B51D-0464
Changes in Amazon Forest Structure and Canopy Illumination from Multi-temporal Lidar Data

Friday, 18 December 2015
Poster Hall (Moscone South)
Veronika Leitold, NASA Goddard Space Flight Center, Greenbelt, MD, United States
Abstract:
Lidar remote sensing of tropical forests provides unprecedented detail on 3D vegetation structure to support in-depth studies of ecosystem processes and carbon dynamics across large landscapes. Here, we used high-resolution, multi-temporal airborne lidar data from nine terra firme forest sites (total area = 3500 ha) in the Brazilian Amazon to estimate spatial and temporal patterns of forest disturbance and associated changes in canopy illumination. Across sites, we observed large variability in mean canopy height (15.7 m to 28.1 m) and the vertical distributions of forest vegetation and light penetration. At the site scale, lidar-derived canopy height models from repeat surveys showed minimal change in canopy structure over time intervals of 1 to 4 years, with nearly identical initial and final canopy height distributions. Annualized rates of total canopy turnover, based on losses in canopy height between lidar collections, ranged from 0.66 to 2.57% yr-1, with a mean value of 1.59% yr-1 across sites. Field estimates of tree crown sizes were used to classify canopy turnover into branch fall, tree fall and multiple tree fall events. Partial crown losses occurred most frequently across the landscape (40% of all events), but accounted for only a small fraction of the total turnover area (10%). Size-frequency distributions of canopy turnover followed a power-law distribution with a decline in the number of events with increasing size across all sites (range of λ between 1.26 - 1.35). The distributions of illumination conditions before and after disturbance events were inverted, as fully-illuminated crowns were replaced by low-light conditions within patches of canopy loss. Estimates of the spatial and temporal patterns of Amazon forest disturbance and recovery from multi-temporal lidar data complement information from plot-scale (≤ 1ha) studies to provide a more complete understanding of regional variability in ecosystem structure and function under current climate.