Heat index algorithms and heat stress hazard in Limpopo Province, South Africa and Guangdong Province, China

Anthropogenic climate change has increased the frequency and magnitude of heatwaves globally, and projections indicate this trend will continue. Adequately quantifying heat stress hazard is essential for understanding human health impacts and informing adaptation efforts, yet no universally accepted metric exists, and commonly used metrics differ substantially in their scientific foundations and behaviour under extreme temperature-humidity conditions. This study addresses this gap by analysing the performance of the novel Extended Heat Index (EHI) relative to eighteen widely used heat stress algorithms across two contrasting regions: the hot-dry climate of Limpopo Province, South Africa, and the hot-humid climate of Guangdong Province, China, and assesses how algorithm design influences the magnitude, spatial distribution, and sensitivity of the estimates. Although all indices reproduce broad spatial structures, their responses diverge sharply under upper-tail conditions, particularly in humid environments. The EHI shows stable, monotonic behaviour across extremes, avoiding non-physical artefacts present in several empirical formulations. Historical trends in EHI-derived heat stress hazard reveal little long-term change in Limpopo but a statistically significant increase across Guangdong, driven primarily by rising temperatures. These results demonstrate that methodological choice strongly shapes inferred heat stress hazard and highlight the need for further development and regional validation of physiologically informed metrics.