populations, which are dependent on these frequently burnt short grasslands. In South Africa, concern for the future of the Wattled crane has brought new emphasis to the conservation of the wetlands that support the remaining 100 pairs breeding in the country. Furthermore, 48% of the birds listed in the latest South African Red Data Book—Birds are grassland and wetland species (Brooke, 1984). Even more urgent, however, is the need to rehabilitate wetlands in the catchments of the country’s major rivers, which now carry up to 375 tons per square kilometer per year of soil lost from overgrazed rangelands and cultivated slopes.
Despite the importance of wetlands, conservation efforts have ignored them because of difficulties in defining, identifying, and mapping them. These problems have now been overcome by the use of hydromorphic soils as the key indicator of wetlands (Begg, 1986). Maps of such soils are readily available in most African countries, and their identification on aerial photographs for detailed checking in the field is relatively easy. Because the soils retain their structural characteristics longer than their vegetation cover, it is also possible to obtain a rapid estimate of the rate of change in wetland systems. In the Tugela Basin of Natal, up to 34% of wetland communities have been destroyed in the past 50 years due to cultivation or overgrazing followed by extensive soil erosion.
The most seriously neglected biotic systems in Africa are the tens of thousands of kilometers of streams and rivers that drain the continent. Even those rivers protected within national parks and reserves are subject to severe impacts from developments upstream or downstream of the protected section. Their narrow linear structure and diffuse spread make them difficult to contain within all but the largest protected areas, and detailed information on their biological values, degree of disturbance, and conservation needs are difficult to synthesize. They are, like wetlands, invariably ignored by conservation planners.
During the last 3 years, considerable progress has been made in overcoming the difficulties of analyzing river conservation needs in South Africa. A computer-based expert systems technique developed by O’Keeffe et al. (1986) simulates the logic processes of river ecologists and converts the multivariate probabilities and diffuse intuitions of real-life situations into a simplified expression of river conservation status. The advantage of the system is that the complex calculations of interrelationships are handled by the computer, but the flexibility to take account of unusual situations is retained. A wide range of attributes are included in the system framework, and the users enter the best-available information on each of these, ideally within a workshop situation where the researchers, conservationists, and planners can pool their resources.
Because each river is different and, therefore, all attributes do not always apply in the same way, a number of rules are included in the program to interrelate attributes or modify their effect. The system is designed to assess whole river systems, individual rivers, parts of rivers, or points in a river. The extent of the assessment must be defined by the user beforehand. Obviously, the assessment of a whole river system will be performed at a coarser resolution than that for a small tributary. The flexibility of the system, allowing successive levels of data to be added as knowledge improves, makes it especially useful in Africa, where few river systems