Evaluating Individual Tree Metrics Estimated from UAV Laser Scanning as Input to a Conventional Growth and Yield Model
Dr. Matthew Sumnall (VT) recently published “Evaluating individual tree metrics calculated from unmanned aerial vehicle laser scanning as input to a conventional growth and yield model” in the International Journal of Remote Sensing.
Abstract
Many growth and yield models (GYMs) have been developed in order to allow forest managers to predict future yield and explore potential management strategies. Remote sensing provides a potential alternative to field-based inputs to conventional GYMs, in particular, airborne drone laser scanning (DLS) has been used to accurately classify individual tree locations and derive stem size metrics, such as tree height and diameter at breast height (DBH) and competitive neighbourhoods, across entire stands, rather than plot-level samples. We adopted an older GYM, PTAEDA4.0, for use in R, which incorporated spatially explicit individual tree and local neighbourhood calculations, and was intended for use in the south-eastern US. Both field- and DLS-only inputs were used to estimate four-years of growth and yield on two managed loblolly pine (Pinus taeda) sites located in the south-eastern U.S.A. with variable stem density, genotype, and silviculture. All GYM estimates generally under-predicted actual field-measured values for both field- and DLS-derived inputs; however, the estimates produced by 2017 field and DLS metrics were statistically equivalent. For Bladen, the prediction error was 21–25% for estimating the tree height and 14–16% for DBH. For Critz, the prediction error was 6–8% for estimating tree height and 8–12% for DBH. The results demonstrate that the GYM would need re-parametrization to account for the current study site. Whilst the DLS was unable to account for all trees (98 to 99 % correctly found), the results demonstrate the potential of DLS as an alternative to traditional field measurements.
Full Citation
Sumnall, M. J., Carter, D. R., Albaugh, T. J., Cook, R. L., Campoe, O. C., & Rubilar, R. A. (2025). Evaluating individual tree metrics calculated from unmanned aerial vehicle laser scanning as input to a conventional growth and yield model. International Journal of Remote Sensing, 1–28. https://doi.org/10.1080/01431161.2025.2521069
Highlights
Airborne LiDAR was used to estimate forest metrics and spatial arrangement of stems;
PTAEDA4.0 was used to compare field and LiDAR inputs;
Estimates from the growth and yield model for field and LiDAR were statistically similar;
This demonstrates that airborne LiDAR is viable for individual tree growth and yield models.
Study Sites
Regionwide 201303 - Bladen Lakes State Forest (BL), Bladen County, NC.
Regionwide 201304 - Reynolds Homestead (CR), Critz, VA
