Divergent impact of urban 2D/3D morphology on thermal environment along urban gradients

Optimizing urban morphology is an effective way to alleviate the thermal environment, and it has always been focus of planning. Although previous scholars have explored impact 2D/3D on Land Surface Temperature (LST), they rarely revealed differences along gradients, thus ignoring nonlinear relationship between LST. In this study, we use multi-source data such as night-time lights, Landsat 8, land cover, buildings, employ dynamic threshold, single-window algorithm, boosted regression trees methods reveal gradient marginal effects Results showed that proportion built-up (PCL), digital elevation model (DEM) normalized difference index (NDBI) important contribution Moreover, relative influence PCL, NDBI building density (BD) LST varies significantly which are +25.87%, +14.85% − 10.88%, respectively. Interestingly, found value each variable increased, its changed. For example, when PCL < ∼49%, negatively correlated with LST, so lowering can environment. Our findings inform sound practice planning by accounting for