Wind-energy development had led to large-scale mortality of a poorly understood clade of migratory North American bats: Lasiurus cinereus, L. borealis, and Lasionycteris noctivagans. Understanding potential impacts on bat populations requires an understanding of the migratory behaviors and migration pathways used by these species, information that is largely unknown for these cryptic species. Taking advantage of spatial gradients in the hydrogen isotope (δ2H) values of precipitation and known relationships between δ2H values precipitation and bat hair, we inferred the summer origins of these species via analysis of their hair δ2H values within a Bayesian framework. Samples of bat hair were obtained from live-captured, turbine-killed, and museum-specimen individuals spanning the broadest latitudinal, longitudinal, and seasonal range possible, and pooled with hydrogen isotope hair values from the literature (Lasiurus cinereus n = 1106, L. borealis n = 685, Lasionycteris noctivagans n = 209). Continuous probability surfaces indicating the likely origin of each individual were compared pairwise using metrics of niche overlap, and hierarchical clustering used to define common summer habitat. Assignment of bats as locals or probable migrants was determined using an odds-ratio approach, which combined with data on the date of mortality enabled inference of seasonal migration timing and intensity.
Results/Conclusions
Mean δ2H hair values (transformed to precipitation δ2H values) were significantly different among species (p = 1.8e-10), which indicates differences in summer habitat amongst species across their entire ranges. Spatial similarity of summer origins was also more similar amongst species (p-values ≤ 4.8e-14 from TukeyHSD tests of similarity values), and the mean similarities of L. cinereus origins distinct from the other species (p-values ≤ 1.0e-6), indicating distinct summer habitat uses and the superiority of comparing the spatial probability surfaces rather than δ2H hair values. The majority of bats (98%, α = 0.95) were successfully grouped by similar summer origin using bootstrapped complete-method hierarchical clustering; such groupings of spatial origin were correlated to migratory destination (i.e. sampling region) with respect to species (Lasiurus cinereus Pearson’s Chi-Squared test p < 2.2e-16, L. borealis p = 2.7e4, Lasionycteris noctivagans p = 4.6e-13), strongly indicating the presence of spatial structure in the migratory behaviors of these species. Ongoing work will explore seasonal patterns in the arrival of migrating bats and connectivity patterns between probable origin and migration routes. This approach will illuminate migratory patterns and help inform efforts to conserve these species in the face of emerging threats, including wind-energy development.