Acoustic observations reveal diel behaviour and local abundance in white-edged rockfish ( Sebastes taczanowskii )

Acoustic observations of fish sounds provide a unique window into diel behaviour and occurrence yet their quantitative interpretation remains challenging. Linking sound production to both population size and diel behavioural patterns is complicated by environmental variability, temporal autocorrelation, and observation error inherent in acoustic data. In this study, we address these challenges by examining the relationship between sound production and abundance in commercially important white-edged rockfish in coastal Japan using a Bayesian state-space model that explicitly incorporates time-of-day effects on sound production. Our findings show that sound production increased around dusk and dawn, whereas the estimated number of individuals peaked around dusk. The regression coefficient linking fish sounds to estimated individual counts was positive during most time periods, indicating that the rate of sound production has strong potential as an abundance index for estimating population size. This state-space modelling approach demonstrates the potential to infer diel behavioural patterns and indices of local abundance from daily acoustic observations and offers a flexible framework applicable to a wide range of fish sound datasets.