Evaluation of Hydrodynamic Mixing in an Afterbay Reservoir

Abstract

This study focused on the mixing of a solute, assumed to be conservative, introduced to one arm of an afterbay reservoir, between Keswick and Shasta Dams on the Sacramento River near Redding, California. Rhodamine water tracer (WT) dye served as the solute in a field experiment, and was introduced over 4.5 h and monitored for 4 days by sondes moored in the reservoir. The scenario was modeled numerically using the Delft3D flexible mesh (FM) hydrodynamic and mixing model, with measured inflows, outflows, water level, water temperatures, and bathymetry as input. Manning’s n and horizontal eddy viscosity served as the (constant) model calibration parameters, and each was adjusted an order of magnitude below the default values to force observed and modeled dye hydrographs to match in arrival time and duration. The low friction factor was concluded to be due to a combination of low flow speeds coupled with energy dissipation inherent to the model. The model and surface drifters equipped with dual-frequency Global Navigation Satellite System equipment revealed velocities in the 1–5 cm/s range in much of the domain during the experiment. Simple analytical expressions were shown to be useful for estimating distance to full cross-sectional mixing, steady-state concentrations, and time to reach them, but the numerical model is required for investigation of the approach to steady state, and at locations where flows intersect. Time to steady-state concentrations was 1.5–13 days for 10 simulations that spanned a wide range of inflow conditions. Model sensitivity tests suggest that wind and heat fluxes were not important during the field study, but simulations of a summer scenario with small inflows of cold water upstream in warm weather should consider water temperature. Both field observations and numerical model results showed inflow to one arm of the reservoir reaching full cross-sectional mixing before plunging below the water surface near the intersection of this arm with the reservoir’s main stem. Model results are being used to guide management decisions related to inflows to the reservoir from a relic mining site that is also a USEPA Superfund site.