202 



DRAFT 



(1) Analysis was not stratified by location and species 



The analysis did not stratify the evaluation of the impacts of the factors on salmonid 

 populations by region, or by species. The relative importance of a factor to the 

 productivity of anadromous salmonids will vary by region and watershed according to 

 the distribution and intensity of the respective factor. In addition, the impacts of a 

 factor will vary for different species due to the life history requirements and biological 

 characteristics of each species. Failure to incorporate these stratifications into the 

 analysis lead to incorrect conclusions concerning the relative importance of the factors. 



(2) The impact of factors were arbitrarily ranked 



The criteria for assigning a relative score (1, 2, or 3; Table 2.3-1) as a measure of the 

 population impact of each factor in the matrix was not documented, and appeared to be 

 arbitrary. The basis for assigning these values was not related to the discussions of the 

 various factors, or information presented in the report. Since the relative impact rating 

 for a factor was the sum of scores for all impacts assigned to the factor, the rating was 

 very dependent upon the scores that were assigned. 



(3) Some population impacts were redundant 



Some of the population impacts in the matrix were redundant and were multiple 

 measures of the same effect on salmonid populations. In the analysis, a factor was 

 often given several scores for the "biological papulation impacts" even though the 

 impact actually only represented a single effect on the population. This redundancy 

 over-weighted the impacts of certain environmental or management factors on salmonid 

 production. For example, salmonids harvested in ocean fisheries were given scores for 

 biological impacts titled: "direct mortality-young", "direct mortality-adults", "indirect 

 mortality", "contributes to mixed fishery", and "contributes to reduced spawning 

 escapement". These scores were then added to obtain a relative impact score. The 

 analysis failed to recognize that the five impacts are synonymous and represent a single 

 effect on salmonid prwiuction (reduced adult spawning escapement). Assigning a 

 "harvested fish" to each of these categories resulted in that fish being counted five 

 times, thus overestimating the actual impact of harvest. 



Because the matrix analysis method relied upon a direct summation of impact scores (1, 

 2, or 3) to evaluate the relative importance of the factors, the ranking of the factors was 

 directly dependent upon the number of matrix categories. The selection of these 

 "biological population impact" categories was not supported by any information 

 presented in the report. In addition, these categories did not accurately reflect the 

 complex biology of anadromous salmonids. Using redundant impact categories and a 

 summation method that is dependent upon the number of categories selected results in 

 conclusions that are not scientifically valid. 



(4) The measure of a factors impact was not based on changes in populatio n abundance 



The analysis method appears to be based on an assessment of the "increased mortality" 

 that a factor can exert on a single life history stage of salmonids, and fails to base the 



13 



