Integrating Water Column and Sediment eDNA Metabarcoding Reveals Seasonal Turnover and Habitat-Specific Differentiation of Fish Communities in a Subtropical Coastal Bay.

Environmental DNA (eDNA) metabarcoding provides a non-invasive tool for monitoring marine biodiversity in complex ecosystems. In this study, we applied eDNA techniques to investigate the spatiotemporal dynamics of fish communities in Daya Bay, South China Sea, by analyzing 120 water column and 113 sediment samples collected across four seasons. Our research aimed to characterize the seasonal turnover and habitat-specific partitioning of fish community structures. The results revealed that sediment matrices exhibited significantly higher alpha diversity compared to water samples. However, sediment showed lower community heterogeneity-quantified as within-group Bray-Curtis dispersion-particularly during the warm seasons (spring and summer) compared with the water column. Conversely, during the cold seasons (autumn and winter), sediment matrices showed lower alpha diversity but higher community heterogeneity. Using LEfSe and random forest models, we identified a suite of discriminative indicator taxa that characterized these community shifts. Based on relative read abundance (RRA), Johnius and Pennahia were significantly enriched in the water column, while Cynoglossus and Zebrias were primarily associated with sediment habitats. These findings highlight the complementary roles of water and sediment eDNA in capturing the full spectrum of fish biodiversity. While this study primarily describes observed patterns within a dynamic hydrographic context, it provides foundational data for advancing eDNA-based biomonitoring in subtropical marine ecosystems and supports the development of targeted conservation strategies.