When I last assessed the status of the fall-run salmon population of the Yuba River near Marysville in a 1/31/22 post, I stated: “The population remains in a very poor state – at about 10% of recent historical levels during and subsequent to multiyear droughts 2007-2009 and 2013-2015 (Figure 1).” Since the record low run in 2017, the fall run on the Yuba River has not recovered.

The failure of the four more recent runs to show signs of recovery (Figure 1) is especially concerning because 2017 and 2019 were wet years. The failure to recover may be simply the lingering effects of the drought years 2014-2015 and the ongoing effects of the 2021-2022 drought. More likely, the spawning stock has collapsed and is in dire need of support. The 2022 run appears to be even worse than the past four runs,¹ thus adding to the concern.

This post delves into the many possible causes of, or contributors to, the collapse of the Yuba River fall-run salmon population. The story is a complicated one. It starts with broodyear 2014.

Figure 1. Yuba River fall-run salmon spawning escapement estimates 1953-2021. (Data source: GrandTab)

The first stop in pursuit of the potential causes of the recruitment failure that has occurred not only on the Yuba River, but in most of the other Central Valley fall-run salmon populations, is a close look at the escapement data.  The spawner-recruitment relationship (S/R) shown in the escapement data (Figure 2) provides a closer perspective than the simple histogram of the run sizes (Figure 1).

The S/R figure is a plot of the log of the escapement with the log of the escapement three years earlier. This is because about 80-90% of spawners are three years old.  The three red lines in Figure 2 show that adult spawners in 2014 produced the spawners 2017, which in turn produced the spawners in 2020.  The adult spawners in 2018 produced the spawners in 2021.  Spawners in 2019 produced spawners in 2022, which based on the incidental reports will likely show up to the lower left of 19.  The lower-left quadrant of an S/R plot is usually a place where a salmon population is headed toward collapse and an inability to sustain itself.

Figure 2. Yuba River fall-run salmon spawner-recruit relationship (1978-2021) with recruit number shown in chart for specific years. Red lines point from spawner to recruit year. For example, recruits in 2017 led to recruits in 2020. Recruits in 2014 (12,000) led to only 1500 in 2017.

When Yuba River escapement (recruitment) is adjusted for strays from other rivers, the recruitment level in record low 2017 (Figure 3) shows itself to be even more dire.  The Yuba River receives many strays because it carries a strong cold-water signal into the Feather River and on into the lower Sacramento River in late summer and early fall.  The Yuba River also attracts spring-run and late-fall-run hatchery salmon that are included in the Yuba River’s fall-run spawning counts.

It is helpful to start the analysis of the 2017 population crash by reviewing the early life cycle of broodyear 2014 – as eggs in their mothers.  Their parental stock, broodyear 2011, had been reasonably normal, if not in the range that might be considered the Maximum Sustained Yield 10,000-15,000 (Figures 1 and 2).  The strong numbers of broodyear 2011 spawners (and their broodyear 2014 eggs) arrived in the Bay in summer 2014.  The questions become what happened to:

1. those broodyear 2011 adult females;
2. their broodyear 2014 eggs and their hatchlings in summer-fall 2014;
3. the surviving broodyear 2014 fry in winter 2015; and smolts in spring 2015;
4. the yearlings, two-year-olds and three-year-olds in the ocean; and finally
5. the adults making up the 2017 run counted in the Yuba River spawning grounds.

The answer is that survival conditions were not good for all five categories above.  Each question is addressed below.

Figure 3. Yuba River escapement numbers. Source: PFMC 22, p.49.

1. The first question addresses the conditions that faced broodyear 2014 eggs when they entered the Golden Gate inside their broodyear 2011 mothers that fateful summer of 2014. Water-year 2014 was a critical drought year, during which the State Board weakened Delta water quality standards for the year.

• • Unusually warm water met the salmon when they entered the Bay in summer of drought year 2014 (Figure 4). By the time they reached the mouth of the Feather River at Verona (if they got that far), water temperatures were near the lethal 75º F level through September (Figure 5).  Elevated water temperatures occurred through the entire route from the Golden Gate to the Yuba River.
  • Once on the spawning grounds of the, Yuba the parents of broodyear 2014 eggs encountered drought-year low flows (Figure 6), which in addition to being warm provided minimal available spawning habitat quantity and quality.
  • By the time the parents were ready to spawn in early fall, they were likely compromised by disease and thiamine deficiency, limiting the viability and survival of the broodyear 2014 eggs, and thereby the potential reproductive success of broodyear 2014 and its contribution toward 2017 recruitment.

2. The second question addresses the subsequent fate of surviving broodyear 2014 eggs and the hatched alevins in gravel redds. The eggs and alevins in the spawning beds faced unusual stresses in the form of erratic flow and very low flows (Figure 7).  Eggs spawned in October-November were subjected to scouring flows in December.  Eggs spawned in December under the high flows were subsequently subject to dewatering in January.

Figure 4. Water temperature in San Francisco Bay spring-summer 2014.

Figure 5. Water temperature in Sacramento River below mouth of the Feather River at Verona gage July-October 2014.

Figure 6. Streamflow in Yuba River at Smartsville and Marysville gages July-October 2014.

Figure 7. Yuba River streamflow in water year 2015 and 53-year average at Marysville gage.

3. The third question addresses conditions in the late fall through spring in critical drought water year 2015. Emergent fry likely benefitted from the February flow pulse that facilitated some fry movement out of the Yuba toward the Delta (see Figure 7).  Those fry that did not move were then subjected to extremely low flows (and stressful water temperatures) through the spring in the lower Yuba River and the lower Sacramento River below the mouth of the Feather River (Figure 8).  Delta and then Bay conditions were at their worst for young Yuba salmon on their way to the ocean in the spring of drought year 2015, made worse by the State Board’s continued weakening of water quality standards.

Figure 8. Streamflow and water temperature in the lower Sacramento River below the mouth of the Feather River at Verona gage in winter-spring 2015.

4. The fourth question regarding broodyear 2014 addresses growth in the ocean from 2015 through early summer 2017. In the ocean, they were subjected to strong fishery pressure (Figure 9) in all three years by the commercial and sport fisheries.  This provokes a series of questions.  Why did the Pacific Fisheries Management Council or PFMC allow those 50%+ harvests after the 2008-2009 collapse and fishery closures, and lack of subsequent population recovery?  Had the fall-run salmon populations really recovered sufficiently to sustain 50%+ harvests?  How accurate were those harvest rate estimates?  How hard were the stocks being preyed upon by seals and orcas?  Were the salmon whose diet had largely consisted only of anchovies becoming thiamine deficient by the time they spawned in the Yuba in early fall 2017.  In considering all these questions, one can only conclude that the summer upstream migration of fall-run salmon to the Yuba in 2017 had been highly compromised before it started.

5. The fifth and final question regarding broodyear 2014 salmon addresses conditions adult fish faced when they re-entered fresh to spawn after having been subjected to high harvest rates in the ocean from 2015-2017 (implied in Figure 9).  Upon returning to the Bay in summer 2017, a wet year, they encountered much better conditions during their upstream migration and spawning period.  After a final tweak by the summer river fishery, they spawned in the Yuba River in record low numbers.

Nearly identical circumstances and outcomes occurred with broodyear 2015 in 2018 (see Figure 3).  Broodyears 2016-2019 were subject to similar stresses.  Broodyears 2020 and 2021 were subject in-river to critical drought years 2021 and 2022.

In conclusion, it appears that the damage to broodyear 2014 and broodyear 2015 had been done for the most part by the time they returned as adults to the Bay in 2017 and 2018.  The record-low numbers of spawners estimated from the carcass surveys in 2017 and 2018 (Figure 1 and 2) were the cumulative effect of a series of survival factors, beginning with stresses on their parents in drought years 2014 and 2015, and ending with high harvest rates in the ocean and rivers in 2017 and 2018.  Management decisions by the State Board and PFMC, with acquiescence by federal and state resources agencies, contributed to this fateful series of events.  The events and their consequences were predictable, and the State Board and PFMC should have anticipated them and taken appropriate measures at the time.

It appears the same mistakes were made in regard to broodyears 2016-2021.  The effects of drought in years 2021 and 2022 will likely contribute further to the crash of the Yuba River salmon population, with even lower Yuba River and Central Valley salmon escapement in 2023-2025.

For more on the problems faced by Yuba River fall-run salmon and what can be done about them, see this October 2018 post.

Figure 9. Sacramento River fall-run salmon index 1983-2019. The 122 on y-axis is the target starting population level (122,000) under which harvest is allowed. Note the fisheries were closed in 2008 and 2009. Source: PFMC.

Following some improvement in the numbers of adult fall-run Chinook salmon returning to spawn in the Stanislaus River and the Merced River from 2012-2017, overall escapement in 2020 and 2021 to San Joaquin River tributaries was severely depressed.  Better flows and water temperatures could help reverse this decline.

In February 2017, I wrote about the fall Chinook salmon runs on the San Joaquin River’s three major tributaries over the previous six years.  Salmon counts in San Joaquin tributaries showed an increase in returning adults in the 2012-2015 drought compared to the poor returns in 2007-2009 drought (see Figures 1 and 2).   The numbers of spawners in 2012-2015 were still well below the returns in the eighties and nineties that corresponded to wet water year sequences, but the increase seemed to suggest progress.

In a December 2019 update, I  updated the earlier post with numbers from the 2016-2018 runs. The 2016 and 2017 runs were the product of poor rearing conditions in 2014 and 2015, both critical drought years, but with good fall adult migration conditions.  The 2018 run was a product of normal-water-year rearing (2016), but poor adult migrating conditions.  The 2016 and 2017 runs were strong in the Stanislaus and Merced rivers (see Figure 3), with both rivers benefitting from hatchery production and strays.  The markedly smaller runs in the Tuolumne (typical throughout the last decade) also benefitted from hatchery strays (Figure 4).  One strong component of the strays was the unusually high proportion of strays from the upper Sacramento River’s Battle Creek hatchery, whose managers’ strategy during the 2014-2015 drought was to truck their fall-run salmon smolts to the Bay, a practice that causes high straying rates, including to the San Joaquin tributary runs.

The 2018 San Joaquin run was lower, but still an improvement over the drought-influenced runs in 2007-2011 (Figure 2).  Spring rearing conditions in 2016 and fall adult migration conditions in 2018 were generally better than they were during the critical drought years, although still stressful.  Also, most of the Mokelumne and Merced hatchery smolts were released to the Bay and west Delta, respectively, in 2016, a likely positive factor in contributing strays to overall escapement.  A further explanation for this improvement was better hydrology-related habitat and migration conditions prescribed in the 2008-09 federal biological opinions that generally led to improved habitat conditions.

In the three years (2019-2021) since 2018, runs generally declined (Figure 3) despite being the product of two wet (2017 and 2019) and one normal (2018) year.  One reason for the reductions was that there were fewer strays from hatcheries. For example, the Merced hatchery smolt releases in 2017 were at the hatchery instead of in the Bay, and thus had poor returns.  Battle Creek hatchery returns were also lower, with less straying by smolts released near the hatchery.

The poor returns in 2020 in all three rivers are especially troubling, given they are the product of a good wet year run (2017) and reasonable rearing conditions in winter and spring of normal year 2018.  One factor in the San Joaquin watershed in late summer and early fall 2020 was unusually low flows and high water temperatures for a normal water year (Figures 5 and 6).  Based on the high number of returns of 2018 Merced hatchery smolt releases straying to other rivers (Figure 7), it appears that a compounding factor to these low flows and high water temperatures was high rates of straying by salmon sourced in San Joaquin watershed to the Mokelumne, American, and Feather Rivers.

The relatively high proportion of the Stanislaus River escapement in the 2021 San Joaquin run appears to be a result of attraction to the Stanislaus from a very warm lower San Joaquin River (Figure 8).  The Stanislaus is the first cool tributary encountered by salmon on their journey up the warm San Joaquin in late summer and early fall.

In summary, there is much straying to and from the San Joaquin salmon spawning tributaries.  Adult run size (escapement) is a function of straying, winter-spring flows and water temperature in the San Joaquin and its tributaries during the winter-spring rearing season, and streamflows and water temperatures during the annual late summer and fall spawning run.  The release locations of smolts from the Merced River hatchery and other hatcheries also plays a role.

Salmon runs to the San Joaquin and its tributaries could be improved with better streamflow and water temperature management.

Figure 1. Fall run salmon escapement to San Joaquin River and tributaries 1989-2021. Source: Grandtab.

Figure 2. Plot of 1975-2021 fall run salmon escapement to San Joaquin River tributaries. Data source: GrandTab.

Figure 3. Plot of 2015-2021 fall run salmon escapement to the San Joaquin River tributaries. Data source: GrandTab.

Figure 4. Returns of code-wire-tagged (cwt) salmon to Tuolumne River in 2016-2017 from five Central Valley hatcheries. Source: cwt return data in https://www.rmpc.org.

Figure 5. July-December water temperature in San Joaquin River at Vernalis in 2020, and historical average.

Figure 6. July-December streamflow in San Joaquin River at Vernalis in 2020, and historical average.

Figure 7. Adult spawner returns to four hatcheries and spawning grounds in 2020 of 2018 Merced Hatchery tagged smolts released in Bay. (Note there were no records for Battle Creek returns.)
Source: cwt return data in https://www.rmpc.org.

Figure 8. Water temperatures in the lower San Joaquin River at Vernalis (VNS), Brant Bridge (BDT), and Mud Slough (MSG), and Ripon (RIP) on the lower Stanislaus River in September 2020. Note adult salmon generally avoid 72°F water.

The 2018 run of spawning adult spring-run Chinook salmon in Butte Creek was not abundant by Butte Creek standards.  These were the offspring of the 2015 drought-year run.  California Department of Fish and Wildlife (CDFW) estimated only around 2000 spawning adults in 2018, in the lower 25% percentile of the population counts since 1995. The first egg laying in 2018 happened in late September, and the last of the spawners finished their dance around the middle of October. Carcasses were counted, the wildlife of the creek had been feasting for weeks.

The wait began for the eggs in the gravel to “eye up,” a point where the eggs are known to have been successfully fertilized. Several weeks later, depending on water temperature, the alevin juveniles “button up”, or finish feeding off their yolk sack and started feeding on their own as fry. Fry juveniles were captured by CDFW in November, some early in November. This is fairly rapid salmon development for spring-run salmon, but Butte Creek, where the fish spawn, is one of the warmest stream reaches supporting spring-run Chinook.

Early on November 8^th of 2018, a black cloud rose rapidly over Butte Creek. The Camp Fire began at 6:30 in the morning. It raced across two ridges and the 3-mile-wide West Branch Feather River canyon, and exploded across Paradise and Magalia and down Little Butte Creek Canyon like a torch. Another section burst off the ridge above Centerville and Helltown. By the next morning, 80% of the canyon downstream of Helltown had burned. Hundreds of homes were leveled. The fire meandered through the canyon for two weeks.

As storms began to approach California, a massive effort to try and control the potentially toxic runoff and pollution was initiated by Friends of Butte Creek and the US Fish and Wildlife Service. They placed many miles of straw erosion control wattles around most all of the burned-out structures. When the heavy rains began, it seemed like there was nothing more that could be done to save the emerging salmon. The creek ran black with ash, soil, and debris. Monitoring on Little Butte Creek, which drains much of Paradise, showed high levels of many toxic chemicals, including arsenic. Although the salmon seemed insignificant in light of the destruction from the fire, biologists were worried that there would little success from this small 2018 run.

Early in March of 2021, an amazing sight began surging through the Butte Creek system. Despite a seriously dry winter and relatively low flow in the creek, salmon started showing up in great numbers. Schools of 10-50 fish were sighted throughout the system. By late March, many thousands of spring-run salmon were making their way into the middle canyon reach below Centerville where cool, deep pools provide the most important refuge for these fish to make it through the summer. In April, at one of the monitoring pools with the easiest access, nearly 800 fish held in one pool.

Downstream, water diversion dams in the valley sections of Butte Creek were in pretty good shape after many years of upgrades to the ladders and screens, allowing salmon to reach the upstream holding pools earlier and in better condition than years past. Late arrivals in the past often showed damage to their skin from concrete dams and ladders, and had low spawning success rates.  Fungus often covered the eyes of these late-arriving fish, and many did not make it through the summer.

Figure 1. Weir 1, May 2012. CDFW Photo.

In April 2021, a number of early-arriving fish uncharacteristically showed up with serious damage to their heads. Not much later, word began to spread that water had been shut off at one of the weirs in the Sutter Bypass, and many spring-run salmon had perished. Apparently, some salmon that did make it past the dam suffered injuries in the process. Additionally, the low flow in the Sutter Bypass may have led hundreds of Butte Creek salmon to continue on up the Sacramento River to Colusa where Butte Creek originally entered the river at the Butte Slough Outfall gates. The gates were closed, but the fish sensed this was a potential access to Butte Creek and began bashing their heads and bodies on the outfall gates. It took many days for the Department of Water Resources to open the gates for fish, and many damaged salmon surged into the creek. At times in late April and early May, as many as 10% of the fish showed signs of damage.

Most significant was the size of the run that looked to be the biggest run of spring-run Chinook salmon to ever return to Butte Creek, all from the parent spawning population of about 2000 fish.

California Department of Fish and Wildlife biologists began doing snorkel surveys in June and quickly began estimating over 10,000 adult returnees. Local observers, including this author, estimated over 20,000, rivaling the 20,000 estimated returnees of 1998. The sight was spectacular, and optimism was high that something about the Camp Fire may have contributed to the success of the juveniles from that fateful fall in 2018, along with the wet winter in 2019.  Successful salmon populations may benefit from nutrient releases, sediment and ash cover for their downstream migration, or for other unknown elements of the cycle.  The extra nutrients, and the high flows and water levels in the Butte Basin and Sutter Bypass in winter 2019, were likely beneficial.

The downside of this huge run of fish became apparent in summer 2021. High air temperatures in mid-June pushed the thermometer over 100 degrees F. for several days, and water temperatures in Butte Creek soared. Meanwhile, operations of the imported West Branch Feather River water for PG&E’s DeSabla powerhouse hit a snag when one shallow (and warm) West Branch reservoir (Round Valley Reservoir) ran out of water earlier than expected. This led to a drop in flow that affected Butte Creek for about 24 hours (Figure 2 below). Colder water was released from another West Branch reservoir (Philbrook) a day later and quickly moderated temperature in the creek, but the brief drop in flow came about the same time as the start of the disease outbreak among the holding salmon in Butte Creek. Water temperatures began to rise above 19.8 degrees Celsius soon after the flow drop on June 23 (Figure 3 below). Dead fish afflicted with Ich and Columnaris began turning up in the pre-spawn mortality survey just a week later. Another two weeks later, hundreds of dead fish began rotting in Butte Creek or were dragged off by opportunistic wildlife. By the end of July, when surveys were interrupted by smoky conditions from the Dixie Fire, almost 14,000 pre-spawn mortalities had been counted.

Figure 2. Water import from West Branch Feather River to Butte Creek, June 22-27, 2021.
Source: California Data Exchange Center

Figure 3: Recorded mean water temperature (ºC) within the 3 holding pool locations along with numbers of pre-spawn moralities recorded throughout the pre-spawn mortality survey.
Source: CDFW Butte Creek Spring-Run Chinook Salmon Adult Monitoring Report 2021

Butte Creek is the most productive salmon stream in California and is the home of what is far and away the state’s most important spring-run Chinook population. However, following a year with good production, good juvenile rearing conditions, good ocean conditions, but nearly 92% pre-spawn mortality, one has to wonder if the management and operations of Butte Creek in Butte Creek Canyon under the current configuration of infrastructure are ever going to be able to provide reliable conditions for spring-run salmon to thrive. Although 2021 was the hottest summer on record in California, new records seem to be set almost every year. The Quartz Bowl (Figure 4 below), where most of the 1807 spring-run that survived in 2021 managed to stay alive, is no longer the refuge it has been in the past.

Figure 4. Quartz Bowl Pool, August 2021. Photo by John Sherman.

The salmon spawning reaches of Butte Creek, upstream of the old Covered Bridge near the intersection of Honey Run Road and Centerville Road in Butte Creek Canyon, are largely managed under PG&E’s DeSabla-Centerville Hydroelectric Project, FERC Project No. 803. The Project notably features diversion of water from Butte Creek into the Butte Canal and diversion of water from the West Branch Feather River into the Hendricks Canal and the Toadtown Canal. At the bottom of these canals is DeSabla Forebay (Figure 6 below), a ~200- acre-foot reservoir located next to Skyway (road) on the ridge uphill from the town of Paradise. Water collected in DeSabla Forebay is dropped through a “penstock” (pressurized pipe) into DeSabla Powerhouse, from which it is discharged into Butte Creek. DeSabla Powerhouse is located right next to Butte Creek, about two miles upstream of the Quartz Bowl Pool, the current upstream limit for Butte Creek’s spring-run salmon.

The DeSabla-Centerville Hydroelectric Project’s import of water from the West Branch Feather River helps provide additional cool water for spring-run salmon holding in Butte Creek when managed properly. Maintaining this import of water in some form is essential to the long-term viability of spring-run salmon in Butte Creek.

Historically, PG&E also diverted water at Centerville Head Dam, less than a mile downstream of DeSabla Powerhouse, into Lower Centerville Canal, where water flowed about 6 miles downstream to pass through Centerville Powerhouse, whose outfall re-entered Butte Creek near the community of Centerville. Centerville Powerhouse has been inoperable since 2011, and PG&E has not diverted water into Lower Centerville Canal since 2013.

Figure 5. Map of the DeSabla-Centerville Project and area, from February 2017 PG&E flyer distributed simultaneous to PG&E’s request to FERC to withdraw its application for a new project license.

In October 2004, PG&E began the process of seeking a new hydropower license from the Federal Energy Regulatory Commission (FERC) for the DeSabla-Centerville Project. The relicensing was largely complete with the State Water Board’s issuance of a final water quality certification for the relicensing, revised  in 2016. A biological opinion from the National Marine Fisheries Service (NMFS) for protection of Butte Creek’s spring-run salmon and steelhead under the Endangered Species Act was the last major remaining step before FERC’s issuance of a new hydropower license.

However, in February 2017, PG&E withdrew its application for a new license, announcing its intention to sell the project. FERC disallowed the withdrawal, but held the licensing process in abeyance pending potential sale.

Five and a half years later, on August 16, 2022, PG&E informed FERC that negotiations to sell the project had ended without sale, and PG&E requested that FERC complete the relicensing process. In the interim, none of the conditions that FERC and the State Water Board were poised to require of PG&E in the new project license have been implemented. Most notable among measures not yet implemented is a device to reduce heating of water as it is held in and passes through DeSabla Forebay. NMFS called out the need for such an infrastructure improvement in a Preliminary Biological Opinion in 2006.

Figure 6. DeSabla Forebay and intake tower to DeSabla Powerhouse. The reservoir is almost totally unshaded; ambient summer temperatures are frequently in the 90’s and above. Photo by C. Shutes.

The mass pre-spawn mortality in 2021 put an enormous exclamation point on the urgency of completing upgrades in the DeSabla hydroelectric project. It should also cause fisheries managers and advocates to revisit decisions about whether a large-scale reconfiguration of infrastructure is needed to keep water imported from the West Branch cold enough to benefit spring-run salmon in Butte Creek. Admittedly expensive options like piping all or part of the Hendricks and Toadtown canals, or bypassing them with a tunnel, may be required. Outside funding may also be required.

Perhaps the best and most durable solution, and one that is being tested elsewhere, is to get the fish upstream to colder water. Three separate studies completed in 1997, 1998, and 2000 by Holtgrieve (CSU Chico), Kier Institute for Fisheries Resources, and Watanabe (CDFG) indicated that there is good habitat upstream, and all recommended further study.

There are problems with fish passage to the upper reaches of Butte Creek. These include natural barriers, such as that at Quartz Bowl, about which fisheries managers have traditionally been squeamish (a notable exception is a recently completed fish ladder past a natural barrier on Deer Creek), and also the no-longer-used Centerville Head Dam (Figure 7 below), which has no fish ladder.

There are also problems with diversion of water out of Butte Creek by the DeSabla Project (into the Butte Canal) and by the nearby Forks of Butte Project, the latter recently offered for sale.

In the past, fisheries managers concluded that the difficulties and costs of fish passage and major new infrastructure outweighed the potential benefits. But given the reality of climate change, increasing temperatures, and greater frequency of drier dry years, it is high time to revisit the tradeoffs. Large-scale improvements may provide value that is well worth the costs to save the extraordinary run of spring-run salmon in Butte Creek.

Figure 7. Centerville Head Dam. Photo by Allen Harthorn.

Figure 8: Butte Creek annual escapement and pre-spawn mortalities, 1956-2021. Figure created by Friends of Butte Creek.

Allen Harthorn

Executive Director

Friends of Butte Creek