Estimating the Number of GEDI-footprints required for an Accurate Representation of Canopy Heights in the Amazon

Estimations of canopy height and its variations provide insights into species composition, forest structure, timber yield, biodiversity hotspots, forest disturbance impacts and net primary production, and can even serve as inputs for ecosystem and carbon cycle models. Traditional methods, dependent on field surveys, are not feasible for large study areas, making approaches such as remote sensing more attractive. LiDAR (light detection and ranging) has the ability to extract accurate elevation and vertical forest structural information even in areas of dense canopy covers. NASA's latest space-borne LiDAR, GEDI (Global Ecosystem Dynamics Investigation), now makes it possible to gather three-dimensional data of forests globally. The aim of this study is to determine the minimum number of high-quality GEDI footprints required to extract height information (RH98; relative height) of tropical forests with satisfactory accuracy for different canopy cover percentages. The study area comprised four forest locations in Amazon - one site each in the states of Para and Mato Grosso and two sites in Acre - which experienced fire and logging disturbances between 2005 and 2015. The validation process included a comparison of the GEDI-derived RH98 (i.e., average from composites of foot-print level measurements) with maximum canopy height derived from airborne laser scanning (ALS) data (i.e., average from grids encompassing the respective GEDI foot-prints); the ALS data was collected as a part of the Sustainable Landscapes Brazil initiative. Preliminary results approximated canopy heights to be in the range of 0 to 52 m and in general, a minimum of 5 GEDI footprints were required for predicting grid-level heights; tree heights within grids having canopy cover less than 40% had the poorest agreement with the ALS data. We also consider PAVD (plant area volume density), to demonstrate the transferability of our approach to other biophysical variables. We plan to use our workflow to assist future research investigating how climate change and previous forest disturbances affect plant growth and biomass accumulation rates over time.