Response to Bay-Delta Science Paper on The Problems with Coleman Hatchery Salmon Straying

In a Maven 1/17/19 post, US Fish and Wildlife Service hatchery managers discuss problems that have developed from trucking salmon smolts from the Coleman National Fish Hatchery (Coleman NFH) to San Francisco Bay. Chief among these problems is that juvenile fish trucked from Coleman NFH stray into other Central Valley rivers as spawning adults, rather than returning to Coleman NFH. This has become a general argument for releasing Coleman NFH smolts near Redding instead of trucking or barging them to the Bay.

During the 2013-2015 drought, many Coleman NFH smolts were trucked to the Bay for release to meet the hatchery’s goals:

The management goals for the hatchery are to provide about 1% of their production or about 120,000 fish from each brood year to the ocean fishery, the freshwater sport fishery, and returns to Battle Creek for brood stock, all while trying to reduce negative impacts on naturally spawning fish.

From 2013-2015, 35.5 million fall run smolts, about 12 million per year, were reared for release by the Coleman NFH. In 2013, all were released at the hatchery into Battle Creek. In 2014, 3.4 million were released at the hatchery, and the remainder were trucked to the Bay. In 2015, all 12 million were trucked to the Bay. Since 2015, all Coleman NFH fall-run smolts have been released at the hatchery.

Trucking during the drought did meet the 1% production goal, but did not provide adequate hatchery returns for broodstock and did not achieve low straying rates.1 The USFWS has since chosen to ignore the production goal in order to achieve the other two goals. In dry years, the 12 million smolts now provide 10,000-20,000 adults (0.1-0.2%) to fisheries instead of 100,000-200,000 (1-2%). Instead of increasing hatchery production by barging smolts or improving post-release river conditions (higher flows), the fall-run salmon production program at Coleman NFH, designed explicitly as mitigation for loss of salmon from the federal Central Valley Project, is not meeting its primary mitigation goal.

In the response to Maven’s questions on USFWS presentation at the Bay-Delta Science Conference, the USFWS managers offered the following responses:

Is Straying Bad?

  • This occurs because fish that do not naturally emigrate through the river system lose the opportunity to imprint properly and may be unable to locate their natal area when they are ready to spawn. Response: High stray rates are unique to the federal Battle Creek Coleman NFH, partially because many of their returning adults spawn in the upper Sacramento River (as intended) and do not move into Battle Creek, and are thus considered strays. Regardless of imprinting influence, returning adults to the Sacramento River are confronted with poor flows and high water temperatures above the mouth of the Feather-Yuba-Butte Creek, Mokelumne, and American rivers, so they choose those rivers out of necessity. This problem does not occur at the Feather, American, and Mokelumne state hatcheries. These other hatcheries have also transported stray, earlier-returning Battle Creek adults (or fertilized eggs) to the Coleman Hatchery to help Coleman offset its egg shortages.
  • Hatchery fish that stray into natural spawning areas can detrimentally affect natural fish populations through genetic, ecological, and behavioral mechanisms…. Increased straying by hatchery fish may reduce genetic diversity within and among salmon populations. Response: While this is generally true for pristine natural salmon watersheds, this concern was long ago lost in the Central Valley after a century of disturbance and hatchery interference. Eggs have been shared, even from out of state, and even from different runs (hatcheries and dams have forced mixing of spring-run and fall-run spawners). Hatchery strays continue to dominate runs in non-hatchery rivers; without the strays, there would be few if any salmon in these rivers. Straying is also a natural process needed to retain population fitness, resilience, and genetic diversity.
  • Likely due to the prevalence of off-site release practices at Central Valley hatcheries, Central Valley fall Chinook Salmon have lost locally adapted genes and become one large, genetically homogeneous population (see Johnson et al 2012). Response: Adapted genes were lost decades ago before off-site releases. Johnson et al were describing how hatchery dominance covers up underlying problems in wild populations, not necessarily how it causes the underlying problems. This is a real concern in the Central Valley where only 25% of the fall run hatchery smolts are marked. The true wild components of individual river runs need protection. For some rivers, natural spawning areas can be reserved for true wild spawners (e.g., Battle Creek, Yuba River, Mokelumne, Calaveras, Butte Creek, etc.). Wild spawners can also be relocated above rim dams. Conservation hatcheries can use “wild” spawners.
  • Furthermore, salmon that are spawned and reared in fish hatcheries may become quickly adapted for characteristics that favor their survival and reproductive success in the hatchery environment while, at the same time, diminishing their ability to survive and reproduce in the natural environment. Response: while potentially (and historically) true, modern hatchery programs minimize or should minimize, if not reverse, these tendencies.
  • Stray hatchery fish may also compete with natural fish populations for limited resources, including competition for preferred spawning areas by adults and competition for food resources and habitat by juvenile fish. Response: Again, while generally and historically true, hatchery adult returns now take the place of wild adults in increasingly depressed amounts of spawning habitat in Central Valley rivers. Also, it is generally known that in many cases wild adults out-compete hatchery adults for habitat. Also, in most Central Valley rivers, hatchery fall-run adults can be effectively excluded from primary spawning areas. Hatchery smolts can readily be excluded from natural rearing areas by controlling timing and location of releases. In arguing against barging or trucking, USFWS managers are in essence arguing to continue releasing their hatchery smolts into prime natural rearing areas.
  • When hatchery fall Chinook Salmon from Coleman NFH stray at high rates, the freshwater fishery may be redistributed to other parts of the Central Valley and sport fishing in the upper Sacramento River fishery is negatively impacted. Response: Low flows and high water temperatures are the biggest contributor to poor upper Sacramento River salmon fishing.2 Also, Coleman hatchery adults that spawn in the upper Sacramento River or that are caught by sport fisherman in the Sacramento River upstream of Battle Creek are considered strays, even though they are produced to mitigate for Shasta Dam.
  • High stray rates of trucked hatchery fish may impact the ability of Coleman NFH to meet its annual fish production targets. For example, in 2017 (which corresponded to age-3 adult returns of fish that were 100% trucked), less than 350 adult salmon returned to the hatchery. At a minimum, Coleman NFH needs to spawn 2,600 pairs of adult salmon to meet the production target of 12 million juveniles. As a result, Coleman NFH released less than 6 million juveniles in the spring of 2018, which will result in fewer fish available to achieve our management goals in future years. Response: Then why not release a portion of the smolts in the Bay as for brood year 2013 (2014 releases)? Other potential measures to increase egg take outlined by the authors can also be enhanced.

Need for Trucking

  • California’s drought had some negative impacts for Coleman National Fish Hatchery….The one percent contribution goal to the fisheries and returns to Battle Creek was not able to be achieved under the drought conditions. Response: The drought was not to blame. Reclamation’s water right permits require measures to protect salmon, even during droughts. During the 2013-2015 drought, Reclamation delivered too much water to Sacramento River settlement contractors. This left Reclamation too little water to meet the conditions in its water rights permits that are designed to protect salmon. Water management during droughts is the problem.
  • Despite our extra efforts to collect additional eggs, we only met about half of our production target. We’ve undertaken a study to assess fish survival under different release strategies. This information will hopefully allow for adaptive management, based on changing environmental conditions and may give us some flexibility to release fish onsite to meet our multiple management goals in the future. Response: If Reclamation provides environmental flows from Shasta, and if USFWS times releases of hatchery juveniles at Coleman NFH to water conditions, then Battle Creek releases may produce more adult salmon. Some amount of trucking may be necessary to meet overall mitigation goals.


Hatchery program managers should not give up on trucking and on reducing high stray rates. They should consider barging juvenile salmon to better imprint smolts and improve smolt survival. Federal water managers also need to improve lower Sacramento River conditions for returning adults, in order to reduce straying. Water contractors and state/federal purveyors must also collectively limit their water diversions to meet water right permit conditions and water quality standards that are designed to protect fish. Egg taking should be increased if hatchery releases are to continue, in order to meet mitigation goals. These measures will all help hatchery managers meet their adult production goals and mitigation commitments. Trucking, a standard practice at state hatcheries in all year types, may be necessary, especially in periods of drought, to meet mitigation goals.

Washington State increasing hatchery salmon production Major realignment toward wild population recovery

Washington State is increasing salmon production to help in recovery of near extinct salmon populations and orca (killer whales). A Washington state Department of Fish and Wildlife webpage describes the role of hatcheries in restoring wild salmon stocks:

In recent years, state hatcheries also have taken on an equally important role in helping to recover and conserve the state’s naturally-spawning salmon populations. Nearly all the hatcheries in the Columbia River and a number of hatcheries in Puget Sound play a role in wild fish rebuilding programs, whether by rearing juveniles prior to release or holding fish through their lifespan to ensure the survival of depressed stocks. This renewed focus on wild stock recovery represents a major realignment in hatchery operations, as WDFW, the tribes, federal government and independent scientists worked to develop a comprehensive operations strategy for hatcheries in Washington. 1

The main reform actions being taken in Washington’s program are:

  1. Marking all hatchery salmon smolts allows identifying hatchery fish by hatchery and lot group. It also allows mark-selective fisheries that require release of wild salmon.
  2. Developing salmon population-specific recommendations intended to provide scientific guidance for managing each hatchery more effectively in the future.
  3. Keeping hatchery program budgets in pace with increasing operating costs (especially utilities, fish feed and labor costs), and not forcing cutbacks in some programs.
  4. Updating aging hatchery infrastructure.
  5. Supplementing wild stocks: to maximize egg fertilization and fry survival (of wild) and thereby increase the number of “wild-type” smolts heading out to the ocean.
  6. Maintaining captive broodstocks of endangered stocks with dangerously low population levels: juveniles are maintained in a hatchery for their entire life to ensure the stock’s survival.
  7. Minimizing interaction between naturally-produced and hatchery-produced outmigrating juveniles and adult fish returning to streams to spawn where necessary.
  8. Donating surplus adult salmon from hatcheries to non-profit hunger-relief programs.
  9. Introducing hatchery carcasses or analogs back to streams to increase natural productivity to low productivity streams.

While California is slowly moving toward some of these reforms, there is resistance to others. For more information and links on California hatchery programs see , , , .


Late-Fall-Run Salmon Update Record low runs in fall 2015 and 2016

In an April 2017 post, I related factors likely important to late-fall-run Chinook salmon in the reach of the Sacramento River downstream of Shasta Reservoir. In this post, I update that assessment with 2015 and 2016 escapement estimates and coded-wire-tag return data from Coleman Hatchery smolt releases from brood years 2008-2013.

Late-fall-run Chinook salmon escapement reached new lows in 2015 and 2016 (Figure 1). The adult returns in these two years were the product of spawning in brood years 2012 and 2013 and of early rearing conditions in winter of 2013 and 2014 (critical drought years).

Figure 1. Late-fall-run escapement 2000-2016. Escapement refers to adult run counts beginning in late fall of spawning year. For example, 2016 represents late-fall-run for water year 2017 (Oct16-Sep17). Source: CDFW Grandtab.

The low escapement in 2015 and 2016 is also reflected in the spawner-recruit relationship (Figure 2).  There is a continuing significant positive spawner-to-recruit relationship and even stronger effect of water-year type, with poorer recruitment from dry-year winter rearing conditions.

The low 2016 escapement is likely in part a consequence of a very poor return from brood year 2013 hatchery smolts (Figure 3).  Of the approximately one million smolts tagged and released in 2014 at the Coleman hatchery near Redding, less than a tenth of a percent survived to be counted in fisheries and escapement surveys.  A good survival rate would be 1 to 3 percent, as occurred for brood year 2010 (2013 run).

Hatchery and wild smolts from brood year 2013 had poor flow conditions in early winter of 2014 (Figure 4), while brood year 2010 had the best flow conditions.  There were no flow pulses to help the smolts move the 200 miles down to the Bay-Delta in January 2014.

The fact that few late-fall smolts showed up in south Delta salvage in 2014 (Figure 5), compared to higher salvage in 2013 (Figure 6), is compelling evidence that smolt survival to the Delta was very poor in 2014.

Conclusion:  To sustain the late–fall-run salmon population, higher winter flows and flow pulses are warranted in the lower Sacramento River in drier years.

Figure 2. Spawner-recruit relationship for late-fall–run salmon. Number is log10-1.5 transformed escapement (recruits) for the fall of that year. For example, year “16” represents escapement for late fall 2016, which includes spawners from early winter 2017. Spawners represent escapement from three years earlier (brood year). In the example, spawners for year 16 were the progeny of escapement in 2013. Colors represent winter rearing condition two years earlier. In the example, red “16” represents dry winter 2014. Green represents normal years two years earlier. Blue represents wet years two years earlier.

Figure 3. Hatchery smolt survival for brood years 2008-2013 based on coded-wire-tag returns.

Figure 4. January river flow at Wilkins Slough in lower Sacramento River 2009-2014. Note very low flows and lack of flow pulses in January 2014.

Figure 5. Salvage of salmon at the Delta pumps in water year 2014. Note only one late-fall tagged smolt (yellow dot), collected in March after January release.

Figure 6. Salvage of salmon at Delta pumps in water year 2013. Note many late-fall tagged smolts salvaged (yellow dots) beginning in December, coincident with flow pulses after December and January releases of hatchery smolts.


Fall Pulsed Flow Protections for Winter Run Salmon Not in 2017 and 2018

It is well known that juvenile winter-run salmon migrate downstream to the Delta from their upper Sacramento River rearing area near Redding/Red Bluff with the first fall-winter flow pulses.1 Protection of this critical migration of winter-run has been recommended in recent Delta proceedings. In the WaterFix petitions process at the State Water Resources Control Board, such protection is referred to as “pulse protection.” So based on the importance of protecting winter-run during this key life-history stage, one would think such protection would have been applied in the fall of our most recent wet year 2017 and below-normal 2018. A quick check of the facts indicates otherwise.

In 2017, the initial pulse of juvenile winter-run salmon passed Red Bluff (rivermile 240) during September and October, prior to any flow pulses (Figure 1). The first major pulses of flow occurred in the latter half of November 2017. A migration spike of winter-run-sized smolts at Knights Landing screw traps (rivermile 90) did occur during the flow pulse (Figures 1 and 2). Instead of protecting these winter run as they were entering the Delta, south Delta exports were increased to take advantage of the available inflow (Figure 3). The increase occurred with the Delta Cross Channel partially open (Figure 4), increasing the risks of migrating salmon to south Delta exports. Other than a four-day pulse, Delta outflow generally declined before and after the flow pulse (Figure 5). Salinity in the lower Sacramento and San Joaquin Delta channel generally increased with the higher exports and lower outflows (Figures 6 and 7).

In 2018, the initial pulse of juvenile winter-run passed Red Bluff in September and October prior to flow pulses (Figure 7). The first major pulse of flow occurred at the end of November. A migration spike of winter-run-sized smolts at Knights Landing screw traps occurred during this flow pulse (Figures 7 and 8). Instead of protecting these juvenile winter run as they were entering the Delta, south Delta exports were increased to take advantage of the available inflow, with nearly two-thirds of the Delta inflow being exported (Figure 9).

The lack of protection is likely a consequence of the young salmon not showing up in salvage (Figure 2). For me, the high risk factors are clear.

  1. The lack of salvage simply indicates that few of the migrating salmon survive upon entering the Delta under high exports. High late-November and early-December south Delta exports are a high risk to the juvenile winter-run survival. Traditional high December exports of up to 65% of Delta inflow allowed under D1641 water quality standards are a major risk factor for winter-run salmon that must be changed.
  2. Declining flows below Keswick Dam (rivermile 300) during the fall, combined with the lack of flow pulses, are also a real concern. Juvenile salmon that pass Red Bluff (river mile 240) in September and October are sustained in low flow conditions for nearly 200 miles of modified river channel until the first rains in late November or early December. And the spawning reach directly downstream of Keswick gets no added flow even when it rains, because inflows to Shasta Reservoir are all captured.

For additional information, see agency discussions and data presentations in:

Figure 1. Juvenile winter-run salmon estimated passage and stream flow at Red Bluff (rivermile 240) from late summer 2017 to June 2018.

Figure 2. Lower Sacramento River screw trap collections at Knights Landing (rm 90) since August 2017 along with flow, turbidity, and water temperature. Note late November catch event during flow pulse.

Figure 3. South Delta exports in fall 2017. Note peak exports mid-November through early December – total 10,000-11,000 cfs (22,000 acre-ft per day). Chinook salvage is shown as zero.

Figure 4. Flow through Delta Cross Channel August through December 2017. Note sporadic opening late September through November 2017, and closure at end of November.

Figure 5. Delta outflow August through December 2017.

Figure 6. Salinity (EC) near Rio Vista on lower Sacramento River channel in the Delta, August through December 2017.

Figure 7. Salinity (EC) near Jersey Pt on lower San Joaquin River channel in the Delta, August through December 2017.

Figure 8. Juvenile winter-run salmon estimated passage and stream flow at Red Bluff (rm 240) from late summer 2017 to June 2018.

Figure 9. Juvenile winter-run salmon estimated passage at Knights Landing (rivermile 90) and stream flow at Wilkins Slough (rivermile 125) from late summer to fall 2018. Note early December pulse of fish, flow, and turbidity.

Figure 10. Export rates from two south Delta pumping plants (Tracy TRP and Banks HRO) summer and fall 2018.

Figure 11. Lower Sacramento River flow at Freeport (FPT, rm 55) and Delta outflow (DTO) summer and fall 2018.

Mokelumne River Salmon Update – Fall 2018

Dan Bacher’s report on 2018 salmon returns to the Mokelumne River Hatchery provides continuing good news for the Mokelumne River salmon run.

When I last reported on the fall run of Mokelumne River salmon, I described strong runs from 2010-2016. The 2017 run was exceptionally strong (Figures 1 and 2), despite being the product of the 2013-2015 drought. This success story is in large part due to the management of the Mokelumne River Hatchery, as reported in the Bacher article. Trucking hatchery smolts to the lower San Joaquin in the western Delta near Jersey Point greatly improves survival over releases of salmon in the lower Mokelumne River.

Best returns from the spring 2015 releases were from smolts barged and released near the Golden Gate. That group of 101,000 smolts (tag code #060593) had a good return of 1.35 percent compared to 0.06-0.80 % from smolts released in the west Delta and 0.01 % released in the lower Mokelumne River. In early May 2015, soon after river and Delta smolt releases, water temperatures reached 68oF in the west Delta and 72oF in the lower Mokelumne in the central Delta, levels stressful to smolt salmon. Delta outflow in early May 2015 was critically low at ~5000 cfs. Net flow in the lower San Joaquin River in the west Delta, with the Delta Cross Channel closed, was near zero.

The returns from tag group #060593 (shown in Figure 3), which represented only about 3% of the total smolts releases from the Mokelumne Hatchery in 2015, exemplifies the kind of return pattern that can be achieved by a progressive hatchery program. The majority of the adults returning from this tag group to the Central Valley were to the hatchery (192), although a significant number strayed to the neighboring American River (82) and San Joaquin River and its tributaries (40). In previous years, the California Department of Fish and Wildlife achieved similar success barging smolts from the Feather River Fish Hatchery.1

Figure 1. Mokelumne River in-river fall Chinook returns (escapement) 1975-2017.

Figure 2. Mokelumne River hatchery fall Chinook returns (escapement) 1975-2017

Figure 3. Mokelumne River fall Chinook tag returns through 2017 from 101,000 Mokelumne River Hatchery smolts, tag group #060593, barged to and released near the Golden Gate Bridge.