Forest Structure Estimation and Pattern Exploration From Discrete Return Lidar in Subalpine Forests of the Central Rockies

Discrete return lidar has been used to accurately measure and characterize forest structure across a range of forest types, with canopy surface height and canopy profile indices used as explanatory variables in regression analysis. This study evaluates the ability of discrete lidar to estimate forest structure and forest biomass variables using both traditional lidar indices (i.e. mean height, max height, height percentiles, etc.) and statistically derived canonical correlation analysis (CCA) variables across three temperate subalpine forest sites in the Central Rockies. Modeling results with both lidar and CCA explanatory variables performed well with lidar models consistently having slightly higher explained variance, and a lower ratio of mean predicted value, relative to models derived with CCA variables. Adjusted R2 values for mean height, sum of leaf area and all carbon in live biomass were (0.93, 0.93), (0.74, 0.73) and (0.93 and 0.85) for the lidar and CCA explanatory regression models respectively. Investigation of forest complexity patterns using graphs of forest variable correlations with lidar canonicals one and two revealed distinct forest structure clusters within ordination space. Canonical one is highly correlated with forest height, biomass, and total leaf area, and canonical two is highly correlated with tree density. When canonicals one and two are considered in conjunction they represent a continuum of stand age and structure from young to mature forest. The lidar derived biomass estimates will be utilized in the US Forest Service Northern Global Change Research Program, where the extensive lidar derived biomass estimates will be compared with coincident intensive flux tower biomass estimates.