Validating modeled lidar waveforms in forest canopies with airborne laser scanning data

A thorough evaluation of the capability of modeling vegetation lidar returns is a critical aspect of deriving vegetation structure from lidar measurements. This study assesses the performance of the Analytical Clumped Two-Stream (ACTS) canopy radiative transfer model to simulate large-footprint lidar waveforms. Modeled lidar waveforms were compared to airborne Laser Vegetation Imaging Sensor (LVIS) data collected in deciduous and conifer forests: Harvard Forest, MA; Bartlett Experimental Forest, NH; and Howland Experimental Forest, ME. The simulated and LVIS lidar waveforms have coefficients of determination R2 > 0.9 and RMSE ~ 0.01 at both plot and stand level for most sites. The ACTS model also produces realistic multi-peak returns from vegetation for the multi-layer and multi-species canopies with R2 ~ 0.79-0.86 and RMSE ~ 0.01 between the simulated and LVIS waveforms. This validation work lays the foundation to retrieve vegetation structure and above-ground biomass directly from lidar waveforms through model inversion with the ACTS model.