Impact of 2D/3D Landscape Pattern on Urban Ecological Environment Quality: Explainable Machine Learning Methods

It is increasingly important to consider ecological environmental quality (EEQ) in urban design. However, urban expansion is increasingly vertical as well as horizontal. This paper proposes a new set of 2D/3D landscape morphology indicators and a new model for quantifying urban EEQ and applies these within a novel machine learning framework. The preliminary results show that: (1) there is a correlation between landscape morphology indicators and EEQ; (2) the proposed new model has improved accuracy and robustness; (3) a threshold for the landscape morphology indicators can be identified; (4) there are seasonal and scale effects. The above findings may help optimize urban spatial structure, and improve the EEQ in urban areas.