430 



Fishery Bulletin 97(3), 1999 



E 

 3 

 S 



CO 



S 



Z 



100 n 

 80 

 60 

 40-1 

 20 

 



U.S. baitboat 



- predicted 

 Observed 



f* Mf >ftt»ft>tf ■■■!■■ 



Longline 







12 16 20 24 28 32 36 40 



- predicted 

 observed 



16 20 24 28 32 36 40 



Japan baitboat 



-predicted ^^ 

 observed ^g 



• y*»ifttifiii f nf ■■>■■ I 



12 16 20 24 28 32 36 40 



Other 



-predicted 

 observed 



^•^■^••• f Bf ■■■!  f ta f  I 

 16 20 24 28 32 36 40 



Troll 



- predicted 

 observed 



 ■yiiif"iif ■■■! mf a f i 



12 16 20 24 28 32 36 40 



All fleets 



- predicted 

 observed 



12 16 20 24 28 32 36 40 



Time at liberty (quarters) 



predicted 



« observed 



mMmmmmmmmA 

 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 



Year 



Figure 7 



Observed tag return numbers and tag returns predicted with model 8 ( see Table 4 for description ). 



Z^fZ^kti'^ti ' 



kf 



Z^/ Z^k^i'f'^v ' 



kf 



+ M 



(6) 



l-expl-X/-! 



,k^^%T„-^kf 



■M 



X 100%, 



where /? is a quarter occurring in year j, are shown 

 in Figure 9 (model 8, /3=0.9). Estimated exploitation 

 rate has been <10% for most of the time series; it 

 declined between 1976 and 1988 and increased 

 slightly afterwards. 



Effect of assumed tag reporting rate 



Although the tag-reporting rate is thought to have 

 been high for North Pacific albacore (Kleiber and 

 Baker, 1987), this contention is not supported by 

 independent data. We therefore examined the sensi- 

 tivity of the results to different values of j3 (assumed to 

 apply equally to each fleet) for the model 8 fit. The es- 

 timates of M were directly related to p; conversely, av- 

 erage exploitation rate was inversely related to /J (Fig. 

 10). In both cases, the sensitivity was slight for /3 > 0.6. 

 If the tag-reporting rate was above this level for the 

 main fleets, the results of our analysis should have been 

 robust to small departures from the assumed value. 



