Quinsam River Wild Steelhead: Cohort Structure, Outmigration Timing, and Observed Adult Return
Jamieson Atkinson
March 06, 2026
1 Introduction
This report summarizes cohort structure, size structure, downstream travel timing, and observed adult return for wild Quinsam River steelhead PIT-tagged through the Bottlenecks to Survival program. The analysis is framed around the June–July 2021 heat dome as a plausible ecological disturbance that may have affected juvenile steelhead growth, survival, and later smolt production.
The Quinsam River Wolf Trap provides a strong platform for annual comparison because a hydraulic fish fence funnels approximately one-third of the river discharge into the trap. Under a consistent capture method, changes in the number and structure of tagged smolts are therefore interpreted as reflecting real biological variation among years rather than large shifts in trapping efficiency.
Because scale ages were not available for this analysis, fork-length classes are treated as provisional age proxies rather than true ages. They are still useful for comparing cohort structure among years and for assessing whether the population shifted toward relatively younger or older smolts.
2 Methods
2.1 Field program
Wild juvenile steelhead were captured annually in the Quinsam River Wolf Trap, measured for fork length and weight, PIT-tagged, and released. The Bottlenecks to Survival study uses 12 mm PIT tags and downstream detection arrays to examine movement, freshwater survival, and apparent adult return.
2.2 Analytical approach
The analytical workflow underlying this report:
- deduplicated tagging records to one deployment row per tag,
- summarized fork length and weight by outmigration year,
- grouped smolts into literature-informed provisional size classes,
- compared age/size structure among years using raw counts and within-year probabilities,
- summarized downstream travel timing to focal antennas using detections restricted to 0–100 days post-tagging, and
- summarized apparent adult return using detections occurring at least 200 days post-tagging.
Adult return is treated here as observed return probability, not detection-corrected survival. PIT antenna performance was limited until August 2023, so the earliest returning adults from the 2021 and 2022 smolt cohorts were likely under-detected.
3 Results
3.1 Smolt abundance by outmigration year
| Year | Number tagged |
|---|---|
| 2021 | 479 |
| 2022 | 338 |
| 2023 | 130 |
| 2024 | 236 |
| 2025 | 802 |
Smolt abundance varied strongly among years. The 2023 outmigration cohort was the smallest observed in the series, with only 130 tagged smolts, compared with 338 in 2022 and 236 in 2024. The 2025 outmigration year was by far the largest, with 802 tagged fish. This pattern indicates marked cohort-to-cohort variability in freshwater production.
3.2 Fork length among outmigration years
| Year | n_fish | mean_fl | sd_fl | se_fl | cv_fl | median_fl | min_fl | max_fl | q25_fl | q75_fl | iqr_fl |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 2021 | 479 | 184.0 | 22.7 | 1.0 | 0.1 | 181.0 | 122.0 | 463.0 | 170.0 | 195.0 | 25.0 |
| 2022 | 338 | 186.7 | 16.8 | 0.9 | 0.1 | 186.0 | 142.0 | 291.0 | 174.0 | 197.0 | 23.0 |
| 2023 | 130 | 173.2 | 18.4 | 1.6 | 0.1 | 173.5 | 116.0 | 220.0 | 165.0 | 185.0 | 20.0 |
| 2024 | 236 | 176.7 | 18.7 | 1.2 | 0.1 | 176.0 | 120.0 | 299.0 | 165.0 | 186.2 | 21.2 |
| 2025 | 802 | 171.0 | 17.2 | 0.6 | 0.1 | 170.0 | 100.0 | 245.0 | 160.0 | 180.0 | 20.0 |
Fork length distributions among outmigration years.
Mean fork length by outmigration year.
Fork length varied among years, but the most biologically meaningful pattern was not simply a change in mean length. Rather, the strongest signal was a change in size-class composition, especially in 2023.
3.3 Outmigration age structure
3.3.1 Outmigration size structure expressed as probabilities
| Outmigration size structure expressed as probabilities | |||||||
| Year | 120-129 mm (transition / small smolt) | 130-159 mm (age-2/3 overlap) | 160-194 mm (mostly larger smolts; age-3-dominant) | 195-229 mm (very large smolts; age-3/4 overlap) | >=230 mm (possible resident / atypically large) | NA | <120 mm (parr / likely age-1+) |
|---|---|---|---|---|---|---|---|
| 2021 | 0.00 | 0.07 | 0.67 | 0.24 | 0.02 | 0.00 | 0.00 |
| 2022 | 0.00 | 0.04 | 0.66 | 0.30 | 0.01 | 0.00 | 0.00 |
| 2023 | 0.02 | 0.13 | 0.72 | 0.12 | 0.00 | 0.00 | 0.02 |
| 2024 | 0.01 | 0.11 | 0.77 | 0.10 | 0.01 | 0.00 | 0.00 |
| 2025 | 0.00 | 0.24 | 0.68 | 0.06 | 0.01 | 0.00 | 0.00 |
3.3.2 Outmigration size structure expressed as raw counts
| Outmigration size structure expressed as raw counts | |||||||
| Year | 120-129 mm (transition / small smolt) | 130-159 mm (age-2/3 overlap) | 160-194 mm (mostly larger smolts; age-3-dominant) | 195-229 mm (very large smolts; age-3/4 overlap) | >=230 mm (possible resident / atypically large) | NA | <120 mm (parr / likely age-1+) |
|---|---|---|---|---|---|---|---|
| 2021 | 1 | 33 | 323 | 113 | 9 | 0 | 0 |
| 2022 | 0 | 12 | 222 | 101 | 2 | 1 | 0 |
| 2023 | 3 | 17 | 93 | 15 | 0 | 0 | 2 |
| 2024 | 2 | 26 | 181 | 24 | 3 | 0 | 0 |
| 2025 | 1 | 196 | 545 | 52 | 6 | 0 | 2 |
Outmigration size structure expressed as within-year proportions.
Outmigration size structure expressed as raw counts.
Outmigration structure changed meaningfully among years. The 2023 smolt cohort was not only numerically small, but also appeared shifted toward larger provisional age/size classes. In contrast, 2022 had a relatively strong outmigration despite occurring immediately after the 2021 heat dome. This pattern suggests that older juveniles present during the heat dome largely survived to migrate in 2022, whereas younger juveniles that would have contributed strongly to the 2023 cohort were disproportionately affected.
Taken together, the 2022 and 2023 cohorts suggest age- or size-selective filtering associated with the heat dome. In biological terms, the event appears more consistent with loss or delayed development of younger juveniles than with a uniform reduction across all age classes.
3.4 Weight among outmigration years
| Year | n_fish | mean_weight | sd_weight | se_weight | cv_weight | median_weight | min_weight | max_weight | q25_weight | q75_weight | iqr_weight |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 2021 | 328 | 57.1 | 20.0 | 1.1 | 0.3 | 54.0 | 13.9 | 266.8 | 44.7 | 67.6 | 22.9 |
| 2022 | 337 | 61.8 | 15.8 | 0.9 | 0.3 | 59.9 | 26.8 | 117.1 | 50.1 | 71.3 | 21.1 |
| 2023 | 130 | 49.3 | 14.0 | 1.2 | 0.3 | 49.3 | 15.6 | 95.8 | 41.0 | 57.8 | 16.7 |
| 2024 | 236 | 52.0 | 15.0 | 1.0 | 0.3 | 50.0 | 27.0 | 120.0 | 41.6 | 59.7 | 18.1 |
| 2025 | 1 | 15.5 | NA | NA | NA | 15.5 | 15.5 | 15.5 | 15.5 | 15.5 | 0.0 |
Weight distributions among outmigration years.
Mean weight by outmigration year.
Weight varied among cohorts and broadly tracked the patterns observed in fork length and size structure. These shifts are consistent with year-to-year differences in growth opportunity, cohort composition, or both.
3.5 Downstream travel timing
Travel timing was restricted to detections occurring within 100 days post-tagging, so this section reflects downstream movement rather than delayed detections or adult returns. Adult returns were analyzed separately and were defined as detections occurring at least 200 days post-tagging.
| Year | loc_code_chr | n | mean_days | median_days | sd_days | min_days | max_days | q25_days | q75_days |
|---|---|---|---|---|---|---|---|---|---|
| 2023 | 91 | 45 | 1.24 | 1.00 | 1.05 | 0.00 | 5.00 | 1.00 | 1.00 |
| 2024 | 921 | 128 | 1.19 | 1.00 | 2.88 | 0.00 | 31.00 | 0.00 | 2.00 |
| 2024 | 922 | 159 | 1.08 | 1.00 | 2.59 | 0.00 | 31.00 | 0.00 | 1.00 |
| 2025 | 91 | 356 | 2.29 | 2.00 | 1.89 | 0.00 | 12.00 | 1.00 | 3.00 |
| 2025 | 921 | 471 | 0.58 | 0.00 | 1.13 | 0.00 | 18.00 | 0.00 | 1.00 |
| 2025 | 922 | 622 | 0.52 | 0.00 | 0.95 | 0.00 | 13.00 | 0.00 | 1.00 |
Travel time from tagging to first downstream detection at focal loc_codes, restricted to detections within 100 days post-tagging.
Downstream travel timing varied among years, though interpretation should remain cautious until the detection timeline is fully synchronized with antenna performance and the cleaned travel table is confirmed against the final script outputs.
3.6 Apparent adult return
| Year | Tagged | Returns | sar | mean_days_to_return | median_days_to_return | lower | upper | SAR (%) | Lower 95% CI | Upper 95% CI |
|---|---|---|---|---|---|---|---|---|---|---|
| 2021 | 479 | 2 | 0.004175365 | 971.0000 | 971 | 0.0005060567 | 0.015000821 | 0.42 | 0.05 | 1.50 |
| 2022 | 338 | 9 | 0.026627219 | 826.7778 | 937 | 0.0122463911 | 0.049943334 | 2.66 | 1.22 | 4.99 |
| 2023 | 130 | 1 | 0.007692308 | 710.0000 | 710 | 0.0001947334 | 0.042112739 | 0.77 | 0.02 | 4.21 |
| 2024 | 236 | 7 | 0.029661017 | 440.2857 | 583 | 0.0120067497 | 0.060155510 | 2.97 | 1.20 | 6.02 |
| 2025 | 802 | 0 | 0.000000000 | NA | NA | 0.0000000000 | 0.004589038 | 0.00 | 0.00 | 0.46 |
Observed adult return probability by outmigration year.
Observed smolt-to-adult return varied substantially among cohorts. Based on the current table outputs, the 2021 cohort had the lowest observed SAR (0.42%; 2 returns from 479 smolts), 2022 was much higher (2.66%; 9 returns from 338 smolts), 2023 was lower again (0.77%; 1 return from 130 smolts), and 2024 was the highest among currently interpretable cohorts (2.97%; 7 returns from 236 smolts). The 2025 cohort has not yet had sufficient time to contribute meaningfully to adult returns.
The statistical outputs associated with SAR indicated strong heterogeneity among cohorts in the current analysis, but the earliest cohorts should be treated cautiously because PIT antennas were not performing well until August 2023.
3.7 Size structure of adult returns
3.7.1 Size structure of adult returns expressed as probabilities
| Adult return size structure expressed as probabilities | |||
| Year | 160-194 mm (mostly larger smolts; age-3-dominant) | 195-229 mm (very large smolts; age-3/4 overlap) | >=230 mm (possible resident / atypically large) |
|---|---|---|---|
| 2021 | 0.50 | 0.50 | 0.00 |
| 2022 | 0.67 | 0.33 | 0.00 |
| 2023 | 1.00 | 0.00 | 0.00 |
| 2024 | 0.57 | 0.29 | 0.14 |
3.7.2 Size structure of adult returns expressed as raw counts
| Adult return size structure expressed as raw counts | |||
| Year | 160-194 mm (mostly larger smolts; age-3-dominant) | 195-229 mm (very large smolts; age-3/4 overlap) | >=230 mm (possible resident / atypically large) |
|---|---|---|---|
| 2021 | 1 | 1 | 0 |
| 2022 | 6 | 3 | 0 |
| 2023 | 1 | 0 | 0 |
| 2024 | 4 | 2 | 1 |
Size structure of adult returns by outmigration year, shown as proportions.
Size structure of adult returns by outmigration year, shown as raw counts.
There are too few adult returns to interpret shifts in return age structure with much confidence. The available return data are still useful descriptively, but not yet strong enough to make firm inferences about whether the age structure of returning adults changed among cohorts.
3.8 Return rate by tagged size class
| Year | Provisional age/size class | Tagged | Returns | return_rate | lower | upper | Return rate (%) | Lower 95% CI | Upper 95% CI |
|---|---|---|---|---|---|---|---|---|---|
| 2021 | 120-129 mm (transition / small smolt) | 1 | 0 | 0.000000000 | 0.000000e+00 | 0.975000000 | 0.00 | 0.00 | 97.50 |
| 2021 | 130-159 mm (age-2/3 overlap) | 33 | 0 | 0.000000000 | 0.000000e+00 | 0.105762810 | 0.00 | 0.00 | 10.58 |
| 2021 | 160-194 mm (mostly larger smolts; age-3-dominant) | 323 | 1 | 0.003095975 | 7.838023e-05 | 0.017128090 | 0.31 | 0.01 | 1.71 |
| 2021 | 195-229 mm (very large smolts; age-3/4 overlap) | 113 | 1 | 0.008849558 | 2.240263e-04 | 0.048320649 | 0.88 | 0.02 | 4.83 |
| 2021 | >=230 mm (possible resident / atypically large) | 9 | 0 | 0.000000000 | 0.000000e+00 | 0.336267117 | 0.00 | 0.00 | 33.63 |
| 2022 | 130-159 mm (age-2/3 overlap) | 12 | 0 | 0.000000000 | 0.000000e+00 | 0.264648469 | 0.00 | 0.00 | 26.46 |
| 2022 | 160-194 mm (mostly larger smolts; age-3-dominant) | 222 | 6 | 0.027027027 | 9.981657e-03 | 0.057895285 | 2.70 | 1.00 | 5.79 |
| 2022 | 195-229 mm (very large smolts; age-3/4 overlap) | 101 | 3 | 0.029702970 | 6.167860e-03 | 0.084356898 | 2.97 | 0.62 | 8.44 |
| 2022 | >=230 mm (possible resident / atypically large) | 2 | 0 | 0.000000000 | 0.000000e+00 | 0.841886117 | 0.00 | 0.00 | 84.19 |
| 2022 | NA | 1 | 0 | 0.000000000 | 0.000000e+00 | 0.975000000 | 0.00 | 0.00 | 97.50 |
| 2023 | <120 mm (parr / likely age-1+) | 2 | 0 | 0.000000000 | 0.000000e+00 | 0.841886117 | 0.00 | 0.00 | 84.19 |
| 2023 | 120-129 mm (transition / small smolt) | 3 | 0 | 0.000000000 | 0.000000e+00 | 0.707598226 | 0.00 | 0.00 | 70.76 |
| 2023 | 130-159 mm (age-2/3 overlap) | 17 | 0 | 0.000000000 | 0.000000e+00 | 0.195064323 | 0.00 | 0.00 | 19.51 |
| 2023 | 160-194 mm (mostly larger smolts; age-3-dominant) | 93 | 1 | 0.010752688 | 2.721974e-04 | 0.058458165 | 1.08 | 0.03 | 5.85 |
| 2023 | 195-229 mm (very large smolts; age-3/4 overlap) | 15 | 0 | 0.000000000 | 0.000000e+00 | 0.218019361 | 0.00 | 0.00 | 21.80 |
| 2024 | 120-129 mm (transition / small smolt) | 2 | 0 | 0.000000000 | 0.000000e+00 | 0.841886117 | 0.00 | 0.00 | 84.19 |
| 2024 | 130-159 mm (age-2/3 overlap) | 26 | 0 | 0.000000000 | 0.000000e+00 | 0.132274604 | 0.00 | 0.00 | 13.23 |
| 2024 | 160-194 mm (mostly larger smolts; age-3-dominant) | 181 | 4 | 0.022099448 | 6.053421e-03 | 0.055614109 | 2.21 | 0.61 | 5.56 |
| 2024 | 195-229 mm (very large smolts; age-3/4 overlap) | 24 | 2 | 0.083333333 | 1.025634e-02 | 0.269972802 | 8.33 | 1.03 | 27.00 |
| 2024 | >=230 mm (possible resident / atypically large) | 3 | 1 | 0.333333333 | 8.403759e-03 | 0.905700676 | 33.33 | 0.84 | 90.57 |
| 2025 | <120 mm (parr / likely age-1+) | 2 | 0 | 0.000000000 | 0.000000e+00 | 0.841886117 | 0.00 | 0.00 | 84.19 |
| 2025 | 120-129 mm (transition / small smolt) | 1 | 0 | 0.000000000 | 0.000000e+00 | 0.975000000 | 0.00 | 0.00 | 97.50 |
| 2025 | 130-159 mm (age-2/3 overlap) | 196 | 0 | 0.000000000 | 0.000000e+00 | 0.018644808 | 0.00 | 0.00 | 1.86 |
| 2025 | 160-194 mm (mostly larger smolts; age-3-dominant) | 545 | 0 | 0.000000000 | 0.000000e+00 | 0.006745731 | 0.00 | 0.00 | 0.67 |
| 2025 | 195-229 mm (very large smolts; age-3/4 overlap) | 52 | 0 | 0.000000000 | 0.000000e+00 | 0.068482209 | 0.00 | 0.00 | 6.85 |
| 2025 | >=230 mm (possible resident / atypically large) | 6 | 0 | 0.000000000 | 0.000000e+00 | 0.459258126 | 0.00 | 0.00 | 45.93 |
Observed adult return rate by outmigration year and provisional size class.
Return rate by tagged size class is potentially informative, but sample sizes are still small. This part of the analysis is therefore best interpreted as exploratory.
4 Discussion
4.1 The 2021 heat dome as a cohort-filtering event
The most compelling interpretation of the Quinsam dataset is that the June–July 2021 heat dome acted as a cohort-filtering event rather than producing a uniform reduction across all fish. The contrast between 2022 and 2023 is particularly informative. The 2022 outmigration year remained reasonably strong, suggesting that the older juvenile cohort present during the heat dome largely survived and migrated successfully the following spring. By contrast, the 2023 outmigration year was sharply reduced and showed a shift toward larger provisional age classes, suggesting that younger juveniles were more strongly affected.
This interpretation is consistent with steelhead life history. Smolts migrating in 2022 would have been older juveniles during the heat dome, whereas fish migrating in 2023 would have included individuals that were younger and likely more vulnerable to extreme thermal stress, reduced feeding opportunity, habitat compression, and density-dependent carry-over effects. In that sense, the 2023 cohort appears to reflect both reduced abundance and altered cohort structure.
4.2 Why 2023 appears different
The 2023 cohort stands out for three linked reasons:
- it had the lowest smolt abundance observed in the time series,
- it showed an apparent shift in outmigration size/age structure, and
- it exhibited low apparent adult return.
A plausible mechanism is that the heat dome caused mortality and growth suppression among younger juveniles during summer 2021. Some of the smallest or youngest fish may have died outright, while others may have grown too slowly to smolt on schedule and instead delayed outmigration. That would produce exactly the pattern observed here: a weak 2023 outmigration year dominated by relatively larger smolts.
4.3 Apparent adult return and the antenna limitation
Observed SAR differed among cohorts, but those values must be interpreted carefully. PIT antennas were not functioning well until August 2023, so early returns from the 2021 and 2022 cohorts were likely under-detected. For that reason, the SAR estimates for those cohorts are best treated as minimum values.
Even with that caveat, the 2023 cohort still appears biologically important because it combines low smolt production with low apparent return. This suggests that the same processes affecting freshwater survival and age structure may also have had carry-over effects into the marine phase.
4.4 What can and cannot be said about return age structure
The current data are not yet strong enough to make confident statements about whether the age structure of returning adults changed among cohorts. There are too few returns in several years, and the antenna limitations further weaken inference for the earliest cohorts. At present, the strongest conclusions lie in the smolt abundance, outmigration age structure, and cohort-specific survival pattern, especially for 2023.
4.5 Implications for warming climates
This study provides a rare opportunity to evaluate how an extreme climate event may propagate through a wild steelhead population. The patterns observed in Quinsam suggest that severe thermal events can influence not only juvenile abundance, but also the age structure of the smolt population and possibly later survival. As heat waves become more frequent under climate change, understanding these cohort-level effects will be essential for interpreting future changes in steelhead productivity and resilience.