Local tree-tree interactions have been identified as a key mechanism underlying biodiversity-ecosystem functioning relationships. The complementary use of the canopy space by coexisting species increases canopy packing, thereby promoting community productivity. Several studies have applied the theoretical volumetric complementarity approach, using the crown complementarity index (CCI). Despite its effectiveness in quantifying vertical stratification, this approach lacks integration of distance between trees and the interacting crown volume. To address these limitations, we developed the crown overlap index (COI), which quantifies the volume of crown interactions within a specific distance threshold derived from 3D laser scanning point cloud data. We calculated CCI and COI for 115 tree pairs in the BEF-China experiment and found a negative correlation between the two indices. At high CCI values, COI was consistently low; at lower CCI values, however, it varied greatly. This suggests that complementarity cannot be quantified holistically by a single spatial index. We also found that trait dissimilarity and structural complexity between interacting trees strongly influenced canopy structure, increasing CCI and decreasing COI. This promoted vertical stratification, but limited crown overlap. These patterns are likely attributable to crown shyness, i.e., the tendency of trees to keep their crowns separate, with leaf trait dissimilarity reflecting differences in crown plasticity. This allows plastic deciduous tree species to avoid crown contact, while structural complexity dissimilarity enhances CCI and reduces COI through contrasting crown organisation. Overall, the crown overlap index provides a novel quantitative approach for assessing crown interactions and deepening our understanding of forest spatial structure.

The structural data was collected using a “FARO Focus 3D X130” terrestrial laser scanner (TLS). These devices use a technology called light detection and ranging (LiDAR) to create a high-resolution 3D model of their surroundings, in the form of a point cloud (Figure 1D). Scanning was conducted between August and September 2023. A multi-scan approach with seven scans per TSP was utilised. The positioning of these was not according to a rigid design but adapted to the local terrain and vegetation density to reduce occlusion. Physical registration aids (spheres, targets) were used. The single scans were registered automatically and manually by one operator using the “FARO SCENE” software. The two TSP trees were then manually segmented from the point cloud. This was done primarily by one operator and reviewed by two others using the software “RIEGL RiSCAN PRO”. All structural parameters were extracted only from these two single tree point clouds per TSP. The structural parameter extraction and data processing was then performed in R v4.5.2 (R Core Team, 2025). For this a newly developed R-library “coi” was used that is available at: https://github.com/ataeub/coi. For the trait data please refer to: https://data.botanik.uni-halle.de/bef-china/datasets/648

BEF-China main experiment: Site A Xingangshan, Jiangxi Province, Dexing County, China 29°080 29°110N, 117° 900 117°930E (geoname ID 7651294)

Tree individuals on Site A of the BEF China Main experiment

Refer to: https://doi.org/10.5281/zenodo.18443104