Laser Remote Sensing of Canopy Habitat Heterogeneity as a Predictor of Bird Diversity and Abundance
Scott
Goetz, Woods Hole Research Center, sgoetz@whrc.org
(Presenting)
Daniel
Steinberg, Woods Hole Research Center, daniel.steinberg@yale.edu
Ralph
Dubayah, University of Maryland, dubayah@umd.edu
Richard
Holmes, Dartmouth University, richard.t.holmes@dartmouth.edu
Matthew
Betts, Oregon State University, matthew.betts@oregonstate.edu
Patrick
Doran, The Nature Conservancy, pdoran@tnc.org
We explored the utility of lidar imagery to estimate habitat metrics associated with bird species richness and abundance in two temperate forested regions of the Eastern United States: the Patuxent National Wildlife Refuge and Hubbard Brook LTER Site. Our objectives were: (i) to map habitat metrics using lidar, (ii) assess the utility of the metrics for predicting bird species richness, abundance, and habitat occupancy (iii) compare the lidar-derived predictors with other metrics derived from more traditional remote sensing imagery. A secondary objective was to assess the relative utility of different statistical techniques for analyzing the results, including traditional linear regression, general additive models, and binary hierarchical splitting algorithms (regression trees). Species richness was predicted best when stratified by guilds dominated by forest, scrub, suburban and wetland species, with similar lidar variables selected as primary predictors across guilds. The vertical distribution of canopy elements was found to be the strongest predictor of total richness. Generalized linear and additive models, as well as binary hierarchical regression trees produced essentially the same results. The lidar metrics were consistently better predictors than traditional remotely sensed variables such as canopy cover, indicating that lidar provides a valuable resource for biodiversity research.
NASA Carbon Cycle & Ecosystems Active Awards Represented by this Poster: