686 



Fishery Bulletin 98(4) 



the fourteen-year time series (Fig. lA, Table 1). Variabil- 

 ity decreased after 1986, when the fall SEAMAP ichthyo- 

 plankton survey was added which both extended coverage 

 into the time period of peak king mackerel spawning and 

 increased the number of samples. Coefficients of varia- 

 tion (100 X SE/mean) have been less than 20% for both 

 frequency of larval occurrence (1989-95) and larval abun- 

 dance (1992-95) for the most recent years of the time 

 series (Table 1). Also an expansion in the areal distribu- 

 tion of king mackerel lai-vae has been apparent since 1986 

 (Fig. 2). 



A total of 798 king mackerel lai-vae ranging in length 

 from 1.6 to 10.1 mm were collected from 1982 to 1995 

 (Table 2). A quadratic equation best described the rela- 

 tionship between larval king mackerel age and length 

 data (Fig. 3). Catch-at-age was calculated from the esti- 

 mates of larval density and age frequencies and then 

 used to construct annual catch-at-age curves for survey- 

 captured king mackerel larvae. Over the time series, 

 lai-vae ranged in age from 2 to 11 days, and estimates of 

 instantaneous daily mortality rates for individual years 

 ranged from 0.35 "in 1985 to 0.70 in 1992 (Table 3). A 

 single catch-at-age regi'ession for all years combined indi- 

 cated no significant difference among slopes (i.e. instanta- 

 neous mortality rates) or among intercepts. Therefore, a 

 pooled regression model was fitted. The slope of the pooled 

 regression gave an estimate of instantaneous daily mor- 

 tality rate (Z) of 0.53 (Table 3) and was subsequently used 

 to backcalculate the abundance of one-day-old larvae and 

 generate the age-adjusted index of king mackerel larval 

 abundance. 



VPA estimates of king mackerel spawning stock size 

 ranged from 46.03 x 10'^ individuals in 1985 to 101.93 x 



10' individuals in 1995 (Fig. lA). Survey larval frequency 

 of occurrence and estimates of stock size corresponded with 

 a correlation of 0.82 over the entire time series, 1982-95, 

 and 0.92 for the period 1986-95 (Fig. lA). The king mack- 

 erel survey index of lai-val abundance (unadjusted for age) 

 was also highly correlated with spawning stock size (Fig. 

 lA). The correlation, 0.84, was the same for both periods 

 of comparison, all survey years, and years since 1986. Our 

 attempt to adjust larval abundance for age did not improve 

 the correlation between the abundance index and spawning 

 stock size. The correlation with spawning stock size was, 

 however, higher after fall sui^veys began, 0.78 versus 0.65, 

 but these values were both lower than those for the "uncor- 

 rected" abundance index and the index based on frequency 

 of occurrence (Fig. IB). 



Discussion 



Hunter and Lo (1993) asserted that fish eggs and lai-vae 

 can be used not only to estimate the biomass of a fish stock 

 but also to monitor trends in relative stock abundance. Indi- 

 ces of relative abundance are less costly to produce than 

 biomass estimates from ichthyoplankton data, but they are 

 also less precise. The CVs of the most precise biomass esti- 

 mates based on ichthyoplankton data range between 20% 

 and 30%. Whether adjusted for growth and mortality of 

 larvae or not, ichthyoplankton indices are "surprisingly sen- 

 sitive to major changes in stock abundance" (Hunter and 

 Lo, 1993). It should be of no surprise that our indices based 

 on abundance and frequency of occurrence of king mack- 

 erel larvae from SEAMAP collections in the Gulf of Mexico 

 closely tracked trends in adult abundance over the time 



