Using occupancy estimation to assess the effectiveness of a regional multiple-species conservation plan: Bats in the Pacific Northwest Original Research Article

Regional conservation plans are increasingly used to plan for and protect biodiversity at large spatial scales however the means of quantitatively evaluating their effectiveness are rarely specified. Multiple-species approaches, particular those which employ site-occupancy estimation, have been proposed as robust and efficient alternatives for assessing the status of wildlife populations over large spatial scales, but implemented examples are few. I used bats as a model to evaluate design considerations for the use of occupancy estimation to assess population status and habitat associations for eight species of bats covered under a regional conservation plan. Bats were one of the groups expected to benefit from a system of reserves for species associated with late-successional/old-growth (LSOG) habitat designated under the Northwest Forest Plan (NWFP). This study produced the first estimates for probabilities of occupancy and detection for bats at a regional scale. Overall probabilities of occupancy ranged from 0.586 to 0.783 and probabilities of detection ranged from 0.239 to 0.532 among the eight species modeled. Although point estimates of occupancy suggested association with NWFP habitat categories for some species, estimate precision was low. Models that assumed constant occupancy with respect to reserve- and LSOG-status were supported for most species. I used model-averaged estimates of occupancy and detection for each species to estimate survey effort necessary to meet precision targets. Occupancy estimation was best suited to species with the highest detection probabilities. Species that are rare or difficult to detect will require enhancement in survey methods or more intensive survey effort to produce meaningful estimates. Optimizing monitoring efforts to address multiple species requires tradeoffs among survey methods, levels of effort, and acceptable levels of precision.