• Final Report (pdf, 5 mb)

 Executive Summary

The North Saskatchewan River (NSR) in Alberta, Canada, begins at the Saskatchewan Glacier in Banff National Park and continues to  the Alberta–Saskatchewan border. The NSR subsequently joins the South Saskatchewan River and eventually flows into Lake Winnipeg. This report documents the configuration, calibration, and validation of an in-stream hydrodynamic and water quality model for a portion of the NSR from 30 kilometers below Abraham Lake to 38 kilometers downstream of the Alberta-Saskatchewan border. Flow, water temperature, dissolved oxygen, organic carbon, nutrients, and algae interactions were modeled under the influence of tributaries, municipal wastewater treatment plants (WWTPs), industrial facilities, combined sewer overflows (CSOs), and stormwater.

The Environmental Fluid Dynamics Code (EFDC) was selected to model both hydrodynamics and water quality for the NSR in this study. EFDC is a public domain general purpose modeling package for simulating one-dimensional (1-D), two-dimensional (2-D), and three-dimensional (3-D) flow, transport, and biogeochemical processes in surface water systems including rivers, lakes, estuaries, reservoirs, wetlands, and coastal regions. Enhancements were made to the EFDC model to simulate ice cover in the river.

Configuration of the NSR EFDC model involved setting up the model computational grid using available geometric data, designating the model’s state variables, setting boundary conditions, and setting initial conditions. The 1-D NSR model was represented by 778 segments of approximately 1,000 meter (m) lengths. Widths of the segments vary from approximately 50 m to 450 m. External forcing factors, or boundary conditions, specified for the model include upstream boundary conditions (i.e., upstream inflows, temperature, and constituent boundary conditions); tributary inflows (i.e., tributary inflows, temperature, and constituent boundary conditions); loadings from point sources including industrial sources, WWTPs, CSOs, and stormwater; surface boundary conditions (i.e., atmospheric conditions); and downstream boundary conditions (i.e., outflow). Initial conditions were set to ensure stable running of the model, especially the hydrodynamics. This was done by setting the initial water surface elevation, which was assumed to be parallel to the bottom elevation. Once the NSR model was configured, calibration was performed. Calibration refers to the adjustment or fine-tuning of modeling parameters to produce an adequate fit of the simulated output to the field observations. The NSR model simulates all conditions from September 2000 to December 2007.

Hydrodynamics and heat transport were calibrated first, and then water quality was calibrated using available monitoring data. Available in-stream data used for calibration included water surface elevation, continuous water temperature, continuous dissolved oxygen, and grab sample results of water quality constituents. The water surface elevation and continuous water temperature data were used to calibrate the hydrodynamics and heat transport simulation. The continuous dissolved oxygen and other discrete water quality data were used for water quality calibration. Water surface elevations were calibrated to ensure flow balance at two flow stations in the modeling domain (Edmonton and Deer Creek stations). The modeled elevations agree well with the observed elevations during non-frozen seasons. Water temperature was evaluated to ensure correct heat transport in the NSR. Modeled water temperatures were compared to observed temperature data at eight datasonde stations. Overall, the modeled water temperature results agree well with observed data. Water quality calibration involved examining the major reaction parameters and adjusting the parameters until model results agreed with the data. The major parameters adjusted include the ammonia nitrification rate; organic carbon dissolution rates; organic phosphorus hydrolysis rates; and algae growth, death, and respiration rates. Calibration focused primarily on a comparison of modeled and observed dissolved oxygen. Dissolved oxygen is a key water quality indicator and is affected by various processes such as water temperature, reaeration, organic carbon decay, nitrification, and algae growth and death. The model was able to generally capture the trends and magnitudes seen in the dissolved oxygen observations. During ice cover, modeled dissolved oxygen results show a stable level because of the very low algae metabolism rates in low water temperature, no reaeration, and low bacteria activities to decay organic carbon or nitrifying ammonia. In warm weather, phytoplankton or benthic algae grow quickly and dissolved oxygen in the water column varies significantly. The model accurately reproduced such growing patterns for most NSR locations. The other modeled water quality constituents also agree well with observed data. Seasonal variations of nutrients are captured, and the modeled water quality constituents are in reasonable ranges. The model also captures the levels of the dissolved and total organic carbon.

The NSR EFDC model provides a sound basis for conducting scenario simulations. Boundary conditions can be readily changed to evaluate effects on conditions throughout the system. The model can be further improved through refinement of tributary boundary conditions, perhaps through watershed model simulation.