Plantegroeidinamika tydens die prosesse van herstel en degradasie in 'n gedeelte van die grasveld- en Karoobioom van Suid-Afrika

Abstract

Dynamic changes, such as the composition, basal cover and spatial pattern of species take place in a natural rangeland when submitted to disturbances such as an unpredictable or erratic rainfall pattern or different grazing conditions. Depending on the type and degree of disturbance, as well as on the change in the environmental conditions, species changes usually start in small patches, influencing the condition of the vegetation on a macro-scale in the short or long term. Macro-plots were selected in three study sites on a rainfall gradient ranging from a low annual rainfall in the Karoo Biome at the Grootfontein ARI, to Glen and Reitz in the Orange Free State, which are characterized by higher annual rainfall patterns. To study the effect of grazing, macro-plots that were previously grazed at certain levels were enclosed and withheld from further grazing (so-called recovery sites), while neighbouring macro-plots were still being lightly, moderately or heavily grazed. Microplots, representing patches of a certain vegetation composition and conditional states, were selected in these macro-plots. Successional changes, including aspects such as the increase and decrease in basal cover or changes in the spatial pattern of single or groups of species were studied in the micro-plots for a period of three years ( 1990 to 1992). By means of a photographic technique, successive changes could be identified only visually, while changes in the percentage basal cover were quantitatively determined using a point quadrat monitoring technique and drawing species distribution maps of each micro-plot. The construction of transition matrix tables led to the identification of indicator species that either remained constant in basal cover, or were subjected to a species by species replacement process, due to the different rainfall and/or grazing patterns. The largest changes, however, took place between the decrease or increase of the basal cover of a species and the percentage of bare ground. By making use of the Markovian Projection Model, the vegetational changes taking place from 1990 to 1991 were used to predict possible future species change and vegetation compositions. Together with the values calculated to indicate change and the indicator identified species, the degree of stability of a micro-plot represented by a certain vegetation composition could be established. Some species were indicators of

a changing rainfall pattern, while others were more influenced by different grazing treatments. On the whole, micro-plots representing a greater stability were characterized by a more climax type of vegetation with a low species diversity. Micro-plots that had previously been or were still being heavily grazed and which already showed a higher degree of degradation, however, were characterized by a higher species diversity and species turnover, resulting in a greater instability. Except for some of the more widely spread species, such as Themeda triandra and some of the Eragrostis species, most of the species were area-specific and served as indicators of vegetational changes in either the drier Karee or wetter Orange Free State areas. The conditional state of the macro-plots was assessed, using already established range condition models for each study site. Changes in the condition from 1990 to 1992 were indicated by trajectories and veld condition scores. Species contributing to the changes in the conditional states of the micro-plots were in most instances the same as the species used as indicators of stability in the micro-plots. This finding underlines the fact that vegetational changes on a micro-scale can be used to some extent to explain conditional changes on a macro-level. The ecological role and behaviour of single species that are usually driven by seasonal fluctuations or episodic events, can, however, only be established by long term monitoring of permanent plots on a micro-scale. From the results of this study it is evident that vegetation dynamic changes, indicating processes of recovery or degradation, are very complex and have to be studied over a much longer period of time. This is especially the case if the effects of rainfall and/or grazing were to be monitored. Certain deficiencies and recommendations have been identified during this study that have to be taken into account in long term vegetation dynamic studies in a natural rangeland.