Bureau of Reclamation Banner

Technical Service Center
Land Suitability and Water Quality Studies

Summary of Two-Dimensional Water Quality Modeling of Cascade Reservoir

by:

Merlynn D. Bender
Technical Service Center
Environmental Applications and Research Group (86-68220)
Denver, Colorado
Telephone No. 303-445-2460

and:

David Zimmer
Pacific Northwest Regional Office
Ecosystems Analysis Group (PN-6520)
Boise, Idaho
Telephone No. 208-378-5088

Prepared for:
Pacific Northwest Regional Office
Project Manager, Dave Zimmer

Executive Summary

Seasonal variations in water quality of Cascade Reservoir were modeled with a two-dimensional model. Modeling indicated that sediment oxygen demand from decaying organics, aerobic release of nutrients, extensive algal biomass, and insufficient flushing were the primary factors contributing to eutrophication and, consequently, low dissolved oxygen.

The water quality and aquatic biological resources of Cascade Reservoir and its tributaries are being adversely impacted by point and nonpoint pollution sources, reservoir operations, and natural processes. Data indicate that Cascade Reservoir has surplus nutrients, extensive algal biomass, a decaying benthic layer, and is, consequently, eutrophic. To better understand the effects of external and internal nutrient and organic loadings and effects of reservoir operations on Cascade Reservoir, a two-dimensional Box Exchange, Transport, Temperature, and Ecology of a Reservoir (BETTER) model was assembled, calibrated, and used for sensitivity and management simulations.

Using a two-dimensional array of longitudinal and vertical elements, the BETTER model calculates flow exchange, heat budget, and dissolved oxygen (DO). The model simulated Cascade Reservoir's warm surface layer and cold bottom water. Currently, cold bottom water is trapped behind sediment deposition in the inlet channel to the turbine intake. The heat budget includes drybulb (air) and dewpoint temperature, solar radiation, wind mixing, convective cooling, and inflow density distribution. The DO components include sediment oxygen demand (SOD); biochemical oxygen demand (BOD); ammonia; surface reaeration; and algal photosynthesis, respiration, and nutrient recycling.

A seasonal DO pattern was simulated from ice-out to ice-up using a 12-hour (day and night) input-data time step. The North Fork Payette River, four tributaries, and local inflow to the surface layer were modeled.

The Cascade Reservoir model was calibrated with sparse data for years 1989, 1993, 1994, and 1995: 1989 was slightly below average inflow and followed a dry year; 1993 was slightly above average inflow and followed a dry year; 1994 was severely dry; and 1995 was wet.

Model simulations indicated that SOD; anaerobic release of nutrients, which stimulates algal growth; and minimal flushing were the primary factors contributing to low DO in the reservoir.

A simulated 50-percent reduction in nutrient and organic loading had minimal immediate effect on Cascade Reservoir. However, significant loading reductions over several years may improve Cascade Reservoir water quality.

Simulated sealing of the sediments, cutting off anaerobic release of nutrients, and cutting sediment oxygen demand in half significantly improved reservoir water quality. However, the mechanism to seal the sediments was not identified or simulated directly. Simulated aeration of releases improved release water quality. However, aeration of releases would not improve water quality upstream of Cascade Dam forebay.

Simulations indicated several alternatives were technically infeasible. Simulated selective withdrawal of bottom waters via removal of sediment deposits just upstream of the turbine trashrack degraded reservoir water quality. Simulated selective withdrawal of surface waters via increased spillway discharge of surface waters had minimal effect on reservoir water quality.

Sediment deposits in front of the turbine trashrack forces both turbine and spillway discharges to be withdrawn from the mixed surface layer. If the deposits were removed, modeling indicated that cold bottom waters would quickly be flushed out of the reservoir through the turbines and replaced with warmer water. The warmer water would increase decay of organics and anaerobic conditions. Water quality would be worse due to increased bottom temperature, even though more bottom waters would be flushed from the reservoir.

Simulations that decreased the minimum pool resulted in more withdrawal from the surface, more flushing and turbulent mixing, and less volume with high DO and cold temperatures for trout refuge. These changes would result in release of supersaturated surface waters during spring, release of poor quality water during late spring and early summer due to flushing of the reservoir bottom water layers, and release of surface waters saturated with DO from algal photosynthesis during late summer and early autumn due to simulated earlier pool turnover. Winter water quality was not simulated. However, lower winter minimum pools would result in less mass of DO and, consequently, greater potential for winter kill of fish.

Increasing the minimum pool is expected to increase the volume of water with high DO and cold temperatures for trout refuge, increase the total volume with poor water quality, and decrease winter kill of fish. Higher pools would lessen the ability to optimize releases for downstream uses, such as irrigation, downstream water quality, and salmon migration to the Pacific Ocean. Continuous release to improve flushing would drop pool water surface elevation.

Acknowledgments

This project investigation was co-sponsored and directed by the Pacific Northwest (PN) Regional Office of the U.S. Bureau of Reclamation (Reclamation) and Idaho Department of Environmental Quality (IDEQ). Cooperation, effort, information, reviews, and guidance provided by managerial, laboratory, and office staff are greatly appreciated.

Special thanks to Dewey Worth (IDEQ), Jenny Smout (Reclamation, PN Region), Ron Ferrari (Reclamation, Denver Office), Norbert Cannon (Reclamation, PN Region), Gail Ewart (IDEQ), Ted Day (Reclamation, PN Region), and Kelly Redman (Desert Research Institute Climatological Center.

Introduction

The purpose of the Cascade Reservoir Water Quality Modeling Study was to develop predictive models to assist in identifying and evaluating operational and structural measures for improving water quality.

Cascade Reservoir water quality has become of critical interest because of public and agency concerns about nuisance algal blooms, potential for winter fish kills resulting from oxygen depletion under ice cover, and bacterial contamination of water for swimming. Water quality studies conducted by the Bureau of Reclamation (Reclamation) and other Federal, State, and local agencies during the 1970s and early 1980s found that phosphorus and bacterial loadings result in dense algal blooms, unacceptable risk of winter fish kills, and bacterial counts that sporadically exceed recreation standards in late summer. Nutrients and bacteria originate primarily from nonpoint sources, including agricultural, timber, urban, and recreational lands (Zimmer, 1983).

In 1982, the Idaho Department of Fish and Game developed a model to predict the risk of winterkill at different minimum pools. Dissolved oxygen (DO) problems were found to be related to decomposition of decaying organics in summer and a combination of respiration and decomposition of decaying dead algae and other detritus in winter. Based on this study, the Idaho Department of Fish and Game recommended maintenance of a 300,000 acre-foot total minimum pool to keep the risk of dissolved oxygen limitation and fish kill of salmonid fisheries to an acceptable level of risk. The recommended 300,000 acre-foot total minimum pool was administratively established by Reclamation in 1985 (Reininger et al., 1982).

Cascade Reservoir is currently designated as a water quality limited stream segment by the Idaho Division of Environmental Quality under provisions of the Clean Water Act; the State of Idaho has initiated nutrient management planning under the Idaho Nutrient Management Act. The reservoir watershed has been subdivided into a number of sub- watershed planning units. Stake holders in sub-watersheds are developing nutrient reduction plans to improve reservoir water quality (Valley Soil Conservation District et al., 1991).

Opportunities for improving reservoir water quality through changes in Reclamation water and shore lands management practices, nutrient management planning, and managing uncontracted storage space in the reservoir were proposed. Possible actions discussed included dredging nutrient-rich organic bottom sediments, bank stabilization, shoreline fencing, alum treatment, aeration, seasonal changes in releases to reduce internal phosphorus loading, construction of a selective withdrawal system (use of the spillway gates results in downstream migration of resident coho salmon), construction of wetland treatment systems near major tributary inflow points to reduce external phosphorus loading, McCall wastewater treatment plant effluent land application and processing, bathymetric sedimentation survey, monitoring and data collection, modeling studies, recreational storm water measures, varying minimum pools, and seasonal changes in releases for optimizing salmon migration to the Pacific Ocean.

One-dimensional modeling studies done by Dr. Steve Chapra were published by Entranco Engineers in 1991. Additional data and improved reservoir water quality models were needed to further evaluate effectiveness of potential actions (Entranco Engineers, Inc., 199 1).

The Idaho Division of Environmental Quality (1996) published a watershed management plan to improve water quality of Cascade Reservoir. This plan constituted the equivalent of defining a total maximum daily load (TMDL) for Cascade Reservoir. The TMDL process is designed to direct management actions to restore polluted water bodies to a level that achieves state water quality standards.

Two-dimensional water quality modeling studies described in this paper provided additional information for selection and implementation of control measures to improve water quality. The report, "Two-Dimensional Water Quality Modeling of Cascade Reservoir (Bender 1997) fully discusses the additional (BETTER) two-dimensional model description, model inputs, model calibration, and analysis of management alternatives.

Site Description

As shown in Figure 1, Cascade Reservoir is located on the North Fork Payette River near Cascade, Idaho. The dam was constructed by the Bureau of Reclamation between 1946 and 1948 to provide 653,200 acre-feet of active storage for irrigated lands on the Payette Division of the Boise Project. The reservoir area is an increasingly popular year-round recreation resource, which supports boating, fishing, camping, picnicking, skiing, and snowmobiling. Extensive recreation homesite development has occurred since reservoir construction was completed.

Figure 1.  Map of Cascade Reservoir showing modeled segments

The North Fork Payette River above Cascade Dam drains 620 square miles of mountainous terrain in west-central Idaho. The average altitude of the watershed is 5,960 feet above mean sea level (AMSL). Primary land uses are forestry, agriculture, urban, and recreation. The reservoir receives nutrient loads associated with non-point source activities in the watershed and point source discharges from the city of McCall and the McCall Fish Hatchery.

Cascade Lake inundates 26,500 acres of fertile, gently sloping alluvium and glacial outwash in Long Valley, which also contributes to internal nutrient loading. The shallowness of the reservoir (mean depth of 26.5 feet), combined with watershed and nutrient loads released from sediments, results in conditions conducive to development of nuisance algal problems during late summer and oxygen depletion under snow and ice during winter.

Model Description

The Box, Exchange, Transport, Temperature, and Ecology of a Reservoir (BETTER) model (Bender et al., 1990) calculates flow exchange, heat budget, and dissolved oxygen. The model is two-dimensional with longitudinal segments (Figure 1) and vertical layers. Model elements are laterally averaged from bank to bank of the reservoir. The BETTER model simulates water quality constituents as a function of longitudinal and vertical location and time for transient hydraulic, meteorological, and inflow temperature conditions.

Using a 12-hour input-data time-step, the BETTER model simulates the seasonal warm surface layer and cold bottom water. The heat budget includes drybulb and dewpoint temperature, solar radiation, wind mixing, convective cooling, and inflow density distribution. The dissolved oxygen components include sediment oxygen demand (SOD); biochemical oxygen demand (BOD); ammonia; surface reaeration; and algal photosynthesis, respiration, and nutrient recycling. Algal biomass in the model represents one combined assemblage of diatom, green, and blue-green algae.

Model Inputs

Reservoir Geometry

The reservoir model geometry consists of longitudinal segments with vertical layers in each segment derived from the September 1995 bathymetric sediment survey.

Five reservoir branches and local inflow were modeled. North Fork Payette River was modeled as the main branch. The shallow embayment at the south end of the reservoir was modeled as a dynamic branch with minimal inflow and local inflow concentrations. Gold Fork, Lake Fork combined with Mud Creek, and Boulder Creek combined with Willow Creek were the other dynamic branches. Local inflow was distributed according to drainage area touching each segment.

A channel restriction just upstream of Cascade Dam traps cold bottom water. Inspection of the Cascade Dam forebay bathymetry revealed sediment deposition in the inlet channel to the turbine trashrack. Erosion of the forebay shoreline sloughed sediments and rocks into the inlet channel to the turbine trashrack. The sediment deposition extends upward to just over half the height of the turbine trashrack and restricts flow to the turbines.

Initial Conditions

Model simulations were started after ice-out on March 31 (day 90). A completely mixed reservoir initial condition (no stratification) was used; however, some temperature and dissolved oxygen stratification would be likely during day 90 for some years depending on reservoir and weather conditions before and after ice-out.

Flow

Calibration years were chosen so as to cover severely dry to wet conditions. Calendar years 1989 (slightly below average inflow), 1993 (slightly above average inflow), 1994 (severely dry), and 1995 (wet) were modeled. Known major tributary inflows, outflows, and change in reservoir elevation were used to derive unknown local inflow. The unknown local inflow was derived by calculating the change in reservoir storage volume from change in elevation, adding outflow, and subtracting the known inflows. The unknown local inflow can be negative due to evaporation, bank storage, and gage error.

Inflow Temperature

Temperature of inflows greatly affects model calibration. Cool inflows dive below the warm surface layer. Unfortunately, there were minimal inflow and inflow temperature data taken during spring after ice-out. Interpolation between known values provided an estimate. However, rapidly changing weather conditions, uncertainties in solar radiation, and large inflows increase the uncertainty of obtaining an accurate reservoir heat balance.

Inflow Loadings

Concentration and flow data from the Idaho Department of Environmental Quality were supplemented with data from STORET, the Environmental Protection Agency's database, to estimate loadings.

The quantity of bioavailable phosphorus from shoreline erosion was estimated to be minor in comparison with other sources. Therefore, nutrients from reservoir shoreline erosion were considered negligible.

Meteorology

Weather data for Cascade Reservoir were assembled from a variety of surrounding stations. The Cascade Reservoir model requires day (7 a.m. to 7 p.m.) and night (7 p.m. to 7 a.m.) drybulb temperature, dewpoint temperature, windspeed, and solar radiation. The meteorological data were sparse.

Model Calibration

Overview

Efforts concentrated on calibrating the model to 1994, a severely dry year, which would have a significant impact on water quality. The most complete data set was 1994. Both 1989 and 1993, the slightly below average and slightly above average inflow years, had some data gaps. However, a combination of all three years was used to calibrate the model. Initially, reservoir temperature and release temperature were calibrated for all three years. Dissolved oxygen and algae calibrations were calibrated next. Finally, anaerobic release of nutrients were calibrated.

The model was verified with data from 1995, a wet year. However, the model coefficients were adjusted slightly to improve the calibration to fit the range of dry to wet years. The 1995 calibration indicated that anaerobic conditions can occur even during a wet year due to reservoir operations. If summer and autumn flows are reduced, the bottom waters stagnate and poor water quality results.

Temperature

Temperature calibration indicated that Boise solar radiation was more intense than in the mountainous region of Cascade Reservoir. Therefore, model solar radiation was decreased by 25 percent. Considerable simulated vertical mixing was required to calibrate reservoir temperature. One of the most influential factors affecting reservoir and release temperature is the sediment deposition in front of the turbine intake trashrack described previously. Slight changes in conveyance area affect the calibration because cool water collects in the pool behind the deposition.

The top of Figure 2 is a modeled sideview slice of Cascade Reservoir temperatures from Cascade Dam (mile 0.0) to the North Fork Payette River inflow point on August 4, 1994. Warm surface water is withdrawn through the turbines because the sediment deposition restricts the bottom half of the trashrack intake. The top of Figure 3 shows observed penstock release temperature for August 4, 1994 (Day 216) being 21 C (70 F). The deposition extends from the reservoir bottom to slightly above the centerline of the forty foot trashrack opening.

Figure 2.  Modeled temperature and DO contours for August 4, 1994.

Dissolved Oxygen and Algae

The bottom of Figure 2 shows modeled dissolved oxygen stratification and anaerobic bottom waters. This modeled sideview slice of Cascade Reservoir from the North Fork Payette River inflow point to Cascade Dam on August 4, 1994, shows a mid-reservoir low DO "bubble" emanating from the bottom of Cascade Reservoir. Nutrients entering the upstream end of the reservoir stimulate algal growth. The algae drift downstream toward the middle of the reservoir, where deeper water slows flow velocity. The algae die, become detritus, settle to the bottom, and decay. Therefore, the anaerobic conditions are most severe at the sediment layer midway down the reservoir. Sediment oxygen demand behind the sediment deposition in front of the turbine trashrack is also large due to trapped decaying organics and stagnant water.

The lake bottom becomes anoxic during summer and nutrients are released. During fall turnover, extensive nutrients are mixed into the upper water column. Releases were affected by the anoxic bottom waters.

Observed release DO for August 4, 1994 (Day 216) is shown in the bottom of Figure 3. A destratified, completely mixed condition, meaning one DO concentration throughout the reservoir, was used for initial conditions on the first day modeled, March 31 (Day 90). However, observed release DO indicates that some DO stratification may have occurred on that day. Observed autumn maximum DO appears to be supersaturated, which may be caused by excessive photosynthesis, aeration near the penstock DO probe, or data collection errors.

Figure 3.  Observed maximum, mean, minimum, and modeled release temperatures and dissolved oxygen.

The field DO data indicate that three algal blooms might be possible each year. The spring algal bloom is likely due to excessive nutrients in the water column. An early fall blue-green algal bloom occurs due to atmospheric nitrogen fixation. A late fall algal bloom occurs due to release of nutrients from the bottom sediments and reservoir turnover. A zero algal settling rate, an above average phosphorus to nitrogen ratio for algal biomass, and a high anaerobic ammonia release rate were used to simulate blue-green algae. However, atmospheric nitrogen fixation was not directly modeled. Therefore, the model did not reproduce the early fall blue-green algae bloom.

Cascade Reservoir algae are primarily light-limited by algal self shading and are minimally affected by inorganic suspended solids. Inorganic suspended solids are minimal and quickly settle out of the water column.

The amount of flushing is important to Cascade Reservoir water quality. Linear regression of Cascade Reservoir chlorophyll-a (an indicator of algal biomass) to Cascade outflow showed some correlation (r-squared of 0.49) and was used in conjunction with plots of annual outflow to explain differences in yearly water quality. The amount of water in the reservoir system and how it is stored are both important factors affecting Cascade Reservoir water quality. For instance, chlorophyll-a during 1978 was about twice as high as during 1979. Counter intuitively, the inflow in 1978 was much higher than in 1979 (see Figure 3). However, the outflow in 1978 was not as high as in 1979 because the pool was being filled after the 1977 drought. Therefore, flushing of algae from the reservoir was minimal during 1978, the water had a long residence time, and the chlorophyll-a concentrations were high. The data regression indicated that the amount of flushing is important to Cascade Reservoir water quality.

Nutrients

North Fork Payette River is the largest source of loading to Cascade Reservoir. Considerable nutrients are also released from the sediments after the reservoir bottom waters turn anaerobic. Most of the surface nutrients are taken up by algae and settle out of the water column as algal detritus. Phosphorus and ammonia are released from the anaerobic sediments. Phosphorus and ammonia are first released in mid-reservoir, where anaerobic conditions appear during July. Nutrients are released from the sediments of Cascade Dam forebay during August. Autumn lake turnover distributes ample nutrients throughout the entire water column.

Modeled phosphorus loadings for 1989, an average flow year, are too low when compared to the other modeled years. The indicated minimal phosphorus loading for 1989 is apparently due to Department of Environmental Quality Laboratory detection limits of only 0.05 mg/L as P. Many concentrations between 0.00 and 0.05 mg/L were reported as 0.00 mg/L (Entranco Engineers, Inc., 1991). Phosphorus loadings were not increased for this error. Model results for 1989 should be viewed with caution. Phosphorus loadings for 1995, a wet year, also appear to be low.

Analysis of Management Alternatives

Overview

Control technologies and operational changes may enhance water quality of Cascade Reservoir. Entranco Engineers, Inc. (199 1) provided a thorough discussion of urban and agricultural watershed management and in-reservoir management techniques. Entranco discussed agricultural best management practices.

Entranco also discussed in-lake restoration techniques, including aluminum sulfate treatment and hypolimnetic aeration. In addition, Entranco discussed overall cost and water quality benefits to the reservoir for combinations of these techniques.

The following discussion expands or provides additional information based on a water quality modeling sensitivity analysis of Cascade Reservoir. Inflow loading reductions, sealing the sediments, dredging the trashrack inlet channel, increased spillway discharge, aeration of releases, and reservoir operational changes were modeled.

Inflow Loading Reductions

Permanently reducing inflow nutrients and organics would eventually help water quality or slow its degradation. However, modeling indicated that cutting 50 percent of all inflow nutrients and organics for one year would negligibly affect Cascade Reservoir DO or release DO in the shortterm. This is due to release of nutrients stored in the sediments during anaerobic conditions.

As shown in the top of Figure 4, release DO would decrease only slightly if initial reservoir concentrations and inflow loadings would be reduced by 50 percent. During spring, slightly more DO would occur because less decayable organic matter and less ammonia would exist. During summer, slightly less DO would occur because fewer nutrients would be available for algae and, therefore, less DO production. During autumn, there is almost no difference between the calibration and the simulation of 50-percent loading reduction. The slightly reduced amount of algal biomass and detritus has negligible effect on the reservoir bottom DO or anaerobic conditions during autumn. Major loading reductions for several years, natural leaching of readily available nutrients from the bottom sediments, and natural burial of bottom sediments would eventually improve water quality. Also, loading reductions in combination with other management techniques may be more effective than loading reductions alone.

Figure 4. Observed mean and modeled release dissolved oxygen versus time during 1994

Sealing the Sediments

The simulated elimination of all anaerobic release of Cascade Reservoir nutrients and organics and cutting sediment oxygen demand rates in half showed significant improvements in water quality. Fewer nutrients in the water column would reduce algal growth after autumn turnover and, therefore, would reduce settled decaying organics and SOD rates. Cutting SOD rates is a major factor in increasing the DO of the bottom layers. Cutting SOD rates greatly increases DO. The bottom of Figure 4 shows an increase in release DO of about 2 mg/L.

Entranco Engineers, Inc. (1991) discussed using a blanket of aluminum sulfate "alum" floc to retard diffusion of nutrients from Cascade Reservoir bottom sediments. However, the option of treating only the lower reservoir layers downstream of Sugarloaf Island where anaerobic conditions are most severe was not discussed. As discussed by Entranco, a reservoir this large has never been treated with alum, and the low alkalinity may prevent adequate flocculation.

Dredging the Trashrack Inlet Channel

Dredging the inlet channel to the turbine trashrack and possibly part of Cascade Dam forebay would release cold deoxygenated bottom water, warm the bottom, degrade water quality, and export nutrients and organics downstream. Therefore dredging of the inlet channel is not recommended. By dredging the inlet channel, the cold bottom water is released and replaced with warmer water. The warmer bottom temperatures would accelerate decay and sediment oxygen demand. Colder water would be released following dredging of the inlet channel, except after autumn turnover. Dredging the inlet channel in front of the trashrack to a depth of 4765 feet would allow about I C cooler releases. The amount excavated would influence the conveyance area and, therefore, the amount of bottom water withdrawn. The cold bottom water currently in front of the dam keeps the bottom sediments cool, which inhibits decay.

Some flushing of the mid-reservoir lower layers would occur if the sediment deposits in front of the turbine trashrack were removed. However, the decay of the sediment organics near the dam quickly consumes the DO. Also, nutrients and organics trapped behind the sediment deposition would be exported downstream. Exportation of nutrients could cause additional eutrophication of downstream rivers and reservoirs.

Dredging the inlet channel to the turbine trashrack is the maximum selective withdrawal option for releasing cold water. Increased spillway discharge is the maximum selective withdrawal option for releasing warm water. All other selective withdrawal options would fall between these bounds.

Increased Spillway Discharge

Releasing flows through the spillway instead of the turbines would not significantly improve water quality. Modeling indicated minimal differences in release temperature for 1994. Release DO during autumn would increase slightly due to surface water oxygenated by algal photosynthesis. The simulation assumed that all flows except 200 cfs would be discharged from the spillway. Due to the height of the sediment deposition in front of the turbine trashrack, the withdrawals for both turbine and spillway discharges are primarily from the mixed surface layer.

Aeration

Entranco Engineers, Inc. (199 1) discussed full and partial lift hypolinmetic aeration systems for Cascade Reservoir. To supplement Entranco's analysis, the BETTER model was used to simulate continuous aeration of Cascade Dam forebay at a rate of 1000 cfs-mg/L. Figure 28 shows that this rate of aeration would supersaturate releases during spring and would increase summer and autumn release DO about I mg/L. Forebay aeration would only increase Cascade Reservoir DO near the dam and releases. Considerably more aeration, expense, and risk would be required to aerate the lower layers of the entire reservoir. An undersized aeration system provides minimal benefit.

Reservoir Operational Changes

Decreases in the minimum pool volume can increase potential fish winterkill. Past studies (Reininger et al., 1982) indicated that maintenance of a 300,000 acre-foot total minimum conservation pool in Cascade Reservoir helps to minimize fish kills caused by low DO. Four Cascade BETTER model simulations using year 1994 (dry) hydrology at 250,000, 300,000, 370,000, and 400,000 acre-foot total minimum pools (including 50,000 acre-feet of inactive storage) were used to bracket the effects of various minimum pools on water quality. The simulation for 370,000 acre-foot total minimum pool represents 70,000 acre-feet of uncontracted storage added to the current 300,000 acre-foot total minimum conservation pool.

The simulations included the following assumptions:

- flood control was a top priority,

- Idaho Power would receive 200 cfs continuous year-round flow for power production,

- irrigators would receive adequate quantities unless the minimum pool assumptions would be violated,

- Cascade Reservoir would be lowered to minimum pool by November 1,

- additional water would be carried over to the next year, Payette Lake and Deadwood Reservoirs were the same as what occurred historically,

- Cascade inflows were the same as what occurred historically,

- power plant capacity was not exceeded to get back down to minimum pool,

- Cascade Reservoir was forced back to minimum pool by November 1, and

- the feasibility or legality of the minimum pools simulated was not considered.

An additional assumption for a dry flow year, such as 1994, was that irrigation demand would be met until August, when irrigation would be shut off to protect the minimum pool.

Fish refuge volumes based on temperature and DO modeling results should be viewed with caution due to sparse meteorological data. Meteorology significantly affects the size of these volumes. Therefore, results should be used for comparative purposes only rather than for determining the size of fish refuge volumes or operational releases.

Modeling indicated that with lower minimum pools, more surface water supersaturated with DO would be discharged during early spring. However, during late spring, summer, and early fall, lower DO concentrations would be discharged due to less DO in the reservoir forebay. During autumn, turnover occurs earlier with lower minimum pools and, therefore, release of supersaturated surface waters from algal photosynthesis occurs earlier.

The simulated volume of water with DO concentrations of less than 3 mg/L for the four total minimum pool values (which includes 50,000 acre-ft of dead storage) for the dry hydrology of 1994 were compared. With higher minimum pools, a greater volume of low DO water would exist in the main reservoir up the North Fork Payette River arm. However, the forebay hypolimnion has a higher average concentration of DO due to more mass of water. At higher simulated minimum pools, the forebay turns over later, thereby preventing hypolimnetic waters from being replenished with DO until late autumn. This result should be used with caution since the model uses a 12-hour time step, which "dampens" the effects of short-duration wind storms that could turn the reservoir over earlier.

Figure 5 compares seasonal variation in Cascade Reservoir trout habitat. Minimum trout habitat (MTH) conditions adapted from Heimer (1990) and referenced in "Plan of Study of A Minimum Conservation Pool in Cascade Reservoir" by Dewey Worth and Don Anderson (1996) were used as trout survival criteria. MTH is defined as water temperatures of less than 21 C and dissolved oxygen greater than 3 mg/L. While these conditions would provide for short-term survival, the associated stress would disrupt normal metabolic functions, reduce or cease feeding behavior, and affect growth and weight gain. Fish could tolerate short periods of exposure to these conditions at the expense of normal growth and behavior until habitat conditions improved and normal functions and activities resumed (Worth and Anderson, 1996).

Total minimum pools of 250,000 and 400,000 acre-feet were used to bracket the effects of the current operation of 300,000 acre-ft. Figure 5 shows the volume of water available for trout survival. Figure 5 indicates that at higher minimum pools, there would be more volume for trout to live and survive in. Second, at higher minimum pools, modeling indicated there would be more volume with poor water quality. Therefore, the fish would likely survive better with higher minimum pools; however, there would be a larger volume of poor water quality with the higher minimum pool. Trout would get squeezed into a small volume of water with the 250,000 acrefoot total minimum pool. Most of the lost refuge volume is due to warm temperatures. The modeled cool reservoir volume between the total reservoir volume and reservoir volume with temperature of less than 21 C on day 240 would only be 17.4 million cubic meters (5 percent of the total volume on this day for the modeled 250,000 acre-foot minimum pool). If the usable trout habitat temperature of 19 C were used, this volume would be 13.6 million cubic meters (about 4 percent of the total volume). A major conclusion is that operations using low minimum pools will stress trout in dry years. Also, if a dry year follows a dry year, the reservoir may not be filled to the minimum pool desired.

Modeled 1994 (dry year) and 1993 (average year) minimum trout habitat for the entire reservoir were compared to the portion of reservoir upstream of Sugarloaf Island for a total minimum pool of 250,000 acre-feet. Under these shallow minimum pool conditions, trout would have minimal volume in which to survive. During the warm months of a dry year there would be no trout habitat upstream of Sugarloaf Island and minimal habitat in the entire reservoir. During August of an average year there would be more, but still minimal volume for trout upstream of Sugarloaf Island. However, there would be significant volume for trout in the entire reservoir.

Higher minimum pools have some disadvantages. Extra storage for wintering and spawning in Cascade Reservoir and downstream fisheries means less would be available for agricultural, municipal, and industrial needs. At higher minimum pools, more bank erosion and sloughing would be expected to occur.

Mechanisms that result in bank failure and erosion include bank saturation and softening during full pool, weakening of the upper bank by piping of fine sediments, bank sloughing as the reservoir falls to normal pool, erosion due to surface runoff and flooding, and erosion due to wind and vessel traffic waves.

Maintaining a higher minimum pool would also limit flushing potential during warm months. During 1995 (wet year), water for late fall salmon migration was stored in Cascade Reservoir during summer, thereby reducing summer releases and raising the pool level. In Figure 6, minimum trout habitat for the 1995 calibration is compared to that for 1200 cfs continuous release after July 6 (Day 187 ) through September 27 (Day 270), which corresponds to releasing about half of the salmon water during summer. The continuous flushing increases the DO of lower layers and pushes nutrients towards the dam, which stimulates algal photosynthesis, thereby increasing epilimnetic DO concentrations. Correspondingly, the upstream end of Cascade Reservoir has less DO due to less algal photosynthesis. Continuous flushing after July 6, 1995, increases the refuge volume for trout during the critical summer period, as shown in Figure 6. By September 27 (Day 270) the reservoir elevation would have dropped 3 feet with continuous 1200 cfs releases after July 6, 1995, when compared to the calibration.

Figure 6. Modeled minimum Cascade Reservoir trout habitat for 1995 calibration versus 1200-cfs continuous turbine release (wet year).

After July 6, during dry and average years, more than 1200 cfs is typically released to satisfy downstream demands. Therefore, changing operations for continuous flushing, rather than storing water during summer, is most applicable to wet years.

REFERENCES

Bender, M.D., "Two-Dimensional Water Quality Modeling of Cascade Reservoir," U.S. Department of Interior, Bureau of Reclamation, Technical Service Center, Denver Colorado, August 4, 1997.

Bender, M.D., G.E., Hauser, M.C. Shiao, and W.D. Proctor, "BETTER: A Two-Dimensional Reservoir Water Quality Model, Technical Reference Manual and User's Guide," TVA Engineering Laboratory, Report No. WR28-2-590-152, Norris, Tennessee, October, 1990.

Entranco Engineers, Inc., "Cascade Reservoir Watershed Project Water Quality Management Plan," Prepared for Valley Soil Conservation District and Idaho Division of Environmental Quality, January 1991.

Idaho Division of Environmental Quality, "Cascade Reservoir Phase I Watershed Management Plan," Southwest Idaho Regional Office, Boise, Idaho, January 1996.

Reininger, Bruce; Rieman, Bruce; and Horner, Ned, "Cascade Reservoir Fisheries and Limnological Investigations," Idaho Department of Fish and Game, Boise, Idaho, December 1982.

Valley Soil Conservation District (VSCD), Idaho Soil Conservation Commission, Idaho Department of Health and Welfare-Division of Environmental Quality, USDA-Soil Conservation Service, "Agricultural Pollution Abatement Plan, Cascade Reservoir Watershed, Valley County, Idaho, Final Planning Report," January 199 1.

Worth, Dewey and Anderson, Don Jr., "A Plan of Study for a Minimum Conservation Pool in Cascade Reservoir," Idaho Division of Environmental Quality, Cascade, Idaho, and Idaho Department of Fish and Game, McCall, Idaho, August 1996.

Zimmer, David W., "Phosphorus Loading and Bacterial Contamination of Cascade Reservoir," Boise, Project, Idaho, U.S. Department of Interior, Bureau of Reclamation, Pacific Northwest Region, August 1983.