Urban forest tree species and 3D structure seasonal contributions on spatial sensible heat exchange at individual tree and cluster scales

Abstract

Urban forests tree species and 3D structure of urban forests directly shape SHF dynamics and indirectly govern spatial heat distribution. However, the multi-seasonal impact of species and structure on SHF remains unclear, hindering our understanding of urban forest energy flow. To address this, a multi-source data framework was established by studying 72 typical urban forest communities with UAV-based thermal, LiDAR, multispectral sensors and portable weather stations. By integrating Random Forest and Pareto algorithms, we identified key tree species and structural factors. Their influence on seasonal SHF dynamics was then quantified at both individual and cluster scales through marginal effect and partial correlation analysis. Results show: (1) UAV and field data together precisely mapped SHF and its clear seasonal patterns in small-to-medium urban forests. (2) At the individual tree scale, mongolian pine, chinese pine, and larch most influenced SHF in summer, while manchurian walnut and mongolian oak dominated in other seasons. Tree trait-SHF relationships varied seasonally: in spring and autumn, East-West Crown Spread (P-E) > Tree Height (H) > Diameter at Breast Height (DBH), Clear Bole Height (CBH) > North-South Crown Spread (P-N). In summer and winter, P-N > DBH > H > P-E > CBH. (3) At the cluster scale, broadleaf and mixed forests outperformed conifers in SHF regulation, with GFP and Hmax as the key 3D drivers. This study offers key insights into urban forest-SHF dynamics and provides seasonal guidance for managing urban heat flows, supporting future optimization strategies.