Multi-media eDNA Metabarcoding Uncovers Habitat-Specific Fish Community and Monitoring Efficiency: Implications for Ecosystem Management in a River-Lake System (Honghu Lake Basin, China)

​​ Environmental DNA (eDNA) metabarcoding is widely used for monitoring aquatic biodiversity, yet its efficiency across different ecosystems and sample types requires further evaluation. This study assessed fish diversity using eDNA from water and sediment in the river-lake system of China’s Honghu Basin. Results demonstrated that: (1) Water samples from rivers yielded the highest taxonomic richness (7 orders, 37 species), significantly surpassing that of lake water (7 orders, 17 species), river sediment (3 orders, 9 species), and lake sediment (3 orders, 7 species). (2) Sediment eDNA consistently detected a higher proportion of benthic/demersal species than water eDNA in both ecosystems.(3) Riverine communities exhibited greater α-diversity and more complex structure than those in lakes. (4)Linear Discriminant Analysis (LDA) identified Pseudorasbora parva and Carassius auratus as key species in the river, whereas Hypophthalmichthys nobilis was the key species in the lake. (5)Redundancy Analysis (RDA) showed that fish communities in river water were significantly correlated with Total Nitrogen (TN), Nitrate-Nitrogen (NO₃-N), and Total Phosphorus (TP), while communities in lake water were primarily associated with Turbidity (Turb), pH, Chlorophyll-a (Chl.a), and Chemical Oxygen Demand (COD). These findings offer a theoretical basis for optimizing eDNA-based biodiversity monitoring and support the conservation of freshwater ecosystems.