LAUSR

Large-scale distribution models are effective predictors of habitat suitability and connectivity across broad landscapes and are useful management tools, though few large-scale species distribution models exist for medium-sized predators in urban landscapes. We modeled the potential distribution of 4 medium-sized predators in a 17,361-km2 portion of the Chicago Metropolitan Area. We applied a maximum entropy algorithm model (MaxEnt) using presence-only data collected via remote cameras from 54 Lake County, Illinois, forest preserves during August–October 2008–2012. Environmental data layers used to model distributions were distances to forest, grassland, barren land, crops, wetlands, developed open space, developed low intensity, developed high intensity, water, primary roads, secondary roads, and tertiary roads. Coyotes (Canis latrans) had the greatest area of potential distribution followed by opossums (Didelphis virginiana), striped skunks (Mephitis mephitis), and raccoons (Procyon lotor). Models for all species had high AUC values (0.90–0.94) indicating strong predictive performance. More than 50% of the study area was predicted to be within the distributional limit for each focal species. Distance to forest was the most important contributory predictor for all species modeled (82% - 96%) and higher probability of presence for all 4 species was indicated closer to forest and further from tertiary roads. However, coyotes and raccoons were predicted to prefer habitat closer to highly-developed areas. Our research indicates medium-sized predators are highly synanthropic and able to persist within the Chicago Metropolitan Area given adequate availability of non-urban land cover, particularly forest, and ample green space linking forest patches within highly-developed areas.