Summer Shasta Releases Are Too High and Lower Sacramento River Summer Flows Are Too Low: Lessons Learned – #5

Following an introductory post, this is the seventh post in a series on the lessons learned by the National Marine Fisheries Service (NMFS) from the 2013-2015 drought that devastated Sacramento River salmon populations.  This post addresses Lesson #5.

The reason for high summer releases to the Sacramento River from Shasta and Keswick reservoirs is to meet the demands of Sacramento Valley water contractors for stored water.  For example, in 2012, summer releases to the Sacramento were 15,000 cfs, with roughly 7000-8000 cfs diverted in the Sacramento Valley for water supply use (Figure 1).

The Bureau of Reclamation learned during the 2012-2015 drought that, if the previous year was wet, it must still limit water releases in summer of dry years to 12,000-13,000 cfs (Figures 2 and 3) to preserve Shasta Reservoir’s cold-water pool supply.  One consequence of this limitation as implemented has been less flow in the lower river 150-200 miles downstream.  This reduced flow has exacerbated water temperature problems in the lower river.  Reclamation has likely also reduced water deliveries in the Sacramento Valley to some degree, but the extent of any such reductions is difficult to tease out.

Because of the lessons learned, Reclamation in 2018 targeted a 53oF water temperature limit in the main spawning reach of winter-run salmon from Keswick Dam (RM 300) down to the mouth of Clear Creek (RM290) (Figure 4).  In the past, the water temperature limit had been higher (56-58oF).  In 2020, the target was once again set higher to sustain a depleted cold-water pool supply through the summer and fall. The target in the spawning reach in drought years 2014 and 2015 was 56oF, which proved ineffective at providing egg/embryo survival.

One of the actions to sustain the cold-water pool has been to limit June-July Keswick releases (Figure 5) to near 11,000 cfs in wet years (2017 and 2019).  Such action cuts into water supply deliveries and leads to reduced flows (Figure 6) and excessive water temperatures (>70 oF, Figure 7) in the lower 200 miles of river.  Without simultaneous reductions in Sacramento Valley water deliveries, reductions in Keswick releases lead to excessive water temperatures downstream of the upper 10-mile salmon spawning reach.   This violates the Central Valley Basin Plan’s temperature standard for the lower reaches of the Sacramento River (68oF).  It causes stress on rearing and migrating salmon and sturgeon, and if high enough severely retards upstream migration of adult salmon.

The obvious lesson learned is that Reclamation must limit summer Shasta cold-water storage releases and maintain sufficient lower river flows.  This will necessarily require Reclamation to more greatly restrict water supply deliveries in the Sacramento Valley than it has historically and recently.

Figure 1.

Figure 2.

Figures 1-3. Sacramento River flow (cfs) at Keswick Dam (rm 300) and Wilkins Slough (rm 120) in summer of dry years 2012, 2018, and 2020. Note: the difference between

Figure 4. September-October water temperatures in the upper Sacramento River at Clear Creek (river mile 290) from 2012-2020. Note loss of temperature control in 2014 and 2015. Note limited control in 2013, 2018, and 2020. Note meeting target 53oF 2012, 2016, 2017, and 2019

Figure 5. Flow rates below Keswick Dam in June-July 2012-2020.

Figure 6. July-September Sacramento River flow rates at Wilkins Slough (rm 120) 2015-2020. Note flows were compromised in summer 2016 and 2017 to help preserve Shasta’s cold-water pool supply and upper river water supply deliveries.

Figure 7. Water temperatures in Sacramento River at Wilkins Slough (river mile 120) May-Oct, 2015-2020

Shasta Water Temperature Experiment #1 4/10-4/11, 2021

In the evening of 4/10 and early morning hours of 4/11, the Bureau of Reclamation began releasing warm surface water from Shasta Reservoir through the upper river outlets to the dam spillway (see inset at right), bypassing the Shasta Temperature Control Device (TCD) and the Shasta powerhouse,  forgoing power production.

Figure 1 shows the warm water beginning at 20:00 4/10 and continuing to 8:00 4/11.  Reclamation changed the recent pattern of midnight releases, raising releases from about 3000 cfs releases to about 6000 cfs (Figure 2).  This was possible due to the available capacity of the upper-level river outlets on the dam spillway.  Reclamation also curtailed peaking power releases through the TCD and penstocks on the evening of 4/10.  The 55ºF temperature in Figure 1 immediately downstream of Shasta Dam reflects the temperature of the lake’s surface water at the upper level outlets (Figure 3).

During and after the experiment, water temperatures in the Sacramento River downstream increased less than 1ºF, both immediately below Keswick Dam (Figure 4) and 10 miles further downstream at the Clear Creek gage (Figure 5).

Overall, the experiment shows that forgoing some peaking power by bypassing penstocks and releasing warmer upper-level lake water in April can save cold-water-pool volume, although with a small increase (0.5ºF in this case) in downstream water temperatures under these specific conditions.

Figure 1. Water temperature immediately below Shasta Dam at gage SHD 4/7-4/12, 2021. Note substantial increase from 20:00 4/10 to 06:00 4/11 during Exp #1.

Figure 2. Inflow to Keswick Reservoir from Shasta Dam 4/7-4/12, 2021. There was also 400-750 cfs of Trinity water entering Keswick from Spring Creek Powerhouse.

Figure 3. Water temperature profile in lake at Shasta Dam on 4/14/2021. Note water temperature in mid-50s at level of the upper six outlets to the spillway, in contrast to the cooler 50ºF water entering the penstocks via the middle gates of the TCD.

Figure 4. Water temperature of water released from Keswick reservoir 4/7-4/12, 2021. Note temperature increased up to 0.5-0.8ºF during and after the experiment.

Figure 5. Daily average water temperatures at selected gages 4/1-4-12, 2021. Note temperatures at Clear Creek confluence (CCR, in bold), with little increase on 4/10-4/11.

Are Sacramento River Water Temperatures Related to Flow: Lessons Learned – #4

Following an introductory post, this is the sixth post in a series on the lessons learned by the National Marine Fisheries Service (NMFS) from the 2013-2015 drought that devastated Sacramento River salmon populations.  This post addresses Lesson #4.

NMFS’s lesson #4 states that the summer water temperatures in the 10-mile reach of the Sacramento River downstream of Keswick dam, the most heavily used reach for spawning by winter-run salmon, are not “correlated with flow.”  The lesson is important in that if generally true, high summer flow releases are not important in managing summer water temperatures for the salmon spawning and egg-embryo incubation that takes place close to Keswick Dam.  The magnitude of flow releases from the dam appears to have minimal effect on how much summer water temperatures increase in the upper 20 miles of river.  Rather, water temperature in this river reach is more a function of distance from the dam and the temperature of the water when released from Keswick.

Keswick is a small re-regulating reservoir that takes in water as it is released from Shasta Dam’s peaking power plant and from the Spring Creek Powerhouse, which generates power with water imported to the Sacramento from the Trinity River.  The water from the powerhouses is mixed in Keswick Reservoir as it enters the reservoir at times of day when power is most valuable.  “Re-regulation” means that Reclamation holds releases from Keswick to the Sacramento River downstream relatively constant over the course of the day.  The water temperature may be cold enough for salmon in the 20 miles of river immediately downstream of Keswick Dam, but water in the river warms quickly, depending on flow, as it moves further downstream through the remaining 200 miles of the lower Sacramento River.

1) Effect of Flow on the Water Temperature of Shasta/Keswick Releases

Water temperature immediately below Keswick in drier years 2014 and 2020 was most certainly related to flow (Figure 1), but not in the sense one might expect.  In 2014, high flow magnitudes (releases) caused a loss of access to the cold-water pool in Shasta Reservoir.  In 2020, high early summer releases led to reduced late summer releases to conserve the cold-water pool.  In both cases, spring-summer water deliveries to downstream water contractors were excessive, leading to limited access to Shasta’s cold-water pool by fall, resulting in high salmon egg-embryo mortality.

2) Effect of Flow Magnitudes into Keswick Reservoir on Temperature Increases within the Reservoir

There is some evidence that the temperature of water increases as it moves through and mixes within Keswick reservoir.  Keswick releases are often slightly warmer than upstream Shasta Dam releases into Keswick Reservoir (Figure 2), although the relationship is complicated by releases from Spring Creek Powerhouse of warmer water imported from the Trinity River system.  Again, as described above in point 1, the water temperature in Keswick Reservoir and water released from it in the late summer and fall of 2014 was primarily the result of gradual decline in the availability of cold water from the bottom of Shasta Reservoir.

3) Effect of Flow Magnitudes on Water Temperature in Sacramento River immediately below Keswick Dam

Higher flow magnitudes immediately downstream of Keswick Dam do not always translate into lower water temperatures, because the temperature of the source water in Shasta is more important (Figure 3).  Temperature increases a short distance from Keswick Dam are small (Figure 4).  Even in summer of wet-year 2019, water temperature increases over the 10-mile spawning reach were similar to those in 2020 (Figur e 5).  In 2019, flow magnitude had relatively little influence on water temperature even as far downstream as Balls Ferry.

It is worth noting that when dam release temperatures are at or above the upper limit of safe survival (53oF) as they were in summer 2020, then any temperature increase in the 10-mile spawning reach becomes a critical issue for the survival of the eggs and embryos of winter-run salmon.  However, the solution is not to increase flow to reduce warming within the 10-mile spawning reach, because that depletes the Shasta cold-water pool and trades a short-term benefit for the long-term impact of reducing the size and accessibility of the Shasta cold-water pool.  Rather, the solution is to keep releases low enough over the course of the spring and summer to allow those releases to maintain temperatures lower than the 53ºF threshold.

4) Effect of Flow Magnitudes on Water Temperature in Lower Reaches of the Sacramento River

Summer water temperature in the lower reaches of the Sacramento River is heavily influenced by magnitude of flow (Figures 6 and 7).  Lower flows promote higher water temperatures.  Downstream of Red Bluff, water temperature depends on air temperature, the magnitude of flow, and how fast the river flows.  Lower flow magnitudes reduce the speed with which water moves downstream.  Smaller thermal mass of water at low flows, combined with the slower rate with which water moves downstream, causes water to pick up more thermal energy and get warmer quicker.  In the lower reaches of the Sacramento River, water no longer depends on the temperature of the water released from Shasta and Keswick reservoirs.

In summary, water temperature in late spring through early fall in the Sacramento River immediately downstream of Keswick Dam is primarily determined by the water temperature of the water released from the dam, not the magnitude of flow.  In contrast, flow is the primary management tool available to reduce water temperature in the lower 200 miles of the Sacramento River from late spring through early fall.

Figure 1. Comparison of water temperature and flow from Keswick Dam in late summer and early fall between critical drought year 2014 and dry 2020.

Figure 2. Water temperature immediately below Shasta and Keswick dams, and flow rate below Keswick Dam in September 2014, a critically dry year when access to Shasta Reservoir’s cold-water-pool became limited. Blue line is flow downstream of Keswick. Orange line is Keswick release water temperature. Green line is Shasta release water temperature.

Figure 3. Comparison of 2019 and 2020 summer water temperature and flow below Keswick Dam in the upper Sacramento River. Note substantial loss of water temperature control in mid-September 2020, a consequence of source water temperature, not lower flows.

Figure 4. Water temperature in the 10 miles of spawning reach below Keswick Dam in late summer and early fall 2020. Water temperature over the 10 miles increased about 1.0 to 1.5oF. Blue line is Keswick release temperature. Orange line is water temperatures in the Sacramento River at Highway 44, 5 miles downstream Keswick Dam. Green line is the Clear Creek gage on the Sacramento River, 10 miles downstream of Keswick Dam.

Figure 5. Summer water temperature in the upper 20 miles of the Sacramento River below Keswick Dam, 2019. Note the increase is about 2-4 degrees over the 20 miles. CCR is the Sacramento River at Clear Creek. BSF is Sacramento River at Balls Ferry, 20 miles downstream of Keswick Dam. SAC is Sacramento River at the Highway 44 bridge. KWR is release temperature from Keswick Dam.

Figure 6.  Summer water temperature in the lower Sacramento River at Wilkins Slough (RM 120) 2015-2020.  Note the water quality standard is 68oF.

Figure 7. Summer river flow magnitude in the lower Sacramento River at Wilkins Slough (RM 120) 2015-2020.

Low Spring Flows Reduce Survival of Spring-Run Salmon – Lessons 14 and 15

Following an introductory post, this is the fifth post in a series on the lessons learned by the National Marine Fisheries Service (NMFS) from the 2013-2015 drought that devastated Sacramento River salmon populations.  This post addresses NMFS’s Lessons #14 and #15.

NMFS acknowledges that low spring flows may lead to low survival of juvenile spring-run Chinook salmon during their emigration to the sea.  NMFS also suggests that disease may also cause poor survival in dry years.

Some additional observations regarding outmigration are appropriate.  Most wild spring-run fry, fingerling, and pre-smolts emigrate in winter, although spring smolt emigration also occurs.  Another thing to consider is that while disease may be more prevalent in spring of dry years, it may be due to extended rearing in the poor habitat of the upper river in drier years (less food, warmer water, and more stress).  Increased predation by striped bass and other predators under these same habitat conditions is almost certainly another factor.  Lower flows make predators more effective because of lower turbidities, warmer and shallower water, less cover, and slower transport rates.

NMFS’s Lessons 14 and 15 did not consider lower survival of adult salmon migrating upstream in spring of drier, low flow, warmer water years.  However, this is also important.  Run counts are strongly related to water-year types, with poor runs in drier years (Figure 1).  Adults that migrate upstream in spring and over-summer in dry years face more difficult passage in spring and warmer spring and summer water temperatures.  Warmer water temperatures lead to greater stress, energy loss, and disease, poorer pre-spawn survival, and reduced reproductive success (Richter and Kolmes 2005).

An overall lesson is that low winter-spring flows lead to higher water temperatures in the lower Sacramento River migration corridor (Figures 2 and 3) and in spawning tributaries (Figure 4), which in turn reduce survival and reproductive success of spring-run salmon.

A related issue at play relates to Lesson 8 (low flows in the lower Sacramento River at the Wilkins Slough gage).  Low spring and summer flows in the lower mainstem Sacramento River flows lead to warmer spring and summer water temperatures.  This in turn leads to higher predation rates on juvenile salmon and poorer adult survival and reproduction success.  Continuing to ignore long-ago established water quality standards for water temperature and operating norms for flow is putting greater stress on all the salmon runs and on sturgeon.  Spring-run salmon are especially vulnerable because both juveniles and adults are present in the lower Sacramento River in spring.

One especially damaging case occurred in early spring 2018.  A series of storms raised flows in the lower Sacramento River by approximately 50,000 cfs, while at the same time there was only a minimum release of about 3000 cfs into the upper Sacramento River from Shasta Dam.  Substantial flows entered the lower Sacramento River from tributaries and from the outfalls of agricultural basins, including the Butte Basin at the Butte Slough Outfall gates.  Some flow even left the river via overflow weirs into the Sutter Bypass. The sudden surge of urban and agricultural basin stormwater, with high oxygen demand, sediment, and chemical loads, led to a series of fish kills of adult spring-run salmon in the lower Sacramento River at the Butte Slough Outfall (Figures 5 and 6).  It is not clear exactly why high concentrations of stormwater may have increased mortality of spring-run salmon in 2018.  What is clear is that low Shasta releases contribute to higher concentrations of stormwater in river flows at key times during their migration.

In conclusion, Reclamation should consider higher spring releases from Shasta in drier years like 2021 to (1) improve juvenile spring-run emigration survival to the Bay-Delta, and (2) improve adult spring-run survival and reproductive success.  Reclamation should also consider spring pulse flow releases from Shasta-Keswick to emulate natural unimpaired lower river spring flows and to enhance migration success.  Such pulses should be timed with lower river tributary flow pulses.  Finally, as discussed in previous posts, Reclamation’s contractors and others must reduce dry-year water diversions in spring from the mainstem Sacramento and its tributaries to further protect spring-run salmon.

Figure 1. Relationship between Mill and Deer Creek spring run salmon counts 1963-2019. Blue dots represent above-normal and wet years. Yellow dots represent below-normal and dry years. Red dots represent critical dry years. Red margin on blue dots represents a wet and above normal year for adult immigration with a critical dry year two years prior during rearing and emigration. Of special note are the sharply higher run sizes in wet years in Deer Creek, but lower run sizes in Deer Creek in drier years. This shows that production of salmon in Deer Creek suffers proportionally more than in Mill Creek in drier years. In both creeks, salmon runs suffer in years when they are emigrating and immigrating in dry conditions.

Figure 2. River flows (1000s of cfs) in the lower Sacramento River at Wilkins Slough (below the mouths of Deer and Mill Creeks) in winter-spring from 1995-2021. Note very wet conditions in 2006 and very dry conditions in 2014 and 2015.

Figure 3. Water temperatures in the lower Sacramento River at Wilkins Slough (below the mouths of Deer and Mill Creeks) in winter-spring from 2013-2021 (only data available). Note the extraordinarily warm water (near or above 70ºF) in early spring of drier 2014, 2015, 2018, and 2020. State standard is 68ºF. Migrating adult salmon suffer stress and mortality at water temperatures above 60ºF.

Figure 4. River flow and water temperature in Deer Creek 2011-2921. Wet year spring in 2011, 2017 and 2019 shown in blue highlight: other spring conditions shown in yellow highlight. Dry years include 2013-2015 and 2021. The remainder are normal or average year types. Note lower water temperatures in wetter years, with highly stressful temperatures (>68ºF, 20ºC) delayed further into spring than during drier years.

Figure 5. Dead adult spring-run salmon floating at the mouth of Butte Slough Outfall in the Sacramento River in March 2018. Fish likely died from severe stress from low oxygen and heavy suspended sediment and chemical load after being attracted to stormwater flow from outfall gates.

Figure 6. Dead adult spring-run salmon floating at the mouth of Butte Slough Outfall in the Sacramento River in March 2018. Fish likely died from severe stress from low oxygen and heavy suspended sediment and chemical load after being attracted to stormwater flow from outfall gates.
Source: https://youtu.be/t5oB_hWcrzU

Use of Shasta Dam TCD side gates – Lesson #3

Following an introductory post, this is the fourth post in a series on the lessons learned by the National Marine Fisheries Service (NMFS) from the 2013-2015 drought that devastated Sacramento River salmon populations.  This post addresses Lesson #3.

The side gates at the bottom of the Temperature Control Device (TCD) on the inside face of Shasta Dam (Figure 1) allow deeper, colder water in the reservoir to be drawn into the power plant intake penstocks and released to the river below. Use of the side gates allows more colder water to be released for salmon in the river below in summer and fall in years when reservoir levels are low and the cold-water-pool is limited.

In 2014 and 2015, NMFS and the Bureau of Reclamation learned that when the reservoir level is low and the side gates are opened to access cold water, some warmer surface water is also drawn downward into the side gate openings.1 The entrance of some warm water through the side gates compromises temperatures downstream of the dam, thus reducing the effectiveness of the TCD system. Warmer-than-expected release temperatures also limit the ability of Reclamation to meet water demands for downstream contractors when reservoir levels are low. Delaying side gate use with the present structure requires maintaining higher summer storage levels with less summer deliveries.

A comparison of Shasta conditions in mid-August 2014 and mid-August 2015 provides a good example of the problems (Figures 2 and 3). Reclamation used the side gates in August 2014 but not in August 2015. In 2014, Reclamation released more irrigation water in summer, in anticipation of being able to use the side gates (Figure 4). In 2015, lower summer releases conserved storage and cold-water-pool volume, thus delaying use of the side gates. The 2014 operations led to the complete loss of access to the cold-water-pool by October (Figure 5). Water temperatures in 2015 were still too high, but complete loss of control did not occur.

There are other operational changes that may help. Most of the water discharged from the TCD passes through penstocks to powerhouses near the base of Shasta Dam. Those powerhouses are operated based on power demand and prices, not based on temperature management (lower graph of Figure 6). As power demand and power values increase, Reclamation “peaks” the powerhouses to follow these increases. However, the amount of water going through the powerhouses seems to affect the layer of water in Shasta Reservoir that the TCD draws from. and thus the temperature of the dam release water (Figure 6). Low intake rates during non-peak times appear to draw a greater proportion of warmer surface waters from Shasta Reservoir.

It is likely that a more constant rate of release from Shasta Reservoir, rather than peaking, would maintain an overall lower temperature of water that leaves Shasta Reservoir and enters Keswick Reservoir immediately downstream of Shasta. It is also likely that a constant flow of water through Keswick would allow less mixing of Keswick’s warmer surface water with the cooler release from Shasta. This is because water from Shasta would spend less time in Keswick and because a constant release from Shasta would help the cooler water maintain an “underflow” of cooler water at the bottom of Keswick.

Similarly, operation of the powerhouses at Trinity Reservoir and Whiskeytown Reservoir affect the water temperature of water exported from the Trinity River to the Sacramento River system. Most of the water exported from the Trinity to the Sacramento passes through Spring Creek Powerhouse, located on the west side of Keswick Reservoir. All of these powerhouses, too, are operated for power first, not for water temperature. Reduced peaking at these facilities, or simply optimization for water temperature, would help improve water temperatures in the Sacramento River downstream of Keswick Reservoir.

In summary, higher storage levels and greater cold-water-pool volume, delayed use of side gates, changing side gate openings, fixing leaks, and possible changes to peaking power operations may be necessary to protect salmon in the Sacramento River below Shasta Dam.

Figure 1. Shasta Dam Temperature Control Device configuration. Source: Reclamation

Figure 2. Shasta reservoir water temperature profile and TCD operation in August 2014. Note side gate use (curved arrows).

Figure 3. Shasta reservoir water temperature profile and TCD operation in August 2015. Note water level and cold water pool elevations were slightly higher in 2015 than 2014 (Figure 2).

Figure 4. Water temperature and volume of releases to the lower Sacramento River Aug-Oct 2014 and 2015. Note loss of Shasta cold-water-pool access in September 2014. The higher storage releases in August in 2014 compared to 2015 made up much of the storage level differences between 2014 and 2015 shown in figures 2 and 3.

Figure 5. Water temperature profile in Shasta Reservoir and outlet tower Temperature Control Device configuration in October 2014.


Figure 6. Hourly water temperature and release rate from Shasta Dam 10/7-10/12 2014. Note highest water temperatures were during lower non-peaking-power release periods. See Figure 5 for reservoir water temperature profile and TCD operation on 10/8/2014.

  1. TCD leaks at higher elevations were also detected.