Understanding the movement patterns and habitat needs of the endangered Matschie’s tree-kangaroo (Dendrolagus matschiei) is important for their conservation and management. Endemic to the montane cloud forests of the Huon Peninsula in northeastern Papua New Guinea, these elusive arboreal marsupials are tremendously challenging to study using traditional observational methods.
This study is an assessment of novel techniques to overcome the significant challenges to in-situ data collection in remote and rugged tropical cloud forests. Animal locations are remotely tracked using purpose built altitude and motion logging GPS collars and habitat structure data is measured using photogrammetry from small Unmanned Aircraft Systems (UAS) aerial imagery. Leveraging the autocorrelation of regular GPS location sampling, this study applied a Time-Local Convex Hull (T-LoCoH) analysis to investigate particular locations that may be important to D. matschiei as well as potential barriers to movement that would be inside of the home range as identified in previous studies. A novel technique of ground surface interpolation from canopy gaps is presented to overcome the challenges of photogrammetric reconstruction of terrain surfaces under closed canopy forests. From this a variety of forest structure variables were calculated to understand the 3D complexity of these heterogeneous cloud forests.
Results/Conclusions
This investigation found that custom GPS collars can provide high fix success rates in dense multilayer forests found at the research site. The regular sampling intervals resulted in areas of utilization that were notably smaller than with traditional home range analyses, and provided insight into landscape features that the animals do not use. D. matschiei were found to preferentially use trees that were taller than average and were found in closer than average proximity to canopy emergent trees. The reconstruction of 3D habitat data from UAS aerial photogrammetry resulted in forest structure maps that have significant potential to overcome the necessity of manual habitat data collection that hinders large scale habitat research, for this and many other species.