The Namibian savannah ecosystem has experienced considerable anthropogenic pressures, which have resulted in the disruption of key ecological processes, and consequently, the proliferation of a dense vegetation structure, commonly known as bush encroachment. Approximately 45 million hectares of land have been affected, which has caused a significant decline in the grazing capacity, poor economic returns for the farmers, and loss of suitable wildlife habitat. Restoration thinning, a method that involves the selective removal of excess trees/shrubs, has been applied to counteract the negative effects associated with this phenomenon.
This thesis aimed to assess the effect of thinning on (a) the activity patterns of local ungulates and predators, (b) the encroaching woody vegetation in terms of (i) regeneration, and (ii) structure, abundance, and habitat sighting lines, in a savannah habitat in the north-central region. Generalised linear mixed-effects models (GLMM) and linear mixed-effects models (LME) were used for statistical analysis. Results showed that thinned areas had overall greater wildlife activity. Thinned areas also had reduced tree/shrub abundance, which was significant for the mature height classes. Natural regeneration was rapid in thinned areas, where the abundance of young cohorts was 34% greater than non-thinned areas. In the thinned areas, red umbrella thorn (Vachellia reficiens) was significantly reduced but umbrella thorn (Vachellia tortilis) numbers increased. Tree/shrub densities in the thinned areas fell within the commonly accepted range for a 400 mm rainfall area (600–750 tree equivalents (TE) ha–1). Thinning significantly modified the dense thornbush to an open vegetation structure with a low woody canopy cover that favours grass growth and provides greater sighting lines for open savannah wildlife. This thesis demonstrated that thinning was effective in controlling bush encroachment and could be used as a method to restore other affected areas. However, periodic post-thinning management is recommended to control the established samplings.