A one-dimensional coupled model has been proposed to consider the flow exchanges between main stream and tributaries during the alternative calculations between open-channel flow and turbidity currents, which is capable of simulating the whole process of flow and sediment transport covering these three different physical factors during the period of water-sediment regulation in the Xiaolangdi Reservoir. The modified storage-cell method was adopted for a complex tributary system to calculate the net flow exchanges caused by the water surface variation, and the discharge formula considering the effect of bottom slope of tributaries was applied to calculate the turbidity currents intrusion. Meanwhile, the plunge criterion of turbidity is used to connect the calculations of open-channel flow and turbidity currents, the static water surface assumption was no longer used so as to couple the interaction between the upper clear water and lower turbidity currents. The whole event of water-sediment regulation conducted in the Xiaolangdi Reservoir in 2002 was used for parameter calibration and the event of 2012 was simulated using both the proposed coupled model (Model I) and the model without considering the flow exchanges (Model II). The results show that the reservoir can be divided into three regions: the upper region with no tributary, the middle region with just open-channel flow and the lower region with turbidity currents, where both contain tributaries. Model I presented the results to closely agree with the measured processes in three regions. Among which the upper region is the only region the results calculated by Model II are as precise as Model I, However, with effects of the flow exchanges on the sediment transport and the conservation of water mass and flow momentum being omitted, Model II presented incorrect predictions in the lower two regions where there are tributaries: the calculated water level, discharge and sediment concentration in the middle region are rather high during the occasion of water intrusion and rather small during the occasion of water recharge; in the lower region, the location and occurrence of the plunge point were predicted to be delayed, and these turbidity factors were overestimated, covering interface level, discharge, flow velocity and sediment concentration, correspondingly, the sediment delivery ratio are rather large. Therefore, the effect caused by the tributary system is so enormous that it should not be neglected in the simulation. In addition, a detailed sensitivity analysis of coefficients in the turbidity module was also conducted in order to investigate the effects of these coefficients on the model predictions, with the proper variation range being determined for each coefficient. It indicates that the effects on the calculated factors caused by the drag coefficient at interface, the resistance coefficient of intrusion and the mixing coefficient between clear water and turbidity decreased in turn. Besides, with an increase in drag coefficient, the model response shows a time lag at a certain extent; with the bottom slope of tributaries not being constant, the effect of resistance coefficient requires a further discussion based on a specific slope.
Read full abstract