Rangeland condition in the Ceres Karoo: the importance of long-term studies

The impact of the small stock industry on Succulent Karoo vegetation has long been acknowledged, and there is a need for researchers and managers to better understand the dynamics and processes leading to vegetation degradation and recovery. Despite the fact that these arid systems tend to hold high demographic inertia, as well as being prone to sudden and unpredictable events, there are few data-sets that are long enough (>50 yrs) to adequately distinguish 'noise' from true changes. This study examines an extensive data set of plant % cover, recorded using a rapid transect step-point technique, for the period between 1971 and 2002 throughout the Ceres Karoo. Correspondence analysis (CA) ordinations were used to show plant community changes from year to year at two sites: one that has been rested for many decades, and one that has used the Group Camp system since 1970. CA ordinations were also used to depict changes between 1992 and 2002 on three farms using different grazing systems. Results are discussed against the backdrop of the Stock Reduction Scheme initiated in the 1970's. Of the two farms examined from 1971 to 2002, the one using the group camp approach has shown an initial lag-period of about 10 years, and a subsequently steady and directional turnover of plant communities, increasing in cover of desirable species, until 2002. The rested farm showed no identifiable change. Of the farms studied between 1992 and 2002, a clear separation was found between the one using the Group Camp system and the others that are only grazed in winter. The former farm appeared to be showing the greatest amount of change. These results challenge the opinion that rested arid region veld is unlikely to recover. On the contrary, there has been a move toward more desirable veld since the 1970's in land that has not even been rested, but has had relatively reduced stock numbers. The more rested lands appear to be healthier than those that have been more frequently grazed, but they are not showing clear signs of change, supporting theories that arid region vegetation dynamics are characterised by a state and transition type of model. An important pattern to note is that changes from a degraded to a more desirable veld are characterised by a long lag period of more than 10 years, with subsequent changes occurring throughout a 20-year period. Thus the importance of allowing rangelands sufficient time to recover is highlighted, as is the importance of establishing, and continuing existing, long-term data sets