Abstract:Seagrass beds and coral reefs, as highly productive nearshore ecosystems, play a critical role in maintaining regional fish diversity and ecological functions through their ecological connectivity. This study focused on the seagrass bed-coral reef continuum in Wenchang, Hainan, China. Using environmental DNA (eDNA) metabarcoding technology, water samples were collected in June, August, and October 2021 to assess fish diversity and identify potential connectivity species between these two habitats. A total of 193 fish species were detected, with 133 species shared between both habitats. Seagrass beds and coral reefs harbored 41 and 19 unique species, respectively. Community composition analysis revealed significant differences in order- and family-level abundance distributions between habitats, though both were dominated by Gobiiformes. The Chao1 index was higher in seagrass beds, indicating greater potential species richness, while coral reefs exhibited superior Pielou, Shannon, and Simpson indices, reflecting higher species evenness and diversity. Principal Coordinates Analysis (PCoA) and PERMANOVA tests confirmed significant structural divergence in fish communities between habitats (R2= 0.356, P = 0.01). LEfSe analysis identified Lutjanus fulviflamma and Gerres oyena as indicator species for seagrass beds (LDA > 3.0), whereas Trachinotus mookalee and Lutjanus argentimaculatus were indicators for coral reefs. Co-occurrence network analysis highlighted Mugil cephalus as having the highest betweenness centrality in both habitats, suggesting its role as a key connector. Additionally, M. cephalus, Gerres filamentosus, and Hyporhamphus quoyi formed cross-habitat co-occurrence networks, potentially serving as critical linkage species. These findings demonstrate that eDNA technology serves as a robust supplementary tool to traditional fish surveys, offering distinct advantages in identifying ecological connectivity species and deciphering habitat linkage mechanisms. This study provides critical data and methodological references for fish resource assessments and science-based conservation of coastal ecosystems.
