SIGNATURE WORK
CONFERENCE & EXHIBITION 2024

The urban heat island effect intensifies heat stress within city environments. Strategies like cool roofs and urban greening are advocated to mitigate this effect, but assessing their effectiveness is complex, as temperature reduction outcomes are influenced by local features. Traditional eval- uation methods, such as computational fluid dynamics and urban canopy models, demand con- siderable computational resources and expert knowledge. We introduce a data-driven method, fueled by recent advancements in spatial causal inference, to simplify this analysis. This method enables the evaluation of potential interventions aimed at lessening the urban heat island impact. Our methodology is illustrated through a case study in Raleigh, North Carolina, where we use accessible data on air temperature, land cover, and satellite imagery to model evening tempera- tures. Our analysis suggests that interventions like tree-lined streets, cool roofs, and converting parking lots into green spaces could reduce evening temperatures by up to 0.7-0.9°C, noting a decrease in temperature impact with increasing distance from the intervention site. Given the straightforward access to the required data, our approach can be adapted for other U.S. cities, offering them tailored strategies to alleviate urban heat stress.