Within a multi-scale analytical framework, researchers employed a random forest (RF) regression model to examine how urban landscape components in Hefei shape seasonal land surface temperature (LST) fluctuations. Outputs from this model reveal that population density, road network density, and building configurations act as core drivers of urban LST variations - these drivers' influences show marked seasonal shifts and intricate nonlinear traits, with road network density having a defined threshold range of 5-7 km/km2. Shifts in spatial resolution notably modify how landscape components regulate LST; at a 300 m spatial resolution, the combined 2D-3D model delivers the strongest explanatory capacity. Such insights supply critical quantitative direction for easing urban heat island impacts via targeted urban planning and landscape configuration.
Liu et al. (Thu,) studied this question.