Inferring behavior from GPS data: Challenges and successes modeling behavior of a large solitary carnivore

Human curiosity about the habits and behaviors of animals have been documented throughout history, and in modern times legal obligations to conserve and manage wildlife resources have made understanding the habits and needs of many wildlife species a necessity. Advances in modern technology have resulted in a variety of new ways to monitor wildlife activities, with Global Positing System (GPS) tagging becoming a popular technique to record animal locations over time. Animal movement captured via GPS telemetry provides underlying information about animal behavior that cannot be observed directly for many species, but how best to quantify and make inference on these observations is still an evolving field of study. Technological and conceptual challenges still need to be overcome before researchers can truly understand what kind of information these observations provide us. This dissertation critically assesses GPS wildlife collar technology and develops ways to correct for biases this data collection method can introduce in topographically complex regions. I provide a review of terrain metrics available for use in wildlife spatially explicit studies describing aspect, complexity, and landforms, and evaluate how a large solitary carnivore, the cougar (Puma concolor) orients to the terrain in northern Arizona. I found that novel metrics of coarse scale terrain ruggedness and a fine scale metric of view-shed ranked highest in describing where cougars occur, but a multivariate approach outperformed any single metric. I then refined resource selection from a population wide estimate to resource selection for particular behaviors, by identifying changes in the characteristics of cougar movement related to foraging behavior. Cougars are a prey caching predator and this foraging strategy requires advanced metrics to quantify movement. I developed a novel way to quantify movement, useful for describing behavior at a variety of scales from identifying circadian activity levels, periods of foraging, and seasonal movements related to changes in prey distributions and reproductive status. The depth of information GPS telemetry provides and advancements in our ability to map and quantify features of the landscape has great value for future conservation efforts. This dissertation provides correction factors and novel metrics to support these future scientific endeavors.