This is a well-written proposal for a project with a long history of critical data acquisition and adaptive changes to increase information and understanding about steelhead in the Umatilla River. The project provides information on population-level survival, productivity, and life history data that is useful for assessing effects of habitat conditions, and restoration and hatchery programs. Of particular interest to the ISRP are the data which show declining smolts-per-female spawner with increasing female escapement (Fig. 3) and the interpretation of it that freshwater habitat is sufficiently seeded (p. 9). The proponents take the interpretation further, suggesting that supplementing the natural population with hatchery-origin fish may not have been an appropriate management strategy. This is a great example of interpreting M&E data to the point where it can be used by decision-makers.

However, as the data indicate that spawner numbers are not limiting juvenile production, then there should be a sufficient number of natural origin spawners to supply all the broodstock for the hatchery. Thus, the ISRP found it surprising that the hatchery program was using some hatchery origin returns for broodstock. Clearly, this program is providing lots of useful information for decision-makers, though some of the decisions regarding hatchery production appear to be ignoring some of the findings presented in the proposal.

M&E matrix – support. As habitat projects and monitoring projects are not presented as part of an integrated proposal or plan, the need for a crosswalk to identify the linkages between implementation and monitoring is extremely important for basins or geographic areas. The ISRP is requesting a response from the Umatilla Basin Natural Production Monitoring and Evaluation Project (199000501) to summarize the linkages between implementation and monitoring projects in the basin. As a key M&E project and partner in the basin, we ask your project to assist them in creating the summary and provide information to them about what, where, and when your monitoring occurs and what is being monitored for and shared with implementation projects in the basin. A map or maps of locations of monitoring actions would be helpful in this regard.

Q1: Clearly defined objectives and outcomes

The section on goals and objectives was brief and some were not specific enough. Stating objectives using the SMART format would be helpful and should be presented in the next round of proposal reviews and annual reports. For Objective 4 for example, how the diversity of steelhead will be assessed is unclear. Also, for Objective 5, it is not clear how or when this this will be accomplished.

Q2: Methods

The methods are generally sound, but there are four areas for potential improvement.

Smolt run size is estimated by a series of independent closed-abundance estimates for each period (sampling interval). The length of each period was not specified in the proposal (except for TMF where it is one month), and the ISRP is concerned about the assumption that all marked fish pass the trap during the interval and/or that capture probability is constant over the interval. If this is not the case, capture probability and abundance estimates will be biased. To what extent have these assumptions been tested? If the length of the interval has increased to meet the passage assumption, is it likely that capture probability is not constant over the longer period?

A more flexible approach would be to use the Bonner and Schwarz (2011 and 2014, BT SPAS R library) time-stratified estimator. This model can be useful when recaptures for some periods are sparse, or when the trap(s) cannot be operated due to high flows (e.g., Fig. 6 of Hanson et al. 2020), and allows for finer temporal intervals that may lead to more accurate estimates of abundance and run timing. This approach avoids problems with arbitrary pooling of data across periods that is needed if sample sizes are low or trapping is not conducted over some periods.

Given the intense effort to mark fish and trap smolts, this analytical upgrade seems well worth it. The precision of smolt run size estimates at TMFD is very high (CVs 1995-2018 =5.4%) and may be an artifact of the analytical procedure (too much pooling). A better model may be more useful in Birch Creek where there are few strata which cover long periods where capture probability is unlikely to be constant as currently assumed (Table 7 of Hanson et al. 2020). Improved estimates of smolt run size at Birch Creek will lead to improved estimates of survival to TMF, which is highly relevant given concerns about survival rates in low-flow years. See Bonner and Schwartz (2011), Bonner and Schwartz (2014), and Hanson et al. (2020) for possible analytical approaches.

Egg deposition estimates could be improved by using a fork length-fecundity relationship rather than age-specific fecundity average. This would better account for the decreasing size and age-at-return that has been seen in many Chinook populations over the last decade or more (e.g., Lewis et al. 2015).

Would it be possible to develop a corrected SAR value that accounts for losses from fisheries? This would allow for better evaluation of effects of downstream/upstream mainstem passage or marine survival. Currently these effects are confounded with changes in exploitation rate.

Would it be possible to calculate the variance on the hatchery:natural ratio using the same binomial likelihood described for the smolt analysis? This error could be substantial for some tributaries where few spawners are observed or where the presence/absence of an adipose fin is difficult to distinguish.

Q3: Provisions for M&E

The proposal provides very little information on how effects of hatchery supplementation, flow, and habitat improvements on smolt run size or juvenile survival rates will be quantified. We suggest fitting a Ricker model with covariates:

log(R/S) = a + b*S + d*X

where R is the number of smolts from brood year t, S is egg deposition or female escapement that produced those smolts, a is the log of productivity (R/S when there is no density dependence because S is 0), b is a density-dependent effect, X is a covariate such as flow or some measure of habitat restoration, and d is the coefficient for the covariate (the strength of the effect per unit increase in X). Another covariate to assess could be pHOS, though it could also be included through adjustment of S via,

S = S*(1-pHOS) + S*pHOS*e

where the first group of terms on the right side of the equation is the contribution of eggs or females from natural origin spawners, the second group of terms is the contribution from hatchery-origin fish where “e” is the estimated effect of hatchery-origin fish on survival from egg-smolt. Essentially S is a weighted average spawner abundance, that accounts for reduced spawning success or lower survival rates of juvenile fish produced from hatchery-origin spawners. It may be challenging to estimate e, depending on the extent of variation in pHOS and survival rates over time.

Survival rates between release locations could be evaluated using

log(Surv) = b0 +b1*X

where b0 and b1 are estimated and X is the covariate to be evaluated.

Q4: Results – benefits to fish and wildlife

The proposal provides an excellent summary of the many project actions, what was learned from the results, and how the objectives and actions were modified as a result. It also provides information about how these results have influenced management and informed other projects that are closely aligned. The results have contributed to broader efforts in status and trend monitoring, and can be used in future life-cycle modeling.

One key problem in the subbasin is that both habitat restoration and hatchery supplementation affect steelhead abundance and survival in the Umatilla River, and so the effects are confounded. The proponents propose tributary-specific monitoring to allow separating the effects of these actions, and this is a high priority for funding.

References

Bonner, S.J. and Schwarz, C.J. 2014. BTSPAS: Bayesian Time Stratified Petersen Analysis System. R package version 2014.0901.

Bonner, S.J. and Schwarz, C.J. 2011. Smoothed estimates for time-stratified mark-recapture experiments using a Bayesian P-spline approach. Biometrics 67:1498-1507.

Hanson, J.T. Jewett, S.M. and S. Remple. 2020. Evaluation of juvenile salmonid outmigration and survival in the Lower Umatilla River Basin. 2019 Annual Report BPA Project #1989-024-01.

Lewis, B., W.S. Grant, R.E. Brenner, and T. Hamazaki. 2015. Changes in size and age of Chinook salmon Oncorhynchus tshawytscha returning to Alaska. PLoS ONE 10(6):e0130184.

Ohlberger, J., E.J. Ward, D.E. Schindler, and B. Lewis. 2018. Demographic changes in Chinook salmon across the Northeast Pacific Ocean. Fish and Fisheries 19:533-546.

This project proposes status and trend monitoring of ESA-listed Umatilla River steelhead and Chinook salmon, and collaboration in an Intensively Monitored Watershed (IMW) project intended to evaluate the effectiveness of habitat restoration actions in two tributaries of the Umatilla. Work related to status and trends monitoring in Objectives 1-4 meets scientific criteria. A response is needed that expands, clarifies, and provides more detail concerning the IMW project and Objective 5. The study design needs more thorough explanation, and more background information on the reference and treatment streams needs to be provided. Comparative metrics and data analyses need further explanation. Overall, this is a thorough proposal for continuation of a centrally important project in the Umatilla Basin. The investigators describe a highly integrated project to collect critical data on production and survival of wild steelhead and spring and fall Chinook salmon. This project could provide critical data to assess whether the habitat restoration projects in the Umatilla River basin are effective in increasing abundance, survival, and productivity of naturally-spawning steelhead and salmon. In addition, it provides key data to determine the success of the new integrated hatchery supplementation program, whereby separate groups of Conservation and Harvest smolts are produced. These data are necessary to determine if the integrated hatchery program is contributing to the recovery of steelhead and salmon, or just another factor leading to their demise (or no change is detected). 1. Purpose, Significance to Regional Programs, Technical Background, and Objectives The project is consistent with many regional programs and projects including the NPPC Fish and Wildlife Program and the Umatilla/Willow Subbasin Plan. It addresses several RPAs in the BiOp. This work is of great significance to regional programs, because it provides critical data to assess how natural populations of steelhead and two life history types of Chinook are responding to a variety of conditions, including in-river habitat, flow, migration corridors, and ocean conditions. Without it, little will be known about the performance of the newly created Conservation groups of salmon and steelhead. The proposal includes status and trends monitoring and a new Intensively Monitored Watershed project. The main goal of the Umatilla IMW project is to determine whether habitat enhancement results in higher abundance, survival, and productivity of natural spawned steelhead and salmon. A confusing aspect of the proposal is that several of the objectives and deliverables include work related to both status and trends monitoring as well as to the IMW habitat effectiveness evaluation. The objectives and deliverables for the status and trends work and those for the IMW work should be separated so that these two aspects of the project are clearly distinguishable. Several projects are addressing components of the IMW work, although this project seems to have the bulk of the responsibility for its conduct. Dividing the work among projects makes scientific evaluation of the IMW effort difficult. Why was the IMW work not consolidated in a single proposal? 2. History: Accomplishments, Results, and Adaptive Management This project has been ongoing in various forms since 1994, but underwent an extensive review in 2006 by the ISRP. It was restarted in 2009, after reformulating goals. This proposal is characterized by carefully planned sampling designs for the redd surveys and juvenile abundance in tributaries, and for habitat monitoring. The project can point to various results that have allowed managers to make important decisions based on the data that was collected. Based upon the results presented, the project appears to have been productive and has accomplished it objectives since it inception in 1994. Data collected through this project are critical for monitoring salmon and steelhead populations in the basin. A notable conclusion drawn from data analysis was that “habitat enhancement has not resulted in a significant improvement for summer steelhead and that the system may be at capacity for production of the species.” The negative relationship between smolts/female and number of females supports this conclusions and suggests that density-dependence may be affecting smolt survival. This conclusion is tentative but it argues for a more rigorous evaluation of the effectiveness of habitat restoration actions in the Umatilla Basin, which the proponents propose to undertake. In addressing adaptive management, the proponents indicate that the information they obtained has assisted with management decisions and provide some examples. They did not specifically address how their project has changed based on previous results. However, their decision to participate in CHaMP is indicative of their willingness to shift the direction of the project. 3. Project Relationships, Emerging Limiting Factors, and Tailored Questions for Type of Work (Hatchery, RME, Tagging) The proposed project is one of four collaborative BPA funded projects aimed at monitoring the status and trends of Chinook salmon and summer steelhead in the Umatilla River. The project is tied to several other BPA funded projects in the Umatilla Basin. It also relates to several other IMW projects in the Columbia Basin that are collaborating in the development and implementation of CHaMP. In particular, this project and another in-basin project (1990-050-01; Umatilla Basin Natural Production M&E) are cooperating in conducting the IMW habitat evaluation in the Umatilla. Some discussion of the new C & H / Integrated Segregated hatchery production scheme would have been helpful, but it seems that the proposed project, without explicitly discussing it, will deal with it effectively. In addressing emerging factors the proponents make the general statement that the data collected by this project could assist in determination of fish population response to emerging threats but do not offer anything more specific. Climate change and predation by birds and native and non-native fish predators are key emerging limiting factors which are dealt with in other proposals. It will be important to determine how this project can link with those data, such as estimating loss of this DPS of steelhead from Caspian tern and cormorant predation at the mouth of the Columbia River. 4. Deliverables, Work Elements, Metrics, and Methods This proposal has components pertaining to both routine status and trend monitoring and evaluation of habitat effectiveness under the IMW program. Methods and metrics for assessing status and trends in Objectives 1-4 are fairly standard and are appropriate for this type of work. The ISRP views positively the proponent’s willingness to engage in rigorous habitat effectiveness evaluation under the auspices of CHaMP and according to ISEMP protocols. Properly conducted, this evaluation could yield the most valuable information to date on effectiveness of habitat enhancement in the Umatilla Basin. Several issues, however, need clarification. Several objectives and deliverables (e.g., deliverables 4, 6, 9, and 10) in the proposal apparently include work related to both status and trends monitoring and to the IMW habitat effectiveness evaluation, complicating scientific review of the proposal. It would be helpful if the objectives and deliverables for the status and trends work and those for the IMW work could be separated so that these two aspects of the project are clearly distinguishable. The study design for the IMW project needs more thorough explanation, and more background information on the reference and treatment streams should be provided. The proposed approach for evaluation of the effectiveness of habitat enhancement actions is to compare a control or reference stream with each of two treatment streams that have undergone habitat enhancement. A main difficulty is that appropriate treatment and control streams are difficult to find. The Upper Umatilla, a reference stream, receives supplementation, whereas Meacham Creek, a treatment stream, has been subject to habitat restoration and also is supplemented. Steelhead use both tributaries for spawning and rearing. Therefore, a comparison between these tributaries should yield information on the effectiveness of the habitat projects in Meacham Creek, assuming there is no interaction between the habitat work and supplementation, and other physical and biological differences between the tributaries are negligible. In contrast, Birch Creek, another treatment stream, receives no supplementation but connectivity and fish passage has been restored. Since the Upper Umatilla is supplemented, it is not an adequate control stream to compare with Birch Creek, although trend monitoring (i.e., before-after) can be conducted to assess changes. How will this apparent problem be resolved? The proponents need to deal with several other questions pertaining to the IMW project. How do the reference and treatment basins compare physiographically and biologically? The history of land use, habitat loss, and hatchery influence in reference and treatment tributaries should be summarized. What habitat restoration actions have been and will be implemented, and on what time frame? What is the fish distribution and abundance in these streams? Comparative metrics and data analyses need further explanation. What metrics (fish and habitat) will be compared between treatment and reference basins to evaluate the effectiveness of habitat restoration actions? Will the proponents be responsible for collection of habitat and fish data, data integration, and data analysis? What data will be collected by other projects? An extremely large amount of data will be collected. How will it be analyzed? It should be possible to use model selection to assess how, for example, smolt production relates to habitat restoration, by fitting models with and without this covariate. ISEMP proposes a long list of habitat variables that can be measured. How will the decision be made as to which of these variables are most important for this work?