Evaluating acoustic signals to reduce avian collision risk

Abstract

Collisions with human-made structures are responsible for billions of bird deaths each year, resulting in ecological damage as well as regulatory and financial burdens to many industries. Acoustic signals can alert birds to obstacles in their flight paths in order to mitigate collisions, but these signals should be tailored to the sensory ecology of birds in flight as the effectiveness of various acoustic signals potentially depends on the influence of background noise and the relative ability of various sound types to propagate within a landscape. We measured changes in flight behaviors from zebra finches released into a flight corridor containing a physical obstacle, either in no-additional-sound control conditions or when exposed to one of four acoustic signals. We selected signals to test two frequency ranges (4–6 kHz or 6–8 kHz) and two temporal modulation patterns (broadband or frequency-modulated oscillating) to determine whether any particular combination of sound attributes elicited the strongest collision avoidance behaviors. We found that, relative to control flights, all sound treatments caused birds to maintain a greater distance from hazards and to adjust their flight trajectories before coming close to obstacles. There were no statistical differences among different sound treatments, but consistent trends within the data suggest that the 4–6 kHz frequency-modulated oscillating signal elicited the strongest avoidance behaviors. We conclude that a variety of acoustic signals can be effective as avian collision deterrents, at least in the context in which we tested these birds. These results may be most directly applicable in scenarios when birds are at risk of collisions with solid structures, such as wind turbines and communication towers, as opposed to window collisions or collisions involving artificial lighting. We recommend the incorporation of acoustic signals into multimodal collision deterrents and demonstrate the value of using behavioral data to assess collision risk.