The primary role of the University of Washington (UW) hatchery is to maintain chinook (Oncorhynchus tshawytscha) and coho (O. kisutch) salmon runs for research purposes. As the adult fish are harvested and spawned, each fish is measured and weighed and the data are recorded into a database. Similarly, during the juvenile rearing stages, the size of fish is recorded at various stages of development to provide growth rate data. Collectively, these data can be used to look at long-term trends within these populations.
In addition to amassing a long-term data set, the hatchery provides a steady supply of research animals for student and faculty researchers. It is a unique opportunity to have a salmon run this accessible. The following abstracts are a collection of some of the recent and ongoing research projects that take place at the UW hatchery.
Long term Trends in Spawning Date, Life-History Traits, and Survival of University of Washington Coho and Chinook Salmon, with Comparisons to Selected Puget Sound Populations
The UW hatchery was founded decades ago, and the data collected as part of routine spawning operations presented a rare opportunity to examine changes in spawning date, age and size at maturity, egg size and fecundity, and survival over time. Analysis of spawning date of chinook and coho salmon at the UW revealed progressively earlier spawning, similar to trends detected at the Issaquah Creek and Soos Creek hatcheries nearby. These trends towards earlier spawning have taken place despite concurrent warming trends in the water supplies to all facilities, which would have been expected to lead to later spawning. Thus the apparent artificial selection in the hatcheries has been sufficient to counteract the natural environmental trends.
Analysis of the trends in size and age at maturity revealed progressively smaller salmon of both species and both sexes over time. The coho salmon are actually growing more slowly than in past years (in both length and condition factor) whereas the chinook salmon are growing at the same rate but maturing at an earlier age (hence smaller size) as a consequence of the tendency of the hatchery to release larger smolts.
In terms of survival, both species experienced higher survival rates in the 1970s and early 1980s than in the years since then, and generally similar patterns were seen in Soos Creek salmon. In neither species and neither hatchery was there a consistent improvement in mean annual survival rate with smolt size at release. Thus the tendency to release large smolts did not increase survival but resulted in younger (hence smaller) salmon, including two age classes of jacks in chinook salmon. Survival rates tended to improve with cooler ocean temperatures off the coast of Washington and British Columbia, but the relationships were less clear than have been reported for coastal populations. This is presumably because much or the marine mortality takes place within Puget Sound, and some of the salmon may spend much or all of their lives there and not feed extensively along the ocean coast.
Quinn, T. P., J. A. Peterson, V. Gallucci, W. K. Hershberger, and E. L. Brannon. 2002. Artificial selection and environmental change: countervailing factors affecting the timing of spawning by coho and chinook salmon. Transactions of the American Fisheries Society 131:591-598.
Quinn, T. P., L. A. Vøllestad, J. Peterson and V. Gallucci. 2004. Influences of fresh water and marine growth on the egg sizeegg number tradeoff in coho and chinook salmon. Transactions of the American Fisheries Society 133:55-65.
Vøllestad, L. A., J. Peterson and T. P. Quinn. 2004. Effects of fresh water and marine growth rates on early maturity in male coho and chinook salmon. Transactions of the American Fisheries Society 133:495-503.
Other Recent Publications
Vøllestad, L. A. and T. P. Quinn. 2003. Trade-off between growth rate and aggression in juvenile coho salmon Oncorhynchus kisutch. Animal Behaviour 66:561-568.
Consequences of Inbreeding in Chinook Salmon
Inbreeding depression, a reduction in fitness that may follow inbreeding, has for decades been among the most prominent genetic concerns of captive breeding programs involving threatened or endangered species. Experimental work in many species has shown a clear link between the degree of inbreeding and fitness loss. Many wild salmon populations exist at low abundance but it is not yet known to what extent inbreeding has reduced and continues to impede productivity of these populations, which aspects of the life cycle are affected most, and whether inbreeding can limit the effectiveness of recovery efforts involving hatchery supplementation or captive broodstocks.
At the University of Washington hatchery, we are conducting research on the consequences of inbreeding in anadromous chinook salmon, in order to characterize the relationship between inbreeding and inbreeding depression, and determine the environmental sensitivity of inbreeding depression. For captive broodstock programs, this information would help to evaluate the risk of inbreeding depression against other risks (such as the risk of domestication). This information in turn would help to formulate guidelines for determining the following:
- under what population scenarios a captive broodstock or captive rearing program should (and should not) be initiated based on current inbreeding levels,
- what captive population sizes should be maintained, and for how many generations, and
- what characteristics of the captive environment are most important to simultaneously reduce risk of inbreeding depression and domestication.
Three basic hypotheses are being tested in this ongoing research project:
* H01: Inbreeding depression does not reduce viability or alter life-history characteristics of chinook salmon.
Ha11: Inbreeding depression reduces viability during early life history but does not affect development rate, age structure, or reproductive capacity.
Ha12: Inbreeding depression has effects throughout the life cycle.
* H02: The degree of inbreeding has no predictable effect on inbreeding depression in chinook salmon.
Ha21: The relationship between inbreeding and inbreeding depression is linear.
Ha22: The relationship between inbreeding and inbreeding depression is nonlinear
* H03: Inbreeding depression in chinook salmon does not vary between captive (i.e., protective culture throughout life cycle) and hatchery (i.e., protective culture from embryo to smolt) environments.
Ha31: Inbreeding depression is greater in a hatchery than in a captive environment.
Ha32: Inbreeding depression is greater in a captive than in a hatchery environment.
In order to test our hypotheses, we have initiated a multi-generational study at the UW hatchery. We have released tagged lines of chinook salmon of known inbreeding history, and are comparing fitness traits in these lines to an outbred control line. Asubset of the fish produced is reared in captive culture and not released. The rate of inbreeding is deliberately increased with each generation by performing controlled matings between individuals within the same line. Chinook salmon typically return every 3-4 years, and the experiment is necessarily a prolonged affair.
Observing the experiment
A visit to the hatchery in the fall spawning months of 2005 and 2006 will show our researchers collecting life history and morphological data for fish that return as part of this experiment. Typical data collected includes survivorship, length and weight, age at maturity, and fecundity in females. At the same time, we read coded wire tags that were implanted in the snouts of the fish when they were released from the hatchery as smolts. The tags are family-specific, which allows us to match fish within each line, and to perform matings between these fish.
We will collect data on the juvenile rearing stages during Winter and Spring of 2006 and 2007, and a visit to the hatchery will show large numbers of tanks housing individual families. We will measure growth rate and later tag experimental fish using both adipose fin clips and coded wires. The experimental fish will be transferred to our outside pond for imprinting, and released at smolting.
Our work is supported by the Bonneville Power Administration.
Stephanie M. Carlson, Harry B. Rich, Jr., and Quinn, T.P.
Validating the Male Spawning Participation Method Proposed by Schroder (M.S. thesis, University of Washington, 1973)
Reproductive success of salmonids can be estimated through a variety of techniques from genetics, which are very costly, to behavioral observations, which can be incredibly time consuming. Schroder (UW M.S. Thesis, 1973) introduced a technique for estimating spawning participation of male chum salmon that would not only require little effort but would also be cost-effective. Additionally, individuals in the wild could be sampled opportunistically at death and would not be handled during the reproductive process (for instance, to obtain a tissue sample for genetics) or disturbed in any way by individuals attempting to obtain behavioral observations. The goal of this study was to validate the technique prior to applying the technique in the field. Specifically, Schroder (1973) was interested in the relationship between number of spawning events that a male participated in and the extent that his gonad was depleted. He held eight male chum salmon in a tank until death. He spawned the fish a number of times to determine the relationship between number of spawning events and gonad mass at death. Of his eight experimental fish, a single fish was held as a control fish and not spawned at all. This one fish had a spawning mass at death that was only 3% less than his expected pre-spawn gonad mass based on his size. This result suggests that the Schroder’s method may be sound but the technique should be verified in a more rigorous fashion (i.e., with a sample size > 1) before it can be applied to other situations. For instance, this technique has potential widespread applicability to field situations where dead males encountered opportunistically during creek surveys could be sampled and their “spawning participation” assessed. Currently, no other investigators have validated the technique. Both coho and chinook salmon return to the UW hatchery and, in theory, either could be used to assess the validity of this technique. However, the smaller size of coho makes them more amenable to sampling.
To assess the validity of Schroder’s method, we sampled a total of 45 coho salmon returning to the UW hatchery. We sacrificed 17 fish spanning a range of body sizes to determine a pre-reproductive gonad mass versus body mass relationship. As expected, there was a very strong relationship between the two. Specifically, we found that 84% of the variation in gonad mass could be explained by differences in body mass alone. Another 28 fish were individually tagged with unique identifiers and held in the hatchery circulars. We artificially spawned each of these fish on a number of occasions and, at each spawning event, we recorded the mass of milt removed. At death, the gonads from all individuals were dissected from the fish and the mass was recorded. The data from this study has yet to be analyzed but this experimental design will allow us to assess whether the gonad mass at death plus the milt removed during spawning differed from the expected pre-reproductive gonad mass based on the particular individual’s size. Thus, ultimately, this study will allow us to assess whether male coho salmon are able to replenish their milt, which would invalidate Schoder’s method.
Effects of Slope, Substrate, Cover, Predators, and Ontogeny on Lentic Habitat Preferences of Juvenile Chinook Salmon Oncorhynchus tshawytscha in Experimental Arenas (Master’s thesis)
Cedar River Chinook salmon Oncorhynchus tshawytscha rear in Lake Washington and migrate through it before going to sea in the first year of their life. This is a relatively unusual behavior seen only in several lake systems throughout the native range of Chinook salmon. Previous field research has found that lake-rearing Chinook salmon fry from the Cedar River migrate to Lake Washington soon after redd emergence (January-April) and mainly occupy shallow littoral regions (depths ≤ 1 m) through mid-May. To supplement and clarify field observations, our experiments examined how bottom slope, substrate, and cover influenced juvenile Chinook salmon habitat choice during this shallow, nearshore period. Besides these abiotic factors, we also examined how Chinook salmon ontogeny and predation pressure by cutthroat trout O. clarkii and prickly sculpin Cottus asper affected habitat choice. In our experiments, both fry and presmolt life stages demonstrated a significant utilization pattern among bottom slopes, with steeper slopes mostly avoided. Both life stages also displayed a negative response to increasing substrate complexity. This trend reversed for some experiments when cover options were added to the substrate treatments. Generally, predation pressure and diel observation period did not have a direct effect on habitat choice. This may be an artifact of experimental fish behavior or an inherent life history choice by juvenile chinook salmon, which grow very quickly in Lake Washington. All of the observed trends became less significant as fish grew. Overall, our experimental results support those of Lake Washington field observations. Further experiments should be conducted to clarify the importance of cover and predation pressure by piscivores on juvenile Chinook salmon lentic habitat selection.
All bottom slope experiments in 2004 were conducted in the large broodstock raceway at the University of Washington hatchery.
The rearing facility consists of various types of rearing space. The original setup consisted of 24 rearing troughs that were 8” x 1’ x 15’, a 16’-long work trough and 4 circular tanks between 5 and 6 feet in diameter. From a production perspective, this was an excellent setup. However, the demand for mass replication required more individual rearing units. Mark Tetrick, the former hatchery manager, designed a new rearing system to accommodate this demand. In the summer and fall of 2002 the facility was remodeled to be more flexible and allow for replication work.
The new system is composed of concrete benches and drains that are integrated into the floor. These benches accommodate a variety of sizes of tanks depending on the rearing requirements. The primary configuration consists of 126 36” semi-square tanks and 10 18” x 96” troughs. Each tank/trough has three water sources available to it and can be connected via quick release hoses that allow for easy reconfiguration.
The outside rearing area consists of 2 in-ground 16’ circulars, 6 in-ground 5’ x 30’ raceways, one 10’ x 100’ raceway, and the 1/3 acre pond. These areas have remained unchanged other than the application of new coatings.
The egg incubation room contains two independent incubation systems; one containing 12 vertical stack incubator units, and a smaller system of 5 similar stacks; each stack consists of 16 Heath incubation trays. Both systems are semi-recirculating and are refreshed with between 3-5 gpm of water, depending on the egg load. Dechlorinated municipal water is used to eliminate the risks of contaminating the eggs with pathogens and fungus found in the lake water source.
In addition to the incubation stacks, each system incorporates a supply pump, sump tanks, head tanks, and a filtration/chilling loop. The pump pulls water from the sump tank and sends a portion of it up to the head tanks that provide water to the stacks at a constant rate (3-5 gpm). The water runs through the stacks and returns to the sump tank. The remaining portion of the water from the pump runs through a filtration/refrigeration loop. This loop consists of a mechanical filter, denitrification tower, UV sterilizer, and a chiller unit. After the water is passed through a mechanical filter it is divided between a denitrification tower, UV sterilizer and a chiller unit, before returning to the sump tank.
The majority of the water used for hatchery operations is pumped from Portage Bay through a gravel filtration bed located on the north shore of the Montlake cut. There are a total of four 650 gpm pumps, two dedicated to the pond and two that provide water for all the interior spaces. The pond pumps can be diverted to provide water to the other outside rearing spaces as well. The interior rearing spaces have three water sources available: ambient lake water, heated lake water and “well water”. “Well water” is actually ground water that percolates up into subspaces of buildings on upper campus. That water needs to be pumped somewhere, so it has been plumbed to the hatchery through the steam tunnels. As it makes its way here it is warmed to 21°C due to the high temperatures in the tunnels. Using a mixing valve and all three water sources we can adjust the water supply between ambient, ranging down to 8°C, and 26°C depending on demand.