Abstract:
Human appropriation of natural resources, and the consequent loss of habitats, means that ecological restoration may in the future become a vital conservation tool. For this to happen, we have to understand the processes and factors that govern community assembly, and their management. Here I analyze data on community structure (richness, evenness, diversity, composition) of assemblages (millipedes, spiders, dung beetles, plants, birds, rodents), and on soil chemical and physical properties, to describe and evaluate post-disturbance dune forest regeneration patterns. Data were collected from program of a dune mining company), from post-mining rehabilitating dune forests, and from self-regenerating dune forests in northern KwaZulu-Natal. Both succession theory and community assembly theory predict that a species' niche will determine when it will colonize new sites, and by extension, what the eventual species composition will be. This type of control should result in deterministic regeneration patterns for a community. In support of this view, I found that the majority of habitat-age related changes in community structure and ecosystem function were either towards benchmark values (and will reach these values in less than 65 years post-disturbance), or were already equal to the benchmark. Age-related trajectories were repeatable between surveys and post-mining sites were changing as fast or faster than spontaneously regenerating sites. Moreover, detailed analysis of changes in community composition of millipedes, dung beetles, herbs, trees, birds and rodents showed that all of the taxa were also regaining the benchmark's species composition. However, community change was dependent on how it was measured - dung beetles recovered only species presence, but others relative abundances as well. Changes were almost never exponential, suggesting that colonization and extinction are not the orderly events foreseen by equilibrium biodiversity theory. Furthermore, the average abundances of birds, trees and millipedes in undisturbed dune forest patches were correlated with colonization success, suggesting that post-disturbance recovery through colonization may be controlled from outside the local community, rather than by species interactions. The recovery of the spider community appeared to be towards the benchmark forest community, but spider species composition was critically linked to microhabitat structure. Because microhabitat is not necessarily restored concurrently with forest community structure, the spider assemblage (and possibly other invertebrates) may not recover the desired pre-disturbance structure or composition. Dune forests thus seem to be resilient to mining disturbances, since most taxa were recovering structure and composition. However, classic successional and community assembly theories are unlikely to fully explain these community recovery mechanisms. More likely, post-disturbance recovery occurs because a new habitat passively "samples" the rain of dispersing propagules and individuals, leading to a high probability of capturing the average species composition of the region. A conceptual model of dispersal in the landscape suggested that species composition of new habitats might equilibrate to the composition of the closest habitat undergoing the least amount of species compositional change, although this may not apply to all taxa. This model may serve as the basis for directing future research and restoration management.