8 0 5Revista Chilena de Historia Natural

75: 805-818, 2002

The decline of Falkland Islands penguins in the presence of acommercial fishing industry

La disminución de los pingüinos de las Islas Falklands en la presencia de actividadesde pesca comercial

MIKE BINGHAM

Environmental Research Unit, P.O. Box 434, Stanley, Falkland Islands;e-mail: conservation@falklands.net

ABSTRACT

The Falkland Islands are an important breeding site for three species of penguin, gentoo (Pygoscelis papua), southernrockhopper (Eudyptes c. chrysocome) and Magellanic penguin (Spheniscus magellanicus). The total penguin populationfor the Falkland Islands has declined by 84 % during the 1980s and 1990s. These declines did not occur in coastal SouthAmerica, so potential causes of decline in the Falklands have been investigated. The suspected cause of decline is areduction of fish and squid due to large-scale commercial fishing around the Falklands. Since 1995 rockhopper andgentoo populations have ceased declining, and appear to have reached a new equilibrium, albeit at a much lower levelthan before commercial fishing began. This has been matched by improved chick-rearing success and juvenile survival,however Magellanic penguins continue declining in the Falklands. Diet analysis shows that Magellanic penguins havea greater reliance on squid and fish species being taken commercially. In 1998 drilling for oil began around theFalklands, despite warnings that environmental protection was inadequate. Within a month the first of three separateoil spills occurred, killing and contaminating hundreds of penguins. The oil rig completed its drilling operations afterfive months and left the Falklands. Since then no further oil spills have occurred. Oil exploration is due to recommencein the near future, and environmental safeguards have not been improved. Ecotourism has increased rapidly over recentyears in the Falklands, with penguins being the main attraction. Monitoring of the affects of tourism has concentratedon breeding success and population change, and the results indicate no detrimental affects on penguin populations atthe current level. This paper investigates potential causes of penguin decline in the Falkland Islands, drawingcomparison with populations in Chile which appear to be healthy. It concludes by calling on the Falkland IslandsGovernment to exclude large-scale commercial fishing close to penguin breeding sites.

Key words: penguins, Falkland Islands, Falklands.

RESUMEN

Las islas Falklands son un lugar importante para tres especies de pingüinos, pingüino papua (Pygoscelis papua),pingüino de penacho amarillo (Eudyptes c.chrysocome) y pingüino de Magallanes (Spheniscus magallanicus).Recientemente estas especies han disminuido 84 % en estas islas. En la costa de Sudamérica los pingüinos nodisminuyeron. Se sospecha que la causa es una reducción de peces y calamares debido a los barcos de pesca comercialque operan en aguas de las Islas Falklands. En 1995 el pingüino papua y el pingüino penacho amarillo terminaron susdisminuciones y ya parece que sus poblaciones están en equilibrio, pero en un número mucho mas bajo que antes quelos pescadores comenzaron 20 años atrás. El pingüino de Magallanes todavía disminuye en las Islas Falklands. Elpingüino de Magallanes depende más de especies de calamares y peces capturados por barcos de pesca comercial.Avisos de protección de la fauna no fueron suficientes para impedir que 1998 comenzaras exploraciones petroleras enlas Islas Falklands. Tres derrames de petróleo ocurrieron en cinco meses, y cientos de pingüinos murieron. La torre deperforación se fue después de cinco meses, y no ocurrieron más derrames de petróleo. Nuevamente van a comenzar abuscar petróleo, sin mejorar la protección para la fauna. El turismo ha crecido rápidamente en las Islas Falklands, y lamayoría de los turistas llegan para ver los pingüinos. Investigaciones de poblaciones y éxito reproductivo indican queaún no hay efectos perjudiciales para los pingüinos por esta actividad. En este trabajo se investigan las causaspotenciales de la disminución de pingüinos en las Islas Falklands y se hacen compariciones con poblaciones en Chileque parecen saludables.

Palabras clave: pingüinos, Islas Falklands, Malvinas.

8 0 6 BINGHAM

INTRODUCTION

The Falkland Islands lie in the south-west Atlantic,450 km north east of the southern tip of SouthAmerica. The archipelago is made up of over 700hundred islands, comprising a total land area ofover 12,000 km2. The irregular shape and largenumber of islands, gives the Falklands a very longcoastline in relation to its land area, providing awide variety of coastal habitats (Fig. 1). Thisvaried habitat, combined with the productivewaters of the Patagonian shelf, make the Falklandsa good place for seabird reproduction and feeding,especially for albatross, cormorants and penguins.

Five species of penguin breed in the FalklandIslands: king penguin (Aptenodytes patagonicus,Miller 1778), gentoo penguin (Pygoscelis papua,Forster 1781), southern rockhopper penguin(Eudyptes c. chrysocome, Forster 1781), macaronipenguin (Eudyptes chrysolophus, Brandt 1837),and Magel lan ic penguin (Spheniscusmagellanicus, Forster 1781). During the 1980sand early 1990s populations of gentoo, southernrockhopper and Magellanic penguins declineddramat ica l ly in the Falk lands. A lack ofcomparative data made it impossible to determinewhether such declines were part of a regionaltrend, or whether they were due to circumstancespertaining to the Falklands. There was also a lackof basic research data with which to determine theextent to which human activi t ies (such ascommercial fishing, tourism, oil exploration andremoval of eggs for human consumption) were

contributing to the decline. Since the late 1980s,penguin populations around the Falklands havebeen studied in order to address these issues. Thispaper looks at the evidence of population declinein the Falklands, and investigates the role ofhuman activities, using comparative data fromSouth America.

MATERIAL AND METHODS

In the austral summer of 1995-1996, an island-wide penguin census of the Falkland Islands wasconducted. All species were counted except forthe Magellanic penguin, which was not includedbecause of the difficulties of conducting a censuson a species that nests in burrows.

For gentoo and rockhopper penguins, nest countswere made to determine the number of breedingpairs. Counts were timed to correspond with theend of the egg laying period, thereby ensuringthat few pairs were still to lay, and allowing anassessment to be made of the underestimate dueto pairs failing, by using failure rates duringincubation from other studies. Gentoo penguinsconcluded their first egg-laying by the end ofOctober 1995. The 1995/96 census counted 15 %of the Gentoo population between 15th and 31stOctober, and the remainder between 1st Novemberand 1st December. Because Gentoos failing earlytend to re-lay, and failure rates during incubationare low (ca. 1 % per week), the magnitude of anyunderestimates resulting from differences in

Fig. 1: Location of the Falkland Islands in relation to South America.Ubicación de las Islas Falkland en relación a Sudamérica.

8 0 7FALKLAND ISLANDS PENGUINS

survey dates should be wel l be low 5 %.Rockhopper penguins are much more synchronousin terms of egg-laying than Gentoo penguins.Laying was concluded by mid-November 1995,and the 1995-1996 census counted 98 % of therockhopper population between 1st Novemberand 1st December (2 % between 2nd and 18thDecember). Repeated counts of rockhoppercolonies in previous years showed that nest countsdrop at a rate of about 3 % per week for the firstmonth after egg-laying, as a result of failed nests.I t is therefore un l ike ly that the averageunderestimate of rockhopper population exceeded10 %.

For most rockhopper and all gentoo breedingsites, the recorder made two separate counts of alloccupied nests using a tally counter. The mean ofthe two counts was taken as the number of breedingpairs. Where these counts differed by more than10 %, a third count was taken to give a mean ofthree counts. In practice this was rarely necessary,and the spread of results was usually well withinplus or minus 5 %. However, for the very largerockhopper colonies on Steeple Jason, GrandJason, Bird Island and Beauchene Island, directground counts were not possible. These sites werecounted using a total of 60 randomly selectedsample plots to determine the range of nestingdensities, and the areas of the colonies weredetermined to enable estimates of total breedingpairs. A minimum of 10 % and a maximum of 15% of the total colony area was sampled at each ofthe sites. These measurements of area and densitytaken during the site visits were later comparedagainst aerial photographs taken of the colonies.The margin of error for this methodology is greaterthan for direct counts, but should be within plusor minus 10 %.

The breeding cycle of the king penguin isdifferent from that of gentoos and rockhoppers,with chicks over-wintering at the colony, and acomplete breeding cycle lasting over a year. Thistends to result in individual birds having theirfollowing breeding cycle out of phase with itspredecessor; thus large chicks and eggs both occurin a colony at the same time. This complicatesassessment of breeding pairs, so chick countswere taken instead. The estimation of error forchick counts is well below 5 %, but will underestimate the number of breeding pairs by about 20% (Lewis-Smith & Tallowin 1979).

Between 1989 and 2002, a total of 40 breedingsites (one king, 21 gentoo, eight rockhopper and10 Magellanic) were counted annually in theFalklands to determine changes in populationsize. For king, gentoo and rockhopper penguinsthese counts were conducted as described above.

Because Magellanic penguins nest in burrows inthe Falklands, nest counts were conducted byexamining each burrow within the colony forsigns of nest occupation and breeding activity.For difficult burrows the use of a video camera ona long pole was employed to determine thepresence of an active nest. Each burrow occupiedby a breeding pair was marked with a small spotof paint during the count to prevent double-counting.

Annual chick counts were also made at each ofthe study sites prior to fledging, in order to deter-mine annual breeding success. Breeding successwas determined by conducting nest counts withinthe colony at the onset of incubation, as describedabove. Each colony was then revisited just priorto fledging, and all chicks within the colony werecounted. Two counts were made using a tallycounter, with a third count being made where thefirst two results differed by more than 10 %.Breeding success was taken as the mean numberof chicks recorded within the colony divided bythe mean number of occupied nests recorded atthe start of the season (chicks per breeding pair).

For Magellanic penguins around 250 occupiedburrows were also marked with names or numbers,and visited an average of three times per weekfrom October to March, in order to record egg-chick progress and to determine the causes ofegg-chick loss. Eggs that failed to hatch werelater examined to determine the stage of embryodevelopment, and dead chicks were removed fromthe burrows for measuring and weighing, and toassess the causes of death where possible.Fledglings were weighed prior to departure fromthe colony, by suspending birds from a springbalance using a soft cotton loop around the legs.Some adults were marked in their burrows usingred and blue animal marker crayons on long poles,so that each partner could be identified duringhourly observations. These markings were placedon the throat where they were easily visible duringnest inspections, and where they could not beeas i ly removed dur ing preening. Hour lyobservations recorded time spent in and awayfrom the burrow for each partner during incubationand chick rearing.

In addition to monitoring overall populationtrends and breeding success, comparisons weremade to determine the effects of human activityin the Falklands. Population trends and breedingsuccess were compared for colonies that areactively promoted by the Falkland Islands TouristBoard as official tourist destinations, and forcolonies that are not visited by any tourists, inorder to investigate the impacts of tourism. TheFalkland Islands Government also permits the

8 0 8

removal of gentoo penguin eggs for humanconsumption. Unlike most other penguins, gentooswill readily re-lay after loosing their clutch,leading to claims by farmers that the removal ofeggs does not harm the species. Comparisons ofpopulation trends and breeding success were madefor colonies where eggs are harvested for humanconsumption, and for colonies from which noeggs are taken, in order to evaluate the impact ofegging.

Each year diet samples were taken of gentoo,rockhopper and Magellanic penguins using thestomach-flushing technique described in Wilson(1984). Samples were taken during incubationand chick-rearing, from adults returning to thecolony after foraging. Samples sizes varied fromyear to year, but average around 25 individualsper year for each species.

Stomach samples were drained and stored injars with formaline solution or alcohol, ready forlater examination. Once in the laboratory thestomach samples were rinsed with water, drainedto remove any excess liquid, and weighed todetermine the wet weight of food retrieved. Eachsample was then divided up into its appropriatecomponents, which were weighed individually todetermine proportional dietary composition bywet weight. Fish otoliths, cephalopod beaks andcrustacean carapaces (which are not easilydigested) were used to aid species identification,and to estimate proportional composition. Thesedata were then compared with fisheries catchstatistics in order to determine the level of overlapbetween penguin diet and commercial fishingactivities (Falkland Islands Government 2001).From 1998 diet sample analysis in the Falklandswas taken over by Dr. A. Clausen of FalklandsConservation (Clausen 2000), although themethodology remained unchanged (Wilson 1984).

During the austral summer of 1996-1997, apenguin census was conducted in South America,in order to determine whether penguin declines inthe Falklands had occurred elsewhere. It had beenshown during the 1995-1996 census of theFalkland Islands that it requires little extra effortto count all penguin species during the course ofsuch a census. The only exception was theMagellanic penguin. Its widespread, low-densitydistribution in burrows made it impossible tocensus with methods employed for surface nestingspecies. For this reason it was decided that the1996-1997 census would include al l SouthAmerican penguins, except for those of the genusSpheniscus.

During the 1995-1996 Falkland Islands censusit had been possible to conduct ground counts ofincubating pairs at each of the breeding colonies,

because most colonies were relatively accessible.By contrast, many of the South American coloniesare remote and inaccessible, and any attempt toconduct ground counts of each and every colonywould have been doomed to failure. It wastherefore decided from the outset that the censuswould be conducted by light aircraft, therebynegating the need to get ashore at difficult andremote sites.

The location of all the Falkland Islands breedingsites had been known prior to the commencementof the 1995-1996 census, but this was certainlynot the case for South America. Although data didexist for a number of known breeding sites aroundSouth America (Venegas 1984, 1991, Frere et al.1993, Woehler 1993), it was likely that other sitesexisted which had not been previously recorded.This was another reason for favouring an aerialcensus, since it provided the opportunity to coverlarge areas of suitable coastline in search ofpreviously unrecorded colonies. This reduced themargin of error that would otherwise have arisenfrom new sites being overlooked, however themargin of error for aerial counts was higher thanfor ground counts.

In order to quantify the margin of error likely to beexpected from aerial counts, a number of aerialcensuses were made of rockhopper colonies in theFalkland Islands for which the number of breedingpairs was also determined by ground counts. Theseaerial counts differed by a maximum of 14 % fromground counts made of the same colony, giving atotal margin of error of ± 20 % for aerial census data.

The 1996-1997 aerial census was conductedthroughout the known Eudyptes breeding rangesof Chile and Tierra del Fuego. The Atlantic coastof mainland Argentina was excluded from thecensus, since this coastline has been well studied,and does not hold any breeding sites for speciescovered by the census, except for a very smallrockhopper colony on Isla Pingüino, near PuertoDeseado (Frere et al. 1993). This colony isregularly monitored as part of an ongoing researchprogramme, and population data from theirresearch was used in favour of duplicating results.

Since 1997 the annual monitoring of Magellanicpenguins has been extended to include colonies insouthern Chile, in order to compare annualpopulation trends, breeding success, foragingbehaviour and diet composition between theFalkland Islands and Chile.

RESULTS

The 1995-1996 Falkland Islands census recorded65,000 breeding pairs of gentoo penguins at 81

BINGHAM

8 0 9

breeding sites (Fig. 2). This represents a declineof around 45 % since the early 1980s (Croxall eta l . 1984) . A repeat census by Fa lk landsConservation in 2000-2001 (Clausen 2001) showsa population recovery to around 113,000 breedingpairs, equivalent to about 35 % of the worldpopulation, last estimated at 320,000 breedingpairs (Woehler 1993).

The 1995-1996 Falkland Islands census recorded297,000 breeding pairs of southern rockhopper at36 breeding sites (Fig. 2). This represents an 88 %decline since the early 1980s (Croxall et al. 1984).A repeat census by Falklands Conservation in2000-2001 (Clausen 2001) recorded a populationof 272,000 breeding pairs. The current Falklandspopulation represents about 60 % of the worldpopulation, with the remaining 40 % located at 15breeding sites in Chile and Argentina (Bingham& Mejias 1999).

Although an island-wide census has never beenconducted for Magellanic penguins, annual countssince 1989-1990 indicate that Magellanic penguinshave undergone a 76 % decline in the Falklandsbetween 1989-1990 and 2001-2002 (Fig. 2). Nodata exists prior to 1989-1990, but since gentooand rockhopper penguins underwent their greatestdeclines during the 1980s, it is likely that theoverall decline of Magellanic penguins is muchgreater than the 76 % recorded. Unlike gentooand rockhopper penguins, Magellanic penguinsare still declining.

Gentoo penguins have averaged 0.84 chicks perbreeding pair since studies began in 1990-1991(SD = 0.21, n = 12). Prior to 1993-1994 theaverage was 0.73 chicks per breeding pair, butbetween 1993-1994 and 1999-2000 the averageincreased to 0.99 chicks per breeding pair (Fig.3). A Mann-Whitney U-test showed this differenceto be significant at the 5 % level. Gentoopopulations stopped declining around 1993-1994,and by 2000-2001 they had recovered to their pre-fisheries level of around 115,000 breeding pairs(Bennett 1933, Croxall et al. 1984). Since thenbreeding success has slumped to an average ofonly 0.59 chicks per breeding pair, with the lasttwo seasons data showing the lowest breedingsuccess ever recorded.

Magellanic penguins have averaged 0.71 chicksper breeding pair since recording began in 1989-1990 (SD = 0.25, n = 13). Prior to 1992-1993 theaverage was 0.43 chicks per breeding pair, butbetween 1993-1994 and 1999-2000 the averageincreased to 0.92 chicks per breeding pair (Fig.3). A Mann-Whitney U-test showed this differenceto be significant at the 5 % level. Despite thisimprovement in breeding success, Magellanicpenguins have continued to decline in the Falkland

Islands, and over the last two seasons breedingsuccess has averaged just 0.53 chicks per breedingpair, the lowest level since 1992-1993.

Magdalena Island and Seno Otway are theclosest major Magellanic penguin colonies to theFalkland Islands with comparable breedingconditions (ie. nesting occurring in burrows asper the Falklands). Monitoring began at these twosites in 1996-1997 to provide comparison withthe Falklands. Magdalena Island and Seno Otwayare situated in areas where large scale commercialfishing does not occur. Commercial fishing didoccur around Magdalena Island until a few yearsago, since when fishing has been banned in orderto protect penguin populations (Radl & Culik1998). These colonies have shown increases inMagellanic penguin population since studiesbegan in 1996-1997, using identical methodologyto that which has shown a population decease inthe Falklands.

Breeding success for Magellanic penguins hasaveraged 0.71 chicks per breeding pair in theFalklands (SD = 0.25, n = 13) (Fig. 3), whilst thetwo Chilean sites have averaged 1.40 chicks perbreeding pair (SD = 0.08, n = 10) (Fig. 4). AMann-Whitney U-test showed these differencesto be significant at the 5 % level, with the lowestvalue for Chile being higher than the highestvalue for the Falklands.

Comparison shows that lower breeding successin the Falklands results from a two fold increasein nest abandonment during the incubation phase,and a two and a half fold increase in chick mortalityafter successful hatching (Fig. 5). This increasein chick mortality in the Falklands was due toincreased levels of starvation and malnutrition.Chick mortality during hatching showed nodifference between Chile and the Falklands.

Fig. 2: Falkland Islands penguins populationtrends.Tendencias poblacionales en los pingüinos de las IslasFalkland.

FALKLAND ISLANDS PENGUINS

8 1 0

Fledgling weights averaged 3.3 kg in Chile, butonly 2.7 kg in the Falklands, differing significantlyat the 5 % level using a Mann-Whitney U-test.Chicks in the Falklands fledge around 10 dayslater than in Chile. Mean foraging duration duringchick rearing averaged 33.9 h in the Falklands,

and 13.5 h in Chile, differing significantly at the0.05 level using a Mann-Whitney U-test.

Since recording began in 1993-1994, rockhopperbreeding success has remained within the rangeof 0.63 to 0.80 chicks per breeding pair, with anoverall average of 0.73 chicks per breeding pair

Fig. 3: Falkland Islands penguin breeding success.Éxito reproductivo en los pingüinos de las Islas Falkland.

Fig. 4: Comparison of Magellanic penguin breeding success in the Falklands and Chile.Comparación del éxito reproductivo del pingüino magellánico en las Malvinas y en Chile.

BINGHAM

8 1 1

(SD = 0.05, n = 9) (Fig. 3). A lack of data makesit impossible to determine whether rockhopperbreeding success was lower prior to 1993, duringtheir rapid population decline. However prior to1995 virtually no juveniles were returning toundertake their annual moult, suggesting lowoverall recruitment.

In addition to the three main penguin species,the Falkland Islands have small numbers of kingand macaroni penguins. The 1995-1996 penguincensus recorded 339 king penguin chicks, whichallowing for losses during incubation and chick-rearing, and the staggered breeding cycle, givesan estimated Falklands population of around 400breeding pairs. This is a tiny proportion of theestimated 1,500,000 breeding pairs world-wide(Woehler 1993). The 2000-2001 census recorded275 chicks (Clausen 2001), a reduction of 19 %since 1995-1996.

There are no macaroni penguin colonies in theFalklands, but a few macaroni penguins breed inrockhopper colonies around the islands. The totalpopulation of macaroni penguins is estimated tobe no more than 50 breeding pairs (Bingham &Mejias 1999). Two very small breeding coloniesof gentoo penguins were found in South America,on Staten Island (about 30 breeding pairs) and onHammer Island, near Ushauia (about five breedingpairs). These are the first breeding gentoos

recorded for South America, but other smallcolonies probably await discovery.

Studies of breeding success in the Falklandsshowed no harmful effects from tourism or theremoval of gentoo eggs for human consumption(Fig. 6). However, it should be noted that onlygentoo eggs can be legally taken in the Falklands,as they are the only Falklands penguin that can re-lay after loosing the first clutch of eggs.

Diet sample analysis shows that gentoo,rockhopper and Magellanic penguins all rely onspecies of fish and squid that are currently takencommercially by the Falkland Islands’ fishingindustry, especially loligo squid (Loligo gahi,d’Orbigny 1835) and blue whiting (Micromesistiusaustral is, Norman 1937) (Falkland IslandsGovernment 1989, 2001). These species make upa small proportion of the diet of gentoos (5.9 % ofobserved diet) and rockhoppers (10.2 % ofobserved diet), but 26.5 % of the observed diet ofMagellanic penguins (Table 1). Magellanicpenguins continue to decline, while gentoo androckhopper penguins appear to have stoppeddeclining.

For all three penguin species, the observed levelof competition with commercial fisheries will bean under-estimate. If there were no commercialfishing activity the abundance of loligo squid andblue whiting would be considerably higher. Since

Fig. 5: Analysis of breeding failure, chick survival and fledging weights.Análisis de fallas reproductivas, sobrevivencia de polluelos y peso de los juveniles.

FALKLAND ISLANDS PENGUINS

8 1 2

penguin diet in the Falklands has only ever beenstudied under conditions of reduced abundance ofcommercially harvested species due to commercialfishing, the importance of such species to penguinsunder natural conditions will be greatly under-estimated.

Further analysis shows that as the decline ofgentoo and rockhopper penguins bottomed out inthe mid 1990s, a gradual change of diet occurred,with less loligo squid being taken. Magellanicpenguins have also shown a change in diet awayfrom lol igo squid, but Magellanic penguinpopulations have continued to decline despitethis change.

DISCUSSION

The Falkland Islands is an important breedingsi te for gentoo, southern rockhopper andMagellanic penguins, but over the last 20 yearsall three have undergone population declines inthe Falklands. Despite a recent recovery in thegentoo penguin populat ion, to ta l penguinpopulations in the Falklands now total just 16 %of that estimated 20 years ago. This decline coin-cides with the development of a large scalecommercial fishing industry around the FalklandIslands over the same time scale.

Southern rockhopper and Magellanic penguinsare only found in the Falkland Islands and southernSouth America, but population declines appear tohave occurred only in the Falkland Islands. Censusand monitoring work in Chile indicates thatsouthern rockhopper and Magellanic penguinpopulations in Chile are stable, despite their closeproximity to the Falklands.

Rockhopper penguins do not change the locationof their breeding sites, so prolonged occupationkills off grasses and other vegetation, leavingbare ground and lichen covered rock. In theFalkland Islands most rockhopper colonies arenow found amidst much larger areas of bareground, where vegetation has been destroyed bycolonies that were once much larger. They appearrather like ponds that have dried out to leave asmall puddle at the centre. These visual signs oflarge-scale decline are supported by census data,which shows the current rockhopper populationin the Falklands to be just 11 % of that recorded20 years ago (Fig. 2).

By contrast most colonies in South Americaoccupy the entire area laid bare, and new nestingareas can be found where vegetation has not yetbeen destroyed. On Staten Island (Isla de LosEstados) colonies have expanded at such a ratethat large numbers of rockhoppers are nesting indense grass which has not yet been killed off by

Fig. 6: Effects on breeding success of tourism and egging 1993-1994 and 2001-2002.Efecto del turismo y la recolección de huevos sobre el éxito reproductivo 1993-1994 y 2001-2002.

BINGHAM

8 1 3

nesting activity. These visual signs suggest thatthe South American population has remainedstable, or in the case of Staten Island, undergonea rapid increase. It is possible that some emigrationhas occurred from the Falkland Islands to sites oncoastal South America, such as Staten Island.

There is no data on rockhopper breeding successprior to 1993, since when populations in theFalklands appear to have stopped declining, andreached equilibrium. There is however strongobservational evidence that recruitment was verylow in the Falklands prior to 1995. For bothrockhopper and Magellanic penguins, juvenilesand non-breeders return to moult at their breedingsite during January and February, a time whencolonies are being monitored to conduct chickcounts. Virtually no rockhopper or Magellanicpenguin juveniles were observed around thebreeding colonies during studies prior to 1995,suggesting very low juvenile survival, despite thefact that colonies had been observed producingfledglings during previous seasons. Since 1995an increase in juveniles has been observed forrockhopper penguins, but not for Magellanicpenguins. Colonies monitored in Chile have shownhealthy numbers of juveniles each year.

Breeding success for Magellanic penguins hasbeen recorded annually since 1989 in the FalklandIslands where large-scale commercial fishingoccurs, and since 1996 for two sites in southernChile, where large-scale commercial fishing isprohibited in order to protect penguins. Breedingsuccess in the Falkland Islands has averaged 0.71chicks per breeding pair (Fig. 3), whilst the ave-rage for the two Chilean colonies has been 1.40chicks per breeding pair (Fig. 4). Large variationsin breeding success are usually associated withchanges in food availability (Boersma et al. 1990),and for the Falklands the main factor influencingbreeding success and recruitment is assumed tobe local food supply (Putz et al. 2001).

Low breeding success in the Falklands resultsfrom greater nest abandonment during eggincubation, and higher chick mortality resultingfrom starvat ion and malnutr i t ion (Fig. 5).Surviving chicks in the Falkland Islands are ofpoor body condition and low weight at the time offledging (average 2.7 kg) compared to Chile (ave-rage 3.3 kg). This suggests that not only do fewerchicks fledge in the Falklands, but that fledglingsalso have a lower chance of surviving their firstyear by virtue of being less well nourished.Comparison of juvenile numbers returning to theirnatal colony to moult support this theory. OnMagdalena Island over four thousand juvenilesare counted on the beach each year, whilst in theFalkland Islands very few juveniles are ever seen.

Gentoo, rockhopper and Magellanic penguinsall compete directly for loligo squid (Loligo gahi)with the Falkland Islands fishing industry (Putzet al. 2001). These penguins also take blue whiting(Micromesistius australis) which is the main fishspecies targeted by the Falkland Islands fishingindustry. Dietary overlap for commercially takenspecies is greatest for Magellanic penguins whichare still declining rapidly, less for rockhopperpenguins which have levelled off at around 11 %of their pre commercial fishing population, andlowest for gentoo penguins which have recoveredfollowing an initial decline.

Over the last 10 years, the Falklands haveexperienced dietary changes in all three penguinsaway from Loligo gahi. Putz (2001) suggests thatthese dietary changes have been forced by reducedabundance of Loligo gahi. These changes may beharmful in terms of chick survival, since lobsterkrill, ( Munida gregaria, Fabricius 1793), whichnow makes up one fifth of Magellanic penguindiet in the Falklands, is not easily digested byMagellanic penguin chicks (Thompson 1993). Thetheory that Magellanic penguins prefer not totake Munida gregaria if more suitable prey areavailable, is supported by diet sample studies inChile. Magellanic penguins at the two Chileanstudy sites show a complete lack of Munida gre-garia in their diet, even though Munida gregariaare present in great abundance, forming a majorpart of the diet of king cormorants nesting onnearby Marta Island (Radl & Culik 1998).

Not only does the removal of preferred preyforce penguins to feed their chicks less favourablespecies, but it can also lead to greater foragingduration, with chicks receiving less food (Radl &Culik 1998). This increase in foraging duration isevident around the Falklands.

At the two Chilean study sites, where large-scale commercial fishing is no longer permitted,Magellanic penguins average 13.5 h to returnwith food for their chicks (Radl & Culik 1998). Inthe Falkland Islands, foraging trip durationaverages 33.9 h. This huge increase in foragingduration coincides with a 50 % reduction inbreeding success, higher chick starvation, lowerfledging weight, and a substantial reduction injuvenile survival, compared to populations inChile.

Reproduct ive success depends on foodavai labi l i ty (Crawford & Dyer 1995), andcompetition for food with commercial fisheries isa recognised cause of population decline in otherregions (Brown & Nettleship 1984, Duffy et al.1987, Culik & Luna Jorquera 1997). Breedingsuccess and fledging weights are recognisedbio ind icators for moni tor ing the mar ine

FALKLAND ISLANDS PENGUINS

8 1 4

environment and marine food supplies (Cairns1987, Furness & Camphuysen 1997), andMagellanic penguins suggest that the marineenvironment around the Falkland Islands may bein poor health.

Prior to 1988 fishing around the Falkland Islandswas intensive and totally unregulated, threateningfish and squid stocks (Patterson 1987). No catchdata or diet sample data exists for this period,during which huge penguin population declinesoccurred. Following a mass starvation of penguinsin 1986, when many rockhopper colonies lostover half their adult population (Keymer et al.2001), it was agreed that commercial fishingaround the Falklands needed to be regulated. Thisregulation was introduced in 1988, since whencatch rates have gradually been reduced bycontrolling the number of boats licensed to fish(Falkland Islands Government 1989, 2001).

For squid, the catch rate per unit effort of fishingvessels is now monitored on a daily basis to

determine when the target of 60 % biomass hasbeen removed each year. The remaining 40 %biomass is deemed by the Falkland IslandsGovernment to be adequate as prey for seabirdsand marine mammals, and as breeding stock forthe following season. Whilst this may ensuresustainable use of stocks as a financial resource,it seems unreasonable to suppose that 60 % of thebiomass can be removed prior to the breedingseason of seabirds and marine mammals, withouthaving an impact on those species which rely onsuch prey for successful breeding.

Diet sample analysis shows that there iscompetition between penguins and commercialfisheries for loligo squid (Loligo gahi) (Putz et al.2001) and potential competition for blue whiting(Micromesistius australis) (Table 1). Since dietsamples have only ever been taken in the presenceof commercial fishing activities, the proportionof commercially harvested species found inpenguin diet will be greatly reduced. Diet analysis

TABLE 1

Diet of three species of penguins in the Falkand Islands. Data for each prey categorycorrespond to the percentage of weight out of the total stomach content. The total number of

stomachs examined is indicated in parentheses for each species

Dieta de tres especies de pingüinos en las Islas Malvinas. Para cada presa se indica el porcentaje del peso enrelación al contenido estomacal total. Para cada especie se indica el número total de estómagos examinados

Prey item Gentoo Rockhopper Magellanic(n = 238) (n = 187) (n = 297)

Squid 12 27 41Loligo gahi 5.40 7.02 18.45Gonatus antarcticus 5.04 15.66 20.50Moroteuthis ingens 0.12 0 0.41Sepiola 0.72 0 0Octopod 0.72 4.05 0.82Onychoteuthid 0 0.27 0.41Ommastrephes 0 0 0.41

Fish 54 23 40Sprattus fuegensis 9.18 9.89 15.60Micromesistius australis 0.54 3.22 8.00Thysanopsetta naresi 1.62 0 0Agonopsis chiloensis 0.54 4.37 2.40Eleginops maclovinus 0 3.22 0Patagonotothen 18.36 2.30 0.40Nototheniodei 10.80 0 13.60Gadiformes 6.48 0 0Perciformes 5.94 0 0Raja 0.54 0 0

Crustaceans 34 50 19Munida gregaria 33.66 10.00 18.43Themisto gaudichaudi 0.34 1.50 0.57Euphausia 0 33.50 0Amphipoda 0 5.00 0

BINGHAM

8 1 5

will therefore greatly under-estimate the level ofcompetition between penguins and commercialfisheries.

Any competit ion for food is l ikely to beexacerbated by the fact that the commercial fishingseason runs until the end of October, whenpenguins and other seabirds begin breeding. Sincestocks are managed by recording the daily declineof catch per unit effort, it stands to reason thatpenguins will also encounter this decline in catchper unit effort as they enter their breeding phase.

Breeding places great demands on adultpenguins to find sufficient food for themselvesand their chicks. Each adult must first catchsufficient food for its own needs, and a surplusfor its chicks, at a time when its foraging range isgreatly reduced by the need to return to the nest.A reduction in food abundance can therefore leadto greater foraging duration, less food beingbrought back to chicks, and reduced chick survival.

It is probable that the rapid penguin declinesobserved in the Falklands during the 1980s werea result of uncontrolled commercial fishing.Elephant seals and southern sea lions, which alsofeed on species taken by the commercial fishingindustry, also underwent huge population decli-nes in the Falklands during the same period(St range 1992, Thompson & Duck 1996,Galimberti 2000). As is the case for penguins,these species also occur in South America whereno such declines occurred. King penguins and furseals which feed on prey not commerciallyharvested, increased in population around theFalklands over the same period.

During the late 1970s and early 1980s, fish andsquid stocks were heavily fished without anymonitoring or control. In 1986 a lack of food ledto mass starvation of adult penguins, and publicconcern called for controls over commercialf ishing. Fol lowing the establ ishment of aregulatory body in 1988, the effects of over-fishing have been greatly mitigated throughreduction of fishing effort and control of speciestaken.

Chick and juvenile survival have shown somesigns of improvement, and gentoo and rockhopperpopulat ions have stopped decl ining. Thesepopulations appear to have reached equilibriumwith the current fishing regime, albeit at a muchlower level than prior to the 1980s. Magellanicpenguins have continued to decline however, andit is likely that their greater dependence on specieswhich are still commercially harvested is a majorfactor.

Concerns about the removal of gentoo penguineggs for human consumption were not borne outby the data, which suggest little effect on breeding

success. Unlike most other penguins, gentooshave the ability to relay within a few days ofloosing their clutch. This enables them to rearchicks successfully after losing their first brood.Although egging could hardly be advocated byany conservationist, in the Falklands it is atradition which is gradually dying out, due to animproved infra-structure that allows people inremote areas to purchase a wider range of fooditems. Conservationists have taken the view thatit is preferable to let the tradition die a naturaldeath, rather than risk a resurgence by threateningpeople’s right to continue the practice.

Tourism is another potential cause for concern.The Falkland Islands are now one of the world’smost popular destinations for penguin spotters,and this growing tourist industry is a potentialthreat to Falklands penguins. The Falkland Islandsreceived around 35,000 visitors during the 2001-2002 season (Fowler 2002).

Over recent years a number of scientific reportshave demonstrated that even well-behaved visitorscan cause stress and increased heart-rate inpenguins and seabirds, but these factors are notnecessarily harmful to the bird or its fecundity.Seabirds are subjected to varying levels of stressin their natural environment, so it was importantto monitor the effects of tourism in its widerperspective, by conducting long-term studies ofpopulation trends and breeding success. Breedingsuccess in particular provides a useful measure ofvisitor disturbance. Careless visitors have thepotential to disturb breeding penguins in a numberof ways: (1) incubating birds may be frightenedaway allowing predators to take eggs or young,(2) raised metabolic rates brought on by stressmay lead to greater food requirement, (3) naturalbehaviour, such as courtship or the feeding ofyoung, may be disrupted, (4) adults could bescared away completely, causing them to abandoneggs or young, (5) severe disturbance could leadto adults or young being killed or injured, and (6)birds that live in burrows may be killed if theburrow collapses under human weight.

These potential consequences of disturbanceshould all lead to reduced breeding success ifthey are occurring at a significant level, howeverstudies of penguin breeding success in theFalklands and southern Chile show no harmfuleffects from tourism so far. Other penguin studieshave reached similar conclusions (Cobley &Shears 1999).

Ecotourism undoubtedly has a number ofbenefits. It provides wildlife with a commercialvalue, giving support for its protection within thecommercial sector. It also educates and entertainsthe people who see the wildlife, raising awareness

FALKLAND ISLANDS PENGUINS

8 1 6

and gathering support for wildlife protectionwithin the community as a whole. It is difficult toprovide strong argument for wildlife protectionunless people can relate to wildlife on a personallevel. It is therefore important to promoteecotourism, whilst at the same time ensuring thatsuch tourism does not damage the wildliferesources that people come to see.

There are clearly a number of threats facingpenguin populations in the Falkland Islands, butpenguin populations appear very robust todisturbance and moderate levels of exploitationon land. Major changes to the landscape broughtabout by livestock, the removal of eggs for food,and exploitation as a resource for tourism, allappear to have had low impact on penguinpopulations in the Falklands. Marine pollutionaround the Falklands has so far been limited to aspate of oil spills that occurred during drillingoperations in 1998, and high levels of cadmiumcommon to the Antarctic region in general, thatmay be due to natural factors (Keymer et al.2001). The only major predators of penguins,southern sealions, have suffered a 97 % decline inthe Falklands (Thompson & Duck 1996) makingthem an unlikely cause of penguin decline.

Even in a healthy population, starvation is themain cause of chick mortality for Magellanicpenguins (Scolaro 1990, Boersma 1991*), and inthe Falkland Islands low breeding success, highchick mortality, low fledging weight and lowrecruitment are largely due to low food supply(Putz et al. 2001). On at least one occasion thislow food supply has also led to mass starvation ofadult penguins (Keymer et al. 2001).

Whilst it is unrealistic to expect the FalklandIslands Government to halt commercial fishingactivity, which is a major source of revenue to theislands, minor changes could be adopted whichwould mitigate the effects of commercial fishingon penguin populations, without greatly effectingrevenue. At present commercial fishing vesselsare permitted to fish within 5 km (3 miles) of thecoastline, even where penguin breeding sites arelocated. It was proposed at the Spheniscus PenguinConservation Workshop (September 2000, Chile)that fishing vessels in the Falkland Islands shouldbe excluded from within 48 km (30 miles) ofpenguin breeding sites during the breeding season.Such measures would protect feeding areas within

the penguins’ daily foraging range, whilst reducingthe total area available to fishing vessels by just4 %.

Satellite tracking has shown that around theFalkland Islands, Magellanic penguins have amean foraging range of about 16 km duringincubation, and 7 km during chick rearing, with amaximum distance of 39 km being recorded duringchick rearing. Gentoo penguins have a meanforaging range of 6 km, with a maximum of lessthan 25 km being recorded (Boersma et al. 2001).These foraging distances lie well within the 48km (30 mile) exclusion zone requested at theSpheniscus Penguin Conservation Workshop.Rockhopper penguins were recorded foragingoutside the requested exclusion zone, with shortforaging trips averaging less than 6 km, beingsupplemented by long distance foraging trips ofwell over 100 km. Even so rockhoppers mademore use of inshore waters during the criticalchick-rearing stage (Boersma et al. 2001).

Magellanic penguins fitted with time-depthrecorders at the two Chilean study sites showedsimilar results, with foraging beginning at around7 km from the breeding site. Mean foraging rangewas 25 km, and maximum foraging ranges for thetwo sites were 33 km (Seno Otway) and 47 km(Magdalena Island) (Radl & Culik 1998).

Inshore fishing around the Falklands has anegative impact on gentoo, rockhopper andMagellanic penguins (Boersma et al. 2001). It istherefore probable that if a fishing exclusion zonewere established around penguin breeding sitesin the Falkland Islands, as has been done aroundMagdalena Island in Chile, that this would allowan increase in food availabil i ty within thepenguins’ foraging range. This in turn shouldlead to a decrease in foraging duration, an increasein food brought back to chicks, and an increase inchick survival and fledging weights, as has beenobserved since the exclusion of large-scalecommercial fishing from around MagdalenaIsland. The exclusion of fishing would only berequired just prior to and during the breedingseason, and it might also help protect Falklandspenguins from their other major threat, marinepollution.

During 1998 an oil rig was sent to the FalklandIslands to look for oil. During the five months ofoperation three separate oil spills occurred killinghundreds of penguins. It is unlikely that the oilcame from the rig itself, which claimed never tohave found oil in commercially viable quantities.The oil is presumed to have come from oil rigsupply vessels operating in Falkland waters at thetime. No further oil spills have occurred since theoil rig and its supply vessels left the Falklands.

* BOERSMA PD (1991) Asynchronous hatchingand food allocation in the Magellanic penguin.Acta XX Congressus Internationalis Ornithologici:961-973.

BINGHAM

8 1 7

In 1995 the United Kingdom, Argentina and theFalkland Islands set aside a Special Area of Co-operat ion for fu ture o i l exp lorat ion anddevelopment, so there is little doubt that oilexploration around the Falklands will recommencein the near future. Unless environmental protectionis greatly improved, it is probable that many morepenguins will die in unnecessary oil spills, ashappens each year along coastal Argentina.

The Falkland Islands are an internationallyimportant breeding site for penguins, and it isvital that the Falkland Islands Government accepttheir responsibility to protect this natural resource.Rockhopper penguins now number just 11 % ofthe population recorded in the Falklands 18 yearsago, and they are now classified as “vulnerable”under International Union for Conservation ofNature (IUCN) guidelines. Magellanic penguinshave undergone a similar magnitude of decline,and are still declining. Sea lions and elephantseals have also undergone major declines sincethe establishment of a commercial fishing industryin the Falklands.

It is very difficult to prove the link between thedecline of these species, and the establishment oflarge-scale commercial f ishing around theFalklands, just as it is difficult to prove linksbetween smoking and individual cases of heartdisease or lung cancer. However, many take theview that for non-target species, and especiallyprotected species such as penguins, the burden ofproof for no harm lies with the exploiter (Boersmaet al. 2001).

On a per capita basis, the Falklands is one of thewealthiest places on earth, with an annualgovernment income of over US$ 30,000 for everyperson living in the islands. As such there is noreason why financial interests should outweighthe need for adequate protection of Falklandswildlife.

ACKNOWLEDGEMENTS

Thanks go to the Environmental Research Unit,Falklands Conservation, Falkland Islands FisheriesDepartment, Corporación Nacional Forestal (Chi-le), Fundación Otway (Chile), Aerovias DAP, ArneRadl, the rangers of Magdalena Island, and the crewsof “Melinka”, “Tierra Australis”, “Hundy” and “DonJorge” for logistical support, and to researchassistants Nidia Mendez and Cici Legoe. Thanksalso to the Conservation & Research Foundation(Connecticut, USA), the PADI Foundation (BeverlyHills, USA), Fauna and Flora International (UK),and the Darwin Initiative (British Government) forproviding funding.

LITERATURE CITED

BENNETT AG (1933) The penguin population of theFalkland Islands in 1932/33. Government Press,Falkland Islands. 4 pp.

BINGHAM M & E MEJÍAS (1999) Penguins of theMagellan region. Scientia Marina 63 (Suplement) 1:485–493.

BOERSMA PD, DL STOKES & PM YORIO (1990)Reproductive variability and historical change ofMagellanic penguins at Punta Tombo, Argentina. In:Davis LS & JT Darby (eds) Penguin biology: 15-43.Academic Press, San Diego, California.

BOERSMA PD, DL STOKES & IJ STRANGE (2001)Tracking breeding penguins at New Island SouthReserve: an overview of work carried out between1997 and 2001. New Island South Conservation TrustSupplement 2, Stanley, Falkland Islands. 4 pp.

BROWN RG & DN NETTLESHIP (1984) Capelin andseabirds in the Northwest Atlantic. In: NettleshipDN, GA Sanger & PF Springer (eds) Marine birds:their feeding ecology and commercial f isheriesrelationships: 184-195. Canadian Wildlife ServicesPublication, Ottawa, Canada.

CAIRNS DK (1987) Seabirds as indicators of marine foodsupplies. Biological Oceanography 5: 261-271

CLAUSEN A (2000) Falkland Islands Seabird MonitoringProgramme Report. Falklands Conservation ReportSMP8, Stanley, Falkland Islands. 56 pp.

CLAUSEN A (2001) Falkland Islands Penguin Census2000/01. Falklands Conservation Report, Stanley,Falkland Islands. 27 pp.

COBLEY ND & JR SHEARS (1999) Breeding performan-ce of gentoo penguins at a colony exposed to highlevels of human disturbance. Polar Biology 21: 355-360.

CRAWFORD RJM & BM DYER (1995) Responses byfour seabird species to a fluctuating availability ofCape anchovy Engraulis capensis off South Africa.Ibis 137: 329-339.

CROXALL JP, MCINNES SJ & PRINCE PA (1984) Thestatus and conservation of seabirds at the FalklandIslands. In: Croxall JP, PGH Evans & RW Schreiber(eds) Status and conservation of the world’s seabirds:271-291. ICBP Technical Publication No.2, ICBP,Cambridge, United Kingdom.

CULIK BM & G LUNA-JORQUERA (1997) Satellitet rack ing o f Humbold t pengu ins (Spheniscushumboldti) in northern Chile. Marine Biology 128:547-556.

DUFFY DC, RP WILSON, RE RICKLEFS, SC BRONI &H VELDHUIS (1987) Penguins and purse seiners:competition or coexistence? National GeographicResearch 3: 480-488.

FALKLAND ISLANDS GOVERNMENT (1989) FalklandIslands interim conservation and management zonefisheries report ’87/88. Falkland Islands GovernmentReport, Falkland Islands. 74 pp.

FALKLAND ISLANDS GOVERNMENT (2001) FisheriesDepartment Fisheries Statistics, Volume 5, 2000. FIGFisheries Department, Stanley, Falkland Islands. 68pp.

FALKLAND ISLANDS PENGUINS

8 1 8

FOWLER J (2002) Review of trends and changes inFalklands Tourism 1995-2001. Falkland IslandsTourist Board, Stanley, Falkland Islands. 46 pp.

FRERE E, M GANDINI, P GANDINI, T HOLIK, VLICHTSCHEIN & MO DAY (1993) Variación anualen el número de adultos reproductivos en una nuevacolonia de pingüino penacho amarillo en Isla Pingüi-no (Santa Cruz, Argentina). Hornero 13: 293-294.

FURNESS RW & CJ CAMPHUYSEN (1997) Seabirds asmonitors of the marine environment. ICES Journal ofMarine Science 54: 726-737.

GALIMBERTI F (2000) Elephant seals of the Falklands.Elephant Seal Research Group report, Milano, Italy.31 pp.

KEYMER IF, HM MALCOLM, A HUNT & DT HORSLEY(2001) Health evaluation of penguins followingmortality in the Falklands. Diseases of AquaticOrganisms 45: 159-169.

LEWIS-SMITH RI & JRB TALLOWIN (1979) Thedistribution and size of king penguin rookeries onSouth Georgia. British Antarctic Survey Bulletin 49:259-276.

PATTERSON KR (1987) Fishy events in the FalklandIslands. New Scientist 1562: 44-48.

PUTZ K, RJ INGHAM, JG SMITH & JP CROXHALL(2001) Population trends, breeding success and dietcomposition of gentoo, Magellanic and rockhopperpenguins in the Falkland Islands. Polar Biology 24:793-807.

RADL A & BM CULIK (1998) Foraging behaviour andreproductive success in Magellanic penguins: acomparative study of two colonies in southern Chile.Marine Biology 133: 381-393.

SCOLARO JA (1990) Effects of nest density on breedingsuccess in a co lony o f Magel lan ic penguins(Spheniscus magellanicus). Colonial Waterbirds 13:41-49.

STRANGE IJ (1992) Wildlife of the Falkland Islands andSouth Georgia. Harper Collins, London, UnitedKingdom. 188 pp.

THOMPSON D & CD DUCK (1996) Southern sea lions inthe Falkland Islands. Sea Mammal Research UnitReport, Cambridge, United Kingdom. 48 pp.

THOMPSON KR (1993) Variation in Magellanic penguindiet in the Falkland Islands. Marine Ornithology 21:57-67.

VENEGAS C (1984) Estado de las poblaciones de pingüi-no de penacho amarillo y macaroni en la Isla Noir,Chile. Informe Instituto de la Patagonia 33, PuntaArenas, Chile. 25 pp.

VENEGAS C (1991) Estudio de cuantificación poblacionalde pingüinos crestados en Isla Recalada. InformeInstituto de la Patagonia 55, Punta Arenas, Chile. 27pp.

WILSON RP (1984) An improved stomach pump forpenguins and other seabirds. Journal of FieldOrnithology 55: 9-12.

WOEHLER EJ (1993) The distribution and abundance ofAntarc t ic and subantarc t ic pengu ins. SCAR,Cambridge, United Kingdom. 76 pp.

BINGHAM

Associate Editor: P. MarquetReceived October 26, 2001; accepted July 8, 2002