Implementing policies to mitigate urban heat islands: Analyzing urban development factors with an innovative machine learning approach

Abstract

Taiwan is located in a hot and humid subtropical climate. Urban development, building heat emissions, and human activities lead to the urban heat island (UHI) effect in which urban air temperatures are significantly higher than those in the surrounding suburban areas. In this study, the factors contributing to the UHI effect were investigated in a case study of Taichung City, Taiwan by using high-resolution geographic climate data. Geographic Information System tools were used for grid-based spatial data analysis, and a physiological equivalent temperature index map was calculated and overlaid to identify heat zones. The results revealed that environmental factors have complex, nonlinear effects on temperatures in in urban areas; hence, the analysis was challenging. The decision tree algorithm was used to identified key environmental factors contributing to urban heat in areas with different levels of development; a weighting analysis was then performed to select effective urban cooling strategies. Finally, backpropagation neural network models tailored to different development levels were used for simulations to validate the proposed cooling policies. The results of this study were incorporated into the urban planning policies of Taichung City, Taiwan to mitigate the UHI effect and create a more suitable urban living environment.