Impact of diurnal variation in 3D urban landscape metrics on land surface temperature in Shanghai: A local climate zone perspective

With rapid urbanization, cities have shifted from two-dimensional (2D) to three-dimensional (3D) expansion, presenting challenges in understanding how urban 3D structures impact ecological environments. This study uses ECOSTRESS LST data, 3D building and vegetation data, Spearman correlation analysis, and a boosted regression tree (BRT) model to examine the diurnal effects of 3D landscape metrics on LST across various LCZs in Shanghai's outer ring. The results show that a 1050 m spatial scale is optimal for analysis. Mean architecture height (MAH) was positively correlated with LST at 02:39, 05:42, and 09:27 but negatively at 12:58. At 12:58, the relative influence of mean architecture projection area (MAPA) was highest in LCZ 5 and LCZ 8. Larger building volumes (AVs) reduced LST at night, but increased it at 12:58, particularly in LCZs 4 and 8. The MAPA had the strongest positive effect on LST at 12:58, while mean architecture height standard (AHSD) had a cooling effect during the day but increased LST at night. This study offers new insights into the temporal impacts of 3D urban morphology on thermal conditions, providing actionable guidance for urban planning and heat island mitigation.