n Ekologiese studie van die suidelikste mopanieveld in die Nasionale Krugerwildtuin

Abstract

A comprehensive inventory of the abiotic and biotic components and the interpretation of the interactions between such components, is a prerequisite for any form of monitoring and wildlife management in a National Park. As a contribution to such a programme this information was collected for a chosen study area in the Kruger National Park. The study area covers an area of 2701 km2 in the northwestern corner of the Central District of the Kruger National Park between 24° and 24° 30' south and 31° 15' and 31° 45' east. The underlying parent rock of the study area is Nelspruit Granite Formation in the west, shales of the Ecca Group, Clarens Sandstone Formation and Elliot Formation in the central areas and Sabie River- and Letaba Basalt Formations in the east. The soils that have developed on these parent materials range from shallow, sandy soils on granite to deep clayey soils on basalt. The rainfall ranges from 450 to 550 mm per annum and occurs mainly during the summer months. The study area experiences a subtropical climate with warm, wet summers with temperatures of up to 40°C and cool, dry winters with frost a rare occurrence. The soil types of the study area were classified and described by using the Binomial Soil Classification System of South Africa (MacVicar et al., 1977) and selected profiles were sampled for physical and chemical analyses. Sixteen soil forms and 45 soil series were identified. The classified soil types were then used for correlations with plant communities. The correlations were good for deeper soil forms, but poor for those soil forms with an A-horizon only (Mispah- and Milkwood Forms). This correlation could be improved by altering the classification system. As proof of the reliability of the soil classification system, which was only based on physical characteristics, the physical and chemical characteristics of the soil types were ordinated by means of a Principal Component-analysis. The ordination showed that profiles of a specific soil form differed markedly in physical and chemical characteristics when the parent material was not the same. The physical and chemical characteristics of soil forms on the same parent material, for example granite, showed a continuum which ranged from sandy, leached soils (Hutton and Clovelly Forms) on the crests to clayey eutrophic soils (Valsrivier and Swartland Forms) on the bottomlands. The vegetation of the study area was classified in a hierarchical, plantsociological system by using the Braun-Blanquet technique. The data from 250 releves were classified into three alliances, seven associations, 19 sub associations and six variations which were described in terms of vegetation composition and structure. The habitats of these associations and subassociations were ecologically interpreted and mapped. A list of all plant species which occurred in the study area was included. The vegetation structure at each releve was analysed on the basis of crown cover per height class by using a Principal Component-analysis. Vegetation structure was classified into 36 structural classes which could be determined by means of a key. The names of the structural classes were used as suffixes in the denomination of the individual sub associations in the plant sociological study. This classification of vegetation structural classes could also be used to describe animal habitats. As the study area included the southernmost occurrence of Colophospermum mopane, the distribution of this plant species was accurately mapped. The habitats where Q. mopane occurred and where it was absent, were ordinated by means of a Stepwise Discriminant-analysis. From this analysis the habitat preference of Q. mopane could be determined. It was concluded that there are some areas where Q. mopane does not occur at present, but which is suitable habitat for the species. However, there was also some evidence that Q. mopane is at present extending its distribution. Landscapes were identified on the basis of homogeneity in geomorphology, macro-climate, soil- and vegetation patterns and associated fauna. A total of 13 landscapes were identified, described and mapped in the study area. The preferences that certain game species showed for these landscapes, as well as the habitat characteristics whose contribution was largest to these preferences, were determined by means of a Correspondence-analysis. The geographical distribution, size and relative homogeneity of the landscapes, make them practical units for wildlife management programmes.