This research is carried out by Timothy Chambers1 with funding support from the Global Disaster Preparedness Center.
Extreme temperatures have severe impacts on human health. In a climate that is projected to significantly warm in future, these impacts become exacerbated. Southern Africa in particular is a hotspot for future warming and combined with large populations of vulnerable people, it is at great risk to heat events. There are however ways of reducing exposure and adapting to or mitigating the effects of this heat, but these require extensive resources. To identify and prioritise those at highest risk, this study maps and models the temperature distributions in 10 hot South African cities at a city-level scale and tests two potential cooling scenarios.
Landsat 8 imagery is used to produce high resolution Land Surface Temperature (LST) maps across the cities on hot days (above 30℃ air temperature). These identify significant (6-10℃) LST differences between parts of the cities, with this strongly related to the land cover type, socioeconomic status and historical settlement planning in the area. Wealthier and historically “white” areas tend to be located in more desirable positions and have far greater tree cover and cooler LST’s. Poorer, less developed areas are often located in more exposed positions with far less tree cover and warmer LST’s. Strong trends of study areas being cooler than their rural, less vegetated surroundings are also seen across the board.
The Weather Research and forecasting model (WRF) was implemented to improve the versatility of the research and test cooling scenarios. This was combined with an Urban Canopy model (UCM) as well as a new high resolution landcover dataset- converted from the 2020 South African landcover (SANLC).The simulations show good skill in reproducing observed spatial patterns and values in both surface and air temperatures, with LST’s much improved under the new landcover. To test the effectiveness of potential cooling strategies, two realistic scenarios were implemented into the model a) increasing tree coverage to 40% across the urban area and b) increasing the albedo of all building roofs to 0.4. Results for both scenarios show reductions in urban LST’s of ~2-3℃ and minor reductions in air temperatures (up to 1℃).
1 Department of Environmental and Geographical Science and Atmospheric science at the University of Cape Town