The SLCM includes the option of integrating hatchery-released fry or 

 subyearlings into the simulated natural populations. In the base model, 

 the number of fry released by the hatchery (FRYREL) is first adjusted for 

 post release mortality using the survival parameter, hfry_sv. The surviv- 

 ing fry interact with the natural fry according to the parameter, fryint. 

 This parameter expresses interaction on an arbitrary scale that ranges 

 from to 100. If fryint = 0, then hatchery fry have no effect on survival of 

 natural fry; if fryint = 100, hatchery fry affect the survival of natural fry to 

 the same extent as additional natural fry. Expected survival of naturally 

 spawned fish and hatchery-released fish is calculated in a multistep pro- 

 cess outlined in appendix A. Admittedly, this process is arbitrary and 

 rather convoluted. The method has two desired properties: (1) survival of 

 naturally spawned fish decreases when hatchery-released fish are present 

 and (2) the impact of hatchery releases on naturally produced fish is di- 

 rectly proportional to the degree of competition. 



The base SLCM accommodates up to four year-classes of migrants. Sub- 

 yearling migrants (SMOLT0) exit the tributary during the same year they 

 are spawned; yearling and later migrants (SMOLT1 - SMOLT3) exit in 

 succeeding years. Each additional year that the presmolts remain in the 

 tributary they are subject to a mortality factor reflected in inbsmsv. 



Alternative Model 

 of Juvenile 

 Production 

 (Version 2) 



In many instances, we do not have the stock-recruitment information 

 needed to unambiguously choose one form of the juvenile production func- 

 tion over other alternatives. Available data may be scarce, providing no 

 hint of the underlying relationship, or the data may be so noisy that no 

 single function provides a superior fit. Where few or no data are available, 

 little can be done to inspire confidence in the quantitative output of any 

 model. Information from similar stocks might help, but in the end the 

 chosen relationship must be viewed as a hypothesis entailing considerable 

 uncertainty. 



In instances where data are available but the underlying relationship is 

 obscured, an empirical approach based on conditional probabilities may be 

 used. This offers a way to capture the variation in recruitment, and to 

 some extent, the density-dependent relationships that might exist. Data 

 on spawning stock size and the estimated number of smolts produced from 

 each spawning are compiled. Lognormal distributions are independently 

 fit to the spawning and smolt production data; these are known as the 

 marginal distributions. Smolts produced is plotted against spawning es- 

 capement in a two-dimensional graph. Reference lines are drawn to sepa- 

 rate the marginal distributions into thirds, thus dividing the plot into nine 

 cells (fig. 4). Dividing the number of points in a given cell by the sum of 

 the three cells in its column provides an estimate of the conditional prob- 

 ability that the number of smolts produced in a given year will fall within 

 that region of the graph, given that the corresponding spawning escape- 

 ment is within the indicated range. These probabilities, along with the pa- 

 rameters defining the marginal distributions, must be specified in the 

 SLCM parameter file if this option is used. 



The SLCM uses these probabilities in a three-step process. First, it 

 takes the number of spawners and determines which third of the spawning 

 distribution is represented. Then, using the conditional probabilities, it 

 randomly selects from a multinomial distribution the region of the smolt 

 distribution that will be used to generate the number of smolts produced. 



10 



