| Efficient sampling of a large proportion of a population is a key feature of a powerful trend monitoring program. Here, we use adaptive sampling to revise a monitoring program for Townsend's big-eared bats (Corynorhinus townsendii) hibernating in volcanic caves at Lava Beds National Monument in northeastern California. Although dozens of known cave hibernacula are surveyed annually, additional effort to locate new hibernation sites is limited by the number (n = 423) of unsurveyed caves. Abundance of bats within hibernacula is negatively correlated with cave temperature (n = 31 caves over 4 years), but comprehensive temperature measurement is not feasible for so many caves. Instead, we investigated the connection between cave morphology and microclimate to identify caves with cold, stable temperatures potentially used by hibernating bats. We built and validated a model using direct measurements of cave morphology (trench depth, entrance area, passage constriction area) to predict bat abundance and assessed model performance at a coarser scale, assigning cave types (bridge, surface tube, talus, and lava tube). Hibernacula surveys completed in January 2018 will validate the predictive model, potentially influencing site selection for future surveys. Improvement in power and efficiency of bat population monitoring is critical to understanding potential impacts of the fatal disease white-nose syndrome. |