Balancing thermal comfort and pollutant mitigation in street canyons: a multi-objective optimization of tree and building configurations.

In urban street canyons, pedestrian health is critically affected by the interplay of thermal stress and pollutant exposure. While urban morphology and vegetation can mitigate these adverse effects, their interactions are complex and often lead to design trade-offs. This paper develops a comprehensive framework to navigate these challenges. We use computational fluid dynamics (CFD) to model a street canyon under varying aspect ratios (AR), local solar times (LSTs), tree canopy sizes (Sc), and leaf area densities (LADs). To holistically assess environmental quality, we introduce a novel Microclimate Environmental Quality Index (MEQI), which integrates the Universal Thermal Climate Index (UTCI) and pollutant concentration (K). A surrogate model based on a Long Short-Term Memory (LSTM) network is coupled with a Multi-Objective Grey Wolf Optimizer (MOGWO) to efficiently search for optimal design configurations. The results reveal a significant and non-linear trade-off between thermal comfort and air quality, highlighting that the environmental performance of vegetation is highly context-dependent. Notably, the cooling benefits of trees can be counteracted by their negative impact on pollutant dispersion, particularly under strong solar radiation. This study presents a robust and transferable optimization framework capable of identifying Pareto-optimal solutions tailored to specific local conditions. This approach provides a powerful decision-support tool for urban designers and planners to quantitatively balance competing environmental objectives and develop context-specific, high-performance streetscape designs.