The response of biological communities to spatial and temporal changes in a regenerating coastal dune forest along the north-east coast of South Africa

Abstract

Ecological restoration that aims to reinstate indigenous processes may be constrained by regional and local conditions, especially those that drive dispersal and colonisation. Local conditions can be managed, while regional conditions cannot. The management of costly rehabilitation programmes is considered best practice when scientifically informed. My thesis documents the responses of biological communities to a range of local conditions developing in coastal dune habitats in response to ecological restoration. Here, landscape-level (spatial structure of patches of tree canopies) local conditions were distinguished from site-level (topography, soil nutrient content, woody plant community richness, microclimatic variables) local conditions. The spatial structure of tree canopies varied over time and differed between the mining lease site and the relatively undisturbed benchmark site prior to (1937- 1970), and after (1990-2006) mining. For example, approximately 20% of the mining lease site and 40% of the benchmark site was covered by tree canopies prior to mining. However, after mining and rehabilitation, the structure of tree canopies began to converge towards that of the benchmark. The topographic profile of coastal dunes was less heterogeneous after mining and rehabilitation than before. Aspect, elevation and gradient of dune slopes were also different and had shifted in space. Variability in the structure of tree canopies could not explain variability in species richness, forest-associated species richness and proportion of benchmark species for the millipede, dung beetle and bird communities ashad been expected. However, species composition did change (though idiosyncratically) with age-related changes in soil nutrient availability and tree species diversity. Temperature, relative humidity and light intensity varied with dune topography, but soil nutrient content (C: N ratio and pH) was better accounted for by the age of the regenerating forest than by dune topography. Similarly, analysis of covariance suggested that tree canopy density, woody plant richness and millipede species richness only responded to the aspect, elevation and gradient of restored coastal dunes when age was taken into account as a covariate. However, the response of keeled millipedes to dune topography, regardless of regeneration age, suggested that the microclimatic variability brought about by topographic heterogeneity may provide these specialists with suitable microhabitats. Throughout my thesis, the age of regenerating patches of indigenous canopies was often more important as an explanatory variable than habitat conditions per se. Age itself is not a determinant of biological communities, but merely the axis along which habitat conditions change with succession, and later, patch dynamics. Therefore, as elsewhere, my thesis has highlighted age as a useful proxy for the response of biological communities to local conditions. It seems that managing local variables such as those considered in my assessment is not an avenue through which to enhance restoration. After kick-starting initial conditions, best practice rehabilitation management should therefore focus on minimizing external disturbances rather than interfering with natural processes.