Increasing the Efficiency of LiDAR Based Forest Inventories: A Novel Approach for Integrating Variable Radius Inventory Plots with LiDAR Data.

Abstract:

LiDAR data are increasingly applied to support forest inventory and assessment across a variety of spatial scales. Typically this is achieved by integrating LiDAR data with forest inventory collected at fixed radius forest inventory plots. A well-designed forest inventory, one that covers the full range of structural and compositional variation across the forest of interest, is costly especially when collecting fixed radius plot data. Variable radius plots offer an alternative inventory protocol that is more efficient in terms of both time and money. However, integrating variable radius plot data with LiDAR data is problematic because the plots have unknown sizes that vary with variation in tree size. This leads to a spatial mismatch between LiDAR metrics (e.g., mean height, canopy cover, density, etc.) and plot data, which ultimately translates into errors in LiDAR derived forest inventory predictions. We propose and evaluate and novel approach for integrating variable radius plot data into a LiDAR based forest inventories in two different forest systems, one in the inland northwest and another in the northern lakes states of the USA. The novel approach calculates LiDAR metrics by weighting the point cloud proportional to return height, mimicking the way in which variable radius plot data weights tree measurements by tree size. This could increase inventory sampling efficiency, allowing for the collection of a greater number of inventory plots, and ultimately improve the performance of LiDAR based inventories.