The paper evaluated the Landsat Automated Land Cover Update Mapping (LALCUM) system designed to rapidly update a land cover map to a desired nominal year using a pre-existing reference land cover map. The system uses the Iteratively Reweighted Multivariate Alteration Detection (IRMAD) to identify areas of change and no change. The system then automatically generates large amounts of training samples (n > 1 million) in the no-change areas as input to an optimized Random Forest classifier. Experiments were conducted in the KwaZulu-Natal Province of South Africa using a reference land cover map from 2008, a change mask between 2008 and 2011 and Landsat ETM+ data for 2011. The entire system took 9.5 h to process. We expected that the use of the change mask would improve classification accuracy by reducing the number of mislabeled training data caused by land cover change between 2008 and 2011. However, this was not the case due to exceptional robustness of Random Forest classifier to mislabeled training samples. The system achieved an overall accuracy of 65%–67% using 22 detailed classes and 72%–74% using 12 aggregated national classes. “Water”, “Plantations”, “Plantations—clearfelled”, “Orchards—trees”, “Sugarcane”, “Built-up/dense settlement”, “Cultivation—Irrigated” and “Forest (indigenous)” had user’s accuracies above 70%. Other detailed classes (e.g., “Low density settlements”, “Mines and Quarries”, and “Cultivation, subsistence, drylands”) which are required for operational, provincial-scale land use planning and are usually mapped using manual image interpretation, could not be mapped using Landsat spectral data alone. However, the system was able to map the 12 national classes, at a sufficiently high level of accuracy for national scale land cover monitoring. This update approach and the highly automated, scalable LALCUM system can improve the efficiency and update rate of regional land cover mapping.

Land-cover change and habitat loss are widely recognised as the major drivers of biodiversity loss in the world. Land-cover maps derived from satellite imagery provide useful tools for monitoring land-use and land-cover change. KwaZulu-Natal, a populous yet biodiversity-rich province in South Africa, is one of the first provinces to produce a set of three directly comparable land-cover maps (2005, 2008 and 2011). These maps were used to investigate systematic land-cover changes occurring in the province with a focus on biodiversity conservation. The Intensity Analysis framework was used for the analysis as this quantitative hierarchical method addresses shortcomings of other established land-cover change analyses. In only 6 years (2005–2011), a massive 7.6% of the natural habitat of the province was lost to anthropogenic transformation of the landscape. The major drivers of habitat loss were agriculture, timber plantations, the built environment, dams and mines. Categorical swapping formed a significant part of landscape change, including a return from anthropogenic categories to secondary vegetation, which we suggest should be tracked in analyses. Longer-term rates of habitat loss were determined using additional land-cover maps (1994, 2000). An average of 1.2% of the natural landscape has been transformed per annum since 1994. Apart from the direct loss of natural habitat, the anthropogenically transformed land covers all pose additional negative impacts for biodiversity remaining in these or surrounding areas. A target of no more than 50% of habitat loss should be adopted to adequately conserve biodiversity in the province. Our analysis provides the first provincial assessment of the rate of loss of natural habitat and may be used to fulfil incomplete criteria used in the identification of Threatened Terrestrial Ecosystems, and to report on the Convention on Biological Diversity targets on rates of natural habitat loss.

The loss of natural habitat resulting from human activities is the principal driver of biodiversity loss in terrestrial ecosystems globally. Metrics of habitat loss are monitored at national and global scales using various remote sensing based land-cover change products. The metrics go on to inform reporting processes, biodiversity assessments, land-use decision-making and strategic planning in the environmental and conservation sector. We present key metrics of habitat loss across South Africa at national and biome levels for the first time. We discuss the spatial patterns and trends, and the implications and limitations of the metrics. Approximately 22% of the natural habitat of South Africa has been lost since the arrival of European settlers. The extent and the rate of habitat loss are not uniform across South Africa. The relatively mesic Grassland, Fynbos and Indian Ocean Coastal Belt biomes have lost the most habitat, while the arid Nama-Karoo, Succulent Karoo and Desert have lost the least. Rates of loss increased across all biomes in recent years (2014–2018), indicating that the historical drivers of change (i.e. expansion of croplands, human settlements, plantation forestry and mining) are intensifying overall. We should caution that the losses we report are conservative, because the land-cover change products do not capture degradation within natural ecosystems. Preventing widespread biodiversity losses and securing the benefits we derive from biodiversity requires slowing and preventing further habitat degradation and loss by using existing land-use planning and regulatory tools to their full potential. Significance: The loss of natural habitat resulting from human activities is the principal driver of biodiversity loss in terrestrial ecosystems in South Africa. Monitoring trends and patterns of habitat loss at a national scale provides a basis for informed environmental decision-making and planning, thus equipping civil society and government to address habitat loss and protect biodiversity while also meeting key development and socio-economic needs. Open data set: https://doi.org/10.15493/SAEON.FYNBOS.10000011