B127-07
Landscape-scale Controls Over Tropical Forest Gap Dynamics Elucidated Using Repeat Drone Photogrammetry

Thursday, 17 December 2020: 04:24
Virtual
KC Cushman1, Helene C Muller-Landau1, Jeff Chambers2, Stuart J Davies3, Matteo Detto4 and Milton N Garcia1, (1)Smithsonian Tropical Research Institute, Balboa, Ancon, Panama, (2)Lawrence Berkeley National Laboratory, Berkeley, CA, United States, (3)Smithsonian Tropical Research Institute, Washington DC, United States, (4)Princeton University, Princeton, United States
Abstract:
Topography and soils are expected to influence tropical forest dynamics because they affect trees’ water availability, access to nutrients, and exposure to disturbance. However, small field plots generally lack sufficient variability in topography and soils to evaluate their importance for controlling rates of tree- and branch-fall, episodic events requiring large samples to accurately quantify.

Here we evaluate the importance of topography and soils for gap dynamics across Barro Colorado Island (BCI), Panama, using three-dimensional measurements of forest structure from drone photogrammetry. BCI contains >1500 ha of lowland tropical forest with substantial variation in topography, geology, and soil types. Drone imagery of the whole island was collected in June-August of 2015, 2017, 2018, 2019, and 2020. We aligned photogrammetric point clouds, created canopy elevation raster models, and quantified gap creation events using the canopy elevation loss between successive years.

We compared rates of gap creation across soil types, forest age, and a suite of topographic metrics and resolutions. We found that rates of gap creation vary annually, are highest in north- and west-facing areas, and are highest in areas with high curvature. Gap creation rates varied approximately two-fold among soil types, from < 2 % yr-1 of the landscape to > 3 % yr-1 . Our results will inform efforts to model tropical forest disturbance dynamics by providing knowledge about the mechanisms driving gap formation, and by providing improved benchmarking data for this much-studied tropical forest site.