Abstract
Harbour porpoises are common in European shelf waters and are sensitive to anthropogenic disturbance. Increasing noise levels from industrial developments may pose a risk for this small and cryptic species. We model data from 8-years of visual and Passive Acoustic Monitoring (PAM) between 2009 and 2016 to investigate the effect of multiple covariates, including construction-related acoustic surveys, on the occurrence of harbour porpoises in an Irish bay. We compare outputs for (1) visual-only, (2) corrected-visual, and (3) PAM models. Comparison of concurrent visual and acoustic monitoring suggests visual surveys underestimated harbour porpoise occurrence by 108%. Visual-only and corrected-visual models agreed on the effect of seasonality and yearly variability, while the importance of sampling site was not significant. In both models higher sea state, low tide and the presence of acoustic surveys had a negative influence on harbour porpoise presence. PAM enabled data to be collected during unsuitable weather or darkness. The PAM model better predicted porpoise seasonal occurrence and showed higher sighting probability near deeper waters. This study highlights the value of complementary monitoring methods and PAM’s added-value in locations where visual surveys are limited by environmental conditions and when working with cryptic species.