Humans have altered landscapes throughout the world to affect the behavior of wildlife species. Yet, we lack a nuanced understanding of how animals navigate the complex agricultural, residential, and natural environments that now characterize much of our planet. Here, we examined coyote (Canis latrans) resource selection in a mixed-use agricultural landscape in Mendocino County, California, USA. We deployed GPS collars on coyotes (n = 8) and used Hidden Markov Models to estimate three behavioral states for coyotes: resting, foraging, and traveling. We then tested whether coyote selection for land cover, cropland, or risk factors differed by behavioral state and time of day.
Results/Conclusions
We found that coyotes selected for different resources depending on time of day and which behavior they were exhibiting. Overall, coyotes selected for livestock pasture, open grassland habitat, and water, with no effect of cropland or development. However, while foraging, coyotes selected for areas far from development, and while resting, coyotes selected for cropland. This research highlights the critical importance of considering both behavioral and diel patterns when assessing how animals balance risk-foraging tradeoffs while living in a human-modified environment.
