Seismic-based Estimation of River Discharge in a Low-flow Tributary: Field Evidence from Riverside Monitoring

Accurate discharge monitoring in low-flow rivers remains challenging due to the limitations of conventional gauging stations and the reduced reliability of remote sensing at small spatial scales. This study investigates the feasibility of estimating river discharge using riverside seismic monitoring as a non-contact hydrometric approach in a low-flow river system. Four field experiments were conducted along a low-discharge tributary (approximately 0.5 m³/s) in southern China, combining continuous seismic observations with synchronous flow velocity measurements. Time–frequency analyses indicate that riverinduced seismic signals span a broad frequency range (2–50 Hz), while seismic energy within the 2–10 Hz band consistently dominates under stable, flow-dominated conditions with negligible sediment transport. Temporal variations in low-frequency seismic energy closely follow measured discharge variations, indicating a strong linkage between turbulence-driven flow processes and seismic signal generation. Based on this relationship, an empirical linear regression model was developed to estimate discharge from seismic power spectral density. Model calibration and validation demonstrate that predicted discharge values agree well with measured values in flow-dominated river sections, with relative errors generally within 10.3%. In contrast, model performance degrades in reaches affected by bed irregularities and sediment–bed interactions. These results demonstrate that riverside seismic monitoring can provide a practical and low-cost alternative for discharge estimation in low-flow rivers when applied under appropriate hydraulic conditions, offering potential benefits for water resources monitoring in small and data-scarce catchments.