Sediment transport modeling for run-of-river hydropower in the Madeira River: Calibration with conventional and remote sensing data

Abstract

This study aims to evaluate sediment transport processes in the Madeira River, a high-load Amazon tributary altered by the Jirau run-of-river hydropower dam. A methodology for sensitivity analysis and calibration of the HEC-RAS one-dimensional morphodynamic model is developed. It integrates multiple model to measured comparisons, including conventional monitoring and water color remote sensing data. The study underscores the value of employing products derived from satellite imagery, refining model differentiation and improving the spatial and temporal resolution of sediment transport predictions. A simple, regionally significant method of estimating depth-integrated concentrations form surface index concentrations is discussed, showing that for high concentrations a 1.10–2 multiplicative factor suffices. Sensitivity analysis highlights the dominant influence of sand content in the upstream sediment load and the necessity of using the Krone–Partheniades transport formula to simulate fine sediment retention. The calibrated model estimates a sediment retention efficiency of 21.3% in the backwater-affected reach over a five-year period, with over 90% of the sand fraction being deposited. Results suggest that the wash load threshold for this system is medium to coarse silts and clay-silt flocs larger than 0.016 mm. These are the key size classes to understand deposition of fines. Flocculation processes may play a role, requiring adjustments in the input sediment load grain size distribution. A multivariate sediment rating curve, incorporating tributary discharge dynamics, enhances model performance, particularly in reproducing seasonal concentration variations in the backwater reach. These findings provide insights into the best practices for sediment modeling in high-load rivers impacted by hydropower and highlight the importance of multi-objective calibration approaches.