Temporal Tracing of Pollution Source Using a Continuous eDNA Approach

Effective watershed pollution management is often hampered by temporal variations in hydrological conditions. However, most existing studies rely on static scenario assumptions facilitated by nucleic-acid sequencing source tracking tools, limiting their applicability to dynamic pollution events. Here, we developed a time-series pollution source tracing framework by leveraging high-resolution environmental DNA (eDNA) data obtained through a field-compatible eDNA system, incorporating both public and local DNA source libraries. During the 2024 flood season (June 13-July 15) in the Yan Tietang watershed, we deployed an eDNA autosampler to collect 65 semidaily water samples, quantifying temporal contributions from six major pollution sources. The eDNA autosampling system was validated, demonstrating microbial community richness and structure remained stable over 21 days across different storage temperatures (4 °C, 25 °C, 35 °C). Using the public-source library, wastewater treatment plants (WWTPs) were the dominant source, contributing 60.9% (spatial) and 54.8% (temporal) of total loads. These contributions were temporally correlated with key water quality parameters, such as NH₃-N and TP. Further validation with the local-source library confirmed source-specific signals (Pearson R = 0.83), and strengthened the association between WWTP effluents and regulatory exceedances. This study highlights the potential of high-frequency eDNA monitoring for real-time pollution diagnostics and adaptive watershed management.