Effect of Grid Spacing on the Prediction of Bank Erosion in a Large Braided Alluvial River

M. Mamum and M. Islam (Bangladesh)


Jamuna river, 2D morphological modelling, curvilinear computational grid, bank erosion, schematization, stretching.


TABLE I SUMMARY CHARACTERISTICS OF THE JAMUNA RIVER Description Parameter Length in Bangladesh 240 km Average bankful width 11 km Discharge 2,400 ~ 102,000 m/s Average discharge 20,400 m/s Bankful discharge 48,000 m/s Dominant discharge 38,000 m/s Braiding index 4 6 Water surface slope 8.5 6.0 cm/km Grain diameter (d50) 0.2 0.162 mm Average yearly sediment transport 600 Million tons Average yearly coarse suspended sediment transport 200 Million tons Maximum erosion rate 100 to 400 m/year Minimum erosion rate 30-40 m/year A method for developing quantitative prediction of riverbank erosion is presented where a detailed comparison has been made among coarse grid, intermediate grid and fine grid predictions for the bank erosion in a large alluvial river in the Jamuna in Bangladesh. In the intermediate grid modelling, the area of interest has further been stretched where the grid sizes are further reduced for higher resolution having more smaller mesh sizes than that of even the fine grid schematization. Overall reduction of the mesh size in fine grid model in the entire computation domain by more than one half than the coarse grid model has limitations in computational time and runtime. The intermediate grid modelling with sufficient stretching function applied at particular locations become necessary for representation of engineering measures in the model grid that produces better results thus saving the computational time and memory. In the present study, a two-dimensional (2D) mathematical model has been developed as a morphological prediction tool and sensitivity analysis has been made for model prediction by different computational grid sizes and resulting bank erosion at the area of interest. The comparison of prediction on the bank erosion has been done for the condition prevailing in the erosion affected river stretch. It has been observed from the model results that intermediate grid prediction provides much better representation of bank erosion along the banks where grid resolution is highest among all other conditions at the erosion prone area. The comparison has provide a basis for validation of the model and see the changes in the detailed flow pattern in the erosion prone area and short term bank erosion due to differences in computational grids.

Important Links:

Go Back