natural fry can be incorporated. Alternatively, Version 2 describes the 

 steps between spawning and smolt production less explicitly, relying on 

 historical information on adult production to generate smolt numbers. 

 Each version is described in more detail below. 



The following description applies to the base model. Specific changes in 

 model structure have been made in various versions to accommodate spe- 

 cial circumstances or life history traits peculiar to individual stocks, or to 

 simulate specific hatchery scenarios. Such changes are easily made. 



The number of female spawners (FSPAWN) determines the number of 

 naturally produced eggs. The number of female spawners is assumed to be 

 binomially distributed with N = the number of spawning adults, and p = 

 ffem, the probability of being female. The number of eggs is then simulated 

 as a random normal deviate with mean = FSPAWN • eg fern and standard 

 deviation = FSPAWN *stdegg, rounded to the nearest positive integer. 

 Since the number of eggs produced is the sum of reasonably large numbers 

 of independent random variates (individual spawnings), the normal ap- 

 proximation seems justified. 



The SLCM's only density-dependent mechanisms are in the egg-to- 

 presmolt transition. Density-dependent mechanisms operating within the 

 first year or two of life are assumed to play an important role in the dy- 

 namics of salmonid populations. The expected probability of survival from 

 the egg stage to just before leaving the subbasin (PRESMOLT) is calcu- 

 lated as a decreasing function of initial egg numbers and hatchery fry re- 

 leases (details below). 



Numerous studies of the relationship between population density and 

 survival indicate that the relationships — when detectable at all — are 

 rather imprecise (Peterman 1987; Rothschild 1986). In addition, factors 

 other than population density can contribute to large variations in early 

 freshwater survival (Alderdice and others 1977; Kocik and Taylor 1987; 

 Randall and others 1987). For this reason, the SLCM uses a binomial-beta 

 distribution to generate the number of presmolts produced each year. The 

 binomial-beta distribution increases the variance in the number of pre- 

 smolts produced from a given number of eggs beyond that introduced by 

 the binomial process alone. The coefficient of variation in egg-to-presmolt 

 survival (cvegsv) is specified by the user as part of the input parameters. 

 The model uses the expected probability of survival calculated from the 

 density-dependent relationship and cvegsv to calculate the parameters of 

 the beta distribution used in this transition. 



The user can choose among three options for the density-dependent rela- 

 tionship used to calculate the expected egg-to-presmolt survival. The 

 choices include the Beverton-Holt relationship (Beverton and Holt 1957), 



E(p) = 



alpha 



(4) 



the Ricker relationship (Ricker 1954), 

 E(p) = alpha[e^ etaEGGS) ] 



(5) 



8 



