Unraveling long-term trends and drivers of fish biodiversity change using eDNA metabarcoding of archived samples

Long-term biodiversity data are critical for detecting and understanding ecosystem change, yet systematic long-term datasets remain scarce. We present the longest eDNA metabarcoding time series for freshwater fishes based on archived suspended particulate matter (SPM) from Germany's Environmental Specimen Bank, covering 17 years (2005–2021) across six major European river systems (Rhine, Danube, Elbe, and tributaries). From 211 annual samples collected according to highly standardized procedures at 13 sites, we detected 63 fish and lamprey species and traced their spatiotemporal trajectories. eDNA metabarcoding captured significantly declining trends at multiple sites with no recoveries. Common and sensitive species declined at nearly half of sites while invasive species like round gobies expanded consistently. Site occupancy and relative read abundance trends were correlated, supporting their use as semi-quantitative biodiversity indicators. Alpha and beta diversity shifted significantly over time and Bayesian hierarchical models revealed site specific trends and that 11% of species showed significant declines in relative read abundance at the national level. The eDNA-derived trends showed strong concordance with long-term regulatory fish monitoring trends, supporting that eDNA metabarcoding data can generate comparable bioindicaton results. Random forest models and multivariate analyses identified anthropogenic stressors such as pesticides, heavy metals, temperature, and fine sediment as key correlates of biodiversity patterns, with site-specific stressor combinations explaining 29% of community variation. Critically, these chemical and biological signals derive from the same archived SPM samples, demonstrating a dual-purpose monitoring matrix that integrates pollution and biodiversity assessment simultaneously. Archived SPM samples thus represent an underexploited resource for reconstructing freshwater biodiversity baselines and supporting evidence-based environmental management.