Jue Wang, Zhewei Liu
Urban heat exposure poses significant risks to human well-being and contributes to emerging social inequity. While previous research has addressed urban heat exposure, mitigating the inequity associated with such hazards remains a significant challenge and an understudied area. To address this issue, this study proposes a machine-learning approach to assess the effectiveness of various urban development strategies in mitigating urban heat exposure inequity. The results from real-world datasets demonstrate that our approach effectively quantifies urban heat inequity, and the proposed machine learning model can accurately capture urban heat exposure based on various urban environmental variables. The season-specific simulation is further conducted by evaluating the outcomes of various urban development strategies across different seasons. The simulation results show that while modifying certain built features may help mitigate heat exposure inequity based on annual data, their effects are significantly more pronounced in winter. In contrast, improving natural features yields more substantial effects in summer and consequently serves as an optimal strategy for mitigating urban heat exposure inequity, which may be overlooked when relying solely on annual data. Compared with prior research that primarily focuses on hazard exposure itself, this study advances the field by addressing the urban heat inequity through a novel and effective machine-learning framework. The proposed methodology not only offers a robust framework for evaluating mitigation strategies but also leverages a season-specific perspective to uncover inefficiencies in approaches that may seem effective in annual analyses yet prove inadequate in critical seasons. By recognizing these seasonal variations, this study enables the development of targeted and cost-effective strategies that optimize urban planning interventions, ultimately reducing urban heat exposure and promoting social equity.