Wind turbines (WTs) frequently kill bats worldwide. During environmental impact assessments, consultant ecologists often use automated ultrasonic detectors (AUDs) to estimate the activity and identity of bats in the zone of highest mortality risk at WTs in order to formulate mitigation schemes, such as increased curtailment speeds to prevent casualties. While acknowledging the potential of acoustic monitoring, we evaluate the limitations of AUDs for monitoring bats at WTs and highlight directions for future research. We show that geometric and atmospheric attenuation of ultrasonic echolocation calls, in conjunction with limited sensitivity of ultrasonic microphones, severely constrain detection distances of bats at WTs. Taking into account the acoustic shadow produced by the nacelle, AUDs cover only approximately 23% of the risk zone for a bat calling at 20 kHz and 4% for a bat calling at 40 kHz, assuming a 60 m blade length. This percentage will further decrease with increasing blade lengths in modern WTs. Additionally, the directionality of echolocation calls and the dynamic flight behaviour of bats constrain the detectability of bats. If a call can be detected, the low interspecific and high intraspecific variation of echolocation call characteristics may impair species identification, limiting the power to predict population-level effects of fatalities. We conclude that technical, physical, and biological factors severely constrain acoustic monitoring in its current form. We suggest the use of several AUDs, installed at complementary sites at WTs, and the testing of other techniques, such as radar, cameras, and thermal imaging, to inform stakeholders on the mortality risk of bats at WTs.

Attenuation of sound pressure level (SPL) of echolocation calls as they propagate through air. Estimates were developed for two bats calling at 20 kHz and 110 dB peSPL and 40 kHz and 104 dB peSPL (source SPLs are referenced to 20 µPa at 1 m in front of the bat). The received SPLs were calculated from source SPL by subtracting geometric attenuation (-6 dB per doubling of distance) and frequency- and weather-dependent atmospheric attenuation (see insets on the right).  The received SPLs are calculated for the full range of atmospheric attenuation, 

Percentage of the rotor-swept area of wind turbines covered by automated ultrasonic detectors (AUDs) with increasing blade length, for bats calling at 20 kHz  and 40 kHz. For this estimate, we assumed the risk zone to equal the circular rotor-swept area with the blade length as radius. AUDs usually do not cover the upper half of this zone, because AUDs are installed at the nacelle bottom (between tower and blades) with the microphone pointing downwards. We further used the maximum distance at which bats can be detected by AUDs (trigger threshold: 60 dB SPL), which only applies for bat calls that are directed towards the AUD.