In our preliminary review, we requested a response on the three topics listed below. Our final comments based on the response are provided after each topic.

1. SMART Objectives. Objective 2 of Goal 3 was modified as we suggested to include the list of actions required to generate pre-season and in-season forecasts of run size, and the creel survey required to estimate harvest. We are disappointed that the actions subsumed by revised Objective 2 still do not explicitly specify quantitative rules for controlling harvest. However, the proponents’ response to Topic #2 adequately explains the process for setting and adjusting harvest regulations. Note that Action 3 is missing the word “survey” after “tribal creel.” In future work plans and proposals, we recommend including the PNI target as part of Objective 2 and specifying that Actions 1-3 will be conducted annually. 
   
   
2. Methods

• Generating and evaluating forecast models and predictor variables

As requested, the proponents provided a copy of the original unpublished report by West. Inc. (Griswold et al. 2009) as well as their annual progress report for 2020. Together these documents provide a detailed account of (and sufficient justification for) the method being used to generate preseason forecasts of adult returns. Statistical “bootstrap” procedures are used to estimate confidence intervals for the abundance forecast associated with each regression model based on variability in the historic data series. However, decisions about which regression models to include for evaluation each year seem a bit ad hoc. The annual report points out (on page 30): “We also continue to investigate a wide range of predictor variables that display correlation to Hood River Chinook returns and may be used to improve our forecasts,” and (on page 32): “A lesson we have learned is that prediction models are not static. To maintain a good statistical fit and effective predictive value, we must continue to assess model performance and explore alternative predictor variables.” For these reasons, and because 12 years of additional data are available since Griswold et al.’s 2009 report, we urge the proponents to summarize and compare (in a future report) the retrospective performance of their alternative models and predictor variable sets. We suggest retrospectively fitting each candidate model in each year of the entire time series (for which calculation was possible) to compare how the candidate models would have actually performed had they been used in each case. This method of retrospective analysis provides a robust evaluation of any forecasting procedure (Haeseker, S.L., R.M. Peterman, S. Zhenming, and C.C. Wood. 2011. Retrospective Evaluation of Preseason Forecasting Models for Sockeye and Chum Salmon. North American Journal of Fisheries Management, 28:12-29).

• Setting and adjusting harvest regulations based on run size forecasts.

The response adequately addresses our concerns by providing a detailed (2-page) description of the decision pathway for setting harvest seasons and fishery regulations for spring Chinook salmon. Relevant parts of that text should be included in future proposals and reports to improve clarity and to complete the documentation of methods used in the project.

3. Evaluation of productivity and recolonization.
The response acknowledges ISRP concerns about the uncertain future of monitoring in the Hood River and whether information will be available to effectively assess productivity, progress towards recolonization, and relationships among fish stocks. The proponents also explain that after their original proposal had been submitted, ODFW decided to completely defund the monitoring activities previously conducted by the ODFW Hood River M&E project 198805304, some of which had been expected to continue in some capacity under the ODFW Hood River O&M project 198805308.

To address these new and unexpected challenges in monitoring, the proponents worked with a consultant to develop an alternative study plan that relies heavily on genetic analyses. They submitted this new monitoring plan, entitled “Winter Steelhead and Spring Chinook Population Monitoring in Response to Restoration Measures in the Hood River, Oregon” as part of their response to the ISRP. The ISRP considers the new monitoring plan to be well designed, and a significant improvement to that proposed in the original proposal. The statistical analyses for evaluating fish responses to hatchery releases and habitat restoration, and the models for estimating population abundance and productivity parameters are specified with commendable clarity and conciseness, together with supporting references. Even so, there is still considerable uncertainty about the feasibility of implementation. Because this monitoring plan will be very important to the project’s success, monitoring results should be reported and evaluated as soon as possible in future annual reports.

Two issues that warrant further consideration and clarification in subsequent reports or proposals are:

(1) The plan to evaluate fish responses to restoration activities specifies that some metrics will be recorded pre-restoration as well as two- and five-years post-restoration. Have these data already been collected for restoration activities already underway? To improve clarity and precision, we suggest adding a table of restoration activities and a timeline for recording metrics.

(2) It is not clear why “habitat carrying capacity” is defined as the 95th percentile of the Poisson distribution for Ni (i.e., specified by the Poisson parameter ?i in equation 1 on page 6). What relationship is being assumed between abundance, recruitment (i.e., intrinsic productivity), and habitat carrying capacity? We agree that habitat carrying capacity can be estimated from maximum observed abundance, but only when recruitment is not limiting and the system has reached equilibrium. For example, the capacity of a bucket can be estimated from the volume of water it contains when it is being filled to the brim continually by rain. However, that measurement could be misleading during a drought, or before it has had time to fill to its capacity, such as immediately following intervention that increased the capacity of the bucket.

Preliminary ISRP report comments: response requested

Response request comment:

This project is currently undergoing review as part of the Council’s Three-Step process. In December 2019, the proponents responded to qualifications in the ISRP’s review (ISRP 2019-3) of the 2019 Addendum to the 2008 Revised Master Plan for the Hood River Production Project (HRPP). The ISRP then recommended (ISRP 2020-2) that the spring Chinook salmon component of the HRPP Master Plan program meets scientific review criteria with two qualifications remaining to be addressed in the next phase of the review:

Qualification 1: Develop quantitative harvest objectives for hatchery origin spring Chinook salmon returning to the Hood River. The response to the first of previous qualifications for the spring Chinook program (i.e., SCP 1) does not adequately explain or justify the harvest targets for the terminal fishery in terms of the average number of hatchery origin returns (HOR) to be harvested or the proportion of years in which the terminal fishery will be opened. Quantitative objectives should also specify how the target harvest rate would change with adult abundance (e.g., a “sliding scale” decision rule). Quantitative harvest objectives are needed to provide a basis for evaluating the program and for informing stakeholders about the level of harvests that might be expected from the program.

Qualification 2: Develop a plan for monitoring and reducing the proportion of hatchery origin adults that spawn naturally (pHOS) prior to demonstrating success in re-introducing spring Chinook (see previous qualification SCP 3). The ISRP remains concerned that hatchery supplementation efforts are proceeding and expanding without adequate monitoring to detect and respond adaptively to unexpected outcomes (e.g., HOR exceeding harvest demand, excessive straying, poor spawner distribution, or low natural productivity), and without decision rules to change the scale or objectives of the program. Monitoring density effects on productivity (previous qualification SCP 2) is likely the most expedient way to determine if total spawner abundance is exceeding the capacity of the watershed.

To help the ISRP evaluate progress in addressing the two remaining qualifications, the proponents are requested to provide a detailed point-by-point response on the following specific issues in the current proposal:

1. SMART Objectives. Objectives 2-5 of Goal 2 do not meet SMART criteria (see proposal instructions). We suggest reframing them as a single new “Objective 2,” with sub-objectives as necessary to annually implement quantitative harvest rules (see comments in the “Clearly defined objectives and outcomes” section below). 
   
   
2. Methods. More explanation and quantitative detail are needed on the following procedures related to Qualification 1 from ISRP 2019-3 (see additional comments in the “Methods” section below). Specifically:
   • Generating and evaluating forecast models and predictor variables. The ISRP could not easily find the report by Griswold et al. (2009) and would appreciate receiving a copy or a digital link, as well as details of subsequent modifications to the methods used by Griswold et al. A descriptive summary is requested to permit a review of the adequacy of these methods.
   • Setting and adjusting harvest regulations based on run size forecasts. Description of the pathway and decision criteria for setting of harvest levels is requested. 
     
     
3. Evaluation of productivity and recolonization. More explanation and details of analyses and results are needed to demonstrate how data from monitoring will be used to evaluate productivity and progress in recolonization (i.e., progress toward achieving Goal 1), and potential impacts on winter steelhead trout. The proponents state “estimating [natural origin] spawner abundance is relatively straightforward because the majority of adult natural-origin spring Chinook transit through the Moving Falls Fish Facility.” Why then are there no estimates available since 2011 (excluding 2016)? Will future monitoring of natural-origin spring Chinook salmon not enumerated at Moving Falls be adequate to justify ignoring them in calculations of HOR, NOR and PNI? 
   
   The proponents indicated in their presentation that they would continue to monitor steelhead productivity and growth for negative correlations with Chinook salmon smolt releases. How will steelhead productivity be monitored if this project will no longer enumerate adult steelhead abundance after 2024? What analytical methods will be used to determine if negative correlations are attributable to Chinook smolt releases versus other co-variables?

Q1: Clearly defined objectives and outcomes

The problem statement provides helpful background about the merging and reorganization of this project and previous O&M projects 198805304 and 198805308.

Both objectives associated with Goal 1 meet SMART criteria.

Objectives 2-5 of Goal 2 are not quantitative, but all concern the management of adult returns from hatchery releases to be achieved as Objective 1 (and related to Qualification 1 in ISRP 2020-2). The ISRP suggests reframing Objectives 2-5 as a single new “Objective 2” to annually implement quantitative harvest rules. Actions associated with this new objective should include the activities required to generate pre-season and in-season forecasts of run size, and the creel surveys required to estimate harvest. What is missing (and needed) is to explicitly specify the quantitative rules to control harvest. Each action or task should be linked to a description of the appropriate analytical or operational methods.

Objectives 3 and 4 of Goal 3 are not strictly quantitative, but the terms “estimate” and “assess” imply quantitative elements. Timelines are not always specified (i.e., should be more explicit) but are presumed to be annual and continuing indefinitely when not specified.

Q2: Methods

Section 4 of the proposal provides a succinct overview of methods. Table 5 provides helpful conceptual links to the objectives and other sources of information. Adequate details are provided for most methods in the annual reports, appendices, or in other references. Exceptions include methods for generating and evaluating pre-season and in-season forecasts of run size, and methods for setting or adjusting harvest rates based on information from forecasts and creel surveys.

The most recent annual report (2018) states (page 31) “The final version of the run forecast models produced, and the accompanying report Forecast Models for Hood River spring Chinook and Steelhead (Griswold et al. 2009), was submitted to the CTWS in May 2009. Since then, the HRPP has been using these multiple regression models to forecast runs and continue to further refine prediction models with alternative predictor variables using the template produced by WEST, Inc.” The ISRP could not easily find this report and asks the proponents to provide a copy or digital link, and to provide more detailed explanation of the subsequent methods for generating and evaluating forecast models and predictor variables. Have analyses been undertaken to compare the retrospective performance of alternative models and predictor variables over the time series (e.g., Haeseker, S.L., R.M. Peterman, S. Zhenming, C.C. Wood. 2011. Retrospective Evaluation of Preseason Forecasting Models for Sockeye and Chum Salmon. North American Journal of Fisheries Management, 28:12-29)?

The description for harvest adjustment (page 14 of the proposal) is too vague: “When survival rates are relatively good and run forecasts suggest that there will be an adequate return of hatchery-origin spring Chinook salmon to the Hood River, a subsistence fishery is opened to Warm Springs tribal members.” More quantitative details are needed to explain the procedures for adjusting harvest and how and when this was administered in the past (e.g., at what levels of survival and run size). This explanation should be linked to revisions recommended for Objectives 2-5 of Goal 2.

Q3: Provisions for M&E

The proposal should describe how SARs and other quantitative metrics are used to evaluate performance in achieving objectives. For example, the proponents state on page 14 of the proposal that SARs “provide an important population performance metric for the program" but do not explain how or why. The proponents should describe the application of these metrics to project management in future work plans and annual reports.

The framework for monitoring and managing the selective terminal fishery appears to be well conceived and implemented. However, the ISRP requests more detail on procedures used to forecast run size and to adjust the harvest control rules. Statements in the annual report for 2018 allude to the utility of various monitoring activities without explaining how these data are used or why they are useful. For example:

• Page 11. “Marking fish with PIT tags has been extremely useful in adaptive management. Detection of Hood River fish at Bonneville Dam allows managers to gauge run strength before the fish arrive to the Hood River. This in turn informs managers whether run forecasts are realistic so as to adjust harvest regulations or other management activities accordingly.”
• Page 34. “A lesson we have learned is that prediction models are not static. To maintain a good statistical fit and effective predictive value, we must continue to assess model performance and explore alternative predictor variables.”
• Page 37. “Harvest monitoring is of critical importance for fisheries managers. The results from this monitoring are used for determining seasons and regulations annually. In some cases, this data may even be used to make fishery changes mid-season.”

The ISRP is also concerned about the apparent lack of analysis of biological data to evaluate progress toward achieving Goal 1 (“Re-establish and maintain a naturally self-sustaining spring Chinook Salmon population in the Hood River subbasin”). The following statements in the annual report for 2018 emphasize the utility of certain biological monitoring activities, without demonstrating how these data are being used to evaluate productivity or progress in recolonization:

• Page 45. "Snorkel surveys have been a useful tool for documenting fish distribution in the Hood River basin. This is especially germane for the spring Chinook population that was reintroduced, and we expect to increase in distribution, if the reintroduction is progressing successfully."
• Page 50. "Following the two years of higher redd counts we observed a substantial increase in Chinook parr estimated during snorkel surveys. Repeating this protocol following a range of spawning and water years could reveal patterns of spawning abundance to juvenile recruitment, and at what point the carrying capacity of Chinook parr in this stream section is reached."

The ISRP does not understand how the observation that many fish enter the Moving Falls fish ladder but do not continue upstream is evidence of deterrence by the ladder rather than homing to the acclimation site below the ladder. Presumably smolts are released below the ladder? More explanation is needed than given on pages 54-55: “To a certain extent Chinook may be staging in the area below the falls due to homing and attraction to the smolt release site at the facility. Thus, they may not be motivated to proceed any further up the West Fork. Even considering this homing behavior that may cause fish to stage below Moving Falls, we believe there is evidence of passage delays or deterrence to enter the trap. […] We documented numerous instances where tagged fish were detected on these PIT antennas at the upstream end of the ladder, but never documented as passing the site or being captured in the trap; ultimately ending up at a final location below Moving Falls.“

The proponents state “estimating [natural origin] spawner abundance is relatively straightforward because the majority of adult natural-origin spring Chinook transit through the Moving Falls Fish Facility.” If this is true, why are estimates not available since 2011 (excluding 2016)? Will future monitoring of other natural-origin spring Chinook (i.e., those not enumerated at Moving Falls) be adequate to show that their abundance can be safely ignored when calculating overall pHOS (the proportion of natural spawners that are of hatchery origin) and PNI (proportionate natural influence)?

The proponents also point out that the total natural spawning abundance remains lower than the target of 400 so that more hatchery origin spawners are necessary to support the recolonization goal. They argue that it is premature to be concerned about monitoring pHOS and more cost-effective to improve PNI (i.e., reduce genetic risks) by improving pNOB (the proportion of hatchery broodstock that is of natural origin). Should pHOS become a concern, hatchery origin fish could be culled at Moving Falls and the East Fork Diversion ladder, providing additional hatchery fish for tribal food distribution.

The proponents have not yet presented data or analyses to determine if juvenile growth of winter steelhead, or the productivity (i.e., smolts per spawner) of the winter steelhead population are negatively associated with the magnitude of spring Chinook salmon releases. They say they will continue to monitor and report on any association between performance metrics for these species. But how will potential effects of Chinook salmon releases on steelhead productivity be monitored if this project is scheduled to stop enumerating adult steelhead abundance after 2024? What analytical methods will be used to determine if negative correlations are attributable to Chinook smolt releases versus other co-variables?

Q4: Results – benefits to fish and wildlife

The proposal provides a good overview of progress achieved since the HRPP began in 1991. In particular, the timeline of milestones (Figure 1) and the history of production goals (Table 1) provide helpful context for reviewing results.

The program is providing tribal and recreational fishers with increased fishing opportunities and harvests of hatchery fish in a selective terminal fishery. Total adult returns of spring Chinook salmon have generally increased over the course of the program. However, the recent 10-year average return to the mouth of the Hood River is only 1,522 adults, and the goal of 1,700 adults has only been met three times. Hatchery feeding and rearing regimes have been adjusted to ameliorate low rates of smolt-to-adult survival and high rates of precocious maturation (“minijacking”) based on recommendations from the 5-year comparative survival study conducted as part of the HRPP’s 2008 Revised Master plan. In 2019, the program was able to meet its original release target, increasing annual Chinook salmon releases from 150,000 to 250,000 yearling smolts, following completion of Moving Falls Fish Facility (MFFF) in 2013 and expansion of Parkdale Fish Hatchery in 2017. Meanwhile, hatchery propagation of both summer-run and winter-run steelhead has been discontinued (in 2008 and 2021, respectively) for a variety of reasons, including concerns about their adverse effects on productivity of the wild ESA-listed population.