Influence of distance, frequency, habitat and angle of incidence on sound levels in passive acoustic monitoring with <scp>AudioMoth</scp> recorders

Abstract Passive acoustic monitoring (PAM) is a major tool for the study of vocalizing animals. As data can be collected by non‐experts and stored for documentation, PAM appears ideal for environmental monitoring. However, the comparability between recordings from different sites can be affected by differential sound attenuation through the sampled habitat and by the target sounds' incidence angle. These effects depend on the sound frequency and may interact with each other. We measured sound attenuation in open and closed habitat with AudioMoth sound recorders, using playback experiments at frequencies between 1.25 and 50 kHz, covering the typical vocalization range of birds and bats. Sound attenuation was stronger in forest than in open land, especially for ultrasound, resulting in up to 52% higher detection ranges for bats in open land than in forest. In addition, fixed‐distance playback experiments covering the same frequency range in the same habitats showed that the orientation of our single‐channel recorders had a major effect, particularly for ultrasound: Recorded sound levels were up to 7 dB lower for sources behind the microphone at 1.25 kHz, and were 15 dB lower at 40 kHz, the latter corresponding to a more than sixfold difference in distance. Taken together, these effects can lead to substantial variation in the effective detection ranges depending on the recording situation, which needs to be considered for the sampling design and interpretation of results from passive acoustic monitoring. However, it should be possible to alleviate at least the effects of directionality by using multiple AudioMoth per sampling point or by using stereo recorders or devices with external omnidirectional microphones.