It is possible to be an armchair meteorologist. While the rain is pouring down outside, one can imagine oneself travelling and analyse in detail the climatological statistics of the arid regions of the world. By so doing, one can identify those places that ought to be, and those that should not be, a desert.
If you do this you make an odd discovery. Climatology suggests that a little more than one-third of the world's land surfaces could be classified as desert. But if you leave your armchair and actually go and look at the actual state of the soil and vegetation, you find that the desert proportion rises to something like 43 per cent. The difference represents the extent of "desertification" due to human activity; it is equivalent to over nine million square kilometres, or more than half the area of Brazil.
Scientists have been divided on the question of whether this area of man-made desert is, or is not, increasing. In the case of the Sahara, for example, some, probably the majority, believe that the southern flank is moving systematically towards the Equator at a rate in places reaching 10 km a year; they variously blame diminishing rains, unwise farming and irrigation practices, and overgrazing by nomadic cattle herders. Others are of the view that the southern boundary of this desert is dynamic - in the scientific sense of being mobile - and that it oscillates in tune with the region's notoriously variable rainfall, which waxes, wanes and even disappears not just from year to year, but from decade to decade. They argue that although the desert may make gains in certain years, fertility is quickly re-established, sooner or later, when the rains return.
New information on this topic is to hand. Satellite data, gleaned over two decades, shows clearly, it seems, that it is variations in rainfall and not the direct mishandling of land by humans that mainly governs the size of the Sahara. A satellite instrument called the Advanced Very-High Resolution Radiometer, or AVHRR for short, measures the infra-red radiation reflected by plants, and from these data a "greenness" index can be calculated showing the areas of ground covered by vegetation. Images collected from 1980 to 1998 show that there is no systematic trend in the advance of the Sahara into the semi-arid grasslands of the Sahel along its southern border.
It seems the Sahara-Sahelian boundary moves south during very dry years, but rebounds northwards when the rainfall has been plentiful. For every "bad" year since 1980 there has been a corresponding good year; they balance out, and there is no overall trend in the Sahara's size.