Change in diversity patterns of fish by cascade dams: comprehensive dataset of eDNA and traditional evidence from the Jinsha River.

The large-scale construction of cascade dams worldwide has profoundly altered river hydrological regimes and habitats, producing significant threats to fish diversity. This study integrated environmental DNA (eDNA) and traditional fishing data to construct a fish dataset from the middle and upper reaches of the Jinsha River (2003-2024). It was used to systematically reveal the spatiotemporal patterns of fish diversity and analyze key environmental drivers. Additionally, the Isolation Forest algorithm was employed to remove 29 outliers from the historical data. Distance-based redundancy analysis showed that flow velocity was the strongest environmental driver for species composition (R2 = 0.51, p = 0.005) and functional composition (R2 = 0.41, p = 0.018). This suggests that flow velocity reduction, closely associated with cascade dam construction, may act as an important agent of environmental filtering. β-diversity analysis revealed that local-scale functional nestedness (β-nes: 37.4%) and a directional functional shift toward resident and limnophilic groups suggest incipient functional homogenization in the cascade dam system, although temporal evidence was lacking due to staggered dam construction. Notably, although species richness increased, functional evenness (FEve) declined significantly, revealing an imbalanced functional diversity despite higher taxonomic diversity. These findings enhance our understanding of the long-term ecological impacts of cascade dams and provide a scientific basis for mitigating incipient biodiversity homogenization induced by hydropower development.