Prey selection of a widespread carnivore relative to predator-prey co-occurrence in space and time.

Understanding how prey selection relates to food availability and the environment is fundamental to conserving wildlife populations, especially in relation to anthropogenic factors. Most assessments of diet consider the abundance of prey, but rarely temporal availability, prey biomass or the co-occurrence of species in space and time, all of which are particularly important in systems where predator and prey species are spatially ubiquitous. Using a novel method that combines camera trap and scat data, we assessed how the spatial distribution and/or activity of prey taxa influenced prey selection by bobcats (Lynx rufus), a widespread carnivore in urbanized (exurban and wildland-urban interface) and wildland areas of Colorado, USA. We developed prey availability estimates using asymmetric species interaction models (spatial) and activity overlap (temporal) analyses, then used a model selection framework to identify if co-occurrence (i.e. availability) between bobcats and prey species in space, time or both better predicted prey consumption. We further investigated whether different measures of prey consumption (i.e. the frequency of occurrence of prey items in scats vs. the proportion of biomass consumed) influenced the relationship between prey availability and use. Across study areas, bobcats consumed a variety of smaller mammals and birds, with cottontail rabbits appearing as the most important diet item. Our results indicated that incorporating time into availability estimates often better predicted prey consumption compared with spatial availability, particularly in wildland areas. Additionally, we found that both the frequency of occurrence of prey remains in scats and the proportion of biomass consumed had similar relationships to availability in our study system. Our results suggest that including temporal availability in analyses of prey consumption may be particularly informative in areas where prey species are spatially abundant. This study provides a framework for researchers and wildlife managers to consider temporal availability and species co-occurrence in studies evaluating prey selection by carnivores.