Estimating sediment transport capacity on sloping farmland on the Loess Plateau considering soil particle size characteristics

Abstract

The sediment transport capacity (Tc) is essential for understanding soil erosion and creating soil erosion prediction models. Although many Tc equations exist, they are not universally applicable to sloping farmland on the Loess Plateau. To tailor Tc equations for this region, Tc for four soils (Shenmu Inceptisol (SMI), Shenmu Entisol (SME), Ansai Inceptisol (AS), and Yongshou Alfisol (YS)) from sloping farmland on the Loess Plateau was determined through flume tests with slopes ranging from 15.84 % to 38.39 % and flow discharges of 0.05 to 0.11 m2 min−1. Additionally, appropriate particle size characteristics were selected to distinguish the soil types. The results revealed that the Tc of SME exceeded that of YS, followed by SMI and AS in descending order. The Tc of the four individual soils increased exponentially with flow discharge (Q) and slope (S). Tc for individual soils could be satisfactorily predicted using power functions of mean velocity, shear stress (τ), stream power (ω), or effective stream power (R2: 0.661–0.950, MAE: 0.020–0.110, RMSE: 0.023–0.151), with ω being identified as an optimal hydraulic parameter for simulating Tc. In contrast, the unit stream power was a poor predictor of Tc for individual soils (R2: 0.403–0.700, MAE: 0.035–0.123, RMSE: 0.040–0.161). The median particle size (D50), sorting coefficient, and kurtosis (Kg) exhibited highly significant positive correlations with Tc (P < 0.01). Finally, a quadratic power equation for Q, S, D50, and Kg, along with two ternary power equations for ω (or τ), D50, and Kg were newly constructed, all with satisfactory accuracies (R2: 0.808–0.826, MAE: 0.061–0.063, RMSE: 0.086–0.090). The inclusion of σ and Kg rectified the inaccuracies in predicting Tc for all four soils using Q and S equations or ω (or τ). The results of this study offer valuable insights into the construction of the Tc equations.