Patterns of amphibian and reptile distributions across an urban to rural gradient within the Santa Monica Mountains National Recreation Area

Background/Question/Methods Urbanization is a process widely recognized for affecting the distribution and abundance of wildlife. Urbanization represents a syndrome of changes to the biophysical environment associated with increasing human population density; however, in ecological studies, urbanization is often treated as a qualitative, dichotomous landcover state (e.g., urban v. rural) or represented by a single covariate (e.g., area of impervious surface). This is potentially problematic when studying wildlife in urbanizing landscapes because wildlife will respond to the suite of specific changes to the biophysical environment that accompany increased human density. The objectives of this study were to analyze changes in habitat attributes in relation to human population density within the Santa Monica Mountains National Recreation Area and to relate patterns of amphibian and reptile distributions to specific environmental attributes or principal components of urbanizing habitats.

Results/Conclusions We used four years of capture data from 79 pitfall arrays. Among our study sites, human population density was negatively correlated with protected area patch size, proximity to paved surfaces, human-made structures and artificial light sources. Models of reptile and amphibian occupancy and abundance generally did not include consistent urban attributes nor the composite index of urban attributes. For the most common lizard species, Western fence lizards (Sceloporus occidentalis) occupancy was positively correlated with human population density and slender salamander (Batrachoseps spp.) occupancy increased within greater proximity to human-made water sources. Occupancy of all other less common reptile species was related to vegetation and/or elevation, not directly related to the anthropogenic impacts of urbanization. Abundance model fit was generally poor; however, results suggest that abundance varied with vegetation among most herpetofaunal species. Our survival models generated similar results for Western fence lizards under count-based modeling approaches; however, both the count-based and capture-mark-recapture approaches may have generated inflated estimates due to low recapture rates and sampling design constraints.