Origins of fairy circles and heuweltjies (earth mounds) in Southern Africa

Both the Namibian fairy circles and South African heuweltjies (‘little hills’) are large (5-20 m diameter) and prominent, regularly-spaced bare patches that occupy up to 25% of the landscape. In contrast to their obvious appearance, their origins and nature remain hotly-debated. Many of the earlier hypotheses for their origin (allelopathic effects, gas emissions, fossorial rodents) have largely been discounted, leaving the termite and self-organising vegetation dynamics hypotheses as explanations for both fairy circles and heuweultjies. The termite hypothesis links fairy circle formation with the subterranean and polycalic nest system of the Sand termite (Psammotermes allocerus), and heuweltjie origin to the very large nest of the Southern harvester termite (Microhodotermes viator). In support of the termite hypothesis, the raised earth mound of the heuweltjies are homologous with the very large nests of Microhodotermes, whose hives can have a diameter of one metre. Nutrient enrichment and turnover of plant communities on heuweltjies are consistent with termite-altered soils, and mound density increases with increasing rainfall and NDVI, suggesting an interaction between mound density, precipitation, vegetation cover and intraspecific competition. Micropedological studies of heuweltjies also show evidence of geochemical soil modification by termites in the localised petro hardpan which is restricted in the landscape to heuweltjies themselves. In contrast, the self-organising vegetation dynamic hypothesis is a model-based explanation for heuweltjie formation which argues that historically bush clumps accumulated soil, attracted termite colonies, and were then subject to peripheral soil erosion, resulting in the elevated heuweltjie. The association between heuweltjies and Microhodotermes termites is considered secondary. Self-organising vegetation pattern models have also been used to predict periodic or even hexagonal lattices that match spotted landscape patterns, where the bare patches are considered the result of competition for resources or nutrients. These models have been used to generate similar spacing patterns observed in real fairy circles. The value of such models is discussed in relation to their limitations (relating to assumptions, predictions and lack of supporting data for model variables). Similar under-dispersion patterns can also be demonstrated for the nests of certain social insects (e.g. heuweltjies) which display hexagonal lattice patterns where saturated populations occupy homogeneous landscapes. Thus models using both plant or social insect variables are capable of generating the same under-dispersed pattern, and the competing hypotheses need to be evaluated using the available field data.