Optimizing urban green space cooling with a multi-objective NSGA-II framework

Using urban green spaces(UGS) to mitigate the urban heat island effect is an effective nature-based strategy. However, previous studies often relied on single-objective threshold optimization when analyzing regulatory thresholds, leading to conflicting optimization objectives. Critically, there is a lack of a systematic framework to reconcile these trade-offs, forcing planners to rely on rigid efficiency thresholds that ignore spatial coverage needs. To solve this problem, we developed a Pareto front–based optimization model to balance multiple objectives for UGS optimization. Ninety UGS patches in Urumqi were selected for analysis. Then, we analyzed the importance differences between patch properties and environmental factors in explaining the cooling effect. The results showed that patch area (PA) and landscape shape index (LSI) dominated cooling area (CA) and cooling efficiency (CE), respectively, while the proportion of external bare soil (BAREout) was an important variable affecting cooling intensity (CI) and cooling gradient (CG). More importantly, we emphasized the need to balance multiple cooling objectives, as the existing threshold value of efficiency (TVoE) only considers the maximization of cooling efficiency. Our model revealed that a “small area with high complexity” configuration strategy serves as an effective approach to achieving balance among cooling objectives. The final optimal solution required only 11.8 ha, which is much lower than the traditional efficiency threshold. This work opens up a new avenue for more effective urban heat mitigation and provides scientific support for UGS planning.