Cooling of bat hibernacula to mitigate white-nose syndrome

White-nose syndrome (WNS) is a fungal disease that has caused precipitous declines in several North American bat species, creating an urgent need for conservation. We examined how microclimates and other characteristics of hibernacula have affected bat populations following WNS-associated declines and evaluated whether cooling of warm, little-used hibernacula could benefit bats. During the period following mass mortality (2013-2020), we conducted 191 winter surveys of 25 unmanipulated hibernacula and 6 manipulated hibernacula across Pennsylvania (USA). We joined these data with additional datasets on historical (pre-WNS) bat counts and on the spatial distribution of underground sites. We used generalized linear mixed models and model selection to identify factors affecting bat populations. Winter counts of Myotis lucifugus were higher and increased over time in colder hibernacula (those with midwinter temperatures of 3-6 °C) compared with warmer (7-11 °C) hibernacula. Counts of Eptesicus fuscus, Myotis leibii, and Myotis septentrionalis were likewise higher in colder hibernacula (temperature effects = -0.73 [SE 0.15], -0.51 [0.18], and -0.97 [0.28], respectively). Populations of M. lucifugus and M. septentrionalis increased most over time in hibernacula surrounded by more nearby sites, whereas Eptesicus fuscus counts remained high where they had been high before WNS onset (pre-WNS high count effect = 0.59 [0.22]). Winter counts of M. leibii were higher in hibernacula with high vapor pressure deficits (VPDs) (particularly over 0.1 kPa) compared with sites with lower VPDs (VPD effect = 15.3 [4.6]). Counts of M. lucifugus and E. fuscus also appeared higher where VPD was higher. In contrast, Perimyotis subflavus counts increased over time in relatively warm hibernacula and were unaffected by VPD. Where we manipulated hibernacula, we achieved cooling of on average 2.1 °C. At manipulated hibernacula, counts of M. lucifugus and P. subflavus increased over time (years since manipulation effect = 0.70 [0.28] and 0.51 [0.15], respectively). Further, there were more E. fuscus where cooling was greatest (temperature difference effect = -0.46 [SE 0.11]), and there was some evidence there were more P. subflavus in hibernacula sections that remained warm after manipulation. These data show bats are responding effectively to WNS through habitat selection. In M. lucifugus, M. septentrionalis, and possibly P. subflavus, this response is ongoing, with bats increasingly aggregating at suitable hibernacula, whereas E. fuscus remain in previously favored sites. Our results suggest that cooling warm sites receiving little use by bats is a viable strategy for combating WNS.