Unveiling the cooling effects of urban green spaces in Eastern China: spatial heterogeneity, diurnal asymmetry, and nonlinear margins of 2D/3D features

Abstract

Urban green spaces (UGS) play an important role in mitigating urban heat island effects. However, the mechanisms underlying their cooling effects are still not fully understood, particularly regarding the role of three-dimensional vegetation structure and nonlinear threshold responses. This study employed a multidimensional analytical framework integrating two-dimensional vegetation cover, three-dimensional canopy structure, and landscape configuration to investigate UGS cooling effects across nine representative Chinese cities spanning subtropical to semi-arid temperate climates. Multi-source remote sensing datasets, including Landsat-8, ECOSTRESS, and high-resolution canopy height data, were integrated with an interpretable machine learning approach (XGBoost–SHAP) to quantify the spatial heterogeneity, diurnal asymmetry, and nonlinear responses of UGS cooling. Compared with conventional linear models, this framework was more effective in capturing nonlinear relationships and quantifying the relative contributions of individual predictors. The results indicated notable nonlinear threshold effects associated with vegetation structure. Cooling intensity increased rapidly when vegetation cover exceeded approximately 0.33 and gradually approached saturation near 0.65. Mean canopy height also exhibited a threshold effect around 6.33 m, above which cooling intensity generally increased across cities. Clear diurnal asymmetry was observed. Daytime cooling was stronger and mainly governed by vegetation cover and canopy height, whereas nighttime cooling was weaker and more strongly associated with landscape configuration metrics. In addition, southern cities generally exhibited stronger cooling effects (mean: 1.05 °C ± 0.52 °C) than northern cities (mean: 0.81 °C ± 0.25 °C). These findings provide new insights into the nonlinear and spatially heterogeneous mechanisms underlying UGS cooling and offer methodological support for climate-adaptive urban green infrastructure planning and urban thermal environment management.