northern gannets in the north sea: foraging distribution

60 © British Birds 104 • February 2011 • 60–76

Northern Gannets in the North Sea: foraging distributionand feeding techniquesaround the Bass RockKees (C. J.) Camphuysen

Abstract Field observationsaround the largest NorthernGannet Morus bassanuscolony in the North Sea, theBass Rock, showed that 66%of all Gannets foraged inareas with very low densitiesof conspecifics, more than100 km from the colony.When one forager foundprey, even distant Gannetsresponded by joining thefinder to obtain a share ofthe bounty but, because of the low densities of Gannets far from the colony,feeding opportunities were typically exploited by small flocks, with relatively fewcompeting birds. Intraspecific competition was thus less intense than it would havebeen nearer the colony. Searching and feeding tactics of Gannets, as well asforaging associations with other top predators, were different between sea areas.Low numbers of Gannets per flock occurred within inshore multi-species feedingassociations, where Gannets hampered feeding opportunities for other seabirds(and themselves) by plunge-diving into compact schools of small prey fish. Largerflocks of competing Gannets formed in situations where an escape response inprey fish was absent (discards behind commercial trawlers) or weakened (fishschools herded by marine mammals). The association of Gannets with marinemammals was typically an offshore phenomenon, despite the abundance ofcetaceans in inshore waters. Behind trawlers, Gannets focused mostly onroundfish, between 22 and 30 cm in length. Discards were, however, a fairlyunimportant source of food during the breeding season and natural feedingopportunities were widespread.

Dan

Pow

ell

IntroductionNorthern Gannets Morus bassanus (hereaftersimply ‘Gannets’) have shown a long-termpopulation increase and range expansion inthe North Atlantic (Nelson 2002). Theirsuccess may have been driven by a relaxationof direct exploitation (Fisher & Vevers 1951)and a concurrent increase in prey fish, forexample as a result of the overfishing of com-peting predatory fish (Sherman et al. 1981;Furness 1982; Camphuysen & Garthe 2000).Within the North Sea, new colonies havebeen formed and breeding numbers at long-established sites have increased spectacularly(Wanless et al. 1986, 2005; Wanless 1987). This population increase may have had

important consequences on the foragingopportunities around colonies. Positive cor-relations between colony size and feeding-trip duration have been reported, suggestingthat at larger colonies birds need to travelfarther, or search longer, for prey. It is ques-tionable whether seabirds could actuallycause significant prey depletion around theircolony (Birkhead & Furness 1985; Croxall1987; Nelson 2002). However, Lewis et al.(2001) suggested that although schools offish close to colonies were perhaps notentirely removed by predators, they would bemore frequently attacked and disturbed bydiving Gannets (thus becoming less available,

as a result of avoidance response behavioursby the fish) than fish schools farther away. Inthat situation, intraspecific competitionamong Gannets might make feeding moreprofitable in areas farther afield with ratherlower densities of competitors. An obvious question to ask is whether

there is any evidence that Gannets triggersuch an avoidance response in fish schools,with adverse effects on the birds’ own oppor-tunities for profitable feeding. Nelson (2002)suggested that Gannets are ‘positivelyadapted to highly communal fishing’,perhaps because the species is commonlyobserved in (large) groups leaving andarriving at their colonies. Communal feedingis, however, difficult to understand in thecontext of a ‘prey-disturbance hypothesis’where Gannets are assumed to disturb eachother’s foraging opportunities rather than toenhance them.So, how and where do Gannets feed

around their colonies? For a spectacularseabird such as the Gannet, in a well-studiedarea such as the North Sea, there is remark-ably little published material and first-handdocumentation on their foraging behaviourat sea. Feeding Gannets are well known fortheir spectacular plunge-dives when targetingshoaling fish (Boddington 1959; Reinsch1969; Nelson 1978, 2002), but they also

61British Birds 104 • February 2011 • 60–76

Northern Gannets in the North Sea

24. The Bass Rock gannetry, off the Lothian coast, July 2001.

David Tipling

commonly scavenge behind fishing vessels(Camphuysen et al. 1995b) and associatewith prey-driving marine mammals (Evans1982; Camphuysen et al. 1995a) or otherseabirds (Camphuysen & Webb 1999;Camphuysen et al. 2006). Several atlases andpapers have described distribution patternsbased on ship-based surveys within theNorth Sea (Tasker et al. 1985; Skov et al.1995; Stone et al. 1995) but none of thesehave been very specific with regard to spatialpatterns in foraging activity, let alone behav-iour. Recent studies using satellite transmit-ters and data loggers have been highlysuccessful in tracking individual birds andtheir foraging activities around the Bass Rock(Hamer et al. 2000; Humphreys et al. 2004).These devices have produced unique andvaluable information on location, depth (inthe water) and prey ingestion of foragingindividuals, and even on particular hydro-graphic properties of the areas these birdswere feeding in. However, data on foraginginteractions with other marine wildlife(including conspecifics) or on precise for-aging and feeding techniques in relation tothe foraging environment and specific feedingopportunities therein can be collected effec-tively only during visual observations at sea. This paper describes the behaviour of

Gannets at sea in multi-species feeding fren-zies and elsewhere, from visual observationsduring ship-based surveys and sessions ofexperimental discarding on board fisheriesresearch vessels in the North Sea.

MethodsThe data from midsummer ship-basedseabird surveys (carried out in June–July, forten seasons between 1991 and 2004, in thenorthwestern North Sea – an area covering54–59°N and 3°W–2°E) were analysed (seealso Camphuysen 2005). These surveyscovered most of the feeding range of Gannetsnesting on the Bass Rock (Hamer et al. 2000;Nelson 2002). The work was conducted inconjunction with the annual acoustic herringsurvey of FRV Tridens (ten seasons between1991 and 2004), and during a dedicatedsurvey on board RV Pelagia in 2003, usingstrip-transect counts, which have been devel-oped as a standard for the North Sea (Taskeret al. 1984; Camphuysen et al. 2004). Countswere conducted between the periods whenthe survey ship was fishing (these periodswere well spaced, normally no more than 3–4per day), when the number of ship-followerswas low and could be controlled for andwhen the vessel travelled at full speed (8–10knots). The observation platforms providedclear forward views (through 180°) at an alti-tude of 12–15 m above sea level. A total of9,972 km² were surveyed, travelling a dis-tance of 33,601 km, and 44,818 Gannets wereobserved during these surveys. A correctionfactor was not applied, with the assumptionthat all individual Gannets, either swimmingor in flight, were detected effectively withinthe 300-m-wide strip transect (counts werediscontinued in rough conditions, e.g. abovewind force 6 on the Beaufort scale). The birds

were aged according toplumage characteristics (fiveimmature plumage classeswere distinguished, in additionto adult plumage).To distinguish between for-

aging (searching for food),feeding and non-feeding birds,the behaviour of the Gannetsobserved was classified,recorded, coded and stored ina database. Inspired byAshmole (1971), 20 types offeeding behaviour and 16types of non-feeding behav-iour were distinguished. Thefollowing behaviours werecommonly observed in

62

Camphuysen

British Birds 104 • February 2011 • 60–76

25. The author, with co-observer Tanja Weichler, surveyingseabirds from the observation platform of FRV Tridens, July 2001.

Kees C

amph

uysen

Gannets: Actively searching, Deep plunging,Shallow plunging, Scooping prey, Surfaceseizing, Scavenging at fishing vessel, Restingor apparently asleep, Preening or bathing,Carrying nest material, and being Underattack by kleptoparasite (see descriptions intable 1; details in Camphuysen & Garthe2004). In the case of Gannets, the differencebetween direct flight and searching for prey isobvious under field conditions. Travellingbirds follow straight lines, with a fullystretched body and the bill pointing forward.Searching Gannets havesomewhat slower,‘bouncing’ wingbeats, theyoften look down, follow amore or less meanderingpath and/or circle over agiven location. An inci-dental peer down into thesea by a bird otherwiseclearly travelling was notrecorded as ‘searching’.Birds associating with orapparently attracted by theresearch vessels have notbeen included in calcula-tions of densities, patternsof relative abundance, oractivity budgets.

All birds, whether swimming or flying,that operated together in a particular area ormovement were categorised as ‘flocks’. Meanflock sizes and 95% confidence intervals werecalculated after log transformation of thedata to normalise the distributions and com-pared with the z-test (z = d/σ where d is thedeviation from the mean). The data wereanalysed in radius bands, or strata, of 20 km(covering the area 0–100 km from thecolony) and 50 km (100–400 km from thecolony) from the Bass Rock. Densities



26. A Northern Gannet Morus bassanus searching for prey (see text for differences betweentravelling and foraging birds) off the Bass Rock, August 2005.

Mike La

ne

27. Northern Gannet Morus bassanus in flight approaching a feedingopportunity, June 2004.

Kees C

amph

uysen

(Gannets per km2) were calculated for thesestrata and were transformed to total numbersof individuals at sea (density x surface area)to estimate how many individuals wereinvolved.Flocks comprising more than one seabird

species were named ‘multi-species (feeding)associations’, or MSFAs. Small, short-livedMSFAs are an important mechanism forobtaining prey off the east coast of Britain fornumerous species of piscivorous seabirds(details in Camphuysen & Webb 1999). Asso-ciations between seabirds and marinemammals involved birds that were activelydrawn to cetaceans, whose foraging activitiesdrive prey to the surface. Both the number ofbirds and the number (and species) ofmarine mammals involved were assessed andthe type of behaviour was described andlogged. Counts made during fishing opera-tions of the research vessel (FRV Tridens) aswell as counts made at nearby commercialtrawlers were used to examine flock size atfishing vessels. The presence, age composi-tion, flock size and the specific role (behav-iour) of Gannets within MSFAs associatedwith both marine mammals and fishing

64

Camphuysen


operations were recorded systematically.Finally, as part of EC-funded research

projects conducted in February 1993 andMay, August and November 1994, the com-petitive strength and prey selection ofGannets in comparison with other scavengingseabirds was determined experimentally(Camphuysen et al. 1993, 1995b). During 536sessions of experimental discarding (carriedout throughout the North Sea survey area), insituations where Gannets were present, a totalof 61,304 discards were individually identi-fied, measured (fish only) and thrown inamong other discards; these included 1,361benthic invertebrates, 141 cephalopods(squid), 6 jellyfish, 7,735 particles of offal,3,618 flatfish and 48,443 roundfish. The exactfate of each of these particular items(dropped and sunk, or picked up and swal-lowed by a scavenger) was determined,including identifications of the scavenger thatpicked up an item first, perhaps dropped it orhad it stolen by another, until finally a morselor fish was swallowed and could be tallied asbeing consumed successfully. From theseexperiments we could deduce prey-speciesand prey-size selection (composition of prey

Table 1. Types of Northern Gannet Morus bassanus behaviour recorded and logged at sea (see Camphuysen & Garthe 2004).

Direct(ed) flight Straight flight with regular, powerful wingbeats and short glides in calmconditions, in higher winds, shortening wingspan and stroke, often banking and side-slipping. Head points straight forward during normal flight.

Actively searching Reduced speed, often meandering course, head pointing down, more bouncing wingbeats, often circling to inspect situations or sea areas more closely.

Deep plunging Headlong plunge-dive into the sea, bird disappears below surface.

Shallow plunging Headlong plunge-dive into the sea, bird remains visible at surface.

Scooping prey Sitting at the sea surface scooping up water and prey in the manner of a pelican.

Surface seizing Sitting at the surface picking up (dead) prey from the surface.

Scavenging at Any activity in association with a fishing vessel (searching, competing with fishing vessel other scavengers, feeding successfully).

Associated with MSFA Feeding frenzy of more than one seabird species, including Gannets that either check out the situation (circling and approaching closely) or join in the foraging activity.

Associated with Cetaceans or seals followed at close range in persistent searching mode marine mammals (see above), or actively feeding as a result of the surface activity of the mammals.

Resting or apparently Sitting at sea surface, no activity or sound asleep.asleep

Preening or bathing Sitting at sea surface, feather care.

Carrying nest material In flight with seaweeds or plastics (excluding entangled birds).

Under attack by Aerial pursuit by a kleptoparasite, such as a skua or gull.kleptoparasite

taken versus composition ofdiscards offered), as well asthe competitive strength ofeach of the bird speciespresent (number of prey itemsobtained and swallowed,including prey items stolenfrom others versus number ofprey items dropped or lost asa result of kleptoparasitism byother scavengers). This com-petitive strength was calcu-lated as a ‘robbery index’: thenumber of experimental dis-cards stolen by a speciesdivided by the number of dis-cards stolen from that species(Camphuysen et al. 1995b).Observed consumption rateswere compared with expecta-tions based on the numericalabundance of scavengers atthe trawl.For each of the observa-

tions, since geographical co-ordinates were known, thedistance (in km) from boththe Bass Rock colony and thenearest part of the Britishcoast was assessed. Waterdepth (m) was taken as theaverage depth within 10’ lati-tude x 20’ longitude rectan-gles. Throughout the paper,means are accompanied bystandard deviations (SD),derived means with 95% con-fidence intervals (CI), andsignificant test results refer toP<0.05 or less.

ResultsAt-sea distributionThe at-sea distribution ofGannets could be described asa roughly bell-shaped formaround the Bass Rock colony,with very similar declines in density with distancerecorded in all directions (fig.1). Mean densities of Gannetsdeclined from 11.4 per km2

within 20 km of the colony to



Fig. 1. Mean densities (number per km2 per 10’-latitude x 20’-longitude rectangle) of Northern Gannets Morus bassanus aroundthe Bass Rock, based on June–July surveys 1991–2004. The circlesindicate 100-km distance contours from the colony.

Density (n/km2)Blank = no counts

no birds0.01–0.490.50–0.991.00–1.992.00–4.995.00–9.9910.00 +

Fig. 2. Relative abundance (mean densities – n/km2 – and numberper km steamed – n/km) of Northern Gannets Morus bassanus andthe proportion of adults with increasing distance around BassRock based on June–July surveys 1991–2004.

0 20 40 60 80 100 150 200 250 300 350 400 450distance from Bass Rock (km)

18.0

16.0

14.0

12.0

10.0

8.0

6.0

4.0

2.0

0

100

90

80

70

60

50

40

30

20

10

0

relative abundance

prop

ortio

n (%

) adultn/km

n/km2

% adult

less than 0.17 per km2 at over 400 km fromthe colony (fig. 2). About 350 km north of theBass Rock, and within 150 km of the Scottishmainland, densities rose again, indicating aninfluence of the Shetland Gannet colonies onthe numbers of birds at sea in that region.There was little or no discernible effect on at-sea densities around the smaller colonies atTroup Head (North-east Scotland) andBempton Cliffs (Yorkshire), however. Within 200 km of the Bass Rock, over

90% of the birds were adults (fig. 2) but theproportion of adults declined to less than50% at 400 km or more from the colony.From extrapolated numbers of birds (based

on densities multiplied by sea surface areawith increasing distance, but within 400 kmof the colony), it can be estimated that abouttwo-thirds of all adult Gannets at sea aroundthe Bass Rock were recorded in areas wherethe mean densities (of Gannets) were verylow (mean 0.53 per km2). From the age com-position at sea, an uninterrupted zone of over90% adult in all the Gannets observedextended up to 200 km around the Bass innearly all directions, and up to 300 km fromthe colony to the ENE (fig. 3). Rather lowpercentages of mature birds were founddirectly off the North-east Scotland coast, inthe Moray Firth area and around most of the

Dogger Bank (to the south-east).

Densities and flock sizeGannets travel in groups(‘skeins’) and clearly profitfrom the aerodynamics of V-shaped or linear flight forma-tions, just as geese, swans orthe (more closely related) pel-icans do. Overall, group sizedeclined with distance fromthe colony (fig. 4). Flocks ofhomeward-bound Gannets(max. flock size 199) were sig-nificantly larger than flocksleaving the colony (max. 47)in all distance strata within250 km of the Bass Rock (z =2.60–7.49, P<0.01; n = 6,623flocks; fig. 4).Actively searching Gannets

did not operate in cohesiveflocks, but rather dispersedindividually. Circling Gannetsusually triggered an imme-diate response, even fromdistant conspecifics, but only

66

Camphuysen


Fig. 3. Age composition (% adult per 10’-latitude x 20’-longitude rectangle) of Northern GannetsMorus bassanus around Bass Rock, based on June–July surveys 1991–2004. A proportion wascalculated only if n ≥10 per rectangle. Bottom graph: relative abundance of immature (plumage types 1–5) and adult Gannets with increasing distance from Bass Rock (km).

Age composition

95–100% adult90–94% adult75–89% adult50–74% adult25–49% adult<25% adult

0–20 –40 –60 –80 –100 –150 –200 –250 –300 –350 –400

100%80%60%40%20%0%

12345A

when that circling becamemore persistent would those‘neighbouring’ birds comecloser to form a flock. For-aging Gannets, in situationswhere a target (its prey) wasnot visible to the humanobservers, attracted few con-specifics to their searchingarea (mean flock size ± 95%CI = 1.30 ± 1.05), and flocksdid not grow significantlylarger when such individualsbegan to dive (1.65 ± 1.09,max. 112 per flock; fig. 5).Foraging opportunities weretypically very short-lived aftersuch an attack. Gannets par-ticipating in natural MSFAsalso formed relatively smallgroups (1.70 ± 1.46 per flockfor searching individuals, z =1.51, ns; 1.76 ± 1.22 per flockwhen actively feeding, z =0.62, ns; max. flock 35).Gannets were often the lastspecies to arrive and jointhese frenzies, after which thefeeding opportunities sooncollapsed. Significantly largerflocks were observed search -ing over prey-driving dol-phins and porpoises (1.91 ±1.10 per flock searching, z =6.83, P<0.001; 2.96 ± 1.28 perflock feeding, z = 4.51,P<0.001; max. flock 100 indi-viduals) and, more particu-larly, near fishing vessels (1.76± 1.15 per flock searching, z = 4.03, P<0.001;4.63 ± 1.34 per flock feeding, z = 6.77,P<0.001; max. flock 1,100 individuals).

Foraging behaviour and feedingdistributionMost Gannets were recorded foraging andfeeding in areas with relatively low densitiesof conspecifics (fig. 6). From extrapolatednumbers it is clear that within 400 km of thecolony 66.1% of Gannets were feeding inlow-density areas (<1 Gannet per km2; >100km from Bass Rock), 25.8% in areas withmoderate densities (1–5 Gannets per km2;

40–100 km from Bass Rock), and only 8.1%in high-density areas (>5 Gannets per km2;<40 km from Bass Rock). However, whilemost Gannets were observed feeding >100km from the colony, the number of attacks(dives) per km² declined markedly with dis-tance, after a maximum intensity between 20and 40 km from the colony. Nonetheless, for-aging activity expressed as the proportion ofGannets seen was lowest immediately aroundthe colony. The type of feeding behaviour was charac-

teristic for the situation the Gannets were in,as summarised below.



Fig. 4. Flock size (derived mean ± 95% CI) of Northern GannetsMorus bassanus returning to (white) and leaving (grey) the BassRock with increasing distance from the colony (n = 6,623 flocks).

6

5

4

3

2

1

0mean ± 95%

CI

homing inflying out

0– 20– 40– 60– 80– 100– 150– 200– 250– 300– 350–20 40 60 80 100 150 200 250 300 350 400

distance strata around Bass Rock (km)

Fig. 5. Derived mean flock size (± 95% CI) of foraging NorthernGannets Morus bassanus, including searching and actively feedingindividuals in different situations: only with conspecifics (non-ass),associated with cetaceans (cetacean), associated with fishingvessels (trawler) and when joining multi-species foragingassociations (MSFA).

7

6

5

4

3

2

1

0

derived mean flo

ck size ± 95%

CI

non-ass cetacean trawler MSFA non-ass cetacean trawler MSFAsearch search search search feed feed feed feed

Gannets feeding on theirownForaging Gannets away fromother predators or fishingvessels were normally (deep)plunge-diving. In deeper(clear) waters far from thecoast, many of these plunge-dives were from spectacularheights, suggesting that thebirds needed to venture deepinto the water and that fishprey could be observed atconsiderable depth.

Gannets joining MSFAsGannets would join the rela-tively small feeding frenzies ofother seabirds by plunginginto the sea at an angle (thusa relatively shallow plunge-dive), or straight into thefeeding frenzy where auks(Common Guillemots Uriaaalge and Razorbills Alcatorda) had driven small prey,mostly sandeels Ammodytes,to the surface. When the spe-cific role of seabird specieswithin a feeding frenzy couldbe assessed (initiators, joinersor scroungers), Gannets weregenerally seen to join estab-lished feeding frenzies andthey were never recorded as initiators. After joining(usually with a plunge-dive),they would position them-selves on the water surface inthe midst of the feedingopportunity, scooping upsmall fish fry for as long as this was within reach.Scooping (in ‘pelicanfashion’) was an unexpectedtype of behaviour, perhapsfirst recorded with certaintyduring the surveys reportedhere, and this only after yearsof studying MSFAs, but onethat has been seen commonlyever since it was first detected.The presence of Gannets

68

Camphuysen


Fig. 6. Frequency of foraging (% searching for prey or feeding) of all birds observed per 10’-latitude x 20’-longitude rectangle) ofNorthern Gannets Morus bassanus around the Bass Rock, basedon June–July surveys, 1991–2004. A percentage was calculated only if n ≥10 per rectangle.

% foraging

90–100%75–89%50–74%25–49%10–24%1–9%none

28. A feeding frenzy, with the Northern Gannets Morus bassanusbeing joined here by Fulmars Fulmarus glacialis and Great SkuasStercorarius skua, south of Orkney, June 2004.

Kees C

amph

uysen

forced smaller (aerial) seabirds such as Kitti-wakes Rissa tridactyla to position themselvesat the periphery of the feeding frenzy andtypically prompted the auks to give updriving fish towards the surface, so that the

feeding possibility collapsed within seconds.So quick would feeding opportunitiesdevelop and collapse in these MSFAs that,even from a passing ship at speed, much ofthe evolution of individual frenzies could be

recorded, time and time again, includingthe role and effect of each of the partici-pating species (see Camphuysen &Webb 1999 for details). MSFA participa-tion as described here typically occurredwithin 100 km of the Bass Rock, andparticularly within 80 km of the coast(fig. 7). Therefore, the frequency withwhich MSFAs were joined by Gannetsreflected the possibilities at sea (fig. 8).Within that sea area, the proportion ofadult Gannets outside MSFAs amountedto 95.6% (n = 11,093), whereas withinMSFAs the proportion of mature birdswas significantly lower (93.7%, n =1,715; Gadj = 11.67, df = 1, P<0.001).

Gannets associated with cetaceans Gannets joining cetaceans were usuallydeep plunge-diving, but often seemedreluctant to do so in the immediatepresence of the hunting mammals.Cetaceans (certainly dolphins) oftenherded fish prey at speed and itappeared that Gannets needed to posi-tion themselves very carefully to profitmost from the opportunities. On manyoccasions, the birds were merely seen tocircle closer and closer above themammals, trying to keep up with themduring their regular disappearances

deep underwater. During the ship-based

surveys, cetaceans wereencountered 1,825 times (eightdifferent species, 5,565 individ-uals in total). On 286 occasions(15.7%), Gannets were seen tobegin searching for prey oractively feeding in associationwith these mammals (table 2).Gannets were never seen tojoin the (rarely encountered)coastal dolphins (BottlenoseDolphin Tursiops truncatusand Risso’s Dolphin Grampusgriseus), and they seldomexpressed interest in baleen



Fig. 7. MSFAs with (red circles) and without (whitecircles) Northern Gannets Morus bassanus in thenorthwest North Sea around the Bass Rock colony.Total number of MSFAs recorded was 2,080, of which442 (21%) contained Gannets.

1

2–5

6–10

11–25

26+no Gannetsassociated

Fig. 8. Sightings of multi-species feeding associations (MSFAs) –number of sightings per km steamed – and the frequency(percentage) of MSFAs containing Northern Gannets Morusbassanus, with distance from the Bass Rock gannetry.


0.50

0.40

0.30

0.20

0.10

0.00

40

35

30

25

20

15

10

5

0sightin

gs of M

SFAs per km

steam

ed

percentage of M

SFAs with

Gannets associated

MSFAs/km% with Gannets

70

Camphuysen


whales (only five associations with MinkeWhales Balaenoptera acutorostrata, out of 462encounters with this species at sea). The mostabundant cetacean, Harbour Porpoise Pho-coena phocoena, was joined by Gannets in 145of 910 encounters (15.9%), while offshore dol-phins of various species were most frequently‘targeted’ (table 2). The frequency of association with certain

cetacean species was strongly correlated withthe mean pod size of that species (table 2).Furthermore, the mean pod size of the com-monest cetaceans joined by Gannets (White-beaked Dolphin Lagenorhyncus albirostris, 7.5(± 6.1) animals and Harbour Porpoise, 3.4 (±3.1) animals) was significantly larger thanthat of pods of the same cetacean specieswhere Gannets were either not nearby or didnot show an interest (pod size 4.0 ± 3.2, t365 =7.24, P<0.001 and 2.6 ± 3.5, t908 = 2.61,P<0.01, respectively).

Table 2. Sightings of cetaceans (number of sightings, number of animals and mean pod size), and sightings where Northern Gannets Morus bassanus were joining the animals while searchingfor prey or actively feeding (number and frequency (%) of sightings) within 400 km of the BassRock, based on June–July surveys, 1991–2004.

Sightings Animals Pod size Sightings withGannets

Common Dolphin Delphinus delphis 5 138 27.6 3 (60.0%)White-sided Dolphin Lagenorhynchus acutus 26 402 15.5 13 (50.0%)White-beaked Dolphin Lagenorhynchus albirostris 369 1,897 5.1 117 (31.7%)Harbour Porpoise Phocoena phocoena 910 2,514 2.8 145 (15.9%)Unidentified dolphin 36 79 2.2 3 (8.3%)Minke Whale Balaenoptera acutorostrata 462 590 1.3 5 (1.1%)Unidentified whale 11 11 1.0 –Fin Whale Balaenoptera physalus 2 2 1.0 –Bottlenose Dolphin Tursiops truncatus 3 19 6.3 –Risso’s Dolphin Grampus griseus 1 3 3.0 –

1,825 5,655 3.1 286 (15.7%)

Associations with cetaceans occurredmainly well offshore (more than 60 km fromthe colony), and typically over deeper waters(fig. 9). The latter seemed to reflect a genuinepreference for cetaceans in deeper, thermallystratified seawater given that these animalswere widespread and occurred in highnumbers closer to the colony. Cetaceans weremost commonly encountered within 100 kmof the Bass Rock, while records of Gannetsassociating with cetaceans were most fre-quent at greater distances (fig. 10). Offshorefeeding assemblages more than 80 km fromthe coast that involved cetaceans normallyattracted Gannets as the only bird species(90.7%, n = 237). The age composition of allGannets encountered >80 km from the coastand >100 km from the Bass Rock comprised90.5% adults (n = 18,036); in the same dis-tance class, taking solely those associatedwith cetaceans, the proportion of mature

birds was significantly lower(84.9%, n = 2,335; Gadj =62.14, df = 1, P<0.001).

Gannets and fishingvesselsFishing vessels were joinedirregularly and, with decliningfishery effort between 1991and 2003, this associationbecame increasingly rareduring the study. Birds scav-enging at trawlers wouldenter the water with a deep ora shallow plunge-dive and usea variety of techniques to

Fig. 9. Sightings of cetaceans (number of sightings per kmsteamed) and the frequency of association with Northern GannetsMorus bassanus (percentage of pods with associated Gannets), withdistance from the Bass Rock gannetry.


0.20

0.16

0.12

0.08

0.04

0.00

40

35

30

25

20

15

10

5

0sightin

gs of p

ods of cetaceans

per km

steam

ed

percentage of cetacean po

ds with

Gannets associated

sightings/km% with Gannets

obtain discards after that initial dive.Scavenging Gannets were sometimesengaged in surface pecking, swallowingdead fish floating at the surface.However, they often continued withtheir normal strategy of deep plunging.Gannets joining a session of discardingat a fishing vessel would typically divedeep into the water, select and swallowone or two sinking fish in the watercolumn, resurface, take off, circle aroundthe boat, and deep-plunge again, untilsatiated. At the surface (or indeed some-times while still underwater), Gannetswould readily steal fish from competingconspecifics or from other seabirds.Flatfish were often ignored and offal wasseldom taken. Groups of Gannets around commer-

cial fishing vessels could be very large(over 1,000 birds at times), but onlywhen vessels were actively discardingfishery waste. It is important to note,although it was hard to quantify, thatGannets very often ignored the activityof nearby fishing vessels, even whensuitable prey was released in large quan-tities (for example, released from thesurvey vessel itself during periods offishing), to continue on their foragingtrips farther away (see also Camphuysenet al. 1995b). At trawlers in the North Sea, Gannets

were found to be the largest and most powerful scavenging predators(cetaceans excluded), equalled inapparent strength and foraging successonly by Great Black-backed Gulls Larusmarinus. Gannets took much less offal(liver and guts from gutted fish) thanexpected from the numerical abundanceduring discard experiments. In contrast,their feeding success was very high forroundfish, notably gadoids and particu-larly in early summer, and they wereobserved to swallow significantly moreof these fish than expected from theirnumerical abundance during the experi-ments. Experiments showed thatGannets focused more on larger round-fish (22–30 cm long) than any of theother birds joining fishing vessels, whichis consistent with their own overall size



Fig. 10. Pods of small cetaceans with (red circles) and without (white circles) Northern Gannets Morusbassanus in the northwest North Sea around the Bass Rock colony. The total number of pods ofcetaceans observed was 1,825, of which 286 hadassociated Gannets.

1

2–5

6–10

11–25

26+unattendeddolphins

29. A feeding frenzy behind a trawler – note the ‘impact’(the underwater white ‘plume’) that marks the entry of aplunge-diving Northern Gannet Morus bassanus, July 2002.

Kees C

amph

uysen

and swallowing capacity (fig. 11). Gannetsalso preferred smooth fish species (gadoids,clupeids) over spiny species (gurnards). The overall robbery index of Gannets

(17.9 for all categories, 18.3 for roundfish;Camphuysen et al. 1995b) over a total of 536sessions of experimental discarding washigher than in any other seabird species andthe species’ competitive strength was mostpronounced in summer and in the northwestand west of the North Sea basin. The agecomposition of all Gannets encounteredmore than 150 km from the Bass Rock(where all records of birds associating withfishing vessels occurred) comprised 87.7%adults (n = 18,654), whereas for those associ-ated with fishing vessels the proportion ofmature birds was significantly higher (96.7%,n = 1,764; Gadj = 171.4, df = 1, P<0.001).

Gannets disturbing fish schools?The effect of a Gannet attack on a fish shoalcould be observed only in cases of surface-driven fish schools. As MSFA participants, tar-geting surface-driven ‘balls’ of sandeels,Gannets normally attacked the fish fromabove while swimming at the surface (typi-

cally after surfacing from a fruitless plunge-dive). The driving auks normally quicklyabandoned their herding when Gannetsjoined in, and fish ‘balls’ driven to the surfaceby the auks were seen to disintegrate and sinkdeeper within seconds of a Gannet attack.Gannets were seen to reach deeper and deeperinto the water until they gave up feeding, indi-cating that soon after their arrival the herdedfish would disperse into waters too deep forthe birds to have access to from the surface.Late arrivals (Gannets or other seabirds)would, as a rule, have no access to the fish andintake rates for those birds were zero.

DiscussionDuring the fieldwork described here, indi-vidual adult Gannets breeding on the BassRock were tracked with satellite transmitters,GPS, time–depth recorders and other devicesto study the foraging behaviour of breedingbirds and to log individual foraging trips(Hamer et al. 2000; Humphreys et al. 2004).The ship-based surveys showed that Gannetsdispersed widely around the colony, withhigh densities near the colony and low densi-ties over vast areas 100–450 km from the Bass

(fig. 1). The telemetrystudies confirmed that thisentire range and beyond isused by breeding birdsfrom the Bass Rock colony.They revealed that themaximum foraging rangewas 540 km from thecolony, and that the meanfarthest distance from thecolony on any one trip was232 km. Foraging tripslasted between 13 and 84hours. Destinations of for-aging trips mapped withsatellite transmitters andGPS data loggers covered awide area of the North Sea,with a non-random distri-bution and a higher-than-expected proportion oftrips to the northeast (gen-erally in the vicinity ofBuchan Deep and HalibutBank) and to the southeast(mostly between Farne

72

Camphuysen


Fig. 11. Roundfish consumption (% of total discarded) by seabirdsduring experimental discarding from fisheries research vessels in theNorth Sea with Northern Gannet Morus bassanus consumptions inred (from data collected in 1993–94; Camphuysen et al. 1993, 1995b).Bars indicate the frequency discarded for each size class (cm total fishlength), colours represent the fraction successfully swallowed by eachof the main groups of seabirds.

8.0

7.0

6.0

5.0

4.0

3.0

2.0

1.0

0.0

roundfish consum

ption (%

)

3 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46roundfish total length (cm)

not consumedKittiwakelarge gullssmall gullsskuasFulmarGannet

n = 48,443 discarded roundfish

Deep, Dogger Bank and OuterSilver Pit) of the colony. Theauthors of these telemetry studieshave suggested that Gannetsbreeding at the Bass Rock focusedtheir activity on bathymetric fea-tures that were probably associatedwith high primary production.From the at-sea surveys describedin this paper, a preference for desti-nations to the northeast and south-east of the colony is more difficultto detect in the overall distributionmap (fig. 1), but could be inferredfrom the recorded age compositionof Gannets at sea around the colony(fig. 3). However, an association offoraging activities with bathymetricfeatures is not clear from ship-basedsurvey results (fig. 6), and, in fact,foraging is a seemingly randomprocess, certainly so if the type offeeding is ignored.The observations summarised in

this paper show that differentfeeding areas are visited during long-distance trips, requiring differentforaging techniques and probablyyielding different prey. Gannetscommonly deployed three rather differenttypes of feeding in the North Sea: (1) deepplunge-diving, presumably mainly for large,shoaling fish; (2) surface feeding for small fish(scooping up prey); and (3) scavengingbehind fishing vessels. Deep plunging wasobserved in monospecific feeding events, inMSFAs, at trawlers and when associated withcetaceans. Surface feeding (includingscooping) was characteristic within coastalMSFAs and at fishing vessels. Monospecificfeeding bouts were seldom seen to build uplarge congregations of birds, but were never-theless generally short-lived (several minutes).There was no evidence to suggest that

Gannets joining inshore MSFAs (<80 kmfrom the coast) were targeting anything butsmall shoaling fish (notably sandeels), whereasbirds joining cetaceans or trawlers were clearlytargeting larger prey. The inshore MSFAsprobably formed fairly predictable feedingopportunities relatively close to the colony(Camphuysen et al. 2006). The developmentof MSFAs was a characteristic phenomenon in

mixed coastal waters inshore of the shallowsea front (the separation zone between mixedcoastal waters and thermally stratified waters).However, Gannets were encountered both inmixed coastal waters and in deeper, stratifiedwaters farther out. They used different for-aging techniques in the two areas and, in con-trast to the case for many other seabirds, theshallow sea front was relatively unimportantas a feeding area for Gannets. High primaryproduction is associated with oceanic fronts,so the suggestion (from the analysis of loggerdata) that Gannets focused their activity onfeatures that were associated with highprimary production (Hamer et al. 2000) couldnot be confirmed.Prey (fish schools) occurs in discrete

patches, which are probably non-randomlydistributed and difficult to find. In offshorewaters, Gannets often fed on fish schools thatwere herded towards the surface by dolphinsor porpoises. The location of cetacean activityprobably varies unpredictably but is at leasthighly visible at the surface, and could thus be



30 & 31. Northern Gannets Morus bassanus and FulmarsFulmarus glacialis gathering above hunting cetaceans – in this case Killer Whales Orcinus orca in the northern part of the North Sea study area, June 2003.

Kees C

amph

uysen

Kees C

amph

uysen

74

Camphuysen


a good guide for Gannets searching for acces-sible prey patches. Otherwise, the potential torespond to the behaviour of even very distantconspecifics will enhance feeding opportuni-ties at greater distances from land. Activelysearching Gannets did so mainly individually,while simultaneously monitoring the behav-iour of conspecifics, such that when oneforager found a patch, other individuals soonjoined the finder. The signalling function ofthe bright white plumage of Gannets was cer-tainly appreciated by the ship-based observers,who were able to detect Gannets at muchgreater distances than they could all otherspecies of seabirds. The majority of Gannetsforaged in areas with low densities of con-specifics, so that the typically short-livedfeeding bouts attracted relatively few competi-tors. These observations provide furthersupport to the hypothesis formulated by Lewiset al. (2001) that intraspecific competitiondrives birds from large colonies farther afield.Lewis et al. (2001) suggested that schools

of fish attacked by diving Gannets wouldbecome temporarily unavailable as a result ofan avoidance response. The question thatremained was whether evidence could beprovided that Gannets would indeed triggersuch an avoidance response in fish schools,with adverse effects on their own opportuni-ties for profitable feeding. The response of afish school was rarely observed directly, butthe short duration of most feeding eventssuggests that feeding was no longer profitablesoon after an attack by one or just a fewGannets. Moreover, an attack-avoidanceresponse by surface-driven fish within someMSFAs could actually be observed afterGannets entered the scene. Gannet flock sizes were significantly larger

over prey-herding cetaceans (even thoughthis took place mainly in areas with very lowdensities of Gannets) and near trawlers. Ineach of these situations, an attack-avoidanceresponse by prey fish was either blocked(herding mammals) or non-existent (dis-cards), and feeding bouts were apparentlyrelatively longer-lived, attracting more birds.Field observations thus support the predic-tion that feeding frenzies over persistent preypatches grow larger than those over small,easily disturbed fish schools and seem tosupport the assertion that Gannets could be

32 & 33. Plunge-diving Northern Gannet Morus bassanus in Shetland waters, June 2008.

David Tipling

David Tipling

important in inducing an avoidance responseby their prey in certain conditions.Competition for prey is probably most

intense behind fishing vessels, wherenumerous species congregate and large flocksmay be formed. The competitive strength ofscavenging Gannets at trawlers is considerableand the species has been identified as a highlysuccessful kleptoparasite, being capable ofstealing 18 times more items from otherseabirds than were lost by them throughrobbery (Hudson & Furness 1989;Camphuysen et al. 1995b; Garthe & Hüppop1998; this paper). A combination of bruteforce and aggression, the ability to swalloweven the largest discards, plus an ability todive to reach sinking (dead) prey makesGannets highly successful competitors inthese situations. The same avian competitorsmeet one another in MSFAs targeting fishballs driven towards the surface by auks. Here,the same ranking in the dominance hierarchywas established. Smaller, surface-feedingseabirds gave way when Gannets arrived onthe scene. An important difference was thatthe prey facilitation mechanism collapsedwhen Gannets arrived (auks gave up driving)and the attack-avoidance response by the fishprompted them to swim down, beyond reach.The scooping behaviour of Gannets at fishballs appears to be a new discovery and thevisual observations of this has elucidatedsome hitherto unexplained findings fromstudies employingtime–depth activityrecorders and stomach-temperature loggers(Humphreys et al.2004).Where intake rates

are expected to fall withincreasing numbers ofcompetitors (Stephens& Krebs 1986), it doesnot pay to arrive late ina feeding frenzy. Wherean attack would triggeran immediate avoidanceresponse by the prey, itmay be predicted thatonly the first few preda-tors experience a prof-itable feeding bout.

Widespread feeding in a patchy environmentwith very low densities of competitors maythus be a sensible option for an aerial andhighly mobile species such as the Gannet.Feeding individuals can be quickly joinedeven if they are several kilometres away, whilelarge flocks of competitors over feedingpatches are unlikely to develop, simplybecause the birds are too far apart. A feedingbout is thus profitable but short-lived. Recentstudies using instrumentation to track indi-vidual Gannets have supported the sugges-tion of wide (individual) dispersal offoraging birds over large distances at sea.

Ackowledgments

The work was funded by EC projects EC DG XIVresearch contracts 92/3505 and BIOECO/93/10,‘Interactions between the marine environment,predators and prey: implications for sustainable sandeelfisheries (IMPRESS; QRRS 2000-30864)’ and ‘Modelling the impact of fisheries on seabirds (MIFOS;CFP 96-079)’. I would like to thank numerous co-observers and crews of FRV Tridens and RV Pelagiafor help in the field. This work benefited from fruitfuldiscussions with all partners of the IMPRESS andMIFOS projects. Stefan Garthe, Mardik Leopold andTheunis Piersma kindly commented on earlier drafts of this paper.

References

Ashmole, N. P. 1963. The regulation of numbers of tropical oceanic birds. Ibis 103b: 458–473.

— 1971. Sea bird ecology and the marineenvironment. In: Farner, D. S., & King, J. R. (eds.),Avian Biology I, pp. 224–286. Academic Press, New York.

Birkhead, T. R., & Furness, R. W. 1985. Regulation of



34. A multi-species feeding frenzy (MSFA) – Northern Gannets Morusbassanus have entered the scene and are scooping prey (one of the keyfindings in this study), Kittiwakes Rissa tridactyla have been displaced andCommon Guillemots Uria aalge (top), which had been driving prey fishto the surface, swim off away from the scene; July 2002.

Kees C

amph

uysen

76

Camphuysen


seabird populations. In: Sibly, R. M., & Smith, R. H.(eds.), Behavioural Ecology, Ecological Consequences ofAdaptive Behaviour, pp. 147–167. Blackwell Science,London.

Birt, V. L., Birt, T. P., Goulet, D., Cairns, D. K., &Montevecchi, W. A. 1987. Ashmole’s Halo: directevidence for prey depletion by a seabird. Mar. Ecol. Progr. Ser. 40: 205–208.

Boddington, D. 1959. Feeding behaviour of Gannetsand Great Black-backed Gull with Mackerel schools.Brit. Birds 52: 383–384.

Camphuysen, C. J. 2005. Seabirds at sea in summer inthe northwest North Sea. Brit. Birds 98: 2–19.

— & Garthe, S. 2000. Seabirds and commercialfisheries: population trends of piscivorous seabirdsexplained? In: Kaiser, M. J., & Groot, S. J. de (eds.),Effects of Fishing on Non-target Species and Habitats:biological, conservation and socio-economic issues, pp. 163–184. Blackwell Science, Oxford.

— & — 2004. Recording foraging seabirds at sea:standardised recording and coding of foragingbehaviour and multi-species foraging associations.Atlantic Seabirds 6: 1–32.

— & Webb, A. 1999. Multi-species feeding associationsin North Sea seabirds: jointly exploiting a patchyenvironment. Ardea 87: 177–198.

—, Scott, B., & Wanless, S. 2006. Distribution andforaging interactions of seabirds and marinemammals in the North Sea: multi-species foragingassemblages and habitat-specific feeding strategies.In: Boyd, I. J., Wanless, S., & Camphuysen, C. J. (eds.),Top Predators in Marine Ecosystems: monitoringchange in upper trophic levels, pp. 82–97. Cambridge Univ. Press, Cambridge.

—, Heessen, H. J. L., & Winter, C. J. N. 1995a. Distantfeeding and associations with cetaceans of GannetsMorus bassanus from Bass Rock, May 1994. Seabird 17: 36–43.

—, Calvo, B., Durinck, J., Ensor, K., Follestad, A., Furness,R. W., Garthe, S., Leaper, G., Skov, H., Tasker, M. L., &Winter, C. J. N. 1995b. Consumption of Discards bySeabirds in the North Sea. NIOZ-Report 1995–5.Netherlands Institute for Sea Research, Texel.

—, Fox, A. D., Leopold, M. F., & Petersen, I. K. 2004.Towards Standardised Seabirds at Sea CensusTechniques in Connection with Environmental ImpactAssessments for Offshore Wind Farms in the UK.Royal Netherlands Institute for Sea Research, Texel.

—, Ensor, K., Furness, R. W., Garthe, S., Hüppop, O.,Leaper, G., Offringa, H., & Tasker, M. L. 1993. SeabirdsFeeding on Discards in Winter in the North Sea.NIOZ-Report 1993–8, Netherlands Institute forSea Research, Texel.

Croxall, J. P. (ed.) 1987. Seabirds: feeding ecology and rolein marine ecosystems. Cambridge University Press,Cambridge.

Evans, P. G. H. 1982. Associations between seabirdsand cetaceans: a review. Mamm. Rev. 12: 187–206.

Fisher, J., & Vevers, H. G. 1951. The present populationof the North Atlantic Gannet (Sula bassana).

Proc. Int. Orn. Congr. 10: 463–467. Furness, R. W. 1982. Competition between fisheries

and seabird communities. Adv. Mar. Biol. 20: 225–307. Garthe, S., & Hüppop, O. 1998. Foraging success,

kleptoparasitism and feeding techniques inscavenging seabirds: does crime pay? Helgol. Meeresunters. 52: 187–196.

Hamer, K., Phillips, R., Wanless, S., Harris, M. P., & Wood,A. 2000. Foraging ranges, diets and feeding locationsof Gannets Morus bassanus in the North Sea:evidence from satellite telemetry. Mar. Ecol. Progr.Ser. 200: 257–264.

Hudson, A. V., & Furness, R. W. 1989. The behaviour ofseabirds foraging at fishing boats around Shetland.Ibis 131: 225–237.

Humphreys, E. M., Hennicke, J. C., Hamer, K. C., Garthe,S., Peters, G., Grémillet, D., & Wanless, S. 2004.Foraging behaviour of Northern Gannets: newinsights from remote-sensing. Oral presentationManagement of Marine Ecosystems: monitoringchange in upper trophic levels. Zoological Society ofLondon Conference, April 2004.

Lewis, S., Sherratt, T. N., Hamer, K. C., & Wanless, S.2001. Evidence of intra-specific competition forfood in a pelagic seabird. Nature 412: 816–819.

Nelson, J. B. 1978. The Gannet. Poyser, Berkhamsted.— 2002. The Atlantic Gannet. Fenix Books, Great

Yarmouth.Reinsch, H. H. 1969. Der Basstölpel. Wittenberg

Lutherstadt: Neue Brehm-Bücherei 412. Sherman, K., Jones, C., Sullivan, L., Smith, W., Berrien, P.,

& Ejsymont, L. 1981. Congruent shifts in sand eelabundance in western and eastern North Atlanticecosystems. Nature 291: 487–489.

Skov, H., Durinck, J., Leopold, M. F., & Tasker, M. L. 1995.Important Bird Areas for Seabirds in the North Sea,including the Channel and the Kattegat. BirdLifeInternational, Cambridge.

Stephens, D. W., & Krebs, J. R. 1986. Foraging Theory.Princeton University Press, Princeton.

Stone, C. J., Webb, A., Barton, C., Ratcliffe, N., Reed, T. C., Tasker, M. L., Camphuysen, C. J., & Pienkowski,M. W. 1995. An Atlas of Seabird Distribution in North-west European Waters. JNCC, Peterborough.

Tasker, M. L., Jones, P. H., Dixon, T. J., & Blake, B. F. 1984.Counting seabirds at sea from ships: a review ofmethods employed and a suggestion for astandardised approach. Auk 101: 567–577.

—, —, Blake, B. F., & Dixon, T. J. 1985. The marinedistribution of the Gannet in the North Sea. Bird Study 32: 82–90.

Wanless, S. 1987. A Survey of the Numbers andBreeding Distribution of the North Atlantic GannetSula bassana and an Assessment of the ChangesWhich Have Occurred Since Operation Seafarer1969/70. NCC, Peterborough.

—, Murray, S., & Harris, M. P. 1986. Gannets – a boomin numbers and distribution. BTO News 145: 8.

—, — & — 2005. The status of Northern Gannet inBritain & Ireland in 2003/04. Brit. Birds 98: 280–294.

Kees (C.J.) Camphuysen, Royal Netherlands Institute for Sea Research (NIOZ), PO Box 59, 1790 AB Den Burg, Texel, The Netherlands; e-mail [email protected]

Kees Camphuysen is co-founder of the Dutch Seabird Group and has worked as an ornithologist at the RoyalNetherlands Institute for Sea Research (NIOZ) since 1992. Since the late 1980s he has participated innumerous at-sea surveys during which ship-based seabird censusing techniques were refined to includebehavioural data.

northern gannets in the north sea: foraging distribution

Documents