Source identification methods in riverine systems: a combined analysis of transient storage and fractional dispersion

ABSTRACT This study investigates source identification using different transport models. A data series, operated by the United States Geological Survey and gathered from eight monitoring stations over a reach length of 41.45 km along Antietam Creek, is used. In this research, advection–dispersion equation (ADE), transient storage model (TSM), and fractional advection–dispersion equation (FADE) models were compared to identify the most suitable model for accurate source identification and pollutant transport prediction. The statistical analysis indicates the best performance by TSM since it gives the lowest RMSE of 0.52 ppm and the highest NSE of 0.95, as it allows accounting for temporary retention in the storage zone. With an RMSE of 0.52 ppm and a high NSE of 0.95, the TSM represents the solute transport process more accurately than the other models. In this study, a new relationship about injection distance has been presented using dimensional analysis with R2 = 0.94. It is also found that the dispersion coefficient scales with distance as D ∝ x0.7. Sensitivity analysis indicates that the model is most sensitive to the dispersion–advection ratio, having an elasticity of −0.18. Monte Carlo simulation shows that 95% confidence intervals range from ±3.5% for near-field to ±7.2% for far-field predictions.