Towards remotely sensed forest dynamics to track and understand forest change and its consequences for the atmosphere

Abstract:

Forest structure and dynamics are changing rapidly around the world as a consequence of climate-change-related pests, droughts and wildfires, and changing anthropogenic forest usage. It is critical to understand these changes at the process level (e.g. demographic performance of trees) to accurately predict subsequent forest responses, including impacts on carbon and water cycles, surface-atmosphere energy dynamics, and ecological trajectories. Remote sensing may offer a critical tool to assess and understand these changes at large scales if it can connect observations of forest canopies to underlying biological processes such as demography. Furthermore, too be effective such an approach must be able to assess the full spectrum of forest demographic groups, including those that fall primarily in the shade of larger trees.

Here we describe recent advances that address this problem by reconstructing high-resolution canopy structure in 3D from airborne LiDAR data and then associating LiDAR derived leaf area strata and leaf area light environments with the stem frequency, biomass and performance of trees in different demographic groups. Focusing on Amazonian forests, we demonstrate how forest structure (e.g., size distributions) may be retrieved by accounting for tree architecture over light environments. Furthermore, by linking LiDAR derived light environments with LiDAR derived (or ground measured) size classes, even a single LiDAR survey can help reveal forest dynamics, while multiple LiDAR surveys across years allow for direct assessment of the role of canopy light environments in demographic dynamics. New advances may unlock the full potential of this approach by including estimates of canopy function, and biophysical environments, and by integrating across scales utilizing measurements from tree plots to space-borne LiDAR. We conclude by illustrating the potential of this approach to inform ecosystem models that explicitly represent forest dynamics and vertical canopy structure, and thereby significantly advance our understanding of forest disturbance and its consequences in the Anthropocene.