John A. Kershaw Jr.
Canopy surface-based and height distribution-based attributes are primarily used to model area-based forest attributes such as volume or biomass per ha, but those attributes are less effective at estimation of other commonly needed forest attributes, including basal area, density, mean tree size, canopy volume, leaf area, and many others. There is limited generalized theory that has emerged regarding the relationship between LiDAR point clouds and forest structure. To have better predictive capability for stand parameters such as leaf area distribution, structural diversity, and biodiversity, a system of biologically-relevant and generalized LiDAR attributes are required. These attributes should be derived from our biological understanding of forest structure and productivity. The approach we propose builds upon current forest inventory practices, stand development theories, and measures of stand structure and complexity. With these traditional measures, we propose to use them to inform the types of LiDAR attributes which should be extracted or calculated from point clouds. Initially, we perceive four types of LiDAR point cloud attributes: 1) height-based; 2) horizontal density-based; 3) horizontal and vertical variability; and 4) individual tree segmentation. The initial goal is generalizable forest inventory predictions. Implications for sample design, growth and yield modeling, and forest management decision making are explored.
No datasets are available for this submission.