FISHERY BULLETIN: VOL. 75, NO. 3 



30 



§ 25 



E 



</) 

 T> 



c 

 o 

 l/> 



3 

 O 



CO 



Q 



20 



15 



O OBSERVED EFFORT 

 • C, 5 = 1870 



1945 



1950 



1955 



I960 



1965 



FIGURE 12. — Simulated landings of yellowtail flounder with ob- 

 served effort and with effort by Equation (20) using c 15 = 

 1,870. 



6.0 



O OBSERVED LANDINGS PER DAY 



• C, 5 = 1540 



A C, 5  1870 



-A C15 = 2200 



O 



1 



a 



UJ 



a. 



w 



z 



o 



a: 



H 



UJ 



£ 

 v> 



13 



Z 



o 



z 



4.0 



_L. 



1945 



1950 



1955 



1960 



1965 



FIGURE 13. — Observed catch of yellowtail flounder per day of 

 fishing and simulated catch per day with effort set by Equation 

 (20) using c 15 = 1,540, 1,870, or 2,200. 



At present, annual catch quotas for the South- 

 ern New England yellowtail flounder stock are 

 based on a prerecruit index (Brown and Henne- 



480 



muth 1971). The index is calculated from the catch 

 of 1-yr-old fish in an autumn bottom trawl survey 

 (Grosslein 1969). Thus the major source of vari- 

 ability in production resulting from the influence 

 of temperature on recruitment is accounted for in 

 current stock assessments. This model should not 

 be considered as an alternate method of manage- 

 ment of the fishery on a year to year basis without 

 further verification and refinement. 



Walters (1969) developed a yield optimization 

 procedure for his generalized fish simulator using 

 the steepest ascent method. The development of 

 an optimization procedure for the model reported 

 in this paper would be more difficult because this 

 model is driven by two exogenous factors, temper- 

 ature and the rate of fishing, while Walters's 

 model is only driven by fishing mortality. This 

 model is generally more complex than Walters's 

 model and much more expensive to run. There- 

 fore, the development of an optimization proce- 

 dure is beyond the scope of the present work. 



DISCUSSION 



A complex compartmentalized simulation 

 model of the Southern New England yellowtail 

 flounder fishery has been described, verified 

 against catch statistics, and used to examine 

 methods of increasing yield. The relationships 

 and parameters upon which the model was based 

 do not appear to be unreasonable since most vari- 

 ability was explained. Nevertheless, in retrospect, 

 some consideration of alternate approaches to 

 estimating parameters and of modifications of the 

 model is appropriate. It is important to remember 

 that there may be numerous other models or 

 parameter values equally as successful at explain- 

 ing variation in catch as the one proposed here. 



An average maximum length (L m4 ) for the sim- 

 ulated population of 480 mm was assumed. This 

 value was used in order to assure that few fish 

 would exceed 500 mm in length. When fishing 

 pressure was applied to the simulated population, 

 its average maximum length was suppressed. For 

 some years, the average length of the older age- 

 groups converged to about 460 mm. Since the 

 growth rate coefficients (k t ) of adult fish were 

 based on Lux and Nichy's (1969) work where a 

 maximum length of 500 mm was assumed, the 

 model tends to underestimate the length of older 

 fish. In order to compensate for this effect, the 

 growth rate coefficient of fish younger than 2 yr 

 of age was overestimated. The result was that 



