Abstract
<title>Abstract</title> <p>Background Coastal bays experience strong hydrographic and nutrient variability that can reorganize microbial biodiversity and alter carbon transfer through the microbial food web. We investigated Dinghai Bay, a mariculture-influenced subtropical coastal bay in southeastern China, by integrating environmental DNA metabarcoding, community assembly analysis, co-occurrence network inference, and dilution-based measurements of phytoplankton and bacterioplankton growth and microzooplankton grazing. Results Bacterial 16S and microeukaryotic 18S rRNA gene datasets revealed clear seasonal changes in diversity and community composition, and these compositional differences were consistently associated with hydrographic and nutrient-related variables. Temperature, pH, and dissolved inorganic nitrogen showing particularly strong relationships. Neutral community model analyses suggested that bacterial assemblages were more compatible with neutral expectations than microeukaryotic assemblages, which showed stronger departures from the model. Integrated co-occurrence networks were highly connected, strongly modular, and dominated by positive associations, with connectivity peaked in summer. Dilution experiments showed pronounced seasonal variation in microbial food-web functioning: phytoplankton primary production and bacterioplankton production were highest in summer, whereas producer-derived carbon flow efficiency was generally lower in summer and higher in winter. Bacterioplankton-derived carbon flow efficiency also tended to be reduced in summer. Across all seasons, phytoplankton production contributed substantially more grazing-mediated carbon transfer than bacterioplankton production. Conclusions Seasonal environmental forcing restructured microbial biodiversity, assembly signals, co-occurrence organization, and grazing-mediated carbon transfer in Dinghai Bay. Together, these results show that microbial community structure and carbon-flow dynamics vary in parallel across seasons, while rate-based carbon transfer remains less tightly coupled to broad community patterns than community composition itself.</p>