244 



Fishery Bulletin 97(2), 1999 



overlap between billfish bycatch and target species 

 catch for the U.S. pelagic longline fishery and develops 

 a method for selecting candidate time— area cells whose 

 closure would selectively enhance billfish survival 

 and minimize impacts on the target species catch. 



culating the percent reduction in the total billfish 

 and marlin bycatch that would have occurred each 

 year if the time-area cells identified in the 1986-91 

 data had been closed. This analysis assumed that 

 the displaced effort would not shift elsewhere. 



Methods 



Results 



The longline catch data were inspected 1) to determine 

 the species composition of the catch to identify the set 

 of species which, taken together constitute the primary 

 targets of the fishery, and 2 ) to identify possible trends 

 in the reporting of billfish bycatch. The reported bycatch 

 of billfish declined markedly in 1992 compared with 

 earlier years. Consequently, the data were partitioned 

 into two periods: 1986-91 and 1992-95. Data fi-om the 

 first period were used to characterize the spatial and 

 temporal distributions of the catch. Data from the sec- 

 ond period were used to test temporal stability of the 

 results of analyses of the first period. 



The billfish bycatch and the combined billfish (sail- 

 fish, spearfish, and blue and white marlin) bycatch 

 as well as the target species catch for each longline 

 set of each reported trip were summed into cells of 

 1°, 2°, and 5° latitude and longitude and month over 

 the period 1986-91. The analysis was also repeated 

 and only blue and white marlin bycatch was consid- 

 ered along with the target catch. Separate cumula- 

 tive frequency distributions of the percentage of bill- 

 fish or marlin bycatch in the combined billfish and 

 target species catches by number by cell were con- 

 structed for each level of spatial resolution. These 

 frequency distributions were used to estimate the 

 percent reduction in billfish bycatch and target spe- 

 cies catch that would have occurred if the effort in 

 time-area cells with billfish or marlin bycatch rates 

 above an arbitrary threshold had been completely 

 eliminated during the 1986-91 fishery. The distri- 

 bution of the minimum number of cells producing 

 any specific reduction in billfish or marlin catches 

 with the least impact on target species catch could 

 then be determined from these data. 



As an example, the distribution of cells that, if they 

 had been eliminated from the 1986-91 U.S. fishery, 

 would have reduced the billfish and marlin bycatches 

 by SO'/f , were plotted for examination. Although the 

 reporting rate declined after 1991, the temporal and 

 spatial relative distributions of billfish in the total 

 longline catch would be unaffected, so long as the 

 change in the reporting rate was random. Conse- 

 quently, the temporal stability of the reduction in 

 billfish and marlin bycatch for the time-area cells 

 identified with the 1986-91 data was evaluated by 

 using data from 1992 to 1995. This was done by cal- 



The catch by species in the data file from U.S. pe- 

 lagic longline logbooks from 1986 through 1995 is 

 presented in Table 1. Inspection of these data indi- 

 cates that the predominant target species of this fish- 

 ery included swordfish, yellowfin tuna, bigeye tuna, 

 albacore, and dolphin fish. Together they constituted 

 90% of the nonbillfish harvest that was reported in 

 these logbooks. For this analysis, billfish bycatch is 

 the combined catch of sailfish, spearfish, and blue 

 and white marlin, and marlin bycatch is the com- 

 bined catch of blue and white marlin. Inspection of 

 the data in Table 1 indicates a sharp drop in the 

 numbers of billfish caught after 1991. This is possi- 

 bly a consequence of a change in reporting trends 

 following the U.S. billfish management plan's prohi- 

 bition of the sale of billfish, or the restrictions on 

 landing undersized swordfish imposed in mid-1991, 

 rather than a consequence of a true reduction in the 

 bycatches of billfish by the U.S. pelagic-longline fleet. 

 Although no obsei-ver data are available before 1992, 

 this conclusion is supported by data collected by ob- 

 servers on longline vessels that indicate billfish 

 bycatch rates after 1992 were much higher than those 

 reported in the logbooks ( Cramer M. For this reason, 

 the longline logbook data were grouped into two pe- 

 riods: 1986-91 and 1992-95. Data from 1986 through 



1991 were used to characterize the distribution of 

 billfish and target species catches, and data from 



1992 through 1995 were used to test the spatial and 

 temporal stability of the patterns observed in the 

 1986-91 data. 



The percentage of billfish in the combined billfish 

 bycatch and target species catch for each month-area 

 cell in strata of 1°, 2" , and 5° was estimated, and the 

 results were sorted to obtain cumulative frequency 

 distributions of the billfish bycatch percentages for 

 each spatial resolution. At the same time, the cumu- 

 lative catches of billfish and target species were com- 

 piled for each observation in the cumulative fre- 

 quency distributions of percentage of billfish in the 

 catch in each time-area cell. The results of this pro- 

 cedure are presented in Figure 1 and Table 2. Note 



Cramer. J. 1996. Pelagic longline billfi.sh bycatch. Inter- 

 national Commission for the Conservation of Atlantic Tunas 

 (ICCATi working document SCRS/96/97. 



