Biogeographical insights from ecotones and phytogeographic regions in southern Africa: case studies on invertebrates and alien plants

This thesis addresses questions within the research fields of invasion biology and spatial ecology, with a focus on species distribution patterns, biogeographical regions and ecological transition zones, or ecotones. More specifically, species distribution patterns in alien plants at large spatial scales using atlas data, and invertebrate patterns making use of field data collected at a smaller scale (total extent ca. 30 km). First I show that alien plants form large-scale geographically differentiated species assemblages in southern Africa (i.e. South Africa, Lesotho, Swaziland, Namibia and Botswana). I demonstrated this by mapping and describing several alien phytogeographic regions at a quarter-degree spatial resolution, and further suggest possible environmental and human-caused determinants of each of these regions. Second, at the same spatial resolution (for South Africa and Lesotho combined, and each of the plant biomes), I show that relatively higher levels of alien plant species richness occur at or near to ecotones, compared to areas that are spatially further away from these ecotones and that typically represent the core habitat of the ecoregions examined. This finding remained valid after taking into account the underlying positive relationships between alien plant richness and energy availability. I also suggest that it is the relatively higher environmental heterogeneity at ecotones (represented here by spatial variation in altitude, rainfall and geology) that promote high alien plant richness. Third, at a smaller spatial scale I report several examples of change in beetle and spider species composition across a savanna-grassland ecotone in the west of South Africa's Free State Province, with the ecotone itself supporting comparatively lower levels of species richness and abundance. This contrasts with a popular assumption that ecotones are characterised by high species richness. Data gained from long-term intensive sampling is preferable for ecological studies, but not always available or practical to acquire; however with the three studies in this thesis I show that data from existing species atlases and feasible short-term surveys can be successfully applied to answer a variety of ecological questions.