Enhancing the cooling effectiveness of green spaces (GSs) is crucial for improving urban thermal environments in the context of global warming. Increasing GS coverage and optimizing its spatial distribution individually proved to be effective urban cooling measures. However, their comparative cooling effectiveness and potential interaction remain unclear. Here, using the moving window approach and random forest algorithm, we established a robust model (R2 = 0.89 ± 0.01) to explore the relationship between GS and land surface temperature (LST) in the Chinese megacity of Guangzhou. Subsequently, the response of LST to varying GS coverage and its spatial distribution was simulated, both individually and in combination. The results indicate that GS with higher coverage and more equitable spatial distribution is conducive to urban heat mitigation. Increasing GS coverage was found to lower the city's average LST by up to 4.73 °C, while optimizing GS spatial distribution led to a decrease of 1.06 °C. Meanwhile, a synergistic cooling effect was observed when combining both measures, resulting in additional cooling benefits (0.034-0.341 °C). These findings provide valuable insights into the cooling potential of GS and crucial guidance for urban green planning aimed at heat mitigation in cities.
Keywords: Random Forest model; Urban heat mitigation; cooling effect; green space coverage; spatial distribution; synergistic effect.