Stevens et al.: Radiometric validation of age, growth, and longevity of Sebastes melanostomus 



719 







length (Love et al., 2002). Accord- 

 ing to an experienced rockfish age 

 and growth researcher, "some of 

 the oldest specimens [rockfish] are 

 rarely the largest (lengthwise), 

 and most, if not all, are males." 

 (Munk 2 ) The reasons for this age- 

 length pattern are beyond the 

 scope of this study, but the impli- 

 cations for stock dynamics and 

 management are that it is worthy 

 of further consideration. 



Age prediction, age-group 



determination, 



and core extraction 



The use of otolith weight as a proxy 

 for age has benefits over conven- 

 tional otolith aging methods by 

 reducing cost, increasing sample 

 size, and allowing greater objectiv- 

 ity (Boehlert, 1985; Pawson, 1990; 

 Fletcher, 1991; Pilling et al., 2003). 

 In this study, predicting ages from 

 otolith weight increased the number 

 of unsectioned otoliths that could 

 be used in the radiometric analysis, 

 but the prediction model also ampli- 

 fied the uncertainty associated 

 with estimates of age. especially 

 in older fish. The variance around 

 the regression line increased with 



otolith weight, and log normalizing the data did not 

 eliminate this problem. Older predicted ages, there- 

 fore, were more uncertain than younger ages (Fig. 3). 

 Although limited to a specific otolith weight range, the 

 prediction model presented here may provide managers 

 with a more efficient and less costly way to investigate 

 the age structure of blackgill rockfish stocks. 



In an ideal study, otoliths from the entire estimated 

 age range for blackgill rockfish would be available in 

 the sample set. Otoliths from fish with predicted ages 

 greater than 70 years, however, were not present in our 

 study in sufficient numbers to allow age determination 

 by radiometric methods. This was so, even though more 

 than half of the 1200 otolith pairs obtained for ourstudy 

 were sampled directly from commercial fishing vessels 

 in 1985 along the coast of central and southern Cali- 

 fornia, where the bulk of the fishery occurred. Because 

 fishermen often target adult aggregations, the absence 

 of these older individuals may be an indication that the 

 population had already experienced depletion of older 

 age classes at the time of sample collection, particularly 

 if natural mortality is thought to be low for most rock- 



1:1 line 



Regression 



y=0.823x + 2.534 



r =0.89 



20 



40 60 



Predicted age (yr) 



80 



100 



Figure 5 



Direct comparison of mean predicted age and radiometric age for 14 pooled 

 otolith age groups for blackgill rockfish (Sebastes melanostomus). A regres- 

 sion of the data points and 1:1 line of agreement are included for comparison. 

 Horizontal error bars represent the error associated with age estimation 

 (average predicted age multiplied by 14.6% CV), plus the standard error of 

 the regression (Is) used to predict age for radiometric samples. Vertical error 

 bars represent high and low radiometric age estimates based on the analytical 

 uncertainty associated with 210 Pb and 22fi Ra measurements. 



fishes (Bloeser 3 ). However, it is possible that the largest, 

 oldest fish are naturally rare, even at the start of an 

 intensive commercial fishery. Knowledge of blackgill 

 rockfish pre-exploitation stock structure and population 

 dynamics would help to elucidate which (depletion of 

 older age classes or a natural situation of low numbers 

 of older fish) is the more likely scenario. 



Radiometric analysis 



In previous studies the analytical uncertainty of 226 Ra 

 was the limiting factor in radiometric age determina- 

 tion (Andrews et al., 1999a). Typically, TIMS determi- 

 nation of 226 Ra reduces error to less than 1-3% of the 

 determined value, but technical difficulties (improperly 

 mixed nitric acid) led to poor recovery and loss of radium 

 in seven samples. The remaining seven samples were 

 deemed reliable because of relatively high radium recov- 

 ery, longer run times, and low analytical uncertainty 

 as determined by the TIMS analysis routine. The 226 Ra 

 activity determined for these samples was consistent 



2 Munk. K. 2002. Personal commun. Alaska Department 

 of Fish and Game, P.O. 25526, Juneau, AK 99802. 



3 Bloeser, J. A. 1999. Diminishing returns: the status of 

 West Coast rockfish, 94 p. Pacific Marine Conservation 

 Council, P.O. Box 59, Astoria, OR 97103. 



