Dynamic occupancy in a peripheral population of Myotis septentrionalis during disease outbreak

Context Once common, the northern long-eared bat ( Myotis septentrionalis ) has experienced declines > 90% due to white-nose syndrome (WNS). This severe mortality has resulted in their listing under the Endangered Species Act and made them a conservation priority. However, their broad distribution, variability in habitat use, and population instability make developing range-wide conservation strategies difficult. To understand how conservation of M. septentrionalis may vary across its range, we examined shifts in M. septentrionalis site occupancy on the edge of their historic range following severe WNS mortality. Objectives Our goal was to determine how mortality from white-nose syndrome affects site occupancy of M. septentrionalis at the southern edge of their historic range. Understanding which areas will remain occupied during disease related mortality will help inform management during disease outbreaks. Methods We used 11 years (2007–2017) of mist-netting records to construct dynamic occupancy models of M. septentrionalis in northern Georgia, USA. Occupancy was updated annually, with the initial period defined as all years prior to the winter of 2012–2013, which corresponds to when white-nose syndrome entered the study area. We assessed occupancy using landscape metrics at the home range (65 ha) and landscape (491 ha) scales, year since WNS occurrence in the study area, distance to karst, and distance to nearest WNS positive county. We estimated probability of detection using Julian date and sampling effort. Results Initial site occupancy was positively associated with percent deciduous forest cover at the home range scale. As M. septentrionalis populations declined from WNS, a site becoming unoccupied was negatively correlated with mean contiguity of forest at the home range scale and largest patch of forest and mean elevation at the landscape scale. Site occupancy declined precipitously in the years following WNS, dropping from 70.75% (41.76–96.98% [95% CI]) occupancy pre-WNS to 0.3% (0.3–20.5% [95% CI]) by the final year of the study. Conclusions Our results demonstrated that sites occupied by M. septentrionalis closer to the historic range edge were more vulnerable to becoming unoccupied after disease arrival. We recommend managers on the periphery of the historic range focus conservation efforts on high elevation forested areas nearer to the range core.