Tropical forests were largely intact until colonial times, when large tracts were removed to provide raw materials for railroads, ships, etc., in the period following the industrial revolution. The loss of tropical forests escalated in the second half of the 20th century. According to the UN Food and Agriculture Organization (FAO, 1996), about 15.4 million ha of natural tropical forests are lost each year. Of this, 42% occurs in Latin America, 31% in Africa, and 27% in Asia. Brunner et al. (1998) estimated tropical deforestation at 19.1 million ha/yr during the period 1990 to 1995. There has, however, been a large increase in area devoted to forest plantations. By 1990, there were 61.3 million ha under plantations and the rate of establishment is now about 3.2 million ha/yr (FAO, 1996).
As pointed out by the IPCC (IPCC, 1996) global estimates of C emissions from deforestation have remained highly uncertain and show high geographical variability. The magnitude of forest regeneration (particularly secondary forest regrowth and regrowth of abandoned lands) and forest degradation processes is not well documented. Improving the accuracy of these estimates remains an urgent and challenging task (Houghton et al., 2000).
Estimates of C emissions from land-use change and forestry activities in the
tropics during the1990s range from 1.1 to 1.7GtC/yr, with a best estimate of
1.6GtC/yr (Brown et al., 1996b; Melillo et al., 1993; Bolin et al., 2000). These
estimates may change with improved information on biomass densities and land-use
conversion. Detailed studies for major tropical countries in the early 1990s,
studies that include forest regeneration and afforestation, show lower net emissions
for most countries than those from aggregate estimates (Makundi et al., 1998).
A review of scenarios of future land-use changes in the tropics, and their implications for greenhouse gas (GHG) emissions, shows a wide range of estimates, particularly for the first part of the 21st century, where estimates differ by a factor of 14 (Alcamo and Swart, 1998). These disparities reflect a lack of agreement on the definition of deforestation, and a lack of knowledge and agreement on the estimation of C emissions (Alcamo and Swart, 1998). These scenarios can be divided into two groups: in one group emissions decline smoothly after 1990; in the other group emissions increase for a few decades after 1990.
The rates and causes of land-use change vary by region and scale (Kaimowitz and Angelsen, 1998). Deforestation is often considered a one way process, but the landscape is a dynamic mosaic of land uses and vegetation types, with transitions both to and away from forest (Houghton et al., 2000). Natural factors, such as forest fires and pests, as well as socio-economic processes, many of which are not seen at the local level, interact in complex ways, complicating analysis. Understanding the causes of this mosaic of land-use and/or land-cover transitions in order to understand and predict the net effect on deforestation rates and C emissions remains a key research challenge.
Conversion of forests to pasture and cropland has been the most important proximal cause of tropical deforestation. Non-sustainable logging has been the leading factor in parts of Southeast Asia, whereas excessive harvest of wood fuel has been important only in specific sub-country regions and in some African countries (Kaimowitz and Angelsen, 1998). According to Bawa and Dayanandan (1997), the causes (correlates) of deforestation are many and varied, with complex interactions. Overall, Bawa and Dayanandan found that population density, cattle density, and external debt were the key factors. In Africa, the most important factors were extraction of fuelwood and charcoal and demand for cropland; in Asia, it was cropland; and in Latin America, it was cattle density.Most analyses of land-use change and forestry have concentrated on proximal reasons for land-use and/or land-cover change; that is, on land uses such as agriculture, pasture, and timber extraction that replace forests. But Meyer and Turner (1992) have identified six underlying forces: (1) population, (2) level of affluence, (3) technology, (4) political economy, (5) political structure, and 6) attitudes and values. The influence of each varies by region and country.
The rate of population growth is now apparently declining, but the population, and hence the demand for food and other land services, is still growing (Roberts, 1999). Population growth has been widely cited as a major cause of deforestation (Myers, 1989), but the relationship between population and deforestation is not simple. Population growth exerts increasing pressure on resources, but whether these pressures lead to forest degradation or to positive changes (e.g., afforestation, improved forest management, and better technology) depends largely on social structure. Extensive migration may also lead to deforestation and soil erosion. Simplistic assumptions about population and deforestation also do not apply where high population densities and/or growth rates are accompanied by forest conservation and reforestation programmes. In India, for example, deforestation rates have declined since 1980, despite population growth, owing to effective forest conservation legislation (Ravindranath and Hall, 1994).
Patterns that affect land-use are changed by economic development. Affluence usually increases consumption, but it does not necessarily decrease terrestrial C stocks. The maintenance of ecosystems tends to improve with increasing and better distribution of wealth, as well as with proper institutional structures and sound development strategies. The demand for and interest in forests and their services is the driving force for the technological and economic capacity to maintain forests. Also, wealthy societies tend to be urbanized and this may reduce destructive pressures on forests. Technological development provides efficient tools for land-use change and for high-value, alternative uses. Technology can also limit encroachment. As seen by the green revolution in agriculture, technological development can increase productivity on intensively managed land, thereby releasing other land areas from agriculture (Waggoner, 1994). Nevertheless, there is always the risk of leakage (i.e., tendencies to transfer destructive operations from the developed to less developed areas and countries), or the possibility that technology development and transfer will have positive spillover effects (Brown et al., 2000; Noble et al., 2000)
In many countries, especially those seeking development of frontier areas, subsidies are provided for activities promoting economic development. Land clearing may be subsidized directly or by providing property rights to cleared land. Frontier development is often considered desirable for security or where there is a disputed area.
Land-use change is driven largely by efforts perceived as best and highest use of the land. But benefits of the land that are non-market and/or external to the direct user (e.g., watershed protection, biodiversity, and carbon mitigation) may be ignored by land managers. For example, the decision to convert forestland to agriculture may ignore the many external and non-market benefits lost. Moreover, where long-term land rights are insecure, lands may be used to generate short-term benefits, with disregard for long-term benefits.
Factors related to social structure and political economy have not been studied widely, but studies at the country and regional levels suggest that deforestation is favoured by the following factors: growing landlessness and persistent inequalities in access to land, insecure land tenure, land speculation, rising external debt, large-scale expansion in commercial agriculture, erosion of traditional systems of resource management and community control, and widespread migration of impoverished people to ecologically fragile areas (Hecht, 1985; Palo and Uusivuori, 1999; Tole, 1998).
Other reports in this collection