Photo of adult spring run salmon in upper Butte Creek canyon in summer awaiting fall spawning. Source: California Department of Fish and Wildlife (CDFW).

Butte Creek supports the largest population of spring-run Chinook salmon in California’s Central Valley.¹  The recovery of Butte Creek spring-run salmon is one of the few modern success stories in the Sacramento River watershed. Efforts to restore fish passage and river habitats over the past several decades have paid off quite remarkably, but those efforts are now in jeopardy due to the recent drought and impending changes in water management in the Central Valley and Butte Creek.²

Butte Creek drains a portion of the mountains southwest of Mt. Lassen, on the east side of the Sacramento Valley (Figure 1). The creek’s steep canyon and falls prevent spring-run salmon from passing upstream of Quartz Bowl Pool. Spring-run spawn in about 13 miles of creek downstream of Quartz Bowl from mid-September through October. The adults arrive in the spawning area from about March through May or June. They hold in the deep pools in Butte Creek over the summer until they spawn in the fall. Fall-run salmon ascend the creek on high flow events at any time from September through December. However, CDFW intentionally blocks fall-run from migrating upstream of Parrot-Phelan Dam (location T6 in Figure 1), in order to prevent fall-run from interbreeding with spring-run and from spawning on top of redds that spring-run have already created.

The Butte Creek spring run has increased over the past 30 (Figure 2) in response to extensive active management. The Central Valley Project’s Anadromous Fish Restoration Program and the CalFed Program funded and implemented many fish passage projects and habitat improvements, in cooperation with DFW and local landowners. These programs purchased key properties, screened water diversions, constructed fish ladders, and restored floodplain habitats. PG&E improved management of its DeSabla – Centerville Hydroelectric Project at the upstream end of salmon habitat, increasing flows in the upper seven miles of Butte Creek’s salmon habitat and implementing focused management of the cold water that the hydro project moves through canals from the West Branch of the Feather River into Butte Creek. management improvements. DFW jump-started the recovery of Butte Creek spring-run by stocking spring-run smolts from the Feather River Hatchery in Butte Creek in the mid-1980s. Over that past ten-plus years, PG&E has funded CDFW to closely monitor salmon in Butte Creek.³

I took a close look at the recruit-per-spawner relationship (Figure 3) to portray long-term trends and factors related to success. The key findings are as follows:

1. There is a strong positive recruit-per-spawner relationship with strong time (year) component – recruitment has increased steadily over the years with the buildup of the population.
2. Recruitment per spawner was stronger for brood years with good conditions during critical time periods: fall spawning for their parents, winter-spring rearing and emigration, and subsequent over-summering holding conditions prior to their spawning.
3. The initial stronger runs in the mid- to late-1980s (1986, 1988, and 1989) were jump-started with the initial stocking of Feather River hatchery smolts from 1983-1985 and optimal migration, summer holding, spawning and rearing conditions in the very wet years in the 1982-1986 period.
4. The runs were again depressed during the extreme drought years of 1990-1992, only to recover and expand in the 1993-1999 period of wet years.
5. Recruitment in 1984, 1985, 1999, and 2008 likely suffered from redd scouring in late fall (Nov-Dec) floods of 1981-1983, 1996, and 2005 (Figure 4).
6. Recruitment per spawner in 2015 was poor due to drought rearing and migration conditions in winter-spring 2013, poor ocean conditions in 2014-15, and poor adult migration and over-summering conditions in 2015. In contrast, the 2016 recruitment per spawner was much higher, likely because of better adult migration and over-summer holding conditions (the drought had broken in 2016).

Present management focuses on protecting over-summering adults, primarily by ensuring they have adequate cool water to sustain them until fall spawning. This is not possible without the cool water from the West Branch of the Feather River provided by the PG&E hydro project near Paradise, CA. The future of that project and the Butte Creek spring run salmon are now in limbo.

Figure 1. Butte Creek spawning and monitoring locations for spring-run and fall- run salmon. Spring salmon can reach as far upstream as Quartz Bowl (T1). CDFW intentionally blocks fall-run at the Parrot-Phelan Dam (T6). Juveniles emigrate from spawning grounds in Butte Creek to the Sacramento River (below T9) via Butte Slough, the Sutter Bypass canals, and various other natural and man-made channels in the Butte Sink, west of the Sutter Buttes Source: CDFW.

Figure 2. Escapement estimates (spawners) observed in the spawning reach of Butte Creek from 1975-2016. Source: CDFW GrandTab.

Figure 3. Recruit-spawner relationship for Butte Creek spring-run Chinook salmon (log10 transformed). Year noted is recruit year. Color red denotes dry year rearing (year two years before). Blue denotes wet year. Green is average or normal year. Conditions encountered during spawning can also affect recruit survival. For example high pre-spawn mortality can occur that detracts from escapement estimate. Examples include 1987, 1992, 2003, 2008, and 2015.

Figure 4. Mean monthly flows in Butte Creek as measured at Chico, CA. Daily flows were high – up to 4600 cfs in Nov-81 and 5000 cfs in Dec-81, and 11,000 cfs in Dec-05.

In an April 10 post on the Klamath Chinook salmon run, I discussed an expected record low run in 2017.  The Klamath run has six subcomponent runs, including the Scott River.  Improving the Scott River run is one means of improving the Klamath run.

Like the adjacent Shasta and Salmon Rivers, the Scott is a unique ecological gem, sitting high in the Marble and Trinity mountains before plunging north down the volcanic escarpment into the Klamath River canyon (Figure 1).  Like the Shasta River, the Scott flows through a mountain rimmed glacial valley not unlike those in the North American Rockies or European Alps.  Scott Valley is one of those “beautiful places.”  It is also one of the last great places for salmon and steelhead in California.  Unlike the Shasta River whose flow is supported by large volcanic springs from Mt. Shasta, the Scott depends on snowmelt from the Marbles and Trinities, as well as on springs from its alluvial valley.

Figure 1. The Scott River Valley in northern California west of Yreka, CA (Yreka is located in the Shasta River Valley). The Scott and Shasta Rivers flow north into the Klamath River, which runs west to the ocean. The Salmon River watershed is immediately west of the Scott River watershed. The upper Trinity River watershed is immediately to the south of the Scott River watershed.

The Scott River is home to wild runs of Chinook salmon, Coho salmon, and steelhead trout that make up significant components of the Klamath River runs of these species. In this post I address the Scott River fall-run Chinook salmon. The California Department of Fish and Wildlife has estimated the annual run size since 1978 (Figure 2).

Figure 2. Escapement of adult fall-run Chinook salmon to the Scott River from 1978 to 2016. Data source: CDFW GrandTab.

The run size, or “escapement” in fisheries science vernacular, is a consequence of the previous number of spawners; their success; survival of eggs, embryos, fry, and smolts in rivers; survival for up to several years in the ocean; and finally, the success of adults migrating back from the ocean to river spawning grounds. There is a lot that can happen at each of these life stages that may affect the ultimate escapement. I show the effect of several key factors in Figure 3, which starts from the escapement numbers in Figure 2 and shows the recruits-per-spawner relationship.

I hypothesized the following from the Figure 3 recruits-per-spawner relationship:

1. Recruits-per-spawner is generally higher for wet (blue) rearing conditions – the winter/spring conditions of the year that followed the spawning or brood year. Note that smolts generally reach the ocean by their first summer, so conditions early in their rearing year likely affect survival prior to entering the ocean. Survival may be affected by the rearing conditions in the Scott River and/or those downstream in the Klamath River. Low late-winter and spring flows affect river rearing survival as well as the overall survival during emigration to the ocean. Ocean survival can be a consequence of success during river rearing or emigration: of the smolts that reach the ocean, larger healthier smolts generally survive better in the ocean.
2. Recruits-per-spawner is generally lower as a consequence of dry conditions during the spawning run (red circle years have lower recruits-per-spawner). The lower Klamath River die-off of adult salmon in 2002 was an example of dry year mortality during the spawning migration.
3. Recruits-per-spawner may be depressed in very wet rearing years when floods disturb spawning beds of salmon. An example is 1999. The number of recruits was depressed by the 1997 New Years flood, which affected the fall 1996 spawn. Similarly, floods in winter 1982, 1983, 1996, 1998, and 2006 may have reduced survival and run size in 1984, 1985, 1998, 2000, and 2008.
4. The number of spawners three years earlier has little or no apparent effect on the number of recruits (at least at these levels of spawners). For example: recruits in 2007 and 2008 were relatively high despite low number of spawners three years earlier.

I derived the wet or dry water year designations using Figure 4. I derived the wet or dry August-September streamflow designations for the spawning run from the average monthly Scott River streamflow for those months and years (Figure 5 shows a sample range of years). Note that there is not a lot of difference among years in the August-October flows – they are all relatively low. That is because by August, the snow-melt season is over and base flows are occurring from springs and hay-field and pasture runoff or seepage.

There is also the negative effect on flows from wells and surface diversions, the predominate forms of irrigation in Scott Valley. In the drier years the late summer and early fall river flows exiting the Valley below Ft. Jones can be extremely low (less than 10 cfs – see Figure 5) because of extensive well use, driven by lower available surface water. Low late summer and early fall flows can block salmon from entering the river for several months (Figure 6). This results in loss of stored energy, lower egg viability, and high pre-spawn mortality. It also results in delayed spawning, increasing the likelihood that salmon will spawn in the lower sections of the Scott River, where there is poor spawning and rearing habitat. Low flows in the river upstream can further hinder migration and access to prime spawning tributaries (Figure 7).

It takes about 100 cfs or higher to provide full access to the upper river’s spawning areas. In most years, flows are too low to provide good access. The Scott River Water Trust purchases irrigation water in some years to help the salmon migration. In most years, the river’s baseflows increase soon after the irrigation draw on groundwater ends in late October, allowing unhindered migration.

Part of the solution to the problem of low flows during the spawning run is to cease irrigation earlier. Hay irrigators generally cease pumping water by October 1, and some say there is minimal benefit of irrigating after August. Ranchers often irrigate pastures into November, if only for stock watering. Because many wells are not operated after August, idle wells have sufficient total capacity to readily keep the river watered after September 1 to pump groundwater directly into the river (ranchers would consider this if the costs of pumping were covered). I believe the effect of the extra pumping on groundwater levels would be minimal, because the Valley aquifer recovers from well pumping over the winter-spring recharge season.

Survival during the late winter-early spring rearing and emigration period could be enhanced in drier years by limiting early spring irrigation use and using selected idle wells prior to the irrigation season to add water directly to the river.

In summary, the Scott Chinook fall-run is reduced in dry years when spawning and rearing success are compromised by low river flows in the Scott and Klamath Rivers. The Scott River Chinook salmon population would likely benefit from (1) improved late-summer and early-fall flows that would improve access of spawners to upriver and tributary spawning grounds, and (2) higher river flows in drier years during the late–winter and early-spring rearing and emigration period. A record low run is expected in fall 2017 because of low fall and spring flows in 2015, which limited the survival of the juveniles from the 2014 run.

Figure 3. The recruit-per-spawner relationship for the Scott River fall-run Chinook salmon from 1978 to 2016. Escapement by year (recruits) is plotted against escapement three years earlier (spawners). The escapement values plotted are transformed (Log10X-2). The number shown is the escapement year or recruit year – the year the run was tallied. Number color denotes rearing year water supply type – two years prior to recruit year. Red is dry. Green is average. Blue is wet. Circle represents escapement year water supply during the spawning run (August-September). For example: “04” is run size in fall 2004, which had an average winter-spring rearing year (water year 2002) and dry conditions during the late summer run in 2004 (red circle).

Figure 4. Water years (10/1-9/30) 1978-2016 average annual flow in Scott River measured at Ft. Jones gage. Source: USGS. I designated years above the blue line as wet years, years below the red line as dry years, and years between the lines as average years.

Figure 5. Scott River average monthly flow (cfs) below Ft. Jones for selected years. Source: USGS.

Figure 6. Adult fall-run Chinook salmon waiting at the mouth of the Scott River in the Klamath River in late summer for flows to improve before attempting to migrate up the Scott River to their spawning grounds.

Figure 7. Trickle of flow in the mainstem Scott River in Scott Valley during late summer below Young’s Dam irrigation diversion near Etna, CA. The dam can be seen in the distance at upper center of photo. The fish ladder at the dam is not functional at such low flows. The dam is located approximately 50 river-miles upstream from the river mouth. There is approximately 20 miles of additional Chinook salmon spawning habitat upstream of Young’s Dam.

The Klamath River Chinook salmon fall run is expected to be a record low in 2017.¹ Predictions are near or below the record low run in 1992. These record low runs followed extended droughts from 2013 to 2015 and 1990 to 1992, respectively.

A very low run in 2016 prompted the Yurok Tribal Council to cancel its commercial fishing season to protect future fish populations. The 2016 salmon allocation was the second lowest on record, and failed to provide each tribal member a salmon. The Tribe did not serve fish at the annual Klamath Salmon Festival for the first time in the event’s 54-year history. In January 2017, the federal government issued a disaster declaration for the 2016 Yurok Tribe fishery.²

An April 6, 2017 article in the Eureka Times Standard stated:

• Tribal fishery scientists such as Michael Belchik of the Yurok Tribe stated the low return of spawners is the result of several severe years of drought conditions and river management practices, which caused the waters to warm and become hot beds for toxic algae and deadly parasites. In 2014 and 2015, up to 90 percent of juvenile Chinook salmon on the Klamath River are estimated to have died from an intestinal parasite, believed to be a major cause for this year’s low run, as were poor ocean conditions…. “All these things together conspire to create a real catastrophe for fisheries,” Karuk Tribe Natural Resources Policy Advisor Craig Tucker said.
• Organizations see dam removal and changes to the federal government’s management of the river as being key solutions to the underlying causes of this year’s low salmon return.” “The solution for this problem is to remove the Klamath dams now,” Pacific Coast Federation of Fishermen’s Association Executive Director Noah Oppenheim said.

A Yubanet article described the expected ancillary effect on the whole California coastal fishery:

The disaster stems from a crash of Klamath salmon stocks, but in order to protect the few Klamath fish that are in the ocean, fisheries regulators have little choice but to close or nearly close the economically valuable commercial and sport fishing seasons along the length of the Northern California and Oregon coastlines. This will impact tribal and non-tribal families alike.

CDFW stated: “Chinook that will be harvested in ocean fisheries in 2017 hatched two to four years ago, and were deeply affected by poor river conditions driven by California’s recent drought.”

A UC Davis study placed some of the blame on hatcheries. “My results suggest that hatcheries’ harm to wild salmonids spans the entire Klamath River basin. For fall Chinook salmon, the decline is concurrent with increases in hatchery returns – a trend that could lead to a homogenous population of hatchery-reared Chinook”.

Having been involved in the Klamath River for 40 years, I provide some of my own insights in this post. In follow-up posts, I will take a closer look at the Scott and Shasta rivers, the two main salmon tributaries of the upper Klamath that contribute substantially to the overall upper Klamath salmon run.

A summary of the overall Klamath salmon run escapement numbers or spawner estimates for the past 40 years is shown in Figure 1. The spawning numbers in 2016 were low, yet this drop came only two and four years after near record runs. Contributions for all six upper Klamath subcomponents in 2016 were down substantially from 2014. Predictions of a poor run in 2017 come from the low number of two-year-old “jack” salmon in the 2016 spawning run.

The question is: why did the strong run in 2014 produce the expected record low run in 2017? And why did the strong run in 2012 produce the weak run in 2015? And on the flip-side, why were the runs in 2012 and 2014 so strong, especially given they occurred during the recent multiyear drought?

A close look at the spawner-recruit relationship (Figure 2), how recruits are related to the number of spawners three years earlier, provides further insight into factors controlling long-term recruitment.

1. The spawner-recruit relationship is weak at best, reflecting the fact that estimates might be poor and/or that other factors are more important than just the number of spawners. The 1995 recovery after the record low 1992 run provides compelling evidence that survival and recovery can be strong even from the weakest of runs (with strong hatchery support – see hatchery component for 1995 in Figure 1). Unfortunately, 2017 appears to suggest that strong runs can produce very weak returns three years later if other factors such as drought are dominant.
2. The population crashes (2016, 2004, 1992) occurred after multi-year droughts (Figure 3). Multiyear effects compound changes to sediment, pathogens, and water quality, the suggested causes of these crashes.
3. Population expansions (2012-2014, 2007-2009, 2000-2003, 1995-1997, 1985-1988) occur after a series of wetter years.
4. There may be some underlying effect of floods, as indicated by the poor run in 1999, a consequence of the New Year 1997 flood that washed out the fall 1996 spawn.
5. The poor run in 2016 and the expected record low run in 2017, in addition to the effects of the 2013-2015 drought, may have been affected by poor ocean conditions, as was believed to be the case in the poorer than expected 2004-2006 runs.
6. Several factors potentially affect production or survival per spawner: conditions during the spawning run (flows, water temperature, disease, upstream passage hinderances, etc), first year rearing and emigration conditions (flows, water temperature, predators, prey, disease, toxins, etc), and ocean conditions. It is likely that flows throughout the water year (Figure 4) have some effect on survival of the affected or subsequent brood years.
7. The contribution of the Shasta River appears to have increased in recent years, likely as a result of the Nature Conservancy’s efforts at Big Springs (more on this in an upcoming post).

Overall, the droughts of 1990-1992 and 2013-2015 (Figure 3) were likely the single most important factors in the upper Klamath Chinook salmon population dynamics. The role of the Irongate Hatchery contributions seems relatively stable and a likely important contributor to recoveries after drought. I was unable to determine the contribution of hatchery salmon to the other components of the run, but it is likely a large factor in the Bogus Creek and upper Klamath elements. It is possibly a lesser factor in the Salmon, Scott, and Shasta river elements, which speaks to the importance of these potentially “wild” runs.

In closing, some thoughts on potential solutions:

1. Knowing a good run was occurring in drought year 2014, managers could have done more to protect the spawners, eggs-embryos, and subsequent rearing-emigrating juveniles with better flows and water quality. Perhaps the recent federal court decision may help ensure future protections. In poor water supply years like 1990-1992 and 2013-2015 (Figures 3 and 4), water managers simply must provide protections for salmon.
2. Future removal of the four dams may help reduce the adverse multiyear effect of droughts on disease and water quality and may provide additional spawning and rearing habitat.
3. Much more could be done to increase run components from the Scott and Shasta rivers (more on this in upcoming posts).
4. The hatchery program is long overdue for reform and upgrade. The program should shift from production to conservation of fall-run and spring-run Chinook, Coho and steelhead.
5. These and other suggestions are discussed in a prior post.

Figure 1. Chinook salmon escapement estimates to the upper Klamath River including Irongate Hatchery, Bogus Creek, Scott River, Salmon River, Shasta River, and Klamath River mainstem below Irongate Dam. The preliminary prognosis for fall 2017 total escapement is 11,000. Source: http://www.pcouncil.org/salmon/background/ document-library/#EnvironmentalAssessmentsalLib

Figure 2. Spawner-Recruit relationship for upper Klamath River fall-run Chinook salmon population. The number is the transformed (log10X – 3.5) escapement estimate for the fall of that year as shown in Figure 1. The color represents winter-spring hydrology conditions in the Klamath River two years earlier when this brood year was rearing in river habitats. Red is dry, green is intermediate, and blue is wet (from Figure 3). Circle color represents late summer water year conditions in numbered year. For example: year 92 represents the recruits in fall 1992 from brood-year 1989 spawn that reared in 1990 winter-spring (red dry year); the red circle represents dry conditions in late summer of that water year (1992). Note that the spawning run for 2002, the year the large die-off of adult salmon occurred in the lower river due to low flow and high water temperatures, likely contributed to the poor returns (recruits) in 2004 and 2005.

Figure 3. Average annual discharge by water year (10/1-9/30) of Klamath River as measured at Link River near Klamath Falls, Oregon. Data source: https://waterdata.usgs.gov/nwis/ annual?site_no=11507500&agency_cd=USGS&por_11507500_113138= 545477,00060,113138,1962,2017&year_type= W&referred_module=sw&format=rdb

Figure 4. Monthly average flow (cfs) in Klamath River below Irongate Dam in selected years. Year 2011 was a wetter year. Year 1992 was a critically dry year. Years 2002, 2005, and 2013 were dry years. Year 2016 was an intermediate water year. Source: www.waterdata.usgs.gov.