Abstract
Three-dimensional structural diversity (3SD) directly influences the distribution and flow of heat within the canopy. However, the nonlinear effects of 3SD of different species on the cooling effects remain unclear. Here, we proposed an analytical framework to explore this relationship at the single tree and community scales. Results indicated that: (1) A benchmark dataset for individual tree segmentation was established, with the best-performing algorithm achieving an accuracy of 77.36% (F-score=0.75), the UAV-based LiDAR, multispectral and thermal infrared imagery using a data fusion approach achieved a better species classification accuracy of 80.41% (kappa=0.78); (2) At the single tree scale, the cooling effects are controlled by vertical structure, heterogeneity, and leaf density (15.36%<rel.inf<26.84%). Entropy, VAI, and Hmax exhibited the largest seasonal relative importance change rates (7%<|Δrel.inf|<11%); (3) At the community scale (10m × 10m), VAI contributed the most to coniferous cooling in summer, while Hmax had the greatest impact on broadleaf cooling in winter. Species’ spatial connectivity had a significantly greater impact on the cooling effects in broadleaf in summer and coniferous in winter compared to structural diversity. This study supports optimizing urban forestry by demonstrating UAV-based data fusion for species classification and highlighting structural diversity's role in regulating temperature across scales and seasons.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call