This is one of three BPA-funded projects that address the critical uncertainty of ocean effects on survival of Columbia River salmon. The ISRP appreciates that project proponents have followed some of ISRP’s past recommendations to develop approaches tailored specifically to Columbia River salmon in the estuary, plume, and ocean. Coordination with other ocean and estuary projects has improved. However, a number of past issues raised by ISRP and ISAB have not been addressed. In addition, there are new issues resulting from proposed changes in project design and methods that need to be addressed. Although the ISRP is not requesting a response at this time, we do have one major qualification.

Qualification 1: Address the issues listed below during the contracting process and in the project’s 2011 annual report, which will be reviewed by the ISRP:

1. Feasibility of COAST Approach. How can the proposed objectives be achieved if the open-coast acoustic array is still being developed? Are there other approaches that would be more cost-effective for estimating life-stage specific open ocean distribution and survival of salmonids?

The proposed work could yield important new data on coastal and estuarine distribution of Columbia River Basin salmonids and endangered ESUs. However further information is requested on how the proponents view the strategic balance of this project between assessing broad “offshore” distributions (where it appears more development work is needed as mentioned below) versus detailed monitoring to estimate survival between closely spaced reaches in the estuary.

After several years of research the project is still in the process of demonstrating “proof of concept” of the effectiveness of the open coast arrays to detect tagged Columbia River and Snake River spring Chinook salmon (no other species have been evaluated). The project now recognizes some of its current limitations. For example, recent results (May 2010) showing incomplete detection histories for several jacks returning to the Columbia River in 2009 and two adults returning in 2010 have highlighted this uncertainty, together with the findings of a fairly uniform spatial distribution along the Willapa Bay sub-array to its current offshore limit of 250 meters. These observations affect a key assumption: 100% or consistent detection by the acoustic array. It will be important to evaluate how this issue could affect key studies involving mortality of transported versus in-river smolts.

The project claims that its methodology is the only experimental technique available for addressing these issues, including early marine survival of salmon. While the approach is innovative and more direct, other studies have used incremental scale and otolith growth to examine size- and life-stage dependent mortality during specific periods at sea.

Estimation of survival at sea is important for separating freshwater versus marine-related factors of survival. Smolt counts and coded-wire tags (CWT) have been traditionally used to estimate survival at sea. This project provides an estimate of mortality for yearling Chinook salmon smolts specific to each phase of early marine life, instead of release to recovery survival (CWT estimates). But results to date seem to show an exponential decline in survival with distance from the river, as expected. This seems to be the primary unique quality of this study of interest to managers. Migration rates are also unique, though other researchers have used short-term tracking to document travel rates.

Interannual and seasonal variability in migration rates, estimated survival, etc. all need to be closely related to measured ocean conditions, either from ships or satellites. What have the proponent’s results shown so far? The proponent’s decay model of survival seems too simplistic, based on the research of others, and this needs further evaluation by the proponents.

The low reported survival at sea is not surprising given the history of low survival rates of yearling and subyearling Chinook salmon based on CWT data. Chinook salmon are well-known to have lower ocean survival rates that other salmon species. The declining survival with distance from the Columbia is expected. For fisheries management, the key information is the evaluation of survival of in-river versus transported smolts. It would be of interest to compare results from CWT and PIT tagged salmon with those from this study and evaluate the benefit of the acoustic tag versus CWT and PIT tag for this management question because the acoustic tag approach is much more costly.

2. Strategic Plan for COAST Array Location. Does COAST have a strategic plan for COAST array location, developed in cooperation with other Columbia River projects that use or plan to use BPA-funded arrays for their projects? If not, such a plan needs to be developed.

We reiterate our suggestion in past reviews that the proponents coordinate development of the COAST acoustic array design with other projects in the Columbia River Basin and ocean, as this issue was inadequately addressed in the proponent’s previous response. The proponents assume that Columbia River spring Chinook salmon migrate northward along a coastal corridor that is adequately sampled by the acoustic arrays. However, evidence exists for migratory patterns in other directions (southward, straight offshore). COAST proposes to remove the only listening line located south of the mouth of the Columbia River. The ISRP reiterates our previous recommendation that two ocean listening lines located to the south are needed to demonstrate the feasibility of this project. If COAST is to provide accurate estimates, arrays to the south of the Columbia River and additional tags to expand the proposed study to a 2-directional design are necessary. In addition, a closer examination of the location of arrays with respect to hypothesized locations of juvenile salmon survival bottlenecks is important to developing a strategic plan for potential future locations of arrays.

3. Coordination with other projects. What specific process is used by COAST to coordinate with other projects to estimate survival of Columbia River salmon?

Coordination with other projects has improved, but it could be better. The proponents promise to tie in closely with the CDFO Shelf Survival proposal (#200300900) and the NOAA Ocean Survival of Salmonids proposal (#199801400). All three projects promise a key deliverable - survival. However, the coordination appears rather loose and further information on exactly how the three projects will work together is required. The proposal presents a possibly unbalanced review of VEMCO tags relative to JSATS, and no discussion is provided in reference to McMichael et al. (2010) regarding their survival estimates. Nevertheless it is encouraging to see the increased discussions and joint work with USACE contractors and others working on survival estimates in the lower river and estuary. The ISRP recommends increased coordination with JSATs research in the estuary, since all COAST smolts are proposed to be released below Bonneville Dam. A component linking COAST to the nearshore studies and restoration work in the estuary, however, is missing. As well, the inner estuary proposals (e.g., LCREP, #200300700) should be tied in to the propose COAST work.

4. COAST Study Design. What are the likely magnitudes of the effects of assumptions in the COAST study design on results and what are the consequences for conclusions?

The ISRP reviewed the appendix attached to the proposal with extensive documentation of the power analysis and resulting study design implications for the POST project. The development of the likelihood approach is clear and the assessment of the results via simulation is useful. The explicit statement of assumptions is particularly appreciated by the ISRP. These assumptions include:

• Survival per day is the same in the estuary, plume, and ocean
• All surviving fish travel a given segment in the same time
• Detection probabilities are the same for the groups being compared
• Observed high and low survival rates bracket rates that are likely to be observed in future
• Effect of dam passage or transport is fully expressed by one month after migrating.

The validity of these assumptions may be debated, but it is clear that they have effects on results from this analysis, although the likely magnitude of these effects is not presented nor the consequences for conclusions. Nevertheless, it appears that the study design and power analysis presents an approach for planning tagging effort and array deployment.

5. Deliverable V. Testing the Delayed Mortality Theory. Can the proponents provide stronger justification for continuation of work on this deliverable? If the work continues, are there other more cost-effective methods for achieving the objective?

The Independent Scientific Advisory Board (ISAB 2007-1) advised against continuing efforts to measure absolute latent mortality, suggesting instead that the focus should be on estimating processes such as in-river versus transport mortality that can be measured directly. Proponents acknowledge the ISAB recommendation but argue for continuation in part by citing Welch et al. (2008; (http://www.plosbiology.org/article/info:doi/10.1371/journal.pbio.0060265), a comparison of the un-dammed Fraser/Thompson River with the dammed Columbia. The ISAB (2007-1) concluded that determining latent mortality relative to a damless reference is not measurable. The argument in the proposal does not convince the ISRP that this ISAB conclusion warrants reconsideration.

Can acoustic tags provide a more accurate and precise estimate of differential delayed (latent) mortality than a similar study approach that used greater numbers of coded wire tagged fish (at a much lower cost)? The acoustic tags estimate survival after a few months, but CWTs measure survival to adults. Has a comparison of the two approaches been made? If research on this objective continues, it would be important to incorporate survival of hatchery versus wild fish into the analysis. Will Chinook salmon tagged by COAST below Bonneville be identified as hatchery versus wild fish? The proposal notes that wild salmon tend to have higher survival rates; therefore, the ratio may affect the survival findings. What is the expected hatchery/wild tagging ratio? It would be interesting to compare data of tagged and untagged Chinook. Also, the study might compare survival rates with those from CWT salmon. This could tell us the fraction of mortality that occurs during early versus late marine life.

6. Detection Efficiencies. The ISRP has a number of questions about tag detection efficiencies that were not addressed in the proposal. What percentages of fish are detected only once, for example, and not again? Are these deemed mortalities or did fish residualize in areas outside of the detection range of arrays? Along the arrays in the ocean, what about fish that migrate close inshore where there are no receivers? And how often are receivers down or lost? On page 22 - the detection range for V7 tags is less than 300m. The detection probability for V7 tags is about 70%. The accuracy and precision of the estimates is questionable. It seems that COAST has given up a lot by going from the V9 to the V7 tag. The depth of a proposed new array at Cape Elizabeth would extend to 500m, but is this depth is beyond the detection range of the V7 tags? Are tagged fish easily detected if they are at or near the surface and the cable is in 500m deep water? What is the effect of wave action on detection of tagged fish?

7. Genetic stock identification (GSI). How many genetic stocks of juvenile spring Chinook salmon can be identified by the proposed GSI? Procedures for GSI need to be described. Proponents need to demonstrate that current techniques are capable of identifying origins of individual fish that are tagged and released. Ocean studies should advance toward designs that can also evaluate differences/similarities in survival of hatchery vs. wild fish of the same genetic stock. Is there a way to standardize genetic stock identification methods so that results of the three BPA-funded ocean projects are directly comparable (different labs are using different methods)?

8. Definition of the plume. Why is the plume defined as Sand Island to Willapa Bay? The proponents’ definition of the plume (Sand Island to Willapa Bay) is very different than accepted terminology, and the proposal would be improved by an explanation as why they chose this definition. The plume is usually described as outside the Columbia River bar, and the plume disperses both to the north along the Washington coastline and to the south along the Oregon coastline. See for example:

Fielder, P.C., Laurs, R.M., 1990. Variability of the Columbia River plume observed in visible and infrared satellite imagery. Int. J.Remote Sens. 11, 999-1010.

9. Alternatives to Fixed Arrays. Are there other more innovative techniques than fixed acoustic arrays that could be employed in the future to track open coast and ocean distribution, migration patterns, and survival of Columbia River spring Chinook? For example, what about the use of robotic vehicles to measure ocean conditions and track tagged salmon to extend coverage beyond the detection range of fixed listening lines on the continental shelf/slope?

10. Scientific workshop. The ISRP recommends a scientific workshop in 2011 focused on estimation of estuarine and ocean survival, forecasting of adult returns, and adaptive estuary, plume, and ocean environmental assessment for Columbia River Basin salmon and steelhead. Perhaps the proposal should include this workshop. A workshop would help to improve coordination and collaboration, standardization of methods (e.g., genetic stock identification), development of simulation and predictive models, and integration of results among Columbia River Basin estuary/ocean projects. One aspect of all projects that needs work is how to include more detail on sub-stock structure, including hatchery versus wild fish, hatchery release time, area comparisons, in-river migration and associated ocean migration, and more in the models.

11. Adaptive management. Are the proponents overselling their ability to use this approach to improve real-time management of spill and transport? How can adaptive management with respect to estimates of ocean survival be implemented in the Columbia River system? Is it possible that tagging experiments could be designed in concert with hatchery, hydrosystem, and harvest managers to test specific hypotheses related to estuarine and early ocean survival?

The proposal would be improved by further details on how POST results have influenced on-the-ground management decisions by fishery or hydrosystem agencies. For example, has the Welch et al. (2008) paper (“Survival of migrating salmon smolts in large rivers with and without dams”) resulted in any changes in operations of the Columbia River Basin dams? How do COAST indicators tie in with those being developed by CDFO, NOAA, and other projects in this review?

In the past several reviews, the ISRP asked, “How would the fully-implemented ocean array and long-term monitoring data on seasonal and interannual variations in survival rates or migration rates among years or stocks actually be used by managers of the Columbia River Basin hydrosystem?” The ISRP agrees with the proponents’ past response that estimates of ocean survival for tagged release groups of hatchery fish can be used to inform policy makers, fishery managers, and researchers. However, the proponents have never answered the ISRP’s question about how hydrosystem managers would actually use the data. The proponents still do not seem to recognize that ocean variability will make the concept of tracking the geography of ocean mortality and subsequent adjustment of hydropower system very difficult to manage.

The project is clearly significant to regional programs, but the proposal could be improved by attention to unrealistic objectives and expectations that implementation of acoustic tagging technology would result in improved real-time management of spill and transport. The proponents state that the latter two options could be decided upon by measurement of marine survival with their methods: “For example, if marine survival is exceptionally low, transportation and/or increased spill may not be beneficial, as smolts would reach the ocean sooner thereby exposing them to unfavorable ocean conditions (e.g., increased predation or decreased food supply), leading to lower survival.” Explain the specific processes that would be used to achieve real-time management. Do managers think this process would work?

12. Communicating Results. Can the proponents develop more effective approaches for communicating their results directly to Columbia River Basin hatchery, hydrosystem, and harvest managers? Websites, scientific meetings, and peer-review scientific publication are excellent methods for communicating with scientific peers, other government agencies, educational institutions, and conservation organizations, but are likely not effective tools for communicating directly with hydro, harvest, and hatchery managers.

13. Update Online Proposal Format. The format of this proposal was confusing and difficult to follow. Proponents should reformat their online proposal to better conform to the specific information requested in each section of the online form. The repetition of the same deliverables under several objectives seems unnecessarily repetitive. Objectives providing the same deliverables could be combined into one objective. The important information on study design that was included only as an attachment should be incorporated into the online form. The online form should present the complete proposal as a stand-alone document.

This is one of three BPA-funded projects that address the critical uncertainty of ocean effects on survival of Columbia River salmon. The ISRP appreciates that project proponents have followed some of ISRP’s past recommendations to develop approaches tailored specifically to Columbia River salmon in the estuary, plume, and ocean. Coordination with other ocean and estuary projects has improved. However, a number of past issues raised by ISRP and ISAB have not been addressed. In addition, there are new issues resulting from proposed changes in project design and methods that need to be addressed. Although the ISRP is not requesting a response at this time, we do have one major qualification. Qualification 1: Address the issues listed below during the contracting process and in the project’s 2011 annual report, which will be reviewed by the ISRP: 1. Feasibility of COAST Approach. How can the proposed objectives be achieved if the open-coast acoustic array is still being developed? Are there other approaches that would be more cost-effective for estimating life-stage specific open ocean distribution and survival of salmonids? The proposed work could yield important new data on coastal and estuarine distribution of Columbia River Basin salmonids and endangered ESUs. However further information is requested on how the proponents view the strategic balance of this project between assessing broad “offshore” distributions (where it appears more development work is needed as mentioned below) versus detailed monitoring to estimate survival between closely spaced reaches in the estuary. After several years of research the project is still in the process of demonstrating “proof of concept” of the effectiveness of the open coast arrays to detect tagged Columbia River and Snake River spring Chinook salmon (no other species have been evaluated). The project now recognizes some of its current limitations. For example, recent results (May 2010) showing incomplete detection histories for several jacks returning to the Columbia River in 2009 and two adults returning in 2010 have highlighted this uncertainty, together with the findings of a fairly uniform spatial distribution along the Willapa Bay sub-array to its current offshore limit of 250 meters. These observations affect a key assumption: 100% or consistent detection by the acoustic array. It will be important to evaluate how this issue could affect key studies involving mortality of transported versus in-river smolts. The project claims that its methodology is the only experimental technique available for addressing these issues, including early marine survival of salmon. While the approach is innovative and more direct, other studies have used incremental scale and otolith growth to examine size- and life-stage dependent mortality during specific periods at sea. Estimation of survival at sea is important for separating freshwater versus marine-related factors of survival. Smolt counts and coded-wire tags (CWT) have been traditionally used to estimate survival at sea. This project provides an estimate of mortality for yearling Chinook salmon smolts specific to each phase of early marine life, instead of release to recovery survival (CWT estimates). But results to date seem to show an exponential decline in survival with distance from the river, as expected. This seems to be the primary unique quality of this study of interest to managers. Migration rates are also unique, though other researchers have used short-term tracking to document travel rates. Interannual and seasonal variability in migration rates, estimated survival, etc. all need to be closely related to measured ocean conditions, either from ships or satellites. What have the proponent’s results shown so far? The proponent’s decay model of survival seems too simplistic, based on the research of others, and this needs further evaluation by the proponents. The low reported survival at sea is not surprising given the history of low survival rates of yearling and subyearling Chinook salmon based on CWT data. Chinook salmon are well-known to have lower ocean survival rates that other salmon species. The declining survival with distance from the Columbia is expected. For fisheries management, the key information is the evaluation of survival of in-river versus transported smolts. It would be of interest to compare results from CWT and PIT tagged salmon with those from this study and evaluate the benefit of the acoustic tag versus CWT and PIT tag for this management question because the acoustic tag approach is much more costly. 2. Strategic Plan for COAST Array Location. Does COAST have a strategic plan for COAST array location, developed in cooperation with other Columbia River projects that use or plan to use BPA-funded arrays for their projects? If not, such a plan needs to be developed. We reiterate our suggestion in past reviews that the proponents coordinate development of the COAST acoustic array design with other projects in the Columbia River Basin and ocean, as this issue was inadequately addressed in the proponent’s previous response. The proponents assume that Columbia River spring Chinook salmon migrate northward along a coastal corridor that is adequately sampled by the acoustic arrays. However, evidence exists for migratory patterns in other directions (southward, straight offshore). COAST proposes to remove the only listening line located south of the mouth of the Columbia River. The ISRP reiterates our previous recommendation that two ocean listening lines located to the south are needed to demonstrate the feasibility of this project. If COAST is to provide accurate estimates, arrays to the south of the Columbia River and additional tags to expand the proposed study to a 2-directional design are necessary. In addition, a closer examination of the location of arrays with respect to hypothesized locations of juvenile salmon survival bottlenecks is important to developing a strategic plan for potential future locations of arrays. 3. Coordination with other projects. What specific process is used by COAST to coordinate with other projects to estimate survival of Columbia River salmon? Coordination with other projects has improved, but it could be better. The proponents promise to tie in closely with the CDFO Shelf Survival proposal (#200300900) and the NOAA Ocean Survival of Salmonids proposal (#199801400). All three projects promise a key deliverable - survival. However, the coordination appears rather loose and further information on exactly how the three projects will work together is required. The proposal presents a possibly unbalanced review of VEMCO tags relative to JSATS, and no discussion is provided in reference to McMichael et al. (2010) regarding their survival estimates. Nevertheless it is encouraging to see the increased discussions and joint work with USACE contractors and others working on survival estimates in the lower river and estuary. The ISRP recommends increased coordination with JSATs research in the estuary, since all COAST smolts are proposed to be released below Bonneville Dam. A component linking COAST to the nearshore studies and restoration work in the estuary, however, is missing. As well, the inner estuary proposals (e.g., LCREP, #200300700) should be tied in to the propose COAST work. 4. COAST Study Design. What are the likely magnitudes of the effects of assumptions in the COAST study design on results and what are the consequences for conclusions? The ISRP reviewed the appendix attached to the proposal with extensive documentation of the power analysis and resulting study design implications for the POST project. The development of the likelihood approach is clear and the assessment of the results via simulation is useful. The explicit statement of assumptions is particularly appreciated by the ISRP. These assumptions include: • Survival per day is the same in the estuary, plume, and ocean • All surviving fish travel a given segment in the same time • Detection probabilities are the same for the groups being compared • Observed high and low survival rates bracket rates that are likely to be observed in future • Effect of dam passage or transport is fully expressed by one month after migrating. The validity of these assumptions may be debated, but it is clear that they have effects on results from this analysis, although the likely magnitude of these effects is not presented nor the consequences for conclusions. Nevertheless, it appears that the study design and power analysis presents an approach for planning tagging effort and array deployment. 5. Deliverable V. Testing the Delayed Mortality Theory. Can the proponents provide stronger justification for continuation of work on this deliverable? If the work continues, are there other more cost-effective methods for achieving the objective? The Independent Scientific Advisory Board (ISAB 2007-1) advised against continuing efforts to measure absolute latent mortality, suggesting instead that the focus should be on estimating processes such as in-river versus transport mortality that can be measured directly. Proponents acknowledge the ISAB recommendation but argue for continuation in part by citing Welch et al. (2008; (http://www.plosbiology.org/article/info:doi/10.1371/journal.pbio.0060265), a comparison of the un-dammed Fraser/Thompson River with the dammed Columbia. The ISAB (2007-1) concluded that determining latent mortality relative to a damless reference is not measurable. The argument in the proposal does not convince the ISRP that this ISAB conclusion warrants reconsideration. Can acoustic tags provide a more accurate and precise estimate of differential delayed (latent) mortality than a similar study approach that used greater numbers of coded wire tagged fish (at a much lower cost)? The acoustic tags estimate survival after a few months, but CWTs measure survival to adults. Has a comparison of the two approaches been made? If research on this objective continues, it would be important to incorporate survival of hatchery versus wild fish into the analysis. Will Chinook salmon tagged by COAST below Bonneville be identified as hatchery versus wild fish? The proposal notes that wild salmon tend to have higher survival rates; therefore, the ratio may affect the survival findings. What is the expected hatchery/wild tagging ratio? It would be interesting to compare data of tagged and untagged Chinook. Also, the study might compare survival rates with those from CWT salmon. This could tell us the fraction of mortality that occurs during early versus late marine life. 6. Detection Efficiencies. The ISRP has a number of questions about tag detection efficiencies that were not addressed in the proposal. What percentages of fish are detected only once, for example, and not again? Are these deemed mortalities or did fish residualize in areas outside of the detection range of arrays? Along the arrays in the ocean, what about fish that migrate close inshore where there are no receivers? And how often are receivers down or lost? On page 22 - the detection range for V7 tags is less than 300m. The detection probability for V7 tags is about 70%. The accuracy and precision of the estimates is questionable. It seems that COAST has given up a lot by going from the V9 to the V7 tag. The depth of a proposed new array at Cape Elizabeth would extend to 500m, but is this depth is beyond the detection range of the V7 tags? Are tagged fish easily detected if they are at or near the surface and the cable is in 500m deep water? What is the effect of wave action on detection of tagged fish? 7. Genetic stock identification (GSI). How many genetic stocks of juvenile spring Chinook salmon can be identified by the proposed GSI? Procedures for GSI need to be described. Proponents need to demonstrate that current techniques are capable of identifying origins of individual fish that are tagged and released. Ocean studies should advance toward designs that can also evaluate differences/similarities in survival of hatchery vs. wild fish of the same genetic stock. Is there a way to standardize genetic stock identification methods so that results of the three BPA-funded ocean projects are directly comparable (different labs are using different methods)? 8. Definition of the plume. Why is the plume defined as Sand Island to Willapa Bay? The proponents’ definition of the plume (Sand Island to Willapa Bay) is very different than accepted terminology, and the proposal would be improved by an explanation as why they chose this definition. The plume is usually described as outside the Columbia River bar, and the plume disperses both to the north along the Washington coastline and to the south along the Oregon coastline. See for example: Fielder, P.C., Laurs, R.M., 1990. Variability of the Columbia River plume observed in visible and infrared satellite imagery. Int. J.Remote Sens. 11, 999-1010. 9. Alternatives to Fixed Arrays. Are there other more innovative techniques than fixed acoustic arrays that could be employed in the future to track open coast and ocean distribution, migration patterns, and survival of Columbia River spring Chinook? For example, what about the use of robotic vehicles to measure ocean conditions and track tagged salmon to extend coverage beyond the detection range of fixed listening lines on the continental shelf/slope? 10. Scientific workshop. The ISRP recommends a scientific workshop in 2011 focused on estimation of estuarine and ocean survival, forecasting of adult returns, and adaptive estuary, plume, and ocean environmental assessment for Columbia River Basin salmon and steelhead. Perhaps the proposal should include this workshop. A workshop would help to improve coordination and collaboration, standardization of methods (e.g., genetic stock identification), development of simulation and predictive models, and integration of results among Columbia River Basin estuary/ocean projects. One aspect of all projects that needs work is how to include more detail on sub-stock structure, including hatchery versus wild fish, hatchery release time, area comparisons, in-river migration and associated ocean migration, and more in the models. 11. Adaptive management. Are the proponents overselling their ability to use this approach to improve real-time management of spill and transport? How can adaptive management with respect to estimates of ocean survival be implemented in the Columbia River system? Is it possible that tagging experiments could be designed in concert with hatchery, hydrosystem, and harvest managers to test specific hypotheses related to estuarine and early ocean survival? The proposal would be improved by further details on how POST results have influenced on-the-ground management decisions by fishery or hydrosystem agencies. For example, has the Welch et al. (2008) paper (“Survival of migrating salmon smolts in large rivers with and without dams”) resulted in any changes in operations of the Columbia River Basin dams? How do COAST indicators tie in with those being developed by CDFO, NOAA, and other projects in this review? In the past several reviews, the ISRP asked, “How would the fully-implemented ocean array and long-term monitoring data on seasonal and interannual variations in survival rates or migration rates among years or stocks actually be used by managers of the Columbia River Basin hydrosystem?” The ISRP agrees with the proponents’ past response that estimates of ocean survival for tagged release groups of hatchery fish can be used to inform policy makers, fishery managers, and researchers. However, the proponents have never answered the ISRP’s question about how hydrosystem managers would actually use the data. The proponents still do not seem to recognize that ocean variability will make the concept of tracking the geography of ocean mortality and subsequent adjustment of hydropower system very difficult to manage. The project is clearly significant to regional programs, but the proposal could be improved by attention to unrealistic objectives and expectations that implementation of acoustic tagging technology would result in improved real-time management of spill and transport. The proponents state that the latter two options could be decided upon by measurement of marine survival with their methods: “For example, if marine survival is exceptionally low, transportation and/or increased spill may not be beneficial, as smolts would reach the ocean sooner thereby exposing them to unfavorable ocean conditions (e.g., increased predation or decreased food supply), leading to lower survival.” Explain the specific processes that would be used to achieve real-time management. Do managers think this process would work? 12. Communicating Results. Can the proponents develop more effective approaches for communicating their results directly to Columbia River Basin hatchery, hydrosystem, and harvest managers? Websites, scientific meetings, and peer-review scientific publication are excellent methods for communicating with scientific peers, other government agencies, educational institutions, and conservation organizations, but are likely not effective tools for communicating directly with hydro, harvest, and hatchery managers. 13. Update Online Proposal Format. The format of this proposal was confusing and difficult to follow. Proponents should reformat their online proposal to better conform to the specific information requested in each section of the online form. The repetition of the same deliverables under several objectives seems unnecessarily repetitive. Objectives providing the same deliverables could be combined into one objective. The important information on study design that was included only as an attachment should be incorporated into the online form. The online form should present the complete proposal as a stand-alone document.

Compliance with RPA 58.1 (Monitor and evaluate smolt survival and/or fitness in select reaches from Bonneville Dam through the estuary)

Kintama in its 2011 and 2012-14 proposed work has had to balance the priority of the following factors within the allotted budgets:

1. Certainty of survival data. That is, are tagged fish are representative of run of the river fish, and are all surviving fish detected on our coastal ocean sub-arrays?

2. Precision and power of survival data. Can we estimate survival to a level that can be of use in hydrosystem management decisions?

Certainty of survival data

Kintama’s consistent focus has been on possible tag effects (we are currently writing up our analysis of a salt-water exposure tag effect experiment carried out in 2010) to ensure that tagged fish do not shed tags or die as a result of tagging, thus underestimating survival. Recent evidence from our ocean tracking studies (2008-2010) however, indicate that a small proportion of tagged smolts may have migrated further offshore, beyond the limit of the acoustic sub-arrays, which may also lead to underestimated survival estimates. The offshore extent of the COAST sub-arrays was based more than a decade of NOAA and other seine and trawl surveys; however, recent trawl data also indicates that smolts may migrate farther offshore than had been observed in the past. We therefore proposed to extend these arrays from the 200 m isobath to the 500m isobath in our 2011 and 2012-14 proposals.

Recognizing that smolts may also migrate south (which is another factor that may lead to underestimated survival), Kintama deployed a sub-array south of the Columbia River at Cascade Head in 2009.  Only three fish (out of 1600) were detected on this line, and all three of these fish were from an experimental group of early released transported fish.  On this basis, redeploying this array was determined to be a lower priority than other options for 2011-2014.  However,

i)  Information received the week of Nov 29^th 2010 from NOAA and PNNL suggests that for 2010 at least a higher portion of fish went south in the plume than perhaps in previous years.  Whether some or all of these fish then went north once they exit the plume is unknown.

ii)  In 2009, Kintama tagged Cle Elum (Yakima River) and Dworshak (Snake River) hatchery fish.  It is likely that the behavior of these individual fish stocks is not representative of run-of-river fish Kintama is now tagging (in 2010 Kintama tagged run-of-river fish at Lower Granite and John Day dams).

iii)  The ISRP review of our 2012-2014 proposal very much favors at least one array to the south of the Columbia.

Precision and power of survival data

To obtain finer scale early ocean survival data that can be directly associated with NOAA data on ocean conditions (the long-term aim being to develop ocean triggers, based on actual smolt survivals, that can be used for management purposes), it was proposed for 2012-14, and hence 2011, to add an array between the existing arrays at Willapa Bay (WA) and Lippy Point (BC) – likely at either Cape Elizabeth or Cape Alava on the Washington coast.  However, to accommodate the expense of the new array, and the extension of all of them out to the 500m isobath, it was necessary to reduce the number of tagged fish for 2011 to 800, balanced by the fact that the majority of these would be tagged and released at Bonneville Dam, thereby removing the impact of losses through the mid-upper Columbia and Snake rivers. Kintama conducted extensive power analyses to determine that given a predicted detection probability and survival rate, it is feasible to detect survival differences in two tagged experimental groups. 

Balancing priorities

When we completed the 2012-2014 and 2011 proposals, an optimized balance, supported by our contract managers at BPA, seemed to be to prioritize a new array north of the Columbia, rather than redeploying the southerly Cascade Head array.  However, given the new information we have recently received, it is likely more imperative to reverse that decision, for 2011 at least, in order that we can groundtruth the extent of southern migration of smolts based on the tagging of run of the river fish.  This means that the power of detecting a difference in survival of our experimental groups will be reduced for northerly migrating fish because of the removal of the Cape Elizabeth/Alava array (survival will be estimated to the northern most array at Lippy Point on Vancouver Island, therefore sample size will be greatly reduced as a result of early marine mortality).  We have made this proposal to our BPA contract managers and are awaiting their review.

Deleted association with RPA 61.1 (Continue work to define the ecological importance of the tidal freshwater, estuary, plume, and nearshore ocean environments to the viability and recovery of listed salmonid populations in the Columbia River Basin)

Kintama is unclear why this has been deleted.  It seems to us that comparative smolt survival data for the hydrosystem, estuary, plume and coastal ocean, particularly if linked to other physical/chemical/biological measurements, does help to define the ecological importance of each zone.

The proponent has provided adequate responses with some notable exceptions as mentioned below. The ISRP's initial (June 2, 2006) review remains largely unchanged. The ISRP continues to recommend that this project be funded in part at a reduced level of funding and deployment of the proposed acoustic tracking arrays, until the proponent's results can demonstrate "proof of concept" of the effectiveness of the open ocean sites to detect tagged Columbia River and Snake River spring Chinook salmon. Results of the 2004 and 2005 field seasons were inconclusive because of incomplete coverage of the continental shelf on the Cape Elizabeth and Brooks Peninsula lines. In addition, detection efficiencies could not be calculated due to significant loss of receivers on the Cape Elizabeth and Brooks Peninsula lines (only 18 of 26 units recovered), as well as the lack of detections on the Alaska line. Somewhat troubling is that BPA-sponsored listening lines installed in previous fiscal years have already required replacement by new lines and new technologies. The ISRP appreciates that our previous comments about placing arrays in the estuary and plume have been used by the proponent to adjust his research. An approach tailored to Columbia River and estuary needs is now apparent. This aspect of the work should be emphasized and more collaboration encouraged between the proponent and other researchers working in the lower river, estuary, and ocean.

The ISRP advises reducing (from 4 lines to 1 line) the number of proposed new listening lines on the open ocean coast. We reiterate our previous recommendation that only four open ocean listening lines (two located north of the Columbia River mouth and two located south) are needed to demonstrate the feasibility of this project. Three of these BPA-sponsored lines have already been funded in FY 2006 (Willapa Bay, WA; Lippy Point, BC; Cascade Head, OR), and installation of a second line south of the Columbia River mouth at Tillamook, OR, is proposed for FY 2008. Scientific justification is not adequate for installation of additional new BPA-sponsored lines in the open ocean at Graves Harbor, AK (FY 2007; 23 nodes), Cape Alava, WA (FY 2009; 80 nodes), and Coos Bay, OR (FY 2009; 31 nodes). The ISRP does not recommend funding permanent upriver acoustic listening lines (above Bonneville Dam). However, the proponent's response notes that upriver arrays have already been installed as part of the 2006 work plan. The ISRP advises that upriver research described in the proponent's response to compare different tagging technologies (PIT tags vs. Vemco acoustic tags) is well beyond the original biological objective of this project, i.e., "tracking smolts in the ocean to resolve how to better manage the Columbia hydropower system." The ISRP reiterates its previous suggestion that the proponent coordinate development of the final acoustic array design with other projects in the Columbia River Basin and Plume, as this issue was inadequately addressed in the proponent's response.

Additional comments are as follows:

1. The ISRP stated that its "primary concern is that results to date indicate effectiveness of detecting tagged juvenile salmon along open coast arrays is not always high . . ." In the open ocean, survival rates can be estimated only if all juvenile salmon movements are confined within the area of the continental shelf where acoustic listening arrays are located . . . " The proponent responded, "We believe we have addressed the question of a significant offshore movement of smolts through a separate manuscript now in review." This response was not adequate, as the data in this manuscript were not provided to the ISRP. In addition, the proponent's response "that the Juan de Fuca line . . . showed that Snake River spring chinook do not use that potential migration route" is not in agreement with data in the 9 January 2006 proposal (Fig. 4), which show the detection of a Snake River Chinook on the Juan de Fuca line in 2005. Was this a false detection?

2. Table 1 of the proponent's response is useful supplemental information to Table 2 of the 9 January 2006 proposal, because it provides data on the specific BPA-sponsored arrays proposed for 2007-2009. However, a prioritized list including data on equipment and maintenance costs, as requested by the ISRP, was not provided. Information on month of deployment would also have been useful, as it is not clear whether new arrays would be installed in time to detect releases of tagged fish in the year of deployment. The project design would be improved if installation of the second array south of the mouth of the Columbia River (presumably at Tillamook, OR, not "WA" as listed in Table 1) occurred at the beginning of the proposed project (early in 2007) before tagged smolts are released. This would provide three years of data at two stations south of the mouth of the Columbia River rather than only two years of data. If the FY 2007 results show that tagged Columbia/Snake spring Chinook smolts are detected at the outermost nodes, then curtain lengths of the arrays would need to be extended well beyond the 200-m isobath.

3. The ISRP asked, "How would the fully-implemented ocean array and long-term monitoring data on seasonal and interannual variations in survival rates or migration rates among years or stocks actually be used by managers of the Columbia River Basin hydrosystem? The ISRP agrees with the proponent's response that estimates of ocean survival for tagged release groups of hatchery fish can be used to inform policy makers, fishery managers, and researchers. The proponent did not answer ISRP's question about how hydrosystem managers would actually use the data. The proponent mentioned the possible over-emphasis of other past projects on freshwater mortality. A balanced approach would consider habitat and environment needs for the community of salmonid fishes, which after all show a wide diversity of life history types. For example, even very good ocean conditions apparently did not enable survival of sockeye in Redfish Lake.

4. The ISRP asked, "Are the proponents relying on these other studies (DFO "Canada-USA Salmon Shelf Survival" project #200300900 and NOAA/NMFS "Ocean Survival of Salmonids" project #199801400) to provide data needed on ocean conditions . . . that might affect survival? The proponent responded, "The goal of the POST project is not to address how the fish die, but to provide hard numbers on where the mortality occurred—and how great the mortality actually is." The ISRP notes that the "hard numbers" will be estimates (statistical probabilities) of survival of two hatchery stocks of spring Chinook salmon. Collaboration with other projects would provide multiple lines of scientific evidence based on different methodologies, including mechanistic approaches and results to explain causality. This would strengthen support for the proponent's hypotheses about the relations between fish passage over dams, barging, and ocean survival of Columbia River salmon. The proponent's response used partial preliminary unpublished data from POST lines to refute alternative hypotheses and technological approaches of other projects, which is not good scientific methodology, even though parenthetical cautions were provided.

Although the emphasis in this research is survival, as stated in the response, and not the causes per se, the ISRP considers it important to note that the estimated ocean locations or ages of high or low survival of Columbia River Basin salmon may not be the same in different years because of ocean variability. Therefore, it will be important to correlate minimal ocean survival rates with ocean conditions in the future by collaborating with other research programs. The proponent does not seem to acknowledge that ocean variability will make the concept of tracking the geography of ocean mortality and subsequent adjustment of hydropower system management very difficult to operationalize. For example, the proponent's response regarding one year of results along a Kintama-sponsored Alaska line: "No Snake R. smolts appear to have migrated over the (Alaska) line, providing a very useful boundary on where the Snake R spring Chinook survival problems must occur." The ISRP advises that this "boundary" is not a fixed line in the ocean.

The ISRP asked, "What specific efforts are underway by the proponents to collaborate with these and other BPA-funded estuary, plume, and ocean studies on salmon survival?" The proponent responded, "We look forwards to closer collaboration in future as POST is proven and we can devote greater time to looking at the linkages." The ISRP advises that the achievement of common biological objectives of the various BPA-funded ocean distribution and survival projects would benefit from better coordination. The ISRP also reiterates its previous suggestion that the proponent coordinate development of the final acoustic array design with other projects in the Columbia River Basin and Plume, as this issue was inadequately addressed in the proponent's response.

5. The ISRP noted that "survival rates will be calculated as a combination of mortality, non-detection, and tag shedding," and asked: "Can the proponents distinguish between detections of tags in live salmon, tags in dead salmon that are drifting with the current, and tags in live predators that ate tagged salmon?" The ISRP agrees with the proponent that a technological solution (mortality sensor) to distinguish between tags in live salmon vs. dead salmon is not feasible at this time. More to the point, the response would have been improved if the proponents had provided information on the acoustic data analysis or interpretation methods that they use to distinguish between tags in live and dead salmon.

6. The ISRP asked for an evaluation of the effect of the acoustic tags on the behavior and survival of spring Chinook salmon smolts. The response partially addressed the ISRP's concerns about behavior by presenting data from an experiment on coho salmon (Chittenden's M.Sc. thesis), but did not adequately address Chinook salmon survival over the period of study for the V6, V7, and V11 tags. The proponent's response included useful information on new Vemco V7 (7 mm) and a 6-mm acoustic tag (Vemco-developed by spring of 2007) for use on smolts down to approximately 10-10.5 cm in length. These tags will have at least a four-month life span, but the geometry of the array's nodes will have to be re-configured to achieve a high detection efficiency for 6-mm tags. The ISRP notes that the size of the 6-mm tags will still limit the data from this project. The proponent provided a letter documenting good cooperation and involvement of hatchery managers in the project but did not respond to ISRP's request for more detailed methods, timelines, and schedules for releases of tagged smolts from the two hatcheries participating in the project (Kooskia National Fish Hatchery and Chandler Juvenile Monitoring Facility). The ISRP advises that differences between hatcheries in rearing and release conditions and schedules could affect experimental results.

The ISRP asked, "How comparable is the ocean distribution of tagged Snake River hatchery fish to wild Snake River Chinook? Is there a size difference? If so, how much will this influence their results and interpretation?" The proponents did not answer this question adequately: "To our knowledge, the answer to this question is currently impossible to ascertain. We hope to address such questions with the POST array over time." Surely the literature could have provided at least a partial answer to this question. Size data are published and extrapolation from Chittenden's thesis work could have been interpreted.

7. In response to ISRP's request, the proponent provided useful and detailed information on permits and permitting processes required to deploy the POST array on the ocean floor. However, the response did not demonstrate ISRP-requested coordination and cooperation with coastal fishing communities through Washington, Oregon, and Alaska Sea Grant.

8. The ISRP requested more information on the proposed method for recovering lost acoustic receivers. Previous ISRP reviews raised concerns about detecting lost receivers and the use of expensive ROVs and side-scan sonar. The proponent's response was informative with respect to problems with acoustic releases. The proponent stated, "as a percentage of the POST array, operations costs for ROVs are reasonable, and the POST array's data is invaluable." Although requested by ISRP, a breakdown of these costs was not provided.

9. The ISRP asked, "How will the data from other investigators who used VEMCO tags be made available to them and at what cost? How will VEMCO and Kintama facilitate other research programs that want to use the coastal receiver network?" The proponent noted that Kintama would probably handle scientific consultation and financial charging for use of the POST array by other researchers who own Vemco tags. It is not clear, however, if other BPA-funded projects that want to use the BPA-sponsored listening lines will be also be charged a fee for these services. Charging (the cost of membership) for use of POST array is troublesome given the significant BPA funding. The use of BPA-funded lines by other researchers should be specified by the proponent and evaluated by the Council and BPA.

The proponents remain optimistic that State, Federal, Provincial, and International agencies will buy into the idea of a continental-scale array and support it in the long term. The cooperation of these agencies is key to the long-term success of POST in this part of the ocean. However the difficulty of continuing long term and expensive monitoring in the ocean may be underestimated.

10. The ISRP noted that "justification for expensive equipment described in the narrative was insufficient" and asked: "What are the specific costs of tags and acoustic nodes? What are the costs of the ROV and additional equipment needed for the ROV, including high-resolution optics, and manipulator, plus surface electronics? What are the projected costs for the single special-purpose vessel that may be required in the future? What are the costs for the wireless (cell, satellite) communications, and other marine electronics? Are these costs shared with other programs funding the POST array? If so, how is BPA's share determined?" The proponent did not provide the requested estimates of specific costs for expensive equipment.

11. The ISRP requested justification for the PI's allocation of 100% FTE to this BPA-funded project. The proponent's response explained that FTE is allocated "between the various POST project sponsors" and "is difficult to precisely define." The ISRP is concerned that FTEs allocated to the proposed BPA-funded project will not be adequate. The ISRP notes that there is a patchwork of FTEs and associated costs that cannot be explained.

Qualifications: The ISRP's "Fundable in Part" recommendation is qualified because the response and proposal were unresponsive with regard to several critical elements of collaboration and cost. These elements are described in the comments and eleven items above. The ISRP recommends that the Council and BPA secure this information as part of the final project selection process and development of this project's statement of work, if funding is continued.