Minimizing Downstream Temperatures Through Modeling of Dam Operations

By Marcela Politano, Ryan Laughery, and Larry Weber

In this study, an unsteady three-dimensional, non-hydrostatic model was used to predict the flow field and temperature dynamics in the forebay and turbine intakes of the U.S. Army Corps of Engineers’ 1,127-MW McNary plant, in order to help curb temperatures in the Columbia River.

A numerical model was developed to simulate the temperature dynamics in the McNary forebay. Most thermal models in lakes are based on a one- or two-dimensional approach. However, hydropower forebays are characterized by complex 3D flow patterns as a result of dam operations and unsteady heat exchange between the atmosphere and water, requiring the use of a 3D model. Those that assume hydrostatic pressure have been widely used over the past decade to simulate temperature dynamics. But, a more complex model was required because both vertical motion and vertical accelerations are important near turbine intakes.


The McNary project includes McNary Dam, the Lake Wallula forebay, 14 turbine-generator units, 22 spillway bays, a navigation lock and two fish ladders. Each turbine unit includes three intake bays, main and intermediate piers, vertical barrier screens (VBS), extended submerged bar screens (ESBS) and turning vanes. Two temporary spillway weirs (TSW) were installed in spillway bays 20 and 22 to provide a less stressful downstream passage route for juvenile salmon.

Juvenile fish mortalities associated with elevated water temperature within the fish facilities at McNary Dam were documented since the 1980s. A significant fish loss, believed to be primarily the result of high water temperature in the juvenile fish facility, occurred in July 1994. Since then, warm water problems were successfully minimized by continuous monitoring and adjusting dam operations.

In 2015, air temperatures and solar radiation levels exceeded the 32-year average, according to Bureau of Reclamation data, while low snowpack and precipitation caused the third-lowest runoff volume in the Columbia and Snake rivers in 56 years. These factors led to water temperatures in the McNary forebay at 0.5 meters below the surface peaking at 26.9 degrees Celsius on July 9.

Snake River sockeye salmon – listed as protected in the Endangered Species Act – migrate within this time period and suffered losses exceeding 95% between Lower Granite and Bonneville dams. Juveniles suffered delayed migration, but travel times were faster than those prior to TSW installation.

Temperature dynamics were modeled using computational fluid dynamics.
Temperature dynamics were modeled using computational fluid dynamics.

Model overview and validation

The numerical model includes 2 miles of Lake Wallula and the main features of the McNary project, including turbine units, spillway, TSW, Oregon and Washington fishways, and navigation lock.

The model used is based on the commercial code ANSYS Fluent. The flow field is solved with the incompressible unsteady RANS equations using the Boussinesq approach to account for buoyancy forces. The turbulence is modeled with a standard k − model with wall functions. The fish screens were modeled as porous media of finite thickness with screen porosities calibrated using experimental data from a physical model.

The temperature was computed from the energy conservation equation. The difference between incoming and reflected measured radiation were used to adjust a quadratic sinusoidal function for the incident radiation.

The radiation absorbed by the water was modeled accounting for the attenuation of solar radiation with depth given by Beer’s Law. The average longwave radiation measured at night was included as a negative source term for elements contiguous to the free surface.

The model was validated against temperature measurements obtained by the Corps at 46 stations along six transects in the McNary forebay, and in the gatewells at 15-minute intervals during a warm day in the summer of 2004. After validation, the model was run with atmospheric and river conditions observed on July 9, 2015. On that day, water temperature at 5m beneath the surface was the season’s highest. The model captured well the measured temperature profiles in the forebay under this extreme condition.

The model shows the curtain is the best option to reduce downstream temperature.
The model shows the curtain is the best option to reduce downstream temperature.

Evaluation of operational and structural modifications

Possible mitigation measures to reduce temperature at the turbine intakes were evaluated with the model. Hypothetical conditions with 60% spill, spill with TSW and the inclusion of a thermal curtain were simulated (see Figure 1 on page 30). According to the model, the curtain is the best option to reduce temperature in the river downstream of the dam while maintaining gatewell temperature in the tolerance zone for migrating salmonids (see Figure 2).


In this article, the temperature dynamics in the McNary Dam forebay and turbine intakes were numerically studied using a CFD model. The model captured the general observed diurnal stratification. Moreover, the model was able to reproduce the period of time when the maximum temperature in the forebay and in the gatewells were reached.

The model was used to quantify the impact of atmospheric, operational and structural changes on the water temperature in the forebay and intake units. According to the numerical results:

  • The model reproduced the measured temperature profiles in the forebay for an extreme atmospheric condition observed July 9, 2015. Gatewell temperatures were below the upper incipient lethal temperature for salmon. Although the air temperature in July 2015 was 6.5 C higher than on Aug. 18, 2004, gatewell temperatures were only about 0.5 C higher. Numerical results seem to indicate that powerhouse operation with the northern units can minimize the transport of the warm surface water into the units, resulting in less potentially damaging temperatures for salmonids migrating downstream;
  • Temperature within the southern gatewells increased about 0.2 C when 60% of the river discharge was spilled and the northern units were closed;
  • Operating with TSWs, the temperature in the gatewells of Units 11 and 14 increased by about 0.5 C; and
  • The presence of a forebay curtain upstream of the powerhouse reduced the temperature within the gatewells by about 1 to 2 C, preventing fish from entering the “resistance zone” in which their behavioral mechanisms allow them to survive short-term in extreme temperatures.

Based on the earlier efforts the Corps try to run the units at lower discharges and space active units apart in an attempt to reduce their ability to be efficient at drawing down the warmer surface water. By doing this, withdrawals seem to be more lateral and less vertical in nature. The operators at the project call this a saw tooth pattern. Looking at the operation on July 9, this seems to be the case.

Editor’s Note: This is a digest of a paper delivered at HydroVision International 2016. The complete paper can be found in the Hydro Library at

Marcela Politano is a research engineer for IIHR at the University of Iowa. Ryan Laughery is a hydraulic engineer for the U.S. Army Corps of Engineers’ Walla Walla District. Larry Weber is the director of IIHR.

Previous articleROV Inspection of the Intake Gate Shaft at Lower Baker
Next articleTech Briefs

No posts to display