Estimation of uncertainty in the hydro-morphodynamic characteristics of the Meghna Estuary, Bangladesh
The coastal zone of Bangladesh is a part of the Ganga–Brahmaputra–Meghna basin, and is home to about 38.5 million people. This coastal community is particularly vulnerable to natural disasters, such as catastrophic floods from extreme river flows, cyclones, land erosion, and sea level rise. Although advances in computational hydraulics facilitate the numerical simulation of extreme events in the coastal zone, informing risk assessment, the numerical models themselves propagate uncertainty from input to output parameters. This thesis presents a numerical derived distribution approach for uncertainty propagation through a computational model of tidal and fluvial processes in Meghna estuary of Bangladesh. The approach involves discretization of an estimated probability distribution function of a key input variable, the computation of a response function linking a single input parameter to a single output variable of interest, and the use of conservation of probability to determine the probability distribution of the output variable. The method requires only a few simulations to be conducted, and so it is very efficient to implement. In the thesis, Delft3D, a well-established computational tool, is verified for a series of standard tests. Then, Delft3D model is set up for the Meghna estuary and Bay of Bengal. Uncertainty propagation is then examined by studying the effect of uncertain bed roughness on the estimate of maximum water level, the effect of uncertain sea level rise on the maximum water level, and the effect of uncertain floc size on the sediment deposition rate at selected sites around the Meghna estuary of Bangladesh. It is found that with a 50% increase of mean Manning’s n, the maximum water level can increase from 24% to 80% at various locations. For several IPCC scenarios of sea level rise, the standard deviation of maximum water level increases about 33-40% at the same locations. For a mean floc size of 227 micron and standard deviation of 171 micron, the coefficient of variation of cohesive sediment deposition rate is estimated to range from 19.8% to 37.6% at three locations in the Meghna estuary.