Habitat mapping and multi-scale resource selection of pygmy rabbits with unmanned aerial systems

Abstract

Assessing habitat quality is a primary goal of ecologists. However, evaluating habitat features that relate strongly to functional habitat quality at fine-scale resolutions across broad-scale extents is challenging. Unmanned aerial systems (UAS) bridge the gap between relatively high spatial resolution, low spatial extent, field-based habitat quality measurements and lower spatial resolution, higher spatial extent satellite-based remote sensing. In chapter one, I mapped structural habitat quality (i.e., shrub height, shrub volume, and aerial concealment) at two study sites with UAS and compared that with estimates derived from terrestrial laser scanning, a highly accurate ground-based sensor. I demonstrated that UAS was able to accurately estimate habitat heterogeneity for a key terrestrial vertebrate at multiple spatial scales. In chapter two, I classified sagebrush structural morphotypes and mapped forage quality including both nutrients (i.e., crude protein) and plant secondary metabolites (i.e., coumarins and monoterpenes) with UAS in both summer and winter. With these maps, I generated foodscapes representing high and low forage quality for herbivores that use sagebrush landscapes such as the sagebrush-obligates, pygmy rabbits (Brachylagus idahoensis) and greater sage-grouse (Centrocercus urophasianus), and generalist herbivore mountain cottontails (Sylvilagus nuttallii). In chapter three, I analyzed habitat selection by pygmy rabbits at multiple scales. I found that pygmy rabbits select for food variables (i.e., crude protein and total monoterpenes) while foraging (bite marks) at the plant scale, and selected for both security cover (i.e., aerial concealment and distance to burrow) and food variables at the patch scale. While moving and resting (fecal pellets), pygmy rabbits selected for a combination of food, security, and thermal properties, with thermal variables appearing more frequently in top models at the plant scale, and primarily for security at the patch scale. Finally, I applied my patch-scale model of pygmy rabbit habitat selection across the landscape using UAS-derived maps of security cover and food quality. These results show scale-dependent habitat selection in a central-place browsing herbivore. Making explicit predictions of trade-offs affecting habitat use, and assessing those predictions at distinct scales, will be necessary for improving knowledge of habitat selection patterns in herbivores.