All deserts have evolved under water scarcity; drought does not destabilize them. But humaninduced degradation does occur - through overgrazing, clearance of woody vegetation, farming, irrigation-induced salinity, soil and water contamination by agrochemicals, groundwater exploitation, and traffic and urban-industrialmining occupation (Figure 4.8). Apart from grazing and wood collection, these pressures tend to be restricted to relatively small areas. It is important to note, however, that these areas are generally of higher productivity, and in them the myriad needs of human societies must be met alongside those of native biota. By definition, dryland farming (that is, not under irrigation) is very restricted in true deserts. There is hardly any natural accumulation of organic matter to lose, dunes are naturally mobile, and large areas have developed a stony surface lag of stones and gravel that resists further deflation and rain splash. The most important sources of dust are ephemeral river courses, the active margins of alluvial fans, and dry lake beds.

Desertification paradigms

In deserts, land degradation may manifest in a variety of ways: as changes in vegetation composition, as erosion of soils by wind and water, or as salinization and pollution associated with irrigation. The major causes of land degradation in deserts are overgrazing, wood collection and deforestation, and non-sustainable agricultural practices. The Global Assessment of Human- Induced Soil Degradation (GLASOD), based on expert opinion (UNEP 1997), estimated that 20 per cent of the world's deserts are affected by some type of land degradation.

To frame a discussion on land degradation of deserts it is important to differentiate between desertification (that is, land degradation in drylands that may be semi-arid or dry sub-humid areas) and degradation of deserts proper. As noted throughout this report, desertification is ecosystem degradation and not a process whose end product is the rich and specialized desert ecosystems that are the subject of this report. The discussion of degradation in deserts should then be placed within the context of two differing points of view: the desertification paradigm versus the counter paradigm (Safriel and Adeel 2005).

The traditional desertification paradigm analyzes soil and vegetation degradation as the triggers of a set of negative interactions that lead to a generalized impoverishment of the environment. According to this view, desertification takes place mainly where agriculture and intensive grazing are the major source of local livelihoods - that is, in dry-farming areas outside the boundaries of the true deserts. The loss of soil and vegetation cover leads to an associated decline in the provision of ecosystem services and a rise in poverty. Additionally, the changes in land cover and soil are also linked to increased aridity as part of a negative feedback loop, a fact that makes desertification, within this paradigm, practically irreversible, and its inevitability increases with aridity (Cleaver and Schreiber 1994).

The main tenet of the counter paradigm is that the interaction of direct and indirect drivers combined with the local situation, can create different outcomes. While the desertification paradigm focuses only on the negative interactions, the counter paradigm approach considers both the negative and the positive interactions, depending on how humans respond to the direct and indirect biophysical and anthropogenic drivers of change. According to the counter paradigm, the drivers, processes, and events described in the desertification paradigm do exist, but the chain of events that leads to desertification and the chainreaction cycle of reduced ecosystem productivity and poverty are far from inevitable. The message of the counter paradigm is that the interacting direct and indirect drivers combined with the local situation can create a range of different outcomes, and that raising a general alarm based on often insufficient scientific understanding or evidence is, in the end, much less effective than identifying individual problem areas where large influxes of refugees or other complicating factors have led to unsustainable local responses. This alternative paradigm also postulates that it is crucial to distinguish between problems originating from the natural harsh and unpredictable conditions of dryland ecosystems, such as cyclic droughts, and problems caused by unsustainable management of the environment, since the remedies will often be different (Safriel and Adeel 2005).

In general, desertification is a problem associated more with semi-arid dryland agriculture than with true arid and hyper-arid deserts. Where desertification does occur on deserts, it takes place mostly on the desert margins and less in their vast interiors. Other sections of this book discuss in detail the problems of land degradation for particular cases of desert rangelands and pastoralist societies; cultivated lands, soil erosion, desert irrigation, and oases; mining and mineral resources; urbanization, industry, and infrastructure; tourism in deserts; military testing areas and historic battlegrounds; and low impact activities in protected areas.