Sensitivity of Sediment Transport Analyses In Dam Removal Applications

Dam removal has become a widespread river management practice in the United States for multiple goals including ecosystem restoration, removal of aging infrastructure, flood risk management, and recreation. The ability to forecast the sediment impacts of dam removal is critical to evaluating different management alternatives that can minimize adverse consequences for ecosystems and human communities. Tullos et al. (2016) identified seven Common Management Concerns (CMCs) associated with dam removal. Four of these CMCs; degree and rate of reservoir sediment erosion, excessive channel incision upstream of reservoirs, downstream sediment aggradation, and elevated downstream turbidity are associated with stored sediment release and changing fluvial hydraulics. There are a range of existing qualitative and quantitative tools developed to infer or quantify geomorphic implications of disturbances like these in river environments (McKay et al., 2019). The purpose of this study was to investigate how a 1D sediment transport model can inform these four CMCs and develop an approach to assess sediment transport model sensitivity in the context of the Simkins Dam removal case study. Model outputs including mean effective invert change (MEIC), eroded sediment volume from reservoir, and sediment concentrations could inform CMCs associated with stored sediment release. After developing a sensitivity analysis approach, sensitivity analyses were then performed to identify key uncertainties, which can inform data collection and model building for other dam removal projects. Sensitivity analyses are commonly used to examine how model outputs deviate from model calibration results because of the variation of input factors (Pianosi et al, 2016). Model inputs used in this study for the sensitivity analysis included reservoir sediment gradation, transport functions, sorting methods, and erosion methods for reservoir sediment deposits. Calibrated model results were compared with results from the sensitivity analysis. Model outputs for the selected case study, including the MEIC and eroded sediment volume from reservoir, were highly sensitive to the variation of the reservoir sediment gradation and sorting method selection. These model outputs also showed some sensitivity to the selected transport functions. Erosion method sensitivity using the channel evolution method will vary depending on side slope and channel parameter selection. This study concluded that 1D sediment transport models can be a useful quantitative tool for assessing post dam removal sediment related CMCs. Collection of sediment field data along the study site and at the reservoir is recommended to develop gradations that can better estimate degradation/aggradation patterns and volume of eroded sediments from the reservoir. Also, the appropriate selection of model algorithms particularly sorting methods is critical to estimate model outputs informing CMCs related to sediment.