Predicting possible effects of dams on downstream river bed changes of a Himalayan river with morphodynamic modelling

The effect of building large dams on downstream channel morphology was reported by several studies on the basis of historical gauged data on discharge, sediment load and channel cross-sections records for a number of large rivers in Asia. However, there has been a notable absence of scientific literature dealing with the prediction of such changes, particularly in the context of South Asia, through model building. Establishing a predictive framework will be an essential prerequisite for including the projected climate change scenarios within the ambit of this type of investigation. Thus, this study aims to investigate the role of dam construction on the downstream channel morphology of the River Teesta, an important tributary of the Brahmaputra River, with a morphodynamic model that considers the process of erosion, deposition and sediment transport. A differential GPS-aided survey, conducted in the pre-monsoon months of 2015, provided an accurate initial model geometry for Delft3D morphodynamic model. Actual discharge (Q) and suspended sediment concentration (SSC) records of monsoon 2015 were used as model inputs. Unlike the majorities of the studies that compared results of morphodynamic models with flume-based laboratory experiments, this investigation used field and satellite-based observations, collected during the monsoon of 2015, for the distributed model calibration and validation. The predicted channel bed at the end of the monsoon of 2015 was compared with spot heights surveyed during the same time with a low RMSE of 0.26 m. In order to predict the effect of dam construction in a medium-term, Delft3D was run with gauged Q and SSC data of four consecutive monsoons belonging to the pre- and post-dam period using the identical channel geometry of the pre-monsoon 2015 channel. A comparison of the final channel bed topography of the pre- and post-dam scenarios revealed a lower relative relief and decreasing channel cross-sectional areas for the later, probably caused by attenuated flood peaks.