Search

The sediment transport rate refers to the quantity of sediment moved by natural processes such as water flow, wind, or ice within a specific timeframe. Rivers, being dynamic systems, undergo constant changes in hydraulics and channel morphology, leading to significant variations in sediment transport rates. The complexity of these dynamics makes it challenging to establish linear relationships between input factors and sediment transport rates characteristics. To address the non-linearity and multi-dimensional behaviour, this study adopts a comprehensive approach incorporating techniques like data mining, using various algorithms such as Natural Gradient Boosting, Random Forest Regressor, Decision Tree Regressor, Multi-Layer Perceptron, Extreme Gradient Boosting, and Extra Tree Regressor. In the context of sediment transport rate characteristics prediction, both bed load prediction and friction factor prediction are essential components. Bed load prediction estimates the sediment that moves along the bed of a water body through processes like rolling, sliding, and saltation, which is crucial for determining the overall sediment transport rate. Similarly, friction factor prediction pertains to the roughness elements on the bed surface that affect water flow resistance. Therefore, precise sediment transport rate predictions enhance our understanding of channel morphology, bed material size, and flow characteristics. To evaluate the performance of the models, several metrics were employed, including the Pearson Correlation Coefficient (PCC), Root Mean Square Error (RMSE), R-squared (R2), Mean Square Error (MSE), and Nash–Sutcliffe Efficiency (NSE). The study's findings reveal that the Random Forest Regressor method is particularly effective, achieving high accuracy in predicting bed load rates, with a Correlation Coefficient (CC) of 0.957, RMSE of 0.239, R2of 0.913, MSE of 0.057, and NSE of 0.913. Moreover, Extra Tree Regression offers the best prediction among the suggested models for friction factor prediction, as evidenced by performance metrics such as RMSE of 0.287, MSE of 0.083, NSE of 0.909, CC of 0.975, and R2of 0.909. These results provide valuable insights for authorities, enabling better anticipation and mitigation of flood risks caused by sedimentation in rivers.