Using UAV-LiDAR for stem volume phenotyping in the genetic selection of radiata pine

Abstract

Abstract Context : Radiata pine breeding programmes rely on stem volume as a key objective, but phenotyping constraints limit selection intensity. UAV-LiDAR offers a scalable alternative to labour-intensive field measurements. Aims : We evaluated UAV-LiDAR-derived metrics as genetic selection proxies for stem volume in radiata pine genetic trials and quantified their utility relative to field-measured diameter at breast height (DBH). Methods: LiDAR metrics describing tree height and size were assessed against allometric stem volume (ASV) across 11 genetic trials (~27,000 trees, two trial series) using single-step genomic best linear unbiased prediction (ssGBLUP) with ~9,500 SNPs. Results : The 3D surface area of the individual tree convex hull (convexhull3D_area) had the highest correlation with ASV (up to r = 0.87) and similar heritability to DBH (mean h 2 = 0.26). LiDAR tree height had the highest heritability (mean h 2 = 0.37) and moderate to high genetic correlation with DBH. Selecting the top 100 genotypes by convexhull3D_area recovered 67-86% of potential ASV genetic gain, versus 88-96% for DBH. Including malformed trees in the genetic analyses of LiDAR traits marginally reduced their performance as stem volume proxies. Conclusion UAV-LiDAR-derived tree height and 3D convex hull surface present desirable properties to complement field phenotyping for stem volume selection in radiata pine. Their scalability, repeatability and high heritability support lower phenotyping costs, better early selection and accelerated genetic gain in radiata pine breeding.

Plant phenotyping relevance

UAV-LiDARを用いて樹高・樹体サイズなどの形質を抽出し、茎体積の遺伝選抜プロキシとして相関・遺伝率・選抜効果を検証しており、植物フェノタイピング手法が中心である。

Code and data availability

The supplied blocks describe UAV-LiDAR point clouds, field DBH measurements, and genetic evaluation of radiata pine trials, but contain no data availability statement, no public repository deposit, and no author code/workflow URL. The analysis relies on proprietary ASReml-R and cited third-party tools (LAStools, lidR,

No evidence-backed public reproduction asset is currently recorded.