Résumé
According to the acoustic niche hypothesis, the soundscape is a finite resource, theoretically implying that soniferous species whose sounds compose it within a given habitat must share this resource to minimize competition. However, depth-related patterns in this partitioning are still poorly understood. This study aims to compare the diel dynamics and realized acoustic niches of fish sounds at two depths (60 and 120 m) in mesophotic coral reefs of the Tuamotu Archipelago, French Polynesia, over a 62-h period. Two complementary objectives are pursued: (1) to investigate depth-related patterns in the diel cycle of fish sounds using generalized additive models, and (2) to assess whether their realized acoustic niches change between depths. Although most sound types were shared between 60 and 120 m, the study revealed that the dynamics of fish sounds varied significantly between the two depths. It also supported clear diel patterns. As depth increased, frequency-modulated sounds became more prominent at night, while the opposite trend was observed for pulse series sounds. At the community level, acoustic activity was more concentrated during short time periods at night at 120 m. Different sound types were either produced at distinct times or differed in pulse period range or frequency spectrum during periods of temporal overlap. The structure of realized acoustic niches also differed between the two depths. The consistent increase in the abundance of frequency-modulated sounds under reduced light conditions (both at night and in deeper waters) may reflect the need for more stereotyped acoustic signals, suggesting a potential adaptation to low-light environments.