Sediment Grain Size Distribution under Quasi-Unsteady Flows through River Reaches

This study explores the temporal and spatial dynamics of sediment transport and bed morphology under quasi-unsteady flow conditions, with an emphasis on the mean sediment grain size (d50). The experiments were conducted in an 18-meter-long, 1-meter-wide, and 1-meter-deep laboratory flume with a mixed sediment supply. Four sediment feeding scenarios were tested: no feed, constant feed, rising limb feed, and falling limb feed, under a symmetric hydrograph comprising seven flow stages. Each stage lasted one hour, with discharges ranging from 50 to 100 L/s. Data were collected to analyze temporal variations in d50 and the influence of discharge on sediment sorting. Comparative analyses of sediment transport during rising and falling limbs revealed distinct behavioral patterns, with flow deceleration promoting deposition. Hysteresis loops highlighted temporal asymmetries between accelerating and decelerating flows, emphasizing the critical role of flow history in shaping bed composition. Bed stability assessments indicated that rapid discharge changes induce transient instability, evidenced by increased d50 variability during abrupt transitions. However, the bed exhibited resilience as flow conditions stabilized. A linear regression model demonstrated the ability to estimate d50 as a function of discharge and time, offering preliminary insights into sediment dynamics. However, limitations inherent to linear models- such as their inability to capture nonlinear interactions- suggest that advanced machine learning approaches could improve predictive accuracy. By integrating empirical analysis and predictive modeling, this study advances sediment forecasting capabilities under variable hydraulic conditions, providing valuable insights for river management and sediment transport processes.