Quantifying intra-urban socio-economic and environmental vulnerability to extreme heat events in Johannesburg, South Africa

TL;DR

This study assesses how extreme heat affects different areas of Johannesburg, showing that historical urban planning and poverty increase heat vulnerability.

Contribution

The study introduces a framework combining environmental and socio-economic data to quantify heat vulnerability in rapidly urbanizing Global South cities.

Findings

Historically disadvantaged areas like Alexandra Township show the highest heat vulnerability due to limited healthcare access and extreme heat exposure.

Higher poverty levels correlate with increased land surface temperatures (LST), indicating a socio-environmental link to heat vulnerability.

Northern suburbs have lower vulnerability due to better healthcare access and higher vegetation coverage.

Abstract

Urban populations face increasing vulnerability to extreme heat events, particularly in rapidly urbanising Global South cities where environmental exposure intersects with socioeconomic inequality and limited healthcare access. This study quantifies heat vulnerability across Johannesburg, South Africa, by integrating high-resolution environmental data with socio-economic and health metrics across 135 urban wards. We examine how historical urban development patterns influence contemporary vulnerability distributions using principal component analysis and spatial statistics. Environmental indicators (Land Surface Temperature (LST), vegetation indices, and thermal field variance) were combined with socioeconomic and health variables (including indicators on crowded dwellings and healthcare access, self-reporting of chronic diseases) in a comprehensive vulnerability assessment. Principal Component Analysis revealed three primary dimensions explaining 56.6% (95% CI: 52.4–60.8%) of the total variance: urban heat exposure (31.5%), health status (12.8%), and socio-economic conditions (12.3%). Built-up areas showed weak but significant correlations with heat indices (ρ = 0.28, p < 0.01), while higher poverty levels demonstrated moderate positive correlations with LST (ρ = 0.41, p < 0.001). The spatial analysis identified significant clustering of vulnerability (Global Moran's I = 0.42, p < 0.001), with distinct high-vulnerability clusters in historically disadvantaged areas. Alexandra Township showed the highest vulnerability(HVI score: 0.87, LST: 29.8°C ± 0.4°C, NDVI: 0.08 ± 0.02), with factors characterising the high vulnerability in that area including limited healthcare access and extreme heat exposure. Northern suburbs formed a significant low-vulnerability cluster (Mean HVI = 0.23 ± 0.07, p < 0.001), benefiting from greater vegetation coverage and better healthcare access. These findings demonstrate how historical planning decisions continue to shape contemporary environmental health risks, with vulnerability concentrated in areas of limited healthcare access and high extreme heat exposure. Results suggest the need for targeted interventions that address both environmental and social dimensions of heat vulnerability, particularly focusing on expanding healthcare access in identified hotspots and implementing community-scale green infrastructure in high-risk areas. This study provides an evidence-based framework for prioritising heat-resilience initiatives in rapidly urbanising Global South cities while highlighting the importance of addressing historical inequities in urban adaptation planning.