Juvenile Pacific salmon enter marine waters of the Strait of Georgia from April until June, depending on the species. Mid-water trawl surveys conducted by the Department of Fisheries and Oceans in July and September provide insight into the population dynamics of juvenile Pacific salmon during their first summer. It has been suggested though that up to 90% of juvenile Pacific salmon may die within their first eight weeks at sea, and yet no surveys have been conducted during this time in recent years. Project 1 examines the early marine distribution, growth, and survival of juvenile Pacific salmon, focusing on Chinook and coho salmon. The objective is to obtain information on the young juvenile salmon before they move offshore to deeper areas of the Strait of Georgia. Project 2 examines habitat usage and early marine mortality of juvenile Chinook and coho using PIT tags. Project 3 provides complementary information on the distribution and behaviour of juvenile steelhead, sockeye and chinook immediately upon ocean entry and provides direct estimates of survival using acoustic tag technologies. The latter is enabled by using new tags that allow researchers to tag salmon down to 8.5 cm and track individual fish using fixed arrays of receivers. Salmon that can be tagged include sockeye (Chilko and Cultus Lake smolts), coho salmon generally, Spring Chinook, and Steelhead. There are three key projects in the juvenile salmon studies category:
1. Juvenile salmon surveys (in-river surveys, PIT tagging, beach seining, purse seining, trawl)
Team: Marc Trudel (DFO-PBS), Chrys Neville (DFO-PBS), Mike Dempsey (DFO-IOS), Carol Cooper (DFO), Svetlana Esenkulova (PSF), Oline Luinenberg (PSF), Dave Preikshot (Madrone Environmental), Lana Fitzpatrick (DFO), August Jones (PSF), Francis Juanes (Uvic), Rana El-Sabaawi (Uvic), Azit Mazumder (Uvic), West Vancouver Laboratory (Ian Forster), Pacific Biological Station (Strahan Tucker, Stewart Johnson).
Current studies include examination of:
- In-river mortality, habitat use, and migration patterns
- Spatial and temporal distribution of salmonid juveniles in the estuary and bay (using beach seines and purse seines)
- Presence-absence of predators (fish predators collected via purse seines, small trawls and gillnet studies, seals and bird impacts on juveniles assessed using on-ground observations and scat analyses.)
- Stomach analyses (to assess diets, quality of prey) of salmon and predators of salmon.
- Plankton densities/composition
- Genome analyses and physiology studies (to assess health and condition of fish, and to assess the links between marine survival and growth, physiology, disease, etc.)
- Hatchery-wild fish interactions
a. Freshwater Survival: Rotary Screw Trap & Freshwater PIT tagging
Team: BCCF (Kevin Pellet, James Craig, Wightman, Stenhouse, Dambourg, Atkinson), Mel Sheng DFO-PBS, Steve Baillie DFO-PBS, Don Elliott Cowichan Hatchery Manager, Dr. J. Taylor, Dave Key Key Mill Construction Ltd. dba Pisces Research Corps Elliot, Cowichan Tribes.
In the last 5 years, research has been undertaken to monitor the spatial and temporal distribution and ratio of hatchery and wild Chinook entering the Cowichan Estuary and through their first months in the ocean. This information has been used to estimate the early marine survival of hatchery and wild Chinook salmon from this system. The estimates assumed that all hatchery fish released 40 km upstream from the estuary in the Cowichan River survive and enter the estuary. However, based on previous freshwater work, this assumption is unlikely to be valid. Therefore, determining the freshwater mortality of the Chinook salmon released from the hatchery, is essential in determining the level of early marine mortality of these fish.
The level of freshwater mortality is being determined by operating a Rotary Screw Trap (RST) in the mainstem of the Lower Cowichan River and conducting a mark-capture population estimate on hatchery released fish. If it is concluded that hatchery smolts released a significant distance upstream from the estuary experience high losses prior to saltwater entry, a follow-up study could be implemented to determine if lower river hatchery releases improve overall and river return survival.
In 2013, staff from BBCF conducted snorkel surveys throughout Cowichan River and estimated that tens of thousands of larger wild Chinook smolts remained in the river in May-June. In contrast, the presence of Chinook juveniles in the estuary started in April, peaked in May and decreased to negligible numbers by June. The migration entry timing into the estuary and the percent contribution of these later freshwater rearing wild juveniles to the adult returns is unknown.
The RST is also being used to estimate the migration timing and numbers of these wild juvenile Chinook salmon. The survey design for the RST and mark recapture was developed by Dr. J. Taylor (JA Taylor and Associates Limited) following similar surveys conducted in BC. The work has been undertaken by DFO and Cowichan First Nations. Daily and weekly sampling activities were consistent with statistical program requirements outlined in the design. Unfortunately there were issues with placement of the RST in 2014 so it was not possible to assess the population size of wild Chinook smolts migrating that year. Based on what was learned during 2014, some amendments to trap design and location were made, and the study was repeated during 2015, and again in 2016.
The RST work was supplemented with a freshwater PIT tagging study during 2015-16. Cowichan Hatchery chinook were PIT- tagged and released at the usual hatchery release location (upstream at Road Pool), and their survival estimated using a RFID array at the bottom of the Cowichan River, which records each PIT tag as the fish crosses the array. The survival estimate using this method can be compared with that from the RST.
Rotary Screw Trap
There is a high level of mortality (98%) during the first two years of life of Cowichan Hatchery chinook salmon, with many possible reasons. This project examined the level of mortality in the initial stage associated with the hatchery fry releases, between the release in the upper river and their arrival downstream to the lower river where they reside for up to 2 months.
A rotary screw trap was used to catch a sample of the hatchery chinook moving downstream. The proportion of the run that were caught is estimated by releasing a small group of stained chinook just upstream of the trap, far enough so that they are randomly distributed. The proportion of the stained release caught in the RST provides a value for trap efficiency, and is used to expand the number of hatchery chinook to estimate the number that arrived to this location. This result is compared to the number of released chinook fry to estimate the mortality rate. This was accomplished for both the early and late releases.
Two groups of Cowichan Hatchery Chinook were released on April 22nd (348,485 CWT) and May 20th (465,438 CWT). A rotary screw trap was operated between April 8th and May 30th 2015 to monitor wild salmonid juvenile migration and to recover and estimate the population of hatchery chinook smolts that successfully migrated to the lower river (40 kms downstream). 9940 dye marked fish were released April 23rd approximately 1 km above the RST to estimate the trap efficiency and 4746 dye marked fish were released May 20th.
Trap efficiency during the two releases was estimated at 27.2% and 19.9%, respectively. Based on this data, they estimated that 65.5K chinook fry (95% CI 63.3K-67.8K) had arrived in the lower river from the early release group of 338.5K. Similarly, 49.7K chinook fry (46.7K-52.8K) arrived in the lower river from the late release group of 460.4K. Thus, they estimated that 19.4% of the early release and 10.8% of the late release represent the portion of the population that arrived in the lower river during the extent of this project. The data from the late release in 2014, using much poorer quality data, indicated a survival rate of 15-20%, similar to 2015.
Several factors are confounding the results. Although the trap efficiency was much higher that their target of 5%, they expected a higher level of trap efficiency with the late release over the early release due to a lower rate of water discharge. Second, they were forced to terminate field operations due to low water, with hatchery release chinook still being caught. In addition, early release chinook were still being caught when the late release was initiated so all hatchery chinook caught after that date were deemed to be from the late release.
The mechanism for the low survival rate is unknown and will need to be studied. Possible factors include:
- Residualization/slow migration rate
- Freshwater based predation (non-native brown trout)
The project was repeated during 2016. Field operations were started on 6 April 2016 and completed on 24 June 2016. Again, a rotary screw trap was installed on the Cowichan River to catch a sample of downstream migrating salmon, including released hatchery origin chinook fry. Two paired releases were completed, with the early release on 12-April and 28-April and the late release on 12 May and 25 May. One of the early groups and one of the late groups were released in the Road Pool, which is the normal location upstream. The other early and late groups were each released in one of two other locations (Stoltz Bluff and Horseshoe Bend) mid-way (approx 20kms) down the Cowichan River. The table below shows the number of fish released by date/location.
In addition, approximately 140K CWT hatchery smolts were also released directly from Cowichan Hatchery May 24th to determine if overall adult survival improves compared to smolts released 40 and 20 kms upstream. The smolts were supplemented with river water 2-3 hours per day, 2-3 times per week for a month period before release to assist with imprinting. We will have more information on whether the hatchery release was successful when the fish return to the river over the next few years.
The individual survival estimates for the upstream (Road Pool) versus downstream releases are shown in the table above. Between 10.7-15.2% of fish survived to the RST from the upstream release site, while fish released lower in the river showed higher survivals (31 -47%). The early releases showed lower survival overall. It is recommended that the lower release sites be used in all future Cowichan Hatchery release plans.
This project provided evidence to support the hypothesis that a high level of mortality is occurring in the freshwater phase of hatchery reared fry. This stage of the life history occurs in the period after release in the upper reach of the Cowichan River, to the lower river upstream of tidal influence. Further investigation is warranted to compare this result to other similar studies to examine whether there are unique influences in the Cowichan River, and/or whether this observation is common to other situations. Further investigation also is warranted to identify whether the observation is due to hatchery release practices, habitat degradation, predator communities or other hypotheses. A detailed study of river predators is taking place during 2017.
Freshwater Pit Tagging
One of the key findings to date is that in-river survival is lower than expected for both wild and hatchery Cowichan River Chinook. Results from the RST program have been validated with PIT tag detections in-river and in the early marine environment.
In 2015, hatchery chinook were PIT- tagged and released at the usual hatchery release location, and their survival estimated using a special RFID array at the bottom of the Cowichan River, which records each PIT tag as the fish crosses the array. The survival estimate using this method could then be compared with that from the RST.
- Despite very low spring flows during 2015, in river migration timing and rearing behavior appeared to be similar to 2014 (a period of normal flows).
- Migrating hatchery Chinook from the release site to the lower river appeared to endure a high degree of mortality. They estimated survival using detection of PIT tags at the RST (22%) and RFID array (27%). Thus the PIT tagging results showed similarly high mortality rates in the freshwater phase of juvenile Cowichan chinook life history as found using the Rotary Screw Trap.
- Migrating wild Chinook also experienced significant mortality although survival appeared to be 2X higher than hatchery fish (estimated at 49% over 41 km)
- Freshwater growth rates were calculated at 0.70 mm/day in 2015 compared to 0.60 mm/day in 2014. This difference was not statistically significant but suggests growth was good despite above average river temperatures and low flows.
Activities in 2016 focused on paired releases of wild/hatchery Chinook at five upstream locations. Subsequent detections at a permanent detection array on the Cowichan River (installed May 2016) revealed losses of 70-80% over a distance of 40 km with a very strong linear trend between sites. Incidental recaptures of PIT tagged Chinook in the early marine environment (Cowichan Bay/Sansum Narrows) showed a dramatically reduced probability of encountering fish tagged in upper reaches of the river in 2015 and 2016.
Abnormal instream detections were noted at a lower river side channel detection array in early June of 2016. Tags released in the mainstem were appearing to migrate into the bottom end of the side channel in clusters. A trail camera was deployed to test the hypothesis that tags were actually inside of another animal that consumed multiple tagged fish. A family of raccoons was captured on the camera within 8 hours of deployment and appeared to interact with the antennas in a way that would result in tag detections. Approximately 4,300 hatchery fish from the late release at Road Pool were estimated to have been consumed over a 15 day period. Abnormalities were discovered upon further analysis of other detection arrays including 12 tags detected 4.5 km upstream of their release location. A river otter, a raccoon and two mergansers were later captured on a trail camera at the mainstem array. A full scale study of chinook predation will take place on the Cowichan River during 2017.
These observations provide evidence that:
1) Large losses were sustained by downstream migrating wild and hatchery Chinook smolts in 2016. Survival to the marine environment for the late hatchery release group was estimated at 12% (23,250 of 193,748).
2) RST work suggests this trend is observed in most/all years where hatchery Chinook survival was investigated.
3) There is mounting evidence that predation accounts for the majority of losses based on abnormal tag behavior on multiple arrays as well as photographic evidence of animals interacting with detection fields.
4) A large portion of the marine survival estimate for hatchery fish released in the upper river can be attributed to freshwater losses in most or all years. They estimate that only 10-15% of the late hatchery release group (CWT indicator stock) made to the marine environment in 2016 assuming the rate of loss observed over the first 40 km continued through the last 7 km.
5) The hypothesis that larger smolts have a higher survival rate in freshwater is not supported by tag detections in the lower river.
b. Marine Juvenile Sampling Studies
The primary objective of the marine juvenile salmon sampling studies is to understand the factors and processes affecting the marine survival of juvenile Chinook salmon. Through a combination of laboratory and field programs (i.e., PIT tagging, beach seine, purse seine, trawl), these studies aim to: 1) estimate the survival of juvenile Chinook salmon throughout their marine life, and 2) to test specific mechanisms that have been hypothesized to affect the marine survival of juvenile Chinook salmon.
For 2014, a pilot study focused sampling on Cowichan Chinook salmon. Taking a comprehensive approach, a number of bottom up factors (e.g. possible impacts of temperature, salinity, availability, timing and quality of zooplankton etc.) and top down factors (e.g. impacts of fish and avian predators, impacts of seals and sea lions, impact of disease, contaminants etc.) were examined in a coordinated program. The top-down projects are discussed in a separate section (Top Down Studies). During 2014, intensive sampling of juvenile salmon in Cowichan estuary and bay occurred concurrently with oceanographic sampling, and this project was continued during 2015. These juvenile sampling studies are designed to address certain key questions, including the distribution and abundance of juvenile salmon (all species, but Chinook is the focus species), the growth and health of the fish, and the differences in these metrics between hatchery and wild fish. For the juvenile salmon collections, this has taken the form of beach seines, purse seines, and small boat trawls to collect juveniles. Beach seining was begun in time to catch the first pulse of outmigrating salmon and ends once the fish have moved off the beach (late June). Purse seining extends into July to catch the fish in the waters Cowichan Bay and Satellite channel. Trawls using small trawl nets are also carried out using a different vessel, the Neocaligus, which also samples a variety of areas in the Bay and further offshore.
In 2015-present, studies were also carried out on an east coast of Vancouver Island population (Big Qualicum River), as well as a Fraser River population (Chilliwack). Attempts will be made to focus on both Chinook salmon and coho salmon, as well as hatchery and naturally spawned fish, for a potential combination of 12 species/populations/life history.
All the juvenile Chinook salmon and coho salmon are scanned for coded wire tags (CWT) and PIT tags (Chinook salmon only). Otoliths of juvenile Chinook salmon are removed to estimate marine growth and the timing of ocean entry. Fish are also assayed for RNA:DNA ratios to provide a short-term measure of growth. Stomach contents are removed to determine their diet. Muscle tissues are also removed for stable isotope and fatty acids analyses. Stable isotopes will be used to complement diet analyses to determine the proportion of fish in their diet and as an overall indicator of ecosystem productivity. Fatty acids will be used to determine the nutritional status of juvenile Chinook salmon. Cesium concentrations are measured in both wild and hatchery fish to estimate their feeding rate.
These analyses will be performed in collaboration with the University of Victoria (Francis Juanes – otoliths, Rana El-Sabaawi – fatty acids, and Asit Mazumder – stable isotopes and fatty acids), the West Vancouver Laboratory (Ian Forster – fatty acids), the Pacific Biological Station (Strahan Tucker – Cesium analyses, fatty acids, and stable isotopes, Stewart Johnson – fatty acids and RNA:DNA ratio).
- Many juvenile Chinook salmon appear to remain in the inner Cowichan Bay for an extended period of time. Purse seine fishing was extended into mid-August during 2015. The catches were lower than July but many Chinook were still captured and the largest catches still remained inside Cowichan Bay.
- Diet and feeding of juvenile Chinook salmon changes in the presence of harmful algae (higher incidence of empty stomach). This was seen in 2014 when harmful algae were observed. They cannot say if this reduction in feeding (increase in empty stomachs) is due to a change in the fish behaviour or a change in food abundance. There was a reduction in their primary diet item (crab larvae) during these periods of time.
- The level of growth hormones in juvenile coho is higher at the north end of the Strait of Georgia than at the lower end, and also higher in even years when there are lots of juvenile pink around. The growth appears to be related to herring in the diet: high growth when they are feeding on herring, and lower growth when they are feeding on crab larvae (data from 2012-2014 – work done in collaboration with B. Beckman). Growth was lower in the Strait in 2013 for coho and chum – they are still waiting for the results on sockeye. One consideration was whether this might explain the poor performance of Fraser River sockeye this year.
- YOY herring were big during 2015 and silvered up earlier than normal. They observed YOY herring in the first purse seines off Big Qualicum River (BQR). Typically at this time of year they tend to see few of these and more larval herring. This “larger” size persisted through September. Results from herring survey in September has corroborated this with the largest YOY herring recorded in their time series during 2015.
- Juvenile chum salmon were also large early in the year and again first noted in sampling off BQR. Verbal reports from hatcheries on ECVI indicated earlier movement of these juveniles down river and likely cause for increased size.
- The feeding of all juvenile salmon in the Strait of Georgia this year was good. There was an increase in larval fish other than herring in many regions.
- Juvenile coho were also big, though they still need to perform DNA analyses to determine their origin. The size of coho in the September survey were the largest in their 15 year time series for this time period. CWT were analyzed to confirm they had entered the ocean this year and also showed that there was the typical mix of SOG and Puget Sound stocks. DNA analysis will help finalize the proportion. In addition to size, there was a high number of jacks in the sample (7.5% of all males). The number was higher for hatchery fish (AD clipped, 9.5%) than wild (unclipped, 6.3%). Subsequent to their sampling, the hatcheries around the SOG have been reporting high jack return.
- There are some intriguing results with respect to fall catches of juvenile coho in 2013 and 2014. Chrys Neville observed the highest CPUE of juvenile coho in the Strait of Georgia in the fall of 2013, but that was not accompanied with high survival. They do not know why, but need to do DNA analyses to determine their origin (either the fish did not survive the winter, or the Strait was invaded by Puget Sound coho). The fall catches of coho were average in the Strait of Georgia in 2014, but returns were very poor. This may be due to the warm temperature anomalies this winter, spring and summer. They would suggest that the relationship seen between September abundance of coho and early marine survival is not as strong in the last couple of years. They have noted average (2014) to above average (2013) numbers but not the same response in total returns. This is complicated by a change in behaviour of the sub-adults with more of these available to sport fisheries in the SOG, changes in the hatchery/wild proportions (yet wild fish still not allowed to be retained in fishery although many caught), and changes in size. The size of the fish is increasing and they do not know how this is related to survival or distribution. In addition, the warm water off ECVI and huge harmful algal blooms are anomalies out of our experience to know how they may affect fish. Overall, in 2014 the fish appeared healthy and were growing very will (biggest until this year). Something happened to them after this point and the result is very variable returns to many systems in the SOG (some average and some poor to dismal).
Pacific pomano were caught in Cowichan Bay during 2015. This species has not been seen before by their surveys in the Strait of Georgia.
The salmon ecology and physiology in the Strait of Georgia study has now completed 2 years of sampling from juveniles in freshwater through their early months in the ocean and into their first fall at sea. The figure below shows the marine sampling locations for 2016.
DNA for stock analysis has been submitted for all marine captured Chinook salmon over the two-year period. Based on DNA results, several stocks of Chinook salmon have been identified to conduct more detailed fatty acid, stable isotope and cesium analyses. For 2015 these stocks include Cowichan, Big Qualicum River, Puntledge and the group of stocks originating from the South Thompson. Initial results from these analyses are now being received and it is anticipated that analysis of these first results will be completed in the early spring of 2017. Additional 2015 samples and samples from 2016 sampling will be processed between January and March of 2017.
Over 2016, 2 staff members were trained in the preparation of otoliths for circuli reading and analysis. Currently, otoliths from 2015 study are being processed and the reading of the circuli from these otoliths began in January 2017.
Analysis of the diet of Chinook salmon is indicating that the majority of the diet is composed of 4-6 key diet items (see figure above). However, there is variation in the proportion of these diet items by area within the Strait of Georgia and possibly by stock or size of fish. This analysis is currently in progress but results of the initial two years of study are anticipated to be complete by the end of April 2017. Integration of diet and survival data with studies in Puget Sound is ongoing to improve our understanding of variability between the basins.
Diet results and laboratory analysis results from the first two years of juvenile salmon sampling in conjunction with results from studies examining the zooplankton production in the Strait of Georgia and fatty acid and stable isotope analysis of plankton samples and diet samples will be used to test a number of hypotheses.
The analysis being conducted on biological samples from 2015 and 2016 will provide information on the change in the condition of juvenile salmon over their first summer at sea. The juveniles captured by trawl in September and October are the survivors of the first marine summer. Therefore, variation between these fish and the fish caught by trawl in May and June will be an indicator of the requirements needed for survival.
Analysis of CWT recoveries from juvenile Coho and Chinook salmon in the Strait of Georgia was conducted for all surveys between 1998 and 2015.
Stable isotope analyses have been performed on a subset of the juvenile Chinook salmon caught in 2015 (Figure 5). The increase in the carbon and nitrogen isotopic signature observed in South Thompson Chinook salmon indicates the transition from freshwater-derived nutrients to marine-derived nutrients. This was not apparent in other stocks due to their larger size at capture. The higher carbon isotopic ratio of Puntledge Chinook salmon suggests that these fish are feeding in different regions of the Strait. This is supported by the CWT recovery data (Figure 4), which shows a more northerly distribution. The proportion of fish and shift to piscivory will be inferred once stable isotopes are determined in both zooplankton and larval fish.
- The salmon ecology and physiology in the Strait of Georgia study has completed 3 years of sampling from juveniles in freshwater through their early months in the ocean and into their first fall at sea. This included freshwater downstream sampling at multiple locations on Vancouver Island and Fraser River, beach and purse seining in Cowichan Bay, purse seining off BQR and charter trawl surveys in SOG in June and September and Puget Sound in July.
- The field objectives for 2015 to 2017 were completely met with samples collected over time. In 2018 auxiliary sampling using standard surveys was completed in June 2018 and will be conducted in September 2018 to support the research program. In addition to fishing, bongo (plankton) and CTD casts were conducted on all surveys.
- Stock analysis using DNA and CWT recoveries have been conducted for coho and chinook salmon from the three year study. Based on DNA results, several stocks of Chinook salmon have been identified to conduct more detailed analyses including Cowichan, Big Qualicum River, Puntledge, Chilliwack/Harrison and the group of stocks originating from the South Thompson.
- Prior to this study they had identified the distribution of Cowichan River Chinook salmon as being predominantly around the Gulf Islands region through the summer months with very few juveniles captured outside of the Gulf Islands (Beamish et al. 2010). Further analysis of distribution patterns in this study is showing that other stocks also have relatively distinct distributions through September. Big Qualicum/Puntledge juveniles were predominantly captured along the east coast of Vancouver Island. South Thompson juvenile Chinook salmon were seen in low numbers in the early summer survey but small numbers appeared around Howe Sound. By September they were predominantly captured along the mainland coast from the Fraser River through Malaspina Strait. Cowichan Chinook salmon were predominantly captured within the Gulf Islands. For each of these stocks, about 80-90% of the catch of each stock were within these regions. Harrison/Chilliwack chinook salmon were captured in the southern SOG. However, the numbers captured were low (about 5% of all chinook salmon) and did not represent the 30-40% that would be expended based on adult escapements. These fish are captured in both the San Juan and Puget Sound (purse seine and trawl survey) indicating a southerly Salish Sea distribution for these fish. Detailed analysis on stock distribution is ongoing.
- Diet and condition are being examined by year and by stock group. In addition, analysis of fatty acid, stable isotopes and energy density is being conducted on the key stocks to examine differences in condition from ocean entry through the first fall, differences in condition between stocks and between years and differences between condition of fish that remain in natal rearing areas with fish that migrate out of these regions. In addition, differences between hatchery and wild fish are being examined for Cowichan Chinook salmon. To date, fatty acid and stable isotope analysis is completed for 2015 and 2016 and analysis of results is ongoing. Preliminary results from the fatty acid analysis indicate differences time in the ocean however for Cowichan and Big Qualicum Puntledge the levels were decreasing over time. Interpretation of these differences are ongoing and results will be available in the fall.
- Analysis of the diet of Chinook salmon is indicating that the majority of the diet is composed of 4-6 key diet items. However, there is variation in the proportion of these diet items by area within the Strait of Georgia and possibly by stock or size of fish. This analysis is currently in progress but results of the initial two years of study are anticipated to be complete by the end of June 2018. Sampling is designed to ensure we have a cross –section of results from all surveys by March 2018.
- Integration of diet and survival data with studies in Puget Sound continues to improve their understanding of variability between the basins.
- Otolith and scale analysis is ongoing with focus on fish that have been processed for energy density, fatty acid and stable isotope analysis. This cross analysis is expected to help in the interpretation of the resultsGIS mapping of fish catch, ocean conditions and food production across the Strait of Georgia to start informing on the interaction between these metrics. Over the coming year these map layers will be an important component as we move into spatial analysis of stock distribution, growth and condition.
- Initiation of work to incorporate their fish surveys with results from other programs has been undertaken. They are working with researchers from UBC, UVIC as well as other DFO staff to begin to examine the relationship of primary productivity through to the juvenile salmon growth and survival.
2. Habitat usage and early marine mortality studies of juvenile Chinook and coho
Team: BC Conservation Foundation (Kevin Pellet, James Craig, Craig Wightman) and Cowichan Tribes (Wayne Paige Jr.).
BC Conservation Foundation have carried out snorkel surveys and a Cowichan juvenile Chinook passive integrated transponder (PIT) tagging project to better understand:
- Migration timing, behavior, residence time, and survival of fry/fingerlings from mid and upper river spawning areas through lower river reaches and the estuary.
- Differences in distribution, growth, and relative abundance of wild and hatchery Chinook juveniles.
- The importance of the Cowichan estuary to Chinook smolt production and future adult returns.
- Relative survival of different groups of Chinook PIT- tagged in-river, in the estuary and in the Bay
Snorkel surveys in river and estuary were carried out during 2014 and are assisting with determining 1, 2 and 3 above. Question 4 is being answered by using PIT tagging methodology. Several previous research studies have attempted to investigate where mortality is occurring in the marine environment but few answers have resulted. The past approach has been to mark fish in freshwater and evaluate recapture rates in the marine environment. However, this has proven to be logistically and technologically challenging. BCCF’s work in the Cowichan provides a solution to many of these challenges by approaching the problem from a different direction with new technology. By applying PIT tags to fish captured in the marine environment at different times and locations, fewer tags are required as mortality rates decline over time. PIT tag technology allows a cost effective method for tagging many fish without constraints of battery life. Tag recoveries are made in freshwater as fish return to spawn which greatly reduces the need to scan a broad geographic range for marked fish. Finally, the individual codes for each tag allow an infinite range of marking locations, times, species, and sizes down to fish measuring only 65mm. If successful, this study could provide the biological foundation which identifies the critical time and location that controls survival in the marine environment. The application of PIT tags to studying marine survival of multiple tagged groups could become a central aspect of the SSMSP. If mortality in the early marine period is as high as believed, then sequentially tagged groups (identified by tagging time/location/species/size) should have substantially different mortality trajectories that could be estimated using these PIT tags. Results from 2014 showed that this method was successful and this project was continued in 2015-present
Hatchery chinook have been tagged each year 1. In hatchery, 2. In freshwater, 3. By beach seine, 4. By purse seine, 5. By microtrolling. Microtrolling (catching fish using specialized trolling equipment aimed at catching juveniles as they move offshore) will be used to capture sublegal juveniles in the marine environment. This is a novel method, piloted by Will Duguid, PhD candidate at UVic during 2014. The method is labour intensive, so it is proposed to achieve higher sample sizes by forming a collaboration with the “Avid Anglers” which will form a second citizen science project in the Strait of Georgia. Wild fish have also been tagged in all locations (except for in the hatchery).
The key hypotheses that are being addressed are as follows:
- The mortality rate in the early marine environment is initially high but subsides as fish grow and move offshore.
- Survival to the adult stage is controlled mostly by mortality at a specific time and location within the first year of marine life.
- The determinants of marine survival may be geographic and temporal. Fish of a similar size may experience significantly different survival based on when they occupy a specific habitat (early vs. late migrants).
- Understanding how mortality is distributed in the marine environment will lead to the identification of causal factors in conjunction with other research activities.
- The size distribution of fish tagged in freshwater matched that of recaptures in the marine environment. This suggests there is little evidence for size-selective mortality during in-river migration and the transition to salt water.
- Marine recaptures were disproportionally associated with fish tagged in the lower reaches of the river. The survival estimate for hatchery fish migrating from the release site to the lower river (46 km) was only 14%.
- Similarly, survival of wild fish was estimated at 27% over 41 km. These survival estimates are about 55% of those based on in-river detections suggesting that fish which come from upper river habitats experience disproportionally higher mortality rates in the early marine environment.
- These observations collectively place a higher value on smolts originating in the lower Cowichan River despite similar abundance throughout 40 km+ of habitat (BCCF snorkel data 2014).
- Micro trolling was proven to be effective for catching significant numbers of first and second year juvenile Chinook with limited bycatch. Catch per hour averaged 8 fish and was as high as 12. This has opened the door for studying juvenile Chinook in the 120 mm+ range using low-cost and highly mobile platforms. The utility of this tool should be discussed for other work, not just application of PIT tags.
- Sansum Narrows was identified as a “hot spot” for rearing juvenile Chinook and Coho. DNA results indicated this area is shared by many different stocks suggesting it may be a good candidate for fine-scale study. Areas with similar characteristics (i.e. tidal mixing) should be explored to see if juvenile Chinook/Coho are disproportionally abundant in these areas.
In-River Detections – Returning Tags
- 15 unique PIT tags were detected in fall 2015 returning Chinook. 9 were at the counting fence and 7 at Skutz Falls. Of the 7 at Skutz Falls, only 1 was detected at the fence. This suggested that about 63 tagged fish returned in 2015.
- Although the detection efficiency was disappointingly low, this indicated that the study design is sound and should produce a reliable estimate of marine survival by stage as the majority of returns are yet to come.
- Tag returns were distributed throughout each tagging stage including hatchery releases, in-river tagging, purse seining and mictro trolling.
A total of 45,202 PIT tags have been applied to juvenile Cowichan River Chinook since spring 2014. Tag application rates have steadily increased over the study period as have detection probabilities in-river. Activities in 2016 were highly successful as tag targets were met or exceeded at all locations. In-river detections were dramatically improved for out-migrating juveniles while over 130 returning tags were detected in the fall.
Steady improvements in micro trolling locations and gear types have solidified the feasibility of thi.s technique for targeting juvenile Cowichan River Chinook in the marine environment. Provided in-river detection systems are maintained, tag returns over the next 3-5 years are expected to provide the necessary data to satisfy project goals.
Tag application targets were met in the early marine environment (Cowichan Bay) providing resolution in space and time. Recaptures of fish tagged in freshwater were not a primary objective of marine sampling activities due to anticipated low capture rates. However, 70 tags were recovered in 2016 allowing for a comparison of recapture rate (% of tags released) by river km. Results were highly correlated with tag detections in the lower river and lend further support to conclusions regarding low in-river survival.
Capture rates aboard the purse seiner in Cowichan Bay were high in 2016 resulting in the application of 7,912 tags over 5 days compared to the pre-season target of 5,000 tags over 6 days. As a result, beach seining activities were curtailed at 1,980 tags over 7 days (pre-season target 3,150) to conserve tags for micro trolling work. The sixth purse seining day was rescheduled for August 21 in Sansum Narrows/Maple Bay to support micro trolling work. Accordingly, 300 additional tags were deployed resulting in a season total of 8,212 tags aboard the Ocean Venture.
Micro trolling techniques were further refined from 2015 in order to target a higher proportion of Cowichan River Chinook. A significant amount of effort was placed in Maple Bay (63% of total) where Cowichan fish were regularly found in 2015 while the rest occurred in northern/southern Sansum Narrows (37%). 2015 results indicated that Cowichan fish occupied the smaller size bins of all juvenile Chinook captured and that smaller fish were generally found closer to the surface. 2016 activities focused on shallower depths (6-21 m) than 2015 (15-30 m) while terminal tackle was scaled down to spoons measuring 2.0 cm from 4.5 cm. An initial batch of 196 samples was run within one week to allow for calibration of tagging procedures (i.e. avoid tagging non-Cowichan). 139 samples (71%) came back with Cowichan as the first stock ID while the cumulative probabilities of all Cowichan ID’s (more conservative estimate) was 68%. GSI work on the balance of DNA has yet to be conducted but estimate 1333 Cowichan fish were tagged during micro trolling/purse seining outside of the bay in 2016 (assuming a 68% Cowichan GSI). This exceeds the pre-season target of 1050 fish and the 2015 total of 431.
Returning Tags – Freshwater
Tag returns in 2015 were estimated at 63 based on a low detection efficiency of 14% in the lower river and 24% overall. An increased number of tags were expected to return in 2016 from multiple age classes in conjunction with improvements to detection equipment. In total, 134 returning tags were detected in 2016 with an estimated probability of detection exceeding 93%. The upstream sub-sample used to estimate detection efficiency was low at only 14 tags. This was a result of a large increase in flows during the migration period such that the upper threshold for fish using the primary fishway was exceeded on October 15. Abnormally high rainfall during October was observed and this result is not expected to be consistent in future years. Antenna performance within the fishway was excellent with each tag detected an average of 36 times (range 14-100).
1) Detection of tags inside of predators/scavengers (presumably raccoons based on trail camera photos) was unexpected as antennas were designed to scan for fish only. This opens the door to conducting future predation studies with terrestrial antennas.
2) Downstream migration survival was similarly low for both hatchery and wild Chinook based on lower tag detections and marine recaptures. Wild fish were expected to survive at a much higher rate based on the assumption hatchery fish would be disoriented upon release and be targeted by predators.
3) Early returns of age 2 and 3 tag groups suggest the marine survival bottle neck could exist at the age 1 micro troll stage (120-250 mm). The estimated marine survival of this group was expected to be 1050% compared to 2-10% for fish tagged in Cowichan Bay (see low/high end estimates in Figure 4 of the proposal) yet the current estimate is only 4.7%. This suggests that over 95% of the Cowichan Chinook tagged during micro trolling have yet to return or have perished. Age 3 returns in 2017 (from 2015 tagging) and age 2 returns (from 2016) will be highly informative relative to the objectives of this study.
3. Acoustic Tracking Studies
Team: Scott Hinch (UBC), Tony Farrell (UBC), Nathan Furey (UBC), Kristi Miller (DFO), Dave Patterson (DFO), Dave Welch and Erin Rechisky (Kintama Research).
The great number and variety of factors influencing the survival of migrating salmon smolts in the early marine environment makes it difficult to understand the causes of recent declines in smolt survival and salmon population productivity.
As part of the SSMSP, PSF has enhanced the use of acoustic tags and receivers within the Strait of Georgia. Until 2015, the existing acoustic receiver arrays allowed for fish detections in the lower Fraser River, in Juan de Fuca Strait, in north-central Strait of Georgia (NSoG), and in northern Queen Charlotte Strait only. While these arrays have been very useful, they could not provide sufficient resolution to assess the residence time; migration patterns, rate and timing; and survival of juvenile Pacific salmon within the Strait of Georgia.
To address these issues, PSF deployed 43 new acoustic VR4 receivers in the Discovery Islands (northern end of the Strait of Georgia/Salish Sea) and Johnstone Strait near Sayward, BC in 2015. 41 of these were loaned from OTN and the other two were loaned from Kintama. These are all dual array receivers and can pick up both 69KHz and 180KHz frequencies, which are emitted by larger Vemco V7 tags (69 kHz) and new V4 (180 kHz) tags, respectively. The V4 tag is half the size of the V7 and weighs only 0.24 gm in water. The smaller tag is preferred for juvenile salmon but the cost of this development is a reduced range of signal detection at 180 kHz.
The locations for deployment have been mapped by Kintama Research and are shown below.
These locations allow us to unambiguously measure individual juvenile fish migration rate, residence time and survival by specific areas within the Salish Sea and Discovery Islands. Calculation of survival rates requires a receiver array seaward of a tagging location and/or a previous array (i.e., the QCI array enables estimates of survival of tagged fish that pass over the new Johnstone Strait array and then are detected at the QCI array). The information on salmon migration behavior and survival provides direct evidence that can support other studies that infer residence times or survival based on observational studies.
Overall Acoustic Tagging Project Objectives
- Establish new acoustic receiver arrays in Discovery Islands and Johnstone Strait, and test detection efficiency of V9, V7, and V4 Vemco acoustic tags (focus of 2015 studies and use of Steelhead trout). COMPLETED 2015
- To evaluate the behavior (migration rates and patterns) and survival of individual juvenile salmon in the Strait of Georgia, the Discovery Islands, and Johnstone Strait (use of sockeye, steelhead, and coho salmon). 2015-2017. This work is being carried out by Scott Hinch and his group at UBC, in collaboration with Kintama Research. In 2015, due to high water making the Chilko fence inaccessible, Chilko sockeye could not be tagged. Instead, 243 Seymour steelhead were tagged by the Hinch et al. lab at UBC. 50 steelhead were used for a double tagging study carried out by Kintama. The fish were tagged with small V4 and larger V7 tags to determine efficacy of the smaller tags. During 2016, this group was able to tag 300 Chilko sockeye- and tagged both Age-1 (200) and Age-2 (100) fish. In addition during 2016, Kintama carried out a project to examine migration patterns and survival of Chilko chinook.
- For Fraser River sockeye salmon (Chilko Lake), directly assess the migration rate and route to assess exposure of the tagged fish to farm sites within the Discovery Island region. This work was carried out during 2016 by Scott Hinch and his group at UBC, in collaboration with Kintama Research. Brief results are presented below.
- Data by individual tagged fishes will be openly available to those supporting the Salish Sea project to enable more quantitative assessments of risks to juvenile salmon exposed by open-net pen salmon farming. These data will be placed into the Strait of Georgia Data Centre in 2017, and all detection data are sent to OTN.
- Results from 2015 assessments will facilitate array designs at finer scale resolution if necessary; e.g., specific to a farm site within the Discovery Island region. Array designs have been created by Kintama Research but would require additional funding to implement.
1. Kintama Telemetry Studies
Team: KINTAMA, Dr. David Welch, Dr. Erin Rechisky, Paul Winchell
The focus for 2015 was to establish and test the detection efficiency of the new arrays and the use of small V4 tags. As part of this project, Kintama deployed 38 receivers with 180 kHz capability to improve the NSoG line (spring 2015). Kintama double-tagged 50 steelhead smolts with V4 (180 kHz) & V9 (69 kHz) transmitters, and Hinch’s UBC group tagged 274 steelhead smolts with V7 tags (69 kHz). The former double tagging study was designed to comprehensively test the detection efficiency of the smaller tags.
The 69 KHz V9 transmitters had 100% detection efficiency at the NSOG and Discovery Islands subarrays. Detection was also 100% at the Johnstone Strait subarray although this estimate is preliminary. The V4 180 KHz transmitters performed very well at the Discovery Island subarray, showing a detection efficiency of 74%. Data are not yet available for the QCS arrays.
Kintama’s primary conclusion to date is that the V4 tags can achieve a similar level of detection as the VEMCO V7-2L tag which has generally been deemed satisfactory when using release groups of a few hundred smolts per year to achieve baseline survival estimates. If their estimates are applicable to other subarrays and salmonid stocks, then these tags can be used to obtain reasonably precise survival estimates for populations that were previously impossible to study because available tags were too heavy or because the array infrastructure was too costly.
The high detection efficiency of V4 tags opens up the possibility of carrying out tagging studies on numerous other salmonid stocks, and allows for tagging much smaller fish (down to 10g) than has been previously possible. The use of V4 tags and VR4 receivers could be expanded to examine the behavior of tagged juvenile salmon around specific farm locations. Residence times and migration patterns of tagged salmon could be a significant contribution to risk assessments of salmon exposed to open-net salmon farms.
In 2016, Kintama conducted a small acoustic telemetry pilot study on juvenile Chilko River Chinook with the following objectives:
- Provide freshwater survival estimates for Chilko River yearling Chinook smolts from release in the Chilko River to the lower Fraser River, and compare this to past published data for other species and populations originating from the Fraser River.
- For Chilko River Chinook smolts that migrate north after ocean entry during the hypothesized timeframe, fill in the critical uncertainty regarding residence time of upper Fraser River Chinook in the Strait of Georgia.
- For Chilko River Chinook smolts that migrate north after ocean entry during the hypothesized timeframe, provide estimated survival in the Strait of Georgia and Discovery Islands and compare to past published data (Welch et al. 2009; Chittenden et al. 2010; Balfry et al. 2011; Welch et al. 2011; Melnychuk et al. 2013; Neville et al. 2015; Clark et al. in press).
Marine survival of Fraser River Chinook salmon stocks has decreased to ~1% in recent decades and lack of information on downstream and early marine survival hampers their effective management. As part of an effort to establish an indicator stock for management of middle Fraser River summer run 52 Chinook salmon, the Pacific Salmon Commission provided funding to develop a coded wire tag (CWT) indicator project for Chilko River Chinook in 2014. In February 2016, Brian Riddell (PSF) approached Kintama and DFO Salmonid Enhancement Program (SEP) managers to conduct a telemetry study on hatchery reared smolts from this program which were going to be released in April 2016.
Thus, Kintama had a unique opportunity to acoustic tag and monitor these fish as they migrate downstream and into the ocean. They took advantage of this opportunity for several reasons: 1) there is a paucity of information on Fraser River Chinook migratory behavior, residence time, and early marine survival in the Strait of Georgia, 2) few populations of mid-upper Fraser River yearling Chinook salmon are presently accessible for tagging, 3) acoustic tags small enough to be used in this population are now available for use and can be detected with reasonable efficiency on acoustic sub-arrays in the Discovery Islands and Johnstone Strait, and 4) this study will provide information on post-release and downstream mortality of the 52 yearling summer Chinook run.
They successfully conducted a small acoustic telemetry pilot study on hatchery-origin Chilko River Chinook, using 100 acoustic-tagged smolts to 1) estimate freshwater survival, 2) investigate residence timing in the Strait of Georgia, and 3) begin to investigate early marine survival. Because acoustic receiver arrays capable of detecting smolts implanted with small 180 kHz acoustic tags only monitor the northern exit from the Strait of Georgia (SOG), residence time and early marine survival could only be potentially estimated if smolts migrated north before tag batteries expired five months after ocean entry.
Freshwater survival of acoustic-tagged Chinook to the Fraser River mouth (49%) was comparable to other populations or species which migrate the same distance downstream; however, their downstream migration rate (only 18 km/day) was dramatically slower than that of wild Chilko Lake sockeye, which migrate rapidly to the ocean after exit from Chilko Lake (100-170 km/day). Thus, Chinook smolts took more than one month on average to reach the SOG, in contrast to wild Chilko Lake sockeye which generally take under a week. It is unknown whether this behavioural difference is the result of their hatchery origin and transport to Chilko Lake.
Only one fish was subsequently detected in the SOG and none were detected exiting the SOG. Combined with the results from trawl surveys, the complete lack of detections in the Discovery Islands and Johnstone Strait suggest that Chilko Chinook do not migrate directly north after river exit. Instead, they likely remain in the SOG for at least several months.
With only a single fish detected in the SOG, they were not able to estimate early-marine survival or residence time. It is unclear if smolts eventually exited the Strait via the southern route, died during their summer residence, or simply ceased migration to reside in the SOG. This uncertainly can be resolved by either 1) increasing the number of tagged smolts released and instrumenting the southern exit from the SOG with acoustic receivers capable of detecting 180 kHz tags, or 2) capturing larger juvenile salmon in the SOG later in the season (by microtrolling or seine) and implanting them with longer-lived, low-frequency tags that are compatible with all of the receivers in the greater Salish Sea area.
2.Physiological and environmental factors affecting the migratory behaviour and survival of sockeye and steelhead salmon smolts
Team: Scott Hinch (University of British Columbia), Tony Farrell (University of British Columbia), Kristi Miller (Fisheries and Oceans Canada), and Steve Cooke (Carleton University)
This team will combine novel methodologies (biotelemetry, biomarkers, simulation models, etc.) simultaneously to examine a variety of factors influencing the migratory behaviour and survival of sockeye and steelhead smolts in the Salish Sea.
Overall, this team has combined novel methodologies (biotelemetry, biomarkers, simulation models, etc.) simultaneously to examine a variety of factors influencing the migratory behaviour and survival of sockeye and steelhead smolts in the Salish Sea. Using small acoustic transmitters, the behaviour and fate of Chilko sockeye (2016 and 2017) and Seymour steelhead (2015) smolts are tracked from release through the Salish Sea. The condition of these same smolts will be assessed prior to transmitter implantation and release through the use of biomarkers for pathogen presence and load, presence of immune- or stress-related responses, and growth potential, to better understand the links between condition during initial outmigration and survival and behaviour in the early marine environment.
Retrospective analyses on a large database generated from ~10 years of acoustic telemetry studies will determine how migratory behaviour and survival are influenced by oceanographic conditions recorded in the Salish Sea. Lastly, individual-based models (IBM) will be developed to simulate smolt migrations by combining various movement behaviours of smolts with fine-scale ocean simulation models to provide potential migratory pathways through the Salish Sea, and a means of testing what navigation and/or orientation behaviours smolts use in the early marine environment. Together these studies will help to provide a mechanistic understanding of salmon smolt migrations to better understand trends in productivity and survival.
- Seymour steelhead: In May 2015, 243 age-one hatchery steelhead smolts were tagged at the Seymour River Hatchery in North Vancouver and released at two locations to test route- and location-specific survival across outmigration. The early marine period was associated with poor overall survival (9-27% over 400km), and the Seymour River and Burrard Inlet were noted as regions of particularly low survival for migrating smolts. New acoustic receivers around the Discovery Islands revealed smolts showed higher use of Discovery Passage, with survival for this route being ~two times as high compared to channels to the east. Migration rates became faster and more variable through the Discovery Islands to Johnstone strait region, and complex milling patterns were noted in this region, suggesting the influence of strong currents impacting migration movements through this part of the Salish Sea. Lab work for genomics of non-lethal gill biopsies was completed in the Spring 2016, and analyses are currently in the initial phases. 162 non-lethal gill biopsies were taken from tagged Seymour River steelhead smolts in 2015, and these were processed using genomic techniques at Dr. Kristi Miller’s lab at the Pacific Biological Station (DFO) in Nanaimo in spring 2016.
Please see http://kintama.com/animator/dep/Seymour_sthd_combined_2015/ for an animation of the Seymour steelhead tagged during 2015.
Chilko sockeye: In April/May 2016, 300 Chilko sockeye salmon smolts were tagged with acoustic transmitters. 200 age-one smolts were tagged (for the first time) with Vemco V4 transmitters. An additional 100 age-two smolts were tagged with V7 transmitters. In addition, non-lethal gill biopsies were taken from 100 age-one smolts and 89 age-two smolts. Gill clips taken from Chilko sockeye salmon smolts are currently being processed at Dr. Kristi Miller’s lab, and will be completed by early 2017.
Chilko sockeye: In April/May 2016, 300 Chilko sockeye salmon smolts were tagged with acoustic transmitters. 200 age-one smolts were tagged (for the first time) with Vemco V4 transmitters. An additional 100 age-two smolts were tagged with V7 transmitters. In addition, non-lethal gill biopsies were taken from 100 age-one smolts and 89 age-two smolts. Gill clips taken from Chilko sockeye salmon smolts are currently being processed at Dr. Kristi Miller’s lab, and will be completed by early 2017.
This group just recently were given access to detections of these smolts on marine arrays in the Salish Sea, including new arrays in the Discovery Islands and Johnstone Strait. As these data are new, all results are preliminary.
Preliminary results from 2016 Chilko sockeye showed similar patterns of mortality in age-1 and age-2 smolts, though the larger smolts appeared to exhibit slightly higher mortality for the first 80km before they entered the Fraser River. It is not known why this occurred, but could be because the larger smolts are more conspicuous to predators. The patterns of mortality were similar in each, with higher mortality in upper Chilko and Chilcotin Rivers (which appear to be a potential mortality hotspot), lower mortality in the more turbid waters of the Fraser River, and then higher mortality again when smolts initially enter the Strait of Georgia. The data for 2016 was similar to the entire time series for the telemetry results from studies that have been carried out on Chilko sockeye since 2010.
In addition, it (at least qualitatively) appears that sockeye preferentially use Discovery Passage on their way to Johnstone Strait. Migration data are currently being analyzed to assess the influence of smolt size and age, microbes and biomarkers, and timing on migration success. Please see http://kintama.com/animator/dep/Chilko2016_sockeye/ for an animation of the movements of these tagged Chilko sockeye during 2016.
In summary, it appears that migration route was found to influence subsequent survival, particularly for steelhead, indicating that area encountered is an important component of the migration experience. For sockeye, both outmigration timing (particularly Chilko sockeye) and residency were correlated with survival. For neither sockeye nor steelhead did size influence survival, albeit this might be due to the small size range of smolts tagged.
The group were surprised by the proportion of both species (20-40% of sockeye and 30-50% of steelhead) that exhibited westward movements in the Strait of Georgia, evidence of “milling” by fairly large contingents of both species. Quantifying use of the Salish Sea and migration routes through it can help to identify factors influencing survival, and by using telemetry they are able to partition survival into smaller geographic areas, and thereby reducing uncertainty in the role of the marine environment.
During 2017, this group will be continuing tagging studies with Chilko sockeye.