Towards a Better Understanding of Forest Biophysical Parameters - Combining High Fidelity Simulations, Airborne Waveform Lidar, and Terrestrial Lidar Sensing

Abstract:

The science community has come a long way from traditional, 2D imaging approaches to the assessment of ecosystem structure, function and composition. For example, waveform- (wlidar) and terrestrial lidar systems (TLS) present us with exciting opportunities for detailed, accurate and precise, and scalable structural characterization of vegetation. wlidar and TLS generally can be regarded as complementary i.e., airborne wlidar typically digitizes the entire backscattered energy profile at high spatial and vertical resolutions, while TLS samples dense 3D point clouds of the bottom-up vegetation structure. Research teams at Rochester Institute of Technology (RIT) have been collaborating with the National Ecological Observation Network (NEON) to assess vegetation structure and variation in the Pacific-Southwest (San Joaquin Experimental Range and Soaproot Saddle sites, CA) and Northeast (Harvard Forest, MA) domains. The teams collected airborne small-footprint wlidar data and in-situ TLS data for these sites and is taking a two-tiered (top-down and bottom-up) approach to forest structural assessment. We will present our work where we (i) studied wlidar signal attenuation throughout the canopy in a simulation environment - the attenuation correction factor was linearly proportional to the sum of the area under the proceeding Gaussians - and (ii) used the fine-scale stem structure extracted via TLS to reconstruct complex, but realistic, 3D forest environments for refined simulation studies. These studies indicate that we can potentially assess vegetation canopies remotely using a vertically-stratified approach with wlidar and use rapid-scan TLS technology to calibrate models predicated upon synoptic airborne systems. Other outputs of our approaches can be used for typical forest inventory, ecological parameter extraction, and new algorithm validation.