CCAMLR Science, Vol. 2 (1995): 111-121

SEABIRD INTERACTIONS WITH LONGLINING OPERATIONS FORDISSOSTICHUS ELEGINOIDES AROUND SOUTH GEORGIA, APRIL TO MAY 1994

J.R. Ashford* and J.P. CroxallBritish Antarctic Survey

Natural Environment Research CouncilHigh Cross, Madingley Road

Cambridge CB3 0ET, UK

*Current postal address:Applied Marine Research Laboratory

Old Dominion University, 1034 West 45th StreetNorfolk, Va. 23529-0456, USA

P.S. Rubilar and C..A. MorenoInstituto de Ecologia y Evolucion

Universidad Austral de ChileCasilla 567, Valdivia, Chile

Abstract

Longlining operations for Dissostichus eleginoides off South Georgia were assessed forincidental mortality and interactions with seabirds. Twenty-seven line sets wereobserved and 98 deaths of seabirds were recorded over 20 sets; no deaths occurredduring hauls. The 16 sets made at night contributed 15% of overall mortality, all ofwhite-chinned petrels (Procellaria aequinoctialis); the four day sets contributed 85% ofdeaths, with giant petrels (Macronectes sp.), grey-headed albatrosses (Diomedeachrysostoma) and black-browed albatrosses (D. melanophrys) predominating.Grey-headed albatrosses, whose populations at South Georgia are in serious decline,were disproportionately affected in relation to their numbers near the fishing vessel;giant petrels may be disproportionately affected in relation to the size of theirbreeding population at South Georgia when compared to albatrosses. The averagemortality rate for the 20 sets was 0.48 birds/1 000 hooks and maximum mortality3.12 birds/1 000 hooks. Mortality and interactions of birds with operations varied withsite and time of day, and was affected by behavioural interactions between birds.Setting only at night would dramatically reduce the number of albatross deaths, andrestrict mortality to white-chinned petrels. A streamer line made to CCAMLRspecifications may reduce mortality but may be less effective during calm weather,intense feeding activity by seabirds, or when incorrectly constructed.

ROsum,O

Une ~valuation des op@rations de p~che a la palangre de Dissostichus eleginoides au largede la G~orgie du Sud a ~t~ effectu~e pour’ mesurer la mortalit~ accidentelle etl’interaction avec les oiseaux de mer. La pose de 27 lignes a ~t~ observ6e; sur 20 d’entreelles, 98 morts accidentelles d’oiseaux marins ont ~t@ enregistr~es; la remont~e des lignesn’a provoqu@ aucune mort accidentelle. Les 16 poses effectu@es de nuit ont caus@ 15%de la mortalit~ totale, et n’ont touch~ que des p~trels a menton blanc (Procellariaaequinoctialis); les quatre poses de jour ont caus@ 85% de la mortalitY, et ont affect~principalement des p~trels g~ants (Macronectes: sp.), des albatros ~ t~te grise (Diomedeachrysostoma) et des albatros ~ sourcils noirs (D. melanophrys). Par rapport ~ leur nombre~ proximit6 du navire, les albatros ~ t~te grise ~taient particuli~rement affect~s; si on lescompare aux albatros, les p~trels g~ants sont peut-~tre touch@s de manibredisproportionn@e, vu la taille de leur population reproductrice en G@orgie du Sud. Letaux de mortalit~ moyen pour les 20 poses ~tait de 0,48 oiseau/1 000 hame~ons et lamortalit~ maximale de 3,12 oiseaux/1 000 hame~ons. La mortalit~ des oiseaux de mer etl’interaction de ces derniers et des operations de p@che d~pendaient du site et dumoment de la journ~e, et @taient influenc@es par le comportement des oiseaux entre eux.Ne poser les palangres que de nuit r~duirait consid@rablement la mortalit~ des albatros

111

Ashford et al.

et la restreindrait aux p@trels a menton blanc. La ligne de banderoles pr@conis~e par laCCAMLR peut r~duire la mortalitY, mais risque d’etre moins efficace par temps calme,lors de l’approvisionnement intensif des oiseaux de mer, ou lorsqu’elle n’est pas mont6ecorrectement.

Pe3~oMe

~Ipycr~bifi npoMblce.a Dissostichus eleginoides B pafione IOxci~o~ Feoprnn 6b~.rIIqCC.qe,ROBaH C TOqKI4 3peHI~I9 IIO6OqHO~ cMepTHOC’rI4 MOpCKIeIX rrrldI! i4 B3aFIMO,!le~ICTBFII~IMe:KAy FITFIIlaMI/I I4 rIpOMb~C.,qOM. Ha6.,qI~OAeHI, I~ npOBOJlI&qtICb B ’reqeHr~e IIOCTaHOBKF127 ~tpycoB; 6b~ao 3apertic~rpnpoBaI-IO 98 c.ayqaeB cMepTI, I nTnU npn 20 nOCTaI-IOBKaX.CayqaeB cMepTI~ nTnIl npn BbI6OpKe ;~pycoB OTMeqeso .e 6b~ao. 15% o6useficMepTI-IOCTn npI~xo)ltl.qOCb rla 16 nOCTarIOBOK, Bblno~n-ierlrib~x I-IOqbm, Bce novn6mnenTnt!bi - 6e.aorop.abxe 6yper~ecTsiaKn (Procellaria aequinoctialis). 85% o6tuefiCMepTHOCTI4 npnxo)~n~OCb Ha qeTblpe ~3:HeBHBIX HOCTaHOBKFI; cpe~n norn6mnx IITHLInpeo6~aaa.~n rnraHTCKne 5ypeBec~rHnKn (Macronectes sp.), ceporo~oBbiea3bSaTpocbI (Diomedea chrysostoma) ~ qepBoSposbie a3365aTpOCbL Ceporo~osbiea~b6aTpocbI, pa3Mep nony~uunn KOTOpbIX pe3KO coKpaTri~C~, oKa3a33ncbno~Bep)KenbI nOSOqSOfi CMepTHOCT!4 B pa3Mepe, senponopttnOHa~bnOM K I4Xqnc~qennocTI~ n5~n3n npOMbIC.~OBOrO cy)lsa. HO cpaBHennm C a~bSaTpoca~nF~IFaHTCKne 6ypeBeCTSnKI4, BO3MO:~HO, oKa3bIBaK)TC~ senponop~I4OHa.qbSOno~Bep)KeSb~ cMepTHOCTn no OTHOmeHnrO K pa3Mepy nx rHe3~osofi nony~unn HaIO~KnOfi Feoprnn. B Teqesi~e 20 nOCTaHOBOK 9pycoB cpean~ 05ma~ cMepTriOCTbCOCTaBI~-qa 0,48 nTn~bI/1000 KpfOqKOB, fl MaKcaMa37bnau CMepTHOCTb -3,12 nTri~Ibffl000 KplOqKOB. YpoBenb CMepTHOCTI~ ~4 B3anMo~efiCTBn~i Me)K~yHTI4IIaMH I4 npOMblC3~OM Mt2H~!~2TC~/ B 3aB~ICHMOCTI~ OT IlpOMbICYIOBOFO yqaCTKa tlBpCM~2HFI CyTOK, a TaKz~e IIOB¢~2Hq~CKFIX OCO~HHOCT~ Me)KAy HTHId(aMIrl.HOCTaHOBKI~ ~IpyCOB TOZIbKO HOqblO B 3Haq~ITe3~bHOfi M~pe MOFyT CoKpaTttTb qHC.rIOc.qyqaeB CM(~pTHOCTH a-qb6aTpOCOB, a TaKTKe oFpan~lq~ITb CM~pTHOCTb IITFIIICM(~pTHOCTbtO ~YIOFOpYlbIX ~ypeB(~CTHI4KOB. I/ICHO3-Ib3OBaHH~ YII4H~I C OTI-IyFI4BaIOIZII4MI4

C1337a)KKaMn, pCKOMeH~OBaHHOFO AHTKOMoM, MO;KeTb rlOH~i31,ITb ypoBenbCMepTHOCT~i, o~naKo eFO ~efiCTB~Ie MO)KeT OKa3aTbCfl Menee 3q])CI)eKTI4BHbIM B c~yqaeTHXOfi I1OFO~bI, IIHTeHC~IBHOFO KOpMYleHId~I MOpCKHX nTml nan B c~yqae, ec33n OHnenpaBI4~bno CKOnCTpynpoBan.

Resumen

Se hizo una evaluaci6n de las operaciones de pesca de palangre de Dissostichuseleginoides realizadas alrededor de Georgia del Sur con respecto a la mortalidadincidental y alas interacciones con ayes marinas. Se observaron 27 lances y seregistraron 98 muertes de aves marinas en 20 calados de la linea; no se registraronmuertes durante el izado. Los 16 calados de palangre hechos de noche contribuyeronun 15% a la mortalidad total, constituida exclusivamente de petreles de ment6n blanco(Procellaria aequinoctialis). Los cuatro calados de palangre hechos de d~a contribuyeronun 85% a la mortalidad total, constituicta predominantemente de petreles gigantes(Macronectes sp.), albatros de cabeza gris (Diomedea chrysostoma) y albatros de ceja negra(D. melanophrys). Los albatros de cabeza gris, cuyas poblaciones en Georgia del Surestzln disminuyendo en forma grave, fueron afectados de manera desproporcionada enrelaci6n a su ntimero cerca de la embarcaci6n pesquera; es posible que los petrelesgigantes sean afectados de manera desproporcionada al tamafio de su poblaci6nreproductora en Georgia del Sur cuando se le compara con los albatros. La tasa demortalidad media para los 20 calados es de 0.48 aves/1 000 anzuelos y la mortalidadmaxima 3.12 aves/1 000 anzuelos. La mortalidad y las interacciones de las aves con lasoperaciones variaron segt~n el lugar y hora del dia, y fueron afectadas por interaccionesentre las aves. Si los calados de palangre se hicieran s61o de noche, se reducirladram~ticamente el n6mero de muertes de albatros, y la mortalidad se limitaria apetreles de ment6n blanco. E1 uso de lineas espantapajaros fabricadas segtin las normasde la CCRVMA podrla reducir la mortalidad, pero ellas podrian ser menos efectivasdurante los periodos de calma, o de intensa actividad alimentaria de las aves, o cuandohan sido construidas incorrectamente.

Keywords: fishery, incidental mortality, albatross, petrel, streamer line, South Atlantic, CCAMLR

112

INTRODUCTION

Declines in populations of several species ofalbatross, usually associated with high adultmortality rates, have been reported from variousSouthern Ocean breeding sites (Gales, 1993).These declines have been linked to longliningoperations, notably those of the fishery forsouthern bluefin tuna (Thunnus maccoyi). Brothers(1991) calculated that in the southern hemispherethe fishery could be killing 44 000 albatrossesannually, and that the introduction of streamerlines to deter albatrosses could save the lives of30 300 albatrosses and enable fishermen to save$4.9 million in bait lost to birds. Murray et al.(1993) estimated a minimum of 3 652 seabirdscaught in New Zealand waters in 1988; this figuredeclined to 360 in 1992, probably due to theintroduction of mitigating measures.

Within the area covered by the Convention forthe Conservation of Antarctic Marine LivingResources (CCAMLR), longlining operations forDissostichus eleginoides began in the late 1980s andhas so far been confined mainly to the vicinity ofSouth Georgia and Kerguelen Islands.Considerable concern has been expressed aboutthe effect these longline fisheries, in such closeproximity to major breeding grounds of severalspecies of albatross, could have on albatrosspopulations. In response to this concernCCAMLR introduced in 1991 the requirement thatstreamer lines be deployed during longlinefishing and encouraged Members to employscientific observers to acquire accurate data on theeffectiveness of these measures and on the natureand magnitude of incidental mortality of seabirds.Accordingly, in 1992/93 observers, including onefrom the United Kingdom, were placed on boardthe FV Friosur V, a Chilean longliner conductingan exploratory cruise aimed at detectingD. eleginoides around the South Sandwich Islands.Although no seabird deaths were recorded duringthe setting of seven longlines, several species ofseabird were shown to be vulnerable to longliningoperations around the South Sandwich Islands;observations of high numbers of birds nearfishing vessels were also made around SouthGeorgia (Ashford et al., 1994).

At the 1993 CCAMLR meeting, the ScientificCommittee recommended that the Commissionconsider mechanisms for placing scientificobservers on a high proportion of longline vesselsin the CCAMLR Convention Area for at least onefishing season to collect the data required for areliable assessment of the number and species of

Seabird Interactions with Longlining

birds incidentally captured on longlines in the,Convention Area. For the South Georgia area,(Subarea 48.3), in the 1993/94 season, the total,allowable catch of D. eleginoides and the fishing..season were divided up and allocated to five,equal time periods, with only one vesselbelonging to one CCAMLR Member beingpermitted to fish in each period. Reports on the,effectiveness of measures aimed at mitigatingincidental mortality of seabirds were required tobe part of the research plan to be submitted by,each Member operating in this fishery. Chile.agreed to take part in the fishery, and as in1992/93 the FV Friosur V was used, with the sameUK observer on board. The aims of the observerwere to record seabird deaths and interactionswith longlining operations, and the effectivenessof a streamer line as a mitigating measure.

MATERIALS AND METHODS

Fishing equipment and operations were asdescribed in Ashford et al. (1994). Most hooks laidwere of the type Maruto no. 26. Approximately5% were of the type Maguro no. 14 which has acurved point (Figure 1); these hooks were laid aseither one or two continuous series on each line.Hooks were baited individually by hand but asmall proportion of baits was lost during setting.Observations were made during both setting andhauling. Information was recorded on data sheetsas used by the New Zealand Ministry ofAgriculture and Fisheries in its Fisheries ScientificObserver Program for the southern bluefin tuna

Figure 1:Fishing hooks used in longline operations by FV Friosur Vat South Georgia, April to May 1994.

113

Ashford et al.

fishery (Murray et al., 1993). Initially, fourcategories of bird activity were recorded duringsetting: strikes at baits, birds hooked, bait-stealing(i.e., successful strikes), and baits dived for. Thenumber of birds hooked at haul-in and releasedalive was also recorded. As there was too muchbird activity during daytime sets to record allevents accurately, different data were collected forthe different sections of the line, delimited by themarker buoys. For two of the four sections, allstrikes and dives for bait were combined as strikeevents and only these recorded; for the other twosections, only the outcomes (i.e., success rate ofbait-stealing and number of birds hooked) of asample of strike events were recorded.

One observer undertook all settingobservations, standing at the stern hatches or onthe deck directly above. Interactions for theperiod observed were noted, and the number ofbirds of each species in the vicinity of the shipwas recorded every 30 minutes. Of the 20 sets forwhich mortality data are available, 16 took placeat night. Visibility during the hours of darknesswas 50 m when the ship’s lights were on: thisextended beyond the point at which the baits sankand most bird species could be readily identifiedat this distance. White-chinned petrels (Procellariaaequinoctialis) were more difficult to identify andobserve, because of their dark colouration, butcould be readily observed within 20 m of thevessel’s stern. Observations were made for allsetting periods.

A streamer line was used for approximatelysix of the 20 sets and a noise-emitting scarer,based on regular or irregular releases ofcompressed air, for another six. The streamer linedid not meet CCAMLR specifications: althoughattached to the ship at the right height from theaft gantry on the port side, the main line wasmore than 1 cm thick, there were no swivels, andthe streamers consisted of lines more than 3 mmthick, each approximately 1.5 m long, along thelength of the line; they did not therefore reach thesurface of the water until well aft of the ship’sstern. The streamer line was located over andclose to the main line. After the first 20 sets, astreamer line meeting CCAMLR specificationswas used to compare the effectiveness of designs.A streamer line was also hung over the haulingpoint during the last seven hauls, to deter birdsfrom approaching the incoming longline.

Observations of hauling operations were madeon a rota system, the duty observer standing onthe factory deck with a clear view of the hauling

point. Approximately 90% of hauling operationswere observed by at least one observer. Everytwo hours, estimates were made of the totalnumber of birds of each species within 5 m and10 m of the hauling point, and within a squarerunning 250 m out from either side of the vesseland 500 m aft of the hauling point; these estimateswere all made by a single observer. During haulscomparisons were made of (i) the effect ofcontinuous discharge of offal from a waste chute2 m from the hauling point with a singledischarge of offal at the end of a haul and (ii) birddensity when discarding offal from the wastechute with discarding from the opposite (port)side of the vessel. A streamer line was hung overthe hauling point during the last seven hauls.

RESULTS

Fishing in waters around South Georgia beganon 6 April 1994. On 7 May a serious fire on boardprevented any further operations and destroyedall data collected, with the exception of reportspreviously sent by radio. Samples of birds killedduring longline operations were preserved infreezers: 28 of these samples were off-loaded atSouth Georgia and data from the labels saved.Due to lack of freezer space, the remainder werereturned to Chile on FV Friosur V where theywere accidentally discarded during unloadingoperations. The data salvaged are presented inthis paper; other information, based onobservations and data recalled by the observer, isalso reported.

Incidental Mortality

Twenty-seven longline set and haul operationswere observed; their locations are shown inFigure 2. Their timing and duration, as well asdata on incidental mortality of seabirds, areshown in Table 1. The observer recollects a totalof 98 birds being caught during the first 20 ofthese operations. All of these birds were dead onarrival on deck, with most deaths occurring inaggregations. Inspection of the carcasses showedthat deaths were caused by hooks catching in thebeak and oesophagus during swallowing of hookand bait, by hooks catching in the birds’ breasts orby entanglement with hooks and lines. A total of206 720 hooks was laid, giving a mean mortalityrate of 0.47 birds/1 000 hooks laid. The maximumnumber of deaths on a single line was 34, giving amaximum rate of 3.12 birds/1 000 hooks laid.Some baits were lost during setting, so mortality

114

4

Figure 2:

27

V/illis Islands

Location of start and finish of each of the 27 longline operations carried out on FV Friosur V at South Georgia, April to May 1994.

Ashford et al.

Table 1: Ti.m.ing and duration of longlining operations by FV Friosur V at South Georgia, April to May 1994,with data on seabird mortalities (data both salvaged and recalled). All times are in GMT. GHA -~rey-headed albatross, BBA - black-browed albatross, GP - giant petrel, WCP - white-chinned petrel.t2awn at approximately 0830, sunset at approximately 1700. For streamer lines (STR): A - CCAMLRspecification, B - other, N - none, blank - no data.

No. Longline Set Day (D) STR Haul Seabird DeathsStart End Night(N) A/B Start End Total Comments

01 23.30 01.06 N 26.40 05.4002 00.06 01.50 N 03.15 15.5003 00.00 01.37 N 03.00 16.0004 00.09 01.46 N 03.10 15.50O5 01.00 02.45 N 04.00 17.1506 02.38 04.15 N 05.40 17.4007 02.48 04.33 N 05.50 19.00O8 03.12 04.43 N 06.20 19.4009 04.30 05.53 N 07.05 19.35 t15 All WCP; min. 4 WCP in10 03.49 05.25 N 06.50 18.05 ] set 7; streamer B used in11 02.05 03.42 N 05.05 19.25 ] 3 sets; noise-emitting scarer12 03.43 05.20 N 06.55 19.25 ] in 6 sets; no device in 6 sets.13 03.35 05.10 N 06.40 18.2514 02.35 04.08 N 05.35 18.0015 02.13 03.48 N 05.15 17.2516 07.04 08.38 D N 10.05 00.30 2717 08.44 10.25 D B 11.50 01.00 3418 09.15 10.58 D N 12.15 24.50 18 Incl. 7 GHA, 6 BBA, 4 GP19 01.05 02.47 N B 11.50 04.30 020 No data 14.00 D B 18.30 06.35 4 Incl. 2 GHA, 1 BBA, 1 GP21 14.44 16.31 D A 18.00 08.2022 16.26 18.08 D A 19.35 09.0023 09.30 11.25 D A 20.10 12.3024 20.47 22.28 N A 06.00 13.35 / Mortality rate of25 21.55 23.27 N A 00.50 14.00 ] 1-5 birds/line26 23.16 01.02 N A 13.45 04.4027 12.53 14.26 D A 16.00 05.20

in relation to the number of baited hooks laid maybe higher. During the first 15 sets, all made atnight, 15 white-chinned petrels were caught, witha maximum of at least four birds (on set 7) caughton any one occasion. No birds were caughtduring the last night-time set (set 19).

The remaining four sets (16 to 18, 20), all tothe north and northwest of Bird Island, weremade in daylight. A total of 83 birds (85% of theoverall total) was caught; of these, 28 birds wereoff-loaded at South Georgia and identified. Thesecomprised nine grey-headed albatrosses(Diomedea chrysostoma), seven black-browedalbatrosses (D. melanophris), five giant petrels(Macronectes sp.), and four Cape petrels (Daptioncapense). Cape petrels were caught by becomingentangled in the lines. It was recollected that thebirds caught also included several white-chinnedpetrels and one juvenile wandering albatross

(D. exulans). Furthermore, black-browed andgrey-headed albatrosses were caught in similarproportions (as can be seen from the specimensretained), but in general giant petrels were moreabundant than the other two species.

Taking all these data together, a reasonableassessment of the contribution of each seabirdspecies to the overall rate of incidental mortality isprobably as follows:

day-time hauls: giant petrel 35%, grey-headedalbatross 25%, black-browed albatross 25%,white-chinned petrel 7%, Cape petrel 7%,wandering albatross 1%; and

all hauls: giant petrel 30%, grey-headedalbatross 21%, black-browed albatross 21%,white-chinned petrel 21%, Cape petrel 6%,wandering albatross 1%.

116

Seabird Interactions with Longlining

For lines 21 to 27, which were a mixture of dayand night-time sets, one to five birds were killedper set.

Interactions During Setting

Large numbers of white-chinned petrels wereseen to the southwest of South Georgia (sets 5to 15) during night-time setting. These occurredin aggregations of 20 to 30 birds, flying in anddiving at the line within 30 m of the stern. Birdswere submerged for up to five seconds, andtended to dive in groups. Observations could notbe made of the line near the surface beyond 20 m.Some black-browed albatrosses, giant petrels andCape petrels also followed the ship, but rarelyapproached the line.

During setting in daylight to the north andnorthwest of Bird Island, large numbers of giantpetrels (average c. 30), Cape petrels (c. 100) andblack-browed albatrosses (c. 30) were present;wandering and grey-headed albatrosses occurredin smaller numbers (c. 5 and c. 7 to 10,respectively). Behavioural differences wereobserved between species. Black-browed andgrey-headed albatrosses were the most persistentbait-stealers: the former approached the line onthe surface as close as 15 m behind the stern,grabbed the bait and then pulled until it camefree; the latter tended to dive from about 3 mabove the water surface, submerging briefly.Once the bait was taken, but often while still onthe hook, other birds (giant petrels in particular,but also all three species of albatross) wouldattempt to steal the bait, forming aggregationswhich persisted up to 300 m behind the stern.Cape petrels were ubiquitous, approaching closeto the vessel and around the margins of theaggregations. Cape petrels, black-browed andgrey-headed albatrosses all approached the linebefore it entered the water, and were vulnerableto hooks flailing in the air. High mortality ratescoincided with intense feeding activity whenbirds acted in a frenzied manner.

Interactions at night during sets 24 to 26 wereminimal: a few white-chinned petrels, giantpetrels and black-browed albatrosses wereobserved. Daytime numbers were still high butwith few giant petrels and no grey-headedalbatrosses seen, and without the intense activitywhich occurred previously.

No data on bait-strikes survived, but it wasrecalled that success rate varied between species:

approximately 70% for black-browed albatross,40% for grey-headed albatross, and 70% for giantpetrels. Aggregations formed quickly aroundhooked birds, obscuring further observation andmaking it difficult subsequently to distinguishbetween birds pulling on a bait or actuallyensnared on a hook. Consequently, no reliableestimate of birds hooked during setting isavailable.

Interactions During Hauling

No deaths were observed as resulting fromhauling operations. The tail of a single giantpetrel was observed to catch briefly on anincoming hook but the bird freed itself easily withno obvious damage incurred.

From memory, large numbers of birds were inthe vicinity of the vessel during haulingoperations both in daylight and at night. Theseincluded black-browed, wandering andgrey-headed albatross; giant, Cape andwhite-chinned petrels; Wilson’s and black-belliedstorm petrels (Oceanites oceanicus and Fregettatropica respectively). A plume of birds formed aftof the offal chutes on the port and starboard sidesto a distance of up to 1 km aft of the stern of theship, with white-chinned petrels tending to stayon the flanks of the plume. Numbers of albatrossspecies were considerably reduced at night; stormpetrels were not observed at night.

During daytime hauling of lines southwest ofSouth Georgia, the mean total of birds countedwithin a 500 m square aft of the stern wasapproximately 700 birds, although this variedbetween and during operations. Of these, themajority (c. 400) were giant petrels; smallernumbers of Cape petrels (c. 200), white-chinned]petrels (c. 50), black-browed (c. 30) andwandering (c. 10) albatrosses, and Wilson’s stormpetrels (c. 10) were also present.

To the north and northwest of Bird Island,~otal numbers during daytime hauling rose to anapproximate mean of 1 400: giant petrels wereagain in the majority (c. 800), with Cape petrels,~c. 400), black-browed (c. 100), wandering (c. 60)and grey-headed (c. 10) albatrosses, Wilson’s,Ic. 10) and black-bellied (c. 30) storm petrels alsopresent. Very few white-chinned petrels wereobserved, either at night or during the day.

For the last two sets, numbers of birds were:reduced to several hundred: initially, few

117

Ashford et al.

giant petrels but increasing later, 80 to100 black-browed albatrosses and 40 wanderingalbatrosses. Some kelp gulls (Larus dominicanus)were also present but few white-chinned andCape petrels, and no grey-headed albatrosseswere observed.

Up to 70 birds were present within 10 m of thehauling point during operations in daylight andat night. Giant petrels were the most common(approximately 60% of all birds) but Cape petrels(30%) and black-browed albatrosses (10%) werealso present. Grey-headed and wanderingalbatrosses and white-chinned petrels approachedthis close only occasionally. Birds made noattempt to take baits on hooks. However, theywere attracted by offal from the waste chute 2 maway and often encroached on the area aroundthe incoming line, becoming vulnerable toactivities associated with gaffing of incoming fish.

Effectiveness of Mitigating Procedures

Little quantifiable information survivedconcerning the effectiveness of mitigatingprocedures: the maximum mortality for a set wassustained while using the first streamer line; incontrast, low daytime mortality rates for sets 20 to24 were associated with use of the streamer linemade to CCAMLR specifications.

Both types of streamer line interfered withbirds’ approach to the baited hooks: the streamers;broke up the sweeping flight of white-chinnedpetrels across the mainline at night, and similarlyfor grey-headed albatrosses during the day. TheCCAMLR-designed streamers and line movedmore unpredictably, distracting birds more. Incalm weather or during intense feeding activity,the heavier streamers of the initial design had[very little effect; birds frequently, however,collided with the CCAMLR-designed line bul~were not harmed. Both streamer lines caught inthe main line and broke, the CCAMLR line alsolosing a streamer on a couple of occasions.

A noise-emitting scarer was found to have amitigating effect early in each set but this wore offquickly. A streamer line suspended off thehauling hatch was observed to reduce the numberof birds approaching the hauling point.:discarding offal from the opposite side of thevessel also reduced the numbers of birds at thehauling point.

DISCUSSION

Seabird Interactions with Longlining

No birds were hooked during hauling:all deaths must therefore have occurredduring setting operations. Brothers (1991)found a mean seabird mortality rate of0.41 birds/1 000 hooks and a maximum mortalityrate of 1.8 birds/1 000 hooks for the southernbluefin tuna fishery off Tasmania. These figureswere sufficiently high to substantiate claims thatserious declines in populations were probablydue to pelagic longlining. Both rates are lowerthan those found in this study, which alsoindicated a large impact on grey-headedalbatrosses, one of the species whose populationsare most seriously in decline at South Georgia(Prince et al., 1994). Wandering albatrosses,whose populations are also in decline at SouthGeorgia (Croxall et al., 1990), showed a lowfrequency of interaction which may be a result of(i) a marginal role during intense feeding activitydue to lack of manoeuvrability, and (ii) atendency, when on the sea surface, to remain wellaft of the ship. With the Japanese system oflonglining on board FV Friosur V, the line isweighted using stones and the baits tend to sinkquickly, reducing the opportunity for wanderingalbatrosses to interact; with other systems wheremore line or monofilament is used, baits may stayon the surface longer, allowing wanderingalbatrosses to interact more with settingoperations.

Murray et al. (1993) showed that a number ofvariables influenced the mortality of seabirds inthe southern bluefin tuna fishery off NewZealand. These included time of day, location,species present and behavioural interactions. Thisstudy provided evidence of similar effects offSouth Georgia: considerably more birds, andspecies (6), were caught per line during daytimeoperations to the north and northwest of BirdIsland than during night-time operations to thesouthwest of the main island, when only a singlespecies (white-chinned petrel) was caught.

Observations of birds in the vicinity oflonglining operations also showed variations invulnerability. More white-chinned petrels wereobserved during both night and day operations tothe southwest of the main island than in otherareas, whereas albatross species were mostcommon during daylight hours north andnorthwest of Bird Island and to the northeast ofSouth Georgia; the highest numbers of giant

118

Seabird Interactions with Longlining

petrels were found off Bird Island during the day,though large numbers were also observed tothe southwest of the main island and duringnight-time operations at both sites.

Colonies of white-chinned petrels are widelydistributed on headlands all around SouthGeorgia, including the southwest coast; albatrosscolonies are concentrated around northwest and,to a lesser extent, southeast South Georgia(Croxall et al., 1984). These distributions, linked tospecies’ feeding strategies, may contributesignificantly to the observed distribution of birdsaround the longline vessel. The timing of thebreeding season when longlining takes place mayalso be important: the decline in numbers ofwhite-chinned petrels observed coincided withthe end of their breeding season.

Behavioural strategies may also be significant.Thus, for grey-headed albatrosses, the proportionof deaths in relation to the number of birdsfollowing the ship was very much higher than forother species and may be linked to their diving onthe bait from the air; alternatively, there may be ahigher turnover in birds observed. In relation tothe size of the local breeding population, giantpetrels (8 500 pairs; Croxall and Prince, 1987) arecaught disproportionately frequently incomparison with black-browed (100 000 pairs)and grey-headed (80 000 pairs) albatrosses(Prince et al., 1994), and white-chinned petrels(c. 2 million pairs) (Prince and Croxall, 1983).White-chinned petrels approached the line anddived in groups, with deaths also appearing to beaggregated. High mortality during daytime setscoincided with intense feeding activity: this wasgoverned by complex behavioural interactionsrather than by any observable increase innumbers following the ship; interactions betweenbirds taking bait and those then stealing itappeared important and may also influence theproportion of deaths by entanglement in relationto those by hooks catching in the birds’ breasts,beaks or oesophagi.

No evidence of deaths or injury to birdsduring hauling was recorded. A problem mayexist where crewmen use gaffs to disperse birdscrowding around the hauling point. Numbers ofbirds following a ship may be related to theamount of offal thrown overboard; large numbersof birds attracted during a haul may theninfluence the number of birds present during thesubsequent set, especially if this occurs in thesame area or follows immediately after the haul.

lvlitigating Devices

Streamer lines were observed to interruptbirds’ behaviour when taking baits, and to reducemortality. However, these may not be effective ifthey do not conform to CCAMLR design, or ifthey are used during calm weather or periods ofintense feeding activity. The CCAMLR design,however, is complicated and labour-intensive tomake, and its fragility made it vulnerable tobreaking by catching on the mainline: fishermenmay therefore be less than willing to use andmaintain it. Noise-emitting scarers wereunreliable and had a limited effect on birds.

Offal may attract birds, but discarding it onlyat the end of a haul may be counter-productive:large numbers of birds were still attracted,remaining astern of the ship during hauling.When bait was discarded after hauling, birdsconverged on the discarding point to compete forthe offal, generating intense activity anddisplaying frenzied behaviour. The problem ofbirds obscuring visibility of the incoming lineduring hauling operations was considerablyreduced by suspending the streamer line off thehauling hatch; redesigning offal chutes so thatthey divert all offal to the side of the vesselopposite fishing operations may also reduce theproblem.

CONCLUSIONS

From this study, the single measure that couldmost effectively limit seabird mortality is to setlonglines at night only; this would substantiallyreduce the number of birds taken, and confine theeffect to white-chinned petrels. Further work tominimise mortality of white-chinned petrelsduring night-time setting is required, andmonitoring of the population breeding at SouthGeorgia would be desirable.

The use of streamer lines is also effective, butfurther work is needed on their design.Understanding the behaviour governing thepresence of bird species affected and the nature ofl~heir interactions with longlining operations alsorequires further studies.

Discharging all offal at or near the end of ahaul may have some adverse effects, in particularas it promotes very intense feeding activity andmay attract very large numbers of seabirds.Furthermore, any vulnerability of birds to hauling

119

Ashford et al.

operations may be reduced by discharging offalon the side of the ship away from fishingoperations.

ACKNOWLEDGEMENTS

We thank Roberto Arce; the Fishing Master,.Sr Oomiya; and the crew of the FV Friosur V for’their invaluable help in this project; also for their’courage and resourcefulness during and after thefire on board. We also honour the memories ofthose who died: Capitan Federico Horn Wheeler’and Patricio Oyarzun.

REFERENCES

Ashford, J.R., J.P. Croxall, P.S. Rubilar andC.A. Moreno. 1994. Seabird interactions withlonglining operations for Dissostichuseleginoides at the South Sandwich Islands.and South Georgia. CCAMLR Science, Vol. 1:143-153.

Brothers, N. 1991. Albatross mortality andassociated bait loss in the Japanese longlinefishery in the Southern Ocean. BiologicalConservation, 55: 255-268.

Croxall, J.P., and P.A. Prince. 1987. Seabirds aspredators on marine resources, especially krill,at South Georgia. In: Croxall, J.P. (Ed.).Seabirds: Feeding Ecology and Role in Marine’Ecosystems. Cambridge University Press,Cambridge: 347-368.

Croxall, J.P., P.A. Prince, I. Hunter, S.J. McInnesand P.G. Copestake. 1984. The seabirds of theAntarctic Peninsula, islands of the Scotia Seaand Antarctic continent between 80°W and20°W; their status and conservation. In:Croxall, J.P., P.G.H. Evans and R.W. Schreiber(Eds). Status and Conservation of the World’sSeabirds. ICBP, Cambridge.

Croxall, J.P., P. Rothery, S.P.C. Pickering and P.A.Prince. 1990. Reproductive performance,recruitment and survival of wanderingalbatrosses Diomedea exulans at Bird Island,South Georgia. Journal of Animal Ecology, 59:773-794.

Gales, R. 1993. Cooperative Mechanisms for theConservation of Albatross. National Parks andWildlife, Tasmania, Australia.

Murray, T.E., J.A. Bartle, S.R. Kalish andP.R. Taylor. 1993. Incidental capture ofseabirds by Japanese southern bluefin tunalongline vessels in New Zealand waters1988-1992. Bird Conservation International, 3:181-210.

Prince, P.A., and J.P. Croxall. 1983. Birds ofSouth Georgia: new records and re-evaluationof status. British Antarctic Survey Bulletin, 61:27-38.

Prince, P.A., P. Rothery, J.P. Croxall andA.G. Wood. 1994. Population dynamics ofblack-browed and grey-headed albatrosses atBird Island, South Georgia. Ibis, 136: 50-71.

Tableau 1:

Liste des tableaux

Heure et dur~e des op@rations de p~che a la palangre r@alis@es par le navire de p@che Friosur V enG~orgie du Sud, d’avril a mai 1994, ain~si que donn~es sur la mortalit~ des oiseaux de mer (tant lesdonn~es r~cup~r~es que les donn~es dont on a pu se souvenir). Toutes les heures sont en GMT.GHA - albatros a t@te grise, BBA - albatros a sourcils noirs, GP - p@trel g@ant, WCP - p~trel a mentonblanc. L’aube apparait a environ 8h30, le coucher du soleil a environ 17h00. Lignes de banderoles(STR): A - selon les specifications de la CCAMLR, B - autre, N - aucune, blanc - sans donn~es.

Figure 1:

Liste des figures

Hame~ons utilis~s au cours des operations de p~che du navire de p~che Friosur V en G~orgie du Sud,d’avril a mai 1994.

Figure 2: Position du d@but et de la fin des 27 operations de p@che a la palangre r~alis~es par le navire dep~che Friosur V en G~orgie du Sud, d’avril a mai 1994.

120

Seabird Interactions with Longlining

Ta6am~a 1: BpeMa npoBe~eHH~ I4 rlpojxo.rlYKIqTe.rlbHOCTb npoMbIC,qOBbIX onepaunfi cy]IHa Friosur V B pafioneIOZKHOfi Feoprrm, anpe:n,-Mafi 1994 r., BMeCTe C ~anH~,iMn no no6oqnofi cMepTHOCTn (KaKBocnpon3Be)lennbiMn, TaK n BOCCTanoB~enrlbiMn). BpeMa B~,Ipa~KeBo no FpnnBnqy.GHA - ceporo~oB~i~ a~b6aTpoc, BBA - qepno6poabl~ anb6aTpoc, GP - rarauTCKn~6ypeBecTnn~, WCP - 6e~orop:n, ifi 6ypeBeCTnnK. BOCXO)~ co.~nt~a - OKO:~O 0830, 3aKaT - OKO:~O1700. flrInn c oTnyrnBarO~nMn qb-~ax~KaMn (SRT): A - no cnetmqbnKatmzM AHTKOMa,B - npoqne, N - oWCyTCTByrOT, npo6e:I - ~aHm, lx neT.

PIqCyHOK 1:

P~CyHOK 2:

CHHCOK pHCyHKOB

KpioqKn, nCnO~b3OBaHH~e B xo~e npOM~CflOB~X onepa~nfi cy~naFriosur V B pafioneIO~HofiFeoprnn, anpeab-~afi 1994 r.

MecTonoflo:KeHne Haqa~a ri OKOHqaHH~q Ka:~K~Ofi ri3 27 IIOCTaHOBOK 14 Bbl~OpOK apycoB,rIpoBe~enHbiX c 6opTa cy~na Friosur V B pafioHe IOzKnofi Feoprnn, anpe~b-Mafi 1994 r.

Tabla 1:

Lista de las tablas

Horario y duraci6n de las operaciones de pesca de palangre efectuadas por BP Friosur V en Georgiadel Sur, de abril a mayo 1994, con datos de las mortalidades de las aves marinas (datos salvadosy recordados). Las horas est~in dadas en GMT. Se utilizan los siguientes c6digos de especies:GHA - albatros de cabeza gris, BBA - albatros de ceja negra, GP - petrel gigante, WCP - petrel dement6n blanco. E1 amanecer y el ocaso ocurren aproximadamente a las 0830 y alas 1700 horasrespectivamente. Lineas espantap~jaros (STR): A- especificaci6n de la CCRVMA, B - otra,N - ninguna, (espacio vacio)- no hay datos.

Figura 1:

Figura 2:

Lista de las figuras

Anzuelos utilizados en operaciones de pesca de palangre efectuadas por BP Friosur V en Georgia delSur, de abril a mayo 1994.

Ubicaci6n del comienzo y t6rmino de cada una de las 27 operaciones de pesca de palangreefectuadas por BP Friosur V en Georgia del Sur, de abril a mayo 1994.

121