Acoustic Characteristics and Seasonal Variations of Photosynthetic Sounds in Coral Reefs of Dongshan Island, China

The acoustic contributions of primary producers to ocean soundscapes are poorly understood compared to those of marine animals. This study investigated acoustic signals generated by photosynthetic bubbles produced by algae in a coral reef ecosystem on Dongshan Island, China. Using continuous passive acoustic monitoring (PAM), we recorded bubble sounds and applied a Python-based code to extract candidate signals, which were then classified using a trained neural network. Subsequent analysis revealed distinct seasonal patterns in signal rates, with the highest in summer (103.2 pulses/30 min), followed by spring (60.4) and winter (41.3). Data from autumn (4.2) were considerably lower and were excluded from interseasonal comparisons because the recorder was deployed far from the reef. The highest rates of acoustic detection often occurred at night. We ascribed this phenomenon to the production of bubbles of comparable sizes via other processes and to the delayed detachment of photosynthetic bubbles, modulated by additional rhythmic factors. We developed a ridge regression model using environmental parameters to examine photosynthetic events during the day, showing that signals directly related to photosynthesis peaked around midday and were positively correlated with dissolved oxygen (DO) concentrations on most sampled days, thereby partly supporting the predictive model. This work may advance research on assessing ebullitive oxygen fluxes from primary producers and offers a novel, noninvasive tool for the rapid, continuous monitoring of ecosystem health.