Evaluating the impact of land use and land cover change on unprotected wetland ecosystems in the arid-tropical areas of South Africa using the Landsat dataset and support vector machine

The study explored the impact of Land Use and Land Cover (LULC) change dynamics in relation to the condition and status of an unprotected wetland located in the arid-tropical parts of the Limpopo Province, South Africa. The long-term Landsat archival data series was used to map and quantify the impacts of LULC change on the wetland over a period of 36 years (1983-2019). A multi-source satellite image analysis was performed, using the support vector machine (SVM) algorithm and advanced spatially- explicit geographic information system tools. Landsat data series covering the entire study area was used to assess, map and monitor LULC change that occurred over-time. Post-classification maps for the Maungani wetland area were analysed to provide a quantitative assessment and a detailed overview of the rate of change. The generated wetland detection maps for four temporal phases (i.e., 1983-1992, 1992-2001, 2002-2010) were analysed. This study found that the spatial extent of the wetland area declined severely during the period under study with 728 300 ha. The findings of this work provide critical insights and baseline information about the state of unprotected wetlands in the rural parts. This information is useful for the development of tailor-made wetland management strategies and a possible rehabilitation framework for unprotected wetland ecosystems.