Evaluation of Water Year 2011 Glen Canyon Dam FLow Release Scenarios on Downstream Sand Storage along the Colorado River in Arizona

Home Browse Search

because tributary inputs cannot be forecasted accurately. Also, Wright and others (in press) showed that the sand routing model is not particularly sensitive to the initial bed conditions for the sensitivity range studied therein. Uncertainties in the model results were evaluated by conducting simulations with estimated errors incorporated into the boundary conditions (tributary inputs) and model calculations (sand transport rates). Following the methods used by Topping and others (2010) for constructing error bars for sand budgets based on high - resolution monitoring data, the tributary inputs were varied by t10 percent and the sand transport rates were varied by t5 percent. That is, for each scenario, two additional simulations were performed: 1) tributary inputs increased by 10 percent and sand transport rates decreased by 5 percent, providing an upper uncertainty bound; and 2) tributary inputs decreased by 10 percent and sand transport rates increased by 5 percent, providing a lower uncertainty bound. This technique is particularly appropriate here because the sand routing model was calibrated to sand transport measurements with comparable error estimates. It is noted that these uncertainty bounds are most useful for evaluating whether there is net accumulation or erosion for a given scenario. For comparing scenarios to each other, it is important to only compare simulations with the same boundary conditions and model parameters. For example, it would not be appropriate to compare one operating scenario with tributary inputs increased by 10% with a different operating scenario with tributary inputs decreased by 10 %. Results The sand routing model predicts sand concentrations at the three computational nodes shown in figure 1 (RM 30, RM 61, RM 87). The sand concentrations were combined with the flows at these locations to compute sand fluxes, and the sand fluxes were then used to construct sand budgets for the three reaches bounded by the computational nodes (flux is assumed to be zero past Lees Ferry). For this report, the two Marble Canyon reaches were combined for sand budgeting purposes. The sand budgets are simply a cumulative accounting of sand inputs to a reach minus sand export from a reach; thus, a positive sand budget indicates net sand accumulation within the reach and a negative sand budget indicates net sand erosion within the reach. The primary results of the simulations are shown in the series of figures 3 -8. Each figure represents a different operational scenario and consists of three panels. The top panel shows the hourly release hydrographs for the given operational scenario for both the 8.23 and the 11.0 MAF annual volumes. The middle panel shows the modeled cumulative sand budgets for Marble Canyon, again for 8.23 and 11.0 MAF, including the uncertainty envelopes. The bottom panel shows the simulated sand budgets for the eastern Grand Canyon reach. The model results show some expected trends that are common to most of the operational scenarios. First, the higher annual release volume (11.0 MAF) consistently leads to more sand export and thus less sand in the reaches projected at the end of WY 2011. The sand budgets are also seen to reflect variations in flow releases (patterns and volumes) and tributary inputs throughout the year. For example, the Marble Canyon sand budgets tend toward accumulation during August and September when Paria sand inputs are greatest (fig. 2Q. An example of the impact of flow volume on the sand budgets is apparent in the results for SAS (fig. 7), where it is seen that extended periods of relatively high flows in the spring drive the sand budgets substantially in the negative direction (that is, erosion). The dashed lines define the modeling uncertainty envelopes and allow for determination of the sign of the sand budget for each scenario at the end of WY 2011 (that is, for determination of a positive or negative sand budget, the uncertainty envelope must not span zero). In order to compare the scenarios more directly, annual sand budgets, including the uncertainty envelopes, were computed for each simulation. The annual sand budgets are equivalent to the cumulative sand budgets shown in figures 3 -8 at the end of WY 2011. These results are shown in figures 9 and 10; in these figures, the horizontal lines represent the "base" simulations (that is, without uncertainty estimates) and the vertical lines denote the range based on the uncertainty simulations. 13