Knock Knock, who's there? Identifying wild species-specific fish sounds with passive acoustic localization and random forest models

This dataset accompanies a publication that identifies wild rocky reef fish sounds from British Columbia, Canada using a passive acoustic localization array paired with underwater video. We localized fish sounds using cross correlation of acoustic transients and time difference of arrival. We manually matched sound coordinates to specific fish species from underwater video. We then extracted 47 sound features from each fish call to identify species-specific fish sounds using random forest models. We identified fish calls for eight fish species. Representative sound clips for each species are included in the repository folder fish_sound_audio_samples. We have also include our full dataset and all R and python scripts we used to analyze and visualize our data.