Exploring heatwaves and urban heat islands: implications for water wesource availability in Gauteng Province, South Africa
Loading...
Date
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
University of the Witwatersrand, Johannesburg
Abstract
Surface water is a scarce resource in Gauteng, which leads to diverse research in analysing possible implications on water resources in the region. Further, heatwaves and urban heat islands are a prevalent topic in urban studies in an attempt to understand and curb urban heat stress in the future. The study analyses heat and water indices along vegetation and land use land cover dynamics to uncover spatio-temporal trends in Gauteng. Using Landsat 5TM and Landsat 8 OLI/TIRS at 30m resolution with NASA POWER over a period of 4 decades. This uncovered trends in urbanisation, surface water and urban heat. Maximum air temperature and NDVI displayed an increasing trend according to the Mann-Kendall analysis, with the water ratio index and precipitation decreasing over the years. Additionally, MNDWI and LST obtain a negative Spearman correlation of ρ = –0.40, p < 0.001. Similarly, NDVI and WRI observe a negative correlation of ρ = –0.69, reflecting the inverse spectral response of vegetation and water surfaces as detected by the sensors. Moreover, heatwaves in Gauteng have increased gradually, with average frequency being ~1 from 1984 to 2009; however, from 2014 to 2024, the average frequency is ~4. While the average UHII for Gauteng in 1984 was –2.18 °C, it increased to 0.25°C in 2024. This further highlights a negative result of UHII, indicating a cooling response in 1984 as opposed to the positive result in 2024, indicating a warming sensation in Gauteng. The mean land surface temperature of vegetated areas exhibits a decreasing trend over time, corresponding with increasing NDVI trends in the region. The increase of NDVI by anthropogenic greening and cultivated land, with the increase of heatwave frequency and duration, contributes to thermal surface water stress. The paper makes use of random forest classification to depict LULC trends over 40 years. The overall accuracy was ~ 96% while Kappa coefficients remained above 90% in general. However, the confusion matrix indicates comparatively higher misclassification among v waterbodies, grasslands, and barren land classes. This confusion is often attributed to spectral similarity in multispectral remote sensing data. This study provides crucial insights into converging indicators of heat stress, as depicted by the increasing LST with decreasing WRI and precipitation. Additionally, the positive correlation between LST and WRI offers strong evidence that water resources in Gauteng are experiencing heat stress. Therefore, the region is vulnerable to climate-land cover interactions.
Description
A research report submitted in fulfillment of the requirements for the Master of Science, in the Faculty of Science, School of Geography, Archaeology and Environmental Studies, University of the Witwatersrand, Johannesburg, 2025
Citation
Kuwaza, Tamiranashe. (2025). Exploring heatwaves and urban heat islands: implications for water wesource availability in Gauteng Province, South Africa [Master’s dissertation PhD thesis, University of the Witwatersrand, Johannesburg]. WIReDSpace. https://hdl.handle.net/10539/50288