Can a Wild Coho Salmon Population Recover Following Closure of a Hatchery Program

In California, Idaho, Oregon, Science Friday, Washington by Nick Chambers

Today’s post is the conclusion of our two part guest series on the recovery of Coho in Oregon’s Salmon River. (Click here for last weeks post) Lately we have shared several studies on Pink and Coho salmon, which provide important lessons for salmonid recovery efforts across a range of species and watersheds. Perhaps the most important lesson is that decisions about how we manage wild steelhead, and other species, should be based on empirical data and a solid understanding of their fundamental biology. The current status of Salmon River Coho reflects a perfect example of scientists asking the right questions, collecting solid data to answer those questions, and utilizing study outcomes to inform management decisions. And the result? A resounding success in a watershed many thought was beyond the point of no return for wild coho.


This article summarizes the population dynamics of the wild Coho Salmon population in Salmon River following the termination of the Coho Salmon hatchery program. The full publication is available in the North American Journal of Fisheries Management. Some modifications have been made to the original figures for clarity. Please cite as:


Jones, K.K., T. J. Cornwell, D. L. Bottom, S. Stein, and K. J. Anlauf-Dunn. 2018. Population viability improves following termination of Coho Salmon hatchery releases. North American Journal of Fisheries Management 38: 39-55. Available:



Can a Wild Coho Salmon Population Recover Following Closure of a Hatchery Program

Authored by:

Kim K. Jones

Daniel L. Bottom


The Oregon Department of Fish and Wildlife discontinued a Coho Salmon (Onchorynchus kisutch) hatchery program at Salmon River in 2007 to support recovery of a wild Coho Salmon population in the Oregon Coast Evolutionarily Significant Unit (ESU). The ODFW had already reduced annual releases of Coho Salmon in coastal basins from 4 – 5 million in the early 1990s to fewer than a million by 1999. Additional reductions, including that in Salmon River, from ½ million down to ¼ million occurred as part of the Oregon Coho Conservation Plan in 2007.



The ODFW reduced the coast Coho Salmon hatchery programs out of concern for potential negative effects on wild populations including increased predation and competition, and reduced productivity from interbreeding with hatchery salmon. While genetic and retrospective studies supported this management change, this is the first published study to describe the dynamics of a salmon population in real time as a hatchery program ended.



Here, we describe the response of the first four generations of Coho Salmon after the closure of the program at Salmon River.



The termination of the Coho Salmon hatchery program in Salmon River was significant and opportune for two reasons: (1) the Salmon River population was considered an independent (self-sustaining) population, and (2) we received funds to study the change (if any) in population dynamics during the transition from a hatchery- dominated population to a wild population. This study, the first of its kind, constituted a unique “management” experiment.



Salmon River is a small watershed on the central Oregon coast (195 km2) with 81 and 110 km of Coho Salmon spawning and rearing habitat, respectively (Figure 1). The basin is similar to others on the Oregon coast; it has a mix of federal, state, and private forestland in the uplands and rural residential areas along the lower river. The head of tide extends to river kilometer (rkm) 6.5, and the hatchery is located at rkm 8. Because of ODFW’s long running adult spawning survey program, we were able to compare population changes in Salmon River with those in the adjacent Siletz (south) and Nestucca (north) basins before and after the hatchery releases ended.



This type of study is referred to as a before-after-control-impact (BACI) design. We had data before and after the hatchery releases ended and for nearby “control” basins (Siletz and Nestucca) to distinquish changes in the Salmon River population related to the hatchery program. The advantage of including the adjacent control basins is that we could account for natural variations in river (e.g., temperature, river flow) and ocean (e.g., productivity) conditions that may have altered the Salmon River population independent of any hatchery influence.


Figure 1. Location of the Salmon River watershed and Salmon River Hatchery on the north-central Oregon coast. The distribution of randomly selected juvenile rearing and adult spawning survey sites is shown as an example of the annual sampling design. The inset map of the Oregon Coast Evolutionarily Significant Unit displays the Nestucca, Salmon, and Siletz River basins.



Coho Salmon have a 3-year life cycle in Oregon. Approximately 93% of the adults return at age 3; the remaining 7% return as precocial males (jacks). The spawning year is also referred to as the brood year, and the progeny of the spawners return 3 years later (return year = brood year + 3). The hatchery program ended with the 2005 brood year (2008 return year). We refer to the 1995 – 2005 brood years as the “Hatchery” period, an interval when the adult population spawning in the watershed was primarily comprised of hatchery origin adults (50 – 100%).



During the subsequent “Transition” period, the spawning populations from the 2006 – 2008 broods (i.e., 2009 – 2011 return years) were also comprised primarily of hatchery origin adults. However, the juveniles from these broods were naturally produced and migrated through the estuary and to the ocean without encountering any hatchery fish.  All adults returning from the 2006 – 2008 Transition broods were naturally produced (by hatchery origin and wild origin spawners). The 2009 – 2013 broods (adults return 2012 – 2016) were comprised of 100% naturally produced adults from naturally produced parents (“Wild” period).



We hypothesized that the hatchery program may have reduced the abundance, productivity (population growth), and diversity of spawning times of the wild Salmon River population. These three indices, or metrics, are considered indicators of population viability, or the ability of the population to be self sustaining without additional input from other basins.



We estimated abundance as the number of adults that returned to spawn plus the number that were harvested. This approach eliminated the effects of variable harvest on the year-to-year comparisons. Productivity, or population growth, was the ratio of the number of “recruits”—the total number of returning adults—to the number of spawners (i.e., parents) that produced them. That is, how many adults returned 3 years later for each adult in the spawner population. A value greater than one indicates positive population growth or productivity for that brood year because more than enough adults returned to replace the existing number of spawners in the population. A value less than one indicates declining productivity wherein fewer adults return than spawned 3 years prior. Spawn timing was determined by the number of adults counted each week in the basin throughout the spawning period (October through January).



We were concerned that observed changes in Salmon River may be due to factors not associated with the hatchery closure, such as ocean conditions. Salmon abundance and productivity in the nearby Siletz and Nestucca basins offered useful “controls” for natural variability because no hatchery fish were released in either basin during our study (1995 – 2016). Abundance was standardized across basins as the number of adults per kilometer of spawning habitat to account for differences in total spawning area among the basins. Relative changes in the viability metrics of the three populations (Salmon, Siletz, and Nestucca) were key to understanding whether the Salmon River population showed signs of recovering after the Coho hatchery program was discontinued.



Prior to the termination of the hatchery program, most of the adults spawning in the Salmon River basin were of hatchery origin. Following program closure, 100% naturally produced adults began returning to the basin in 2009 in numbers similar to the total abundance (hatchery and wild combined) during the hatchery period. The variability in wild returns remained high due to ocean conditions but the average and range of values were similar to those observed during hatchery operations. The most important finding was that the spawning population—previously dominated by hatchery fish—did not collapse when the hatchery program ended, but in fact, remained abundant and tracked the pattern of the neighboring basins.


Figure 2. Number of hatchery-reared and naturally produced adult Coho Salmon spawning in the Salmon River, 1995–2016. All hatchery-origin fish were marked, allowing their identification on the spawning grounds.




Changes in the abundance viability metrics are best interpreted by comparing trends in the neighboring control basins (Figure 3). Only the Salmon River population had significantly more adults return during the post-hatchery period (2009 – 2013 broods). This indicates that the Salmon River population increased in abundance after hatchery closure as the neighboring populations remained constant across time periods.


FIGURE 3. Preharvest adult abundance per kilometer (log10 transformed) of wild (unmarked) Coho Salmon in the Salmon River, Siletz River, and Nestucca River for the Hatchery (dark boxes) and Wild (white boxes) brood years (BY) (line within box = median; ends of box = first and third quartiles; ends of whiskers = sample minimum and maximum).




The second measure of viability, productivity, was evaluated by the number of adult recruits produced by the number of spawners three years prior (Figure 4). During the 1995 – 2005 brood years, the Salmon River population was well below replacement levels (i.e., recruits/spawner < 1.0) for most years, whereas the Siletz and Nestucca populations exceeded replacement in many years. The 2009 – 2013 brood years fared poorly coastwide because of poor ocean conditions, but the Salmon River population improved significantly from the hatchery period and was slightly higher than the Siletz and Nestucca populations. This measure indicates that none of these populations were highly productive in recent years, but the Salmon River population has improved since hatchery closure and is performing similar to others on the mid-coast.


Figure 4. Recruits per spawner (log10 transformed) in the Salmon River, Siletz River, and Nestucca River populations for the Hatchery and Wild brood years (line within box = median; ends of box = first and third quartiles; ends of whiskers = sample minimum and maximum). A value of “0” on the Y-axis indicates 1 recruit for 1 spawner (replacement).

Figure 4. Recruits per spawner (log10 transformed) in the Salmon River, Siletz River, and Nestucca River populations for the Hatchery and Wild brood years (line within box = median; ends of box = first and third quartiles; ends of whiskers = sample minimum and maximum). A value of “0” on the Y-axis indicates 1 recruit for 1 spawner (replacement).



A third measure of viability of the Salmon River population—diversity—was indicated by spawn timing. Conveniently, a 1975-1977 ODFW survey in Salmon River provided baseline data for the Coho Salmon spawning population immediately before the hatchery began operating. The pre-hatchery spawn timing in Salmon River (1975 – 1977) was very similar to that of wild populations currently spawning in other coastal basins. Historical spawn timing extended from early October through late January with a peak in December. This pattern spread the deposition of eggs across 4 months and extended the timing of fry emergence in the spring.



During the active hatchery program in Salmon River, Coho Salmon spawning was concentrated in October and early November, prior to the onset of large winter storms. This placed the whole population at risk of being washed out during November and December storms. In addition, early emergence from the gravel may have increased fry mortality during spring flood events. Figure 5 shows an expanding spawning period with later spawning in recent years. Although the duration of spawning activity in Salmon River remains shorter than during the pre-hatchery years, it has steadily expanded since the hatchery program ended, with 25% of the population now spawning in December.


Figure 5. Timing of adult Coho Salmon spawning in the Salmon River for selected return years representing Prehatchery (1975–1977), Hatchery (2007–2008), Transition (2009–2011), and Wild (2012–2014) generations. Tick marks indicate the beginning of each month.






Even though hatchery-origin spawners previously had accounted for most of the adults returning to Salmon River, the naturally produced population did not collapse, and abundance increased and spawn timing expanded significantly after the Coho Salmon hatchery program ended. Recruit – spawner ratios in the Salmon River, although initially low, are now approximately equal to those of neighboring populations. The positive response of coastal Coho Salmon to a reduction of hatchery releases provide direct evidence that the adverse effects of hatchery fish on wild population abundance and productivity may be reversible.


The ODFW’s decision to discontinue the Coho hatchery program at Salmon River has enabled re-establishment of a naturally reproducing population with no decline in adult abundance, supporting ESA recovery goals without adversely affecting fisheries management. The population dynamics within the generations after hatchery releases ended in Salmon River were consistent with the ODFW’s assessment that the hatchery program was the principal factor limiting Coho Salmon population viability in Salmon River.


The Salmon River results have important implications for other populations where hatchery programs dominate salmon production. Despite the poor survival and recruitment of naturally produced fish when the Salmon River hatchery was operating, the rapid increase in natural production during the post-hatchery period suggests that hatchery removal can be an effective strategy for salmon recovery.


Ultimately, long-term resilience of the Salmon River population may depend on whether natural selection processes re-establish a characteristic spawn timing distribution and whether the recovering population can adapt to changes in freshwater, estuary, and ocean environments predicted under climate change scenarios for the region. We recommend continued monitoring of the dynamics and life histories of the Salmon River population to track the long-term response to hatchery removal and to identify factors that strengthen or undermine population resilience.