Beyond flat cities: dramatic 2D/3D differences in urban gray–green space configurations and the nonlinear thermal regulation mechanisms

Le résumé fourni par la source

Traditional 2D morphological indicators are inadequate for characterizing vertically expanding cities, which hampers the identification of key predictors of urban heat under extreme temperatures and within the complex interactions between gray and green spaces. Therefore, we developed a suite of multidimensional morphological indices and leveraged an XGBoost-SHAP framework to decipher the complex, nonlinear relationships and synergistic interactions between these indices and land surface temperature (LST). We selected Xi'an, the only megacity in Northwest China, as a case study and introduced Lanzhou as a case for verifying transferability. The findings revealed the following: (1) Within the Xi'an Ring Expressway, 73.84% of the blocks underestimated the relationship involving 3D gray and green space configuration, while this proportion was higher in the Lanzhou main urban area, and the underestimated regions were concentrated in the core area of high-density cities. Spatial variations in 2D and 3D gray–green ratios (GGR) effectively delineate urban functional zoning. (2) Counterintuitively, average building height (ABH) emerged as the most critical factor influencing LST. At the same time, 3D indicators contributed 1.86 times more to LST changes than 2D indicators. (3) The complex nonlinear correlation between urban spatial morphology and LST is exhibited, and the marginal effect results reveal the key threshold of urban spatial morphology indexes affecting LST, 3D GGR is more sensitive than the 2D influence on LST. Our findings reveal that the relationship between urban 3D gray–green space allocation in high-density urban areas is generally underestimated, confirming the significant advantages of three-dimensional indicators in addressing urban thermal environment problems and providing a transferable framework for proposing targeted strategies.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.