356 



For the Snake River evaluations, the C8FWA study focused on the spring/summer 

 Chinook transportation evaluations for smelt releases during 1986 and 1989. with 

 subsequent adult returns. These studies indicate that the average ratio of transport to 

 non-transport salmon was between 1 .6 to Z5. Or stated differently, the NMFS estimates 

 that transportation of juvenile salmon provides between 60% to 150% more adult returns, 

 back to Lower Granite Dam, than fish that remain in the river and pass through the 

 Snake-Columbia river hydro projects and reservoirs. Evaluations for other stocks 

 (steelhead), over the years of research, display similar positive transportation benefits. 



The C8FWA Review group analyzed very small data samples for "tagged" Chinook salmon 

 in upriver areas above Lower Granite Dam, with most return fish collected at hatchery 

 sites. Their findings suggest that hatchery fish did respond positively to transportation, 

 but limited samples for wild fish suggested no direct transportation benefits. The key 

 problem surrounding the CBFWA Review Group analysis stems from sample size and 

 sampling methodology-both of which are unacceptable for a basis to make valid 

 sdentific conclusions. To their credit even the Review Group noted: "NMFS contends 

 that tinis information [above Lower Granite data samples] is not reliable and should not 

 be used for analysis because sampling was not complete and may be biased." 



Specific Technical Comments: 



Clarifying Mid-Columbia Versus Snake River Evaluations: 



The reviewers-or other parties-use of tiie Priest Rapids fransportation studies to draw 

 any conclusions regarding the Snake River transportation program is totally inappropriate. 

 The Priest Rapids studies, testing experimental procedures, used different transportation 

 techniques than tine Snake River evaluations of an operational system. 



Inappropriate Marked Release Group Analysis: 



In part of the CBFWA review, the Review Group attempts to analyze the transportation 

 data from an approach that is inconsistent with tiie evaluation stijdies' intended 

 experimental design. It is highly questionable to analyze individual marked subsets within 

 the treatment and corrtrd groups, because recoveries from these indivkjual groups 

 provide sample sizes too small for reliable analysis. The individual, marked subgroups 

 should not be considered as separate ti-eatment/control groups but rather as lots within 

 the complete ti^atment and control groups-accepting that subgroup variation does exist 



It is recognized that varying conditions do occur across each smolt release group . But 

 given the very small sutsgroup sample sizes, possessing unknown variation levels across 

 multiple variables, aggregated analysis offers the most acceptable level for interpretation. 

 Until specific variables can be independentiy measured and larger sample sizes obtained, 

 it would be imprudent to reject the aggregated statistical parameters. 



