Scientific Investigations Report 2006–5060
U.S. GEOLOGICAL SURVEY
Scientific Investigations Report 2006–5060
A total of 36 model scenarios, including the 2001 and 2002 base cases, were run to examine the effect of three dam raises, two water delivery levels, various outlet configurations, and two schemes for augmenting inflows to Hagg Lake. These scenarios are summarized in table 1.
Increases in dam height of 6.1 m (20 ft), 7.6 m (25 ft), and 12.2 m (40 ft) resulted in simulated full-pool total storages of 88,800, 95,800, and 118,000 acre-ft, respectively, with the Hagg Lake model grid. The simulated base case storage at full pool was 63,900 acre-ft (slightly larger than the Bureau of Reclamation’s reported full-pool storage of 62,216 acre-ft). These increases in dam height required the determination of new “fill curves,” the targeted daily lake elevations used to optimize reservoir operation. This curve requires water levels to be below full pool in winter to provide for 20,600 acre‑ft of available storage for flood control. The fill curve then specifies an increase in water level through May to reach full pool, thereby maximizing water availability for summertime withdrawals. Three new fill curves were generated based on the current Hagg Lake fill curve. The maximum elevation of the fill curve was increased by 6.1, 7.6, or 12.2 m (20, 25, or 40 ft), and the minimum (winter) fill curve elevation was determined as the lake elevation that provides space for 20,600 acre-ft of flood control storage.
In order to fill the increased volume of an enlarged reservoir, all or part of the water (above deliveries and instream flow requirements) released from the reservoir in December of the previous year to make room for flood control storage was kept in the reservoir. This resulted in an initial (January 1) water surface elevation that was 2.1 m higher than the 2002 base case for all scenarios involving a dam raise (but without augmented inflows). For 2001, the drought year, keeping the excess spilled water from late 2000 only raised the water surface elevation by 0.02 m (late 2000 was fairly dry). In the model runs, water was spilled from the reservoir when the water surface elevation reached the applicable fill curve, thus keeping the water level from exceeding the fill curve.
According to one analysis, although a 12.2 m (40 ft) increase in the height of Scoggins Dam provides sufficient storage, unaugmented inflows into the lake would meet future water demands in only about 20 percent of years (Ryan Murdock, Montgomery Watson Harza, written commun., 2005). Therefore, two additional sources of inflow water are considered in this study in an attempt to more reliably fill the enlarged lake created by the larger [12.2 and 7.6 m (40 and 25 ft)] dam raises. Additional inflows were not provided in the 6.1 m (20 ft) dam raise scenarios (table 1), because these scenarios were only run for year 2002, when there was enough natural flow to fill that reservoir size.
One option to augment inflows to Hagg Lake is to use water in the upper Tualatin River and route it via gravity through a tunnel into Sain Creek, a tributary of Hagg Lake. Available water would be defined as the amount that is above both the Tualatin River’s environmental needs and downstream water users’ needs. The water would be diverted from Hillsboro Reservoir on the Tualatin River, at approximately the same location as Haines Falls [river mile (RM) 73.3, fig. 1]. The length of the tunnel would be between 2.8 and 3.9 mi.
Model scenarios were completed with two flow schedules for the tunnel: “low” and “high” inflow. The low inflow version would use available water after considering water rights. In these scenarios, the diversion was 150 ft3/s in December and January, 120 ft3/s in February, and 60 ft3/s in March, corresponding to a total diversion of approximately 28,800 acre-ft. The high inflow schedule would use water availability plus a portion of “natural flow” water rights allowed to the Joint Water Commission, a consortium of Tualatin River Basin water providers. This would divert 250 ft3/s in December and January, 220 ft3/s in February, and 160 ft3/s in March, for a total of 52,800 acre-ft (Tom VanderPlaat, Clean Water Services, written commun., 2005). In these scenarios, the Sain Creek tunnel was used from January through March.
In addition to these winter inflows, which would help fill the reservoir, water released from Barney Reservoir (fig. 1) to the upper Tualatin River in summer was modeled to be routed into Hagg Lake via the Sain Creek tunnel but not kept in the lake; an identical flow was released from Hagg Lake. This arrangement allowed the water released from Barney Reservoir to be discharged into future pipelines connected to the Scoggins Dam outlet works. The amount of these Barney Reservoir releases/inflows were obtained from the annual flow reports for the appropriate year (Tualatin River Flow Management Technical Committee, 2001; 2002). These flows generally occurred between early June and mid-November and ranged from 14 to 54 ft3/s.
Water temperatures assigned to Sain Creek tunnel flows for 2002 were taken from measurements in the Tualatin River from below Lee Falls (RM 70.5). Besides temperature, no other water quality data were collected there, so data from Cherry Grove (RM 67.8) on the upper Tualatin River were used. Water of this composition was mixed together with Sain Creek water to obtain a mixed flow, temperature, and water quality input from Sain Creek for the simulations with the new inflows. The Lee Falls water temperatures were similar to Sain Creek temperatures for most of the year, but as much as 4°C cooler in August and September. The upper Tualatin River water may be cooler in part due to releases from Barney Reservoir upstream. The conductivity of upper Tualatin River water was lower in summer, by as much as 40 µS/cm, probably for the same reason. Dissolved oxygen concentrations usually were higher (by about 1 mg/L) through most of the year, and nitrite/nitrate concentrations were approximately 0.1 mg/L lower in the upper Tualatin River. Orthophosphate concentrations in both waters were similar.
Another proposed method of augmenting inflows to Hagg Lake during the winter is to pump water into the lake from downstream. This method would take advantage of a proposed pipeline connecting Scoggins Dam to the Joint Water Commission’s (JWC) drinking water treatment plant south of Forest Grove. The main purpose of that pipeline would be to efficiently deliver water from Hagg Lake to the drinking water treatment plant via gravity during summer months when JWC requires water from the lake. CWS also may use the pipeline to transport some of its flow-augmentation water.
This pump-back option proposes to run the pipeline in reverse during the winter (December to May) using the Tualatin Valley Irrigation District’s (TVID) equipment at its Springhill Pumping Plant (RM 56.1, fig. 1) to draw water from the Tualatin River and pump it uphill through the pipeline and into Hagg Lake. The Springhill Pumping Plant is located adjacent to the JWC drinking water treatment plant; both TVID and JWC currently draw water from the Tualatin River at that location. The TVID pumps are not used in the winter due to the lack of irrigation demands, so they could be used for pump-back.
In this pump-back scenario, available flows in the Tualatin River could be delivered to Hagg Lake via the TVID pumping array and the JWC pipeline. Modeling studies of this option by Montgomery Watson Harza (Ryan Murdock, written commun., 2005) considering the capacity of the pipe, the size of the pumps, competing needs for the pipeline and pumps, and minimum Tualatin River streamflows, resulted in the following constraints on pump-back water:
The quality of pump-back water was estimated using a flow-weighted mix of water from Gales Creek and the Tualatin River at Dilley (RM 58.8), as the pump plant is just downstream of that confluence. Total suspended solids, total phosphorus, and orthophosphate concentrations were higher in the winter of 2001–02 than the winter of 2000–01. Although these pump-back scenarios did not limit pumping during storms, it is possible that pumping might be turned off if the source Tualatin River water had excessive loads of suspended sediment, in order to minimize the importation of phosphorus and sediment into the lake and to minimize mechanical damage to the pumping equipment.
One of the concerns of the pump-back option is that Tualatin River water at the Springhill Pumping Plant probably has higher phosphorus concentrations than water that originates in the Scoggins Creek drainage upstream of Hagg Lake. This extra phosphorus load to the lake might stimulate additional algal growth in the lake during the summer, thus creating higher levels of organic matter that, in turn, could cause problematic levels of disinfection byproducts or taste or odor problems once that material is delivered to the JWC drinking water treatment plant. Model scenarios that use the pump-back option were evaluated for this effect.
Another potential concern is that pumping would import contaminants into Hagg Lake. Sampling for pesticides by the USGS at the Springhill Pumping Plant over the course of a year (88 samples from March 2002 through February 2003) showed some detections of common pesticides such as atrazine, simazine, and diazinon, but at very low concentrations (<0.16 µg/L; U.S. Geological Survey, unpub. data, 2002-03). None of the detections appeared to come close to their regulatory criteria for the protection of aquatic life or for drinking water. Treatment processes may further reduce or remove these materials from finished drinking water. So, although low concentrations would be imported into Hagg Lake via pumping, the concentrations do not appear to be of great concern.
Pump-back scenarios were run for 2002 and 2001 for the 12.2 m (40 ft) and 7.6 m (25 ft) dam raises and for initial and future maximum water deliveries. Use of the pump-back option to augment inflows to Hagg Lake in these model runs was implemented in a manner consistent with the monthly inflow triggers, Tualatin River water rights, the pump-back target fill curve, and maximum pumping rates. In 2002 with the 12.2 m (40 ft) raise, this scenario resulted in the use of the pump-back option intermittently from December 3 to March 10, delivering 32,100 acre-ft of water to the lake. A total of 18,400 acre-ft was spilled between January 7 and May 31, when natural inflows raised the lake level above the fill curve. In 2002 with the 7.6 m (25 ft) raise, pump-back occurred intermittently between December 4 and May 1, delivering 22,300 acre-ft of water to Hagg Lake. A total of 30,400 acre-ft was spilled between December 11 and May 31. Alternate rules for pump-back might have minimized the amount of water that was spilled in these 2002 pump-back scenarios, thereby also reducing the amount of water that was pumped and reducing the importation of phosphorus and other constituents from the Tualatin River downstream. These scenarios, then, can be considered as a test of those conditions that might occur if the volume of pump-back were maximized —a useful test when exploring the limits of an influence on water quality. In 2001 with the 12.2 m (40 ft) raise, deliveries of 25,900 acre-ft were made between December 1 and May 19. In 2001 with the 7.6 m (25 ft) raise, 24,000 acre-ft of water was pumped into Hagg Lake between December 1 and May 19. No spill occurred in 2001, as the lake level did not reach the fill curve.
In all model scenarios, water was released from Hagg Lake to meet instream flow requirements in Scoggins Creek below the dam, to meet requirements of downstream users, and to spill excess water when lake levels rose above the fill curve. Annual deliveries for instream requirements and downstream water users are shown in table 2 for each scenario.
For the base case in 2002, minimum discharges to meet instream flow requirements in lower Scoggins Creek were 10 ft3/s from December through September and 20 ft3/s in October and November. New instream flow requirements would accompany a dam raise, resulting in a minimum flow of 25 ft3/s through the entire year. Some of the water releases for downstream users were delivered into Scoggins Creek downstream of the dam and thus could be used to meet instream flow requirements. Other releases to downstream users (all JWC water and a portion of CWS water) were routed into the not-yet-built JWC pipeline and therefore could not be used to meet instream flow requirements. If water deliveries to Scoggins Creek for downstream customers were insufficient to meet instream flow requirements, then water was released from the lake specifically to meet those requirements. Minimum instream flows were given first priority before apportioning water to future users.
CWS, JWC, TVID, and the Lake Oswego Corporation (LO) all hold water rights to stored water in Hagg Lake. All these entities received water in the base case, but only CWS and JWC would receive additional water from an enlarged Hagg Lake. For the 12.2 m (40 ft) dam raise option, CWS would receive as much as an additional 15,000 acre‑ft, and JWC would receive as much as an additional 37,000 acre-ft, depending on whether the lake filled. For the 6.1 m (20 ft) dam raise, CWS would receive as much as an additional 6,540 acre-ft, and JWC would receive as much as an additional 16,100 acre-ft.
As lake level is known to affect water quality in Hagg Lake (Sullivan and Rounds, 2005), two levels of water deliveries to downstream users were simulated in all model runs involving a dam raise. The first level of water delivery (“a” scenarios) was characterized by lower outflows, simulating water demands just after the dam raise is finished. Most deliveries in these scenarios were set to actual deliveries from the base case year, either 2001 or 2002, with one exception. CWS, which releases water from Hagg Lake to improve water quality in the lower Tualatin River, would immediately take delivery of its new allocation. The timing and amount of new releases for CWS was determined by meeting minimum flow targets in the Tualatin River at Farmington Bridge (RM 33.3). These new targets were set to 156 ft3/s in July, 195 ft3/s in August, and 234 ft3/s from September through November 15 for all dam height increases. To meet these targets under the hydrologic conditions of 2002, CWS would need 25,500 acre-ft of water to be released from Hagg Lake. This is less than the total amount of water allocated to CWS with the dam raise; when available, any extra water was kept in the lake. In 2001, CWS would have required 22,000 acre-ft of water to meet minimum flows targets at Farmington Bridge. In this drought year, no new water was available, scale-backs were necessary, and this amount of water was not released in any of the scenarios.
The second delivery level (“b” scenarios) had higher outflows, simulating maximum future water demands. In these scenarios, all current users were allocated deliveries up to their contract levels; at present, some users are not using all their allocated water. Additional stored water from the dam raise was then allocated. In some scenarios, new deliveries had to be scaled back according to the following procedure. On May 31, the volume of new water available was determined from lake level and volume-elevation curves. If the new storage was greater than the new allocations and instream rights, then all users received 100 percent of their new water. However, if the new storage was less than or equal to the instream rights, users did not get any new water, and cutbacks to old water were made, if necessary, in order to fulfill the instream rights. If the new storage was greater than the instream rights, but less than that required to meet new allocations and instream rights, then the instream rights were met first, by setting aside an appropriate amount of water for the rest of the year, and the remaining new storage was distributed to users on the basis of a fraction of what was left.
The number and configuration of outlet structures and the operational rules for their use can have an important effect on the lake’s water quality dynamics. Scenarios were run with several outlet configurations to examine this effect. Presently, the lake has one main outlet (plus a spillway), with an opening elevation between 69.8 and 72.5 m above sea level. Several model scenarios were run with this original fixed-elevation outlet (table 1). In other scenarios, additional fixed or sliding-gate (variable-elevation) outlets were added and used in conjunction with the original outlet (fig. 2). In some cases, a sliding-gate outlet was used alone. The scenarios with more than one fixed-elevation outlet, or any scenario with a sliding‑gate outlet, allowed water to be withdrawn from different elevations in the lake. In this way, water could be withdrawn from different depths and blended, if necessary, to attain an outflow with desired water quality characteristics. These scenarios are said to have “selective withdrawal” capability. Some dams have been retrofitted to add this capability in the form of multiple fixed gates, such as at Shasta Dam in northern California, or in the form of an assembly of multiple sliding gates, such as at Cougar Dam east of Eugene, Oregon; both of these retrofits were done to help attain downstream temperature targets.
Hagg Lake model scenarios with selective withdrawal capability were run in an attempt to meet a downstream water temperature target. The target temperature used in this study was created to (1) provide a more-natural seasonal temperature pattern for Scoggins Creek downstream of the dam, and (2) comply with Oregon’s temperature standard. One could simply use the temperature standard as a target, but that leaves no room for warming in downstream reaches, and does not match conditions upstream of the reservoir. In this study, the temperature target was the 30-day running average of the 30-day running average of the 7-day running average of the measured daily maximum water temperature in Scoggins Creek upstream of Hagg Lake for 2002. Water temperature standards in Oregon are based on the 7-day running average of the daily maximum (7dADM); the two 30-day running averages were used to smooth and eliminate the influence of shorter timescale variations due to weather patterns. In this way, the target has a close similarity to upstream conditions. The State maximum water temperature standard in Scoggins Creek downstream of Hagg Lake is 13°C from October 15 to May 15, and 18°C for the rest of the year (Oregon Department of Environmental Quality, 2003). The target temperature meets these criteria throughout the year (fig. 3). Other temperature targets could easily be created and tested, but this one provided a link between upstream and downstream conditions.
The 7dADM water temperature downstream of the dam in 2002, as in most years, did not comply with Oregon’s temperature standard for Scoggins Creek (fig. 3). It is not known whether additional lake outlets and selective withdrawal might be required as part of a dam raise; these model scenarios were included simply to explore the consequences of including selective withdrawal as a means of complying with the temperature criteria in Scoggins Creek. The use of these scenarios is not meant to imply that temperature regulations would require selective withdrawal, only that this is one means of addressing the downstream temperature issues.
Version 3.12 of CE-QUAL-W2 does not contain algorithms that allow the dynamic blending of withdrawals from multiple outlets to meet a target release water temperature. Nor does it have algorithms to automatically select the optimum depth for a sliding-gate outlet. In some past applications of CE-QUAL-W2, these necessary operational and blending decisions were made outside of the model and optimized in an iterative manner (Hanna and others, 1999; Bartholow and others, 2001), a process that required many model runs. For this work, a new subroutine was written for CE-QUAL-W2 that internalized these procedures, allowing blending rates and sliding-gate positions to be set during the model run; in this way, iteration with many model runs was avoided. Details of this new subroutine are described in Appendix A of this report. When selective withdrawal and blending was used, the primary goal was to meet the downstream temperature targets, using operational rules that are detailed in Appendix A. The operational rules tested in these scenarios are but one of many potential sets of rules and may not be an optimal design. They are sufficient for use in these scenarios, however, as a demonstration and proof of concept.
For more information about USGS activities in Oregon, visit the USGS Oregon Water Science Center home page.
U.S.
Department of the Interior | U.S. Geological
Survey
Persistent URL: https://pubs.water.usgs.gov/sir20065060
Page Contact Information: Publications
Team
Page Last Modified:
Thursday, 01-Dec-2016 19:05:02 EST
