The world's forests, especially the tropical forest belt, have been conceived as part of the planet's cooling system (Lovelock 1990). Although less significant than the oceans, (estuaries and) tropical forests are key regulators. The forests remove atmospheric carbon dioxide and transform it into wood, soil, perhaps eventually peat, coal, etc. Forests also pump water into the atmosphere. The moisture produced creates clouds, which cool both the forests and the planet by reflecting solar radiation. The rain from these clouds helps to sustain more forest in areas that would otherwise be too dry (Dickinson 1987).

Lovelock (1991) warns that the loss of the tropical forests could contribute to a sudden and dramatic failure of the planet's current system of climatic regulation, regarded as already being close to the margins of its stability. Using analyses based on the mathematics of dynamic systems theory, and images from the liveliest traditions of environmentalism, Lovelock predicts a future of unprecedented and violent environmental fluctuation (Lovelock 1990). This would be followed by an abrupt jump to an equilibrium state, with a very much higher stable planetary temperature. Lovelock (1990) links the destabilising influences of deforestation and accelerated soil erosion together as a planetary disease, "exfoliation". The guru's advice to environmentalists is to plant trees and to minimise the release of greenhouse gases to the atmosphere, even if that means using nuclear energy rather than burning fossil hydrocarbons (Lovelock 1990). Indeed, NASA-based Noever and team (1996) calculate that, to preserve its current temperature regime, the planetary system already requires more forest and less desert.

On the smaller landscape scale, where most geographers work, deforestation affects most of the issues that concern geographers. It impacts on both macro- and micro-scale climatic patterns (Reading et al. 1995). It leads to dysfunctions in landscape systems, which are caused by the interrelated degradation of its climatic, hydrological, edaphic and biological components. Deforestation allows increased soil erosion, increased landslide activity, sediment pollution, changes in fluvial geomorphology, and changes in the hydrological, biogeochemical and climatic regime (Haigh 1984). Tropical forests are enormously complex and highly stable systems. After deforestation, they are replaced with systems that are much simpler and have much reduced biodiversity and much lower stability (Reading et al. 1995). These replacements are, in general, much less efficient in the tasks of self-preservation, they retain and recycle nutrients less efficiently, but they may recover more rapidly from disturbances such as destruction by fire.

Given its significance, current estimates of deforestation are alarming. Of course, these estimates, in common with those for similar global issues - soil erosion, soil degradation, biodiversity loss and the rate of species going extinct - are often suspect. Their range is huge and shifts according to the definition of deforestation applied, the techniques used for estimation and their efficiency, the method employed to convert remotely sensed data into deforestation data, and the geographical areas selected to "ground-truth", that is calibrate empirically the remotely sensed data (Grainger 1993; Parisi and Glantz 1992; cf. Skole and Tucker 1993). This is quite apart from any shift caused by the bias and ambitions of those who would use such data. However, recent years have seen a convergence of estimates (Downton 1995).

Forest, of some description, covers about 40 per cent of the Earth's land surface. The FAO Forest Resource Assessment suggests that the world's forests cover 3454 million ha (1995), a little more than half of which lies in developing nations. The FAO (Food and Agriculture Organisation of the United Nations) definition includes forests with a greater than 10 per cent crown canopy cover in the developing world and 20 per cent cover in the developed world (World Resources Institute 1996). The FAO (1997) also suggests that 15.4 million ha of tropical forest is lost annually. Myers (1993) suggests that the loss to the entire biome is about 2 per cent per year. Murali and Hegde (1997) prefer 1.8 per cent and that the rate is greatest in the smaller nations, especially in Africa. The World Resources Institute (Washington) suggests that the true rate is 0.8 per cent per year. It adds that, between 1960 and 1990, Asia lost nearly a third of its tropical rain forest and Africa and Latin America a sixth each. However, the rate of increase in the area deforested declined everywhere except Latin America, where agricultural extension continued to accelerate (World Resources Institute 1994).