Numerical modelling of sediment erosion from reservoirs

Sedimentation in reservoirs causes loss in the useful water storage volume, threatening the long term sustainability of dams. The number of rivers being dammed worldwide is increasing rapidly, but the favorable locations for dam sites are decreasing with it. So it is necessary to establish some efficient measures to manage the problem of sedimentation in existing reservoirs. Recently, sediment flushing operations have been more common to release part of the stored sediment in reservoirs. However, there is no clear methodology for the study and optimization of sediment removal from reservoirs. Sediment removal from reservoirs is a complex process involving the interaction between hydrodynamic and morphodynamic phenomena, which makes it difficult to study it analytically. Recent advancements in numerical models have provided some opportunity to study such complex processes. One-dimensional numerical models have been employed in some reservoir sediment removal process, but their performance is just limited to longitudinal processes and cannot give information at the cross-sectional scale. So two-dimensional models are necessary to get sufficient information in both the reach scale and cross-sectional scale. But the performance of two-dimensional models in predicting the sediment removal from reservoirs has not yet been verified. Recently some dam removal events are being extensively monitored to document the river morphological changes due to the exposure of large amounts of sediment from the reservoir. Dam removal can be considered analogous to large scale sediment flushing operations, as both activities involve release of large amounts of sediment from the reservoir towards downstream. Similar is the case of Marmot Dam Removal. The Marmot Dam located across the Sandy River in the Oregon State of USA, was removed in 2007, due to economic and environmental considerations. This exposed a large amount of sand and gravel deposited in the reservoir to be eroded by the natural flow of the Sandy River. The removal of the Marmot Dam was monitored extensively for the erosion, transport and downstream deposition of the impounded sediment. This provided an opportunity to conduct a research on the morphological changes occurring in reservoirs during the sediment removal. This study considers the case of Marmot Dam Reservoir Erosion to evaluate the performance of a 2D numerical model built in the open source software Delft3D. The modelling work includes simulation of the cofferdam breaching process and the erosion process of graded non-cohesive sediment from the reservoir after the dam removal. The model results are analyzed to validate the model considering any shortcomings. The validated model is then used to study and optimize the sediment flushing operations in the same reservoir. The results indicate that Delft3D is indeed a good tool in simulating sediment erosion process from reservoirs. There are some limitations in the quantification of the removed sediment amounts, but the overall morphological development is quite well represented by the model. This evaluates Delft3D as a useful tool to simulate different sediment flushing operations, in view of optimization and selection of the best scenario for flushing