Overholtz et al.: Assessment advice for Northwest Atlantic mackerel stock 



123 



10 

 



F=0.05 



u 



-io 



LU 

 DC 

 LU -20- 



Q -30 

 -40 

 -50 



WTO 



m 



as 



^ 

 ^ 

 ^ 



i i i r 



A B C D E F G 



F=0.29 



A B C D E F G 



Figure 5 



Impact on 1987 catch, 1988 spawning-stock biomass (SSB), 

 and 1991 SSB of running the density-dependent simulation 

 (DDM) model with all possible combinations of growth, matur- 

 ity, and predation mortality. (A) maturity; (B) growth; (C) 

 predation mortality; (D) growth, maturity; (E) maturity, preda- 

 tion mortality; (F) growth, predation mortality; (G) maturity, 

 growth, predation mortality. 



A validation run of the density-dependent model was 

 produced for comparison with the observed time-series 

 of catch and VPA biomass for 1967-85 (Fig. 6). The 

 simulated series was produced by using the same fish- 

 ing mortality series as in the VPA, recruitment scaled 

 upward by a factor of 1.5, and density-dependent 

 growth, maturity, and natural mortality. The pattern 

 of simulated versus observed catch and biomass is quite 

 comparable in terms of trend and magnitude, except 

 for a few years in the early 1970s. This occurs even 

 though the density-dependent model has greatly dif- 

 ferent natural mortality rates at ages 1 and 2 and much 

 higher recruitment. Another run with the same F pat- 



tern, recruitment scaled by 1.5, and no density depen- 

 dence produced the same trend in catch and biomass 

 for 1967-85, but the values were approximately a fac- 

 tor of two larger than the observed series. Thus, not 

 including the density-dependent component resulted in 

 values of catch and SSB that were greatly different 

 than the observed series. 



The model components were validated by comparing 

 the different outputs produced by the various mechan- 

 isms with available empirical data. In some cases, rela- 

 tionships were re-parameterized or tuned to produce 

 results in the same ranges as observed in the empirical 

 database. The model was used to investigate a variety 

 of different problems. Runs from the STD model were 

 compared with the results of the different density- 

 dependent model outputs to gauge the changes that oc- 

 curred in catch, total stock, spawning stock, mean 

 weights, and other factors. STD runs were parameter- 

 ized with the same data as that used in the 1986 assess- 

 ment (Overholtz et al 1988). 



