Ecological Characterization Of An Intact Tropical Peat Forest Using Airborne Small Footprint LiDAR

Wednesday, 17 December 2014
HA THANH Nguyen1, Lucy Hutyra1, Steve M Raciti2 and Brady S Hardiman1, (1)Boston University, Boston, MA, United States, (2)Hofstra University, Biology, Hempstead, NY, United States
Tropical peat forests in Southeast Asia have been experiencing climatic and anthropogenic disturbances in the form of drought, fire, deforestation and drainage at an increasing pace and with an increasing extent throughout the past two decades. In this project we aim to improve our understanding of the structural dynamics of tropical peat swamps and the effect of deforestation on the forest structure by (i) characterizing the forest structural parameters (stem density, stem height, crown area, crown roughness, gap size and frequency) of an intact peat dome and (ii) comparing with those from a nearby deforested peat dome. Both are located in Northwestern Borneo. We combine field sampling of 0.8 hectare of forest in 2014 and 84km2 of airborne, small footprint, discrete returns LiDAR acquired in 2010 to extract the parameters of interest. We first process LiDAR data to produce to a Digital Elevation Model (DEM) and a Canopy Height Model (CHM) of the area. Individual canopy stems are extracted through local maxima filtering with varying size and shape of search windows. Canopy crowns are segmented from the CHM via K-means clustering using stem positions as fixed cluster centroids. Canopy crown height and stem density are calibrated with field survey in order to upscale stem density to the whole peat dome. Crown roughness is defined as standard deviation of each cluster (crown). Finally, gaps were delineated from the CHM with 30m as vertical threshold and 40m2 as minimum area. The entire procedure is then repeated for the deforested peat dome.

Across the intact peat dome, we find an increase in stem density but a decrease in canopy stem height, canopy crown area and canopy crown roughness as a function of a 5m elevational change. Gap size frequency follows a Gamma distribution with higher variance in gap percentage for areas closer to the dome center. As a function of canopy stem height, aboveground biomass decreases towards the dome center. For the deforested peat dome, deforestation removes the canopy trees, masking away the trends of increasing stem density but decreasing stem height and crown area against dome elevation. Canopy roughness is higher at the deforested site as a result of regrowth. Deforestation also results in higher number of smaller gaps, created by felling trees, with diffuse spatial pattern throughout the forest.