Effects of landscape covariates on the distribution and detection probabilities of mammalian carnivores

Abstract

Mammalian carnivores are affected by various anthropogenic disturbances near urban environments. Urban expansion and increased anthropogenic activity near and in preserved habitats may cause shifts in the current spatial distributions of those species. To predict the effects of future land-use changes on mammalian carnivores, we modeled their current occurrence across former Fort Ord Army Base as a function of urban proximity and road/trail density. We collected detection/nondetection data for domestic dogs, coyotes (Canis latrans), gray foxes (Urocyon cinereoargenteus), raccoons (Procyon lotor), striped skunks (Mephitis mephitis), and bobcats (Lynx rufus) using scent stations. We analyzed our data with likelihood-based occupancy modeling and used evidence ratios based on Akaike information criterion weights to infer the effect of each variable on occurrence and detection probabilities for each species. We used the estimated weighted model coefficients of the predictive variables to create current and future species distribution maps given proposed landscape changes in the study area. Domestic dogs were more likely to use areas closer to the urban edge, while gray foxes showed a preference toward inland areas. Detection probability was highest in areas closer to the urban edge for striped skunks and in areas with high road/trail densities for raccoons. Our results suggest that the distribution of domestic dogs will most likely expand with future development, while those of gray foxes will contract. We predict that future land-use changes outside of preserved areas will have an adverse effect on gray foxes within the protected areas.