A Hybrid Framework for Evaluating Outdoor Thermal Comfort in Urban Landscapes

Over half of the world's population currently lives in urban areas. It is projected that the urban population ratio will rise 68% by 2050. This urbanization process has notably led to the urban heat island. Given the challenges posed by rapid urbanization and extensive heat stress, outdoor thermal comfort has become crucial for urban livability. However, there is a lack of field survey-based research on large-scale thermal comfort assessment across continuous urban spaces. To address this gap, this study developed a hybrid framework for assessing outdoor thermal comfort. A total number of 668 onsite observations from field studies during the daytime on typical summer days were collected and used for model development. The sites were distributed in diverse local climate zones (LCZs), enabling the prediction of outdoor thermal comfort across the city under different urban settings. A neural network model was trained for predicting daytime outdoor thermal comfort based on both meteorological and morphological variables. The model was then applied to wider urban layouts and dynamic climatic conditions.The results revealed that during extreme hot conditions, approximately 74.8% of areas experienced strong to extreme heat stress, with thermal sensations classified as hot or very hot, while the remaining 25.3% fell under moderate heat stress. High levels of thermal stress were observed in urban layouts of low-rise buildings, with LCZ 3 showing the highest extreme heat stress percentage at 61.3%, followed closely by LCZ 6 at 57.6%. These findings are crucial for identifying urban regions with high thermal stress. The neural network structure also demonstrated accurate estimations for outdoor thermal comfort, with R2 values of 0.95 in the full dataset for the Universal Thermal Comfort Index (UTCI) predictions. The framework could be valuable for cities with similar climates and geographical contexts.