Date of Award:


Document Type:


Degree Name:

Doctor of Philosophy (PhD)


Wildland Resources

Committee Chair(s)

S. Nicki Frey


S. Nicki Frey


R. Douglas Ramsey


David K. Dahlgren


Edd Hammill


Daniel R. MacNulty


In response to ongoing landscape change, wildlife species are likely to respond in varied ways. By studying habitat specialists, we are able to better understand the most likely ways in which the denizens of threatened ecosystems will react to those changes. Among the most threatened ecosystem types in North America are sagebrush ecosystems of the Intermountain West, where one of its most well-known residents, greater sage grouse (hereafter, “sage-grouse), have lost more than 50% of their habitat due to fire, invasive species, climate change, encroachment by coniferous forests and avian predators using it, and human-caused landscape conversion. Sage-grouse rely on sagebrush throughout their lives, and there are ongoing efforts to protect them as emblems of vulnerable species and to preserve a changing landscape. The purpose of my dissertation, as part of the ongoing efforts, was to improve understanding of how sage-grouse select habitat along their southern distribution edge in southern Utah and Nevada, where habitat tends to be fragmented and of poorer quality. In this research, I used more than six years of location data from GPS transmitters on sage-grouse across four study areas to address how sage-grouse respond to the threats they face and by what means of data analysis we are best able to detect those threats and inform effective conservation. My research shows that, despite the risk posed by avian predators, sage-grouse in these study areas select habitat closer to trees than expected and do so when they are likely able to also use dense sagebrush cover and a rugged landscape to be concealed from predators. I also found that sage-grouse may use habitat near to trees for shade and escape from extreme heat and cold when the sagebrush in their habitat is not enough to provide shelter, suggesting that sage-grouse must often make risky decisions to balance the many threats they face. Finally, I found that random forests, an intuitive machine learning method, are able to detect important effects of sagebrush and tree cover on habitat selection, able to predict those effects in new areas, and should be considered among the useful and important tools for measuring wildlife-habitat associations.