For nature reserves in urban settings, wildlife and wildlife habitats may be affected by recreational activities and intensive, adjacent development. Sustaining biodiversity in such reserves is a challenge for land and natural resource managers, but identification of core areas and key resources for wildlife species may help in planning for current and emerging threats. To help identify core areas and resources, we conducted spatial analyses and predictive modeling of vertebrate distributions for a network of nature reserves in densely populated Orange County, California. We primarily focused on bobcats (Lynx rufus), a species with a strong association with natural habitat. Bobcat space use has been correlated with broad, simple land-use categories, but relatively little is known about the influence of greater landscape complexity on habitat suitability for bobcats. To examine habitat selection by bobcats, we developed spatial data layers representing environmental factors that might influence this species, and we used previously collected Global Positioning System tracking data for 30 male and 21 female bobcats to indicate bobcat response to complex landscape factors. We examined these inputs using Resource Selection Function (RSF) modeling and developed spatially explicit models of the probability of bobcat use (selection or avoidance) of landscape characteristics. RSF models highlighted the general importance of reserve habitat for bobcats, but suggested that female bobcats were more dependent that male bobcats on habitat within designated reserves. Male bobcats, which range more widely than female bobcats, were associated with undeveloped areas both within and outside reserves. Small areas were present outside reserves that seemed to provide additional suitable habitat or movement areas for bobcats, potentially through restoration, connectivity, or reduced edge effects.

Although bobcat RSFs suggested areas of high value to this species and potentially other species, taxa can differ greatly in their resource-selection and spatial requirements. Thus, for several species of reptiles, amphibians, and birds, we adapted species distribution models based on occurrence data to examine the response of other vertebrates to the landscape. To identify potential High-Value Areas (HVAs) for single or multiple species, we then developed a step-wise filtering process that can be applied to a series of spatial data layers. We provide examples of alternative decision models for HVAs that capture different elements of biodiversity and a range of management considerations. As landscape and management challenges change, these spatial layers and decision rules can be adjusted based on new information. Our approach thus establishes a general framework for identifying high-value habitat that can be used for current management decisions and refined in the future, depending on management interests and goals and the availability of suitable quality data or adequate surrogate information.