Bioacoustics (2018)

DOI: 10.1080/09524622.2018.1477071


Automatic click detectors and full-bandwidth sound recorders are widely used in passive acoustic monitoring of small cetaceans. Detection of these signals depends on a variety of factors, including signal to noise ratio. Passive acoustic monitoring is often used to study impact of underwater noise on small cetaceans, but as detection probability is affected by changes in signal to noise ratio, variable noise levels may affect conclusions drawn from these experiments. Therefore, we examine how different detectors and filters perform in varying ocean noise conditions. C-PODs and full-bandwidth recorders (Wildlife Acoustics, SM2M+) were deployed at two stations in an environment with fluctuating ambient noise for 42 days. Noise level and harbour porpoise (Phocoena phocoena) click trains simultaneously recorded on both loggers were compared. Overall, we found that porpoise click detections by the algorithm used to analyse full-band recorder data (Pamguard) paralleled detections by the C-POD. However, Pamguard detected significantly more clicks than the C-POD. A decrease in detections was seen for both loggers with increasing noise in the band 20 –160 kHz, in particular for levels above 100 dB re 1μPa rms. We also found that the Pamguard detection function changed the least over varying noise conditions when compared to the C-POD detectors. This study sheds light on the fact that inference of animal presence/absence or density that are based on echolocation cues (here, Porpoise Positive Minutes) shall account for the acoustic environments where probability of detecting signals may be affected by variability in ambient noise levels.