The Northwest Power and Conservation Council’s 2009 amendments to the Columbia River Basin Fish and Wildlife Program called for a regular system of independent and timely science reviews of the Fish Passage Center’s (FPC) analytical products. The 2014 Program’s Appendix H maintains this review function. These reviews include evaluations of the Comparative Survival Study’s draft annual reports. The ISAB has reviewed these reports annually beginning six years ago with the evaluation of the CSS’s draft 2010 Annual Report and most recently the draft 2015 Annual Report (ISAB 2010-5, ISAB 2011-5, ISAB 2012-7, ISAB 2013-4, ISAB 2014-5, and ISAB 2015-2). This ISAB review of the draft 2016 CSS Annual Report is the ISAB’s seventh review of CSS annual reports in response to the Council’s Program language.
This ISAB review begins with an overview of the latest report (this section). It then moves on to suggesting topics for further CSS review (Section III), general comments on each chapter of the 2016 CSS Annual Report (Section IV), and ends with specific queries and suggestions (Section V).
The annual CSS report is a mature product, typically including only updates with the latest year of data and expansion of analyses as more data are acquired. Many of the methods have been reviewed in previous ISAB reports and so now receive only a cursory examination. As more data are acquired, new patterns and questions arise on the interpretation of the results—this is now the primary focus of our reviews. The ISAB appreciates the CSS’s detailed response to suggestions provided in previous reviews, and the ISAB does not expect the CSS to necessarily respond immediately to new requests for further analyses.
Chapter 1 is similar to previous years with the 2015 results added. Two new fish populations have been added. In the 2016 report, the size of the PIT tags used is reported as being 11 to 12 mm instead of the 9 to 12 mm in previous reports. If this is a real change, the rationale for the change is needed along with a discussion of potential impacts on the fish (e.g., are larger fish now tagged to accommodate the larger tags?).
In Chapter 2, the existing life cycle model was used in a prospective analysis to simulate the relative benefits of flow/spill modifications to habitat. While the approach is generally well implemented, the ISAB has some concerns about specific aspects of the simulation study that suggest the outcomes may not be as clear cut as indicated in the report. For example, what is the justification of choosing particular years as “representative” of low/medium/high flow conditions? Some variables that could vary are held constant—e.g., powerhouse contact rate derived from PIT tag data (PITPH) and water travel time (WTT)—and so the simulation results may underreport the variability in the response.
Chapter 3 is mainly an update with the latest information on in-river effects on juvenile travel time, instantaneous mortality, and survival. A key finding is that there is large variation in the results among years and among cohorts. The variation among years is understandable; the variation within a year less so. Many figures (e.g., Figure 3.2) show a consistent pattern in fish travel time and survival over cohorts as the year progresses. Mortality tends to increase over the migration season and with water temperature (except for sockeye). The report lists four potential mechanisms: (1) declining smolt energy reserves or physiological condition over the migration season and with water temperature, (2) increasing predation rates on smolts over the migration season and with increased water temperature, (3) increases in disease susceptibility or disease-related mortality over the migration season and with increased water temperature, or (4) some combination of these often interrelated mechanisms. Is there an attempt to test these hypotheses using other approaches, either within CSS or by other investigators? Answers to these questions might lead to improvements in survival. The ISAB agrees that the apparent contradictory response of sockeye warrants further investigation.
Chapter 4 described overall annual SARs and was updated with new data; details are presented in appendices. In addition, an analysis of relationship between the ratio of transport to in-river survival (TIR; transport effect) and in-river survival is now included. It is not surprising that the transport TIR is inversely correlated with in-river survival (Lower Granite Dam [LGR] to Bonneville Dam [BON]). This new analysis identified the value for in-river survival when the benefits of transportation appear to disappear. The CSS also reported on the relatively large absolute difference in SAR based on PIT-tags versus run reconstruction (the values are highly correlated, however). As in previous reports, this report listed various hypotheses. A study is underway to further evaluate PIT-tag effects on salmon survival, but these results will not be ready until after summer 2017 when tagged age-5 Chinook will have returned. Potential bias in survival caused by tagging methodology (or in the run reconstruction methodology) is an important issue to resolve, and the ISAB looks forward to the results of this study.
The material in Chapter 5 was combined with other chapters in previous reports and has now been split out. This is an update with an additional year of data. Chapter 5 continues the examination of the relationships between life cycle productivity and SARs, including the level of SARs needed to reach or exceed population replacement. The findings suggest that pre-harvest SARs of 4%-6% are associated with pre-1970 levels of productivity for Snake River spring/summer Chinook; SARs are much lower in subsequent decades. How might these early SARs (~4-6%) compare with SARs from viable wild Chinook populations in other regions? To what extent might improvements in hydrosystem management, predator control, and estuarine habitat lead to SARs of 4%-6%?
Chapter 6 is mostly an update on Snake River subyearling fall Chinook.
Chapter 7 is a repeat of an analysis done in 2010 with additional years of data. A logistic regression analysis was used to investigate the impact of year, bypass effects, and rearing type on subsequent survival and return as an adult. Estimates (Table 7.4) appear to be on a “per bypass” basis. What is the average number of bypasses encountered by a fish? Wouldn’t that be a more accurate reflection of the impact of bypass on an outgoing smolt?
Chapter 8 examined differences in the mean age of maturity among different stocks, years, and fish type (wild or hatchery) using regression methods. While the analysis mostly seems appropriate, the ISAB has numerous suggestions on improving the presentation of the results.
Appendices A and B are updated with an additional year of data. The ISAB is pleased that electronic versions of many of these tables are now available at the FPC website.
Appendix C reports on the development of a weighted bootstrap procedure to deal with a stratified sampling and tagging of smolts that is not proportional to abundance. The authors describe a bootstrap procedure for estimating parameters, but the ISAB suggests that a stratified approach be incorporated directly into the analysis routines currently used to allow for future expansion of stratified-tagging studies.
Suggested Topics for Further Review
In 2013, the ISAB recommended these topics (ISAB 2013-4, Page 1):
- Hypotheses on mechanisms regulating smolt-to-adult survival rates (SARs)
- Life-cycle modeling questions and Fish and Wildlife Program SAR objectives
- Data gaps
- Rationalization of CSS's Passive Integrated Transponder (PIT)-tagging, and
- Publication of a synthesis and critical review of CSS results
In 2014, the ISAB recommended these topics (ISAB 2014-5, pages 2-3):
- Hypotheses on mechanisms regulating smolt-to-adult return rates (SARs) [update from 2013 review]
- Life-cycle modeling questions and Fish and Wildlife Program SAR objectives [update from 2013 review]
- New PIT/CWT study to further investigate differential survival among these tag types
In 2015, the ISAB recommended these topics (ISAB 2015-2, pages 4-5):
- Use SAR data to examine both intra- and interspecific density dependence during the smolt out migration and early marine periods
- Propose actions to improve SARs to pre-1970s levels
- Explore additional potential relations between SARs and climate and ocean conditions
- Consider ways to explore the variability of inter-cohort response
The CSS group has incorporated many of our suggestions into the current document. For example, the current report has a substantial discussion of correlations among SARs from different regions or effects of transport on SARs (#1 in 2013; #1 in 2014). The life cycle modeling now allows for variation in stream productivity and hydrosystem survival and simulates the correlative impacts of these changes on predicted future population abundances (#2 in 2013; #2 in 2014; #2, #3 in 2015). The ISAB appreciates the CSS efforts to respond to our queries which in turn lead to further questions.
Some of the recommendations from the ISAB appear to be beyond the scope of the CSS but will become increasing important in the future. For example, is there evidence of density dependence during the smolt out-migration and early marine periods (2015 #1)? Could the CSS estimate total smolt abundance of each species, say at Bonneville Dam? Is this a potential mechanism to explain the inter-cohort variation in responses (2015 #4)? This is reflected in our recommendations for future work below.
In 2016, the ISAB recommends the following four topics for future reports:
- Use more realistic and more variable future flow conditions for the study on the impact of flow/spill modifications under future climate change. Simulating only low flows or high flows for decades may not be a realistic scenario. What is the impact of a correlation between Pacific Decadal Oscillations (PDOs) and flows that has not been considered in the simulations presented in the 2016 report?
- What is the impact of the new restricted tag sizes? Are there fish that were previously marked and are now not marked (e.g. smaller fish) due to the larger PIT tags being used? Similarly, conclusions from studies of compensatory mortality (e.g. in relation to predator control) may be affected by the choice of fish that are tagged. A brief review of the PIT tag procedures should be undertaken so that users of the CSS data are fully aware of any limitations in the conclusions of other studies that are related to types/sizes of fish tagged.
- There has been a great deal of interest in the impact of predator control programs on salmon returns, especially northern pikeminnow and birds. A life-cycle model is the natural way to study these impacts, but the current version of the life-cycle model appears to incorporate density dependence only at the spawner-to-smolt stage. The ISAB recommends that consideration be given to modifying the life-cycle model to allow a range of compensatory responses ranging from complete additivity (as now is the case) to plausible compensatory mortality effects related to density dependence and predator selectivity (see ISAB 2016-1). This continues our previous recommendation (#1 in 2015) to investigate impacts of density dependence on subsequent return.
- Both the CSS and NOAA provide estimates for in-river survival. How do these estimates compare to each other? If there are consistent differences in the estimates, can these be explained?
- What factors have led to declining proportions of four and five-year olds and increases in three-year olds in spring/summer Chinook? Models that include ocean factors associated with salmon growth and climate change, differences in hatchery practices, or freshwater environments (tributary temps, or annual differences in migration corridor) may be of interest.