Modeling tree radial growth with competition and crown characteristics from an aerial perspective

Tree growth is a central aspect of stand dynamics; however, traditional methods for measuring tree growth have limitations, including destructive sampling and the need to remeasure plots periodically. Unmanned aerial vehicle laser scanning (ULS) provides new opportunities for precise, detailed, nondestructive, stand-scale reconstruction of tree canopies via three-dimensional structural methods. However, the aerial canopy information generated by ULS has not yet been effectively applied in exploring tree growth. This study was designed to analyze and quantify the influences of crown structural characteristics and competition among individual trees on $\mathrm{DBH}$ increments by combining field-measured growth data with single-temporal ULS data. An automated point cloud processing method was employed to characterize 7112 trees on 71 sample plots of larch in three dimensions to extract information in the form of distinct crown structural factors and competition indices. The results revealed that crown structural characteristics and competition indices determined via ULS were closely linked to tree $\mathrm{DBH}$ growth. Specifically, among the crown variables, the crown exposed surface area was the most strongly correlated with tree $\mathrm{DBH}$ growth (r=0.34), and the distance-independent competition index showed a stronger negative correlation (r=-0.61) with tree growth than the other competition indices. The metrics extracted from ULS can explain 53.9 % of the variation in $\mathrm{DBH}$ growth. This study evaluated the potential for using the latest ULS data as input, in conjunction with an automated point cloud processing algorithm, to forecast future tree $\mathrm{DBH}$ growth, which could inform the development of more effective strategies for forest management.