Die identifisering en karakterisering van veldekotope in die noordwes-Transvaalse suur- en gemengde bosveld

Abstract

The vegetation of the study area, which covers 1407 km2, was mapped on a scale of 1:50 000. The mapping units were simplified to major vegetation types. . Veld condition differences were sampled in the major vegetation types. Various data sets, that consisted of frequency and cover values, were subjected to TWINSPAN and DECORANA classification and ordination algorithms. Similar results were obtained from the different data sets. These results were ecologically interpreted and relationships between the major vegetation types and the habitats were determined. A TWINSPAN classification of a data set that consisted of only grass species, was compared to a TWINSPAN classification that consisted of both grass and woody species. The generally similar results are an indication that grass species composition is also determined by the habitat. However, a better ecological impression of the vegetation was formed when the data sets that consisted of grass and woody species were subjected to classifications and ordinations. A classification and an ordination of soil variables reflected the major vegetation types. It was found that clay content was significantly correlated with most other soil variables. SPOT images were digitally enhanced and classified to major vegetation types that consist of interrelated vegetation, geological and soil complexes. The spatial relationships between vegetation, geology and soil were determined and quantified by using spatial analysis techniques in a GIS environment. These results complimented the relationships determined by means of multivariate techniques. Rainfall does not explain the distribution of the vegetation in the study area satisfactorily, while habitat does. The conclusion is made that it is possible to predict distribution of the vegetation by means of habitat classification. It is shown that habitat has a clear influence on reaction of vegetation changes due to grazing, towards grazing. Species composition in a specific habitat, are much more profound than species composition changes caused by minor habitat differences. The same species reacted differently to grazing in the different habitats. Clear difference in the sensitivity of the different habitats for degradation is shown. SPOT satellite images were also used to map veld utilization patterns. The cover of the herbaceous layer was directly linked to veld condition. The integrated GIS - remote sensing approach that was followed, made it possible to quantify the degradation of the different habitats separately. There were major utilization differences between the habitats identified. The conclusion is made that habitat has a major influence on the vegetation in the study area. It also influences the intensity of utilization of vegetation, and the sensitivity to degradation. The broad divisions of vegetation in major vegetation types are probably valid veld management divisions. The techniques applied in this study are showing a great potential as an effective aid in vegetation mapping and veld management. An adequate knowledge of the ecology, of any study area to be investigated, is a prerequisite for the successful application of these techniques.