26 



of almost seven fish survived to die from fisheries or dams; or, escaped 

 to spawn). This implies that some of the non-natural spawners were at 

 least partially successful in producing progeny. The return per spawner 

 for the 197S-1989 broods, when all spawners over Lower Granite Dam are 

 included, was estimated to average 5.0 (Table 13). 



The estimated returns plotted against natural spawners shows a wide fan- 

 shaped pattern (figure 12). There is little evidence of density 

 dependence as the largest escapement (1,000 spawners) produced an 

 estimated three recruits per spawner. The estimated returns plotted 

 against all Lower Granite Dam spawners is similarly shaped. In both 

 spawner-recruit relationships, the estimated curve is very flat and 

 similar levels of returns are predicted to result from various spawning 

 levels. Horishima (1994) indicates this is typical (from simulation 

 studies) of extreme downstream mortality. It is possible that it is a 

 product of errors in age composition and catches. There are no data 

 points below the replacement line in either of the two spawner-recruit 

 relationships. If recent escapements were on the right hand side of the 

 spawner recruit relationship where density dependence was operating, 

 data points below replacement would be expected, particularly given the 

 level of downstream mortality that occurs. 



If the data are accurate, the estimated number of parents needed to 

 produce a maximum sustained yield return is 440 natural adults or 472 

 total spawners counted over Lower Granite Dam. Another way to determine 

 an appropriate escapement goal (other than an HSY escapement) would be 

 to choose an escapement level that maximires coturns. The number of 

 spawners predicted to produce maximum returns are 516 naturals or 570 

 total spawners over Lower Granite Dam. Recent year escapements are on 

 the order of these estimates. The shape of the spawner-recruit curves 

 indicates that at spawner densities observed, there does not appear to 

 be a strong density dependent spawning or rearing limitation for Sna)ce 

 River fall chinoo)c salmon. Further, recruitment is not below 

 replacement even with the high levels of migration corridor mortality. 

 Based on these analysis it seems lilcely that escapements on the order of 

 double the indicated HSY levels (around 1,000 adults counted over Lower 

 Granite Dam) would produce strong returns. 



In an attempt to determine if the estimated returns were correlated with 

 marine survival and river flows, a multiple regression with three 

 independent variables (escapement, marine survival, and river flow) was 

 run against the returns using both the natural escapements and the total 

 adult escapement past Lower Granite 0am. The marine survival index was 

 taken from CTC (1994) and was for the Salmon River Chinook salmon stoclt, 

 a north Oregon coast stocJc that has a similar ocean distribution pattern 

 to Upper Columbia River brites and Lyons Ferry Hatchery fish. River 

 flows were for the Snake and Columbia rivers during July and August. 

 The multiple regressions improved the modeled fit of the spawner-recruit 

 relationships somewhat, indicating that even with confounding errors, 

 marine survival (surrogate - Salmon River Chinook marine survivals) and 

 river flows were affecting recruitment. 



Residuals for the second spawner-recruit relationship (all adults 

 counted over Lower Granite Dam) were plotted (Figure 13). The residuals 

 are ordered and correlate somewhat with the marine survival estimates 

 for Salmon River chinook ( R^ = 0.25). The residuals are positive at the 

 beginning of the time series and negative at the end of the time series 

 (positive meaning the returns were higher than predicted from the 

 spawner-recruit curve). 



Because of time limits due to the deadline for responding to the 

 December 28, 1995, Federal Register, we could not expend the level of 

 research this topic deserves; particularly with regard to residuals 

 analysis and comparing other variables to these residuals. We recommend 

 that the topic be further investigated as it likely will bear fruit 

 useful to ESA managetnent related activities for the Snake River fall 

 Chinook population. 



