CHAPTER 21 CHAPTER 21 Global Effects of Deforestation
21.3 EFFECTS OF DEFORESTATION ON SOURCES AND SINKS
21.3.1 CARBON
21.3.1.1 Carbon Stored in Vegetation and Soils
The amount of carbon stored in the living plants of the earth is of the same order as the amount held in the earth’s atmosphere. In the eighteenth century, 700–800 Pg (1 Pg = 1 × 1015 g) carbon were held in the earth’s vegetation, and about 600 PgC were in the atmosphere. Today, trees, grasses, and herbs are believed to hold about 500 PgC, and the atmosphere about 760 Pg. The increase in atmospheric CO2 is accurately known; the decrease in terrestrial biomass is not as well known.
The amount of organic carbon stored in the soils of the earth has also decreased (by about 3%) over the last 150 years as a result of cultivation and remains at 1100–1500 PgC. Terrestrial ecosystems (including both the living plants and soils) hold almost three times as much carbon as the atmosphere. Most of this terrestrial carbon is stored in forests. Forests cover about 30% of the land surface and hold almost half of the world’s terrestrial carbon. If only vegetation is considered (soils ignored), forests hold about 75% of the living carbon.
In area, tropical forests account for slightly less than half of the world’s forests, yet they hold about as much carbon in their vegetation and soils as temperate-zone and boreal forests combined.
Considering only vegetation, undisturbed tropical forests hold, on average, about 65% more carbon per unit area than forests outside the tropics. When the carbon in soils is included, temperate zone and tropical forests are more nearly equal. Nevertheless, equivalent rates of deforestation in the two regions will cause considerably more carbon to be released from the tropics than from regions outside the tropics because only a fraction of the soil carbon is lost with cultivation.
The biomass of temperate zone and boreal forests is reasonably well known because foresters have been conducting forest inventories in northern countries for decades. Such inventories are rare in tropical countries, and the distribution of biomass throughout the tropics is poorly known. A recent comparison in the Brazilian Amazon showed that estimates of biomass for the entire region varied by more than a factor of two.8 Uncertainties result from limited data on belowground biomass and variability in the biomass of trees smaller than those routinely sampled, in vines, in nontree components, in palms, in the shape and density of tree boles, and in the amount of woody debris on the forest floor. There is also considerable uncertainty in the spatial distribution of biomass over large areas. None of the five estimates compared in the Brazilian Amazon showed similar distri- butions of high- and low-biomass forests.8
Per unit area, forests hold 20 to 50 times more carbon in trees than the ecosystems that generally replace them (Table 21.2), and this carbon is released to the atmosphere as forests are transformed to other uses. In the tropics, much of the carbon is released immediately through burning. After- wards, decay of soil organic matter and woody debris continues to release carbon to the atmosphere, but at lower rates. If croplands are abandoned, regrowth of live vegetation and redevelopment of soil organic matter withdraw carbon from the atmosphere and accumulate it again on land. To calculate the net flux of carbon from deforestation and reforestation, ecologists have documented the changes in carbon associated with different types of land use and different types of ecosystems in different regions of the world. Annual changes in the different reservoirs of carbon (live vege- tation, soils, debris, and wood products) determine the annual net flux of carbon between the land and the atmosphere. Because of the variety of ecosystems and land uses, and because the calculations require accounting for cohorts of different ages, bookkeeping models have been developed for the calculations.9–13
Table 21.2 compares the relative losses of carbon as a result of converting forests to other uses.
The losses in biomass range from 100% for permanently cleared land to 0% for nondestructive
harvest of fruits, nuts, and latex (extractive reserves). Losses of carbon from soil may also occur, especially if the soils are cultivated.
21.3.1.2 The Global Extent of Deforestation
21.3.1.2.1 Temperate and Boreal Forests
According to the latest Forest Resources Assessment published by the Food and Agriculture Organization (FAO) of the United Nations,14 the area of forest in developed countries (largely outside the tropics) increased by an average of 1.76 × 106 ha/year during the first half of the 1990s.
This increase does not include the Russian Federation because the FAO could not determine a reliable estimate. Reported increases in forest area of 2.4 × 106 ha between 1988 and 1993 are largely a result of reclassifications and definition: forest lands in Russia are defined as lands that trees can grow on and that are not used for other purposes. One estimate finds that the area actually covered by closed canopy forests (more than 30–40% canopy cover) declined by 7.6 × 106 ha between 1988 and 1993.15 However, in the 10 years between 1983 and 1993, only about 30% of the forests in Siberia were inventoried, and few of these areas were inventoried twice, so estimates of change are unreliable for the country with the largest share of the world’s forests.
The average annual change of 1.76 × 106 ha included increases of 0.39, 0.76, 0.05, and 0.56 × 106 ha/year in the developed countries of Europe, North America, Oceania, and countries of the former Soviet Union (not including Russia), respectively, during the period 1990–1995.14 In China, forest area experienced a net loss of 0.087 × 106 ha/year. Despite this net loss of forests, China has established the largest area of forest plantations: 34 × 106 ha by 1995. As mentioned above, it is not clear whether changes in Russian forests would have increased or decreased the average increase of 1.76 × 106 ha/year in developed (largely temperate-zone and boreal) countries.
Table 21.2 Percent of Initial Carbon Stocks Lost to the Atmosphere when Tropical Forests are Converted to Different Kinds of Land Use
Land Use
Carbon Lost to the Atmosphere Expressed as
% of Initial Carbon Stocks Vegetation Soil
Cultivated land 90–100 25
Pasture 90–100 12
Degraded croplands and pasturesa 60–90 12–25
Shifting cultivation 60 10
Degraded forests 25–50 <10
Loggingb 10–50 <10
Plantationsc 30–50 <10
Extractive reserves 0 0
Note: For soils, the stocks are to a depth of 1 m. The loss of carbon may occur within 1 year, with burning, or over 100 years or more, with some wood products. Values are from Houghton et al.9 unless otherwise indicated.
aCroplands and pastures, abandoned because of reduced fertility, may accumulate carbon, but their stocks remain lower than the initial forests.
bBased on current estimates of aboveground biomass in undisturbed and logged tropical forests.61 When logged forests are colonized by settlers, the losses are equivalent to those associated with one of the agricultural uses of land.
cPlantations may hold as much or more carbon than natural forests, but a managed plantation will hold, on average, 1/3 to 1/2 as much carbon as an undisturbed forest because it is generally regrowing from harvest.62
21.3.1.2.2 Tropical Forests
In contrast to the modest increase in developed countries, the average rate of deforestation in developing (largely tropical) countries was 13.7 × 106 ha/year during the first half of the 1990s (Table 21.3), down somewhat from the 15.5 × 106 ha/year rate for the 1980s. The recent rate is equivalent to clearing an area about the size of Georgia or Wisconsin each year. The highest rates (in 106 ha/yr) were reported in Brazil (2.554), Indonesia (1.084), Zaire (0.740), Bolivia (0.581), Mexico (0.508), and Venezuela (0.503). For all developing countries, the annual rate of forest loss was about 0.65% of forest area. Relative rates of loss are somewhat smaller in tropical Latin America (0.62%/year) and largest in tropical Asia (0.98%/year). The rates of deforestation in developing countries were offset to some extent by the establishment of forest plantations. Tropical countries reporting the largest areas of plantations in 1995 were India, Indonesia, and Brazil, with 15, 6, and 5 × 106 ha, respectively. Natural forests and plantations are not readily distinguished in developed countries, and so were combined in the discussion of developed countries above.
The errors in the rates of deforestation reported by the FAO are unknown. The most recent assessment by the FAO14 revised the 1990 estimate of deforestation from 16.3 to 15.5 × 106 ha/year (5%) as a result of new estimates of forest cover in 1980 and 1990. At the other extreme, preliminary national communications from Bolivia and Zimbabwe reported rates of deforestation six times less than reported by the FAO.16 Other countries were more similar to, or used, the FAO estimate.
Mexico reported credible rates that varied between 0.370 and 0.858 × 106 ha/year, a range that is approximately 80% of the mean. It is possible to lower the uncertainty in average rates of defor- estation with the use of high-spatial-resolution data from satellites, such as Landsat. However, two estimates of the total area deforested in the Brazilian Amazon, both based on data from Landsat, differed by 25%.17 The reasons for the difference have not been fully resolved.
One fate of land in the tropics bears special attention from the perspective of deforestation.
Most deforestation is thought to be for agriculture. The expansion of human settlements involves relatively little area, and logged areas, if the logging has not been too destructive and if the area is not colonized by farmers, generally return to forests. However, despite the fact that most deforestation is for some form of agriculture, the annual net expansion of agricultural lands is considerably less than the annual net reduction in forest area.18 For the entire tropics, for example, the expansion of croplands accounted for only 27% of total deforestation. Adding the increase in pasture area accounted for an additional 18% of deforestation. Fully 55% of the deforestation between 1980 and 1985 was explained by an increase in “other land.”19 Although some of this
“other land” is urban land, roads, and other settled lands, these uses are unlikely to have accounted for more than a few percent of the area deforested. Most of the increase in “other land” seems likely to be abandoned, degraded croplands and pastures, lands that no longer support crop or livestock production but that do not revert readily to forest, either.
Forests are not converted directly to degraded areas, of course. The transformation of land is from forest to agriculture and, subsequently, to degraded land. The important point is that only about one half of the area of tropical forest lost each year actually expands the area in agriculture.
The other half is only temporarily useful. After a few years it is lost, neither agriculturally productive nor forested. If these estimates are correct, making agriculture sustainable, so that agricultural lands
Table 21.3 Average Annual Rates of Deforestation (106 ha yr–1) in Developing (Largely Tropical) Regions
1980–1990 1990–1995
Africa 4.28 3.75
Asia-Oceania 4.41 4.17
Latin America and Caribbean 6.77 5.81
Developing world 15.46 13.73
Source: From FAO, State of the World’s Forests 1997, FAO, Rome, 1997.
can be farmed continuously, may be as effective in halting deforestation as increasing the yields of crops.
The fraction of deforestation used to expand the area in agriculture, as opposed to replacing worn-out land, varies among tropical regions. In Africa, the expansion of croplands accounted for only about 12% of the net area deforested. Eighty-eight percent of the decrease in forest area was matched by the expansion of “other land.” In tropical Asia, 40% of the net reduction in forests appeared as an expansion of agricultural lands. In Latin America, about two thirds of the reduction in forests could be accounted for by the expansion of croplands and pastures. If agriculture could be made sustainable throughout the tropics, rates of deforestation could be reduced by about 50%
without reducing the expansion of agricultural lands, and large areas of marginal or degraded lands might be reforested
21.3.1.3 A Broader Definition of Deforestation: Land-Use Change
Because this review is concerned with the effects of forests and forest management on emissions of greenhouse gases, particularly carbon dioxide, it is important to consider changes in the carbon content, or stature, of forests and not just changes in their area. The carbon content of forests, in MgC/ha (1 Mg = 106 g), varies geographically as a function of climate and soil fertility. It also varies temporally as a result of disturbances, including those attributable to human activities, such as logging or shifting cultivation, and as forests respond to management, for example, silviculture or fire suppression (Table 21.2).
21.3.1.3.1 Temperate and Boreal Forests
In temperate and boreal forests, the small net changes in area do not reflect the changes in carbon stocks. Estimates of recent changes in the carbon content of temperate and boreal zone forests suggest that forests were accumulating 0.7 to 0.8 PgC/yr (1 Pg = 1015 g) during the 1980s and 1990s.20–22 The factors responsible for this net accumulation are uncertain. Changes in land use are not thought to have been responsible because releases of carbon from decay of logging debris and wood products approximately balanced accumulations of carbon in growing forests.9,23 Thus, the accumulation of carbon measured in forest inventories has been attributed largely to changing environmental factors or to recovery from past natural disturbances rather than from human activity (land-use change).
A more recent analysis of the United States alone estimated a net uptake (sink) of 0.15 to 0.35 PgC/year,24 but less than half of this sink was in forests, and only 25% of the forest sink was attributed to past changes in land use or management. Although forests in the eastern U.S. were accumulating carbon as a result of fire suppression and the abandonment of agricultural lands, increased rates of logging in the northwest were responsible for a net release of carbon in that region. Another recent analysis of forests in the U.S. found that almost all of the carbon accumu- lating could be attributed to the age structure of the forests, that is, to past disturbances or management rather than to enhanced growth rates.25 Understanding the reasons for the accumulation of carbon is more than academic. If the present accumulation of carbon is the result of environmental changes (for example, increased concentrations of CO2 in the atmosphere, deposition of nitrogen, changes in climate), then the sink may be expected to continue until other factors limit forest growth. On the other hand, if the current sink is the result of past land-use and management practices, a continuation of this sink into the future is limited, because, once regrown, forests no longer accumulate carbon.
There are other reasons to suspect that the current terrestrial sink may not continue in the future.
Indeed, reductions in forest area could be self-amplifying. One of the effects of a global warming, for example, is likely to be an increased rate of respiration (including decomposition of soil organic matter).26,27 Increased emissions of respiratory CO2 and CH4, in turn, would increase the warming,
and the warming might then be beyond human control. Perhaps relatedly, the areas burned annually in Canadian and U.S. forests have increased in recent years. Earlier fire suppression allowed carbon to accumulate in both live and dead plant material, but in the last decades, the frequency of fires in Canada seems to have increased dramatically.28 The year 2000 was also a year in which an unusually large area of forest burned in the U.S. Whether the change is related to global warming and whether it will continue in the future are unknown at present. The net effect of numerous other global changes currently underway is difficult to predict. The danger, of course, is that the changes will reduce the capacity of the earth to support life and, further, that the changes will initiate processes in the earth’s climate system that are irreversible for generations. For the present, the net flux of carbon into or out of northern forests is small relative to the emissions of carbon from deforestation in the tropics. However, these possible feedbacks are not included in any of the models used to predict future climatic change.
21.3.1.3.2 Tropical Forests
There is considerable evidence that carbon is being lost from tropical forests in addition to the losses resulting from outright deforestation. Selective logging, shifting cultivation, and grazing are widespread within tropical forests, and the increasing practice of these activities for both subsistence and economic interests is reducing the standing stock of carbon in trees.29–35
Whether the carbon content is changing in response to factors other than direct human activity is as uncertain in tropical forests as it is in temperate-zone and boreal forests. In fact, changes are more difficult to evaluate in the tropics. In contrast to the national forest inventories of temperate-zone countries, tropical forests have rarely been inventoried. Direct measurement of change is limited to small areas where permanent plots are repeatedly sampled. An analysis of these plots suggests that growth rates may have increased in South America but not in tropical Africa or Asia.36 The finding may be artificial, however, and related more to methods of measurement than to real change.37 Evidence for an accumulation of carbon in undisturbed forests also comes from a small number of sites where the flux of CO2 has been measured over a forest. Measurements suggest that undisturbed tropical forests are indeed a sink for carbon.38,39 However, the method is sensitive to atmospheric conditions, which differ day and night. Because the fluxes also vary day and night, the possibility for a systematic bias in measurement is strong. Indeed, a positive correlation between nighttime respiration and wind speed seems to confirm a methodological, rather than a physiological, basis for variations in respiration and, hence, net uptake. If only those nights with a high wind speed are considered, the calculated net flux of carbon is a source, rather than a large sink (M. Goulden, personal communication).
21.3.1.4 Emissions of Carbon to the Atmosphere Caused by Land-Use Change The net flux of carbon to the atmosphere from deforestation and reforestation (in the broad sense, including activities affecting biomass within forests) are calculated from two types of information: historic and current reconstructions of land-use change (for example, annual rates of forest conversion to croplands and annual rates of wood harvest) and a knowledge of the per-hectare stocks of carbon in vegetation and soils and changes in these stocks as a result of land-use change.
Between 1850 and 1990, about 120 PgC are calculated to have been released globally into the atmosphere from changes in land use.40 This value is a net flux; it includes the uptake of carbon in forest growth following harvests as well as the releases of carbon from burning and decay. The total net flux from changes in land use is approximately half of the amount of carbon emitted from combustion of fossil fuels over this period. Before the first part of the twentieth century, the annual net flux from land-use change was greater than annual emissions from fossil fuels. Almost two thirds of the carbon released from terrestrial ecosystems has come from tropical lands; one third has come from temperate and boreal lands. About 90% of the net flux has been from forests, most of the rest arising from cultivation of mid-latitude prairies.