52 



Fishery Bulletin 105(1) 



60° 



100' 



L. 



90° 



80° 



30° 



20° 



10° - 



30° 



A— ^ 5000 t 



I I Floating-object sets 



^^1 Dolptiln-associated sets 

 Unassociated sets 



X <=5t 



O 







O o 



Q 9 







x^ 



r i\ 



• • • (3 '' ^ 



G G GQGGGQ©© ® ''^' 



Q G G G G G G 

 OOGGGGGQ Q( ^'0 



o o O O G O G G G G G 

 oooooooooG C y v^ 



eastern Pacific Ocean 





5"- 



•30° 



■20° 



■10° 



-0° 



•30° 



140° 



130° 



120° 110° 100° 90" 80° 



Figure 2 



Average annual distribution of the purse-seine catches of skipjack tuna (Katsuwonus pelamis), by set type, in the 

 eastern Pacific Ocean, 1995-2002. 



skipjack tuna catches was high, rather than just areas 

 of high bigeye tuna catches, because we wanted areas 

 where the losses in skipjack tuna catches would be mini- 

 mized. We chose 5°x5° areas by quarter of the year as 

 the scale for the individual hotspots. 



We defined an index for each time-area strata for 

 each year. The index was a ratio of bigeye tuna catch 

 to skipjack tuna catch that was robust to annual fluc- 

 tuations in the abundance of either species. We then 

 summed the annual indices over the time period to 

 find areas that consistently resulted in high bigeye to 

 skipjack tuna ratios. The indices were calculated sepa- 

 rately for each year so that they were not dominated by 

 data from years with exceptionally high or low catches 

 of either species. 



The data used for this and the closed-area analysis 

 were the following: 



B 





bigeye catch in quarter i in area/ in year t; 

 skipjack catch in quarter /' in area^ in year t. 



We standardized catches within a year on the basis 

 of the median catch of the year. 



s,^^ = S,^,/median(S ,). 



The location of hotspots did not differ noticeably if we 

 standardized by the mean or total catch for each year, 

 rather than the median. 



Using the standardized catches, we defined the an- 

 nual index for a single 5°x5°area by quarter, 0,^, as 



e.„,. 



where 



1 where 



f(b, j,,s, J,) otherwise 



b,,,=0 



».,.,.= 



(1) 



b, , , / s, , , - min(6, , , / s, , , ) 



where f(b, , ,s, ,. ,) = '-'■' '/■' .'■■'■' '-'/ (2) 



■•'•' ''J-' max{b,j,/s,j,}-mm(b,j,/s,j,) 



