Exploring the potential of spaceborne LiDAR data for forest biomass estimation in a complex-structured coniferous forest

The development of adaptive forest ecosystem management strategies presumes the existence of aboveground biomass estimates at different spatial scales. In recent years, a series of remote sensing technologies have been developed, providing timely, low-cost, and reliable estimates on biomass and other forest attributes, especially over large and/or remote areas. Among the various technologies, Light Detection and Ranging (LiDAR) employed on aerial, satellite, and even ground platforms typically offer accurate forest biomass estimates. This is due to the ability of the emitted pulses to penetrate the forest canopy, thereby providing 3D information about all forest vegetation layers. While ground-based and aerial LiDAR provide a more detailed characterization of the vertical structure of vegetation compared to satellite-based LiDAR, the latter offers obvious advantages in terms of cost-effectiveness. This study aimed to investigate the potential of LiDAR GEDI (Global Environmental Dynamics Investigation) satellite data to reliably estimate the biomass of stem, bark, branches, and needles at the GEDI footprint level, which are the constituent components of forest aboveground biomass. The study area consists of a dense fir forest characterized by complex structure and intense topography. A regression analysis with the random forest (RF) algorithm was performed to develop prediction models for all parts of the aboveground forest biomass using height and crown profile metrics. Accordingly, five random forest models were created and evaluated for their predictive performance using an independent validation sample. In addition, the influence of forest biomass density and topography on the estimates was examined. The results reveal that the biomass of crown parts can be estimated with satisfactory accuracy and an increase of almost 15% in accuracy was observed in areas with gentle slopes and topography. Overall, GEDI data can be used to estimate forest biomass and its key parts in large extents with mild topography, regardless of the aboveground biomass density.