BEARDSLEY and CONSER: AN ANALYSIS OF CATCH AND EFFORT DATA 



o 



ci 



t^ in 



CD "^ 



1^ 



<h m 



•^ CO 



CM O 1- O 



00 I I I o o o 



I I I •.- CO ■- 



O CD 



I 9 "^ 



a> o -^ 



CO O CD 



d 



I I I I 



CD 



I '°- 

 I O 



I i2 



CD 



I " 



1 1 1 1 ; I ; 1 1 V I 



I o 



• cnmocD-^'<tc\jc>ir^cDco 



;_C\lcDOr^^'-CDC\l>->-CDO> 



ococNjooT^cbcbT-d'-dcbci-r^P 



II II II 11 II II II II II II II 



eg ^ ^ M 



II 



to 9 f^ tn f^ 



II II 



into — c>j(0<OCDLi CDuin ^-cjcm 

 CN* — — <Ncg — (M**- ^r*!*"*- ^ — — 



o) CM in CO ^ ; 

 '- in CO (^ o 



Si 



n-;in-^i^°^-2 

 in^csji-O'-cO'-ocMCO'-incj't- 



N II II II II II II II II II II II II II II 



U.U.U.LLlj,U.U.U. 



U. U. U. 



oO'-i^cjcDinco^H^cDCNi^Trin 



inpocoh-.incDinor^r-^cvjTr<3)tD 



^ oddddddd) l d d d d d d 



II II II II II II II II II II II II II II II 



•-a)tn'-T-aDtn>-'~<-coina><-cu 



fvJCM — ^C«ICNJ'-f->'-(M<M»-(M — t- 



E E E •= E = E E = 



0) g 



If 



- o 



CD 



ggraggSracoS 



01 



a> 



CD 



E 



CO 



f 9 



1= t= S c j^ c 



CO 



o 



CD 



0) 



CJcOOOcOcCJC-'cCJcC-'^ 



OoooooOo°°oOoOo 

 D-jQQQJjQJjQQ-jQJjQ^ 



CO 



CO 



5 



o 



3 iS 



^ 



c 



CO 



c 

 o 



t3 



CO 



to 



0) 



€ £ 



■o 



S 



o 



O 



z 

 < 



0) 



.£e| 



52§ 



<U I 



ID 



z g Q. 



HI 

 111 



3 =iO 



o JS^ 2 

 o *3 CO 



Q^ .= CO ^ 



: " " !8 

 : to a, CB 



- oi 55 o 

 P c Sp 



• . S -D § 



O (D tOr^ 



a 



the New Orleans area. The interaction implies 

 that a better model exists for representing the 

 data, probably one with a cross product term, but 

 since the data were well correlated, we believe a 

 reasonable power estimate could be obtained 

 using the simpler basic catch model. The degree of 

 fit of the regression model used to estimate power 

 is good in both areas, accounting for nearly 90% of 

 the variation in the data, and despite the interac- 

 tion found in the New Orleans area the power 

 estimates appear to be reasonable. So few white 

 marlin are caught in the Texas area that sample 

 size was not adequate to make similar determina- 

 tions. The correlation and ANOVA analyses pro- 

 duced nonsignificant results. 



The results generally indicate that the basic 

 catch model adequately represents the white mar- 

 lin data in the northern Gulf of Mexico. It appears 

 that any one of the data types (dock, tournament, 

 and longline) can be used to index relative abun- 

 dance for local areas but more reliable indices will 

 result if all three are aggregated as was done here 

 and displayed in Figure 8. 



Sailfish 



No significant differences were found in the fish- 

 ing power of dock and tournament data and no 

 interactions were apparent in any of the three 

 areas. CPUE's were significantly correlated in the 

 Panhandle and Texas areas but not in the New 

 Orleans area. Relatively few sailfish are caught in 

 the New Orleans area, especially in tournament 

 fishing during the early years, and sample size was 

 inadequate for comparison of dock and tourna- 

 ment data. Sample size was better from dock data, 

 so dock and longline data were compared in the 

 New Orleans area while dock and tournament 

 data were pooled and compared with longline data 

 in the Panhandle and Texas areas. In all three 

 areas, the CPUE's were significantly correlated, 

 significant differences in power were found be- 

 tween recreational and longline fishing, and no 

 interactions were apparent. The regression model 

 used for estimating power is highly significant, 

 and the degree of fit is good in all three areas, but 

 the recreational fishing power estimate may be 

 somewhat inflated in the Panhandle area because 



61 



