Interacting effects of tree-forest structure and terrain on heavy rainfall-induced tree destruction

Rainstorm events are becoming increasingly frequent due to the impacts of global warming, which results in widespread erosion and associated tree destruction. However, previous studies of forest damage have focused on typhoons or wildfires, largely overlooking the increasing risk of tree destruction caused by rainstorm-induced erosion. It is unclear what scale of tree destruction can be caused by heavy rainfall. In this study, we used a tree segmentation method based on airborne light detection and ranging (LiDAR) technology to accurately quantify the tree destruction caused by heavy rainfall in a representative afforested catchment on the Chinese Loess Plateau. We evaluate the respective and combined contributions of tree structure (tree height, crown diameter, and crown area), forest structure (tree density, gap fraction, leaf area index, and canopy cover), and terrain parameters (elevation, slope, and terrain relief) using machine learning models (random forest and logistic regression). The results show that 3,253 trees in the catchment (0.9 km²) were destroyed due to rainstorm-induced erosion, of which 2,845 trees were located on gully slope landforms, accounting for 87.4% of all destroyed trees. Tree destruction was primarily induced by erosion on steep slopes (45.5°–50.5°) and by sediment deposition along the gully bed. Although the total deposition area (21,265 m²) that resulted in tree destruction exceeded the erosion area (20,020 m²), erosion was more destructive. Importantly, the interaction between increased tree structural parameters and higher canopy density (leaf area index and canopy cover) significantly promoted destruction, likely because the combined biomass and canopy weight increase mechanical load on saturated soils, which can counteract the inhibitory effect of terrain on destruction. This synergy also raises destruction probability at similar elevations. Our study provides a replicable methodology for assessing forest damage under extreme rainfall and highlights the need to avoid overly dense afforestation in vulnerable landscapes. The study underscores the need for improved climate-resilient reforestation strategies that consider both structural and topographic interactions in erosion-prone landscapes.