Recording on the fly: Can acoustic monitoring track bumble bees (Bombus spp.) in remote habitats?

Background/Question/Methods:
Species worldwide are migrating in response to climate change and habitat destruction. Acoustic monitoring has historically been used to monitor behavior and biodiversity of sound producing animals, but could also track migration patterns in response to anthropogenic pressures. Bumble bees (Bombus spp.), which are keystone pollinators in many habitats, are migrating into higher altitudes in response to climate change. Many are also experiencing population declines, making non-invasive methods ideal. Bombus species exhibit characteristic pollination buzz frequencies, suggesting that buzz frequencies may provide a means of identifying them on the fly. Here we test the effectiveness of acoustics for monitoring two dominant alpine species (B. balteatus and B. sylvicola.) We captured individual bees foraging in alpine meadows at the Niwot Ridge Long Term Ecological Research Site (CO, USA), identified them to species and caste, and measured their wing length and intertegular distance to assess body size. We then released them into a 1x1x0.5m flight cage with five equidistant microphones and recorded them as they flew among flowers. Wind speed was recorded to determine if it would affect our ability to detect buzzes. The recordings were filtered and analyzed using Audacity© to identify the fundamental frequency of each species and caste.

Results/Conclusions:
Resident alpine Bombus species and castes differed in fundamental buzz frequency in flight. Discriminant analysis indicates that queens of each species were correctly classified by their flight buzzes 100% of the time. While castes differed in fundamental buzz frequency within species, workers were less reliably assigned between species, with many individuals being incorrectly classified. Results suggest that acoustic monitoring for these alpine species would be most effective early in the flowering season when queens are the predominant foragers. Wing length explained more variation in flight buzz frequency than intertegular distance, with buzz frequency decreasing as wing length increased. Differences between species are only marginally significant when wing length is included as a covariate, indicating that wing length is the primary driver of species-specific buzz frequencies. Winds of low to moderate speed did not obscure the fundamental frequencies of the bee’s flight buzz. Future work will elucidate the validity of acoustic methods for estimating pollinator density in more diverse communities. Our results illustrate the potential for using flight buzz frequency to determine the composition and size diversity of alpine bumble bee communities. If such frequencies are unique across taxa, acoustic monitoring has the potential to track mobile, declining and elusive animals.