Mapping and predicting potential distribution patterns of free-range livestock in the rural communal rangelands of Mgwalana, Eastern Cape, South Africa

Communal rangelands provide habitat to many plants and animals. However, there is evidence that livestock cause range degradation. Range degradation occurs because livestock select grazing based on the availability of resources such as water and forage material, their use of the landscape is non-uniform, consequently causing resource deterioration. Range management is thus necessary because communities depend on range condition for livestock productivity. However, precise quantification of livestock distribution within communal rangelands is lacking. In developed countries, Global Positioning Systems (GPS) collars have been used to monitor wildlife and domestic livestock in pastures and seem to have worked efficiently. However, in a developing country like South Africa, GPS technology to monitor animal behaviour has been used only for wildlife on privately owned land. The high costs of monitoring livestock herds in large open areas such as communal rangelands have resulted in little or no monitoring of domestic livestock using GPS technology. This study links monitored livestock distribution to physical landscape variables in Mgwalana, and uses the modelled relationship to predict livestock distribution in quaternary catchments, T12A and T35A-E. The research addresses the questions (1) where do livestock spend time in the wet and dry seasons? And (2) how can areas of potential livestock distribution be identified in other catchments where actual distribution is unknown? Livestock were tracked during the wet and dry seasons using GPS collars. The resulting distribution data is combined with selected physical landscape variables to identify selectivity. The GPS location data and the physical landscape variables are used to predict potential livestock distribution where distribution is unknown in quaternary catchments (T12A and T35A-E). The ArcGIS Predictive Analysis Tool (PAT) was used to extract the selected landscape variable ranges based on the GPS location data and identify areas with the same conditions in the quaternary catchments were subsequently selected. The key findings are that livestock prefer accessible areas with gentle terrain near water sources, avoiding south-facing slopes which receive less solar radiation and tend to be cooler. Livestock are attracted to vegetation in riparian zones. Rural communal lands are dominated by poverty, and land-based livelihood strategies can potentially contribute to the well-being of the community. Therefore, understanding livestock distribution can contribute to a rangeland management strategy aimed at improving range condition which could increase livestock productivity and contribute to the livelihoods of local people