After the operation of the Three Gorges Project, the Middle Yangtze River has continuously suffered from channel degradation, along with significant bank erosion processes in local subreaches, especially in the Jingjiang Reach, which may cause serious problems to channel stability, flood control management and navigation safety. In this study, a one-dimensional coupled morphodynamic model for bed evolution and bank erosion processes was proposed, including three modules for hydro- and sediment dynamics, groundwater flow and bank erosion, with the mode of plannar failure in the Upper Jingjiang Reach (UJR) and the mode of cantilever failure in the Lower Jingjiang reach (LJR) being considered. The proposed model was applied respectively to simulate the fluvial processes in the Jingjiang Reach in 2005, 2007 and 2010, in order to conduct a detailed process of model calibration and validation. These calculated results show that: (1) The proposed model accurately reproduced the flow and sediment transport processes, with the root mean squared errors of daily averaged discharge and suspended sediment concentration far less than the corresponding mean values. Besides, the model also predicted the major occurrence regions of serious bank erosion processes; (2) bank failure was mainly controlled by fluvial erosion in the Jingjiang Reach, with more than 50% of cantilever failure events in the LJR occurring during the flood period; and (3) the groundwater level variation was delayed, as compared with the river stage change, which could also obviously influence bank erosion processes in the UJR. Furthermore, multiple test runs were conducted to investigate the influences of the variations in outlet river stage and bank soil properties on the processes of bank erosion. These results from test runs indicate that: (1) If the outlet river stage was increased due to a higher lateral inflow from the Dongting Lake, the bank retreat width in the Jingjiang Reach would be decreased, resulting from the reduced longitudinal water surface slope and mean velocity in section; (2) the calculated bank retreat width at a section was most sensitive to an reduction in critical shear stress of bank soil, especially in the LJR. However, the calculated bank retreat width in the UJR was also sentitive to the change in bank soil cohesion, but it did not monototically decrease with an increase in soil cohesion. This is because the varitation in soil cohesion can lead to changes in the number, timing and individual failed block volumes of bank failure events. As the failure number keeps unchanged, an increase in soil cohesion would lead to a delayed occurrence of bank failure event and thus a larger failed block volume, which finally results in a larger total retreat width.
Read full abstract