109 



tem, despite the fact that apphcable technology has been available for at least 20 

 years, should be seen as a stunning indictment of the institutional structures and 



golicy frameworks under which fisheries research is conducted in the Columbia 

 iver basin. The ability to separate survival during hydroelectric passage from sur- 

 vival in the estuary and ocean is fundamental to understanding Uie relative benefits 

 and cost effectiveness of various mitigation measures. Since the historical data base 

 contains so few opportunities to separate the effects of the hydroelectric system fi-om 

 those of the ex-hydroelectric environments, there can be no resolution of competing 

 scientific hypotheses about the benefits and risks of the role of actions such as spm 

 and transportation in salmon recovery. 



For example, the Oak Ridge National Laboratory study by Bamthouse et al., iden- 

 tified fundamental differences in the assumptions governing the Bonneville Power 

 Administration's model of hydroelectric effects (Anderson, University of Washington) 

 and the state, Federal and tribal version of the same model. The Bamthouse report 

 concluded that the differences in these models, which are extensively used to fore- 

 cast the benefits of salmon recovery actions, could only be resolved by collecting ob- 

 servations on the survivals of the fish through tiie hydroelectric system, among 

 other data. 



Hence a good deal of the present salmon recovery controversy is due to the fact 

 that all parties rely on models which contain critical quantities which have never 

 been estimated. These critical quantities are best professional judgments of the sci- 

 entists involved, so the model results are no more, and no less, uian the opinions 

 of the scientists who create and operate the models. There is a widespread mis- 

 conception that "running the model ' with the latest data can add something to the 

 decision making process which is superior to that which can be obtained by simply 

 polling knowledgeable scientists for their views. If the Columbia basin models actu- 

 ally had critical quantities estimated by observation, these model runs would pro- 

 vide insight superior to informed opinion, but they do not, so they cannot. 



It is therefore important to recognize that the decision making process for spill 

 is guided bv qualitative assessments of individual scientists, hence the comparison 

 of benefits oased on the percentage of juveniles supposed by the models to transit 

 the hydroelectric system alive is not a statistically or mathematically valid proce- 

 dure. The figures cited by the Snake niver Recovery Team in its letter of May 30, 

 1995 to Mr. Will Stelle, Northwest Regional Director of NMFS, fall into this cat- 

 egory. The Team notes in its letter that it does not know how to tell if the percent- 

 ages calculated for the spill and no-spill scenarios are statistically significantly dif- 

 ferent fixtm one another, but paradoxically, the Team goes on to conclude that the 

 differences in benefits are, in lact, significant, and the Team rejects spill as a recov- 

 ery option based on these differences. The logically inconsistent position taken by 

 the Team in advising the decision making process is a direct consequence of the log- 

 ical inconsistencies in ita modeling approacn. This modeling approach is an unhappy 

 mixture of observation and bsdd speculation which is impossible to remedy without 

 observations on the survivals of juvenile salmon in the hydroelectric system. 



My contention is that Ae absence of data to estimate quantities so crucial to the 

 validation of tiie modeling process that advises the management decision making 

 process would never have been permitted by an independent scientific peer review 

 process which had influence over the funding of projects. In my experience as an 

 independent peer reviewer of proposed fisheries research for the Exxon Valdez Oil 

 Spill Trustee Council, Anchorage, Alaska, the actions of the Federal and state agen- 

 cies which conduct the research can be modified by independent peer review to con- 

 form to a central theme which produces logically consistent sets of observations for 

 evaluation of specific effects. My experience as a fisheries scientist in the Columbia 

 River basin is that the absence of independent scientific peer review allows Federal 

 and state agencies to overlook critical information in implementing research pro- 

 grams, and to confound the mission of salmon recovery with the missions of the in- 

 dividual agencies. 



Note that I am not presuming that a peer review process could, or would, usurp 

 legislatively mandated agency prerogatives. Rather it is my view that when agencies 

 taie research money in the name of the Endangered Species Act and the closely al- 

 lied Fish and Wildlife Program (NPPC), their proposed actions should be subject to 

 an independent scientific review process which certifies that the proposed research 

 actions are consistent with developing information critictd to the implementation of 

 the Fish and Wildlife Program. I am also counseling that research projects should 

 not be funded without this certification. In my experience with the Trustee Council, 

 the requirement for certification has not cost any agency its share of the oil spill 

 funding, but rather it has imposed standards on the types of research which are 

 conducted by the agencies with oil spill funding which focus the research efforts on 

 the attainment of established restoration objectives. 



