Dr.-Ing. Tian Wang

It is known that soil and water loss is serious in Loess Plateau, China. Soil erosion is a serious threat to the ecological environment and social productivity in watershed. The slope is considered as the main source of soil erosion. As the most basic unit of the watershed, slope is usually the beginning of the soil erosion research process. The sediment production mechanism of slope soil erosion and the quantitative study of sediment transport is a difficult and hot topic of international research. The thesis takes loess slope as the research object to develop the soil erosion numerical model. Firstly, the algorithm of hydrological-hydrodynamic numerical model is optimized. Secondly, the response of sediment yield to hydrodynamic parameters is analyzed through the indoor slope erosion experiment, and then the sediment transport capacity is established for loess slope. Finally, high-performance and robust numerical model is established for sediment transport processes simulation on loessal slope based on the optimized hydrological-hydrodynamic numerical model and established slope sediment transport capacity. The high-performance and robust numerical model is applied to simulate the rainfall erosion experiment for verification model. This research obtained the following main conclusions:

(1) A novel friction source term treatment method was developed for overland flow simulation using shallow water flow model, which improved the calculation efficiency of the hydrological-hydrodynamic numerical model and the accuracy of the surface runoff simulation. The mathematical method is applied to change the implicit solution into the explicit scheme for the friction source term in the shallow water equation by the quadratic equation, which improve calculation efficiency of the friction source term. Compared with the fully implicit algorithm, the proposed algorithm can save 45.16%-93.55% friction calculation time. The proposed algorithm function can be flexibly integrated into the shallow water equation.

(2) Clarified the response of surface runoff simulation accuracy to high-resolution DEM, high-resolution DEM can significantly improve the accuracy of simulation results. By simulating the rainfall-runoff process in the 2m and 5m resolution DEM in watershed, it is shown that the capture accuracy of the peak flow and time for the higher-resolution DEM is increased by 14.71% and 20%, respectively. For the simulation accuracy of the runoff process, NSE and R² increased by 26.03% and 14.67% respectively; RMSE and RB decreased by 22.58% and 84.78%, respectively.

(3) Revealed the response of sediment transport on slope to hydrodynamic parameters under different slope and flow conditions, slope sediment transport capacity formula was established based on unit energy power. Compared to the relationship between runoff shear stress, unit stream power and runoff energy consumption theory and slope erosion, it is shown that the three theories have themselves advantages in slope erosion research. The parameters are easy to obtain for unit stream power calculation, and unit stream power theory is widely used to the erosion model based on physical processes. Based on the slope sediment transport capacity formula of Govers, slope sediment transport capacity was corrected by introducing the viscous sand starting flow rate calculation method, the parameters based on the particle size of the sediment particles were determined through regression analysis of the experimental data. A corrected slope sediment transport capacity formula of Govers was established, which is suitable for the loess slope. Compared to original Govers formula, the established slope sediment transport capacity formula can is better to predict the sediment transport capacity. Regarding the evaluation metrics, P.O.0.5-2.0 and R2 are increased by 28.94% and 5.38%, and RMSE is reduced by 80.00%.

(4) The high-performance and robust numerical model of two-dimensional water and sediment on the slope was developed. The temporal and spatial evolution of slope runoff and sediment was simulated, and the source and change of erosion was identified on the slope. The hydrological-hydrodynamic numerical model coupled the slope sediment transport module established by corrected Govers sediment transport capacity formula, which shaped a high-performance robust numerical model for the two-dimensional water and sediment for the loess slope. The model simulation was verified by simulating the rainfall erosion experiment for slope and slope-gully. Regarding the evaluation metrics for slope simulation accuracy: the NSE were 0.83 for runoff rate and 0.66 for sediment concentration, R² were 0.89 for runoff rate and 0.73 for sediment concentration, respectively, and RB was -5.02% for runoff rate and -1.02% for sediment concentration. Regarding the evaluation metrics for slope-gully simulation accuracy: the NSE were 0.81 for runoff rate and 0.47 for sediment concentration, R² were 0.74 for runoff rate and 0.57 for sediment concentration, respectively, and RB was -5.88% for runoff rate and -1.53% for sediment concentration. The analysis of the slope erosion source in spatial shown that the most severely eroded area is the middle and lower part of the slope, and the erosion contribution rate of the slope 1~4m is 69.59%. It is necessary to carry out some measures for controlling the soil erosion in main erosion source area.