482 



Spatial and temporal patterns In the bycatch 

 of seabirds In the Argentinian longllne fishery 



Patricia Gandini 



Esteban Frere 



Centre de Investigaciones de Puerto Deseado 

 Universidad Nacional de la Patagonia Austral 

 Conseio Nacional de Investigaciones Cientificas y Tecnicas (CONICET) 



and 



Wildlife Conservation Society 



Avda, Prefectura Naval s/n, cc 238, (9050) 



Puerto Deseado Santa Cruz, Argentina 



E-mail address (for P Gandini) pagandinnayahioo com ar 



Longline fisheries have grown 

 throughout the world's oceans for 

 more than 40 years. This type of 

 fisheries has captured high-quality 

 fish (mature individuals rather than 

 unwanted juveniles), has had minimal 

 destructive effects on bottom habi- 

 tats, and has produced a low bycatch 

 of nontargeted fish (Brothers et al., 

 1999). Seabirds, however, are hooked 

 accidentally when they swallow or are 

 snagged on the baited hooks set by 

 commercial longline crews (Brothers, 

 1991; Barnes et al., 1997; Tasker et 

 al., 2000; Belda and Sanchez 2001; 

 Jahncke et al., 2001). 



Population declines of several spe- 

 cies of albatrosses and petrels in 

 the Southern Ocean are linked to 

 longlining operations (Croxall and 

 Prince, 1990; Brothers, 1991; Cherel 

 et al., 1996). The importance of the 

 Patagonian shelf waters as a forag- 

 ing habitat for seabirds is well docu- 

 mented (Cooke and Mills, 1972; Veit, 

 1995), particularly for black-browed 

 albatross (Thalassajxhe melanophi'is; 

 Gales, 1998). We estimated the mag- 

 nitude of seabird mortality in the 

 kingclip {Genyptei-us blacodes) fishery 

 in the Argentine Exclusive Economic 

 Zone (EEZ). 



Materials and methods 



Description of the fishery 



Demersal longline vessels target two 

 species on the Patagonian shelf of 

 Argentina: kingclip and Patagonian 



toothfish (Dissostichiis eleginoides). 

 Six vessels operate within Argen- 

 tina's EEZ in the South Atlantic: 

 three vessels, Marunaka. Estela. and 

 Magallanes II, fish mainly kingclip 

 off Comodoro Rivadavia and Puerto 

 Deseado and three vessels, Antarc- 

 tic I. II and ///, target Patagonian 

 toothfish off Ushuaia. Another twelve 

 vessels based in the Malvinas (Falk- 

 land) Islands and operating outside 

 Argentina's jurisdiction target king- 

 clip and Patagonian toothfish (Blake, 

 2001; Fig. 1). Within Argentina's EEZ, 

 an artisanal longline fleet of 70 ves- 

 sels target hake (Merluccius hitbbsi). 

 fishing mainly in shallower waters 

 within the San Matias Gulf. 



The kingclip fishery operates year 

 round with trips lasting up to 65 

 days. Vessels are equipped with an 

 autoline system; the line is 10 km in 

 length and has up to 20,000 hooks 

 which are baited with thawed squid 

 at night by using a baiting machine. 

 Offal is released strategically to at- 

 tract birds to the opposite side of the 

 vessel away from the main line dur- 

 ing hauling. 



Data collection 



Observers, trained in seabird iden- 

 tification and in quantifying mortal- 

 ity, estimated fishing effort and the 

 bycatch of seabirds. Data were col- 

 lected through a special agreement 

 between a fishing company (Argenova 

 S.A.) and the university (Universidad 

 Nacional de la Patagonia Austral). 

 Observers sampled 156 sets and 



hauls from December 2000 to Sep- 

 tember 2001. Sets were deployed as 

 early as 23:00 and as late as 04:00. 

 Observers recorded the date, num- 

 ber of hooks deployed, position and 

 time of the start and end of each set, 

 water depth, and sea condition. If a 

 bird was caught, or if we observed 

 flapping wings or the struggle of a 

 bird, we recorded the event. During 

 hauling we identified and counted all 

 drowned birds. We collected a sample 

 of drowned birds and brought them to 

 the laboratory to determine their sex. 

 Observers sampled the sets of one of 

 the three vessels targeting kingclip 

 in Argentina's EEZ. We estimated to- 

 tal bird catch using observed bycatch 

 and official information, such us the 

 number of hooks deployed from log- 

 sheets of the other vessels. 



Data analysis 



We used the bootstrapping method to 

 calculate the 95% confidence limits for 

 the mean number of birds caught/1000 

 hooks and made 10,000 simulations 

 (n = 156). Using logistic regression, 

 we determined which variables best 

 explained the "capture events." Ini- 

 tial independent variables included 

 date, number of hooks deployed, water 

 depth, latitude, sea condition, time at 

 which the setting began, and moon 

 phase. Moon phase was scored from 

 to 14, where represents a full moon 

 and 14, a new moon. To determine 

 the contribution of each variable to 

 seabird mortality rates, we used a 

 stepwise multiple regression (Zar, 

 1984). On sets with catch per unit of 

 effort (CPUE) greater than 0, three 

 separate multiple regressions were 

 performed: 1) for total seabirds, 2) 

 for albatrosses, and 3) for petrels. 

 Seabirds whose sex was determined 

 were used to test differences from the 

 1:1 sex ratio by using a x~ test (Zar, 

 1984). We compared, using this test, 

 the number of birds of each species 

 caught during winter, summer and 

 autumn (Zar, 1984). 



Manuscript submitted 28 May 2003 

 to the Scientific Editor's Office. 



Manuscript approved for publication 

 24 October 2005 by the Scientific Editor. 



Fish. Bull. 104:482-485(2006). 



