Relating Occupancy Patterns to Multi-Life-History Scales for Pond-Breeding Anurans in Eastern Virginia

Amphibians are experiencing population declines worldwide due to a multitude of factors, including habitat loss. Therefore, assessments of breeding habitat occupancy are important to determine which environmental characteristics are crucial for conservation of amphibians. While most research assesses habitat conditions at the breeding site (i.e., pond or wetland), for many species this misrepresents annual habitat use as many migrate to terrestrial habitats once breeding is completed. To provide a comprehensive assessment of annual habitat use, we evaluated how anuran (i.e., frog and toad) breeding site occupancy related to three life-history scales: breeding, migration and dispersal. Basically, the three scales define habitat requirements during breeding (i.e., breeding life-history scale), the non-breeding season (i.e., migration life-history scale) when anurans migrate between wintering and breeding areas, and periods when individuals leave their natal area to move to another breeding location (i.e., dispersal life-history scale). We focused our study on eight anuran species, and for each we assessed occupancy nine times during two breeding seasons (March – July 2011 and 2012) at a total of 108 breeding sites. We generated occupancy models to evaluate which variables and their associated extents (i.e., distribution of habitat use within concentric buffers surrounding breeding sites) were important predictors of breeding-site occupancy. We categorized covariates on their order of their importance: important (high AICc weight and 95% CI not including zero); of intermediate importance (intermediate AICc weight and 95% CI including zero); and unimportant (performing worse than the null model). We found that for five species, percent upland-forest cover, relating positively to occupancy, was the most important migration life-history scale variable. For the dispersal life-history scale, occupancy was negatively related to highway density in seven species. The breeding life-history scale was the least important predictor of breeding-site occupancy because it was important for only one species. Our research identified the importance of forest cover surrounding breeding sites ranging in extents from 300 m - 900 m, which exceeds buffer width currently required by law in the Commonwealth of Virginia (30-m buffer). In addition, we also found negative correlations between highway density and breeding-site occupancy at extents ranging from 3 km – 4 km, which suggests meta-population dynamics are undermined by highway development. Our research supports a growing body of evidence that habitat loss and fragmentation are major factors contributing to anuran population declines