas in the Brazilian Amazon. Ranchers reburn pastures at intervals of 2-3 years to combat invasion of inedible woody vegetation. Logs lying on the ground when these reburnings occur are often burned. Some charcoal formed in earlier burns can be expected to be combusted as well. Parameters for transformations of gross carbon stocks are given in Table IX. A typical scenario of three reburnings over a 10-year period would raise the percentage of above-ground C converted to charcoal from 1.9% to 2.4% (Tables IX and X). In addition, a small amount of elemental carbon would be formed as graphitic particulates in the smoke (not considered here); over 80% of the elemental carbon formed remains on the site as charcoal (Kuhlbusch and Crutzen, 1995). The carbon transformations over a typical 10-year sequence are shown diagrammatically in Figure 3. These and other calculations are carried out in a series of interlinked spreadsheets.

(Figure 3 and Tables IX and X here)

5. SOURCES AND SINKS OF GREENHOUSE GASES

5.1. Burning

Biomass carbon not converted to charcoal (or elemental C in the smoke) is released gradually through combustion and decay, the relative importance of each affecting the gases emitted. If an area were burned only once, 33.2% of the pre-burn above-ground carbon would be released through combustion and 64.8% through decay (the remainder enters sinks as charcoal or graphitic particulates) (Table IX). With a typical scenario of three reburnings, 42.0% would be released through combustion and 55.6% through decay. Both combustion and decay release other trace gases such as methane.

In these calculations, the burning efficiency used for the initial burn (33.2%) is the mean of three studies (Fearnside et al., 1993, nd-a,b), adjusted for the effect of removal of trunks by logging (Fearnside, nd-b). The burning studies used found, respectively, efficiencies of 27.6% in a 1984 burn near Manaus, Amazonas, 42.0% as the mean of three 1986 burns near Altamira, Pará, and 28.3% in a 1990 burn near Manaus. Kauffman et al. (1995) have recently found higher burning efficiencies: 51.5% in a 1990 burn at Jacundá, Pará, 51.3% in a 1991 burn at Marabá, Pará, 40.5% in a 1992 burn at Santa Barbara, Rondônia, and 56.1% in a 1992 burn at Jamurí, Rondônia.

Parameters for carbon emissions (CO₂, CH₄ and CO) from the different burning and decay transformations of biomass are given in Table X. Two sets of parameters are given: a "low trace gas" and a "high trace gas" scenario, reflecting the range of values appearing in the literature for trace gas releases from such sources as termites and flaming and smoldering burns. Carbon emissions as CO₂, CH₄ and CO are diagrammed in Figure 4 with parameters for the low trace gas scenario. The low and high trace gas scenarios reflect the range of values for trace gas emissions only, not the level of uncertainty for factors affecting the amount of material that burns or decays, such as forest biomass, deforestation rate and burning efficiency. Parameters for other sources of greenhouse gases from land-use change are given in Table XI, and trace gas release parameters are given in Table XII.

(Figure 4 and Tables XI and XII here)

The amount of methane released is heavily dependent on the ratio of smoldering to flaming combustion; smoldering releases substantially more CH₄. Aircraft sampling over fires (mostly from virgin forest clearing) indicates that a substantial fraction of combustion is in smoldering form (Andreae et al., 1988). Logs consumed by reburning of cattle pastures are virtually all burned through smoldering rather than flaming combustion (personal observation).

Carbon monoxide (CO) is also produced by burning. This gas contributes indirectly to the greenhouse effect by impeding natural cleansing processes in the atmosphere that remove a number of greenhouse gases, including methane. Carbon monoxide removes hydroxyl radicals (OH), which react with CH₄ and other gases.

Burning also releases some nitrous oxide (N₂O), which contributes both to the greenhouse effect and to the degradation of stratospheric ozone. A sampling artifact has made measurements prior to 1989 unusable (Muzio and Kramlich, 1988). Estimates after discovery of the artifact indicate N₂O emissions from biomass burning are substantially lower than had previously been thought (Crutzen and Andreae, 1990). Parameters used in the present estimate (Table XII) are unaffected by the artifact.

5.2. Removal of Intact Forest Sources and Sinks

Deforestation makes an additional contribution to methane by removing a CH₄ sink in the soil of intact forest (Table XII). Removal of intact forest sources and sinks also affect the contribution of deforestation to a variety of compounds of nitrogen and oxygen (NO_x) and to non-methane hydrocarbons (NMHC), especially isoprenes. Effects of removing intact forest sources are included in the parameters for trace gases (Table XII). No forest sink is explicitly included for N₂O because the emission values used for this gas represent the net difference between forest and pasture emissions.

In the case of NMHC, the net effect of deforestation is to decrease the source strength over the 100-year period used in the current calculation. The magnitude of the effect of this reduction on global warming cannot be calculated yet. Indirect effects of trace gases were included in the 1990 IPCC scientific assessment of climate change (J.T. Houghton et al., 1990) but were dropped in the 1992 IPCC supplementary report pending resolution of disagreements over the magnitude of the effects (J.T. Houghton et al., 1992). Some of the indirect effects were restored in the 1994 report in the case of CH₄, raising its 100-year time horizon global warming potential from 11 to 24.5 as compared to CO₂ on a mass basis (Albritton et al., 1995: 222), making 1 t of CH₄ carbon equivalent to 8.9 t of CO₂ carbon [Note these values are expected to decrease to 21 and 7.6, respectively, in the 1995 IPCC report]. No indirect effects are yet agreed upon for NO_x and NMHC (as well as CO), which continue to be considered as having zero global warming potential (Albritton et al., 1995: 222).

5.3. Soil Carbon

Release of soil carbon would be expected when forest is converted to pasture because soil temperatures increase when forest cover is removed, thus shifting the balance between organic carbon formation and degradation to a lower equilibrium level (Cunningham, 1963; Nye and Greenland, 1960). A number of studies have found lower carbon stocks under pasture than forest (reviewed in Fearnside, 1980). For the same reason, naturally occurring tropical grasslands also have much smaller soil carbon stocks per hectare than do forests (Post et al., 1982). Lugo et al. (1986), however, have found increases in carbon storage in pasture soils in Puerto Rico, especially in drier sites, and suggest that tropical pastures may be a carbon sink. The present study treats soils as a source of carbon when forests are converted to pasture. All carbon released from soils is assumed to be in the form of CO₂.

Soil carbon in pasture is taken to be that in a profile equivalent to what is compacted from a 20-cm profile in the forest. Parameters used in deriving soil carbon changes are given in Table XIII. The layer compacted from the top 20 cm of forest soil releases 3.92 t/ha of carbon (the value used in the current calculations).

(Table XIII here)

The 3.92 t/ha release from the top 20 cm of soil represents 38% of the pre-conversion carbon present in this layer. This is higher than the 20% of pre-conversion carbon in the top 40 cm of soil that Detwiler (1986) concluded is released, on average, from conversion to pasture. The difference is not so great as it might seem: since carbon release is greatest nearest the surface, considering soil to 40 cm would thereby reduce the percentage released. One factor acting to compensate for any overestimation possibly caused by using a higher percentage of soil carbon release in the present study is the low bias introduced by having considered only the top 20 cm.

If soil to one-m depth were considered (the usual practice), and the same 38% of pre-conversion carbon were assumed to be released, then the release would be increased to 9.33 t/ha (Table XIII). The calculation to one-m depth considers that the top 20 cm of soil contains 42% of the carbon in a one-m profile (based on samples near Manaus: Fearnside, 1987). Brown and Lugo (1982: 183) have used a similar relationship to estimate carbon stocks in Thailand to a depth of one m from samples of the top 20 cm, considering 45% of the carbon in a one-m profile to be located in the top 20 cm.

Soil carbon release may be substantially higher than the values used here. Nepstad et al. (1994) have found carbon stocks 16 t/ha lower in the top one m of soil under degraded pasture as compared to intact forest, implying a carbon emission from soil about twice the values used. The combined effect of adopting these higher values and extending consideration from the top 20 cm to the top meter of soil would be to increase soil emissions by a factor of four, resulting in an additional net committed emission of about 30 X 10⁶ t of CO₂-equivalent carbon from the area cleared in 1990.

On the other hand, some studies have indicated less carbon loss than the value adopted here. Desjardins et al. (1994) found a net loss of 2.3 t/ha in the top 20 cm of soil in Pará, about 40% less than the value used here. The authors of the study believed that carbon loss in the 10-year-old pasture was relatively low because it had not been actively used "for a few years" (Desjardins et al., 1994: 112). Cerri et al. (1991: 254) found that 12 t/ha were lost from the top 20 cm during the first year, but that 8-year-old pasture had recovered the soil carbon stocks of forest and increased the stock in the top 20 cm to a level 6 t/ha above that in the forest soil. They attributed the strong recovery to "a rather exceptional management" in an "almost ideal well managed pasture" at an agricultural experiment station near Manaus (Cerri et al., 1991: 255). The studies indicating only slight carbon loss or a carbon gain do not include correction for soil compaction under pasture, making them underestimate the carbon loss that occurs from the layer of soil compacted from a given layer of the original forest soil (Fearnside, 1985, 1987; Veldkamp, 1993).

Soil below one-m depth may also be releasing significant amounts of carbon in deforested areas. Nepstad et al. (1994) have found carbon stocks at 1-8 m depth exceeding those in the top meter, and isotopic data indicate that this deeper pool experiences significant turnover, which is expected to result in carbon releases as the elimination of deep roots removes the source of replenishment for the deep soil carbon pool. Major adjustments of soil carbon inventory, including pools in the deeper soil layers, occur within the first decade after deforestation (Trumbore et al., 1995).

The present analyis considers all loss of carbon inventory from the top 20 cm of forest soil to be emissions, whether or not the losses occur through release of CO₂ directly to the atmosphere or by leaching to deeper soil layers and from there to groundwater and river outflow. The atmospheric effect of loss to groundwater is probably almost the same as if the carbon had been released directly. Since carbon in groundwater is in dissolved rather than particulate form, it would not enter ocean sediments and would remain exposed to oxidation after reaching the sea. Half of the carbon entering the Amazon River is oxidized before it reaches the ocean (Richey et al., 1980: 1350). A modest amount of carbon is undoubtedly lost through erosion as particulates and survives transport to long-term pools in ocean sediments (not considered in the current analysis).

5.4. Termites and Decay

A lively controversy surrounds the question of how much methane is produced by termites (Collins and Wood, 1984; Fraser et al., 1986; Rasmussen and Khalil, 1983; Zimmerman et al., 1982, 1984). This stems from differing estimates of the number of termites, the amount of wood consumed per termite, and the methane emission per gram of wood consumed. No measurement exists of the percentage of felled biomass that is ingested by termites in Amazonian clearings. Termite populations increase to a peak approximately 5-6 years after clearing, and subsequently decline as the available wood disappears (A.G. Bandeira, personal communication, 1990). It is assumed that none of the below-ground wood is ingested by termites: a conservative assumption given that termite species that eat buried wood are known to occur (Bandeira and Macambira, 1988) and termites consume underground biomass in other regions, such as Africa (e.g. Wood et al., 1977).

The termite emissions are limited by the ability of the termite populations to expand to consume the large amount of dead wood biomass that becomes available to them after deforestation. Both in the forest and in clearings, the inability of the observed populations to ingest the quantity of wood available makes the amount of wood biomass consumed independent of the amount of wood present. The calculations used here (Martius et al., nd) are based on the one available measurement of wood consumption, which is for the termite Nasutitermes macrocephalus (a wood-feeding species in the Amazonian várzea): 49 mg of wood dry weight consumption per gram of termite biomass per day (Martius, 1989: 228).

Both the high and low trace gas scenarios use the average emission rate for wood-feeders (Martius, et al., 1993) applied to the total termite biomass of wood-, soil- and leaf-feeders. This emission of 3.01.3 μg/hour is equivalent to 0.0022 tons of carbon released as methane per ton of carbon consumed (considering 50% carbon content of wood). This is virtually identical to the value obtained by Seiler et al. (1984) for termites in Africa. It is lower by a factor of four than the 0.0079 t methane C released/t wood C consumed estimated for Amazonian termites by Goreau and de Mello (1987). However, the lower value (Martius, et al., 1993) is believed to be more reliable, as the higher value was based on monitoring a single nest for only two days, whereas the value used here is based on monitoring 15 nests for two years.

5.5. Cattle and Pasture

Methane is produced in the rumens of cattle that occupy pastures in deforested areas. The portion of the deforested area considered to be maintained under pasture is that derived from the equilibrium landscape (Tables VII and VIII). Parameters used to derive methane emissions from cattle are included in Table XI.

Pasture soils in Amazonia emit N₂O in quantities substantially higher than forest soils when measurements are made over a full annual cycle (Luizão et al., 1989). Most emissions are in the wet season, and are not reflected in measurements restricted to the dry season (e.g. Goreau and de Mello, 1987). Unlike emissions from the initial burning, conversion of a given hectare to pasture does not result in a one-time release of this greenhouse gas, but rather a continuous additional flux at this rate for as long as the area is maintained under this land use (the assumption in the present calculation). However, Keller et al. (1993) have found that N₂O emissions from pasture soil declines with pasture age in Costa Rica, falling below emission levels from forest soils after 10 years.

One factor not included in the calculation is production of trace gases by reburning of pasture and secondary forest. Combustion of logs remaining from the original forest is included. Burning of the biomass of the pasture itself and of secondary forest does not contribute to net release of carbon dioxide, as the same amount of carbon is reabsorbed when the vegetation regrows. However, CH₄, CO, N₂O and NO_x do increase as a result of reburnings as these gases do not enter photosynthetic reactions. Methane degrades to CO₂ after an average of 14.5 years (Albritton et al., 1995: 222), and CO degrades after a few months (Thompson and Cicerone, 1986: 10,857), after which the carbon can return to the vegetation. The trace gas inputs of reburning the replacement vegetation represent one of several factors not included in the current calculation of net committed emissions, but which are included in calculations of the annual balance of emissions (Fearnside, in press-b). Burning of cerrado and other savanna vegetation without clearing of trees is a source of trace gas emissions in the annual balance, but does not contribute to net committed emissions (because only the fate of the cleared area is followed).

5.6. Hydroelectric Dams

Because no new reservoirs were filled in 1990, the calculations presented above consider only emissions from conversion of natural forest vegetation to a landscape dominated by cattle pasture--the dominant trend in Brazilian Amazonia today. Hydroelectric dams in rainforest areas release greenhouse gases both by decomposition of dead forest left standing in the reservoirs and by the continuing release of methane from flooded areas (especially in portions that are alternately dried and flooded).

Although hydroelectric dams are commonly believed to have no impact on the greenhouse effect, in contrast to fossil fuel use, the validity of this conclusion depends heavily on the biomass of the vegetation in the flooded areas and on the power output of the dams. In Amazonia, dams are sometimes worse than petroleum from the point of view of greenhouse gas emissions. The worst case is Balbina Dam, which was closed in 1987. Located on relatively flat terrain, Balbina's shallow 3147 km² reservoir can generate only enough power to deliver annually an average of 109 megawatts to Manaus (Fearnside, 1989a). The biomass of the flooded forest is now decomposing, releasing its carbon to the atmosphere. Methane is also released from decay of biomass under the anoxic conditions at the bottom of the reservoir. Generating the same energy from petroleum would take 250 years to equal the carbon release from flooding the Balbina reservoir (based on Junk and Nunes de Mello, 1987; see Fearnside, 1989a). However, most of the carbon release from reservoirs takes place in the first decade, which increases the importance of these emissions relative to those from fossil fuels when calculations include time horizons and/or weighting of emissions for time preference as by discounting. In 1990, Balbina was releasing twenty times more CO₂-equivalent carbon than would have been produced by generating the same power from fossil fuels (Fearnside, 1995c).

Calculating emissions for hydroelectric dams requires estimating vertical distribution of the biomass in order to calculate releases occurring above the water, in the seasonally-inundated zone, and underwater in the surface and anoxic water portions of the reservoir. Transfers from the above-water to the underwater categories occur as branches and trunks fall. Assumptions about decay rates and paths allow estimation of emissions of CO₂ and CH₄ from forest biomass. To these emissions are added CH₄ releases from decay of macrophytes and of organic matter entering the reservoir from the river. The emission in 1990 from the approximately 5500 km² of reservoirs in the region was 38 X 10⁶ t of CO₂ gas and 0.22 X 10⁶ t of CH₄ gas (Fearnside, 1995c). In 1990, no new reservoirs were filled in the Legal Amazon.

Amazonian várzea (white water floodplain) has been identified as one of the world's major sources of atmospheric methane (Mooney et al., 1987). Várzea occupies about 2% of the five X 10⁶ km² Legal Amazon, the same percentage that would be flooded if all of the reservoirs planned for the region are created. A total of 100,000 km² would be flooded (Brazil, ELETROBRÁS, 1987: 150), or about 20 times the present reservoir area. Virtually all planned hydroelectric dams are in the forested portion of the region--and the reservoirs would flood approximately 2.5% of this portion of the region. Were these reservoirs to contribute an output of methane per hectare on the same order as that produced by the várzea, they would together represent a significant contribution to the greenhouse effect. The annual emission of methane to the atmosphere from the várzea has been calculated by Devol et al. (1990) to be 5.1 X 10⁶ t CH₄/year. Like biogenic release of N₂O, this would be a permanent addition to greenhouse gas sources, rather than a one-time input. Unfortunately, CH₄ emission measurements from Amazonian reservoirs do not yet exist, making measurements in várzea lakes the best available surrogate.

6. NET COMMITTED EMISSIONS

The quantities of gases released by each source and absorbed by each sink in the 13.8 X 10³ km² of forest and 5 X 10³ km² of cerrado cleared in 1990 are given in Table XIV for the low trace gas scenario. Table XV presents the corresponding results for the high trace gas scenario. Although the emissions of CO₂ dwarf the absolute quantities of the other gases, the greater greenhouse impact per ton of the latter gives them a significant role in deforestation's contribution to global warming.

(Tables XIV and XV here)

I have converted my estimated emissions of trace gases to CO₂ equivalents (Table XVI). The conversion uses the global warming potentials adopted by the Intergovernmental Panel for Climate Change (IPCC) in its 1994 report (Albritton et al., 1995: 222). These express the impact on global warming of an immediate emission of one ton of each gas relative to emitting one ton of CO₂ gas, considering a 100-year time horizon without discounting. The exclusion of a number of indirect effects, the long time horizon and the lack of discounting all reduce the weight given to CH₄ and CO relative to CO₂, thereby reducing the calculated impact of tropical deforestation relative to fossil fuel combustion (Fearnside, 1992a).

(Table XVI here)

Using this conservative methodology for expressing trace gas impact, the CO₂-equivalent carbon from trace gases represents 5% and 9% of the total in the low and high trace gas scenarios, respectively. The total contribution of 1990 clearing in forest and cerrado was 274-285 X 10⁶ t of CO₂-equivalent carbon (Table XVI). In addition to this, logging in 1990 emitted 61 X 10⁶ t CO₂-equivalent C (Fearnside, nd-b). As this was calculated assuming a constant logging harvest at the 1988 official rate (24.6 X 10⁶ m³/year), it reflects net committed emissions from logging. Adding this to the total from forest clearing raises the net committed emission for 1990 land-use change in the Brazilian Legal Amazon to 335-346 X 10⁶ t CO₂-equivalent carbon. This represents roughly 5% of the total global emissions from deforestation and fossil fuel sources, indicating the high impact of deforestation. At the same time, it makes clear the fact that reducing or halting deforestation is not enough to avoid the most damaging consequences of global warming, and that use of fossil fuels must also be substantially reduced.

* NOTE

(*) Some inconsistency remains in the definition of original forest area used here (Tables III and IV), and that used in the deforestation estimate (Tables I and II). The deforestation estimate used a line between forest and nonforest drawn by INPE from LANDSAT-TM 1:250,000 scale images with some reference to the RADAMBRASIL vegetation maps (but without a list of vegetation types classified as forest and nonforest). The area so defined has not yet been measured by INPE, but a compilation by map sheet (using IBGE 1:250,000 scale maps as a geographical base) was made of the approximate proportions of forest and nonforest in each sheet. The total from this compilation is 4.0 X 10⁶ km², lower than the 4.3 X 10⁶ km² measured from the IBGE/IBDF 1:5,000,000 scale map.

The "present" vegetation is also inconsistent: the IBGE/IBDF mapping totals 3.7 X 10⁶ km² of forest (circa 1988) (Table IV), whereas the original forest area from the same map, less the area deforested by 1988 (Table I), yields a total of 3.9 X 10⁶ km².

ACKNOWLEDGMENTS

Studies on burning in Altamira were funded by National Science Foundation grants GS-422869 (1974-1976) and ATM-86-0921 (1986-1988), in Manaus by World Wildlife Fund-US grant US-331 (1983-1985), and in Manaus and Roraima by the Pew Scholars Program in Conservation and the Environment and by the Fundação Banco do Brasil (Grant 10/1516-2). I thank P. Crutzen, C.A.S. Hall, O. Masera, G.H. Platais and S.V. Wilson for comments.

REFERENCES

Ahuja, D.R.: 1990, `Regional anthropogenic emissions of greenhouse gases', in Intergovernmental Panel on Climate Change (IPCC), Response Strategies Working Group (RSWG), Subgroup on Agriculture, Forestry and other Human Activities (AFOS), Proceedings of the Conference on Tropical Forestry Response Options to Global Climate Change, U.S. Environmental Protection Agency, Office of Policy Assessment (USEPA-OPA, PM221), Washington, DC, 417-461.

Albritton, D.L., Derwent, R.G., Isaksen, I.S.A., Lal, M. and Wuebbles, D.J.: 1995, 'Trace gas radiative forcing indices' in Houghton, J.T., Meira Filho, L.G., Bruce, J., Hoesung Lee, Callander, B.A., Haites, E., Harris N. and Maskell, K. (eds.) Climate Change 1994: Radiative Forcing of Climate Change and An Evaluation of the IPCC IS92 Emission Scenarios, Cambridge University Press, Cambridge, UK, 205-231.

Andreae, M.O., Browell, E.V., Garstang, M., Gregory, G.L., Harriss, R.C., Hill, G.F., Jacob, D.J., Pereira, M.C., Sachse, G.W., Setzer, A.W., Silva Dias, P.L., Talbot, R.W., Torres, A.L., and Wofsy, S.C.: 1988, `Biomass-burning emissions and associated haze layers over Amazonia', J. Geophys. Res. 93(D2): 1509-1527.

Bandeira, A.G. and Macambira, M.L.J.: 1988, `Térmitas de Carajás, Estado do Pará, Brasil: Composição faunística, distribuição e hábito alimentar', Boletim do Museu Paraense Emílio Goeldi: Zoologia 4(2): 175-190.

Barbosa, R.I.: 1994, Efeito Estufa na Amazônia: Estimativa da Biomassa e a Quantificação do Estoque e Liberação de Carbono na Queima de Pastagens Convertidas de Florestas Tropicais em Roraima, Brasil, Masters thesis in ecology, Instituto Nacional de Pesquisas da Amazônia (INPA) and Universidade Federal do Amazonas (UFAM), Manaus, Brazil.

Bogdonoff, P., Detwiler, R.P., and Hall, C.A.S.: 1985, `Land use change and carbon exchange in the tropics: III. Structure, basic equations, and sensitivity analysis of the model', Environ. Manage. 9(4): 345-354.

Brazil, Instituto Brasileiro de Geografia e Estatística (IBGE) and Instituto Brasileiro de Desenvolvimento Florestal (IBDF): 1988, Mapa de Vegetação do Brasil, Map Scale 1:5,000,000, IBAMA, Brasilia.

Brazil, Instituto Nacional de Pesquisas Espaciais (INPE): 1992, Deforestation in Brazilian Amazonia. INPE, São José dos Campos, São Paulo, 3 pp.

Brazil, Ministério das Minas e Energia, Centrais Elétricas do Brasil (ELETROBRÁS): 1987, Plano 2010: Relatório Geral. Plano Nacional de Energia Elétrica 1987/2010. (Dezembro de 1987). ELETROBRÁS, Brasilia.

Brazil, Ministério das Minas e Energia, Departamento Nacional de Produção Mineral (DNPM), Projeto RADAMBRASIL: 1973-1983, Levantamento de Recursos Naturais, Vols. 1-23, DNPM, Rio de Janeiro.

Brown, S. and Lugo, A.E.: 1982, `The storage and production of organic matter in tropical forests and their role in the global carbon cycle', Biotropica 14(3): 161-187.

Brown, S. and Lugo, A.E.: 1984, `Biomass of tropical forests: A new estimate based on forest volumes', Science 223: 1290-1293.

Brown, S. and Lugo, A.E.: 1992, 'Aboveground biomass estimates for tropical moist forests of the Brazilian Amazon', Interciencia 17(1): 8-18.

Brown, S., Gillespie, A.J.R., and Lugo, A.E.: 1989, `Biomass estimation methods for tropical forests with applications to forest inventory data', For. Sci. 35(4): 881-902.

Buschbacher, R.J.: 1984, Changes in Productivity and Nutrient Cycling following Conversion of Amazon Rainforest to Pasture, Ph.D. thesis, University of Georgia, Athens, GA.

Cerri, C.C., Volkoff, B., and Andreux, F.: 1991, `Nature and behaviour of organic matter in soils under natural forest, and after deforestation, burning and cultivation, near Manaus', For. Ecol. Manage. 38: 247-257.

Cofer, W.R., Levine, J.S., Riggan, P.H., Sebacher, D.I., Winstead, E.L., Shaw, E.F., Brass, J.A., and Ambrosia, V.G.: 1988, `Trace gas emissions from a mid-latitude prescribed chaparral fire', J. Geophys. Res. 93: 1653-1658.

Collins, N.M. and Wood, T.G.: 1984, `Termites and atmospheric gas production', Science 224: 84-85.

Crutzen, P.J. and Andreae, M.O.: 1990, `Biomass burning in the tropics: Impact on atmospheric chemistry and biogeochemical cycles', Science 250: 1669-1678.

Crutzen, P.J., Aselmann, I., and Seiler, W.: 1986, `Methane production by domestic animals, wild ruminants, other herbivorous fauna, and humans', Tellus 38B: 271-284.

Crutzen, P.J., Delany, A.C., Greenberg, J., Haagenson, P., Heidt, L., Lueb, R., Pollock, W., Seiler, W., Wartburg, A., and Zimmerman, P.: 1985, `Tropospheric chemical composition measurements in Brazil during the dry season', J. Atm. Chem. 2: 233-256.

Cunningham, R.H.: 1963, `The effect of clearing a tropical forest soil', J. Soil Sci. 14: 334-344.

Desjardins, T., Andreux, F., Volkoff, B., and Cerri, C.C.: 1994, `Organic carbon and ¹³C contents in soils and soil size-fractions, and their changes due to deforestation and pasture installation in eastern Amazonia', Geoderma 61: 103-118.

Detwiler, R.P.: 1986, `Land use change and the global carbon cycle: The role of tropical soils', Biogeochem. 2: 67-93.

Detwiler, R.P. and Hall, C.A.S.: 1988, `Tropical forests and the global carbon cycle', Science 239: 42-47.

Devol, A.H., Richey, J.H., Forsberg, B.R., and Martinelli, L.A.: 1990, `Seasonal dynamics in methane emissions from the Amazon River floodplain to the troposphere', J. Geophys. Res. 95(D10): 16,417-16,426.

Dickinson, R.E. and Henderson-Sellers, A.: 1988, `Modelling tropical deforestation: A study of GCM land-surface parameterizations', Quar. J. Royal Meteorolog. Soc. 114: 439-462.

dos Santos, J.R.: 1989, `Estimativa da biomassa foliar das savanas brasileiras: Uma abordagem por sensoriamento remoto', in IV Simpósio Latinamericano en Percepción Remota, IX Reunión plenaria SELPER, 19 al 24 de noviembre de 1989, Bariloche, Argentina. Tomo 1: 190-199.

Falesi, I.C.: 1976, Ecossistema de Pastagem Cultivada na Amazônia Brasileira, Centro de Pesquisa Agropecuária do Trópico Úmido (CPATU), Belém, Para.

Fearnside, P.M.: 1979, `Cattle yield prediction for the Transamazon Highway of Brazil', Interciencia 4(4): 220‑225.

Fearnside, P.M.: 1980, `The effects of cattle pasture on soil fertility in the Brazilian Amazon: Consequences for beef production sustainability', Trop. Ecol. 21(1): 125-137.

Fearnside, P.M.: 1985, `Brazil's Amazon forest and the global carbon problem', Interciencia 10(4): 179‑186.

Fearnside, P.M.: 1987, `Summary of progress in quantifying the potential contribution of Amazonian deforestation to the global carbon problem', in Athié, D., Lovejoy, T.E., and Oyens, P. de M., (eds.) Proceedings of the Workshop on Biogeochemistry of Tropical Rain Forests: Problems for Research, Universidade de São Paulo, Centro de Energia Nuclear na Agricultura (CENA), Piracicaba, São Paulo, 75-82.

Fearnside, P.M.: 1989a, `Brazil's Balbina Dam: Environment versus the legacy of the pharaohs in Amazonia', Environ. Manage. 13(4): 401-423.

Fearnside, P.M.: 1989b, A Ocupação Humana de Rondônia: Impactos, Limites e Planejamento (CNPq Relatórios de Pesquisa No. 5), Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq), Brasilia.

Fearnside, P.M.: 1990, `The rate and extent of deforestation in Brazilian Amazonia', Environ. Cons. 17(3): 213-216.

Fearnside, P.M.: 1991, `Greenhouse gas contributions from deforestation in Brazilian Amazonia', in Levine, J.S. (ed.) Global Biomass Burning: Atmospheric, Climatic, and Biospheric Implications, MIT Press, Cambridge, MA, 92-105.

Fearnside, P.M.: 1992a, Greenhouse Gas Emissions from Deforestation in the Brazilian Amazon, Carbon Emissions and Sequestration in Forests: Case Studies from Developing Countries, Volume 2, LBL-32758, UC-402, Climate Change Division, Environmental Protection Agency, Washington, DC and Energy and Environment Division, Lawrence Berkeley Laboratory (LBL), University of California (UC), Berkeley, CA.

Fearnside, P.M.: 1992b, `Forest biomass in Brazilian Amazonia: Comments on the estimate by Brown and Lugo', Interciencia 17(1): 19-27.

Fearnside, P.M.: 1993a, `Deforestation in Brazilian Amazonia: The effect of population and land tenure', Ambio 22(8): 537-545.

Fearnside, P.M.: 1993b, `Desmatamento na Amazônia: Quem tem razão --o INPE ou a NASA?', Ciência Hoje 16(96): 6-8.

Fearnside, P.M.: 1993c, `Biomass of Brazil's Amazonian forests: Reply to Brown and Lugo revisited', Interciencia 18(1): 5-7.

Fearnside, P.M.: 1995a, `Global warming response options in Brazil's forest sector: Comparison of project-level costs and benefits', Biomass and Bioenergy 8(5): 309-322.

Fearnside, P.M.: 1995b, `Potential impacts of climatic change on natural forests and forestry in Brazilian Amazonia', For. Ecol. Manage. 78: 51-70.

Fearnside, P.M.: 1995c, `Hydroelectric dams in the Brazilian Amazon as sources of greenhouse gases', Environ. Cons. 22(1): 7-19.

Fearnside, P.M.: 1996, `Amazonian deforestation and global warming: Carbon stocks in vegetation replacing Brazil's Amazon forest', For. Ecol. Manage. 80, 21-34.

Fearnside, P.M.: in press-b, `Amazonia and global warming: Annual balance of greenhouse gas emissions from land-use change in Brazil's Amazon region', in Levine, J. (ed.) Biomass Burning and Global Change, MIT Press, Cambridge, MA.

Fearnside, P.M.: nd-a, `Biomass of Brazil's Amazonian forests', (in preparation).

Fearnside, P.M.: nd-b, 'Tropical forest logging and management: Implications for global warming', (in preparation).

Fearnside, P.M. and Ferraz, J.: 1995, `A conservation gap analysis of Brazil's Amazonian vegetation', Cons. Biol. 9(5): 1134-1147.

Fearnside, P.M. and Guimarães, W.M.: 1996, `Carbon uptake by secondary forests in Brazilian Amazonia', For. Ecol. Manage. 80, 35-46.

Fearnside, P.M., Tardin, A.T. and Meira Filho, L.G.: 1990, Deforestation rate in Brazilian Amazonia, Instituto de Pesquisas Espaciais (INPE), São José dos Campos, São Paulo, 8 pp.

Fearnside, P.M., Leal Filho, N., and Fernandes, F.M.: 1993, `Rainforest burning and the global carbon budget: Biomass, combustion efficiency and charcoal formation in the Brazilian Amazon', J. Geophys. Res. 98(D9): 16,733-16,743.

Fearnside, P.M., Graça, P.M.L.A., Leal Filho, N., Rodrigues, F.J.A., and Robinson, J.M.: nd-a, `Tropical forest burning in Brazilian Amazonia: Measurements of biomass, burning efficiency and charcoal formation at Altamira, Pará', (in preparation).

Fearnside, P.M., Leal Filho, N., Graça, P.M.L.A., Ferreira, G.L., Custodio, R.A., and Rodrigues, F.J.A.: nd-b, `Pasture biomass and productivity in Brazilian Amazonia', (in preparation).

Fearnside, P.M., Barbosa, R.I., and Graça, P.M.L.A.: nd-c, `Burning of secondary forest in Amazonia: Biomass, burning efficiency and charcoal formation during land preparation for agriculture in Roraima, Brazil', (in preparation).

Fearnside, P.M., Graça, P.M.L.A., and Rodrigues, F.J.A.: nd-d, `Biomass, combustion efficiency and charcoal formation in Amazonia: Measurements in a rainforest burn near Manaus, Brazil', (in preparation).

Fraser, P.J., Rasmussen, R.A., Creffield, J.W., French, J.R., and Khalil, M.A.K.: 1986, `Termites and global methane--another assessment', J. Atmos. Chem. 4: 295-310.

Glerum, B.B.: 1960, Report to the Government of Brazil on a Forestry Inventory in the Amazon Valley (Part Five) (Region between Rio Caete and Rio Maracassume), FAO Report No. 1250, Project No. BRA/FO, Food and Agriculture Organization of the United Nations (FAO), Rome, 67 pp.

Goreau, T.J. and de Mello, W.Z.: 1987, `Effects of deforestation on sources and sinks of atmospheric carbon dioxide, nitrous oxide, and methane from central Amazonian soils and biota during the dry season: A preliminary study', in Athié, D., Lovejoy, T.E., and Oyens, P. de M., (eds.) Proceedings of the Workshop on Biogeochemistry of Tropical Rain Forests: Problems for Research, Universidade de São Paulo, Centro de Energia Nuclear na Agricultura (CENA), Piracicaba, São Paulo, 51-66.

Goudriaan, J. and Ketner, P.: 1984, `A simulation study for the global carbon cycle, including man's impact on the biosphere, Clim. Change 6: 167-192.

Greenberg, J.P., Zimmerman, P.R., Heidt, L., and Pollock, W.: 1984, `Hydrocarbon and carbon monoxide emissions from biomass burning in Brazil', J. Geophys. Res. 89(D1): 1350-1354.

Guimarães, W.M.: 1993, Capoeira em Pastagens Degradadas e o Efeito Estufa: Quantificação do Papel da Vegetação Secundária no Balanço de Carbono Resultante do Desmatamento na Amazônia Brasileira, Masters thesis, Instituto Nacional de Pesquisas da Amazônia/Fundação Universidade do Amazonas (INPA/FUA), Manaus.

Hall, C.A.S., and Uhlig, J.: 1991, `Refining estimates of carbon released from tropical land-use change', Can. J. For. Res. 21: 118-131.

Hecht, S.B.: 1981, `Deforestation in the Amazon Basin: Magnitude, dynamics, and soil resource effects', Studies in Third World Societies No. 13: 61-108.

Heinsdijk, D.: 1957, Report to the Government of Brazil on a Forest Inventory in the Amazon Valley (Region between Rio Tapajós and Rio Xingú), FAO Report No. 601, Project No. BRA/FO, Food and Agriculture Organization of the United Nations (FAO), Rome, 135 pp.

Heinsdijk, D.: 1958a, Report to the Government of Brazil on a Forestry Inventory in the Amazon Valley (Part Two) (Region between Rio Xingú and Rio Tocantins), FAO Report No. 949, Project No. BRA/FO, Food and Agriculture Organization of the United Nations (FAO), Rome, 83 pp.

Heinsdijk, D.: 1958b, Report to the Government of Brazil on a Forest Inventory in the Amazon Valley (Part Three) (Region between Rio Tapajós and Rio Madeira), FAO Report No. 969, Project No. BRA/FO, Food and Agriculture Organization of the United Nations (FAO), Rome, 72 pp.

Heinsdijk, D.: 1958c, Report to the Government of Brazil on a Forest Inventory in the Amazon Valley (Part Four) (Region between Rio Tocantins and Rios Guamá and Capim), FAO Report No. 992, Project No. BRA/FO, Food and Agriculture Organization of the United Nations (FAO), Rome, 94 pp.

Houghton, J.T., Callander, B.A., and Varney, S.K. (eds.): 1992, Climate Change 1992: The Supplementary Report to the IPCC Scientific Assessment, Cambridge University Press, Cambridge, UK.

Houghton, J.T., Jenkins, G.J., and Ephraums, J.J. (eds.): 1990, Climate Change: The IPCC Scientific Assessment, Cambridge University Press, Cambridge, UK.

Houghton, R.A.: 1991, `Tropical deforestation and atmospheric carbon dioxide', Clim. Change 19(1-2): 99-118.

Jeffers, J.N.R.: 1978, An Introduction to Systems Analysis: With Ecological Applications, Arnold, London.

Junk, W.J. and Nunes de Mello, J.A.S.: 1987, `Impactos ecológicos das represas hidrelétricas na bacia amazônica brasileira', in Kohlhepp, G. and Schrader, A., (eds.) Homem e Natureza na Amazônia, Tübinger Geographische Studien 95 (Tübinger Beiträge zur Geographischen Lateinamerika-Forschung 3), Geographisches Institut, Universität Tübingen, Tübingen, Germany, 367‑385.

Kaplan, W.A., Wofsy, S.C., Keller, M., and da Costa, J.M.: 1988, `Emission of NO and deposition of O₃ in a tropical forest system', J. Geophys. Res. 93: 1389-1395.

Kauffman, J.B., Cummings, D.L., Ward, D.E. and Babbitt, R.: 1995, `Fire in the Brazilian Amazon: Biomass, nutrient pools, and losses in slashed primary forests', Oecologia 104: 397-408.

Kaufman, Y.J., Setzer, A.W., Justice, C., Tucker, C.J., Pereira, M.G., and Fung, I.: 1990, `Remote sensing of biomass burning in the tropics', in Goldammer, J.G., (ed.), Fire in the Tropical Biota: Ecosystem Processes and Global Challenges, Springer-Verlag, Heidelberg, Germany, 371-399.

Keller, M., Jacob, D.J., Wofsy, S.C., and Harriss, R.C.: 1991, `Effects of tropical deforestation on global and regional atmospheric chemistry', Clim. Change 19(1-2): 139-158.

Keller, M., Kaplan, W.A., and Wofsy, S.C.: 1986, `Emissions of N₂O, CH₄ and CO₂ from tropical forest soils', J. Geophys. Res. 91: 11,791-11,802.

Keller, M., Veldkamp, E., Weltz A.M., and Reiners, W.: 1993, `Effect of pasture age on soil trace-gas emissions from a deforested area of Costa Rica', Nature 365: 244-246.

Klink, C.A., Macedo, R.H., and Mueller, C.C.: 1994, Cerrado: Processo de Ocupação e Implicações para a Conservação e Utilização da sua Diversidade Biológica, World Wide Fund for Nature (WWF)-Brasil, Brasilia.

Kuhlbusch, T.A.J., and Crutzen, P.J.: 1995, `Toward a global estimate of black carbon in residues of vegetation fires representing a sink of atmospheric CO₂ and a source of O₂', Global Biogeochem. Cycles, in press.

Lugo, A.E., Sanchez, M.M., and Brown, S.: 1986, `Land use and organic carbon content of some subtropical soils', Plant and Soil 96: 185-196.

Luizão, F., Matson, P., Livingston, G., Luizão, R., and Vitousek, P.: 1989, `Nitrous oxide flux following tropical land clearing', Global Biogeochem. Cycles 3: 281-285.

Makundi, W.R., Sathaye, J.A., and Masera, O.: 1992, Summary, Carbon Emissions and Sequestration in Forests: Case Studies from Developing Countries, Volume 1, LBL-32758, UC-402, Climate Change Division, Environmental Protection Agency (EPA), Washington, DC and Energy and Environment Division, Lawrence Berkeley Laboratory (LBL), University of California (UC), Berkeley, CA.

Martius, C.: 1989, Untersuchungen zur Ökologie des Holzabbaus durch Termiten (Isoptera) in zentralamazonischen Überschwemmunsgswäldem (Várzea), AFRAA-Verlag, Frankfurt, Germany.

Martius, C., Fearnside, P.M., Bandeira, A.G. and Wassmann, R.: 1996, `Deforestation and methane release from termites in Amazonia', Chemosphere, 33, 517-536.

Martius, C., Wassmann, R., Thein, U., Bandeira, A., Rennenberg, H., Junk, W., and Seiler, W.: 1993, `Methane emissions from wood-feeding termites in Amazonia', Chemosphere 26(1-4): 623-632.

Mooney, H.A., Vitousek, P.M., and Matson, P.A.: 1987, `Exchange of materials between terrestrial ecosystems and the atmosphere', Science 238: 926-932.

Muzio, L.J. and Kramlich, J.C.: 1988, `An artifact in the measurement of N₂O from combustion sources', Geophys. Res. Letters 15: 1369-1372.

Nepstad, D.C., de Carvalho, C.R., Davidson, E.A., Jipp, P.H. Lefebvre, P.A., Negreiros, G.H. da Silva, E.D., Stone, T.A. Trumbore, S.E., and Vieira, S.: 1994, `The role of deep roots in the hydrological and carbon cycles of Amazonian forests and pastures', Nature 372: 666-669.

Nye, P.H. and Greenland, D.J.: 1960, The Soil Under Shifting Cultivation, Technical Comunication No. 51, Commonwealth Agricultural Bureaux of Soils, Harpenden, UK.

Post, W.M., Emanuel, W.R., Zinke, P.J., and Strangenberger, A.G.: 1982, `Soil carbon pools and world life zones', Nature 298: 156-159.

Rasmussen, R.A. and Khalil, M.A.K.: 1983, `Global production of methane by termites', Nature 301: 700-702.

Rasmussen, R.A. and Khalil, M.A.K.: 1988, `Isoprene over the Amazon Basin', J. Geophys. Res. 93: 1417-1421.

Revilla Cardenas, J.D., Kahn, F.L, and Guillaumet, J.L.: 1982, `Estimativa da Fitomassa do Reservatório da UHE de Tucuruí', in Brazil, Presidência da República, Ministério das Minas e Energia, Centrais Elétricas do Norte S.A. (ELETRONORTE) and Brazil, Secretaria do Planejamento, Conselho Nacional de Desenvolvimento Científico e Tecnológico, Instituto Nacional de Pesquisas da Amazônia (SEPLAN‑CNPq‑INPA), Projeto Tucuruí, Relatório Semestral Jan.‑Jun. 1982, Vol. 2: Limnologia, Macrófitas, Fitomassa, Degradação de Fitomassa, Doenças Endêmicas, Solos, INPA, Manaus, Amazonas, 1-11.

Richey, J.E., Brock, J.T., Naiman, R.J., Wissmar, R.C., and Stallard, R.R.: 1980, `Organic carbon: Oxidation and transport in the Amazon River', Science 207: 1348-1351.

Saldarriaga, J.G., West, D.C., and Tharp, M.L.: 1986, Forest Succession in the Upper Rio Negro of Colombia and Venezuela, Oak Ridge National Laboratory, Environmental Sciences Publication No. 2694, ORNL/TM-9712, National Technical Information Service, Springfield, VA.

Seiler, W. and Crutzen, P.J.: 1980, `Estimates of gross and net fluxes of carbon between the biosphere and the atmosphere from biomass burning', Clim. Change 2: 207-247.

Seiler, W., Conrad, R., and Scharffe, D.: 1984, `Field studies of methane emission from termite nests into the atmosphere and measurements of methane uptake by tropical soils', J. Atmos. Chem. 1: 171-186.

Shine, K.P., Derwent, R.G., Wuebbles, D.J., and Morcrette, J-J.: 1990, 'Radiative forcing of climate', in Houghton, J.T., Jenkins, G.J., and Ephraums, J.J. (eds.) Climate Change: The IPCC Scientific Assessment, Cambridge University Press, Cambridge, UK, 41-68.

Skole, D. and Tucker, C.: 1993. `Tropical deforestation and habitat fragmentation in the Amazon: Satellite data from 1978 to 1988', Science 260: 1905-1910.

Shukla, J., Nobre, C., and Sellers, P.: 1990, `Amazon deforestation and climate change', Science 247: 1322-1325.

Thompson, A.M. and Cicerone, R.J.: 1986, `Possible perturbations to atmospheric CO, CH₄, and OH', J. Geophys. Res. 91(D10): 10,853-10,864.

Trumbore, S.E., Davidson, E.A., de Camargo, P.B., Nepstad, D.C., and Martinelli, L.A.: 1995. `Belowground cycling of carbon in forests and pastures of Eastern Amazonia', Global Biogeochem. Cycles 9: 515-528.

Uhl, C. and Saldarriaga, J.: nd, `The disappearance of wood mass following slash and burn agriculture in the Venezuelan Amazon', (manuscript).

United Nations Educational Scientific and Cultural Programme (UNESCO)/United Nations Environment Programme (UNEP)/Food and Agricultural Organization of the United Nations (FAO): 1978, Tropical Forest Ecosystems: A State of Knowledge Report, UNESCO, Paris.

Veldkamp, E.: 1993, Soil Organic Carbon Dynamics in Pastures Established after Deforestation in the Humid Tropics of Costa Rica, Ph.D. dissertation, Agricultural University Wageningen, Wageningen, The Netherlands, 117 pp.

Vermeer, D.E.: 1970, `Population pressure and crop rotational changes among the Tiv of Nigeria', Annals of the Association of American Geographers 60: 299-314.

Ward, D.E.: 1986, `Field scale measurements of emission from open fires', technical paper presented at the Defense Nuclear Agency Global Effects Review, Defense Nuclear Agency, Washington, DC.

Ward, D.E. and Hardy, C.C.: 1984, `Advances in the characterization and control of emissions from prescribed fires', paper presented at the 77th annual meeting of the Air Pollution Control Association, San Francisco, CA.

Wood, T.G., Johnson, R.A., and Ohiagu, C.: 1977, `Populations of termites (Isoptera) in natural and agricultural ecosystems in southern Guinea savanna near Mokwa, Nigeria', in Malaisse, F., (ed.) Structure, Fonctionnement et Amenagement d'Ecosystemes Tropicaux (Geo-Eco-Trop Vol. 1, No. 2), Faculté des Sciences, Université Nationale du Zaïre, Lubumbashi, Zaïre, 139-148.

Zimmerman, P.R., Greenberg, J.P., and Darlington, J.P.E.C.: 1984, `Termites and atmospheric gas production', Science 224: 86.

Zimmerman, P.R., Greenberg, J.P., Wandiga, S.O., and Crutzen, P.J.: 1982, `Termites: A potentially large source of atmospheric methane, carbon dioxide, and molecular hydrogen', Science 218: 563-565.

FIGURE LEGENDS

Fig. 1.Brazil's Legal Amazon region, with locations mentioned in the text.

Fig. 2.Forest and nonforest in the Brazilian Legal Amazon (Source: Fearnside and Ferraz, 1995).

Fig. 3.Carbon transformations through a typical 10-year sequence of clearing, burning and reburning (updated from: Fearnside, 1991).

Fig. 4.Partioning of carbon between type of release and emitted gas (Low trace gas scenario) (updated from: Fearnside, 1991).

TABLE I: EXTENT OF DEFORESTATION IN THE BRAZILIAN LEGAL AMAZON^(a)


	Original	Deforested area					Deforested area
Political	forest	(km² X 10³)					(% of original forest area)
unit	area
	(km² X 10³)	Jan. 1978	Apr. 1988	Aug. 1989	Aug. 1990	Aug. 1991	Jan. 1978	Apr. 1988	Aug. 1989	Aug. 1990	Aug. 1991

Acre	152	2.6	8.9	9.8	10.3	10.7	1.7	5.8	6.4	6.8	7.0

Amapá	115	0.2	0.9	1.0	1.3	1.7	0.2	0.8	0.9	1.1	1.5

Amazonas	1,481	2.3	18.0	19.3	19.8	20.8	0.2	1.2	1.3	1.3	1.4

Maranhão	143	65.9^(b)	90.8^(b)	92.3^(b)	93.4^(b)	94.1^(b)	46.1	63.5	64.5	65.3	65.8

Mato Grosso	528	26.5	71.5	79.6	83.6	86.5	5.0	13.5	15.1	15.8	16.4


Pará	1,139	61.7^(b)	129.6^(b)	137.3^(b)	142.2^(b)	146.0^(b)	5.4	11.4	12.1	12.5	12.8

Rondônia	215	6.3	29.6	31.4	33.1	34.2	2.9	13.8	14.6	15.4	15.9

Roraima	164	0.2	2.8	3.6	3.8	4.2	0.1	1.7	2.2	2.3	2.6

Tocantins	59	4.2	21.6	22.3	22.9	23.4	7.1	36.7	37.9	38.9	39.7


Legal	3,996	169.9	373.9	396.6	410.4	421.6	4.3	9.4	9.9	10.3	10.5
Amazon

		FOREST FLOODED BY HYDROELECTRIC DAMS
Legal
Amazon		0.1	4.4	5.5	5.5	5.5	0.0	0.1	0.1	0.1	0.1

		DEFORESTATION FROM ALL SOURCES
Legal
Amazon		169.9	378.3	402.1	416.0	427.1	4.3	9.5	10.1	10.4	10.7

(a) Sources: Fearnside, 1993a,b
(b) Maranhão values include 57.8 X 10³ km², and Pará values include 39.8 X 10³ km², of
"old" (approximately pre‑1970) deforestation now largely under secondary forest.

TABLE IV: AREA OF NATURAL VEGETATION PRESENT IN THE BRAZILIAN LEGAL AMAZON^(a)

Cate‑	Code	Acre	Amapá	Amazonas	Maranhão	Mato	Pará	Rondônia	Roraima	Tocantins/	Total
gory						Grosso				Goiás	present

Dense	Da‑0		9,011	164,876			76,570	2,704	3,326	2,610	259,097
forest	Db‑0	16,408	2,184	615,203	22,586		164,091	2,066	10,248		832,786
	Dm‑0		113	10,181			3,418		20,661		34,373
	Ds‑0	518	99,220	178,103	1,988	23,154	413,345	14,607	83,692	3,055	817,682

	Subtotal	16,926	110,528	968,363	24,574	23,154	657,424	19,377	117,927	5,665	1,943,938

Non‑	Aa‑0	10,591		65,748			805	2,273			79,417
dense	Ab‑0	114,380		211,052				41,064			366,496
forest	As‑0			37,555		124,620	286,271	77,794	8,430	1,216	535,886
	Cs‑0				3,666	736	5,386			115	9,903
	Fa‑0					3,554					3,554
	Fs‑0					24,317		7,718	1,041	1,328	34,404
	La‑0			14,979					970		15,949
	Ld‑0			37,405					10,967		48,372
	Lg‑0			9,663					9,767		19,430
	LO‑0			172,607					30,184		202,791
	ON‑0					168,069	2,991	4,801	3,045		178,906
	Pf‑0		1,823		2,089		3,894				7,806
	SM‑0				384						384
	SN‑0			1,082	6,570	142,778	27,812	4,781	904	14,465	198,392
	SO‑0		4,226	27,350		22,124	59,734	21,932	4,286	6,551	146,203

	Subtotal	124,971	6,049	577,441	12,709	486,198	386,893	160,363	69,594	23,675	1,847,893

	Subtotal	141,897	116,577	1,545,804	37,283	509,352	1,044,317	179,740	187,521	29,340	3,791,831
	all forests

Non‑	Pa‑0		15,157	12,778	2,517	14,738	27,162	8,690			81,042
forest	rm‑0								390		390
	Sa‑0			1,531	55,758	167,534	5,686	11,028		102,445	343,982
	Sd‑0				15,771	10,840	1,274			2,234	30,119
	Sg‑0					10,490	5,057		15,481	7,113	38,141
	Sp‑0		10,038	5,556	26,980	64,085	12,393	2,664	8,969	48,962	179,647
	ST‑0					6,599					6,599
	Td‑3								1,550		1,550
	Tp‑3								10,671		10,671

	Subtotal	0	25,195	19,865	101,026	274,286	51,572	22,382	37,061	160,754	692,141

	Total	141,897	141,772	1,565,669	138,309	783,638	1,095,889	202,122	224,582	190,094	4,483,972
(a) Areas in km² measured from 1:5,000,000 vegetation map (Brazil, IBGE and IBDF, 1988).
These areas do not reflect losses due to recent deforestation.

TABLE V: FOREST BIOMASS PER HECTARE: MEANS BY ECOREGION, VEGETATION TYPE AND STATE (t/ha)^(a)

Cate‑	Code	Acre	Amapá	Amazonas	Maranhão	Mato	Pará	Rondônia	Roraima	Tocantins/	Mean in	Area‑
gory						Grosso				Goiás	sampled	weighted
											plots	mean
Dense	Da‑0		471	513			412	310	419	500	498	478
forest	Db‑0	438	583	495	458		585	546	417		551	510
	Dm‑0		441	344			441		448		440	417
	Ds‑0	391	613	478	464	402	518	464	429	111	430	506

	Dense	437	601	493	459	402	522	452	431	290	519	502
	forests

Non‑	Aa‑0	425		435			536	431			430	434
dense	Ab‑0	455		466				403			455	456
forest	As‑0			507		426	361	378	376	376	374	389
	Cs‑0				384	384	384			384	384	384
	Fa‑0					370					370	370
	Fs‑0					403		472	424	424	423	420
	La‑0			465					465			435
	Ld‑0			437					437			435
	Lg‑0			437					437			435
	LO‑0			500					437		499	490
	ON‑0					383	401	547	409		401	389
	Pf‑0		437		464		437					435
	SM‑0				437							435
	SN‑0			409	385	383	479	385	336	385	385	397
	SO‑0		404	582		360	410	404	404	404	402	433


Non‑dense		453	414	478	399	394	379	399	425	392	435	423
forests

All		451	591	488	438	394	469	404	429	372	500	464
forests

(a) Values in italics are for ecoregions where no sample exists: values are based on the mean in
sampled plots for the same vegetation type in other states. Source: Fearnside, nd‑a.

TABLE VI: APPROXIMATE BIOMASS CLEARED IN 1990 IN EACH ECOREGION IN THE BRAZILIAN LEGAL AMAZON (10³ t/year)

Cate‑	Code	Acre	Amapá	Amazonas	Maranhão	Mato	Pará	Rondônia	Roraima	Tocantins/	Total
gory						Grosso				Goiás

Dense	Da‑0		944	2,924			15,013	733	125	2,600	22,339
forest	Db‑0	2,846	293	10,492	29,196		44,014	1,106	384		88,331
	Dm‑0		11	77			717		807		1,613
	Ds‑0	80	13,993	2,981	2,978	7,391	99,957	6,400	3,015	689	137,484

	subtotal	2,926	15,242	16,474	32,175	7,391	159,702	8,238	4,331	3,288	249,767

Non‑	Aa‑0	1,780		1,020			205	961			3,966
dense	Ab‑0	20,448		3,472				16,183			40,103
forest	As‑0			669		42,163	49,222	27,044	284	931	120,313
	Cs‑0				4,551	225	984			90	5,850
	Fa‑0					1,044					1,044
	Fs‑0					7,786		3,576	40	777	12,178
	La‑0			239					40		279
	Ld‑0			577					411		989
	Lg‑0			151					383		534
	LO‑0			2,815					1,121		3,936
	ON‑0					51,195	570	2,574	112		54,451
	Pf‑0		28		3,129		810				3,967
	SM‑0				543						543
	SN‑0			16	8,173	42,641	6,337	1,805	27	11,348	70,347
	SO‑0		319	569		6,334	11,647	7,506	155	5,395	31,926

	Subtotal	22,228	347	9,528	16,396	151,387	69,778	59,648	2,573	18,541	350,426

	Subtotal	25,155	15,589	26,002	48,571	158,779	229,480	67,886	6,904	21,829	600,193
	all forests


Average	Dense
biomass/	forests	437	601	494	459	402	522	453	431	288	500
ha
cleared
	non‑dense	453	407	478	399	394	379	399	425	391	397
	forests

	All	451	595	488	437	394	468	405	429	371	434
	forests

TABLE VII: ANNUAL PROBABILITIES OF TRANSFER

Initial state	Subsequent state
	Regener‑	Farm‑	Produc‑	Degraded	Secondary	Secondary
	ated	land	tive	pasture	forest	forest
	Forest		pasture		from	from
					farmland	pasture

Regener‑
ated	0	0.347	0.653	0	0	0
Forest


Farmland	0	0.450	0.468	0	0.082	0



Productive	0	0	0.849	0.008	0	0.143
pasture


Degraded	0	0	0.007	0.926	0	0.067
pasture


Secondary	5.17 X 10^-10	0.065	0.128	0	0.807	0
forest from
farmland

Secondary
forest from	7.16 X 10^‑8	0.061	0.101	0	0	0.838
pasture

TABLE VIII: REPLACEMENT VEGETATION WEIGHTED BIOMASS CALCULATION AT EQUILIBRIUM

Category	Equilib‑	Biomass	Resid‑	Biomass
	rium	(t/ha	ence	source
	propor‑	total)	time
	tion		(years)
Forest	0.000	464	0.0	(a)


Farmland	0.040	0.7	0.9	(b)


Productive	0.438	10.7	4.2	(c)
Pasture

Degraded	0.052	8.0	9.1	(d)
pasture

Secondary	0.020	35.6	3.2	(d)
forest
from agriculture

Secondary	0.449	50.5	3.9	(d)
forest
from pasture


Weighted mean:		28.5

(a)Secondary forest is assumed to be equivalent to original forest from the standpoint of biomass after 100 years. Saldarriaga et al. (1986: 96) calculated recovery in 144‑189 years in Venezuela. Original forest biomass from Fearnside (nd‑a).

(b) Guess: above‑ground biomass=0.5 t/ha; root/shoot ratio=0.3.

(d) Calculated from the residence time and growth rate (Fearnside and Guimaraes, 1996).

TABLE IX: PARAMETERS FOR TRANSFORMATIONS OF CARBON STOCKS


Parameter	Value	Units	Source	Comment


Total biomass	407.0	t/ha dry weight	Fearnside, nd‑b	Weighted mean for areas being cleared in 1990 (after logging).
Carbon content of forest biomass	0.50	fraction of dry weight	Measured at Manaus by Fearnside et al., 1993	Many authors use a value of 0.45, but without reference to experimental data for its origin. Brown and Lugo (1984) use a value of 0.50.


Above ground fraction	0.759		Calculated from Fearnside, nd‑b		Total above‑ground biomass/total biomass of 1990 areas at time of clearing.

Burning efficiency in initial burn	0.332	fraction of C released	Fearnside et al., 1993, nd‑a,b adjusted for logging by Fearnside, nd-b	For areas cleared after preparatory logging.

Char C fraction in initial burn	0.019		Fearnside et al., 1993, nd‑a,b		Average of two studies near Manaus, Amazonas and one in Altamira, Pará.

Fraction of char on biomass following initial burn	0.89		Fearnside et al., nd‑a	Near Altamira, Pará.

Exposed to soil char C transfer fraction during first interval	0.3		Guess	First interval = 5 years.

Fraction surviving decay in first interval	0.400		Calculated from Uhl and Saldarriaga, nd and Buschbacher, 1984^(a)

Burning efficiency in first reburn	0.201	fraction of C released	Fearnside, et al., nd‑d; Barbosa, 1994	Burns of secondary forest and pasture in Apiaú, Roraima

Fraction converted to char in first reburn	0.010		Fearnside, et al., nd‑d; Barbosa, 1994		Burns in Apiaú, Roraima (NB: includes charcoal from secondary forest, but not secondary forest biomass)

Char C combustion fraction in first reburn	0		Assumed zero because char conversion value is net

Fraction surviving decay in second interval	0.543		Calculated from Uhl and Saldarriaga, nd^(a)		Second interval = 3 years

Burning efficiency in second reburn	0.201	fraction of C released	Assumed equal to first reburn

Fraction of C converted to char in second reburn	0.010		Assumed equal to first reburn

Fraction of char on biomass after first reburn	0.89		Assumed equal to initial burn

Exposed to soil char C transfer fraction during second interval	0.3		Guess

Char C combusted fraction in second reburn	0		Assumed zero because char conversion value is net

Fraction of char on biomass after second reburn	0.89		Assumed equal to initial burn

Exposed to soil char C transfer fraction during third interval	0.3		Guess

Fraction surviving decay in third interval	0.767		Calculated from Uhl and Saldarriaga, nd.		Third interval = 3 years

Burning efficiency in third reburn	0.201	fraction of wood C released	Assumed equal to first reburn

Fraction of C converted to char in third reburn	0.010		Assumed equal to first reburn

Char C combustion fraction in third reburn	0		Assumed zero because char conversion value is net

Soil C release from top 20 cm	3.92	t/ha	Fearnside, 1985, 1987

Replacement vegetation biomass	28.5	t/ha	Table VIII.	Weighted average for equilibrium landscape

Carbon content of replacement vegetation biomass	0.45	fraction of dry weight	Based on measurements at Altamira by Guimarães, 1993.

(a) See Fearnside, in press-b

TABLE X: PARAMETERS FOR CARBON EMISSIONS

Scenario

Component

Transformation

Value (t C released in this form / t C present in component)

t gas released/t fuel burned

Basis and reference

Both high and low trace gas scenarios

Above-ground biomass carbon

Combustion release

0.4203

Calculated from parameters in Table IX and Figure 3

Decay release

0.5561

Calculated from parameters in Table IX and Figure 3

Charcoal carbon formation (initial+subsequent burns)

0.0237

Calculated from parameters in Table IX and Figure 3

Carbon released through combustion

Initial burn

0.7905

Calculated from parameters in Table IX and Figure 3

Reburns

0.2095

Calculated from parameters in Table IX and Figure 3

Combustion release of below-ground biomass

Assumption

Carbon released through decay

Decay release through termites (above ground)

0.0297

Martius et al., nd (average for first 10 years for original forest biomass; emission calculations done with a different proportion for each year)

Decay release through other decay (above ground)

0.9703

One minus decay release through termites

Decay release of below-ground biomass

Assumption

Decay release through termites (below ground)

Assumption (unrealistically low)

Decay release through other decay (below ground)

Assumption (unrealistically high)

Low trace gas scenario

Carbon released by combustion in initial burn

CH₄ carbon

0.0078

0.005

Kaufman et al., 1990 from Ward, 1986^(a)

CO₂ carbon

0.8779

1.55

Kaufman et al., 1990 from Ward, 1986^(a)

CO carbon

0.1068

0.12

Kaufman et al., 1990 from Ward, 1986^(a)

Carbon released by combustion in reburns

CH₄ carbon

0.0109

0.007

Kaufman et al., 1990 from Ward, 1986^(a)

CO₂ carbon

0.7930

1.4

Kaufman et al., 1990 from Ward, 1986^(a)

CO carbon

0.1958

0.22

Kaufman et al., 1990 from Ward, 1986^(a)

Carbon released through termites

CH₄ carbon

0.0022

0.001

Martius et al., 1993, nd

CO₂ carbon

0.9978

1.996

Assumed all C not released as methane is CO₂.

High trace gas scenario

Carbon released by combustion in initial burn

CH₄ carbon

0.0093

0.006

Kaufman et al., 1990 from Ward, 1986^(a)

CO₂ carbon

0.8779

1.55

Kaufman et al., 1990 from Ward, 1986^(a)

CO carbon

0.1335

0.15

Kaufman et al., 1990 from Crutzen, et al., 1985^(a)

Carbon released by combustion in reburns

CH₄ carbon

0.0171

0.011

Kaufman et al., 1990 from Greenberg et al., 1984^(a)

CO₂ carbon

0.7930

1.4

Kaufman et al., 1990 from Ward, 1986^(a)

CO carbon

0.2492

0.28

Kaufman et al., 1990 from Greenberg et al., 1984 and Ward, 1986^(a)

Carbon released through termites

CH₄ carbon

0.0079

0.005

Goreau and de Mello, 1987

CO₂ carbon

0.9978

Assumed all C not released as methane is CO₂

(a) Calculated from data presented by Kaufman et al. (1990: 382) as emission/kg fuel.

Carbon content of the experimental fuel is assumed to be 48.2%, this value

being chosen such that all combusted carbon is accounted for, either as CO₂, CH₄, CO,

NMHC or particulate graphitic carbon.

TABLE XI: PARAMETERS FOR OTHER SOURCES OF GREENHOUSE GASES FROM LAND‑USE CHANGE

Factor	Units	Value	Reference	Note

Soil carbon from top 20 cm	t C/ha	3.92	Fearnside, 1985	(a)

Cerrado biomass carbon	t C/ha	17.27	Fearnside, nd‑a	(b)

Cattle CH₄	kg CH₄/head/year	55	Ahuja, 1990, based on Crutzen et al., 1986

Cattle stocking rate	head/ha	0.3	Fearnside, 1979	(c)

Pasture soil N₂O	kg N₂O/ha/year	3.8	Luizão et al., 1989	(d)


(a) For conversion to pasture at Paragominas, based on Falesi (1976: 31 and 42) for carbon contents
and Hecht (1981: 95) for soil densities.

(b) Based on conversion to pasture (total biomass 10.7 t/ha) of cerrado with average total biomass
of 45 t/ha.

(c) Feeding capacity after 3 years.

(d) Full annual cycle under pasture and forest at Manaus.

TABLE XII: TRACE GAS PARAMETERS


Factor	Gases	Units	Value	Source

Intact forest soil sink	CH₄	t C/ha/yr	‑0.0004	Keller et al., 1986

Burning release	N₂O^(a)(b)	t gas/t CO₂ emitted from burn	0.0002	Cofer et al., 1988 cited by Kaufman et al., 1990

Burning release	N₂O^(c)(d)	t gas/t C burned	0.0017	Calculated by Keller et al., 1991: 146 from Andreae et al., 1988

Burning release	NO_x^(e)	t gas/t C burned	0.0079	Keller et al., 1991: 146

Intact forest release	NO_x^(e)	t gas/ha/yr	0.0131	Kaplan et al., 1988; see Keller et al., 1991
Flaming burn release	Total particulates	t/t CH₄ gas from burn	3.33	Calculated by Kaufman et al., 1990: 380 from
				Ward and Hardy, 1984 and Ward, 1986

Smoldering burn release	Total particulates	t/t CH₄ gas from burn	1.67	Calculated by Kaufman et al., 1990: 380 from Ward and Hardy, 1984 and Ward, 1986

Flaming burn release	NMHC^(b)	t/t CH₄ gas from burn	0.67	Derived using factor of 0.2 t NMHC/t particulates calculated by Kaufman et al., 1990: 380

Smoldering burn release	NMHC^(b)	t/t CH₄ gas from burn	0.50	Derived using factor of 0.3 t CH₄/t particulates calculated by Kaufman et al., 1990: 380

Mixed burn release	NMHC^(d)(f)	t/t C burned	0.0131	Keller et al., 1991: 146 from measurements of Andreae et al., 1988

Intact forest release	NMHC	t gas/ha/yr	0.12	Rasmussen and Khalil, 1988: 1420
(a) Intact forest release accounted for in pasture soil calculation.

(b) Used in low trace gas scenario.

(c) results in 0.109 t gas/ha burned, or three times the 0.041 t gas/t C burned obtained using the parameter relating N₂O to CO₂.

(d) Used in high trace gas scenario.

(e) NO_x weight given in NO₂ basis (following Shine et al., 1990: 61).

(f) NMHC emission corresponds to 0.85 t gas/ha burned, much higher than values derived from methane, which are (for high and low trace gas scenarios, respectively): 0.36 and 0.43 t NMHC/ha burned for flaming combustion and 0.10 and 0.16 t NMHC/ha burned for smoldering combustion.

TABLE XIII: SOIL CARBON PARAMETERS AND CALCULATIONS

Parameters

Units

Value

Source

Soil density in forest

g/cm³

0.56

Hecht, 1981: 95

Carbon in forest soil

% by wt

0.91

Falesi, 1976: 31 and 42

Carbon in pasture soil

% by wt

0.56

Falesi, 1976: 31 and 42

Top 20 cm C as fraction of 1 m C

% by wt

Fearnside, 1987

Calculated values

Top 20 cm of soil:

Soil dry weight

t/ha

1,120

Carbon in forest soil

t/ha

10.19

Carbon in pasture soil compacted from top 20 cm of forest soil

Release from top 20 cm

t/ha

3.92

Release fraction of pre‑conversion soil C

% by wt

Top meter of soil:

Soil dry weight

t/ha

5,600

Carbon in forest soil

t/ha

24.27

Carbon in pasture soil

t/ha

14.93

Release from top meter

t/ha

9.33

Release fraction of pre‑conversion soil C

% by wt

TABLE XIV: NET COMMITTED GREENHOUSE GAS EMISSIONS BY SOURCE FOR 1990 CLEARING IN THE LEGAL AMAZON: LOW TRACE GAS SCENARIO

Source			Area	Emissions (10⁶ t of gas)
			affected
			(10³ km²)
				CO₂	CH₄	CO	N₂O	NO_x	NMHC
FOREST
	Initial burn		13.8	228	0.74	17.66	0.05	0.56	0.49

	Reburns		13.8	55	0.27	8.58	0.01	0.15	0.14

	Termites above‑ground decay		13.8	13	0.010

	Other above‑ground decay		13.8	422

	Below‑ground decay		13.8	249

	Cattle^(a)		6.1		0.010

	Pasture soil^(a)		6.1				0.002

	Loss of intact forest		7.3		0.0003			‑0.01	‑0.09
	sources and sinks^(a)

	Soil carbon (top 20 cm)		13.8	20

	Regrowth		13.8	‑65

	Hydroelectric^(a)		0.0		0.00


		Forest subtotal		921	1.03	26.25	0.06	0.70	0.54



CERRADO
	Initial burn		5.0	17	0.06	1.35	0.003	0.04	0.04

	Reburns		5.0	2	0.01	0.26	0.0003	0.005	0.004

	Termites above‑ground decay		5.0	0.1	0.0001

	Other above‑ground decay		5.0	4

	Below‑ground decay		5.0	15

	Cattle^(a)		5.0		0.008

	Pasture soil^(a)		5.0				0.002

	Loss of intact cerrado		5.0		0.0002			‑0.0004	‑0.004
	sources and sinks^(a)(b)

	Soil carbon (top 20 cm)		5.0	7

	Regrowth		5.0	‑9


		Cerrado subtotal		36	0.07	1.62	0.01	0.05	0.04



TOTAL FOR LEGAL AMAZON				958	1.10	27.86	0.06	0.74	0.58

(a) Recurring effects (cattle methane, forest soil methane sink, pasture soil N₂0, hydroelectric methane) summed for 100‑year period for consistency with IPCC 100‑year horizon calculation.

(b) Intact cerrado source for NO_x and NMHC derived from the forest per‑hectare emission assuming emission is proportional to the tree leaf dry weight biomass in each ecosystem. Cerrado tree leaf biomass (dry season) = 0.756 t/ha (dos Santos, 1989: 194). Forest (at Tucuruí, Pará) = 12.94 t/ha (Revilla Cardenas et al., 1982: 6).

TABLE XV: NET COMMITTED GREENHOUSE GAS EMISSIONS BY SOURCE FOR 1990 CLEARING IN THE LEGAL AMAZON: HIGH TRACE GAS SCENARIO

Source			Area	Emissions (10⁶ t of gas)
			affected
			(10³ km²)

				CO₂	CH₄	CO	N₂O	NO_x	NMHC
FOREST
	Initial burn		13.8	228	0.88	22.08	0.12	0.56	0.93

	Reburns		13.8	55	0.43	10.93	0.03	0.15	0.25

	Termites above‑ground decay		13.8	13	0.01

	Other above‑ground decay		13.8	422
	Below‑ground decay		13.8	249
	Cattle^(a)		6.1		0.01

	Pasture soil^(a)		6.1				0.002

	Loss of intact		7.3		0.0003			‑0.01	‑0.09
	forest sources^(a)

	Soil C stock		13.8	20

	Regrowth		13.8	‑65

	Hydroelectric^(a)		0.0		0.00


		Forest subtotal		921	1.33	33.00	0.15	0.70	1.08



CERRADO
	Initial burn		5.0	17	0.07	1.69	0.009	0.043	0.07

	Reburns		5.0	2	0.01	0.34	0.001	0.005	0.01

	Termites above‑ground decay		5.0	0.1	0.0001

	Other above‑ground decay		5.0	4
	Below‑ground decay		5.0	15
	Cattle^(a)		5.0		0.01

	Pasture soil^(a)		5.0				0.002

	Loss of intact		5.0		0.0002			‑0.0004	‑0.004
	cerrado sources^(a)(b)

	Soil C stock		5.0	7

	Regrowth		5.0	‑9


		Cerrado subtotal		36	0.09	2.03	0.012	0.05	0.07



TOTAL FOR LEGAL AMAZON				958	1.42	35.03	0.165	‑0.74	1.16

(a) Recurring effects (cattle methane, forest soil methane sink, pasture soil N20, hydroelectric methane) summed for 100‑year period for consistency with IPCC 100‑year horizon calculation.
(b) Intact cerrado source for NO_x and NMHC derived from the forest per‑hectare emission assuming emission is proportional to the tree leaf dry weight biomass in each ecosystem. Cerrado tree leaf biomass (dry season) = 0.756 t/ha (dos Santos, 1989: 194). Forest (at Tucuruí, Pará) = 12.94 t/ha (Revilla Cardenas et al., 1982: 6).

TABLE II: RATE OF DEFORESTATION IN THE BRAZILIAN LEGAL AMAZON^(a)


Political	Deforestation rate (km² X 10³/year)
unit
	1978‑1988	1988‑1989	1989-1990	1990‑1991

	DEFORESTATION EXCLUSIVE OF HYDROELECTRIC DAMS


Acre	0.6	0.6	0.6	0.4

Amapá	0.1	0.2	0.3	0.4

Amazonas	1.6	1.3	0.5	1.0

Maranhão	2.5	1.4	1.1	0.7

Mato Grosso	4.5	6.0	4.0	2.8

Pará	6.8	5.8	4.9	3.8

Rondônia	2.1	1.4	1.7	1.1

Roraima	0.2	0.7	0.2	0.4

Tocantins	1.6	0.7	0.6	0.4

Legal	20.0	18.1	13.8	11.1
Amazon
	FOREST FLOODED BY HYDROELECTRIC DAMS
Legal
Amazon	0.4	1.1	0.0	0.0

	DEFORESTATION FROM ALL SOURCES
Legal
Amazon	20.4	19.2	13.8	11.1

(a) Sources: Fearnside, 1993a,b

TABLE XVI: NET COMMITTED EMISSIONS FROM 1990 DEFORESTATION WITH CO₂ EQUIVALENT, 100‑YEAR TIME HORIZON


		Low trace gas scenario						High trace gas scenario						Contribution of each gas to total effect (%)

Gas	Global warming potential^(a)	Amount emitted (10⁶ t of gas/year)			CO₂ equivalent (10⁶ t of gas/year)			Amount emitted (10⁶ t of gas/year)			CO₂ equivalent (10⁶ t of gas/year)			Low trace gas scenario	High trace gas scenario


		Forest	Cerrado	Total	Forest	Cerrado	Total	Forest	Cerrado	Total	Forest	Cerrado	Total

CO₂	1	921.23	36.41	957.41	921.23	36.41	957.41	921.23	36.41	957.64	921.23	36.41	957.64	95.26	91.63
CH₄	24.5	1.03	0.07	1.10	25.23	1.80	27.02	1.33	0.09	1.42	32.66	2.19	34.85	2.69	3.33
CO	0	26.25	1.62	27.86	0.00	0.00	0.00	33.00	2.03	35.03	0.00	0.00	0.00	0.00	0.00
N₂O	320	0.06	0.01	0.06	18.83	1.83	20.67	0.15	0.01	0.16	48.83	3.83	52.66	2.06	5.04
NO_x	0	0.70	0.05	0.74	0.00	0.00	0.00	0.70	0.05	0.74	0.00	0.00	0.00	0.00	0.00
NMHC	0	0.54	0.04	0.58	0.00	0.00	0.00	1.08	0.07	1.16	0.00	0.00	0.00	0.00	0.00


Total CO₂‑equivalent gas (10⁶ t)					965	40	1,005				1,003	42	1,045	100.0	100.0

C0₂‑equivalent carbon (10⁶ t)					263	11	274				273	12	285

(a) IPCC 100‑year values expressed as kg of CO₂ gas equivalent/kg of gas. The global warming potentials are from Albritton et al., 1995: 222.