632 



Fishery Bulletin 91(4), 1993 



I TOTAL FLEET = 33 BOATS J 



Fleet Rep (1) 



% Coverage % Coverage 



FtePO) Rep (50) 



I K. = 50) I 



• A 



K K _. 



50 



? K 1 



50 



10 



£ K. 



50 



Figure 4 



Sampling procedure for simulation study. 



Within each replicate, cumulative annual mortality 

 for a given group of dolphins (YHAT) is estimated as 



YHAT, jklm = (K, lkl JDAYS, k J 



TFDAYS 



> 



where K, ]Um is the observed cumulative annual kill on 

 all selected trips, DAYS jklm is the observed number of 

 days, and TFDAYS is the total number of days spent 

 at sea by all vessels selected for that replicate. 



For each set of 50 replicate estimates of annual mor- 

 tality for a given fleet size, percent coverage, and dol- 

 phin group type, we calculated the average estimate 

 (YAVE ha ), the relative bias of that estimate (RBY S0 ), 

 and the coefficient of variation of that estimate (CVY S0 ) 

 as follows 



YAVE rMljtl = (1/50) *lYHAT llkln , 



RBY M „ lk , = 100 * (YAVE^u-known kUl„ k )lknown kill, jk 

 and 



CVY mkl = l^MSE(YAVE S0 , lkl )]/known kill llk \*100 

 where 



MSE(YAVE h0ljkl ) = H(YHAT llklm - known kill ijk f 



m = l 



and 



