Satellite lidar data do not show static greenness in wet equatorial Amazonian rainforests

Abstract:

Lidar waveform instruments have been successfully used to quantify biophysical properties (e.g. height and biomass) of forests by examining the vertical profile of return laser energy from the surface. Although temporally consecutive lidar waveform data are feasible to detect abrupt changes such as clear-cut or forest fire in a large area, their application to study phenology or seasonal dynamics is novel and untested. A recent study, claiming no green-up in Amazon forests, suggested that Geoscience Laser Altimeter System (GLAS)-derived metrics are potential candidates to capture seasonal variations in canopy structure and greenness. Those metrics are independent from sun-sensor geometry. Here, we focus on tests of three GLAS metrics, that is, (a) Waveform Centroid Relative Height (WCRH), (b) 1064 nm NIR apparent reflectance (AR), and (c) Leaf Area Index (LAI) for detecting seasonal dynamics in an already dense tropical rainforest. Our investigations show a seasonal green up in wet equatorial Amazonia (ΔLAI ~0.5 during the dry season). We highlight that the WCRH and AR are not explicitly sensitive over the high LAI condition (LAI ≥ 4.5), which may lead incorrect conclusions (e.g. consistent forest structure and greenness). In the case of WCRH, the metric would be useful to detect seasonal dynamics only when forests allow enough penetration of lidar energy (LAI from 0 to 4.0). On the other hand, AR is not recommended for identifying the seasonal greenness in Amazonian forests, whereas the metric could be used to screen invalid GLAS data.