Hydrological and hydraulic modelling of Brahmaputra Basin using global datasets

Brahmaputra river basin is one of the largest catchment in the world. It is shared by the countries China, Bhutan, India, and Bangladesh. It's a complex catchment in the sense that some of the areas receive rainfall in the form of snow, some areas receive too much rain and others receive little rain. Adding to the complexity is the range of elevations. The highest elevation at the Tibetian Plateau is almost 5800 meters while the lowest point at Bangladesh is mean sea level. This catchment has a big importance to the riparian countries since it supports agriculture, hydropower electricity, and other industrial uses. This study is aimed to generate a good hydrological/hydraulic model. Due to limited observed hydrological data we were using global datasets. One of them is the rainfall data, in this study Tropical Rainfall Measuring Mission (TRMM) version 7 was used as a rainfall estimate. The software used in hydrological modelling was HEC HMS of United States Army Corps of Engineer (USACE). For the approximation of the parameters in HEC HMS global datasets were used. These global datasets such as soil types, rock types, and landcover/landuse were downloaded from open source. The hydrological model was simulated for 15 yrs., calibrated for 8 yrs., and validated for 5 yrs. The model had an acceptable performance with a NashSutcliffe Coefficient of 0.74 - 0.81. The simulated and observed peaks were almost the same but in low flows the simulated discharge values are lower than the observed values. Using the result of hydrological modelling we generated a hydraulic model for the Brahmaputra River. The software used was HEC RAS of USACE. The length considered in the model was about 500 km, the portion above the Bahadurabad Station. The model was simulated for the whole year of 2015. The result was compared with the water level at Bahadurabad Station, and it showed good performance except during the low flow months. It was also compared with the satellite altimetry and produced a good correlation. Simulations were also done for 2 different flood events in 2012. In all of these flood events flood maps were generated. For comparison, satellite observed flood maps were downloaded from Dartmouth Flood Observatory of University of Colorado. The two flood maps were compared using Critical Success Index (CSI) method. The result showed positive output with a high Probability of Detection (POD) although False Alarm Ratio was quite high due to the shallow portion of water present in the simulated flood and not in the observed flood. Flood frequency analysis was conducted for observed flood discharges to determine the frequency of each flood event occurring in the river. Generally, the hydrological model and the hydraulic model produced reasonably good results and can be used by the management of the river basin. However, there are scopes for improving these models. More gauge rainfall will be useful. Also more detailed geologic map of the area will be beneficial since in this study surface rocks types was used to determine parameters for groundwater. Actual survey of the cross sections of the river or LidAR based DEM may help to improve the performance of the model since in this study, a satellite based DEM was used to determine bathymetry of the river