Implications for Dam Operation and Planning


The Detroit Dam model results from this study show the range of release temperatures that might occur under varying hydrologic and meteorological conditions as well as under several operational and structural scenarios. A common theme among all model results is that spring and summer dam operations have an effect on operational flexibility and the extent to which release temperatures can be controlled later in autumn. Model results indicate that as early in the year as April, solar radiation heats the surface of the lake and thermal stratification begins. Because most of the lake vertical profile is still relatively cool at that time, the ability to meet downstream temperature targets during spring is dependent on the ability to access and release warmer water near the lake surface. This can be difficult when the lake surface is either well below or well above the spillway crest in spring and early summer. As the surface of the lake becomes warmer throughout summer and the thermocline moves deeper in the lake, access to cool water below the thermocline begins to decrease from about June until about mid-November, at which point the lake has been drawn down to make room for potential flood storage and typically is isothermal. In general, the release of warm surface water from the lake during summer allows the cooler water deeper in the lake to be reserved until autumn when that cold water is needed most to meet downstream temperature targets. 


The ability to mix and release (warm) lake surface water with (cold) deeper water throughout the year often is the limiting factor in controlling release temperatures from Detroit Lake or other deep reservoirs with similar outlet configurations. The existing outlets at Detroit Dam do not allow near-surface waters to be released during times when the lake elevation is below the spillway crest (spring and autumn). During years in which the reservoir may be late to fill or not fill at all (as in hot/dry and base model scenarios), the spillway may only be a viable release point for a limited time in summer. Immediately after the lake is drawn down below the spillway crest elevation, dam operations with existing outlets must release cool water from below the thermocline using either the power penstocks or the upper ROs. Later in the year, the cool water supply below the thermocline can become exhausted at the elevation of the available outlets, and an uncontrollable rise in release temperatures results from about October through November. Thus, the existing structures allow the managers and operators of Detroit Dam to blend releases from multiple outlets for only part of the year, with less flexibility in drier years.


Power production requirements limit the use of existing structures at Detroit Dam to expel warmer water in summer (June–September) and cooler water in autumn (October–November). Operational scenarios with no minimum outflow requirements to the power penstocks (noppmin) led to outflow temperatures from Detroit Dam that were closer to meeting downstream temperature targets than operations with dedicated minimum flows for power production.


Model simulations indicate that by delaying the drawdown of Detroit Lake in autumn, better control over release temperatures is possible immediately downstream of Big Cliff Dam. Delaying the drawdown of the lake for better downstream temperature management must be balanced against the need to make room in the reservoir to manage storm-related November inflows that might lead to floods. The temperature benefits of delaying lake drawdown result mainly from an extended use of the Detroit Dam spillway until as late as November 1 (see delay_dd1 and delay_dd2 scenarios in figs. 14 and 17). By delaying the date at which drawdown begins, warm epilimnetic water can continue to be released in conjunction with cool water from the hypolimnion, thereby rationing the deeper cool-water supply throughout autumn. As a result of this sustained use of the spillway under these operational scenarios (figs. 15, 16, 18, and 19), the abrupt change in release temperature caused by the loss of spillway usage is not as apparent as with base operational scenarios (fig. 8). Such abrupt changes in release temperatures may or may not be a consideration for downstream salmon habitat during late summer and autumn. Farther downstream of Big Cliff Dam, however, the North Santiam River model results show that decreased releases during September 1–October 15 necessary to keep lake levels high in scenario h8 (see delay_dd2 operations in table 4) cause substantial downstream warming (2–5°C).


Aside from operationally delaying the drawdown of Detroit Lake, a number of simulated structural scenarios showed that the addition of hypothetical floating outlets at Detroit Dam could provide access to warm surface water to be released spring through autumn, allowing better management of release temperatures throughout the season. Adding a floating outlet generally leads to greater control of the outflow temperature compared with existing outlets at Detroit Dam, even under hot/dry conditions. Combining the upper ROs with a floating outlet (uro-float) resulted in greater temperature control in autumn than the combination of the power penstocks and a floating outlet (pp-float) (compare figs. 23 and 24 or fig. 22 with 25). As the elevation of the lower outlet was decreased (going from the power penstock elevation to the upper RO elevation), the amount of outflow temperature control at Detroit Dam increased. As decreased minimum flow requirements were placed on the lower, fixed-elevation outlets in these scenarios (that is, an increase in the allowable “percent spill”), the resulting outflows generally were cooler in autumn. Likewise, warmer outflows during June and July were possible under these scenarios and may have contributed to the relatively large supply of accessible cool water in the lake later in autumn. 


When a hypothetical sliding outlet was used alone (slider1340), outflow temperatures roughly met the max temperature target at Detroit Dam (fig. 20), yet this scenario resulted in more day-to-day temperature variation than equivalent scenarios incorporating both a floating and lower (fixed-elevation or sliding-gate) outlet. This illustrates the value provided by having two outlets to access warm and cold water separately throughout the year. As the thermocline moves up and down in the water column during the day due to seiching, a more variable release temperature results from a single sliding-gate outlet (slider1340) than from a blended combination of one floating outlet withdrawing warmer surface water and one fixed-elevation outlet withdrawing cooler water (slider1340-float). 


The estimated emergence date of spring Chinook salmon was tabulated as a way of comparing the relative success of the model scenarios in this study. Success, as measured in this study, is a delay in the estimated spring Chinook emergence date, as early emergence can be problematic (U.S. Army Corps of Engineers, 2012). The comparison showed that many structural scenarios and scenarios in which no minimum flow was directed to the power penstocks generally led to later emergence dates. Under hot/dry environmental forcing conditions, structural scenarios generally exhibited later emergence dates than scenarios incorporating only operational changes to Detroit Dam, perhaps in large part because of the limited use of the spillway under hot/dry conditions in non-structural scenarios. Downstream in the North Santiam River, estimated emergence dates also were influenced by cool inflows from large tributaries such as the Little North Santiam River, which delayed the emergence date significantly. The emergence date is not the only factor involved in assessing the biological success of an operational or structural scenario, however, as with other streamflow and habitat considerations, may be important. 


Results from the Detroit Lake model show that the ability to control release temperatures and meet downstream temperature targets throughout the year can be more closely attained at the site of the dam by delaying drawdown of the lake in autumn, decreasing the minimum power-generation requirement during summer/autumn, and (or) installing a well-conceived combination of floating and (or) sliding-gate outlets. 


Results from the North Santiam and Santiam River model downstream of Big Cliff Dam show that release temperatures from Detroit and Big Cliff Dams have an important and measurable effect in the North Santiam River, especially in autumn as days become shorter and solar radiation imposes less heating to the river. The river modeling illustrated the importance of both flow rate and water temperature downstream of Big Cliff Dam during autumn to benefit spring Chinook salmon spawning. The temperature effects of altered releases at Detroit Dam tend to diminish with downstream distance, but the effects are large enough to be measurable throughout the North Santiam and Santiam River systems. The temperatures and seasonal temperature pattern downstream of Detroit Dam in the North Santiam River system can be managed and controlled through a variety of changes in dam operations or outlet options at the upstream dams. 


First posted October 30, 2012

Revised June 11, 2013