Multifunctional green infrastructure, a key component of compact sustainable cities, is challenged by the pressures associated with rapid urbanization. In this paper, we present a method that uses remote sensing, GIS modeling and stakeholder engagement to produce a decision support tool that communicates the availability and need for green infrastructure benefits. The case study presented is the City of Tshwane, South Africa, a Global South city facing rapid urbanization. We found that this method of mapping green infrastructure benefits can provide simultaneous oversight on multiple objectives for green infrastructure, including climate change adaptation, biodiversity, and equitable distribution of urban green space. We found that low-scoring benefit areas occur in dense urban areas where small-scale nature-based solutions or rehabilitation activities are required. Moderate benefit scores occurred in parts of the city that are vulnerable to urban expansion and densification activities, warranting the careful planning of green infrastructure provision, and that moderate-to-high-scoring areas can be protected as conservation areas. The results are discussed in terms of the role of decision support tools for urban planning practice. Composite indexes can provide important guidance to decision-makers involved in spatial planning and urban upgrading and expansion activities.

The settled landscape in the Drakensberg Mountain region of South Africa is characterized by increasing urbanization. Some of the supposedly rural settlements in the region have experienced increasing change in their landscapes over the last 3 decades, resulting in significant land use and land cover (LULC) changes. Among such settlements, Phuthaditjhaba and its environs are slowly becoming a metropolitan area. Based on conceptual considerations regarding sustainable urban development, we assessed LULC change using 4 Landsat images from 1989, 1999, 2009, and 2019 and a combination of unsupervised and supervised classification methods. The images were classified into 4 LULC classes. Between 1989 and 2019, the urban built-up area in Phuthaditjhaba increased from about 5% to 19%, representing a total increase of 270%. However, the greatest increase in land cover was in bare surface at the expense of vegetated areas, including farmland, which decreased from about 45% to 15%. The increase in bare surface could be due to fires. Built-up areas also increased due to a consistent increase in population density in the study area. We further described spatial patterns in LULC using selected landscape metrics. A decrease in patch density (PD) and cohesion, coupled with constant edge density (ED) and an increase in the fractal dimensional index (FDI), indicates fragmentation and less connectivity between 1989 and 1999; we interpret this as a sign of unsustainability. An increase in PD and cohesion and fluctuations in ED and FDI show that land patterns were more aggregated between 2009 and 2019. At the class level, an increase in PD, cohesion, and ED also showed more aggregated land patterns, which was confirmed by the mean patch size. The FDI revealed greater connectivity, which we also interpreted as unsustainable because of the increase in bare surface and built-up areas. Integrative coplanning and comanagement of land use and allocation are needed to ensure sustainable development.

Spatial patterns of human settlements, their changes and their geographical implications are important for understanding the drivers of land use and land cover change. This paper examines the spatiotemporal pattern of settlement development in three different settlements of Thabo Mofutsanyane municipal district namely Harrismith, Vrede and Ladybrand using GIS, remote sensing and spatial metrics techniques. The study is based on 30 years of time-series data compiled from satellite images with emphasis on pre and post 1994’s (the year of change in government from apartheid to majority rule) spatial change in settlement development. Landsat 4–5 imageries for 1989, 1999, 2009 and Landsat 8 OLI for 2018 were downloaded and classified for land use/land cover change (LULCC). Also, landscape and class metrics were computed using Fragstats QGIS 2.18.9 to generate spatial analysis. A dynamic spatial pattern is observed in the settlements under study. Urban built-up areas had a rapid trend of growth in Harrismith from 1989 (5 years before transition from apartheid to majority rule) to 1999 (5 years after transition from apartheid to majority rule) but later slowed down in the subsequent years under study while Ladybrand and Vrede had moderate growth trend in the subsequent years under study. The settlement development process has developed fragmented and heterogeneous land use combinations in the years after 1994. At landscape level, land fragmentation occurred due to land use changes and significant urban expansion; Ladybrand experienced more physical connectedness than Harrismith and Vrede. While at class level, Harrismith and Vrede are more aggregated or physically connected than Ladybrand; this means Ladybrand was relatively more fragmented than the other two settlements. The study results show that a LULCC and landscape metrics integrated approach is effective to analyse and describe the spatial patterns of urban landscapes.