Systematically characterizing biodiversity across broad scales can be challenging due to difficulties in observing, mapping, and validating the presence of species individuals and communities. Understanding species interactions therefore can facilitate biodiversity assessments. Woodpeckers are considered a keystone guild of species whose presence has been shown to be indicative of overall forest bird and other vertebrate diversity at the landscape scale. Our objective was to characterize woodpecker habitat via remote sensing methods so as to predict woodpecker presence, and in so doing to identify potentially suitable habitat for a suite of other vertebrate species that depend upon woodpecker cavities. To meet this objective, we derived vertical vegetation structure information using waveform lidar from the Geoscience Laser Altimeter System (GLAS) aboard the ICESat satellite and used it to predict woodpecker presence as sampled via ground based surveys. We used terrain information derived from airborne lidar and SRTM DEM data to remove a simulated ground return signal from the spaceborne lidar return waveform. To test the improvement in feature discrimination in the resulting vegetation signal, an airborne lidar based simulated waveform was used as a baseline measure of accuracy. We modeled relationships between forest structure parameters (e.g. mean canopy height, proportion of foliage to ground return energy), GLAS lidar waveform characteristics (e.g. amplitude and height of dominant peaks) and woodpecker presence/absence with an ecologically based logistic regression approach that relies on habitat features known a priori from the literature to be important in nest and roost site selection for woodpeckers.
Results/Conclusions
Results indicate the most parsimonious model contains the amplitude of the highest peak (~30 m) and the height of the second highest peak (~25 m) in the GLAS waveform to predict presence of Pileated woodpecker (Dryocopus pileatus). These components correspond to the density of foliage at the mean canopy height and height of foliage in the immediate sub-canopy layer, respectively, which are consistent with habitat requirements that include nesting trees 65 – 154 cm DBH and roosting trees 155 – 309 cm DBH. These results are significant because 1) they enable development of methodologies to quantify vegetation structural characteristics and improve our ability to characterize habitats for biodiversity assessments, and 2) create techniques to establish current and baseline vegetation based habitat characteristics for animal taxa at the landscape scale that will be applicable to broader studies at regional and global scales.