Puntledge River Juvenile Chinook Salmon Freshwater Outmigration and Survival Study
Atkinson, J.B., James, S., & Guimond, E.
2025-08-07
Purpose
To address key data gaps identified in recovery planning for Puntledge summer Chinook, a freshwater survival study was initiated as part of the larger Bottlenecks Project in partnership with the Department of Fisheries and Oceans Salmon Enhancement Program (DFO-SEP). The study focuses on assessing juvenile Chinook mortality during downstream migration in relation to hydroelectric infrastructure and habitat conditions.
Paired groups of PIT-tagged fall-run Chinook juveniles (as a proxy for summer Chinook) were released upstream and downstream of the BC Hydro diversion dam to estimate survival and migration rates through this critical reach. This approach is informed by prior PIT-based research in other systems (e.g., Cowichan River) that has shown the effectiveness of this method in identifying areas of loss during emigration.
The study employs multiple in-river and hatchery-based PIT antenna and arrays, including a main permanent array at the Puntledge RV Park and temporary floating antennas deployed upstream (Figure. 1). Three trials were conducted over the course of a week with 800 PIT tagged juveniles released both above and below the diversion dam (N = 4,800 Chinook) to ensure adequate statistical power for detecting differences in survival. Power analyses have demonstrated that this sample size can confidently detect a 5% difference in survival between groups under moderate survival scenarios (40–60%).
Environmental parameters and flow data are also monitored during release windows, which are timed to avoid peak spill events. This enhances consistency across trials and supports evaluation of passage-related impacts, including potential predation or injury linked to diversion dam structures.
Puntledge River
Puntledge River summer-run Chinook salmon are of significant conservation concern. This population was assessed as Endangered by the Committee on the Status of Endangered Wildlife in Canada (COSEWIC) in 2020 and is currently under consideration for listing under the Species at Risk Act. This designation highlights the importance of understanding factors that affect freshwater productivity and survival within the watershed.
The Puntledge River has been extensively modified by hydroelectric infrastructure. The Comox Dam, constructed in 1953, impounds Comox Lake and diverts water to a powerhouse via a 5 km-long penstock, bypassing a significant section of the river. A downstream diversion dam, equipped with a fishway and outfall, regulates residual flow through the lower river reach. These structures are owned and operated by BC Hydro and are capable of generating approximately 24 MW of power.
While the dams have enabled long-term energy production and water management, they have also introduced barriers to fish passage, altered natural flow patterns, and contributed to habitat fragmentation. Juvenile salmon migrating downstream must pass through or around these structures, and may experience injury, predation, or delayed migration associated with altered hydraulic conditions.
The upper watershed is composed primarily of forested Crown land, with significant areas managed under Tree Farm License 39 (TimberWest), as such, it has been heavily logged. Logging activities, both historical and ongoing, have contributed to significant changes in sediment transport, reduced riparian cover, increased surface runoff, decreased water storage in the summer months and increased water temperatures. These land use changes likely negatively affect stream habitat complexity and water quality, particularly in headwater tributaries (Comox, Cruickshank, Reese, Eric, and Upper Puntledge creeks).
Legacy mining activities, including early 20th-century copper and coal extraction, have also left persistent water quality issues in certain tributaries. While most active mining has ceased, abandoned sites continue to contribute to localized acid drainage and elevated metal concentrations in stream sediments.
The lower watershed includes the urban centers of Courtenay and Comox, where residential development, road networks, and recreational use place additional pressures on the river corridor. Comox Lake is a popular destination for boating and camping, and the Puntledge River itself supports a wide range of recreational uses including swimming, tubing, and angling.
Although municipal wastewater is treated, increasing population density in the Comox Valley Regional District has led to growing concerns regarding stormwater runoff, nutrient loading, and shoreline development. Recreational pressures and land use intensification present ongoing challenges for fish habitat and water quality.
The Puntledge watershed lies within the asserted territories of the K’ómoks First Nation and is historically and presently used by several other Indigenous communities including the Hupacasath, Qualicum, We Wai Kai, and Wei Wai Kum Nations. The watershed supports traditional activities such as fishing, cultural ceremonies, and harvesting of plant and animal resources. Collaborative efforts with K’ómoks First Nation are essential to the success of the Bottlenecks Project and for developing sustainable salmon populations.
Puntledge River Hatchery
The Fisheries and Oceans Canada Puntledge River Hatchery has been the cornerstone of Chinook salmon enhancement in the watershed. While the hatchery has been effective at preventing extirpation, natural-origin fish currently make up a small proportion of returning adults. Recognizing the importance of maintaining adaptive traits through natural reproduction, hatchery operations have evolved to support strategies such as altered release timing, reduced in-hatchery mortality, and more refined broodstock management. The hatchery also supports PIT tagging and recapture infrastructure crucial for conducting survival studies under the Bottlenecks Project.
Puntledge River Study Area
The Puntledge River watershed is situated on the eastern slope of Vancouver Island, British Columbia. It drains approximately 586 km² of mountainous terrain, originating from the eastern boundary of Strathcona Provincial Park and flowing southeast into Comox Lake. The Puntledge River flows from Comox Lake to its confluence with the Tsolum River in Courtenay, forming the Courtenay River, which ultimately discharges into the Strait of Georgia.
The watershed exhibits a coastal hydrological regime characterized by winter rainfall, spring snowmelt-driven freshet, and reduced baseflows during summer and early fall. The river supports a diverse suite of anadromous and resident fish species, including Chinook, Coho, Chum, Sockeye, and Pink salmon, as well as steelhead, coastal cutthroat trout, and kokanee.
Puntledge River Mainstem PIT Array Infrastructure for Juvenile Outmigration
The Puntledge River PIT array system consists of one full-river-width and one partial river width mainstem array (701 and 702; see Table 1). The array is comprised of HDPE antennas (Biomark 30-ft units) installed in series across the river’s mainstem. A custom built antenna is also installed in the hatchery bypass fishway (71). A temporary floating antenna (72) was installed ~300 m upstream of the Puntledge Hatchery fence, to increase detections for this emigration study. Additionally, two custom made HDPE antennas (73 and 74) were installed in the Diversion Dam Fishway, located just downstream of the treated release location to determine if fish utilized the fishway during outmigration. The antennas were installed to capture outmigrating PIT-tagged juvenile Chinook and coho and to provide reach-specific survival metrics as part of the Bottlenecks Program.
| Location Code | Location Name | River Kilometer |
|---|---|---|
| 74 | Diversion Fishway Antenna | 14.0 |
| 73 | Diversion Fishway Antenna | 14.0 |
| 72 | Temporary Floating Antenna | 7.2 |
| 71 | Hatchery Fishway Bypass Antenna | 7.0 |
| 702 | Mainstem Array | 3.8 |
| 701 | Mainstem Array | 3.8 |
Detection Efficiency Calculations
Detection efficiency (DE) is calculated by comparing the number of unique tagged fish detected on each transect. To ensure the accuracy of DE estimates, survival is not assumed to be 100% between transects. Instead, detection overlap is used to estimate the likelihood of missed tags, and subsequent calculations follow methods described in Connolly et al. (2008), which were developed for multi-antenna river systems like the Puntledge.
Results (Preliminary)
Puntledge River Juvenile Outmigration Detections
Detections of tagged salmon movements by release group and antenna location within the Puntledge River are shown below (Table 2 & Table 3; Figure 2).These are movements of fish filtered to one distinct detection per day. PIT tag detections from juvenile Chinook released upstream and downstream of the BC Hydro diversion dam provide insight into freshwater migration routes and reach-specific survival.
Here we classify those fish released above the dam as our “treatment” group. Treated release group 1 had the highest number of detections across the system, including 60 detections at antenna 73 and 55 at antenna 74, both located upstream of the BC Hydro diversion infrastructure. This contrasts with the control groups, which had no detections at upstream antennas (73 or 74), consistent with their expected downstream-only migration pattern. All groups were detected at antennas below the diversion dam (antennas 71, 72, 701, 702), with consistently higher detection rates at antennas 72 and 702.
To truly investigate the impacts of damn infrastructure on the outmigration survival and timing (behaviour) of the treatment group, fish which were detected migrating out through the fishway were removed from comparative analysis between the control and treated groups, and are only compared to the behaviour of the treatment released fish which most likely migrated through the penstock (eicher screens) or over the spillway at the end of this report.
These data confirm successful migration and detection through multiple antenna locations, supporting reach-specific evaluation of freshwater survival.
| Release Group | Date and Time of Release | Total Release (N) | Detected | 74 | 73 | 72 | 71 | 702 | 701 |
|---|---|---|---|---|---|---|---|---|---|
| control 1 | 2025-05-30 | 790 | 237 | 2 | 2 | 170 | 1 | 61 | 18 |
| control 2 | 2025-06-03 | 789 | 145 | 0 | 0 | 112 | 0 | 37 | 4 |
| control 3 | 2025-06-06 | 793 | 133 | 1 | 1 | 83 | 0 | 46 | 7 |
| treated 1 | 2025-05-30 | 797 | 279 | 56 | 61 | 184 | 2 | 47 | 15 |
| treated 2 | 2025-06-03 | 793 | 120 | 13 | 12 | 72 | 2 | 35 | 6 |
| treated 3 | 2025-06-06 | 791 | 138 | 21 | 25 | 65 | 0 | 41 | 11 |
| Release Group | Date and Time of Release | Total Release (N) | Detected | 72 | 71 | 702 | 701 |
|---|---|---|---|---|---|---|---|
| control 1 | 2025-05-30 | 788 | 235 | 170 | 1 | 61 | 18 |
| control 2 | 2025-06-03 | 789 | 145 | 112 | 0 | 37 | 4 |
| control 3 | 2025-06-06 | 792 | 132 | 83 | 0 | 46 | 7 |
| treated 1 | 2025-05-30 | 722 | 204 | 163 | 2 | 42 | 14 |
| treated 2 | 2025-06-03 | 776 | 103 | 69 | 2 | 34 | 6 |
| treated 3 | 2025-06-06 | 759 | 106 | 63 | 0 | 38 | 11 |
Puntledge River Discharge
The six tagging groups were released on three different dates, and each release faced a distinct flow regime. The first release-control 1 and treated 1, on 30 May 2025-occurred under the lowest discharge observed, about 23.5 m^3/s at release with a five-day post-release average of roughly 26.9 m^3/s. The second release-control 2 and treated 2, on 3 June 2025-experienced the highest flows, with an instantaneous discharge of 36.6 m^3/s, an increase of 55.7%, and a five-day mean of about 35.5 m^3/s. The third release-control 3 and treated 3, on 6 June 2025-was conducted under flows slightly lower than those of release 2 but still much higher than those of release 1; discharge at release was around 35.1 m^3/s, an increase of 49% from the first release, and the subsequent five-day average was about 34.1 m^3/s.
Figure 2. Puntledge River Discharge. Shaded areas indicate night-time hours.
Outmigration Timing and Duration
Summary plots were developed analyzing the time of day when fish were actively outmigrating and the duration of outmigration (Figures 2-6). Migration timing plots illustrate both diel and multi-day migration patterns across groups.
A large proportion (51%) of detections occurred at night (10PM-6AM) across all release groups (Figure 3).
Figure 3. Timing of all juvenile Chinook detections at PIT antennas 72, 701, and 702 in the Puntledge River. Shaded areas indicate night-time hours.
Mean time from release to first detection at the temporary floating array (72; approximately 7 km downstream) ranged from 28.1 hours (control 3) to 54.2 hours (treated 1; Table 3; Figure 3). Control groups consistently exhibited shorter and more compressed migration durations compared to treated groups. The standard deviation (SD) of migration times was notably high in some groups, Trial 1 (control SD = 42.5, treatment SD = 38.3 hrs), indicating substantial variation among individual migrations.
The time from detection at the floating array to arrival at the downstream mainstem array (701/702; a distance of 3.4 km) further supports this variability. Treated groups tended to have longer and more variable migration durations, with treated 3 showing a mean travel time of 16.1 hours, though no standard deviation was available for this group (data are still being collected). In contrast, control groups ranged from 6.6 to 9.3 hours, with relatively lower variation, suggesting faster and more consistent downstream movement once past the diversion dam.
Thus, there is some evidence that encountering and moving through the diversion dam could be slowing downstream travel speeds, with effects seen all the way to ocean entry.
The difference between release groups is most pronounced in Trial 1 under low flow conditions [insert flow]. Trials 2 and 3 experienced slightly higher flows and outmigration speeds between release groups were more similar.
| Release Group | N Released | n Detected at 72 | n Detected at 701/702 | Mean Time to 72 (hrs) | SD Time to 72 (hrs) | Mean Time 72 to 701/702 (hrs) | SD Time 72 to 701/702 (hrs) |
|---|---|---|---|---|---|---|---|
| control 1 | 788 | 170 | 79 | 40.4 | 50.4 | 6.7 | 7.1 |
| control 2 | 789 | 112 | 41 | 43.6 | 44.0 | 11.1 | 8.8 |
| control 3 | 792 | 83 | 53 | 30.8 | 19.1 | 27.8 | 42.4 |
| treated 1 | 722 | 163 | 56 | 59.0 | 47.4 | 17.2 | 53.6 |
| treated 2 | 776 | 69 | 40 | 54.8 | 38.7 | 7.8 | 11.3 |
| treated 3 | 759 | 63 | 46 | 43.4 | 23.8 | 32.3 | 35.9 |
Figure 4. Duration of the juvenile Puntledge River Chinook outmigration from release above/below the diversion dam to antenna 72 (river KM 7.2).
Figure 5. Duration of the juvenile Puntledge River Chinook outmigration from release above/below the diversion dam to the mainstem array (river KM 3.8).
Figure 6. Puntledge River Chinook Juvenile Mean Freshwater Outmigration Timing (hours) from release to 701/702.
| Unique Tags Detected | Detection Efficiency | |
|---|---|---|
| Puntledge Floating Antenna 72 | 660 | 0.17 |
| Puntledge Mainstem Array Upstream 702 | 258 | 0.02 |
| Puntledge Mainstem Array Downstream 701 | 60 | 0.06 |
| Puntledge 2025 system | 921 | 0.24 |
Preliminary Survival Results
To assess the effect of treatment on juvenile Chinook freshwater survival, a binomial generalized linear model (GLM) was fitted with survival (Survived2; 1 = detected at freshwater array, 0 = not detected) as the response variable and tagging group as the predictor. Fish detected at downstream arrays (loc_code 73 and 74) were excluded from the analysis to focus on in-river survival outcomes.
Survival varied significantly among tagging groups (p < 0.001). The baseline survival probability for the control 1 group was approximately 29%. Both control 2 and control 3 exhibited significantly lower odds of survival relative to control 1 (log-odds estimates: -0.63 and -0.76 respectively; both p < 0.001).
Among treated groups:
Treated 1 did not differ significantly in survival from Control 1 (Estimate = -0.21, SE = 0.11, p = 0.056). While not statistically significant, this result suggests a potential negative effect of treatment.
In contrast, Treated 2 and Treated 3 showed significantly lower survival probabilities compared to control 1 (Estimates = -1.06 and -1.06 respectively; both p < 0.001), indicating a negative effect of treatment on freshwater survival in these cohorts.
The model explained approximately 2.6% of the deviance (McFadden’s pseudo R = 0.026), suggesting that while tagging group contributed meaningfully to variation in survival, other factors likely influenced survival outcomes as well.
These preliminary findings indicate that treated fish in two of the three treatment groups experienced significantly lower freshwater survival compared to controls, supporting the hypothesis that the treatment may negatively affect early outmigration success under certain conditions.
Figure 7. Time-stepped cumulative detections of juvenile Puntledge River Chinook from release to detection at the floating antenna (72).
Preliminary Survival Analysis
A logistic regression model was used to evaluate differences in freshwater survival among six tagging groups of juvenile Chinook salmon. Survival was defined as detection at downstream PIT antennas (71, 72, 702, or 701). Fish detected at upstream antennas (73 or 74) were fully excluded from the analysis to avoid potential confounding effects from atypical upstream movement.
Paired Survival Comparisons by Release Timing
To account for changing environmental conditions across trials, freshwater survival was assessed using three separate logistic regression models comparing survival between control and treated fish for each release period (Trial 1, 2, and 3).
Trial 1: Control 1 vs. Treated 1
The survival difference between control 1 and treated 1 was not statistically significant, although the model showed a trend toward lower survival for treated fish. The estimated coefficient for the treated group was -0.21 (p = 0.056), suggesting a ~5% decrease in survival probability, but with insufficient statistical evidence to confirm a true difference.
Trial 2: Control 2 vs. Treated 2
A significant reduction in survival was observed for treated 2 fish relative to control 2. The estimated coefficient for the treated group was -0.42 (p = 0.0031), representing a 27.0% lower survival probability for treated fish.
Trial 3: Control 3 vs. Treated 3
Results were consistent with Trial 2. The treated group had a significantly lower survival rate, with an estimated coefficient of -0.29 (p = 0.039), reflecting a 25.4% reduction in survival probability relative to the control group.
Results suggest that both control and treated groups released later in the study experienced lower survival, potentially due to environmental changes (e.g., temperature, flow conditions, or predator dynamics). Notably, treated 1 had lower but more comparable survival to its control group, whereas the later treated groups showed significantly lower performance. This reinforces the importance of both release timing and location in influencing freshwater survival of juvenile salmon.
Figure 8. Preliminary Juvenile Chinook Outmigration Survival Results.
Comparison of Treated Fish Detected Outmigrating Through the Upper Fishway Compared to Those which Migrated Through the Penstock (Eicher Screens) or over the Spillway.
A portion of the treated Chinook were detected migrating downstream through the impoundment dam fishway (locations 73 and 74). Fish passing through the fishway may exhibit different survival probabilities and migration behaviours compared to treated Chinook that likely exited via the Eicher screens or over the dam spillway. In total, 124 treated Chinook were detected using the fishway, the majority originating from treated release group 1. For treated release groups 2 and 3, physical release locations were positioned farther from the left bank and closer to the river’s center to better mimic a more natural migration pathway.Table 6 provides an summary of detections for each treated release group for fish detected outmgirating through the fishway and their subsequent detections.
| Release Group | Date and Time of Release | Total Release (N) | Detected | 74 | 73 | 72 | 71 | 702 | 701 |
|---|---|---|---|---|---|---|---|---|---|
| control 1 | 2025-05-30 | 2 | 2 | 2 | 2 | 0 | 0 | 0 | 0 |
| control 3 | 2025-06-06 | 1 | 1 | 1 | 1 | 0 | 0 | 0 | 0 |
| treated 1 | 2025-05-30 | 75 | 75 | 56 | 61 | 21 | 0 | 5 | 1 |
| treated 2 | 2025-06-03 | 17 | 17 | 13 | 12 | 3 | 0 | 1 | 0 |
| treated 3 | 2025-06-06 | 32 | 32 | 21 | 25 | 2 | 0 | 3 | 0 |
Fishway array detection efficiency was calculated to be 67% and 75% efficient for antennas 73 and 74. Subsequent detections of fish downstream were low, due to the high and variable water levels throuhgout the study, and the alteration of antenna location 72, shortly after the treated 1 release, due to B.C. Hydro operations.
Figure 9. Preliminary Juvenile Chinook Outmigration Survival Results for treated fish that outmigrated through the fishway vs those that did not.
Preliminary Survival and Migration Timing Analysis
A logistic regression model was used to assess differences in freshwater survival between treated juvenile Chinook salmon that were detected at upstream antennas (73 or 74) and those that were not detected at these sites. Survival was defined as detection at any of the lower river PIT antennas (72, 701, or 702).
Survival Comparison – Detected vs. Not Detected at 73/74
The model indicated that fish not detected at antennas 73 or 74 had a significantly lower probability of surviving to the lower river relative to those that were detected at these upstream sites. The estimated coefficient for the Not Detected group was -0.53 (p = 0.0108), corresponding to a lower odds of survival by approximately 41% compared to the Detected group. This suggests that detection at 73/74 may be associated with higher survival probability to downstream arrays, potentially reflecting differences in migration path, speed, or condition.
Migration Timing – Detected vs. Not Detected at 73/74
A linear model was used to compare the number of days between release and first detection at the lower river between the two groups. On average, fish not detected at 73/74 arrived at the lower river 0.93 days later than those detected upstream; however, this difference was only marginally significant (p = 0.0679). The effect size suggests a small but biologically relevant delay that could contribute to differences in survival outcomes, though further investigation with larger sample sizes or finer temporal resolution is warranted.
Next Steps
Results presented above are preliminary and should not be utilized in any meaningful discussions regarding management implications. Further analysis on outmigration speed and duration and survival are required. Future survival analysis will need to include the drastic and unexpected changes in flow rates that occurred between each release date as well as an understanding of changes in outmigration corridors which likely impacted Treated Group 3. The final survival analysis will utilize an integrated freshwater residence and capture-recapture model using Bayesian inference developed with Bottlenecks partners (Balfour et al. 2025).
Due to these preliminary results, further research is needed to fully understand the implications for B.C. Hydro infrastructure. This study used large (5 gram) fall Chinook smolts during a low flow regime. Future studies should utilize summer Chinook, at a smaller size (>3 grams) to better replicate the impacts on wild fish that spawn above the diversion dam, particularly in upstream habitats.