Study Design and Budget Attachment_WFC.pdf

Study Plan

Field Protocol

With the modified trap deployed in 2021 at the Clifton Channel study site, WFC and partners will evaluate the ability of the gear to target hatchery fishes in-river while reducing bycatch mortality of ESA-listed wild salmonids. Results will be summarized in a report by WFC and partners for peer and regional review, which may ultimately inform fish trap management and implementation through the ongoing emerging fishery process (RCW 77.65.400).

The fall season study will be conducted from August 15 through October 31, 2021 during the upriver migration period for adult fall Chinook, coho, and steelhead in the lower Columbia

River (Healey 1991). Testing will proceed in the following manner. Four individuals will be needed to operate the gear and record data, including 3 trained WFC employees, a commercial fisher, and/or a potential volunteer. When fishing, the trap pots will be deployed from their suspended positions above the water column to the river bottom and tunnel doors opened to initiate the soak period. Observers will note the set time, tidal stage, tide height (m), water temperature (ºC), and presence of marine mammals.

Trap operators will monitor the pots and upstream live wells to determine fish entrance and occupancy. Once a live well chamber is occupied by one or more fishes, study investigators will passively trap the catch through closure of the capture door. If fishes cannot be captured passively, they may be dip netted from the pot and the capture method noted. Fishes captured through any means will be enumerated, measured (FL), noted for capture/release conditions

(WDFW 1-5 scale; Cox et al. 2019), and identified by species and adipose fin mark status. Fish showing no signs of life at capture or release from any compartment of the trap—as observed from the live well dock and daily underwater video surveys of all deployed trap panels

(Splashcam Deep Blue HD)—will be considered an immediate mortality and potentially donated to a local food bank. Non-lethal 2 mm caudal fin genetic samples will be secured from all

Chinook and steelhead encountered and processed by the WDFW Molecular Genetics Lab.

Furthermore, all salmonids will be scanned for Passive Integrated Transponder (PIT)-tags with a

Biomark HPR Lite reader and tag data recorded through P4 software (PTAGIS 2019). After all data are collected, live salmonids will be released to resume the upriver migration (unless criteria are met for inclusion in mark-recapture or net pen holding studies described below). Fish will be captured and released through this means over each research day. These methods will enable assessment of the following metrics within the final report: genetic stock-composition, stock-specific bycatch encounters, capture/release conditions, immediate survival, CPUE, and hatchery stock exploitation.

Mark Release-Recapture Experiment for Steelhead and Chinook

The final study design for this project may be subject to further revision with input from

WDFW, ODFW, TAC representatives, and the University of WA (UW). Given agency feedback, WFC and partners currently plan to pursue paired mark release-recapture studies to estimate and evaluate the following metrics: post-release survival of wild Chinook and steelhead bycatch, migration delay effects, and cumulative commercial capture effects from multiple fish trap encounters. In general, our methodology will compare rates of upriver fish detection, survival, and migration timing between a control group (unexposed to commercial capture processes) and two treatment groups (exposed to commercial capture, tagging, and release processes) of

Chinook and steelhead to Columbia River mainstem dam detection points.

For the paired mark release-recapture study, all adult fall Chinook (> 570 mm) and steelhead (> 500 mm) encountered in the trap pot / live well and documented through the standardized field protocol will be PIT-tagged with a 12.5-mm, 134.2-kHz, full-duplex PIT tag using an MK-25 Rapid Implant Gun (Biomark). These fish will then be scanned to record PIT-tag information, after which they will be released by hand into an adjacent temporary mesh holding chamber (3.00 × 0.65 × 0.85 m) with continuously circulating river water. PIT-tagged

Chinook and steelhead randomly encountered with the gear will be alternately assigned to two separate treatment groups throughout each sampling day: 1) single exposure, and 2) double exposure. The single exposure group will represent adult Chinook and steelhead that have been exposed to the commercial fish trap capture and release process once, and will serve to estimate the release survival effect relative to a control group unexposed to the commercial capture and release process. The double exposure group will represent adult Chinook and steelhead that have been exposed to commercial fish trap capture and release processes twice, and will serve to enable investigation of 1) migration delay effects from the gear; and 2) cumulative fish trap capture and release effects on fish survival. Comparison of relative survival between the double and single exposure groups may also serve to isolate the post-release survival effect from the second fish trap exposure.

As previously stated, tagged fishes that have been alternately assigned to single and double exposure groups will be released into a temporary holding chamber. The chamber will be equipped with two separate sub-compartments (1.5 × 0.65 × 0.85 m), one for each respective treatment group. Once a sample of five fishes is achieved for each treatment group (or fishes have been held for the maximum duration of 30 minutes), the holding chamber will be drifted with the current to a marked location in the channel near the outside of the heart compartment.

The single exposure treatment group will be passively released via the exit door of the holding chamber for upriver detection at mainstem dam and other Columbia Basin PIT-tag arrays (total group sample size targets: 1000 Chinook, 300 steelhead). At release, the time will be noted by investigators. The remaining fish in the holding chamber assigned to the double exposure group will be released back into the heart of the modified trap for a repeated, second exposure to fish trap commercial capture and release procedures. Once again, the time of release will be noted by investigators for this group. Through the standard field protocol, Chinook and steelhead tagged and assigned to the double exposure treatment group will be re-detected in a live well. If an individual from the double exposure group is re-detected at the trap, tag data will be recorded, and the fish will be passively released from the live well for upriver detection at Columbia Basin

PIT-tag arrays (total group sample size targets: 800 Chinook, 240 steelhead). Those assigned to the double exposure treatment group that escape the trap and are not detected for a second time at the trap will be disqualified from the standard survival analysis described in the “Statistical

Analysis” section below.

A separate control group of Chinook and steelhead unexposed to commercial fish trap operations (and previously PIT-tagged as juveniles in either Upper Columbia or Snake River tributaries) will be passively detected between mid-August and November 2021 at the lower

Columbia River PD-7 array PIT tag detection site, located near rkm 70. PD-7 is the only array available near the project site to estimate survival below Bonneville Dam. In analysis, migration timing and detections of control group fishes may be compared to that of treatment group fishes to estimate migration delay effects and the relative post-release survival effect of both single and double exposure to fish trap gear capture and release processes. Using this control group, we would expect relative survival results to be negatively biased as PD-7 detected fishes are not handled/tagged in the adult life-history stage; since only the treatment groups are handled/tagged in the adult life-history stage, only the treatment groups will be subject to this added research mortality effect and relative survival results for the fish trap should be biased low (Pollock and

Pine 2007).

An additional control group of Chinook and steelhead may be handled, tagged, and released in the adult life-history stage at the Bonneville Dam Adult Fish Passage Facility (AFF).

Assuming PIT-tagging occurs at the AFF between mid-August and September 2021, migration timing and detections of AFF tagged control group fishes will be compared to that of treatment group fishes in upper river reaches above Bonneville Dam to estimate migration delay effects and the relative post-release survival effect of both single and double exposure to fish trap gear capture and release processes. Using this control group for river reaches above Bonneville Dam, we would expect relative survival results above Bonneville Dam to be positively biased as short-term handling and tagging mortality effects for the control group would be detectable while treatment group fishes suffering from short-term handling/tagging research effects would have either perished before Bonneville Dam or partially recovered from the handling/tagging process.

Given the inherent biases of each control group described above, analysis of the relative survival effect between the double exposure and single exposure treatment groups may provide the most unbiased and precise estimate of the isolated post-release survival effect of the fishing gear. Since these treatment groups are handled and tagged in an equivalent manner and assignment / release of each group is alternated throughout each fishing day, use of the single exposure group as a control group will enable unbiased isolation of the post-release survival effect from the second fish trap exposure.

Mark Release-Recapture Post-Release Survival Statistical Analysis

A pair of Cormack (1964) single release-recapture models will be used to estimate relative bycatch survival of the single and double exposure treatment groups of Chinook and steelhead to the selected control group (τ = STreatment/SControl)—as well as the double exposure treatment group to the single exposure treatment group (τ = SDouble.Treatment/SSingle.Treatment)—in upriver reaches between the fish trap site (rkm 55), Bonneville Dam (rkm 234), The Dalles Dam

(rkm 309), and McNary Dam (rkm 470) on the Columbia River (Figure A-1). Above McNary

Dam, the joint probability of survival and detection to upper basin detection sites will also be estimated.

The joint likelihood model for the tag analysis is described in detail in Tuohy et al.

(2019) and is known as the complete capture history protocol by Burnham et al. (1987:112–125).

With four mainstem upstream detection sites, as well as numerous tributary detection sites, the control and treatment releases will each produce 24 = 16 unique capture histories described in their respective likelihoods. While reach survival of the control group will be parameterized by Si

(i = 1, ∙∙∙, 3), reach survival of treatment groups will be parameterized by Si * τi (i = 1, ∙∙∙, 3) in the joint likelihood model. In so doing, near and far field effects of the fish trap on survival of trapped and released Chinook and steelhead (τi) can be directly estimated from the model (τi =

Si.Treatment/Si.Control; alternatively, τi = SDouble.Treatment/SSingle.Treatment may be a preferred or used to confirm relative survival results using the PD-7 or AFF control). Immediate survival (τ0) from capture to the time of release from the gear will be observed directly. Cumulative relative survival from release to McNary Dam will be estimated by the product τ0 * τ1 * τ2 * τ3.

Figure A-1. Map of the Columbia River and mainstem dam PIT-tag arrays.

Detection histories for control and treatment groups will be downloaded from the PIT-tag

Information System (PTAGIS), which provides public access to all PIT-tag detection data throughout the Columbia Basin. The data will be uploaded to Program USER (User Specified

Estimation Routine) to calculate maximum likelihood estimates of reach survival as described in

Skalski and Millspaugh (2006). Likelihood ratio tests (LRT) will be performed to identify the most parsimonious models for describing the capture process at α = 0.05 two-tailed.

Power analysis has been conducted to estimate the precision of the single / double exposure study (in which τ = SDouble.Treatment/SSingle.Treatment). Coupling sample sizes targeted in this study (Single Exposure: 1000 Chinook, 300 steelhead; Double Exposure: 800 Chinook, 240 steelhead) with projected detection rates for pooled groups at McNary Dam arrays (Chinook:

0.300; steelhead: 0.725; Gayeski et al. 2020), we expect standard error (SE) for release survival estimates to McNary Dam to be roughly SE = 0.05 for steelhead and SE = 0.07 for Chinook salmon. Alternatively, survival estimates to Bonneville Dam may be preferred. In this case, we expect SE for release survival estimates to McNary Dam to be roughly SE = 0.04 for steelhead and SE = 0.04 for Chinook salmon given projected detection rates for pooled groups at

Bonneville Dam arrays (Chinook: 0.600; steelhead: 0.825; Gayeski et al. 2020). Nevertheless, precision will depend upon various factors, including species-specific returns to the Columbia

River, stock-specific capture efficiency of the gear, and upriver fishing and predation factors which may impact the sample size achieved and tag detections at upriver arrays.

For this mark release-recapture study, it must be noted that genetic stock assignment techniques will be used to ensure the model assumption of equivalent stock composition is achieved between control and treatment groups (Tuohy et al. 2019; Cox and Sippel 2020). In comparative analysis to the PD-7 control group, only Chinook and steelhead that assign to Snake

River and Upper Columbia populations through parent-based tagging (PBT) and genetic stock identification (GSI) will be used. Since PD-7 control group fishes are tagged as juveniles in either Upper Columbia or Snake River tributaries, this genetic strategy will eliminate potential biases that may result from non-detections of tagged treatment group fishes destined for lower basin tributaries and hatcheries below Bonneville Dam where effective PIT-tag arrays are unavailable (Cox and Sippel 2020). However, for comparison of the double and single exposure treatment groups (τi = SSingle.Treatment/SDouble.Treatment), all tagged fishes will be used in the relative survival analysis as the law of large numbers, random sampling, and alternated assignment should ensure equivalent stock-composition between single and double exposure treatment groups. Despite the high likelihood of equivalence, generalized linear modeling based on a log-link and Poisson error structure will be used to test the null-hypothesis of homogeneity of

Chinook and steelhead population assignment to single and double treatment groups at the α ≤

0.05 significance level (Tuohy et al. 2019).

Migration Delay Analysis

TAC members have expressed interest in evaluating potential migration delay effects from the fish trap gear that could potentially affect spawning success of released bycatch. To this end, the double exposure group will function, in part, to assess the mean period of fish migration delay within the gear. Currently, it is assumed based upon anecdotal WFC / WDFW observational evidence that the gear may delay fish migrations up to the ~6-h length of a tide-cycle (when fish have been observed clearing from the trap during slack tide). Comparison of the double exposure release and re-detection times will result in a robust point estimate of the mean migration delay effect in the fish trap gear and calculation of the 95% confidence interval (Zar

2010). Furthermore, ANOVA analysis of all treatment and control group upriver detection timings to mainstem dams will test the null hypothesis of equivalence between mean group migration durations at the α ≤ 0.05 significance level (Zar 2010).

References

Buchanan, S., P. Farell, J. Fraser, P. Gallaugher, R. Joy, and R. Roughtledge. 2002. Reducing gill-net mortality of incidentally caught Coho Salmon. North American Journal of Fisheries

Management 22(4):1270–1275.

Burnham, K. P, D. R. Anderson, G. C. White, C. Brownie, and K. H. Pollock. 1987. Design and analysis methods for fish survival experiments based on release-recapture. American Fisheries

Society, Monograph 5. Bethesda, Maryland.

Cormack, R. 1964. Estimates of survival from the sighting of marked animals. Biometrics

51:429–438.

Cox, B., Wadsworth, T., and J. Holowatz. 2019. Short-term survival of fall Chinook and coho salmon captured by purse seines in the lower Columbia River, 2017: A holding study.

Washington Department of Fish and Wildlife. Olympia, WA.

Cox, B., and T. Sippel. 2020. Pound net and PD7 2017 survival analysis: recommended release mortality rates for pound net gear. Washington Department of Fish and Wildlife. Olympia, WA.

Gayeski, N. J., Tuohy, A. M., and A. Jorgenson. 2020. Evaluation of pound nets as stock-selective fishing tools in the lower Columbia River Sub-basin. Report to the National Oceanic and Atmospheric Administration Fisheries Service Bycatch Reduction Engineering Program.

Silver Spring, Maryland. available from

<https://wildfishconservancy.org/BREPFinalReport_WildFishConservancy_March2020.pdf >.

Healey, M. 1991. The life history of Chinook salmon. In C. Groot and L. Margolis (eds), Life history of Pacific salmon, p. 311-393. UBC Press. Vancouver, BC.

Holowatz, J., M. Zimmerman, A. Stephenson, D. Rawding, K. Ryding, E. Kinne. 2014. Lower

Columbia River alternative commercial fishing gear mortality study: 2011 and 2012.

Washington Department of Fish and Wildlife. Olympia, WA.

Pollock, K.H., Pine III, W.E., 2007. The design and analysis of field studies to estimate catch-and-release mortality. Fish. Manage. Ecol. 14, 123–130.

Skalski, J. R. 1981. Statistical inconsistencies in the use of no-observed-effect levels in toxicity testing. Pages 377–387 in D. R. Branson and K. L. Dickson, editors. Proceedings of the Fourth

Annual Symposium on Aquatic Toxicology. Special Technical Publication, American Society for Testing and Materials 737, Philadelphia.

Skalski, J. R., and J. J. Millspaugh. 2006. Application of multidimensional change-in-ratio methods using program USER. Wildlife Society Bulletin 34:433-439.

Takata, H., and A. Johnson. 2018. Post-release mortality of coho salmon captured with tangle nets in the lower Columbia River. Oregon Department of Fish and Wildlife. Salem, OR.

Tuohy, A. M., Skalski, J. R., and N. J. Gayeski. 2019. Survival of salmonids from an experimental commercial fish trap. Fisheries 44(6).

Tuohy, A. M., Skalski, J. R., and A. Jorgenson. 2020. Modified commercial fish trap to help eliminate salmonid bycatch mortality. Accepted for publication by the North American Journal of Fisheries Management.

Zar, J. H. 2010. Biostatistical Analysis, 5th Edition. Pearson Publishing. London, England.

Budget

Category Budget

Salary $42,000

Travel $5,000

Supplies $3,000

Total $50,000