q: 10 



79 80 81 



YEAR-CU\SS 



Fig. 16. Nianlic River winter flounder rccruilment indices Tor the 1976-84 year-classes predicted by Ihe 

 three-parameter mode! (line) compared to values actually observed (circles). 



influences on recruitment appear to be phenomena 

 fairly pervasive and occurring over large areas and 

 for a number of different species (Gushing 1973; 

 Sissenwine 1984). The use of the temperature 

 parameter (p in the model as well as the relatively 

 large estimate of the a parameter implies that 

 environmental and other density-independent pro- 

 cesses are important factors in winter flounder 

 reproduction as noted by Roff (1981) for several 

 flatfishes. However, the density-dependent pa- 

 rameter (P) of the model must operate to adjust 

 year-class strength, as the relatively small varia- 

 tions in recruitment suggest some stabilization 

 mechanism. 



Given the relatively cold winters that occurred 

 during the late 1970s (Fig. 1 7) along with moderate 

 parental stock abundance, good to exceptional 

 year-classes resulted in large parent stock sizes 

 from 1981 through 1983. TTie relatively abundant 

 adults coupled with mostly above-average tem- 

 peratures through the present have resulted in the 

 below-average recruitment and decreased winter 



flounder abundance presently seen. Recent warm 

 winters suggest that a large increase in winter 

 flounder will not occur in the near future. How- 

 ever, when parental stock size and prevailing water 

 temperatures for 1985-87 were used with the stock 

 and recruitment relationship, moderately im- 

 proved recruitment seems likely during 1988-90 

 as shown by the three predicted points (Fig. 15). 

 Future values of age 3 CPUE may be compared 

 to those predicted by the three-parameter model 

 to examine its credibility. This would also address 

 a criticism of Sissenwine (1984) in that empirical 

 models often fail to predict post-publication 

 events. However, as the number of data points 

 used for the model was relatively small, the ad- 

 dition of others in forthcoming years will likely 

 change parameter estimates to an unknown degree 

 and model reliability will likely improve. Finally, 

 increasing knowledge of the reproductive process 

 and early life history of the winter flounder may 

 enable the formulation and testing of plausible 

 hypotheses concerning stock and recruitment 

 mechanisms and density-dependent mortality. 



176 



