Testimony of Nicolaas Bouwes Biometrician, Oregon Department of Fish and Wildlife
Before The Subcommittee on Fisheries, Wildlife, and Water Committee on Environmental and Public Works United States Senate
September 14, 2000 Washington, D.C.

Background

Current management of the hydrosystem is guided by a Biological Opinion on 1994-1998 operation of the Federal Columbia River Power System and its supplement for listed steelhead (1995 Biological Opinion). The 1995 Biological Opinion contains specific measures for operating and improving the configuration of the hydrosystem in the near term. However, it deferred decisions about the long-term future operation and configuration of the hydrosystem until late 1999, when it anticipated the completion of a formal and deliberate assessment of three general alternatives. The decision about the long-term future state of the hydrosystem was deferred until 1999 because of significant uncertainties associated with the projected likelihood of survival and recovery of listed fish under each option.

This assessment, referred to as the Plan for Analyzing and Testing Hypotheses (PATH) was shaped by federal District Court orders arising from a challenge of the legal adequacy of the Biological Opinion on the 1993 operation of the Federal Columbia River Power System (1993 Biological Opinion) by the Idaho Department of Fish and Game and the State of Oregon, joined by four treaty Indian tribes. They argued the chosen jeopardy standard and the consideration of the reasonable and prudent alternatives (RPAs) to avoid jeopardy were arbitrary and capricious and otherwise not in accordance with the purposes of the Endangered Species Act (ESA). The District Court agreed and set aside and remanded the 1993 Biological Opinion and records of decision to the federal defendants with instructions that they review and reconsider them. Rather than reconsider the challenged 1993 Biological Opinion, the federal defendants opted to reconsider the newly issued 1995 Biological Opinion. The following District Court orders guided their efforts:

1. NMFS must consider relevant facts and articulate a rational connection between the facts found and the choices made. These choices included the choice of a standard, for which the District Court expressly rejected any attempt to impose bright-line definitions of survival and recovery. Instead, the District Court stated that, with respect to listed Snake River salmon, survival and recovery are virtually indistinguishable. 2. NMFS must conduct a reasoned evaluation of all available information. The District Court found that NMFS arbitrarily and capriciously discounted low range assumptions without well-reasoned analysis and without considering the full range of risk assumptions. This was particularly problematic given the enhanced risk associated with the small size of listed Snake River salmon populations. 3. NMFS must substantively consider significant information and data from well-qualified scientists such as the fisheries biologists from the states and tribes. The District Court directed NMFS to provide analysis and reasoned evaluation of submissions by such qualified scientists, with any rejection of such submissions thoroughly explained. 4. NMFS must provide sufficient reasoned analysis of its consideration of alternatives and measures [for operation of the hydrosystem] to permit judicial review.

In response to the District Court's findings, NMFS agreed in a Joint Statement of the Parties, filed with the Federal Report of Compliance, to several coordinating principles. Three of these principles are particularly germane to the purpose of PATH and led to its development.

1. NMFS, for development of its hydrosystem biological opinion concerning the listed salmon, will use a regional analytical work group, including state agencies and the Columbia River Inter-Tribal Fish Commission, to provide technical analysis of biological parameters affected by fish passage through the hydrosystem and impacts on other portions of their life cycle. 2. The federal action agencies or NMFS will provide state agencies, tribal governments, and others as they deem appropriate, with a reasonable opportunity to provide new scientific and technical information on a draft biological opinion(s). 3. Federal power system operators, the U.S. Fish and Wildlife Service (FWS) and NMFS will provide state agencies and tribal governments and others with an opportunity to meet to discuss the analysis of the expected effects of proposed actions in biological assessment(s) and biological opinion(s) before final decisions are made. In this regard, the federal action agencies, FWS and NMFS will make available to state agencies and tribal governments and others documents containing data, analysis, and other information upon which the biological assessment and biological opinion rely.

The PATH process was developed through a collaborative process and adopted by NMFS in 1995 to provide a biological framework for decisions concerning the listed Snake River salmon and steelhead, and most PATH analyses were completed in 1998. The PATH forum is an inclusive, regional analytical work group developed to provide technical analyses of biological parameters affected by fish passage through the hydrosystem and impacts on other portions of their life cycle. The PATH analyses evaluated factors responsible for the decline of ESA listed Snake River salmon and steelhead (retrospective analysis), and described a range of possible responses to alternative management actions (prospective analysis). The range of population responses to each management action described the ability and uncertainty in meeting the 1995 ESA jeopardy standards developed by the Biological Requirements Work Group (BRWG). In a memorandum from Randall Peterman, a world-renowned fisheries biologist reviewing the PATH process, to the NMFS chaired Implementation Team, stated "it is fair to say that the PATH process, is the most comprehensive analysis of alternative hypotheses and management options that I have ever seen, heard about, or read about."

The PATH retrospective analysis concluded that the most likely factor responsible for the decline of Snake River spring/summer chinook was the development and operation of the lower Snake River hydrosystem. PATH found that the management action most likely to improve Snake River salmon survival enough to meet the jeopardy standards over the greatest range of hypotheses, was breaching of four lower Snake River dams. The improved hydrosystem operations and increased transportation options did not meet the jeopardy standards over the majority of hypotheses evaluated.

In 1999, NMFS announced their intention to delay the 1999 Biological Opinion to 2000. NMFS also announced the beginning of their new analytical process, the Cumulative Risk Initiative (CRI). In the Lower Snake River draft Environmental Impact Statement, NMFS stated in their Anadromous Fish Appendix, that "The CRI approach cannot replace PATH's detailed examination of modifications of transport or fish passage systems, and is not intended to do so." Instead, NMFS has stated that they elected to move away from the PATH process to the NMFS' CRI process because they needed a tool flexible enough to evaluate the impacts of hydro, habitat, hatcheries, and harvest on all listed stocks. We agree that these additional analyses are necessary; however, the established PATH process could have addressed these needs if the Implementation Team, who assigned PATH their analytical tasks, had deemed them necessary. In addition, PATH made considerable progress in investigating the improvements that might be expected from habitat restoration, alternative harvest reductions, and estuary mortality reduction in much greater detail than has been attempted by CRI. In the draft Biological Opinion, NMFS' has ignored the PATH findings and has relied on the CRI for the Snake River listed stocks and the Quantitative Analysis (QAR) for the mid-Columbia listed stocks. The CRI only evaluated "modifications of transport, or fish passage systems" and harvest for the Snake River listed stocks. Therefore, CRI is not used for the purposes NMFS has given to abandon PATH, but only to "replace PATH's detailed examination." The draft Biological Opinion should include a description of why the PATH process and their findings that were meant to provide the analytical basis for the 1999 (2000) Biological Opinion were abandoned in exchange for what NMFS admits is a less "detailed examination".

Review of the draft Biological Opinion

A great deal of effort has been made by Oregon Department of Fish and Wildlife to understand the 700 pages of draft Biological Opinion that describes the analytical approaches and rationale developed for the future operation of the FCRPS to ensure the survival and recovery of the 12 listed ESU salmon and steelhead. This review occurred over the last 6 weeks, and in general we are concerned that the conservation burden of the federal hydropower system has not been adequately defined and has been inappropriately assigned to harvest, hatcheries, and habitat programs. The states and tribes should not shoulder the mitigation responsibility of the federal hydropower system, nor should the responsibility be shifted from the mainstem to the tributaries and estuary without a full accounting of what limits the ability of the hydropower system to meet its mitigation responsibility. The following comments highlight shortcoming of the draft Biological Opinion that has lead us to this conclusion.

In general, the BiOp:

1) overestimates probabilities of survival and recovery for listed salmon and steelhead 2) underestimates survival improvements necessary to ensure the survival and recovery of listed salmon and steelhead 3) overestimates or, for some measures, does not estimate improvements to survival resulting from implementation of the Reasonable and Prudent Alternative (RPA) 4) describes an RPA for operation of the federal hydropower system that does not significantly change the status quo 5) does not adequately explain whether the success of the RPA can be confidently described by proposed performance standards and measures after 5-8 years 6) does not anticipate and have ready an alternative RPA, if the proposed RPA does not produce survival improvements necessary to ensure the survival and recovery of listed salmon and steelhead

Specifically:

1) The BiOp overestimates the probability of survival and recovery for listed salmon and steelhead because analyses are based on optimistic assumptions.

a) optimistic assumptions: The BiOp evaluates jeopardy using only those assumptions that present an optimistic view of the status of listed salmon and steelhead. Assumptions used in the BiOp are not based on the weight of evidence. Nor, in the absence of evidence, are they conservative, i.e. they do not avoid placing undue risk on the listed species.

b) extinction threshold: The BiOp evaluates jeopardy using the probability of an absolute extinction of 1 fish/brood. In reality, populations are at significant risk of extinction well before abundance declines to 1 fish/brood. The National Marine Fisheries Service (NMFS) points this out in their description of Viable Salmonid Populations (McElhany et al. 2000). The Biological Requirements Work Group (BRWG) that NMFS formed to set threshold population levels for survival and recovery of listed salmon and steelhead also points this out. Using an absolute extinction of 1 fish/brood as the survival threshold under-estimates the probability of real extinction for the listed species.

c) definition of high risk: The BiOp evaluates jeopardy by defining high risk as a 5% probability of extinction in 24 and 100 years. This is inconsistent with the definition of high risk previously described by NMFS in the Anadromous Fish Appendix of the US Army Corps of Engineers' Environmental Impact Statement for juvenile fish passage improvements at federal projects in the lower Snake River. In the Appendix, NMFS defines high risk as a 1% probability of extinction in 100 years. Relaxing the definition of high-risk under-estimates the probability of real extinction for the listed species.

d) base time period: The BiOp evaluates jeopardy using a base time period that only includes stock status information for the years after the federal hydropower system was constructed. The evaluation also uses stock status projections (returns that have not occurred) through 2004 in an attempt to reflect affects of recent good ocean conditions. By not including years before construction of the hydropower system, and by including stock status projections for future years, the BiOp under-estimates the decline in population abundance coinciding with construction of the hydropower system, and also over-estimates the probability of survival and recovery.

e) population summary statistic: The BiOp evaluates jeopardy using a metric for population growth that assumes a linear decline in population levels. Evidence suggests that declines in population levels are non-linear (Oosterhout 2000). In failing to correct for a non-linear decline, the approach over-estimates the probability of survival and recovery.

f) hatchery effectiveness: The BiOp evaluates jeopardy based on the assumption that hatchery effectiveness is low. Evidence suggests that hatchery spring and summer chinook that spawn in the wild in the Snake River may be as effective as wild spawners. Assuming hatchery effectiveness is low over-estimates the productivity of listed stocks, and consequently, over-estimates the probability of survival and recovery.

g) density dependence: The BiOp evaluates jeopardy based on the assumption that there is no density dependence, i.e. that populations can grow exponentially without limit. This assumption may be reasonable at low population levels, but not at population levels that approach recovery. Assuming no density dependence over-estimates productivity, and consequently, the probability of recovery.

8) The BiOp underestimates the survival improvements necessary to ensure the survival and recovery of listed salmon and steelhead.

a) necessary survival improvements: Because the BiOp bases its evaluation of jeopardy on optimistic assumptions that over-estimate the probability of survival and recovery, estimates of the necessary survival improvements are too low. Consequently, the BiOp concludes that to meet the 24-year survival standard, necessary survival improvements for Snake River spring and summer chinook are less than 30% over the life-cycle. This is an order of magnitude less than estimates of over 740% by Peters and Marmorek (2000) and of 280 to 850%, based on smolt-to-adult ratios needed to meet the 24-year survival standard used in the 1995 Biological Opinion.

b) delayed mortality: The BiOp evaluates jeopardy, for some stocks, using a "full mitigation" standard that is equivalent to survival through a natural river. This full mitigation standard was calculated based on the assumption that there is no delayed mortality of fish traveling through or transported around the federal hydropower system. This assumption is not consistent with the direct evidence that delayed mortality exists and the indirect evidence that delayed mortality is substantial (NMFS 2000, Bouwes 1999, Schaller et al. 1999, Marmorek and Peters- SRP 1999, Marmorek and Peters 1998, Marmorek et al. 1996). Assuming no delayed mortality under-estimates mortality related to the federal hydropower system, and consequently significantly lowers the full mitigation standard. This, in turn, underestimates the survival improvement needed to meet the standard.

3) The BiOp overestimates, or for some measures, does not estimate improvements to survival resulting from implementation of the Reasonable and Prudent Alternative (RPA).

a) improvements in survival of juvenile salmon and steelhead: The BiOp evaluates jeopardy based on the assumption that estimated improvements in survival of juvenile salmon and steelhead are primarily the result of the measures implemented under the 1995 Biological Opinion, and included as part of the proposed action. These improvements could be a result of using data from recent high flow conditions or an artifact of using different models to describe the base conditions.

b) improvements in survival of adult salmon and steelhead: The BiOp evaluates jeopardy based on the assumption that the RPA reduces losses of adult salmon and steelhead caused by the federal hydropower system by 25%. No data or analyses are presented to support this assumption.

c) hydropower system responsibility: The BiOp does not adequately explain why certain assumptions were used, and not used, to determine the level of impact attributable to the federal hydropower system. The BiOp relies on assumptions that require the least amount of hydropower system improvements by selecting "best case" scenarios.

d) survival improvements from harvest, habitat and hatchery measures: The BiOp evaluates jeopardy based on the assumption that "the greatest opportunity for survival improvements may lie outside the scope of the hydropower corridor". This assumption is based on misleading "numeric experiments" rather than analyses of feasible management actions. No data or analysis is presented to support the conclusion that necessary survival improvements can be achieved from harvest, habitat and hatchery measures. Nor is there an assessment of risks of extinction and associated uncertainties under these measures.

1) harvest rates: The BiOp appropriately concludes that for wild Snake River spring and summer chinook, further harvest restrictions will not produce significant survival improvements and sets the overall fishery impact standard at the spring season 2000 level of 6-9%, which is a similar impact rate to the level of 6-10% set by NMFS and captured in U.S v. Oregon Management Agreements, 1996-99. It inappropriately indicates the majority (if not all) the spring and summer chinook impacts could be allocated to the Treaty Indian tribes because of federal trust responsibility and the federal view that tribal harvest has a priority legal standard over non-tribal harvest. The parties to U.S. v Oregon negotiate Treaty Indian and non-Indian harvest sharing. A non-Indian impact level of 1-3% is considered the minimum to conduct non-Indian selective fisheries on abundant Willamette and Cowlitz hatchery-stock spring chinook.

2) harvest measures benefits: The BiOp implies benefits from harvest restrictions on listed stocks other than Snake River spring and summer chinook, but fails to point out those restrictions must remain in place for decades, and that some require agreement with Canada.

3) selective fisheries: The BiOp does not clearly explain that while much focus of selective fisheries will be towards hatchery origin fish, selective fishery opportunities are available for healthy wild stocks (e.g. Mid-Columbia sockeye and upriver bright fall chinook salmon).

4) fishery effort reduction program: The BiOP does not clearly explain whether buyouts of commercial fishing licenses and permits are voluntary.

5) hatcheries: The BiOp suggests changes to artificial production programs, but only qualitatively assesses how changes will affect listed salmon and steelhead. In addition, the assessment erroneously attributes potential survival improvements to monitoring and evaluation of artificial production programs.

6) habitat: The BiOp does not describe specific measures for habitat protection and restoration in subbasins, nor does it include measures to increase mainstem spawning habitat for fall chinook in impounded reaches. It also does not explain how necessary survival improvements for Snake River spring and summer chinook will be achieved, given that it concludes that habitat measures offer little potential improvement and assigns Snake River subbasins a low priority.

g) feasibility of timely implementation: The BiOp evaluates jeopardy based on the assumption that harvest, hatchery and habitat measures are timely implemented and produce near-term survival improvements. However, it neither evaluates the feasibility and risks of implementing any of these measures, nor offers a "game plan" to ensure timely implementation. Survival improvements from habitat measures likely would not be realized for decades.

8) The BiOp describes an RPA for operation of the federal hydropower system that does not significantly change the status quo. The BiOp does not acknowledge that many measures in the 1995 Biological Opinion were not implemented as intended, or at all, for various anticipated and unanticipated reasons. As a result, the BiOp does not assess the likelihood that individual measures in the proposed RPA will be fully implemented.

a) flow: The BiOp does not designate meeting flow needs of listed salmon and steelhead as at least an equal priority with other uses of the water (e.g. power generation). It does not aggressively seek, nor does it describe steps to acquire additional volumes of water necessary to meet flow targets.

b) transportation: The BiOp does not acknowledge the considerable uncertainty in the potential benefits of transportation. As a result, it does not adequately spread the risk between transporting listed salmon and steelhead and leaving them to migrate in river by limiting the percentage of fish transported to no more than 50%.

c) spill: The BiOp reduces spill at The Dalles Dam from 64% to 40%, despite the fact that no statistically significant results exist that indicate the need for the change.

4) The BiOp does not adequately explain whether the success of the RPA can be confidently described by proposed performance standards and measures after 5-8 years.

a) population summary statistic (l): Although it is appropriate to use a life-cycle summary statistic such as l as a performance measure, using l alone may not incorporate variability. The BiOp does not clearly explain whether and how it incorporates variability in its measurement of performance. The BiOp also does not clearly explain whether the time series used to estimate l is the 1980 to newest years or just the newest years.

b) performance measures: The BiOp does not clearly explain whether it will use consistent methods to compare performance before and after implementation of the RPA.

c) evaluation of uncertainty and error: The BiOp does not evaluate the feasibility of resolving uncertainty, or assess whether the analytical approach will be able to reject the null hypothesis that the RPA results in no survival improvement over current measures. It does not describe feasible experimental design options to manage uncertainty and error.

d) experimental management: The BiOp does not clearly state whether it embraces approaches that evaluate the value of what we can learn from efforts to ensure the survival and recovery of listed salmon and steelhead.

5) The BiOp does not anticipate and have ready an alternative RPA, if the proposed RPA does not produce survival improvements necessary to ensure the survival and recovery of listed salmon and steelhead. The midpoint evaluation is not aggressive enough to avoid jeopardy given the unstated and likely great uncertainty of the RPA and the high probability of extinction.

a) The BiOp does not adequately assess the likelihood of recovery under an alternative RPA (e.g. dam-breaching) after 5 to 8 years, if the proposed RPA does not significantly improve survival (l>0.95). If survival does not improve or continues to decline over the time period, extinction of certain populations may be unavoidable under any action.

b) The BiOp does not assess the lead time needed to implement an alternative RPA, nor does it describe what needs to be done in the interim to ensure timely implementation. It does not describe steps that must be taken now to satisfy NEPA requirements, get congressional authorization, complete mitigation planning etc, and have an alternative RPA ready to go, if needed. These steps could take 5-8 years after their initiation to complete.

c) The BiOp cites significant uncertainty in survival improvements from dam breaching as a basis for deferring its consideration until some point in the future. However, the biological decision analysis completed as part of the Plan for Analyzing and Testing Hypotheses (PATH) project concluded that the benefits from dam breaching were more certain than non-breaching alternatives. The BiOp does not describe decision criteria it would use or the approach it would take to resolve conflicting assumptions, especially with respect to delayed mortality, and reduce uncertainties associated with the proposed RPA or an alternative RPA.

d) The BiOp cites the fact that only Ecologically Significant Units (ESUs) in the Snake River benefit from dam breaching as a basis for deferring its consideration until some point in the future. However, a number of measures in the proposed RPA only affect certain ESUs. The BiOp does not describe why this criterion is valid for one potential measure and not others. Problems with Snake River ESUs are not less significant because other salmon populations have subsequently been listed as threatened or endangered.

Conclusion

The perilous state of these ESA listed stocks is real; last year in two of the Snake River spring/summer chinook indicator stocks that spawn in pristine wilderness areas, zero fish returned. We may have already lost Snake River sockeye, and coho have gone extinct in the Snake River basin since efforts have been made to mitigate for the FCRPS. In the opinion of the ODFW, based on our assessment of the current data and analyses, the draft Biological Opinion does not use the best available scientific information to determine the management actions most likely to recover Snake River ESA listed stocks.

The problems highlighted above underestimate the true risks to these stocks and thus, underestimate the survival improvement needed to avoid jeopardy. We also believe that the analyses used in the draft Biological Opinion do not identify factors most likely responsible for the decline in salmon and steelhead and, therefore, prescribe management actions that may not provide the greatest survival improvement to listed stocks. Specifically, the draft Biological Opinion shifts responsibility of hydrosystem mitigation away from the mainstem and onto habitat restoration, hatcheries, and harvest reductions. The benefits expected to occur from the RPA's offsite mitigation and the aggressive hydrosystem operations are subjective and unsupported. We believe these benefits are overestimated, particularly for the Snake River spring/summer chinook where harvest is already extremely low, are located in good to pristine habitat and thus been assigned the lowest priority for habitat improvements, and have no hatcheries in 6 of the 7 indicator stocks. The aggressive hydrosystem improvements under the RPA provide only a slight increase of the flow targets defined in the 1995 Biological Opinion, which often have not been met in the last 5 years. In fact, the draft Biological Opinion actually decreases flow targets for Columbia River chum. In addition, we do not believe that the described methods to assess the success of the RPA can convincingly determine if the RPA has achieve its goals, over the 5-8 year interim period. Finally, we do not believe the draft Biological Opinion adequately anticipates and has prepared an alternative RPA that can be immediately implemented if at the end of this interim period the current RPA has failed.