Heat stress prediction in Glasgow: Integration of historical data with Machine Learning models

Recurring heat stress has increasingly detrimental effects on human health and well-being. To address this challenge, it is essential to identify the key parameters of urban heat derived from available datasets that contribute to its intensification for supporting urban heat mitigation. Despite the value of technological advancements, data accessibility, and in-depth research aimed at mitigating heat stress and forecasting urban climate, they often fall short in effectively adapting to climate change through spatial planning. This study investigates the usability of key heat stress parameters derived from secondary and historical datasets by employing machine learning algorithms. The developed model predicts Land Surface Temperature (LST), a proxy for heat stress, under different urban greening scenarios on vacant lands at a city scale. The findings underscore the limited yet significant role of urban greenery in mitigating thermal stress, particularly in relation to the diverse characteristics of vacant land. For example, the influence of shading provided by vegetation and buildings can significantly affect thermal comfort, depending on the compactness or openness of areas within Glasgow's central district. Additionally, the new approach highlights opportunities for improving data collection, organization, and public accessibility, which could support urban planners and decision-makers in developing more effective strategies for mitigating urban heat.