Uncertainty in lidar derived canopy height models in three unique forest ecosystems

摘要

NEON conducts annual LiDAR flights over several ecologically unique sites within the continental United States. One of the products derived from the LiDAR acquisitions is a canopy height model (CHM), required to make inferences about vegetation structure and annual changes in growth. It is hypothesized that the flight acquisition parameters are sufficient to allow NEON's CHMs to detect inter-annual change in growth at the NEON forest plot scale. To test this hypothesis, a series of overlapping LiDAR flight lines were collected over a sample forest area at three unique ecosystems. The repeat data allowed empirical quantification of uncertainty in CHMs. When related to the growth rate of the dominant species, this allowed estimation of the time required to separate true change from CHM uncertainty and conclude growth had occurred. Results highlight that each forest eco-system displays unique sources of uncertainty, and that annual changes in NEON forest plots are generally achievable with the exception of plots with minimal canopy cover.