Discovering directional relationships between urban morphology and thermal environment using an anisotropic 3D urban texture model

Abstract

Urban morphology generally refers to the structure and layout of urban space and is often characterized by urban texture. Rapid urbanization has changed the shape of cities and created environmental challenges, especially in urban thermal environments. However, the directional relationship between urban morphology and the thermal environment has yet to be fully explored. Existing studies often rely on 2D models, which fail to capture the complexity of 3D urban morphology. Moreover, the influence of urban thermal anisotropy on this relationship is insufficiently quantified. To address these, this study proposes an anisotropic 3D urban texture model which divides surrounding space into directional sectors and generating distinct texture curves representing variations in 3D building density. The model was applied in Wuhan, China, to analyze the relationship between urban morphology and land surface temperature (LST) using random forest regression. The fitting accuracy (R²) of the proposed anisotropic model reached 0.725, which was 13.46 % higher than the original isotropic model. Furthermore, the study evaluated the contributions of 3D texture features from different orientations to LST. The proposed model provides a quantitative tool for understanding the directional influences of 3D urban morphology on environmental issues like temperature and air quality.