Spatially explicit population estimates for black bears based on cluster sampling

ABSTRACT We estimated abundance and density of the 5 major black bear ( Ursus americanus ) subpopulations (i.e., Eglin, Apalachicola, Osceola, Ocala‐St. Johns, Big Cypress) in Florida, USA with spatially explicit capture‐mark‐recapture (SCR) by extracting DNA from hair samples collected at barbed‐wire hair sampling sites. We employed a clustered sampling configuration with sampling sites arranged in 3 × 3 clusters spaced 2 km apart within each cluster and cluster centers spaced 16 km apart (center to center). We surveyed all 5 subpopulations encompassing 38,960 km 2 during 2014 and 2015. Several landscape variables, most associated with forest cover, helped refine density estimates for the 5 subpopulations we sampled. Detection probabilities were affected by site‐specific behavioral responses coupled with individual capture heterogeneity associated with sex. Model‐averaged bear population estimates ranged from 120 (95% CI = 59–276) bears or a mean 0.025 bears/km 2 (95% CI = 0.011–0.44) for the Eglin subpopulation to 1,198 bears (95% CI = 949–1,537) or 0.127 bears/km 2 (95% CI = 0.101–0.163) for the Ocala‐St. Johns subpopulation. The total population estimate for our 5 study areas was 3,916 bears (95% CI = 2,914–5,451). The clustered sampling method coupled with information on land cover was efficient and allowed us to estimate abundance across extensive areas that would not have been possible otherwise. Clustered sampling combined with spatially explicit capture‐recapture methods has the potential to provide rigorous population estimates for a wide array of species that are extensive and heterogeneous in their distribution. © 2017 The Wildlife Society.