Gaertner et al.: Bycatch of billfishes by the European purse seme fishery 



685 



Longitude 



,20, , , ,1E, , , ,10, 



1 Metric ton 



Figure 2 



Spatial distribution of the bycatch (t) per FAD set for the 

 marlin group from the observers' trips during the Euro- 

 pean Union bigeye tuna program. 



year. That means that we take into account the probabil- 

 ity of setting on the fishing mode k, and the conditional 

 probability that group s is present in this type of set. For 

 each group s, the simulated input (i.e. the bycatch gener- 

 ated by one set) is drawn from the observed distribution 

 of the bycatch by set for the fishing mode k. Note that the 

 calculation of the total bycatch is based on the proportion 

 of total sets (including those made on trips without observ- 

 ers at sea) recorded in commercial logbooks in each strata 

 by fishing mode. That means that the observed bycatch in 

 each set was previously raised by the corresponding rais- 

 ing factor: i?f ,,j. A single set is then generated by random- 

 ly selecting inputs from the distributions of bycatch per 

 set. This is repeated 1272 times (the total number of sets 

 reported by observers aboard) to give the total bycatch of 

 group s during the standard year. 500 Monte Carlo simu- 

 lations were generated to determine the amount of varia- 

 tion expected in the yearly bycatch estimates. Confidence 

 limits for total bycatch by group were performed by the 

 percentile method. 



However, the results obtained in our study may have 

 been affected by the moratorium on fishing with FADs 

 adopted by the French and the Spanish tuna purse-seine 

 companies in the eastern Atlantic Ocean. This temporary 

 ban was adopted over a large area from the African coast 

 to 20°W and from 5°N to 4°S. between 1 November and 

 31 January m 1997-98 and in 1998-99, respectively. With 

 this consideration in mind, the Monte Carlo simulation 

 was performed a second time to simulate a standard year 

 without a moratorium. We assumed that the probability of 

 setting on FADs (estimated at 0.293 during a year with a 

 moratorium) increases to 0.55 during the years without a 

 moratorium (estimated from commercial logbooks). Conse- 

 quently, the second Monte Carlo simulation was conducted 

 with the following probabilities: 0.55 (for FAD sets), 0.419 



Table 2 



Estimated bycatches of billfishes (t) and observed com- 

 mercial tuna catches (t) taken by fishing modes by the 

 European purse seiners in the eastern Atlantic Ocean for 

 a standard year (October 1997-September 1998). School 

 sets were made without FADs. 



Fishing mode 



Marlins 



Sailfish 



Tuna catches 



FAD sets 

 School sets 

 Seamount sets 

 Total 



200.8 



42.8 



1.5 



24.5.1 



7.6 

 34.7 



0.1 

 42.4 



49,214 



88,456 



1,285 



138,955 



(school sets) and 0.031 (for seamount sets, assuming that 

 the percentage of sets made on seamounts was not af- 

 fected by a ban on fishing on logs. Table 1). 



Results 



Table 2 shows the estimated bycatch of billfishes taken 

 by each fishing mode during a standard year. The results 

 indicate that 245 t marlin and 42 t sailfishes were taken 

 as bycatch by the European purse seiners in the eastern 

 Atlantic Ocean. Figures 1 and 2 depict the spatial distribu- 

 tion by fishing modes of the marlin bycatch in the eastern 

 Atlantic for all data collected during the observer program. 

 Figure 2 shows that marlin bycatch in FAD sets were 

 spread across a large area in the eastern Atlantic Ocean. 

 When compared to FAD sets, the occurrence of marlins 

 bycatch in school sets was lower, specifically in central 



