Chapter 1

Chapter 1 Executive Summary BCLME Top Predators Project Steering Committee

1. OBJECTIVES OF PROJECT

The Steering Committee met on three occasions:

Project LMR/EAF/03/02 of the Benguela Current Large Marine Ecosystem (BCLME) Programme investigated landbreeding top predators in the BCLME. The project had objectives to assess the utility of top predators as biological indicators of ecosystem change in the BCLME and to implement an appropriate, integrated, system-wide monitoring programme to support sustainable management of the BCLME.

17 February 2005, Swakopmund, Namibia; 15 May 2006, Robben Island, South Africa; 6–8 March 2007, Swakopmund, Namibia.

2. SCOPE OF PROJECT 2.1 The project considered only land-breeding top predators in the BCLME region, the extent of which was taken to accord with the definition of the BCLME that appears in Article 1 of the Interim Agreement Between the Government of the Republic of Angola and the Government of the Republic of Namibia and the Government of the Republic of South Africa on the Establishment of the Benguela Current Commission (BCC). 2.2 In accordance with Chapter II of the Memorandum of Agreement relating to LMR/EAF/03/02, the top predators on which primary emphasis was laid were Cape Fur Seal Arctocephalus pusillus pusillus, African Penguin Spheniscus demersus, Great White Pelican Pelecanus onocrotalus, Cape Gannet Morus capensis, Cape Cormorant Phalacrocorax capensis, Bank Cormorant P. neglectus, Crowned Cormorant P. coronatus, White-breasted Cormorant P. lucidus, Kelp Gull Larus dominicanus, Hartlaub’s Gull L. hartlaubii and Swift Tern Sterna bergii. Other species considered included Leach’s Storm Petrel Oceanodroma leucorhoa and Roseate Tern S. dougallii. Substantive time series of information were not available for the Grey-headed Gull L. cirrocephalus, Caspian Tern S. caspia and Damara Tern S. balaenarum, which also breed in the BCLME. 2.3 Additionally, the project investigated the mammalian and avian faunas of the marine region of southern Angola, and reports briefly on the status of turtles in Angola. 3. STEERING COMMITTEE The Steering Committee for the project comprised representatives from the Republic of Angola, the Republic of Namibia, the Republic of South Africa and the BCLME (Annex 1).

4. OBJECTIVES FOR ECOSYSTEM MONITORING IN THE BCLME BASED ON LAND-BREEDING TOP PREDATORS The project considered objectives for a monitoring programme for land-breeding top predators in the BCLME region (Chapter 3). These include: accounting for the requirements of predators dependent on species targeted by fisheries, as required in an Ecosystem Approach to Fisheries (EAF); providing information useful in the management of prey resources; providing indices of the state of health of marine ecosystems; monitoring, assessing and updating the conservation status of species of conservation concern; assessing the outcomes of management interventions aimed at improving the conservation status of top predators; monitoring interactions between species where one or more species is of conservation concern; assessing the effect of environmental variability (including climate change) on the ecosystem. 5. COLLATION OF TIME SERIES FOR SEALS AND SEABIRDS IN THE BCLME An inventory of the available time series of data for seals and seabirds in Namibia and South Africa, was made, including instructions as to how data can be accessed by prospective users (Chapters 4 and 5). 5.1 Seals Time series of information that was collated for seals in the BCLME included: diet information from South Africa (stomach samples) and Namibia (scat samples); aerial censuses of pup production in both countries.

Top Predators of the Benguela System

1

Figures 1a and b: The first aerial census of Cape Fur Seals at Baia dos Tigres took place in December 2006. Results confirmed the large northward extension of the breeding range of Cape Fur Seals. (Photos: M.A. Meÿer)

5.2 Seabirds Time series of information that was collated for seabirds in the BCLME included: sizes of regional breeding populations, or portions thereof, for 13 species; sizes of adult and immature populations at five colonies of African Penguins; breeding success of African Penguins at four Namibian colonies and one South African colony and of Cape Gannets and two South African colonies; the contribution of anchovy Engraulis encrasicolus and sardine Sardinops sagax to the diet of Cape Gannets at two Namibian and three South African colonies. 6. REVIEW OF TIME SERIES FOR LAND-BASED TOP PREDATORS IN THE BCLME

Seals increased substantially between 1972 and 1993, at a rate of approximately 3% per annum. However, there was no significant change between 1993 and 2004, suggesting that the population size (all age classes) remained relatively constant over this period. The counts for more recent surveys (2005–2007) have yet to be completed (Chapter 6). (See Fig. 1.) 6.1.1.2 Angola Following reports of increasing numbers of seals at Ilha dos Tigres in recent years, a ground count of seals at the island was conducted in November 2005 (Chapter 39) and aerial photographs were taken in December 2006 (Chapter 40). During the land survey, which took place before the peak of the pupping season, 1 165 pups were counted; on the aerial photographs 4 378 pups were counted. These results confirm the island’s status as a seal breeding location.

6.1.1 Population trend

6.1.1.3 Namibia Pup numbers in Namibia increased substantially between 1972 and 1993, at approximately 3.3% per annum. However, pup counts fluctuated considerably between 1993 and 2004, as a result of the effects of environmental perturbations on feeding conditions in the northern Benguela system. There was an overall decline of 15% in pup counts over this period (Chapter 6).

6.1.1.1 Overall Based on pup counts, the overall population of Cape Fur

6.1.1.4 South Africa In South Africa, pup numbers increased substantially be-

The project reviewed collated time series of information for land-based top predators in the BCLME. 6.1 Cape Fur Seal

2


tween 1972 and 1993, at approximately 2.8% per annum. However, the pup counts then stabilized, with no significant change occurring between 1993 and 2004 (Chapter 6). 6.1.1.5 Distributional shift A northward shift is apparent in the spatial distribution of the seal population in the northern Benguela system. This is confirmed by the growth of the colony at Cape Frio in northern Namibia (approximately 30% per annum between 1993 and 2004, based on aerial censuses), the establishment of a breeding colony at Baia dos Tigres (Chapter 37), and the decline in seal numbers at colonies in the Lüderitz area, which contributed about 60% of pups born in Namibia in 1993 but less than 50% of those born in 2004 (Chapter 6). 6.1.2 Diet South Africa Large increases in the abundance of sardine and anchovy in the southern Benguela during the last 20 years, were reflected in the seals’ diet (Chapter 7). There was no evidence that fluctuations in these favoured prey species affected birth

rates at seal colonies on the west coast of South Africa. However, the effects of recent eastward shifts of these prey species on seal diet and population dynamics, should be monitored. 6.2 African Penguin 6.2.1 Population 6.2.1.1 The overall number of African Penguins breeding decreased by 74% from 141 000 pairs in 1956/57 to 36 000 pairs in 2006/07. 6.2.1.2 Namibia The number of African Penguins breeding in Namibia decreased by 90% from 42 000 pairs in 1956/57 to 3 200 pairs in 2006/07. The more recent trends are documented in Chapter 8. 6.2.1.3 South Africa The number of African Penguins breeding in South Africa

Figures 2a and b: Halifax Island in the 1930s (top) and 2004 (bottom). The African Penguin population in Namibia declined substantially during the latter half of the 20th century. (Photos: Eberlanz Museum, Lüderitz and J. Kemper) Top Predators of the Benguela System

3

Figures 3a and b: Artificial nests provided at Halifax island improved the nesting habitat of African Penguins, and positively influenced breeding success. (Photos: J. Kemper)

decreased by two thirds from 99 000 pairs in 1956/57 to 33 000 pairs in 2006/07. The more recent trends in the Western Cape are documented in Chapter 10. 6.2.1.4 Census techniques Various techniques for counting African Penguins are compared and discussed in Chapter 9.

6.2.4 Energetics It was estimated that the amount of food African Penguin parents have to provide their chick between hatching and fledging on a diet of sardine, would be 13.1 kg. On a diet of anchovy, 16.7 kg would be required (Chapter 15).

6.2.2 Breeding success

6.3 Great White Pelican

6.2.2.1 Namibia 6.2.2.1.1 Estimates of breeding success of African Penguins in Namibia varied between localities, years, seasons and habitats. Nests initiated at the end of October at Mercury, Ichaboe and Halifax islands and at the end of September at Possession Island were more likely to be successful than at any other time of year. Nest position, nest type and vulnerability to flooding were important explanatory variables of breeding success (Chapter 11).

6.3.1 The number of Great White Pelicans breeding in the Western Cape increased from about 30 pairs in 1954 to more than 800 pairs in 2004, but decreased to less than 400 pairs in 2006. The increase followed the removal of controls and an increased food supply provided by artificial dams and agricultural offal. The decrease resulted from a decrease in the supply of agricultural offal (Chapter 16).

6.2.2.1.2 From 2001/02–2003/04 at Halifax Island, the breeding success of African Penguins was improved by the provision of artificial sites (see Fig. 3), which provided better nesting habitat than natural sites (Chapter 12).

6.3.2 Following the decreased supply of agricultural offal, pelicans inflicted high mortality on the nestlings of some seabird species. 6.4 Cape Gannet

6.2.2.2 South Africa The breeding success of African Penguins was measured at Robben Island from 1989/90–2004/05. It was significantly related to estimates of the abundance of anchovy and sardine, the main prey species of penguins in the Western Cape (Chapter 13).

6.4.1 Population

6.2.3 Moult

6.4.1.2 Namibia The number of Cape Gannets breeding in Namibia decreased by 95% from 204 000 pairs in 1956/57 to 10 000 pairs in 2005/06 (see Fig. 6).

The seasonal pattern of moult of African Penguins in adult plumage at islands in Namibia was linked to the season of breeding. Counts of birds moulting provided rough estimates of the breeding population at Namibian localities (Chapter 14).

6.4.1.1 Overall (Chapter 17) The overall population of Cape Gannets was about 250 000 breeding pairs from 1956/57–1968/69 and about 150 000 pairs from 1978/79–2005/06.

6.4.1.3 South Africa The number of Cape Gannets breeding in South Africa increased from 50 000 pairs in 1956/57 to 135 000 pairs in 2005/06. 6.4.1.4 Distributional shifts (Chapter 17) There was a long-term shift to the south and east in the distribution of breeding Cape Gannets, with the largest colony at present in the Eastern Cape. This shift followed a similar shift in the distribution of sardine, an important prey of Cape Gannets. 6.4.2 Breeding success

Figure 4: At Dassen Island, Great White Pelicans surrounded nests of Cape Cormorants. Many Cape cormorants were eaten by the pelicans, after a decreased supply of agricultural offal. (Photo: R. Mullers) 4


6.4.2.1 South Africa The breeding success of Cape Gannets was measured at two colonies in the Western Cape. At Lambert’s Bay, from 1991/92–2004/05, gannets fledged on average 0.58 chicks per pair during a breeding season. In 2005/06, the entire colony abandoned the island and no chicks were fledged. At

Figure 5: The number of Cape Gannets in Namibia decreased by about 95% during the latter half of the 20th century. (Photo: R. Mullers)

Malgas Island, from 1988/89–2004/05, gannets fledged on average 0.44 chicks per pair during a breeding season. This fell to 0.01–0.02 chicks per pair in the 2005/06 and 2006/07 breeding seasons (Chapter 31). Chick growth was measured at Malgas Island in 2003/04 and 2004/05 (Chapter 18). Growth was slower in the latter season. This was associated with a decrease in the proportion of two prey species with high calorific contents, anchovy and sardine, in the diet. Reduced fledging success was associated with the decline in chick growth rates in 2004/05.

Figure 6: A comparison of the proportions of Cape Gannets and epipelagic fish (sardine and anchovy) in Namibia and South Africa over a 50-year period, illustrating the long-term shift in the distributions of these resources (from Chapter 17).

6.4.3 Diet 6.4.3.1 Namibia In the 1950s and 1960s, sardine dominated the diet of Cape Gannets in Namibia. During the late 1970s, the diet consisted largely of anchovy. By 2004, hake Merluccius spp. and saury Scomberesox saurus contributed >70% by mass to the diet at Ichaboe Island, whereas horse mackerel Trachurus trachurus, snoek Thrysites atun and saury dominated the diet at Mercury Island. A considerable proportion of hake in the diet originates from gannets feeding on offal discarded by hake trawlers and long-liners. 6.4.3.2 South Africa Anchovy and sardine contributed 49–77% by mass of the diet of Cape Gannets in the Western Cape from 1978–2003, but just 16% in 2006. In the Eastern Cape, they contributed 22– 76% by mass of the diet from 1979–2001 but 83–98% from 2002–2006. The decreased contribution of these species to the diet in the Western Cape and their increased contribution in the Eastern Cape coincided with an eastward displacement of prey (Chapter 31).

6.5.1.2 Namibia and South Africa The population in Namibia increased between the 1950s and 1970s after additional breeding space became available as a result of a decrease in Cape Gannets at Ichaboe Island and the development of guano platforms off central Namibia. Decreases later took place in Namibia and South Africa, as a consequence of food scarcity, disease and predation of fledglings by Cape Fur Seals and Great White Pelicans. 6.5.1.3 Angola The species was first recorded breeding in Angola, at Baia dos Tigres, in 1996. This represented a large northward extension of the breeding range of the species. Some 2 600 breeding pairs were counted at Baia dos Tigres in November 2005 (Chapter 39). 6.5.1.4 Trends in guano production Cape Cormorants are the main producer of guano at the platforms off central Namibia, where in 2006 the yield of 1 012 tonnes was the lowest recorded since 1979 (see Figs 7). This was less than half the average for the 50-year period 1956– 2005 (Chapter 19).

6.5 Cape Cormorant 6.6 Bank Cormorant 6.5.1 Population 6.6.1 Namibia 6.5.1.1 Overall The overall population of Cape Cormorants increased from more than 110 000 pairs in 1956/57 to about 250 000 pairs in 1978/79, but decreased to less than 100 000 pairs in 2005/ 06 (Chapter 19). However, the population for 1956/57 may be underestimated.

Between 1993 and 1998 the Namibian breeding population is estimated to have declined by 68%, and it has not recovered since. Mercury and Ichaboe islands support 80% of the Namibian and 57% of the global breeding population of Bank Cormorants (Chapter 25). Top Predators of the Benguela System

5

Figures 7a and b: In 2006, the yield of seabird guano at platforms in Namibia attained its lowest level since 1979. In March 2007, guano platforms at Cape Cross were devoid of breeding Cape Cormorants. (Photos: R.J.M. Crawford)

6.6.2 South Africa About 600 pairs of Bank Cormorants bred at eight islands in the Western Cape in 1990, but this fell to about 300 pairs after 1995 (Chapter 20).

cause no recent counts are available for localities north of Walvis Bay, where the majority of the Namibian coastal population breeds, it was not possible to evaluate trends for this species in Namibia. 6.8.2 South Africa

6.7 Crowned Cormorant 6.7.1 Namibia Numbers of Crowned Cormorants in Namibia are considered stable. The breeding population is estimated at roughly 1 000 pairs (Chapter 25).

Numbers of White-breasted Cormorants breeding at nine islands in the Western Cape fluctuated around 200 pairs for the period 1977–2006, with no trend (Chapter 20). 6.9 Kelp Gull 6.9.1 Population

6.7.2 South Africa Numbers of Crowned Cormorants breeding at ten islands in the Western Cape increased from 800 pairs in 1978 to more than 1 200 pairs from 2003/04–2006/07. There was a large eastward expansion in the breeding range of Crowned Cormorant off South Africa from west of Cape Agulhas to Tsitsikamma at some time between 1981 and 2003 (Chapter 20). 6.8 White-breasted Cormorant 6.8.1 Namibia Whitebreasted Cormorants breed at a number of localities along Namibia’s coast, as well as at inland localities. Be-

6.9.1.1 Namibia In Namibia, the highest numbers of breeding Kelp Gulls are found at two islands situated in Lüderitz harbour, where gulls profit from supplementary food provided by a rubbish dump and offal from fish processing. Kelp Gulls also breed in large numbers at Possession Island and at the Swakopmund Salt Works and guano platform. Numbers of breeding pairs in Namibia are generally stable or increasing (Chapter 25). 6.9.1.2 South Africa The number of Kelp Gulls breeding at 11 islands in the Western Cape increased from 6 500 pairs in 1978 to 16 000 pairs in 1999 and then decreased to 13 000 pairs in 2005. The increase came after removal of controls on gulls and was associated with supplementary food provided at rubbish tips and fish factories. The decrease resulted from predation of up to 85% of gull chicks hatched at some colonies by Great White Pelicans. Decreases at the Western Cape’s northern colonies of Kelp Gulls and increases in numbers to the south and east correspond to a similarly altered distribution of several species of seabird and fish off South Africa, thought attributable to long-term environmental change (Chapter 21). 6.9.1.3 Angola At least 950 pairs of Kelp Gulls bred at Ilha dos Tigres in November 2005 (Chapter 39).

Figure 8: A recent decline of Kelp Gull numbers in South Africa is associated with predation of chicks by Great White pelicans (Photo: B.M. Dyer).

6


6.9.2 Densities of nests

6.13 Caspian Tern

At Dassen Island, the density of Kelp Gull nests remained constant as the colony doubled, but decreased by 50% as the colony decreased, suggesting that Kelp Gulls spaced out their nests more when subjected to predation. At this locality the clutch size increased following the initiation of predation (Chapter 21).

The Caspian Tern breeds at inland and coastal water bodies.

6.9.3 Survival

6.13.2 Population in South Africa’s West Coast National Park

The mean annual survival rate of juvenile and adult birds in the BCLME was estimated to be 0.44 and 0.84, respectively. The mean clutch size was 2.2 eggs (Chapter 22).

6.13.1 Population at Ilha dos Tigres In November 2005, there were 88 pairs of Caspian Terns at Ilha dos Tigres, Angola (Chapter 39).

The number breeding in South Africa’s West Coast National Park and adjacent localities decreased from 32 pairs in 1999 to four pairs in 2006.

6.9.4 Energetics The energy budget of a Kelp Gull chick between hatching and fledgling was estimated to be 35.4 MJ. Based on this, it was estimated that approximately 8.4 kg of food needs to be delivered to a chick between hatching and fledging (Chapter 23). 6.9.5 Moult The parameters of the primary moult of adult Kelp Gulls in South Africa were estimated, using moult protocols and an appropriate moult model. The mean duration of moulting was found to be 168 days, with on average 1.9 feathers moulting simultaneously (Chapter 24). 6.10 Hartlaub’s Gull

6.14 Leach’s Storm Petrel 6.14.1 Population 6.14.1.1 The first recorded breeding of Leach’s Storm Petrel in the Southern Hemisphere was at South Africa’s Dyer Island in 1996. Subsequently, breeding was also recorded at Jutten and Dassen islands. With hindsight it is likely that a pair bred at St Croix Island from 1979–1984. 6.14.1.2 The overall number of Leach’s Storm Petrels breeding in southern Africa decreased from 22 pairs in 1999 to five pairs in 2006 (Chapter 26). 6.14.1.3 Larger numbers of Leach’s Storm Petrel from the Northern Hemisphere visit the BCLME in the austral summer.

6.10.1 Population in South Africa’s Western Cape 6.15 Colonial waterbirds Between 1978 and 2006, the number of Hartlaub’s Gulls breeding in the Western Cape fluctuated around a mean level of about 4 000 pairs. Numbers breeding in Cape Town increased when the feral cat population at Robben Island increased and displaced birds to mainland breeding sites (Chapter 26).

The loss of mainland habitat can cause islands to become important breeding localities for colonial waterbirds (Chapter 27). 6.16 The status of turtles in Angola

6.10.2 Eastward extension of breeding range There was a large eastward expansion in the breeding range of Hartlaub’s Gull after 1995, when the species was first recorded breeding east of Cape Agulhas. In 2001, it bred in Port Elizabeth. 6.11 Swift Tern From 1984–1999, less than 5 000 pairs of Swift Tern bred in the Western Cape. Then, as the biomass of epipelagic fish off South Africa increased, the number breeding increased to more than 7 000 pairs from 2004–2006. In 2006, almost all breeding occurred at the eastern colony of Dyer Island, as epipelagic fish became less available in the west (Chapter 26). 6.12 Roseate Tern A small, isolated population of Roseate Terns occurs in the southern Benguela region, where from 1977–2006 numbers breeding fluctuated between 120 and 265 pairs (Chapter 26).

During surveys at sea and along the entire coastline of Angola since 2000, five species of turtle, all listed as endangered or critically endangered by the IUCN, were recorded. Nesting of three of the species was confirmed, but the breeding status of the other two species in Angola is uncertain. The density of nests in the vicinity of Luanda is estimated to have declined by nearly half between the 1980s and the present (Chapter 28). 7. INFLUENCE OF ENVIRONMENT AND FISH STOCKS ON TRENDS IN TOP PREDATORS 7.1.1 Influences on Cape Fur Seals of long-term changes in prey abundance Considerable inter-annual variability in prey availability in the northern Benguela system, caused by environment perturbations, was reflected in censuses of pup numbers since 1993 (Chapter 6). A northward extension of the seal population’s range, into southern Angola (Chapter 40), and a decline in numbers in central Namibia, have apparently resulted from poor feeding conditions in the latter region (Chapter 6). The large increases in the abundance of epi-


7

complete breeding failure by gannets in 2005/06. The contribution of sardine to the diet of gannets fell from an average of 40% during 1987–2003 to 5–7% in 2005 and 2006. The proportions of Cape Cormorants and Swift Terns in the province that bred in the south increased as sardine moved south and east. 7.1.3.3 Eastern Cape The number of African Penguins breeding in the Eastern Cape halved between 2001 and 2003, whereas after 2002 there was an increase in the number of Cape Gannets that bred and in the contribution of sardine to their diet. Cape Gannets have a much greater foraging range than African Penguins when breeding. Figure 9: Predation by seals on seabirds, in this instance an adult Cape Gannet at Lambert’s Bay, has increased recently. (Photo: unknown supplied to A.C. Wolfaardt)

pelagic fish stocks in the southern Benguela in the last 20 years were reflected in the seals’ diet, though the growth of the population in South Africa appears to have stabilised (Chapter 7). 7.1.2 Influences for seabirds of a reduced abundance of prey Seabirds feeding mainly on anchovy and sardine (Chapters 29 and 30) In the BCLME three endemic seabirds feed mainly on anchovy and sardine: African Penguin, Cape Gannet and Cape Cormorant. Following a collapse of the Namibian sardine stock, numbers of penguins and gannets breeding in Namibia decreased by 90% and 95%, respectively. Cape Cormorants initially benefited from an increased availability of breeding space and of pelagic goby Sufflogobius bibarbatus, which partially replaced sardine, but have recently also decreased. In South Africa, when sardine collapsed it was replaced by anchovy. In the Western Cape, numbers of Cape Gannets and Cape Cormorants were initially stable but African Penguins decreased. Cape Cormorants have recently decreased.

7.1.3.4 Mitigating for reduced availability of prey Management measures that may to some extent mitigate the impacts on seabirds of an unfavourable, long-term change in the distribution of prey include the provision of breeding habitat where prey is abundant, spatial management of fisheries competing for prey and interventions aimed at limiting mortality. 7.1.4 Long-term change in carrying capacity The carrying capacity of the BCLME for African Penguins was estimated to decrease from 1.5–3.0 million adult birds in the 1920s to just 10–20% of this value from 1978–2006, as a result inter alia of environmental change and increased competition with fisheries and seals for food (Chapter 32). 7.1.5 Pollution Oil spills have had severe impacts on seabirds, notably African Penguins. Following the Treasure oil spill in 2000, 17 000 penguins were de-oiled and released. The breeding success of these de-oiled penguins was about two-thirds that of never-oiled penguins, with losses of chicks occurring mainly at the time of peak food demand. Attempts were made to raise 3 000 penguin chicks that were “orphaned” as a result of the spill. The return rate to breed of chicks raised in this way is indistinguishable from the return rate of normally raised chicks (Chapter 33). 7.1.6 Predation by Cape Fur Seals (Chapter 34)

7.1.3 Influences for seabirds of a long-term change in fish distribution (Chapters 30 and 31). 7.1.3.1 Between 1997 and 2005, sardine off South Africa was displaced 400 km to the south and east moving its centre of distribution to between seabird breeding islands in the Western Cape and in the Eastern Cape. Sardine became progressively less available to seabirds in the Western Cape but increasingly available to Cape Gannets in the Eastern Cape. Environmental change that influences the abundance and availability of prey can have severe consequences for central-place foragers, such as penguins, if prey is displaced to regions where no suitable breeding localities occur. 7.1.3.2 Western Cape Between 2004 and 2006, the number of African Penguins breeding in the Western Cape decreased by 45%. Between 2003/04 and 2006/07, the number of penguins in adult plumage that moulted at Robben Island decreased by 62%, suggesting a maximum annual survival rate in this period of 0.72 compared to 0.82 estimated for the 1990s. Penguins established a new mainland colony in the east of the province in 2003. Between 2001/02 and 2005/06, the number of Cape Gannets breeding decreased by 38%. There was almost 8


7.1.6.1 Cape Fur Seals were estimated to kill 6 000 Cape Gannet fledglings around Malgas Island in the 2000/01 breeding season, 11 000 in 2003/04 and 10 000 in 2005/06, equivalent to 29%, 83% and 57% of the overall production of fledglings in these seasons, respectively. Modelling suggests this predation is not sustainable. 7.1.6.2 There has been a large increase in predation by Cape Fur Seals on seabirds around southern African islands since the mid 1980s. (Figure 9) 7.1.7 Disease Outbreaks of epornithics (epidemics impacting birds) appear to be occurring with increasing frequencies globally. This is thought to be related to global change. Two avian diseases are of particular concern in the BCLME. 7.1.7.1 Avian cholera Since 2002, avian cholera Pasteurella multocida has been recorded on four islands of the Western Cape (Lambert’s Bay, Dassen, Robben and Dyer). The worst cholera outbreak killed 9 500 adult and 4 500 fledgling Cape Cormorants on

Dyer Island between October 2004 and January 2005 (Chapter 35). See Figure 10. 7.1.7.2 Avian malaria The vector for infections with avian malaria Plasmodium spp. is mosquitos mainly of the genus Culex. Penguins are poorly adapted to deal with avian malaria and many die. The percentage of penguins admitted to the SANCCOB rescue centre that tested malaria positive on blood smear within five days of admission was 30% in 2001 and 35% in 2002. As it takes 14 days from infection with the parasite to blood smears testing positive to the parasite, this implies that the birds admitted to SANCCOB were infected at their breeding colonies, and therefore that Culex mosquitos are breeding at some islands. 8. PREDATORS AS INDICATORS Because top predators are at the apex of the food chain, they are often good integrators of the functioning of the ecosystem in which they live. Land-based predators are readily accessible to monitoring programmes. 8.1 Seals Using otoliths recovered from seal scat samples collected along the Namibian coast, several key parameters of prey species can be monitored including relative abundance, distributional shifts and size frequency distribution. As an example, an annual recruitment index for juvenile Cape hake M. capensis has been developed (Chapter 36). This index is available about a year before the juvenile hakes are available to surveys. In addition, growth parameters and the median birth date of juvenile hake can be estimated for each cohort. This cost effective technique provides data that cannot be obtained by standard fish survey methods and that are now used in the assessment of this hake stock. Similarly, the proportion of juvenile horse mackerel found in seal scats at Cape Cross in Namibia was strongly correlated to the total catch of the fishery one to two years later, and could provide a useful index to predict purse-seine catches (Chapter 37). 8.2 Seabirds (Chapter 38) 8.2.1 Of all the marine top predators, seabirds probably have the greatest potential for regular and reliable monitoring. In the BCLME, dedicated fieldwork has provided data on numbers of breeding pairs of most seabird species on an annual basis, and there is more limited data available on diet and breeding productivity. These latter measures are likely to respond almost immediately to changes in prey abundance, with population sizes tracking longer-term trends. 8.2.2 An index that is sensitive to variability in the abundance of epipelagic fish stocks has been produced, by combining time series for species which feed preferentially on anchovies and sardines. It was significantly correlated to estimates of fish abundance.

Figure 10: Dead Cape Cormorants collected for incineration at Dyer Island, South Africa. At least 14 000 Cape Cormorants died of avian cholera at this location between October 2004 and January 2005 (photo: L.G. Underhill)

9.1.2 The north and south ends of the island were surveyed over a two-day period. There was insufficient time to survey the central area. A total of 1 165 seals were counted, including 52 pups in the north and 1 111 pups in the south. A total of 27 bird species was observed, and four coastal species (Cape Cormorant, White-breasted Cormorant, Kelp Gull and Caspian Tern) were breeding on the island. Analysis of Cape Cormorant pellets and regurgitates revealed that approximately 55% of the diet composition was horse mackerel. Eighty fledglings were banded and a PTT transmitter was placed on an adult bird, which was subsequently tracked. Evidence was found that Green Turtles Chelonia mydas may occur at the island. This needs to be investigated further. 9.2 Aerial survey, December 2006 (Chapter 40) The first aerial census of seal numbers at Baia dos Tigres was conducted in December 2006. In total 4 378 pups were counted on the aerial photographs, confirming the status of Baia dos Tigres as a seal breeding location. A further 17 062 subadult and adult seals were counted. The latter represents an underestimate of numbers of these age classes in the region, since an unknown number of them would have been at sea. 10. AT-SEA DISTRIBUTION OF MAMMALS AND BIRDS OFF SOUTHERN ANGOLA (CHAPTER 41) Observations of seabirds and marine mammals were made from a research vessel during a survey off southern Angola in July–August 2005. Results of this and previous surveys indicate that southern Angola is an important over-wintering ground for Cape Gannets. Given that recruitment of young gannets appears to be insufficient to sustain the declining breeding populations, particularly in Namibia, it is important that Angola be included in a joint conservation effort regarding these seabirds.

9. SURVEY OF BAIA DOS TIGRES 11. STATUS OF TOP PREDATORS IN THE BCLME 9.1 Ground survey, November 2005 (Chapter 39) 9.1.1 The project funded a survey of top predators at Ilha dos Tigres in November 2005. The island was visited using Angola’s R.V. Tòmbwe.

11.1 The Cape Fur Seal (as a subspecies) and nine (seven as species; two as subspecies) of the 15 seabirds that breed in the BCLME are endemic to the region. The African Black Oystercatcher also is endemic to the BCLME. The signatoTop Predators of the Benguela System

9

ries to the BCC have sole responsibility for the conservation of the endemic taxa. Of the six non-endemic seabird species, two occur as small isolated populations in the BCLME. The other four species breed in freshwater as well as marine environments through much of sub-Saharan Africa. None of five turtle species that have been recorded in Angola are endemic to the region. 11.2 Criteria of the World Conservation Union (IUCN) were used to assess the conservation status of land-based top predators in the BCLME (Chapter 42). One seabird endemic to southern Africa, the Bank Cormorant, is Endangered. Three others, the African Penguin, Cape Cormorant and Cape Gannet, are Vulnerable and the Damara Tern is Nearthreatened. The isolated population of Leach’s Storm Petrel in South Africa, is Critically Endangered and that of Roseate Tern Vulnerable. The southern African populations of Great White Pelican and Caspian Tern are Near-threatened. Hence, nine of the 15 seabirds that breed in the BCLME are Threatened or Near-threatened. 12. COMPARISON OF SEABIRDS IN THE BENGUELA AND HUMBOLDT LMES (CHAPTER 43) 12.1 The Benguela and Humboldt LMEs are the two eastern boundary current upwelling ecosystems in the Southern Hemisphere. Both have an endemic penguin, sulid, cormorant and tern that feed mainly on anchovy or sardine or both. For each of these four pairs of species, those found in the Humboldt system have a biology that enables them to increase more rapidly than their Benguela counterparts. This reflects the higher frequency of environmental perturbations that depress seabird populations in the Humboldt system. 12.2 Both the Benguela and Humboldt systems have an endemic cormorant that feeds near the coast and a small endemic tern that breeds in the adjacent mainland desert and feeds at the sea surface. 12.3 Several seabirds endemic to one of these systems have no obvious equivalents in the other system including a diving-petrel, four storm petrels, a pelican that feeds on epipelagic shoaling fish and a gull in the Humboldt system. The epipelagic feeding habit of several of these seabirds accords with the high production of zooplankton and shoaling fish in the epipelagic zone of the Humboldt system. In the Benguela system, endemic species include the Bank Cormorant, which subsists largely on the midwater pelagic goby, which is not found in the Humboldt system, and the Hartlaub’s Gull.

14. CAPACITY BUILDING “Young scientists near the start of their careers” were defined as the recipients of capacity building initiatives in terms of this project (Chapter 44). Seven such persons were trained as part of the project. Two students received BCLME bursaries, and five were supported to attend international conferences relevant to the objectives of this project and present research at these conferences. Six of the seven persons were registered for postgraduate degrees during the lifetime of the project, five of these within the Avian Demography Unit at the University of Cape Town. In addition, nine persons participated in a Time Series Workshop at Lüderitz from 24– 26 May 2005 (Chapter 45). 15. MONITORING OBJECTIVES Recommendations for future monitoring of seabirds and seals in the BCLME, are proposed in Chapters 46 and 47. 16. MANUALS FOR MONITORING Manuals of field methods, detailing a range of techniques used to monitor various aspects of seal and seabird biology, demography and ecology, were produced. The manuals aim to guide monitoring activities in the field and to standardize sampling procedures in the region (Annexes 2 and 3). 17. PUBLICATIONS Twelve papers acknowledging the BCLME project LMR/EAF/ 03/02 have been published in peer-reviewed journals. A further 19 have been submitted or will soon be submitted (Chapter 48). 18. FINANCIAL STATEMENT The original total project budget was US$217 994.00. After consultations and review, the Project Steering Committee at its meeting in 2006 revised the budget to US$207 994.00. Of this amount, US$173 621.69, has been expended to date (actual expenditure and commitments/contracts). It is expected that most of the remaining funds will be spent. The final financial report will be available after the completion of the project 19. IN-KIND CONTRIBUTIONS

13. INTERNATIONAL COOPERATION 13.1 Cooperative surveys of colonies of seals and seabirds were undertaken in Angola, Namibia and South Africa. 13.2 Scientists from the BCLME project presented papers at the Fifth International Penguin Conference in Ushuaia, Argentina in September 2004, the International Ornithological Congress in Hamburg, Germany in August 2006 and the International Conference on the Humboldt Current System in November–December 2006. Papers will also be presented at the Sixth International Penguin Conference in Hobart, Australia in September 2007, the 17th Biennial Conference of Marine Mammal Biology in Cape Town in November– December 2007 and the workshop on Seabirds as Indicators in the Western Indian Ocean in Mahe, Seychelles in December 2007. 10


In kind contributions by South Africa’s Marine and Coastal Management (SA) and Namibia’s Ministry of Fisheries and Marine Resources to the project, amounted to approximately US$ 575 000 and US$ 250 000, respectively. These included vessel time, salary time and administrative logistics and, in the case of the former, flying time for aerial photographic surveys. The Universidade Agostinho Neto, Angola, contributed approximately US$ 50 000 in the form of salary time and logistics. Contributions of bursaries and running costs by the South African National Research Foundation to postgraduate students working within the framework of the project, amounted to approximately US$ 18 000. The University of Groningen contributed approximately US$ 36 000 in fieldwork and laboratory expenses, including doubly-labeled water and associated analyses, of Dutch postgraduate students involved in the project. Approximately US$ 14 300 was

Figure 11: The top predators of the BCLME provide substantial opportunity for development of ecotourism (photo: R.J.M. Crawford)

contributed by Earthwatch, with regard to penguin fieldwork on Robben Island. In terms of salary time dedicated to the project, approximately US$ 8 600 was contributed by the University of Cape Town. The Dr Fridtjof Nansen Programme, Norway, accommodated two to three scientists on board the vessel for a month at a time, during three separate surveys in Angola. 20. RECOMMENDATIONS The Steering Committee of Project LMR/EAF/03/02 of the Benguela Current Large Marine Ecosystem: – AWARE of a long-term decrease in fish catches and of ecosystem degradation over large portions of the BCLME; – NOTING the high degree of endemism of land-breeding top predators in the BCLME (see paragraph 11.1); – further NOTING the unfavourable conservation status of a large proportion of the land-breeding top predators in the BCLME (see paragraph 11.2); – further NOTING concerns about the impact of abundant top predators on commercial fish stocks; – further NOTING that many of the land-breeding top predators in the BCLME exhibit long-term (decadal scale) transborder shifts in the distributions of populations, as well as shorter-term movements across international boundaries; – further NOTING that many of the threats facing land-breeding top predators in the BCLME, for example competition with fisheries for food, incidental by-catch mortality, oil spills and outbreaks of disease, occur throughout the BCLME region and cannot be effectively managed by states acting in isolation;

– further NOTING that consequentially populations of top predators within the BCLME are shared resources that require co-operative management; – further NOTING that states within the BCLME have signed or ratified, or are in the process of signing or ratifying, international declarations or instruments aimed at securing the conservation status of top predators, such as the Reykjavik Declaration, the Johannesburg Declaration, the Agreement on the Conservation of Albatrosses and Petrels (ACAP), the African–Eurasian Waterbird Agreement (AEWA), and the Memorandum of Understanding concerning the Conservation Measures for Marine Turtles of the Atlantic Coast of Africa; – further NOTING that states within the BCLME have developed or are in the process of developing national legislation and/or national instruments, such as NPOAs (National Plans of Action) – Seabirds, aimed at ensuring sustainable utilization and conservation of top-predator biodiversity; – further NOTING that consequentially, objectives for management of top predators for states within the BCLME will share many similarities; – further NOTING the importance of top-predator resources in the BCLME, and the potential to expand sustainable utilization of such resources through, e.g., the development of ecotourism and the collection of seabird guano; – RECOMMENDS: a) that a Top Predator Advisory Group be formed to advise the Ecosystem Advisory Committee of the Benguela Current Commission on matters relating to land-breeding top predators in the BCLME, such as the conservation of top predators, establishing target levels for populations of top predators, accounting for the requirements of top


11

predators in an Ecosystem Approach to Fishing (EAF), developing ecotourism and contributing to relevant regional or international agreements; – further NOTING that top predators in the BCLME are useful indicators of the status of some resources on which they prey and of long-term change in the structure and functioning of the BCLME, including the influence of long-term environmental (climate) change; – further NOTING the potential usefulness of top predators in providing rapid, reliable indices of the state of health of marine ecosystems, and hence inter alia in defining exceptional ecosystem circumstances, as well as in assessing the impact on marine ecosystems and resources of long-term environmental change;

– RECOMMENDS: b) that an integrated, system-wide monitoring programme for land-breeding top predators in the BCLME be implemented that has as its objectives those listed under section 1.4, and c) that the use of top predators for providing indices of the state of health of the marine ecosystem be further developed. – further NOTING the lack of capacity within the BCLME region relating to the implementation of a monitoring programme for land-breeding top predators and the curation of ensuing information; – RECOMMENDS:

– further NOTING the accessible nature of many land-breeding top predators for monitoring programmes and the reliable nature of data obtained from such monitoring; – further NOTING the likely consequential usefulness of toppredator indices in assessing the recovery of degraded portions of the BCLME;

12


d) that capacity within the region for implementing and maintaining a monitoring programme for land-breeding top predators in the BCLME, and curating ensuing information, be further developed and, where necessary, be created, with such monitoring and curation of information to be self-sustaining by 2012.

Introductory chapters

Chapter 2 Introduction BCLME Top Predators Project Steering Committee

A seal, fifteen species of seabird and an endemic shorebird breed in the Benguela Current Large Marine Ecosystem (BCLME). Aspects of their biology may be readily monitored because they breed on land. For several reasons, which are expounded more fully in chapters 3, 46 and 47, it is necessary to undertake monitoring of the BCLME’s land-breeding top predators. By virtue of their position at the apex of the food chain, they integrate processes occurring at lower trophic levels and hence are potentially good indicators of ecosystem changes. Further, several of the BCLME’s predators have a poor conservation status. Many of the BCLME’s predators are endemic to southern Africa, so that states bordering the BCLME have sole responsibility for their conservation. Aware of these factors BCLME Project LMR/EAF/03/02 A Regional Ecosystem Monitoring Programme: Top Predators as Biological Indicators of Ecosystem Change in the BCLME was initiated in March 2004. It had as objectives “to assess the utility of top predators as biological indicators of ecosystem change in the BCLME, and to implement an appropriate, integrated, system-wide monitoring programme to support sustainable management of the BCLME.” Towards this objective, the project: 1 Defined the objectives for an ecosystem monitoring programme in the BCLME based on land-breeding top predators (chapters 3, 46, 47); 2 Collated existing time series for land-based top predators of the BCLME (chapters 4, 5); 3 Analysed existing time series for land-based top predators to establish trends in these time series for Cape Fur Seal Arctocephalus pusillus pusillus (chapters 6, 7), African Penguin Spheniscus demersus (chapters 8, 10, 11, 13), Leach’s Storm Petrel Oceanodroma leucorhoa (chapter 26), Great White Pelican Pelecanus onocrotalus (chapter 16); Cape Gannet Morus capensis (chapter 17), four cormorants Phalacrocorax spp. (chapters 19, 20, 25), Kelp Gull Larus dominicanus vetula (chapters 21, 22, 25), other gulls and terns (Laridae) (chapters 25, 26);

8 Documented primary moult in Kelp Gulls (chapter 24); 9 Examined linkages between time series for top predators and fisheries data and environmental data, including carrying capacity, pollutants, predation and disease, and considered means of mitigating adverse influences (chapters 6, 7, 13, 17, 19, 25, 29, 30, 31, 32, 33, 34, 35); 10 Evaluated the utility of top predators for providing information useful for fisheries management in the BCLME (chapter 36, 37); 11 Investigated the usefulness of top predators in providing indices of ecosystem health in the BCLME (chapter 38); 12 Initiated new time series of information for top predators in the northern part of the BCLME region (chapters 39, 40, 41); 13 Tested the potential for use of satellite transmitters deployed on cormorants in the ecosystem monitoring programme (chapter 39) (such instruments had already been successfully deployed on seals, penguins and gannets); 14 Re-assessed the conservation status of land-breeding top predators in the BCLME (chapters 28, 42); 15 Promoted regional collaboration in monitoring of landbreeding top predators (chapter 39, 40, 41), as well as linkages with monitoring of top predators in the similar Humboldt system (chapter 43); 16 Identified those parameters for land-based top predators that are required to attain the objectives of the monitoring programme (chapters 46, 47) and described the most appropriate monitoring methods to be applied for monitoring (chapter 9, annexes 2, 3); 17 Formulated recommendations for an integrated ecosystem monitoring programme in the BCLME region, based on land-breeding top predators, and guidelines for interpreting data from the monitoring programme and incorporating its results in management at the ecosystem level (chapters 46, 47);

4 Considered the impact of habitat on the breeding success of African Penguins (chapter 12);

18 Provided experience for young scientists (chapter 44) as well as training in time-series analysis (chapter 45) and;

5 Considered the effect of age and breeding status on the moult phenology of African Penguins in Namibia (chapter 14);

19 Considered available data bases relating to landbreeding top predators and made recommendations for maintaining a shared regional database (chapters 4, 5, 46, 47);

6 Considered variability in foraging behaviour and chick growth of Cape Gannets (chapter 18); 7 Considered the energetic requirements of two species of seabirds in the BCLME (chapters 15, 23);

20 Resulted in 12 publications that appeared in peerreviewed journals and 19 additional manuscripts that it is hoped will be published (chapter 48). Top Predators of the Benguela System

15

16


Chapter 3 Objectives for an ecosystem monitoring programme in the BCLME based on land-breeding top predators BCLME Top Predators Project Steering Committee

The project requires the objectives for an ecosystem monitoring programme based on land-breeding top predators in the BCLME to be addressed. The major land-breeding top predators in the BCLME are Cape Fur Seals Arctocephalus pusillus and a suite of seabirds. Objectives for ecosystem monitoring programmes in the BCLME are likely to be influenced by the objectives of states falling within the BCLME region (Angola, Namibia, South Africa) regarding the management of their marine resources. Objectives for South Africa are listed in section 2 of its Marine Living Resources Act (No. 18 of 1998). These include: “(a) The need to achieve … ecologically sustainable development of marine living resources; (b) the need to conserve marine living resources …; (c) the need to apply precautionary approaches in respect of the management … of marine living resources; (d) the need to … achieve … a sound ecological balance …; (e) the need to protect the ecosystem as a whole, including species which are not targeted for exploitation; (f) the need to preserve biodiversity …” (Government Gazette Vol. 395, No. 18930). Objectives are also likely to be influenced by international declarations or agreements. For example, in 2002 at Johannesburg, the World Summit on Sustainable Development adopted a Plan of Implementation that agreed (Article 29d) to “encourage the application by 2010 of the ecosystem approach, noting the Reykjavik Declaration on Responsible Fisheries in the Marine Ecosystem” (FAO 2003). The 2001 Reykjavik Declaration had called for the incorporation into fisheries management of ecosystem considerations “such as predator–prey relationships” (http://www.fao.org/PDF/acroe.htm). Earlier, at its Twenty-eighth Session on 31 October 1995, the FAO Conference adopted the FAO Code of Conduct for Responsible Fisheries. Article 6.2 of this Code states “Management measures should not only ensure the conservation of target species but also of species belonging to the same ecosystem or associated with or dependent upon the target species.” Article 6.5 reads “States and subregional and regional fisheries management organizations should apply a precautionary approach widely to conservation, management and exploitation of living aquatic resources in order to protect them and preserve the aquatic environment, taking

account of the best scientific evidence available. The absence of adequate scientific information should not be used as a reason for postponing or failing to take measures to conserve target species, associated or dependent species and non-target species and their environment” (http:// www.fao.org/DOCREP/ 005/v9878e/v9878e00.htm). The 1982 United Nations Convention of the Law of the Sea requires that “coastal states shall take into consideration the effects on species associated with or dependent upon harvested resources with a view to maintaining or restoring populations of such associated or dependent species above levels at which their reproduction may become seriously threatened” (FAO 2003). The CCAMLR (Commission for the Conservation of Antarctic Marine Living Resources) Ecosystem Monitoring Programme (CEMP) has as objectives: “(ii) the maintenance of ecological relationships between harvested, dependent and related populations of Antarctic marine living resources and the restoration of depleted populations …” (iii) the prevention of changes or minimization of the risk of changes in the marine ecosystem which are not reversible over two or three decades …” In essence CEMP attempts to monitor food availability to predators (CCAMLR 1991). Clearly, both in the South African context and international declarations, objectives for fisheries management include accounting for the requirements of predators dependent on species targeted by fisheries. Monitoring of top predators may seek to advise how this might be accomplished. In the Benguela system some predators may provide useful indices of the abundance of stocks of prey that are exploited by commercial fisheries (e.g. Berruti et al. 1993, Crawford 2003). They may also forecast recruitment to fisheries (Roux in preparation), provide information on changes in natural mortality rates of fished resources (Crawford et al. 1992) and changes in the distribution of fished resources (e.g. Crawford 1998, Crawford et al. 2002). Hence they may provide information useful in the management of prey resources. Additionally, because of their position at the apex of the food chain and susceptibility to factors such as pollution, top predators have the potential to provide indices of the state

Top Predators of the Benguela Benguela System

17 17

Objective

Population size

Breeding productivity

Diet

Account for food requirements of predators

Yes

Yes

Yes

Information for management of prey resources

Yes

Yes

Yes

Indices of ecosystem health

Yes

Yes

Monitor and update conservation status of predators

Yes

Assess efficacy of conservation interventions

Yes

Manage species interactions

Yes

of health of marine ecosystems (e.g. Underhill and Crawford 2005). Several of the top predators in the Benguela ecosystem are classified as Threatened or Near Threatened in terms of criteria of the IUCN (World Conservation Union). It is necessary to monitor the conservation status of species of conservation concern. Indeed, trends in the abundance of species are frequently used to assess and update their conservation status (e.g. Barnes 2000, BirdLife International 2000). Monitoring is useful to assess the outcomes of conservation interventions (e.g. Crawford et al. 2000, Barham et al. 2006). Abundant predators may influence the conservation status of less numerous predators and it may be necessary to manage interactions between species (e.g. Crawford et al. 1989, David et al. 2003). Time series of information that may prove useful in attaining the above objectives include information on: – Population sizes (per species per breeding colony) – Breeding productivity (per species for selected breeding localities, including breeding success and growth) – Diet (per species for selected breeding localities) – Abundance of prey – Telemetry/GPS instruments and banding (for selected species) providing information on foraging areas and dispersal. Those time series groups thought to be necessary to attain the different objectives are indicated in the table above. References BARHAM PJ, CRAWFORD RJM, UNDERHILL LG, WOLFAARDT AC, BARHAM BJ, DYER BM, LESHORO TM, MEŸER MA, NAVARRO RA, OSCHADLEUS D, UPFOLD L, WHITTINGTON PA, WILLIAMS AJ (2006) Return to Robben Island of African Penguins that were rehabilitated, relocated or reared in captivity following the Treasure oil spill of 2000. Ostrich. 77: 202–209. BARNES KN (ed) (2000) The Eskom Red Data Book of Birds of South Africa, Lesotho and Swaziland. BirdLife South Africa, Randburg.

18


Prey abundance

Instruments and banding

Yes

Yes Yes

BERRUTI A, UNDERHILL LG, SHELTON PA, MOLONEY C, CRAWFORD RJM (1993) Seasonal and interannual variation in the diet of two colonies of the Cape gannet (Morus capensis) between 1977–78 and 1989. Colonial Waterbirds 16: 158–175. BIRDLIFE INTERNATIONAL (2000) Threatened Birds of the World. Lynx Edicions and BirdLife International, Barcelona and Cambridge, UK. CCAMLR (1991) CCAMLR Ecosystem Monitoring Program. CCAMLR, Hobart, Australia. CRAWFORD RJM (1998) Responses of African penguins to regime changes of sardine and anchovy in the Benguela system. South African Journal of Marine Science 19: 355–364. CRAWFORD RJM (2003) Population, and size and location, of colonies, of swift terns, in South Africa’s Western Cape, 1987–2000. Waterbirds 26(1): 44–53. CRAWFORD RJM, COOPER J, DYER BM, UPFOLD L, VENTER AD, WHITTINGTON PA (2002) Longevity, inter-colony movements and breeding of crested terns in South Africa. Emu 102: 1–9. CRAWFORD RJM, DAVID JHM, WILLIAMS AJ, DYER BM (1989) Competition for space: recolonising seals displace endangered, endemic seabirds off Namibia. Biological Conservation 48: 59– 72. CRAWFORD RJM, DAVIS SA, HARDING R, JACKSON LF, LESHORO TM, MËYER MA, RANDALL RM, UNDERHILL LG, UPFOLD L, VAN DALSEN AP, VAN DER MERWE E, WHITTINGTON PA, WILLIAMS AJ, WOLFAARDT AC (2000) Initial impact of the Treasure oil spill on seabirds off western South Africa. South African Journal of Marine Science 22: 157– 176. CRAWFORD RJM, UNDERHILL LG, RAUBENHEIMER CM, DYER BM, MARTIN J (1992) Top predators in the Benguela ecosystem – implications of their trophic position. South African Journal of Marine Science 12: 675–687. DAVID JHM, CURY P, CRAWFORD RJM, RANDALL RM, UNDERHILL LG, MËYER MA (2003) Assessing conservation priorities in the Benguela ecosystem: analysing predation by seals on threatened seabirds. Biological Conservation 114: 289– 292. FAO (2003) Fisheries management 2. The ecosystems approach to fisheries. FAO Technical Guidelines for Responsible Fisheries 4 (Suppl. 2): 1–112. UNDERHILL LG, CRAWFORD RJM (2005) Indexing the health of the environment for breeding seabirds in the Benguela system. ICES Journal of Marine Science 62: 360–365.

Collation of time series

Chapter 4 Time series of data for Cape Fur Seals in the BCLME S.P. Kirkman1, 2 1

Avian Demography Unit, Department of Statistical Sciences, University of Cape Town, Rondebosch 7701, South Africa Marine and Coastal Management, Department of Environmental Affairs and Tourism, Private Bag X2, Roggebaai 8012, South Africa

2

Introduction This report contains an inventory of time series of Cape Fur Seal data that are available in South Africa and Namibia. The chapters and sections in the report correspond with the relevant chapters and sections in the Manual of Methods for Monitoring Cape Fur Seals, annexed to this report. Details of study objectives and techniques of data collection, are given in the manual. Ownership of data is given in square brackets below each inventory table. The following definitions apply: MCM: MFMR:

Marine and Coastal Management (South Africa) Ministry of Fisheries and Marine Resources (Namibia)

Where applicable, the name of the database where the data is stored is given in brackets.

1. Monitoring the diet of Cape Fur Seals 1.1

Background

Most studies of Cape Fur Seal diet have followed one of two approaches – analysis of stomach contents of seals sampled at sea, or analysis of feces (scats) collected at colonies. Both techniques are subject to many limitations and biases. In recent years, analysis of scats collected at colonies has been the preferred method of investigating fur seal diet. This technique is cheap and practical relative to the at-sea sampling of seals for dietary information, and can provide important information on spatial and temporal trends in the relative consumption of the main prey species (Tollit and Thompson 1996), and quantitative information which can be useful for the management of fish stocks (Mecenero 2005). 1.2 Scat sample analysis

Queries regarding MCM data should be directed to: The Deputy Director Sub-directorate Ecosystem Utilization and Conservation Department of Environmental Affairs and Tourism Branch: Marine and Coastal Management Private Bag X2 Roggebaai 8012 South Africa Tel: +27 21 4023114, Fax: +27 21 4023639 Queries regarding MFMR data should be directed to: Head of Section Marine Mammal Section National Marine Information and Research Center Ministry of Fisheries and Marine Resources P.O. Box 912 Swakopmund Namibia Tel: +264 64 4101000, Fax: +264 64 404385

Scat samples were collected at three mainland colonies in Namibia, namely Atlas Bay (including Wolf Bay), Cape Cross and van Reenen’s Bay, since January 1994. Sampling was monthly, but there were many months, especially at Cape Cross and van Reenen’s Bay, when no samples were collected. The distribution of sampling at the three colonies is shown in Table 1. 1.3 Stomach sample analysis Dedicated cruises for the pelagic sampling of seals commenced in 1974. Thirty cruises took place from then to 2001. There were no cruises in some years, and multiple cruises in others. There was variation between cruises in timing, geographical distribution, sample size and sex ratio of the samples (Table 2). Stomach contents, as well as morphometric and reproductive data, were collected from sampled animals.

Top Top Predators Predators of of the the Benguela Benguela System System

21 21

2. Estimating Cape Fur Seal pup numbers 2.1 Background Assessments of the size and trend of the Cape Fur Seal population, or parts thereof, have been based on pup numbers, for various reasons:

mated (Reid 2002). Changes in pup mass with time can be determined either via longitudinal sampling, whereby the same individuals are weighed repeatedly over a period of time, or cross sectional sampling, whereby a random sample of different individuals in the population are weighed at pre-defined intervals. 3.2 Birth mass

c Pups are the only demographic category that are all confined to land simultaneously (during their first few weeks of life); c Their small size and the black pelage of their first ten weeks of life permit them to be easily distinguished from other age classes; c Their small size permits them to be physically restrained with relative ease. Two techniques have been used to determine pup numbers, namely tag–recapture and aerial censusing. The aerial census method is judged to be the most practical and cost-effective method of surveying the seal population (Wickens et al. 1991), although Shaughnessy (1993) considered tag– recapture to be more accurate. 2.2 Aerial census Pup counts from aerial photographs, taken during aerial surveys between 1972 and 2004, are shown in Table 3. Surveys are conducted (with noted exceptions) between 17 and 22 December. However, the year by which each census is named is the year in which the pup cohort was weaned, ie, the following year. Comprehensive surveys, where all or nearly all of the major colonies were photographed, were conducted during 13 of the 32 years. Counts of the most recent censuses (2005–2007) are not complete, to date. 2.3 Tag–recapture The tag–recapture method provides an estimate of the number of pups in the colony at the time of tagging (~20 January). Different estimates of pup numbers have been obtained, based on the time of recapture (Table 4). 3. Mass at birth, growth of Cape Fur Seal pups 3.1 Background Annual changes in the average mass of fur seal pups at birth is considered to be a good indicator of fetal growth in the last few months of gestation, and thus a possible indicator of food availability in the area where cows forage before parturition (Reid 2002). During the lactation period, mass gains by pups depend on provisioning by the mothers, therefore the growth rates of pups reflect the rate of energy transfer from mothers to pups, and provide a good indication of prey availability (Beauplet et al. 2004). In addition, because fur seal mothers are central place foragers during the lactation period, the growth rates of pups should reflect conditions whose spatial and temporal boundaries can be readily determined or esti-

22


Pups have been weighed soon after birth at Atlas Bay since the 1989–1990 breeding season (Table 5). 3.3 Longitudinal sampling In some years, it has been attempted to measure the growth of pups, tagged and weighed soon after birth, by serially reweighing them. The method becomes exceedingly difficult after pups are about two months old. Listed in Table 6 are the sample sizes of male and female pups that were serially weighed until they were approximately two months old, and for which three or more data points exist. 3.4 Cross sectional sampling Cross-samples of pup mass have been collected at Atlas Bay (including Wolf Bay), Cape Cross and van Reenen’s Bay, between January and the age of weaning (Table 7). Dead pups were also weighed at the factories, following harvesting, in many years. References Beauplet G, Guinet C, Arnould JPY (2003) Body composition changes, metabolic fuel use, and energy expenditure during extended fasting in subantarctic fur seal (Arctocephalus tropicalis) pups at Amsterdam Island. Physiological and Biochemical Zoology 76: 262–270. Kirkman SP, Oosthuizen WH, Meÿer MA, Kotze PGH, Roux J-P, Underhill LG (2007) Making sense out of censuses and dealing with missing data: trends in pup counts of Cape Fur Seals Arctocephalus pusillus pusillus for the period 1972–2004. African Journal of Marine Science 29(2): 161–176 Mecenero S (2005) The diet of the Cape fur seal Arctocephalus pusillus pusillus in Namibia: variability and fishery interactions. PhD Thesis, Department of Statistical Sciences, University of Cape Town, Cape Town. 219 pp. Reid K (2002) Growth rates of Antarctic fur seals as indices of environmental conditions. Marine Mammal Science 18: 469–482. Shaughnessy PD (1987) Population size of the Cape fur seal Arctocephalus pusillus. 1. From aerial photography. Investigational Report of the Sea Fisheries Research Institute, South Africa 130:1–56. Shaughnessy, PD (1993) Population size of the Cape fur seal Arctocephalus pusillus. 2. From tagging and recapturing. Investigational Report of the Sea Fisheries Research Institute, South Africa 134: 1–70. Tollit DJ, Thompson PM (1996) Seasonal and between-year variations in the diet of harbour seals in the Moray Firth, Scotland. Canadian Journal of Zoology 74: 1110–1121. Wickens PA, David JHM, Shelton PA, Field JG (1991) Trends in harvests and pup numbers of the South African fur seal: implications for management. South African Journal of Marine Science 11: 307–326.


23 23

Month

1 2 3 4 5 6 7 8 9 10 11 12

1 2 3 4 5 6 7 8 9 10 11 12

1 2 3 4 5 6 7 8 9 10 11 12

Year

1994

1995

1996

17 16 19 17 21 17 16 9 19 17 16 17

20 27 12 25 23 14 18 23 18 19 20 20

30

7 6 10 7 11

15 7

AWB

18 29 27 15 9 26 16 21 10 24 28 1

25 30 19 26 16 1

25 25 27 27 26

23 6 29 1

4 11 19 22 27

CC

Day

(continued . . .)

9

13

12

12

12

12

8 8 8 9

15

VRB

1999

1998

1997

Year

1 2 3 4 5 6 7 8 9 10 11 12

1 2 3 4 5 6 7 8 9 10 11 12

1 2 3 4 5 6 7 8 9 10 11 12

Month

Table 1: Dates on which scats were collected at 3 mainland colonies in Namibia, since 1994

31

22 19 19 19 23 17 21 30 27 27

5

14 23 10 17 13 22 17 21 18

20 2 3 4 5 5 1 5 5 6 27 18

AWB

27 15 8

2 22 27

21 23 21 14

15 6 23 23 16 9

20 18 19 24 3 27

28 18 20 23

CC

Day

(continued . . .)

29 12 24 14 31

12

12

24

20

27 20

13

26 18 17 16

VRB

1 2 3 4

2002

18 20 19 19

17 21 20 21 23 21 19 21 25 25 25 25

19 27 19 28 20 28 21

18 30 20

AWB

23 20 19 19

5 17 21 20 17 21 22 20 20

28

15 13 8

16 7

CC

Day

19 22 21 24

22 19 24 18 21 21 17 19 21 18 1 1

10

VRB

[MFMR] {AP_diet from scats database – Access database file}

1 2 3 4 5 6 7 8 9 10 11 12

1 2 3 4 5 6 7 8 9 10 11 12

Month

2001

2000

Year

24


1974 1975 1975 1977 1977 1979 1980 1981 1982 1984 1985 1986 1987 1988 1989 1989 1989 1989 1989 1989 1990 1990 1990 1992 1992 1993 1994 1995 1996 2001

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

Apr Mar Aug Mar Oct Mar Aug Apr Sep Oct Apr May Jul Jan Apr Feb Apr Jun Aug Oct Jan Mar Nov May Nov May Jun Oct Jun May

Month

Cape Town Cape Town Orange River Ichaboe Island Conception Bay Cape Cross Kleinzee Cape Town Cape Town Cape Town Cape Town St Helena Bay Walvis Bay Cape Town Cape Town Cape Town Cape Town Cape Town Cape Town Cape Town Cape Town Cape Town Cape Town Cape Infante Cape Town Hout Bay Quoin Point Cape Point Cape Point Bok Punt

Start point

[MCM] {AP_pelagic sampling database – Access database file}

Year

Cruise no. Dyer Island Algoa Bay Ichaboe Island Conception Bay Cape Cross Mowe Bay Ichaboe Island Orange River Port Nolloth Port Nolloth Algoa Bay Walvis Bay Kunene River Buchu Twins Lamberts Bay Lamberts Bay Lamberts Bay Lamberts Bay Lamberts Bay Lamberts Bay Lamberts Bay Lamberts Bay Lamberts Bay Port Nolloth Cape Town Kleinzee Port Nolloth Mossel Bay Port Nolloth Port Nolloth

End point 9.0 1.0 1.0 2.0 2.0 1.0 1.0 6.0 0.5 0.5 1.0 1.0 1.0 1.1 0.5 0.7 2.0 1.5 4.2 1.6 3.0 4.0 1.3 1.7 8.0 2.1 1.4 4.8 3.0 1.5

Closest inshore (km) 11.0 17.0 26.0 23.0 15.0 14.0 35.0 53.0 3.0 3.0 7.0 37.0 15.0 18.0 33.0 35.0 14.0 18.0 41.0 48.0 41.0 42.0 12.0 38.0 25.0 30.8 47.0 11.2 30.0 20.0

Furthest offshore (km) 3 14 35 45 57 61 14 54 58 94 43 62 34 54 18 9 17 11 12 29 18 14 26 72 17 22 61 4 48 41

n Females

1 6 18 62 49 42 1 29 38 81 55 65 79 39 38 26 31 25 27 23 42 40 9 55 5 24 28 13 19 47

n Males

Table 2: Details of all stomach-sampling cruises from 1974 to 2001. Sample sizes exclude animals with empty stomachs, and animals estimated to be less than one year old. Month denotes the month in which the greater part of sampling occurred


25 25

14 449a

1971

2680 3746 3237 1703

30 450 2496 4808 2427 0

3243 15 772 2769

0 17 839 5042 0 755 2910 1691 2875 7443b 8879b 12 228 3722

*1972

27 776 1095 3376

1973

8805a 23 295a

1974

1730 1262 904

1975

4952

1629

1976

1978

12 199 6638 1090 1176 86 765 957

12 297

22 097 2772 0 378 258 2114 2472 978 971 779 920 15 017b 23 759b 36 453 9840 2393 205 3208a 10 879 9461 3248 0 52 075 1398 3772 1273

1977

17 852 55 852 13 361

0

1979

8188 4099 630 380 442

473

59 165 1826

12 252 4632 278 3591 11 370

1236 528 616

0

16 327

1980

1981

8574 6137

1982

0

10 017 9151 1074 899 561

2748 3132 929

1167 883 26 669 61 438 13 223 5254 216 4953 9419 1614

1 26 623 1945 0

1983

[South African colonies – MCM; Namibian colonies – MCM in conjunction with MFMR] {AP_census database – Access database file}

1984

6954

83 469

* 1972 photographs taken on 4–6 December, therefore values are adjusted by a factor of 1.5539 a Considered undercounts because photographs taken in January (Shaughnessy 1987) b Considered undercounts because unrealistic in comparison with harvest figures (possibly due to incomplete coverage of colonies) c Aerial surveys preceded by large bull harvests that resulted in breeding disturbance, thereby probably reducing pup production (Wickens et al. 1991) d Breeding population deliberately disturbed during breeding season therefore pup production reduced (Wickens et al. 1991) e Considered undercount due to high altitude of pgotographs f Missing areas filled in proportionally using 1999 counts of the same colony g Missing areas filled in using 2001 counts of the same colony h Digital photographs used

Cape Frio Cape Cross Hollams Bird Island Mercury Island Marshall Reef Staple Rock Boat Bay Rock Dumfudgeon Rock Wolf Bay Atlas Bay Long Island Albatross Rock Black Rock van Reenen Bay Sinclair Island Lion’s Head Buchu Twins Kleinsee Elephant Rock Jacob’s Reef Robbesteen Duikerklip Seal Island, FB Geyser Rock Quoin Rock Seal Is, MB Black Rocks, AB Pelican Point Sandwich Harbour Klein Ichaboe Bird Island, LB Paternoster Rock

Colony

Table 3: Pup counts from aerial surveys (1972–2004), taken from Kirkman et al. (2007). The underlined years are “full-census” years

0

3270 1515 11 11 010 8345

1792 45

407 6701

1128

1985

3828

1987

3331 202 393 4820 6900 8011 1817 19 38 43 267c 47 113c 2612 1086 1616 1368 22 75 12 116 5218d 9584 8643 1644 1496 1170 808 0 672 27 10 8 127 74

3606 398 2212 1618 1623 10 616c 16 860c 12 812

3 35 590

1986

8 943

46 850c 3740 1971 1575 15 14 105 10187 1756

491 6235

29 454 43 923

5178

37 882

1988

800 0 4 6 0 1098

55 247 5216 3114 666 2325 2066 2003 24 548 42 223 22 160 4354 439 5590 11 139 3437 29 74 620 3326 3886 1224 6 13 503 12793 2041

1989

26


d–f

See preceding page

Cape Frio Cape Cross Hollams Bird Island Mercury Island Marshall Reef Staple Rock Boat Bay Rock Dumfudgeon Rock Wolf Bay Atlas Bay Long Island Albatross Rock Black Rock van Reenen Bay Sinclair Island Lion’s Head Buchu Twins Kleinsee Elephant Rock Jacob’s Reef Robbesteen Duikerklip Seal Island, FB Geyser Rock Quoin Rock Seal Island, MB Black Rocks, AB Pelican Point Sandwich Harbour Klein Ichaboe Bird Island, LB Paternoster Rock Cape Columbine Dolphin Head Conception Bay Baie dos Tigros Mowe Bay Torra Bay Toscanini Sylvia Hill Cliff Point Jutten Island Vondeling Island

Colony

35

0 18 17 3 1527

18 15 484 10 749 1676 1238

16 78 809

4990

19 286 41 607

51 890 3267 460d

1990

7 1697

1487 14 13 898 9651

63 246 3476

461 5232

552

384 2314

44 636

1991

0 1877

79 301 3841 3606 1722 11 17 522 11 522 2367

25 680

65 557

1992 477 61 891 4902 35 942 1405 1240 1667 39 534 62 823 20 170 1715 200 5293 8703 6121 107 72 203 2193 1265 964 8 12 974 11 743 1834 754 463 3 292 68 14 758 107 2500 231

1993

768 112

75

15 235 10 324 1694

3813

28 476

1994

1576

24 502 63 66 2476 42

59 370 2110 1566 707 9 17 144 11 616 2080

96 42 17 031 8809 2031 252 3011 6992 4501

3044 29 990 961 0

1995

12 756 157 352 1932 96 141 693

13 528 11 939 1520

69 669 3092 2221 976

9158 13 581 7751 1152 210 3317 7967 4623

4337 35 498 2026 0

1996

19 396 12 266 1639 989 296 0 26 89 314 3233 91 158 0

69 930 4074 2064 1212

12 098 2451 100 3989 7186 5529

4419 38 564 1827 0 242 1542 693

1997

16 806 11 184 1779 691 142 0 33 128 40 1200 57 0 52

87 841 2165 1650 1155

7191 48 993 3478 0 146 1899 883 465 36 700 46 225 14 835 2785 206 5783 10771 8308

1998

1999

2000

91 641

2001

Table 3: continued. The underlined years are “full-census” years. “Y” denotes that counting of pups on the aerial photographs has yet to be completed

0

130 1014

78 12 298 5659e 1223 658 505

79 710 4293

10 880 37 394 2285 0 106 1462 669 1099 15 184 18 193 7822 1335 84 2953 7472 6163

2002

2003

612

592 908

71

423

18 339h

80 897g 4398 3376h 908

10 543 9603

259

29 531 45 155f 12 648

16 608 54 546 2305 0

2004

Y Y

Y Y Y Y

Y Y Y Y

Y Y

Y Y

2005

Y Y Y Y Y Y Y

Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y

Y Y Y Y Y Y Y Y Y Y Y Y

Y Y Y

2006

Y Y

Y

Y

Y

Y

Y

Y

Y Y Y

Y Y

2007

Table 4: Estimates of pup numbers from tag–recapture experiments, based on recaptures in January or at harvesting* Colony Cape Cross

Wolf Bay

Atlas Bay

Kleinsee

Marshall Reef

Recapture

Year

Estimate

Colony

January January Harvest Harvest

1980 1986 1980 1986

18 27 20 35

Staple Rock

260 583 036 541

January January Harvest Harvest

1979 1985 1974 1979

30 30 21 29

278 174 655 462

January January January January Harvest Harvest Harvest

1974 1979 1985 1989 1974 1979 1985

37 54 59 62 52 74 65

931 151 713 493 286 609 861

January January January January January January January January January Harvest Harvest Harvest

1973 1981 1984 1987 1991 1993 1995 1998 2002 1973 1981 1984

30 62 80 33 49 61 71 75 69 37 78 99

006 535 902 756 135 341 895 870 675 719 649 914

January

1978

Recapture

Year

Estimate

January Harvest

1978 1978

2043 2376

Boat Bay Rock

January

1978

870

Dumfudgeon Rock

January Harvest

1978 1978

661 872

Long Islands

January January Harvest

1977 1983 1977

15 155 18 025 20 809

Albatross Rocks


1977 1983 1977

5586 6955 5577

Sinclair Island


1978 1983 1978

11 931 12 589 15 083

Elephant Rock

January Harvest

1976 1976

1196 1548

Seal Island (FB)


1971 1982 1971

1196 9611 14 072

Geyser Rock

January January

1976 1982

6533 8163

Quoin Rock

January Harvest

1975 1975

2292 1740

Seal Island (MB)

January Harvest

1975 1975

2095 2552

248

* As tag-induced mortality is apparently displayed more strongly in females than in males, the harvest estimate is based on recapture of male pups only. [MCM] {AP_census database – Access database file}

Table 5: Sample sizes of male and female pups weighed shortly after birth at Atlas Bay or Wolf Bay Season 1989–1990 1990–1991 1991–1992 1992–1993 1993–1994 1994–1995 1995–1996 1996–1997 1997–1998 1998–1999 1999–2000 2000–2001 2001–2002 2002–2003 2003–2004 2004–2005

Number of females

Number of males

340 341 45

338 308 65

24 69 163 377 101 576 246 327 287 195 385 615

27 62 120 364 181 585 279 260 289 199 405 532

Table 6: Sample sizes of male and females pups measured at Atlas Bay for longitudinal growth, between birth and 60 (50–70) days of age, with three or more growth points Year

Number of females

Number of males

1997 1999 2001 2002

42 25 41 19

83 36 54 44

[MFMR] {AP_pup mass and growth – Access database file}



27 27

Table 7.1: Sample sizes of male and female pups weighed during cross-samples, at Atlas Bay or Wolf Bay (A = alive; D = dead, ie, harvested) Date

Julian days

No. No. State Date males females (A/D)

13 14 18 08 24 10 11 14

Apr 87 May 87 Jun 87 Jul 87 Jul 87 Sep 87 Sep 87 Sep 87

103 134 169 189 205 253 254 257

37 36 38 23 23 23 13 10

30 18 26 17 18 14 8 11

A A A A A ? ? ?

20 10 26 18 28 13 15 05 19 10 12 14 28 29 12

Jan 88 Feb 88 Feb 88 Mar 88 Mar 88 Apr 88 Apr 88 May 88 May 88 Jun 88 Jul 88 Jul 88 Jul 88 Jul 88 Aug 88

20 41 57 78 88 104 106 126 140 162 194 196 210 211 225

37 39 33 32 42 42 52 49 35 35 42 38 39 41 25

24 28 26 28 28 32 30 29 27 25 28 27 29 30 28

A A A A A A A A A A A A A A ?

19 27 15 25 13 04 04

Jan 89 Feb 89 Mar 89 Apr 89 Jun 89 Jul 89 Aug 89

19 58 74 115 164 185 216

120 60 54 55 44 41 47

121 30 34 33 30 29 31

A A A A A A ?

19 19 16 18 08 18 27

Jan 90 Feb 90 Mar 90 Apr 90 May 90 Jun 90 Jul 90

19 50 75 108 128 169 208

89 61 58 46 41 54 53

60 38 34 33 37 36 53

A A A A A A A

18 21 15 18 03 23 20

Jan 91 Feb 91 Mar 91 Apr 91 Jun 91 Jul 91 Aug 91

18 52 74 108 154 204 232

68 74 53 52 54 46 51

55 40 33 46 53 34 34

A A A A A A ?

18 19 24 26 07

Feb 92 Mar 92 Apr 92 May 92 Jul 92

49 79 115 147 189

64 85 46 66 55

37 32 47 38 46

A A A A A

19 23 12 21 18 09

Jan 93 Feb 93 Mar 93 Apr 93 May 93 Aug 93

19 54 71 111 138 221

73 62 73 35 71 31

55 50 31 26 64 35

A A A A A ?

18 24 21 13 11 12 13 17

Jan 94 Feb 94 Apr 94 May 94 Jun 94 Jun 94 Jun 94 Jun 94

18 55 111 133 162 163 164 168

172 47 68 42 2 1 3 49

102 42 48 43 5 2 3 57

A A A A A A A A

28

Julian days

No. No. State Date males females (A/D)

20 Jan 95 17 Feb 95 07 Mar 95 15 Mar 95 02 May 95 10 May 95 15 Jun 95 18 Jul 95 10 Aug 95 14 Aug 95

20 48 66 74 122 130 166 199 222 226

66 19 42 6 35 6 37 44 162 157

54 9 31 1 48 6 50 40 187 209

A A A A A A A A ? ?

18 Jan 96 20 Mar 96 24 May 96 24 Jun 96 1 Jul 96 2 Jul 96 1 Aug 96 14 Aug 96 15 Aug 96 16 Aug 96 19 Aug 96 20 Aug 96

18 80 145 176 183 184 214 227 228 229 232 233

115 65 66 56 46 5 47 84 95 76 97 82

97 52 56 59 42 2 45 105 83 77 114 97

A A A A A A ? ? ? ? ? ?

21 19 24 25 20 27 28

Jan 97 Feb 97 Mar 97 Mar 97 Aug 97 Aug 97 Aug 97

21 50 83 84 232 239 240

106 49 8 10 65 61 26

105 25 18 8 83 47 24

A A A A ? ? ?

9 Jan 98 20 Jan 98 17 Feb 98 5 Mar 98 18 Mar 98 16 Apr 98 7 May 98 18 May 98 18 Jun 98 7 Jul 98 10 Aug 98 11 Aug 98 12 Aug 98 13 Aug 98 1 Oct 98

9 20 48 64 77 106 127 138 169 188 222 223 224 225 274

172 233 75 51 47 70 51 54 47 20 63 73 66 67 81

94 140 38 41 21 54 28 33 38 20 54 82 51 48 70

A A A A A A A A A A D D D D D

7 Jan 99 8 Jan 99 18 Jan 99 20 Jan 99 1 Feb 99 9 Feb 99 16 Feb 99 8 Mar 99 30 Mar 99 1 Apr 99 15 Apr 99 16 Apr 99 12 May 99 17 May 99 16 Jun 99 17 Jun 99 9 Aug 99 10 Aug 99 11 Aug 99 12 Aug 99 18 Aug 99

6 8 18 20 32 40 47 67 89 91 105 106 132 137 167 168 221 222 223 224 230

49 60 51 141 28 50 15 60 51 41 51 44 47 46 100 64 160 86 126 135 40

49 46 52 134 21 40 7 57 56 40 50 34 45 41 98 71 159 70 138 102 37

A A A A A A A A A A A A A A A A D D D D A


Julian days

No. No. State males females (A/D)

7 Sep 99 8 Sep 99

250 251

105 29

95 45

D A

5 Jan 00 6 Jan 00 20 Jan 00 10 Feb 00 28 Feb 00 23 Mar 00 19 Apr 00 30 Jun 00 1 Aug 00 21 Aug 00 22 Aug 00

5 6 20 41 59 83 110 182 214 234 235

61 79 108 83 56 54 50 42 116 62 83

39 41 64 50 57 54 32 39 75 70 76

A A A A A A A A D D D

6 Jan 01 9 Jan 01 20 Jan 01 8 Feb 01 23 Feb 01 19 Mar 01 12 Apr 01 31 May 01 22 Jun 01 23 Jul 01 25 Jul 01 25 Jul 01 26 Jul 01 31 Jul 01 31 Jul 01 1 Aug 01 2 Aug 01 3 Aug 01 6 Aug 01 7 Aug 01 17 Aug 01 20 Aug 01 24 Aug 01 29 Aug 01 3 Sep 01 7 Sep 01 13 Sep 01 17 Sep 01

6 9 20 39 54 78 102 151 173 204 206 206 207 211 211 212 213 213 216 217 227 230 234 239 244 248 254 258

84 67 104 21 59 54 60 52 49 54 61 17 55 50 19 15 42 16 42 11 20 55 20 24 58 22 22 28

67 62 96 23 44 47 40 56 57 41 41 15 47 49 23 15 61 20 56 12 14 46 10 26 43 10 6 27

A A A A A A A A A D D A D D A A D A D A A D A D D A A D

4 Jan 02 18 Jan 02 8 Feb 02 4 Mar 02 27 Mar 02 19 Apr 02 21 May 02 26 Jun 02 19 Jul 02 26 Jul 02

4 18 39 63 86 109 145 177 201 207

61 158 59 64 73 52 53 57 60 51

50 121 47 36 41 49 47 44 45 41

A A A A A A A A D A

20 Jan 03 21 Jan 03 18 Mar 03 3 Jun 03 7 Jul 03 9 Jul 03 14 Jul 03 15 Jul 03 17 Jul 03 21 Jul 03 23 Jul 03 25 Jul 03 28 Jul 03 30 Jul 03

20 21 77 154 188 190 195 196 198 202 204 206 209 211

193 108 62 53 87 74 85 32 56 59 67 119 78 85

169 59 36 18 75 73 72 25 42 64 55 64 65 45

A A A A D D D D D D D D D D

(Table 7.1: continued) Date

Julian days

Table 7.3: Sample sizes of male and female pups weighed during cross samples, at Cape Cross (A = alive, D = dead, ie harvested)

No. No. State males females (A/D)

6 Aug 03 7 Aug 03 13 Aug 03 20 Aug 03 21 Aug 03

218 219 225 232 233

48 65 52 64 57

54 41 53 39 42

D D D D D

7 Jan 04 9 Jan 04 11 Jan 04 19 Jan 02 18 Feb 04 2 Mar 04 8 Apr 04 6 May 04 05 Jun 04 22 Jun 04 5 Jul 04 9 Jul 04 13 Jul 04 3 Sep 04 8 Sep 04

7 9 11 19 49 62 125 153 184 200 213 217 221 273 278

62 54 72 64 62 55 52 76 68 68 124 66 100 114 113

40 47 32 38 34 55 36 58 47 65 105 74 102 101 108

A A A A A A A A A A D D D D D

20 Jan 05 31 Mar 05 26 Apr 05

46 116 142

121 72 51

88 35 63

A A A

Date


Table 7.2: Sample sizes of male and female pups weighed during cross samples, at van Reenen’s Bay (A = alive, D = dead; ie, harvested) Date 27 20 22 12 14 22 19

Jan 98 May 98 Jan 99 Mar 99 May 99 Jan 01 Feb 01

Julian days 27 140 22 71 134 20 50

No. males No. females 68 52 49 50 41 49 51

State (A/D)

48 49 30 46 34 48 51

A A A A A A A

28 Apr 87 22 Jun 87 22 Mar 88 23 Mar 88 23 Jun 88 26 Jun 89 15 Aug 89 2 Jun 90 25 Apr 90 25 Jun 90 8 Aug 90 21 Sep 90 2 Jun 91 22 May 91 20 Feb 92 6 Mar 92 4 Aug 93 4 Mar 94 19 May 94 2 Jun 94 25 Jan 95 27 Feb 95 26 Apr 95 6 Apr 95 25 Jun 95 17 Aug 95 23 Feb 96 9 Jun 98 23 Sep 98 2 Mar 99 22 Apr 99 27 May 99 2 Sep 99 28 Feb 01 17 May 01 12 Aug 01 13 Aug 01 13 Feb 02 5 Jun 02 11 Feb 04

Julian days

No. males

118 173 82 83 175 146 227 37 115 176 220 264 37 142 51 155 98 63 139 153 25 58 116 155 176 229 54 157 266 61 112 147 245 59 137 224 225 44 156 42

17 9 34 51 45 39 52 77 49 41 51 47 64 60 62 61 39 74 58 52 67 55 56 53 53 98 76 13 50 58 49 39 88 82 15 52 50 57 52 66

No. females State(A/D) 16 13 20 50 48 52 24 68 49 39 30 37 40 46 54 31 38 58 48 51 48 37 42 40 41 91 55 18 50 50 50 35 68 47 18 43 50 43 48 40

A A A A A A ? A A A A D A A A A A A A A A A A A A D A A D A A A ? A A A D A A A




29 29

30


Chapter 5 Report on the availability and quality of seabird information in Namibia and South Africa J. Kemper1, 2 and R.J.M. Crawford1, 3 1

Avian Demography Unit, Department of Statistical Sciences, University of Cape Town, Rondebosch 7701, South Africa African Penguin Conservation Project, c/o Lüderitz Marine Research, Ministry of Fisheries and Marine Resources, P.O. Box 394, Lüderitz, Namibia 3 Marine and Coastal Management, Department of Environmental Affairs and Tourism, Private Bag X2, Rogge Bay 8012, South Africa 2

1. NAMIBIA 1.1 Historical seabird information in Namibia before 1994 Early, anecdotal, accounts of seabird numbers in Namibia stem from whalers, sealers and guano collectors traveling along the coast of western Africa. Notable among those are descriptive accounts from Morrell (1844; in Shelton et al. 1984), Ex-member of the Committee (1845), Eden (1846), Anonymous (1885) as well as those listed in Kinahan (1990, 1992). The first comprehensive surveys of the Namibian islands as well as the artificial breeding platforms were done using aerial photography in 1956 (Rand 1963). Since then, aspects of seabird population trends, biology and diet have been investigated in Namibia, mainly by the then Sea Fisheries Institute (e.g. Rand 1960, Matthews 1961, Crawford et al. 1985, 1991) as well as Cape Nature Conservation (e.g. Berry 1974, Williams 1985, Williams and Dyer 1990) and the University of Cape Town (e.g. Cooper 1985). Largely anecdotal information was also collected, mainly in diaries, by island headmen, tasked with guarding guano stocks and overseeing guano and seal harvesting operations. In addition to subsequent aerial surveys in 1967 and 1969 (Shelton et al. 1984), ground surveys of most islands were conducted in 1970 (Frost et al. 1976), and roughly annually between 1977 and 1979 and during the 1980s (e.g. Shelton et al. 1984, Crawford et al. 1995). Counting methods differed between surveys, including aerial surveys and ground counts during which individuals, nests or potential nest sites were counted (see Appendix 1 for an example of different techniques used by various institutions to count African Penguins at Ichaboe Island between 1956 and 1990). Monitoring effort of seabird populations intensified when Ichaboe Island became permanently staffed for that purpose in 1989, followed by Mercury Island in 1992. 1.2 Seabird information in Namibia since 1994 Namibia achieved independence in 1990 and all offshore islands were transferred from South Africa to Namibia by the Transfer of Walvis Bay to Namibia Act, 1993. Since 1994, the Ministry of Fisheries and Marine Resources (MFMR) has been responsible for the management of the islands; while mainland sites fall under the authority of the Ministry of Environment and Tourism (MET), Namdeb or local authorities.

The artificial platforms near Swakopmund and Cape Cross are privately owned. In 1996, Possession Island became permanently staffed (it had also been briefly staffed during the early 1990s) for the purpose of seabird monitoring. Halifax Island, easily accessible from the town of Lüderitz, has been regularly visited by staff of MFMR and J. Kemper since 1996. Most Namibian seabird information stems from the four staffed / regularly visited islands. In addition, photographs were taken of the guano platforms and gannet colonies during aerial seal censuses, undertaken in mid December in most years since the early 1980s. This is a cooperative effort between MFMR and Marine and Coastal Management (MCM). The aerial photographs are developed, counted and housed by MCM. Occasional surveys of other seabird breeding localities in Namibia, particularly of the non-staffed islands are undertaken. Ad hoc sightings of seabirds at sea have been recorded by D. Boyer (formerly of MFMR) during pilchard research cruises in the 1990s. More recently, staff of the seabird section of MFMR have been involved in occasionally monitoring seabirds at sea during trawling and long-lining activities (this includes collecting data for a BCLME project on the effects of long-lining on seabirds). A large number of seabirds, particularly African Penguins and Cape Gannets, have been and are being ringed in Namibia. A comprehensive data base containing all Namibian seabird ringing records, as well as records of ringed birds recaptured or recovered in Namibia are kept at MFMR. 1.3 Seabird information collected by MFMR 1.3.1 African Penguins Spheniscus demersus 1.3.1.1 Population estimates Monthly count data of active nests (containing eggs or chicks) are available from Mercury Island since 1994, Ichaboe Island since 1992 and from Halifax and Possession Islands since 1996. These four islands support approximately 96% of the African Penguin population in Namibia. Counts are done every two weeks at Possession Island. Separate counts of nests containing eggs and those containing chicks have been done at Halifax Island since 2000 and at Possession Island since 1996. Additional nest counts from other penguin breeding localities are collected during annual seabird surveys and during ad hoc visits. Counts done at all localities prior to 1992 are mostly single annual counts of nest sites (nests containing nesting material, eggs or chicks Top Predators of the Benguela System

31 31

or those defended by an adult). Two-weekly counts of moulting individuals (with juvenile individuals undergoing moult counted separately from those in adult plumage) have been done at Mercury Island since 1991 (with a gap between April 1993 and March 1994), Ichaboe Island since 1992, Halifax Island since 1996, with some gaps during the late 1990s and between November 2005 and February 2007, and Possession Island since 1996. All penguin nest and moult counts have been collated, computerised and vetted and are regularly updated. A comprehensive review of all count data for African Penguins in Namibia was done in 2006 (Kemper 2006). 1.3.1.2 Banding Large numbers of African Penguins have been banded and subsequently monitored in Namibia. A total of 8 623 penguins were banded between 1991 and 2005; of these, 7 346 were banded as chicks (Table 1). Monitoring intensity of banded penguins in Namibia was poor during the 1980s and early 1990s. As islands became staffed or were visited regularly, monitoring effort increased. At all four islands, banded penguins are currently recorded during fortnightly moult counts as well as during monthly active nest counts, and during daily island rounds at the three staffed islands. Since 2000, additional, dedicated band searches have been carried out at all four islands. Re-sighting effort at these four localities is relatively uniform throughout the year. Since 2002 only rehabilitated penguins and fledglings from monitored study nests are banded, pending the design and supply of better quality penguin bands. Banding, re-sighting and recovery records are regularly updated on MFMR’s seabird ringing database (in Excel format), which contained nearly 38 000 records in December 2006. Copies of records are sent electronically to SAFRING and MET at the end of each ringing year. In addition to the information required by SAFRING, the database provides details of the bird’s activities, such as breeding or moult status. 1.3.1.3 Breeding success Individual penguin nests have been monitored throughout the year at Mercury and Ichaboe Islands since December 1996, and at Possession and Halifax Islands since December 1999 and April 2000 respectively. Since January 2000, a number of nest site characteristics have been recorded for each monitored nest. Nest contents of monitored nests were

Table 1: Number of African Penguin chicks banded in Namibia between 1991 and 2003. Year refers to 1 July of that year to 30 June the following year Year

1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 32

Banding locality Mercury

Ichaboe

Possession

Other

305 886 240 76 43 35 118 78 275 118 60 93 34 9 0

366 598 309 192 181 84 123 122 228 180 20 307 228 76 78

83 190 2 0 0 15 240 172 220 223 165 47 125 131 2

33 37 5 0 0 6 12 14 3 49 47 19 18 11 15


checked every 5–7 days. A total of 2 780 nests have been monitored until April 2004 at the four islands. Nest monitoring data (including information on the date at which the nest was first monitored, estimated hatching and fledging dates of individual eggs or chicks, and estimated egg / chick mortality dates), associated nest site characteristics, chick weights taken shortly before fledging and banding details of study nest fledglings banded during this period have been collated, computerised and vetted. Study nest data collected after April 2004 still need to be collated and computerised. Since nests rather than individual pairs are followed, these data do not allow for a calculation of annual breeding success per pair. 1.3.1.4 Diet sampling Diet sampling of African Penguins is regularly done at Mercury Island (since 2000), and more sporadically at Ichaboe and Possession Islands. Between 2001 and 2005, an average of eight stomach samples were collected each week at Mercury Island. Diet samples have been processed by island staff. Records of diet samples are largely still written in island log books and need to be computerised. Records include information on date and time of collection, weight of the bird, mass of the diet sample and identity and mass of the individual prey species. 1.3.1.5 Transmitter studies Satellite PTTs (Platform Terminal Transmitters) were fitted to one breeding penguin each at Mercury and Ichaboe Islands in January 2003. Rough summary maps of the recorded tracks were forwarded to MFMR. During a seabird survey in 2005 one breeding penguin each at Halifax and Possession Islands were equipped with PTTs. Electronic data for all four deployments are available at MFMR (Janine Basson) and MCM (A.B. Makhado). GPS loggers (recording the individual’s position at regular interval in addition to other information such as diving depth and water temperature) have been deployed on a number of penguins at Mercury, Halifax and Possession Islands in 2004 and 2005 to obtain information on foraging strategies at different localities. This information forms part of a PhD project by Ms. Katta Ludynia, who is based at the University of Kiel, Germany. Data are currently collated, analysed and interpreted. A copy of all data will be given to MFMR to be incorporated into the seabird data base. 1.3.2. Cape Gannets Morus capensis 1.3.2.1 Population estimates Owing to the size of the gannetries, particularly at Ichaboe Island, and the associated difficulties of doing accurate ground counts of nests, estimates of breeding population size are derived from aerial photos. In most years, aerial photos are taken of all gannetries in Namibia (i.e. at Mercury, Ichaboe and Possession Islands) during the aerial seal census in mid-December, which, in some years might be prior to the breeding peak. The photos are developed and analysed by staff at MCM. The area (in ha) occupied by breeding gannets is estimated using an Ibas interactive imageanalysis system; numbers of breeding pairs at a particular breeding locality are extrapolated from the estimated area occupied and ground-based density estimates. Population estimates based on this method have been forwarded to MFMR, but copies of the aerial photos still need to be made available to MFMR for further analysis. 1.3.2.2 Ringing Between 1998 and 2006, 5 513 Cape Gannet chicks and 834 adults have been ringed at Mercury, Ichaboe and Possession Islands. These records, in addition to 985 re-sighting or re-

covery records between 2002 and 2006 have been computerised and vetted. Additional gannet ringing records from Namibia prior to 1998 have not been computerised yet; in some cases, paper copies of ringing records are missing. These records could be obtained from SAFRING. 1.3.2.3 Breeding success Individual gannet nests have been monitored annually between since the mid 1990s at Mercury and Ichaboe Islands. No study nests have been monitored at Possession island, as the single, small gannetry there is particularly vulnerable to disturbance. Nest contents of study nests are recorded weekly. Study nest information is recorded in island logbooks. None of these data have been collated, computerised or analysed. 1.3.2.4 Diet sampling Diet samples have been regularly collected from Cape Gannets at Mercury Island (between 2000 and 2005). Diet sampling has been sporadic at Ichaboe Island. At Possession Island, virtually no diet sampling has been done in recent years owing to the vulnerability to disturbance of that colony. Some of the samples have been processed; few data have been collated and computerised. 1.3.2.5 Transmitter studies In January 2003, one breeding Cape Gannet each was fitted with a PPT at Mercury, Ichaboe and Possession Island. The transmitter of the gannet at Possession Island failed and another gannet was fitted with a transmitter there in January 2005. Rough summary maps of the recorded tracks of the individuals equipped in 2003 were sent to MFMR. Electronic data for all four deployments are available at MFMR (Janine Basson) and MCM (A.B. Makhhado). GPS data loggers have been deployed on a number of breeding Cape Gannets at Ichaboe Island during the 2004/ 05, 2005/06 and 2006/07 breeding seasons by Ralf Mullers, PhD student at the Avian Demography Unit, University of Cape Town. This data yields detailed information on foraging flight paths. Data are currently collated and analysed; copies of the data will be given to MFMR upon completion of the project. Similar loggers were used on a smaller sample of Cape Gannets at Mercury Island by K. Ludynia from Kiel University, Germany in 2004. GPS loggers have also been deployed as part of a collaborative, international study. The data, collected by Drs Sue Lewis in December 2003 at Ichaboe and Mercury Islands, were partially analysed by B. Dundee as part of his MSc project during 2004 and 2005. 1.3.3 Bank Cormorants Phalacrocorax neglectus 1.3.3.1 Population estimates Monthly counts of Bank Cormorant active nests (containing eggs or chicks), active nest sites (nests under construction, but not yet containing eggs or chicks), as well as head counts have been collected at Mercury Island (since October 1990), Ichaboe Island (since January 1990) and Possession Island (since June 1996). Nests at other breeding localities are counted during (annual) seabird surveys or on an ad hoc basis. Nest counts have been collated, computerised and vetted. 1.3.3.2 Ringing A total of 586 Bank Cormorants, including 489 chicks, have been ringed at Mercury and Ichaboe Islands as well as Neglectus Islet between 1998 and 2005. Altogether 87 resighting records and nine recovery records (all but one from chick which never fledged) have been reported since 1996. These records are all incorporated into MFMR’s seabird ringing database. Computerised records have been vetted; more

ringing records prior to 1998 may exist. These need to be traced, possibly through SAFRING. 1.3.3.3 Breeding success A number of Bank Cormorant nests have been monitored weekly at Mercury Island since 2001. None of these data have been computerised or analysed yet. 1.3.3.4 Diet sampling Samples of Bank Cormorant pellets are collected at irregular intervals, with data gaps between for some years at Mercury and Ichaboe Islands since 2001. Few pellets have been processed and otoliths extracted, sorted by species and counted. Processed and unprocessed pellets are stored at MFMR. No Bank Cormorant diet sampling data have been collated, computerised or analysed. 1.3.4 Crowned Cormorants Phalacrocorax coronatus 1.3.4.1 Population estimates Monthly counts of Crowned Cormorant active nests (containing eggs or chicks), active nest sites (nests under construction but not yet containing eggs or chicks) as well as head counts have been collected at Mercury Island (since August 1996), Ichaboe Island (since January 1990) and Possession Island (since June 1996). Nests at other breeding localities are counted during (annual) seabird surveys or on an ad hoc basis. Nest counts from these localities have been only partially collated, computerised and vetted. 1.3.4.2 Ringing Only one Crowned Cormorant has been ringed by MFMR staff since 1996; Potential ringing records prior to 1998 need to be traced, possibly through SAFRING. 1.3.4.3 Breeding success No Crowned Cormorant nests have been monitored in southern Namibia. During 2002, some nests were monitored at Bird Rock platform as part of a final year project of a student studying at the Polytechnic of Namibia. These data have not been made available to the MFMR seabird database. 1.3.4.4 Diet sampling No diet samples have been collected from Crowned Cormorants in Namibia. 1.3.5 Cape Cormorants Phalacrocorax capensis 1.3.5.1 Population estimates Monthly counts of Cape Cormorant active nests (containing eggs or chicks), active nest sites (nests under construction but not yet containing eggs or chicks) as well as head counts have been recorded at Mercury Island (since March 1996), Ichaboe Island (since January 1990) and Possession Island (since June 1996). Peak nest counts, as well as estimates of numbers of nests at other breeding localities are obtained from aerial photographs taken during the seal census in midDecember, during seabird surveys or on an ad hoc basis. Counts made during mid-December may not necessarily represent peak breeding activities. Monthly counts have been partially collated, computerised and vetted. Estimates of numbers of breeding pairs from aerial surveys, as well as from correlations with guano production of the platforms in central Namibia have been provided by MCM. 1.3.5.2 Ringing Of the 54 Cape Cormorants ringed in southern Namibia since 1998, 50 were ringed during the January 2003 seabird cruise with rings issued to MCM. These ringing records have been incorporated into the MFMR seabird ringing database. Potential ringing records prior to 1998 need to be traced and Top Predators of the Benguela System

33 33

incorporated, possibly with the assistance of SAFRING. 1.3.5.3 Breeding success No Cape Cormorant nests have been monitored in southern Namibia. 1.3.5.4 Diet sampling Some pellets have been collected from Cape Cormorants in Namibia, particularly at Mercury and Ichaboe Islands; few have been processed or analysed. 1.3.6 Kelp Gulls Larus dominicanus 1.3.6.1 Population estimates Monthly counts of Kelp Gull active nests (containing eggs or chicks) and active nest sites (nests under construction but not yet containing eggs or chicks) have been recorded at Ichaboe, Halifax and Possession Islands. Monthly head counts are made at these three islands as well as at Mercury Island. Much of these counts are still in island logbooks. So far, data for Ichaboe have been computerised for the period July 1998 to October 2004, Halifax (January 2000 to November 2005) and Possession Island (parts of 1998 and 1999, 2004). Nests at other breeding localities are counted during (annual) seabird surveys or on an ad hoc basis; some of these data have been computerised. 1.3.6.2 Ringing A total of 784 Kelp Gulls, including three juvenile and three adult individuals were ringed in southern Namibia between 2000 and 2005. These data, as well as 153 re-sighting and recovery records have been incorporated into MFMR’s seabird ringing database. Ringing records from before 2000 need to still be entered on the database. 1.3.6.3 Breeding success A few nests have been monitored at Ichaboe Island as well as at Possession Island. These data have not been computerised, vetted or analysed yet. 1.3.6.4 Diet sampling No diet samples have been collected from Kelp Gulls in Namibia. 1.3.7 Other species Other seabird species for which monthly counts exist from Mercury, Ichaboe, Halifax and Possession Islands include Whitebreasted Cormorant (Phalacrocorax lucidus), Hartlaub’s Gull (Larus hartlaubii), Swift Tern (Sterna bergii), Damara Tern (Sterna balaenarum) and African Black Oystercatcher (Haematopus moquini). Information collected may include counts of individuals, active nests and active nest sites. The presence of other species of sea- and shorebirds encountered at the staffed islands are recorded during monthly counts or during daily island rounds. 1.4 Recommendations It is suggested that basic but regular monitoring of seabird populations in Namibia, particularly those which are threatened locally or globally, is continued at least at Mercury, Ichaboe, Halifax and Possession Islands. Where possible, monthly counts are advantageous, although single annual counts at remote localities are also valuable. It is imperative that good quality photocopies of ALL island logbooks are made. This task should be a priority, to be effected by the seabird section at MFMR. Copies need to be carefully checked for quality, collated and stored properly. On several occasions logbooks have been lost or became illeg34


ible, and valuable information has been lost. Similarly, poorquality photocopies of logbooks have caused data to become illegible. All data need to be computerised in a way which maximises the detail of data collected, yet allows easy summarising, filtering and sorting of data. Data headings must be explicit to avoid confusion and to ensure a standardised monitoring approach. An original database should always be kept separately, in case of a diskette becoming corrupt or infected with a virus, or through accidental “mis-sorting” (a classic example involved a large Excel spreadsheet containing detailed ringing and re-sighting information from Mercury Island. During a sorting operation, only one column was sorted instead of the whole sheet, causing the information on the spreadsheet to become jumbled. The problem was only noticed until after the jumbled version had been saved. The file had not been backed up anywhere else and had to be retyped). Computerised data need to be backed up regularly in more than one location. A number of monitoring techniques have been used to estimate seabird populations since comprehensive counts of seabirds were first made in the 1950s. Although monitoring protocol has been standardised in Namibia, there are still inter-regional differences, particularly with regard to the definition of what constitutes a nest. Differences must be considered and acknowledged when interpreting long-term population trends at a particular locality or between localities and regions. 1.5 Contacts and queries regarding Namibian data All data collected at seabird breeding localities in Namibia by staff of the Ministry of Fisheries and Marine Resources are collated and curated by the Seabird Section, Lüderitz Marine Research, Ministry of Fisheries and Marine Resources, PO Box 394, Lüderitz, Namibia. Ownership of these data rests with MFMR. Queries regarding all MFMR seabird data should be directed to: Head of Section, Seabird Section National Marine Information and Research Center Ministry of Fisheries and Marine Resources P.O. Box 912, Swakopmund, Namibia Tel: +264 64 4101000, Fax: +264 64 404385 Queries regarding data collected in Namibia prior to 1994 by the South African Sea Fisheries Research Institute (SFRI), now named Marine and Coastal Management (MCM), and not yet made available to Namibia, should be directed to: The Deputy Director Sub-directorate Ecosystem Utilization and Conservation Department of Environmental Affairs and Tourism Branch: Marine and Coastal Management Private Bag X2, Roggebaai 8012, South Africa Tel: +27 21 4023114, Fax: +27 21 4023639 Additional information regarding seabird abundance at coastal localities not monitored by MFMR is collected by staff of the Ministry of Environment and Tourism, as part of their bi-annual Wetland Bird Count. For further details, contact: Holger Kolberg Principal Conservation Scientist, Survey Unit Directorate Scientific Services Ministry of Environment and Tourism Private Bag 13306, Windhoek, Namibia Tel. +264 61 2842554, Fax +264 61 259101

2. SOUTH AFRICA Similarly to Namibia, information on seabirds breeding in South Africa was largely anecdotal until surveys were undertaken by Rand (1963a) in the 1950s. Little information was collected in the 1960s and early 1970s (for sources of these sporadic studies refer to Hockey et al. 2005).

sporadic information is collected from Bird Island, Eastern Cape. Sampling of the diet of African Penguins has been conducted at Robben Island since 1989 and at Dassen Islands since 1991. More sporadic information is collected from Dyer Island and the Eastern Cape. For other seabirds, diet samples are collected opportunistically, e.g. regurgitations by chicks when they are being banded.

2.1 Population estimates

2.5 Transmitter studies

Commencing in 1977, the numbers of active nest sites of all seabird species have been counted at many South African colonies in many years. The raw data are stored on maps of islands and island visit cards at MCM (contact person B.M. Dyer). Results for African Penguins up until 1994 have been published in Crawford et al. (1995) and for the Western Cape for 1987 to 2005 in Underhill et al. (2006). For African Penguins, two-weekly counts of moulting individuals (with juvenile individuals undergoing moult counted separately from those in adult plumage) have been done at Robben Island since 1989 (Underhill and Crawford 1999). Shorter time series exist for some other islands. Information for Robben Island is computerised at MCM (R.J.M. Crawford) and for Dassen (contact person A.C. Wolfaardt) and Dyer (contact person L. Waller) islands at CapeNature. As in Namibia, owing to the size of the gannetries, estimates numbers of Cape Gannets breeding are derived from aerial photos and measures of nest density (Crawford et al. 2007). Large colonies of Cape Cormorants have also been counted from aerial photographs (methods described in Shelton et al. 1982). The photographs are developed and housed at MCM (contact person B.M. Dyer).

PTTs have been fitted to African Penguins and Cape Gannets by MCM. Electronic data are stored on a database at MCM (contact person A.B. Makhado). GPS loggers (recording the individual’s position at regular interval in addition to other information such as diving depth and water temperature) have been deployed on a number of penguins and gannets by the University of Cape Town (contact person P.G. Ryan, Percy FitzPatrick Institute, also R. Navarro, Avian Demography Unit).

2.2 Banding Large numbers of seabirds have been banded and subsequently monitored in South Africa. Information on banding, as well as recoveries and re-sightings, is stored at SAFRING. Recent analyses for African Penguins include Whittington et al. (2005a, 2005b). Other species that have been intensively banded include Great White Pelican, Cape Gannet, Cape and Crowned Cormorants, Kelp and Hartlaub’s Gulls and Swift Tern. Some results are reported in Hockey et al. (2005). 2.3 Breeding information Breeding success of African Penguins has been monitored over several years by MCM and Earthwatch (L.G. Underhill, University of Cape Town) at Robben Island (Crawford et al. 2006) and by CapeNature at Dassen Island (contact person A.C. Wolfaardt). For Cape Gannets, it has been monitored by MCM at Lambert’s Bay and Malgas Island. The data collected by MCM and Earthwatch are stored at MCM (contact person L. Upfold). Colony census cards have recorded information on the nest contents of most seabirds over several years. These cards are stored at MCM (contact person B.M. Dyer). They provide information on the numbers of eggs and chicks present at nests that were investigated. 2.4 Diet sampling MCM maintains a computer database that records information pertaining to the diet of seabirds. A manual for use of the database is available (contact person L. Upfold). Sampling of the diet of Cape Gannets has been undertaken in most months at Lambert’s Bay and Malgas Island since 1978 (about 50 regurgitations being collected each month). More

2.6 Queries Queries regarding MCM data should be directed to: The Deputy Director Sub-directorate Ecosystem Utilization and Conservation Department of Environmental Affairs and Tourism Branch: Marine and Coastal Management Private Bag X2, Roggebaai 8012, South Africa Tel: +27 21 4023114, Fax: +27 21 4023639 References Anonymous 1885. Proceedings of the Angra Pequeña and West Coast Claims Joint Commission. March–September, 1885. Saul Solomon & Co., Cape Town. Berry, H.H., Seely, M.K. & Fryer, R.E. 1974. The status of the Jackass Penguin Spheniscus demersus on Halifax Island of South West Africa. Madoqua Series II, 3: 27–29. Cooper, J. 1985. A note on the diet of Cape Cormorant Phalacrocorax capensis at Mercury Island, South West Africa, in November 1978. South African Journal of Marine Science 3: 129–130. Crawford, R.J.M., Cruickshank, R.A., Shelton, P.A. & Kruger, I. 1985. Partitioning of a goby resource amongst four avian predators and evidence for altered trophic flow in the pelagic community of an intense, perennial upwelling system. South African Journal of Marine Science 3: 215–228. Crawford, R.J.M., Williams, A.J., Randall, R.M., Randall, B.M., Berruti, A. & Ross, G.J.B. 1990. Recent population trends of Jackass Penguins Spheniscus demersus off southern Africa. Biological Conservation 52: 229–243. Crawford, R.J.M., Ryan, P.G. & Williams, A.J. 1991. Seabird consumption and production in the Benguela and Western Agulhas Ecosystems. South African Journal of Marine Science 11: 357– 375. Crawford, R.J.M., Williams, A.J., Hofmeyr, J.H., Klages, N.T.W., Randall, R.M., Cooper, J., Dyer, B.M. & Chesselet, Y. 1995. Trends of African Penguin Spheniscus demersus populations in the 20th century. South African Journal of Marine Science 16: 101–118. Eden, T.E. (Jr). 1846. The search for nitre and the true nature of guano, being an account of a voyage to the south-west coast of Africa; also a description of the minerals found there, and of the guano islands in that part of the world. R. Groombridge & Sons, London. Ex-Member of the Committee (1845). The African Guano Trade. Being an account of the trade in Guano from Ichabo, and other places on the African Coast, more particularly, the Proceedings of the Committee of Management. Nautical Magazine 11: 616– 666. Frost, P.G.H., Siegfried, W.R. & Cooper, J. 1976. Conservation of


35 35

the Jackass Penguin (Spheniscus demersus (L.)). Biological Conservation 9: 79–99. Hockey, P.A.R., Dean, W.R.J. & Ryan, P.G. 2005. Roberts Birds of Southern Africa. VIIth ed. John Voelcker Bird Book Fund, Cape Town. Kemper, J. 2006. Heading towards extinction Demography of the African Penguin in Namibia. PhD thesis, University of Cape Town, South Africa. Kinahan, J. 1990. The impenetrable shield: HMS Nautilus and the Namib coast in the late eighteenth century. Cimbebasia 12: 23– 61. Kinahan, J. 1992. By command of their Lordships. The exploration of the Namibian coast by the Royal Navy, 1795–1895. Namibia Archaeological Trust, Windhoek, Namibia. Matthews, J.P. 1961. The Pilchard of South West Africa (Sardinops ocellata) and the Marsbanker (Trachurus trachurus) – bird predators, 1957–1958. Investigational Report of the South West African Marine Research Laboratory 3: 1–35. Rand, R.W. 1960. The biology of guano-producing sea-birds. 2. The distribution, abundance and feeding habits of the Cape Penguin, Spheniscus demersus, off the south-western coast of the Cape Province. Investigational Report Sea Fisheries Research Institute 41: 1–28. Rand, R.W. 1963a. The biology of guano-producing sea-birds. 4. Composition of colonies on the Cape islands. Investigational Report Sea Fisheries Research Institute 43: 1–32. Rand, R.W. 1963b. The biology of guano-producing sea-birds. 5. Composition of colonies on the South West African islands. Investigational Report Sea Fisheries Research Institute 46: 1–26. Shaughnessy, P.D. 1977. Jackass Penguins on the northern guano islands. Cormorant 2: 18–19.

Shelton, P.A., Crawford, R.J.M., Kriel, F. & Cooper, J. 1982. Methods used to census three species of South African seabirds, 1979 – 1980. Fisheries Bulletin South Africa 16: 115–120. Shelton, P.A., Crawford, R.J.M., Cooper, J. & Brooke, R.K. 1984. Distribution, population size and conservation of the Jackass Penguin Spheniscus demersus. South African Journal of Marine Science 2: 217–257. Underhill, L.G. & Crawford, R.J.M. 1999. Season of moult of African Penguins at Robben Island, South Africa, and its variation, 1988– 1998. South African Journal of Marine Science 21: 437–441. Underhill, L.G., Crawford, R.J.M., Wolfaardt, A.C., Whittington. P.A., Dyer, B.M., Leshoro, T.M., Ruthenberg, M., Upfold, L. & Visagie, J. 2006. Regionally coherent trends in colonies of African Penguins Spheniscus demersus in the Western Cape, South Africa, 1987–2005. African Journal of Marine Science 28: 697–704 Whittington, P.A., Randall, R.M., Crawford, R.J.M., Wolfaardt, A.C., Klages, N.T.W., Randall, B.M., Bartlett, P.A., Chesselet, Y.J. & Jones, R. 2005a. Patterns of immigration to and emigration from breeding colonies by African Penguins. African Journal of Marine Science 27(1): 205–213. Whittington, P.A., Randall, R.M., Randall, B.M., Wolfaardt, A.C., Crawford, R.J.M., Klages, N.T.W., Bartlett, P.A., Chesselet, Y.J. & Jones, R. 2005b. Patterns of movement of the African Penguin in South Africa and Namibia. African Journal of Marine Science 27 (1): 215–229. Williams, A.J. 1985. Seabirds breeding on Halifax Island off the coast of South West Africa/Namibia, June 1984. Cormorant 13: 75–76. Williams, A.J. & Dyer, B.M. 1990. The birds of Hollamsbird Island, least known of the southern African guano islands. Marine Ornithology 18: 13–18.

APPENDIX Example of different counting methods used for African Penguins at Ichaboe Island, Namibia, between 1956 and 1990 Date

Method

Oct. 1828 28 May 1845 Dec. 1851 1853 1860 1885 20 Nov. 1956 20 Nov. 1956 15 Nov. 1967 25 /26 Nov. 1969 1970 12–13 March 1977 24–28 Nov. 1978 24–28 Nov. 1978 28 Nov. 1978 28 Nov. 1978 9 July 1979 3–7 Feb. 1980 24 Nov. 1985 25 Nov. 1986 27 Nov. 1986 29 Aug. 1987 2 Dec. 1987 June 1990

anecdotal anecdotal anecdotal anecdotal anecdotal anecdotal aerial aerial aerial aerial rough estimate visit transects aerial aerial aerial head count visit visit visit visit visit visit

What counted

individuals individuals individuals individuals total population adults nest sites breeding population individuals individuals individuals nest sites nest sites nest sites nest sites nest sites nest sites nest sites

1

Count literally covered about 100 000 innumerable numbers much reduced numbers recovering present 4179 8400 2882 3226 2000 several thousand 3598 7196 10 437 12 207 2120 4200 2070 739 739 1372 827 2427

“more numerous than at any other part of the coast which I have seen” p. 34 actual count from photos 3 estimated population 4 extrapolated from nest sites 5 aerial count adjusted for absenteeism 2

36


Source Morrell 1844; in Shelton et al. 1984 Eden 1846 1 Keane 1851, in Kinahan 1992 Anonymous 1885 Anonymous 1885 Anonymous 1885 Rand 1963 2 Rand 1963 3 Shelton et al. 1984 Shelton et al. 1984 Frost et al. 1976 Shaughnessy 1977 Shelton et al. 1984 Shelton et al. 1984 4 Shelton et al. 1984 Shelton et al. 1984 5 Shelton et al. 1984 Shelton et al. 1984 Crawford et al. 1990 Crawford et al. 1995 SFRI unpubl. report 1986 Crawford et al. 1995 Crawford et al. 1995 Crawford et al. 1995

Cape Fur Seal

Chapter 6 Making sense of censuses and dealing with missing data: trends in pup counts of Cape Fur Seal Arctocephalus pusillus pusillus for the period 1972–2004 SP Kirkman1, 2 *, WH Oosthuizen2, MA Meÿer2, PGH Kotze2, J-P Roux3 and LG Underhill1 1

Avian Demography Unit, Department of Statistical Sciences, University of Cape Town, Rondebosch 7701, South Africa Marine and Coastal Management, Department of Environmental Affairs and Tourism, Private Bag X2, Rogge Bay 8012, South Africa 3 Lüderitz Marine Research, Ministry of Fisheries and Marine Resources, PO Box 394, Lüderitz , Namibia *Corresponding author, e-mail: [email protected] 2

Trends in the population of Cape Fur Seals Arctocephalus pusillus pusillus were estimated from counts of pups on aerial photographs of colonies, taken between 1972 and 2004 to determine trends in the overall population and sub-populations. Incomplete coverage resulted in missing data in some years. Various methods of determining proxy values for missing data were assessed, and it was concluded that different methods were applicable to Namibian and South African colonies. This reflected variation in trends of pup counts between the countries, which was associated with differences in productivity between the southern and northern Benguela ecosystems. In Namibia, temporal changes in pup num-

bers were non-linear in some years and there was correspondence in fluctuations at most colonies. This appeared to be on account of an effect of periodic, widescale prey shortages that reduced birth rates. There was a northward shift in the distribution of seals in the northern Benguela system. In South Africa, pup counts were much less variable between years, probably on account of a relative stability of food supply. A linear approach was therefore suitable for determining proxy values for missing data at South African colonies. Pup counts suggest that there has been little change in the overall population of the Cape Fur Seals since 1993, when it was estimated at about two million animals.

Keywords: abundance, Cape Fur Seal, distribution, missing data, Namibia, pup counts, South Africa

Introduction Cape Fur Seals Arctocephalus pusillus pusillus occur along the southern and western coasts of southern Africa (Figure 1). The size of the seal population before the arrival of Europeans in southern Africa is unknown, but it is thought that seals occurred on most, if not all, of the islands off South Africa and Namibia (Shaughnessy 1982, 1984). However, seal hunting (sealing) between the 17th and 19th centuries caused a marked decline in the population size (Rand 1952, Shaughnessy and Butterworth 1981). The effects of uncontrolled sealing, together with the activities of guano collectors and the management of many islands for guano and other seabird products after the discovery of guano, resulted in the extirpation of seals from many of their former breeding locations. In general, the remaining seals were restricted to islets not utilised by guano-producing birds, and not easily accessible to seal hunters (Rand 1952). By the beginning of the 20th century, Cape Fur Seals had disappeared from at least 23 offshore locations (Best and Shaughnessy 1979, Shaughnessy 1982). At its most reduced level, the population size is thought to have been below 100 000 individuals (Shaughnessy and Butterworth 1981). The most recent assessment of the Cape Fur Seal population size, estimated about 2 million animals (including pups) at the beginning of 1993 (Butterworth et al. 1995),

indicating that the population had grown about 20-fold during the 20th century. The recovery in numbers followed the imposition of legal controls on sealing at the beginning of the 20th century, and has been perceived as the normal response of a population recovering from overexploitation (Shaughnessy and Butterworth 1981). The recovery was notwithstanding that seals have been unable to re-colonise most of the offshore locations from where they were previously extirpated (Shaughnessy 1984). Instead, new breeding colonies that formed on the mainland during the 20th century have accounted for most of the growth (Rand 1972). It is thought that mainland-based seal breeding colonies were not viable before the arrival of Europeans in southern Africa, owing to the presence of terrestrial mammal predators, not only lions Panthera leo, brown hyaenas Hyaena brunnea and black-backed jackals Canis mesomelas, but also humans such as early hunter–gatherers (Shaughnessy and Butterworth 1981). The large mainland colonies have been established in the coastal diamond mining zones of Namibia and South Africa, where terrestrial seal predators had been largely exterminated and human access and disturbance was minimised (Rand 1972). With the seal population estimated to have doubled in size between 1970 and 1990, it was mooted that the seemingly unlimited breeding space presented by mainland locations, compared with offshore locations, may have caused the seal Top TopPredators Predatorsof ofthe theBenguela Benguela System System

39

Figure 1: Distribution of Cape Fur Seal in South Africa and Namibia, showing mainland and island breeding colonies. Regions 1–6 are indicated as R1–R6, and separated by straight lines drawn inland from the coast

population size to surpass its pre-sealing level (Griffiths et al. 2005). The outcomes of modelling exercises conducted in 1990, around the time when a moratorium was placed on seal harvesting in South Africa, predicted that the seal population would again double within 10 years, and treble in 20 years, unless the population was subjected to density dependent effects (e.g. food deprivation) or further sealing (Butterworth and Wickens 1990, Butterworth et al. 1991). Since then, however, sealing has continued off Namibia, where approximately 60% of the Cape Fur Seal population occurs (Wickens et al. 1991), but not in South Africa. Furthermore, since 1993, there have been at least two mass die offs of seals in Namibia, apparently related to the effects of unfavourable environmental conditions on the distribution and abundance of their prey (Roux 1998, Roux et al. 2002). The first, in which tens of thousands of seal pups and thousands of adults starved to death in 1994–1995 (Roux 1998), was the largest mass die-off recorded for any seal species (Harwood 2002). Consequently, there is interest in recent trends of the seal population. On the one hand, there is concern for the conservation status of the population, in view of the mass die offs and continued harvesting in Namibia. On the other hand, many fishers motivate for a reduction in seal numbers, because they perceive this as benefiting their livelihoods (Wickens et al. 1992, Best et al. 1997). Moreover, seabird conservationists claim that the seal population at the start of the 21st century exceeds its pristine level, and are concerned that seals negatively impact locally breeding seabird species classified as “threatened” according to IUCN criteria (e.g. Crawford and Robinson 1990, Ward and Williams 2004). Censuses of Cape Fur Seals have been conducted frequently since the early 1970s. The censuses were based on counts of pups on aerial photographs, taken systematically of seal breeding colonies when the numbers of new born pups of the year were expected to be at their maximum. Although these censuses inherently underestimate the numbers of pups in each colony, Shaughnessy (1987) found them to be useful indicators of pup production. However, where trends in pup numbers over time are determined from collective pup counts of colonies, complete aerial coverage in each census year is desirable. Where this was not achieved and counts of one or more colonies are lacking, values need to be inferred for the missing data. Otherwise, 40


censuses of different years are not directly comparable, particularly if one or more of the larger breeding colonies are concerned. Missing data is a recurrent problem in the timeseries of Cape Fur Seal censuses, and the problem has been approached differently between some previous assessments of the population (e.g. Butterworth et al. 1987, 1995). However, no attempt has been made to empirically assess the accuracy of alternative approaches for estimating missing data values. In this study, we assessed the accuracy shown by four different methods in approximating the correct values of available counts, with a view to determining the best-suited approach. Once this was achieved, and we had inferred missing data values were inferred accordingly, the trends in pup numbers of the whole population and various subpopulations were investigated, based on all the censuses that have been completed to date (1971–2004). The pup count trends are interpreted and the relationship between recent trends (especially since 1993) and the status of the seal population (all age-classes inclusive) are discussed. Material and Methods Background The procedure for censusing the Cape Fur Seal pup population using aerial photography is detailed by Shaughnessy (1987). Briefly, near vertical, serial overlapping photographs are taken from aircraft flying parallel flight paths over colonies at a height of c. 100 m. For large colonies, high altitude pictures (c. 300 m) of the colony are also taken, to assist with fitting of the lower altitude prints to map the colony. The timing of photographing is standardised, taking place during 16– 22 December each census year, except where otherwise indicated (see Appendix). After printing, pictures are laid out in frame sequence and a photographic mosaic of each colony is arranged. Boundaries between neighbouring, overlapping photographs are delineated on each photograph, using landmarks or seals that are in common between the photographs, to prevent counting repetition. Duplicate photographs are eliminated. The seal pups on each photograph are counted by two people, and the arithmetic mean of the counts are taken1. In the few instances where counts differ by more than 20%, additional counts are conducted until two counts are within 20% of each

other, and the other counts are discarded. Whereas this approach precludes estimates of variance, it means that the amount of variability between the counts used is always 3 000 pairs), Hartlaub’s Gulls (several colonies) and ground-nesting Crowned Cormorants on Schaapen Island. Pelicans were present in the area and small groups of up to 22 birds were seen eating cormorants and terns, but they did not cause serious damage to their colonies and they also regularly foraged on fish.

Cape Gannet At Malgas Island, pelicans patrolled the edge of the Cape Gannet colony in groups of 6 to 12 individuals (Figure 2 (3a)). Sometimes they would stand on prominent rocks to search for chicks of adequate size. When they identified a potential prey they would stop, staring at the nest attentively for minutes at the time (Figure 2 (3b)). Often smaller pelican clusters would join in larger groups. Frequently a predation would be followed by a scuffle, in which several pelicans would fight over the chick. Pelicans from nearby would flock to the site where the predation occurred, frequently chasing the pelican in possession of the chick and occasionally stealing the prey. Gannets aggressively defend their nesting site against neighbouring gannets. However they showed no defensive behaviour towards approaching pelicans, which probably indicates that pelicans were not seen as a threat. Nevertheless, pelicans were cautious when approaching nests and never ventured into the midst of the gannet colony. Only chicks on the edge of the colony were targeted by pelicans. Possibly due to reduction in food availability, the number of gannet breeding pairs diminished in the 2005/06 and 2006/ 07 breeding seasons (Crawford et al. 2007). This resulted in the colony becoming fragmented, which favoured chick and egg predation by gulls and pelicans (L. Pichegru & R. Mullers, pers. comm.). Gannet chicks eaten ranged in size between small naked chicks brooded by their parent, to chicks 3–4 weeks of age. Later in the season, pelicans attempted to prey on chicks near fledgling size (possibly >1.5 kg in mass) which were still covered in down but beginning to grow feathers. On several occasions groups of two to three pelicans attempted to eat extremely large gannet chicks. Occasionally, a pelican that had captured a large chick would be barely able to lift its bill off the ground, struggling for up to 90 min to swallow the chick, not always successfully. One chick survived a 75 minute struggle inside the pouch and even the throat of a pelican before being released. One pelican managed to swallow a chick of 2.05 kg but dropped it when disturbed in order to take flight (L. Upfold pers. comm.). Swift Tern To be able to capture highly mobile Swift Tern chicks, pelicans would surround the colony by forming a tight circle around it, enclosing a large number of chicks in the centre (Figure 2 (2)). Almost synchronously, pelicans thrust their bills forward and then lifted them upwards in order to swallow the chicks. Adult terns flew frantically around the group, uttering loud alarm calls. However, they did not attack and hit the pelicans as Kelp Gulls did. Chicks ran in all directions when attacked by pelicans. Pelicans also looked for chicks between and under bushes. Older tern chicks moved to the shore with their parents; they jumped into the water and swam in compact groups when disturbed. Pelicans were observed chasing and swallowing escapee chicks in the waters near the tern colony on Schaapen Island (P. Nel pers. comm.). 136


Individuals involved in predation events The maximum number of pelicans counted during simultaneous counts (the same day) at the three islands monitored in Saldanha Bay, was 220. These pelican counts are in likelihood underestimates because in the case of Jutten and Malgas, it was not possible to view the entire islands from the vantage points. Pelican numbers were highest between the beginning of November and the beginning of December 2006. Jutten Island had the highest number of pelicans, with a peak of 176 on 22nd November 2006. The numbers dropped abruptly once the island was depleted of breeding seabirds (Figure 3). Pelicans roosted at Jutten Island throughout the study period, even after mid December, by which time both Jutten and Schaapen islands were devoid of breeding seabirds. Pelicans continued to forage on gannet chicks at Malgas Island by day, and then returning to Jutten Island to roost. The straight line distance between Malgas and Jutten Islands is 4.3 km. Pelicans would usually fly from Jutten to Malgas early in the morning, most of them arriving with the first light before sunrise. Time of departure was variable.

Figure 3: Pelicans counted on Jutten, Malgas and Schaapen islands during the period 1 October 2006 to 31 January 2007

When left undisturbed they would leave Malgas Island any time between 16:00 and 17:00. However, sudden changes in weather conditions (mainly wind strengthening) sometimes prompted them to leave earlier. Travelling time from Malgas to Jutten Island averaged 6.5±1.4 min (n = 37) or 41.5 km/ hr, and was clearly dependent on wind strength and direction. The maximum speed recorded was 64.5 km/hr against a mild S–SE wind, and the minimum was 23.5 km/hr against a strong SW wind. Pelicans would use different flight pathways and strategies to cross the sea between the two islands, depending on the wind conditions, often changing direction to fly in the lee of the island. Mostly they flew low over the water in a horizontal line, beating wings synchronously and gliding for short periods in between wing beats. About 10% of the Western Cape pelicans carry either colour or metal rings, or both. Ringing effort started in 2001 at the breeding colony on Dassen Island. Six cohorts of pelican chicks have been ringed thus far; hence all ringed birds are aged 6 years old or younger. At most, seven individually recognisable pelicans were observed preying on seabirds during the entire observation period (Table 3). This proportion is low in comparison to feeding sites on the mainland, where 10% of birds are regularly observed to have rings. Almost all pelicans observed on the islands had adult plumage, and many were in full breeding plumage. Age of first breeding for Great White Pelicans is 3 or 4 years old (Hockey et al. 2005). Discussion Why do pelicans eat seabirds? Although all pelican species are largely piscivorous, they are able to exploit other sources of food and even scavenge on dead animals or fish discards (Johnsgard 1993). However, feeding on prey other than fish is uncommon behaviour for pelicans. Yet, in the case of the Great White Pelicans in the Western Cape, feeding on chicks of other bird species has increased noticeably in intensity and extent in the last decade. A much smaller pelican population in the past and a recent shift in dietary preferences could have been the rea-

sons for this behaviour receiving so little attention until recently. The Western Cape pelican population exhibits some characteristics that make it unique compared to Great White Pelican populations elsewhere in the world. With the exception of the western African population, the Benguela population is the only one that is wholly coastal in terms of its breeding distribution. The situation in western Africa (specifically Mauritania and Senegal) is different to the Western Cape in that pelicans live in vast wetlands that exist in the coastal regions. Sandy islands separated from the mainland by narrow and shallow channels of water offer protected sites for breeding, while the adjacent waters are suitable for foraging. In the Western Cape, where most marine resources surrounding the colony at Dassen Island are not available to the pelicans because of their inability to dive, pelicans rely mainly on the fish they capture in the estuaries and freshwater systems on the adjacent mainland, notwithstanding artificial food sources. However, the nutrient rich west coast of southern Africa is the only place throughout the range of Table 3: Ringed pelicans seen on the islands involved in predation activities Date 13 Oct 06 8 Nov 06 10 Nov 06 10 Nov 06 11 Nov 06 8–12 Jan 07 8 Jan 07 9 Jan 07 9 Jan 07 14 Mar 07 14 Mar 07 15 Mar 07 4 May 07 16 May 07 17 May 07

Island Schaapen Jutten Jutten Jutten Jutten Malgas Malgas Malgas Malgas Schaapen Schaapen Schaapen Schaapen Jutten Schaapen

Ring Blue on left, no metal Green on left, metal Blue on left Red on left, no metal Red on left, no metal 2 × Red on left, no metal Hole in pouch Black/white CS on left Black/white (other) Blue on left, no metal Hole in pouch, no metal Blue on left, no metal Blue on left, no metal Blue on left, no metal Blue on left


Years old 6 3 6 5 5 5 – 2 2 6 – 6 6 6 137

the Great White Pelican where ground-nesting seabirds are available in large densities nearby breeding pelicans. This and the fact that pelicans are opportunistic feeders (Hockey et al. 2005) are probably the most significant causal factors of predation on seabirds by pelicans, which has occurred in the region for as long as records are available (although probably not at the present scale). The marked population increase of pelicans in the Western Cape since the 1980s has probably contributed to the steep increase in the observed incidence of predatory interactions. When the pelican population was much smaller (up until the late 1980s), the impact of predation on seabird colonies would have been orders of magnitude lower, and less noticeable. The rapid increase in the size of the pelican population in the region has been attributed to the increased availability of offal on pig farms in the mid-1980s (Crawford et al. 1995, de Ponte Machado et al. in prep.). The first large-scale predation events in recent decades were observed on Dassen Island in the 1990s. Since then, predation on seabirds by pelicans in the region intensified and reached extremely high levels in 2005/06 and 2006/07. Up until January 2005, an average of 1 500 pelicans fed daily on a pig farm near Stellenbosch (Figure 1) (de Ponte Machado et al. in prep.), which therefore sustained some 60% of the pelican population in the region. Offal became increasingly scarce at this site from the end of 2004 (de Ponte Machado et al. in prep.) therefore pelicans had to rely more and more on ‘natural’ food sources, i.e. fish and other vertebrates from fresh water lakes and dams. However, while there is no apparent shortage of fish in coastal water bodies and some farm dams have been stocked with fresh water fish (Guillet & Crowe 1981, Guillet & Furness 1985, Crawford et al. 1995), the number of fresh water food sources in range of the breeding colony is limited (Figure 1, Table 4). The 25-fold increase in pelican breeding pairs, together with the sudden shortage of artificial food supplies, would certainly increase intra-specific competition for food between breeding pelicans, which have to cover distances of 120–180 km between their colony and freshwater bodies to satisfy their chick’s high energetic demands. Ashmole (1963) postulated that seabird populations are limited by densitydependent competition for food during their nesting period.

Table 4: Return trip distance (km) from the pelican breeding colony on Dassen Island to the most utilized pelican foraging sites in the Western Cape, indicating main source of food for pelicans (a: straight line; b: following the most common route to Yzerfontain and then straight to the foraging site; * fish indicates aquatic organisms other than seabirds) Foraging sites Geelbek Schaapen Island Jutten Island Malgas Island Rietvlei Rondevlei Strandfontein Paarl Stellenbosch Pig farm Atlantis Malmesbury Mamre Berg River Verlorenvlei Jakkalsfontain Olifants 138

Main prey

a

b

fish* birds birds birds fish fish fish fish fish offal fish fish fish fish fish fish fish

52 74 78 88 120 162 175 180 182 160 82 118 78 144 252 300 382

58 84 90 100 150 189 202 198 204 178 102 130 94 168 274 326 396


In support of this theory, a number of studies have indicated that seabird populations undergo density-dependent regulation of population size as a consequence of limited prey resources in range of the colony (Hunt et al. 1986, Furness & Birkhead 1984, Birt et al. 1987, Lewis et al. 2001, Votier et al. 2007). The diminishing artificial supply of food since 2005 and increased competition due to the population increase, may have led to depletion of fresh-water fish at sites near the pelican colony. However, data on trends in fresh-water fish populations and evidence of competition for food among pelicans are required to test this hypothesis. The observed increased predation pressure by pelicans on seabirds in recent years could be a response of foodstressed pelicans following the reduction of artificial sources of food since 2005. Interactive effects of food shortages (affecting both predator and prey), and increased predation rates by Larus argentatus and L. marinus, have been described as a cause of widespread breeding failure of Blacklegged Kittiwakes (Rissa tridactyla) in Newfoundland, Canada (Regehr & Montevecchi 1997). Also, diminishing availability of fish discards, together with a reduction of pelagic stocks, caused increased predation rates on seabirds by Great Skuas (Stercorarius skua) (Oro & Furness 2002, Votier et al. 2004). For Great White Pelicans in the Western Cape, it was predicted that reduced availability of food at the pig farm could cause intensified predation on seabird chicks (de Ponte Machado & Hofmeyr 2004). Until 2004, predatory behaviour was mostly restricted to the proximity of the pelican breeding colony on Dassen Island, but from 2005 it expanded to new localities (e.g. at Rondevlei and Malgas Island, with prospecting for food occurring at Robben and Dyer islands). Pelicans also targeted a wider array of species and increased the intensity and impact of predation on the prey species. According to Votier et al. (2004) the suppression of artificial feeding sources could in the long-term reduce breeding performance and population size of the predator, although their ability to find alternative prey may maintain their populations for some time. Seasonal progression of predation Results show that pelican predation on seabirds is mostly seasonal, peaking in the summer season (October to January), coincidental with the Great White Pelican breeding season in the region (Hockey et al. 2005). Kelp Gulls, and to a lesser extent Cape and Crowned Cormorants, also experience breeding peaks during the austral summer (Hockey et al. 2005). Predatory interactions between pelicans and seabirds attempting to breed in autumn–winter (March to June) were of a much smaller magnitude compared to summer. This suggests that predation on seabirds could be triggered by the elevated energy requirements of the pelican chick raising period, coupled with intraspecific competition for food in accessible water bodies. Further research is also needed to corroborate this theory. Coloniality and other prey species defensive mechanisms Colonial breeding has been proposed as a mechanism to reduce predation pressure on birds and their eggs and chicks (Wittenberger & Hunt 1985, Siegel-Causey & Kharitonov 1990, Clode 1993, Danchin & Wagner 1997). Reproductive synchrony would produce a superabundance of chicks, and satiation of the predator would allow some to survive (Darling 1938, Gochfeld 1982). In this study, colonial breeding appeared ineffective at improving the individual survival probability of chicks in the face of pelican voracity and predator–prey ratios; breeding rates were extremely low with cer-

tain species at some of the monitored islands experiencing near total breeding failure for the season. It has also been postulated that a high concentration of chicks and eggs in a reduced area acts as an attraction to predators (Becker 1995; Hernández-Matías & Ruiz 2003). This was apparently true for this study, with pelicans prospecting for foraging sites and identifying conspicuous aggregations of prey species. They were able to identify prey availability and optimum chick size in the different colonies, eating all the chicks in a cluster before moving to the next one. Nest location was fundamental with regard to protection from predatory pelicans. Crowned Cormorants that selected sites out of reach of pelicans on the top of trees or high structures were safe from pelican predation and fledged many chicks. At Jutten and Schaapen islands, Cape Cormorants breeding on isolated high rocks or ledges – where pelicans were unable to land – were also able to fledge some chicks. Kelp Gulls chicks that survived predation were those near by to hiding places. Furthermore, Cape Cormorants colonies located in close proximity to the researcher’s housing on Malgas Island showed normal recruitment levels, as pelicans would avoid human presence. After over a decade of pelican predatory pressure on seabirds on Dassen Island we could expect some adaptive responses on the behaviour of the prey species. On this island, Cape Cormorants seem to be selecting more protected breeding sites, often building nests under rocks (J. Visagie pers. comm.). Also, the Cape Cormorant breeding population on Dassen Island has shown a sharp decline, while other sites (e.g. Dyer Island) show an increase in the number of breeding pairs (Cape Nature unpublished data). However, pelican predation is not the only factor that could have instigated this numerical shift. Fluctuations in the distribution and abundance of pelagic fish, the main prey of Cape cormorants, may also have contributed. Conclusions and management interventions As discussed earlier, feeding on live avian flesh is not unique behaviour among pelicans, although most incidences of pelican predation outside the study area were anecdotal, isolated cases. Seabird chicks have become an important part of the diet of Great White Pelicans in the Western Cape. It has been suggested that predation is a learned behaviour that has expanded in the population by cultural transmission, triggered by an artificially increased population and scarcity of food during the breeding season near the colony. This behaviour, currently expanding, is causing increasing concern for the conservation of resident seabirds. Pelican dispersal and movements are not fully understood, but this behaviour could be exported to other areas of their distribution. Due to the delicate conservation status of the prey species, intervention may be necessary to counteract the negative impact of pelican predation on local breeding seabird populations, especially considering that agricultural offal has triggered the pelican population expansion, and therefore may be an underlying cause of the intensification of pelican’s predatory behaviour. Different management strategies have been considered, and some preliminary trials performed at Dassen and Malgas islands in 2006/07 (de Ponte Machado et al. in prep, Musangu et al. in prep). Here I compile some of the management options, with the aim of providing a preliminary platform to discuss their effectiveness, pros and cons and ethical considerations. They include: (a) cutting down subsidised sources of food; (b) preventing pelicans from breeding; (c) culling pelicans; (d) negative conditioning for pelicans; (e)

active chasing of pelicans in order to prevent predation; and (f) building artificial structures and enclosures to protect prey species. Cutting down subsidised sources of food may in the long term be an adequate measure to reduce the artificially increased pelican population. Pons & Migot (1995) measured body condition, fecundity and adult survival rates in a population of Herring Gulls (Larus argentatus) after the closure of a refuse tip that provided abundant food for the species. They found a sharp decrease in fecundity while adult survival rates remained unchanged. This result is consistent with the predictions of life-history theory, which predicts that in longlived species a decrease in food supply should affect fecundity before affecting adult survival (Ashmole 1963, Goodman 1974, Stearns 1976, Birkhead & Furness 1985). In the short term however, due to the capacity of the pelicans to switch prey, the elimination of artificial sources of food it will (and has) contributed to increased predatory pressure on nesting seabirds. Therefore, although this is a worthwhile recommendation, it should not be implemented in isolation but combined with other measure/s. Preventing pelicans from breeding on Dassen Island has been presented as a solution to reduce pelican numbers towards a more ‘natural’ pre-subsidised level, by reducing recruitment into the population, but more effectively by removing the higher energetic demands of brooding pelicans. However, pelicans are able to change breeding sites in response to disturbance, as shown by their historical breeding distribution (Crawford et al. 1995). This action could result in exporting the problem to other islands; therefore this option should be considered only in combination with other management actions. Removing avian predators (i.e. gulls) from an area has been described as having positive effects on the reproductive success of threatened seabird species (Guillemette & Brousseau 2001). However, proper evaluations of the effects of such programmes are scarce, in some cases causing undesirable secondary effects like transference of the problem to neighbouring sites (Vidal et al. 1998), high costs in term of manpower and material resources (Bosch et al. 2000), or demonstrated reduced success in the long term (Oro & Martínez-Abraín 2007). The implementation of such a management strategy to control pelican predation would be a controversial and highly unpopular measure. Additional data on the numbers and identity of pelicans involved in predation events should be obtained to assess whether it will be feasible to reduce or eradicate the behaviour from the local populations by removing individuals. Attempts at identifying ‘rogue’ pelicans during this study were not successful due to the low number of marked pelicans involved in predation events. Also a number of ethical, legal and conservationrelated considerations need to be explored. Pelicans are classified as protected species both internationally and in South Africa. Furthermore, pelicans are charismatic animals with a vast potential for use as flagship species. Any measure involving pelican culling would require large amounts of additional effort and funds to for an information campaign justifying the action to the public. Non-lethal methods to reduce predation by birds have been used successfully to modify the feeding behaviour of species to meet management objectives, mostly for ravens and starlings (Avery et al. 1995, Cowan et al. 2000, Neves et al. 2006). Negative conditioning of pelicans using bad tasting seabird chicks or other kinds of deterrents present logistical, technical and economical challenges, given the susceptibility of breeding seabirds to disturbance (this action would involve manipulating chicks on the nest) and the large area and number of islands that would need to be covered to make this measure effective. However, it may be worthTop Predators of the Benguela System

139

while exploring the efficacy of such intervention on Malgas Island, to deter pelicans and gulls from feeding on gannet’s eggs and chicks. Of the species targeted by the pelicans, gannets enjoy the highest priority for conservation actions due to their current decline in population numbers. During November and December 2006, MCM personnel and researchers tested the effectiveness of chasing pelicans away from cormorant and gannet colonies at Malgas Island, in order to prevent them from eating their chicks. The preliminary results were positive, as it is possible to chase pelicans off the colonies without drastic disturbance to nesting birds. Human presence is often enough to keep pelicans at a safe distance. However, constant monitoring of the pelicans’ movements is necessary, as they tend to return quickly after people are out of sight. This would therefore require a permanent presence on the islands by two or more people wholly dedicated to monitoring pelican activities. Not all islands (or parts of them) are suitable for this type of intervention, so balanced decisions will have to be made on priority prey species and areas. Both a scientific assessment on the effectiveness of this management intervention, and monitoring of the breeding success of prey populations in the absence of pelican predation, will be necessary. At Dassen Island, two fenced perimeters were erected in 2006 and Kelp Gull breeding performance was monitored inside and outside the fenced areas. Breeding success was found to be slightly higher inside than outside the enclosure (details in Musangu et al. in prep). Building artificial structures for cormorants to nest on is another option to be considered. The design should mimic the topographic characteristics and location of the nest sites where chicks fledged successfully at the monitored islands; and any evident shortcomings (due for example to human disturbance or flawed design) should be carefully evaluated. Careful experimental design is required to avoid bias due to human disturbance while erecting the structures or monitoring the nests (Prieto et al. 2003). Based on this discussion, it is recommended that management interventions should include a combination of the following: eliminating artificial food supplies for pelicans; chasing of pelicans from the breeding grounds of threatened species; and designing artificial structures to provide protection from predation to prey species. Further, an integrated ecosystem-based form of management that addresses food shortage problems for prey species (gannets and cormorants) and habitat restoration in order to improve the conservation of threatened seabirds is advocated. The aim of this paper is to assist the decision-making process, contributing data and an ecological insight to the problem of pelican predation. The solution/s to this conservation conundrum, in which a protected species negatively affects the conservation of other threatened species, is not simple, and should be carefully examined. There is also wide scope for further research and hypothesis testing. Management strategies should be further discussed and decisions be made by the pertinent authorities. Acknowledgements – Thanks to Les Underhill, Peter Ryan, Rob Crawford and Steve Kirkman for comments on an earlier drafts of this manuscript. Thanks to Steve Kirkman for providing the map. I am grateful to Mwema Musangu, Sipho Jivindava, Ralf Mullers, Lorien Pichegru, Colin and Pam Codner, Azwianewi Makhado and Eddie Papier for observations on pelican numbers and predatory behaviour; and to Pierre Nel, Tony Williams, Phil Hockey, Bruce Dyer, Johan Visagie and Anton Wolfaardt for insights and discussion on pelican predation and management options. I am indebted to SANPARKs for transport, accommodation and support on the islands; and to the project LMR/EAF/03/02 of the Benguela Current Large Marine Ecosystem (BCLME) Programme for funding. Many other people contributed with their observations and comments,

140


which have helped to shape some of the ideas included in this manuscript: thanks!

References Altmann, J. 1974. Observational study of behaviour: sampling methods. Behaviour 49: 227–267. Ashmole, N.P. 1963. The regulation of numbers of tropical oceanic birds. Ibis 103b: 458–473. Avery, M.L., Pavelka, M.A., Bergman, D.L., Decker, D.G., Knittle, C.E. & Linz, G.M. 1995. Aversive conditioning to reduce raven predation on California least tern eggs. Colonial Waterbirds 18: 131–138. Barnes, K.N. 2000. The Eskom Red Data Book of birds of southern Africa, Lesotho and Swaziland. BirdLife South Africa, Johannesburg. BirdLife International. 2000. Threatened birds of the world. Lynx Editions & BirdLife International, Barcelona & Cambridge. BBC News. 2006. Pelican swallows pigeon in park. 25 October 2006 http://news.bbc.co.uk/2/hi/uk_news/england/london/ 6083468.stm Becker, P.H. 1995. Effect of coloniality on gull predation on common tern (Sterna hirundo) chicks. Colonial Waterbirds 18: 11–22. Berry, H.H. 1976. Physiological and behavioural ecology of the Cape Cormorant Phalacrocorax capensis. Madoqua 9(4): 5–55. Berry H.H. & Berry C.U. 1975. A check list and notes on the birds of Sandvis, South West Africa. Madoqua 9(4): 5–55. Birt, V.L., Birt, T.P. Goulet, D., Cairns, D.K. & Montevecchi, W.A. 1987. Ashmole’s halo: direct evidence for prey depletion by a seabird. Marine Ecology Progress Series 40: 205–208. Birkhead, T.R. & Furness, R.W. 1985. The regulation of seabird populations. In: R.M. Sibly & R.H. Smith (eds.) Behavioural Ecology. Blackwell Scientific Publications, Oxford, pp 145–167. Blokpoel, H. & Spaans, A.L. 1991. Superabundance in gulls, causes, problems and solutions. Acta XX Congressus internationalis ornithologici, vol. 4. New Zealand Ornithological Congress Trust Board, Wellington, New Zealand, pp 2359–2398. Bosch, M., Oro, D., Cantos, F.J. & Zabala, M. 2000. Short–term effects of culling on the ecology and population dynamics of the Yellow-legged Gull. Journal of Applied Ecology 37: 369–385. Bosch, M., Oro, D. & Ruiz, X. 1994. Dependence of Yellow-legged Gulls (Larus cachinans) on food from human activity in two western Mediterranean colonies. Avocetta 18: 135–139. Burleigh, J. 2006. Pelican makes a meal out of a pigeon. Weekly Telegraph, 1–7 November 2006. Clode, D. 1993. Colonially breeding seabirds: predators or prey? Trends in Ecology and Evolution 8: 336–338. Cowan, D.P., Reynolds, J.C. & Gill, E.L. 2000. Manipulating predatory behaviour through conditioned taste aversion: can it help endangered species? In: Gosling, L.M., Sutherland, W.J., Avery, M. (eds.) Behaviour and conservation. Cambridge University Press, Cambridge, UK, pp 281–299. Crawford, R.J.M., Cooper, J. & Dyer, B.M. 1995. Conservation of an increasing population of Great White Pelicans Pelecanus onocrotalus in South Africa’s Western Cape. South African Journal of Marine Science 15: 33–42. Crawford, R. J. M., Dundee, B. L., Dyer, B. M., Klages, N. T. W., Meÿer, M. A. & Upfold, L. 2007. Trends in numbers of Cape Gannets (Morus capensis), 1956/57–2005/06, with a consideration of the influence of food and other factors. ICES Journal of Marine Science 64: 169–177. Crawford, R.J.M., Nel, D.C., Williams, A.J. & Scott, A. 1997. Seasonal patterns of abundance of Kelp Gulls Larus dominicanus at breeding and non-breeding localities in southern Africa. Ostrich 68: 37–41. Crivelli, A., Catsadorakis, G., Hatzilacou, D., Hulea, D., Malakou, M., Marinov, M., Mitchev, T., Nazirides, T., Peja, M., Sarigul, G. & Siki, M. 1998. Status and population development of Great White Pelican Pelecanus onocrotalus and Dalmatian Pelican P. crispus breeding in the Paleartic. In: P. Yésou, and J. Sultana (eds.) Monitoring and conservation of birds, mammals and sea turtles of the Mediterranean and Black Seas. Proceedings of the 5th Medmaravis Symposium. Environment Protection Department, Gozo, Malta, pp 38–45. Danchin, E. & Wagner, R.H. 1997. The evolution of coloniality: the

emergence of new perspectives. Trends in Ecology and Evolution 12: 342–347. Darling, F.F. 1938. Bird flocks and breeding cycles. Cambridge University Press, Cambridge. Del Hoyo, J., Elliott, A. & Sargatal, J. 1992. Handbook of the Birds of the World vol. 1. Lynx Edicions, Barcelona. De Ponte Machado, M. & Hofmeyr, J. 2004. Great White Pelicans – waterbirds or farm birds? Bird Numbers 13: 11–12. De Ponte Machado, M., et al. (in prep.). Great White Pelican predation on South African West Coast Islands and its impact on seabird populations. De Ponte Machado, M., Crawford, R.J.M., Underhill, L.G., Hofmeyr, J. (in prep.). Trends in the population of Great White Pelicans in South Africa’s Western Cape – role of human subsidisation. De Villiers, M., le Roux, J. & du Toit, M. 2002. The islands of Saldahna Bay and Langebaan Lagoon: a trip in time. http:// web.uct.ac.za/depts/stats/adu/travel/island-trip1.htm Du Toit, M., Boere, G.C., Cooper, J., de Villiers, M.S., Kemper, J., Lenten, B., Petersen, S.L., Simmons, R.E., Underhill, L.G., Whittington, P.A. & Byers, O.P. (eds.) 2003. Conservation Assessment and Management Plan for Southern African Coastal Seabirds. Avian Demography Unit, Cape Town & Conservation Breeding Specialist Group, Apple Valley, USA. Emslie, S.D., Karnovsky, N. & Trivelpiece, W. 1995. Avian predation at penguin colonies on King George Island, Antartica. Wilson Bulletin 107: 317–327. Furness R.W. 1987. The skuas. Peyser, Calton. Furness R.W. & Birkhead, T.R. 1984. Seabird colony distributions suggest competition for food supplies during the breeding season. Nature 311: 655–656. Finney, S.K., Harris, M.P., Keller, L.F., Elston, D.A., Monaghan, P. & Wanless, S. 2003. Reducing the density of breeding gulls influences the pattern of recruitment of immature Atlantic Puffins Fratercula arctica to a breeding colony. Journal of Applied Ecology 40: 545–552. Gochfeld, M. 1982. Reproductive synchrony and predation satiation: an analogy between the Darling Effect in birds and mast fruiting in plants. Auk 99: 586–587. Goodman, D. 1974. Natural selection and cost ceiling on reproductive effort. American Naturalist 108: 247–268. Guillemette, M. & Brousseau, P. 2001. Does culling predatory gulls enhance productivity of breeding Common Terns? Journal of Applied Ecology 38: 1–8. Guillet, A. & Crowe, T.M. 1981. Seasonal variation in group size and dispersion in a population of Great White Pelicans. Gerfaut 71: 185–194. Guillet, A. & Furness, R.W. 1985. Energy requirements of a Great White Pelican (Pelecanus onocrotalus) population and its impacts on fish stocks. Journal of Zoology of London 205: 573–583. Groombridge, B. 1994. IUCN Red List of Threatened Animals. IUCN, Gland. Green, L.G. 1947. Tavern of the seas. Howard B. Timmins. Cape Town, South Africa. Green, L.G. 1950. At daybreak for the islands. Howard B. Timmins. Cape Town, South Africa. Hartley-Brewer, J. 1999. Murder most fowl. Evening Standard (London), 11 June 1999. Hernández-Matías, A. & Ruiz, X. 2003. Predation on Common Tern eggs by the Yellow-legged Gull at Ebro Delta. Scientia Marina 67: 95–101. Hockey, P.A.R., Dean, W.R.J. & Ryan, P.G. (eds.) 2005. Roberts Birds of Southern Africa, 7th ed. Trustees of the John Voelcker Bird Book Fund, Cape Town. Hunter, S. & Brooke, M. de L. 1992. The diet of giant petrels Macronetes spp. at Marion Island, Southern Indian Ocean. Colonial Waterbirds 15: 56–65. Hunt G.L., Eppley, Z.A., Scheider, D.C. 1986. Reproductive performance of seabirds: the importance of population and colony size. Auk 103: 306–317. Johnsgard, P.A. 1993. Cormorants, Darters, and Pelicans of the World. Smithsonian Institution Press, Washington and London. Knechtel, H.A. 2003. Brown Pelican disturbances at two central California Common Murre colonies. Pacific Seabirds 30: 42. Layard, E.L. & Sharpe, R.B. 1884. The birds of South Africa. Quaritch, London. Lewis, S., Sherratt, T.N., Hamer, K.C. & Wanless, S. 2001. Evidence

of intra-specific competition for food in a pelagic seabird. Nature 412: 816–819. Lehner, P.N. 1996. Handbook of ethological methods. 2nd edition. Cambridge University Press. Cambridge, UK. Martínez-Abraín, A., González-Solis, J., Pedrocchi, V., Genovart, M., Carles Abella, J., Ruiz, X., Jiménez, J., Oro, D. 2003. Kleptoparasitism, disturbance and predation of yellow-legged gulls on Audouin’s gulls in three colonies of the western Mediterranean. In: E. Mínguez, D. Oro, E. de Juana & A. MartínezAbraín (eds.) Mediterranean seabirds and their conservation. Vol. 67 (2): 89–94. Migot, P. 1992. Demographic changes in French Herring Gull Larus argentatus populations, a modelling appracah and hypothesis concerning the regulation of numbers. Ardea 80: 161–169. Mora, M. A. 1989. Predation by a Brown Pelican at a mixed-species heronry. Condor 91: 742–743. Musangu, M., de Ponte Machado, M. & Ryan, P.G. (in prep.). The effect of Great White Pelican predation on breeding seabirds at Dassen Island, South Africa. Neves, V.C., Panagiotakopoulos, S. & Furness, R.W. 2006. A control taste aversion experiment on predators of roseate tern (Sterna dougallii) eggs. European Journal of Wildlife Research 52:259–264. Oro, D. 1999. Trawler discards: a threat or a resource for opportunistic seabirds? In: Adams, N.J. & Slotow, R.H. (eds.) Proceedings 22 International Ornithological Congress, Durban BirdLife South Africa, Johannesburg, pp 717–730. Oro, D. & Furness, R.W. 2002. Influences of food availability and predation on survival of kittiwakes. Ecology 83: 2516–2528. Oro, D. & Martínez-Abraín, A. 2007. Deconstructing myths on large gulls and their impact on threatened sympatric waterbirds. Animal Conservation 10: 117–126 Phillips R.A., Thompson, D.R. & Hamer, K.C. 1999. The impact of Great Skua predation on seabird populations at St Kilda: a bioenergetics model. Journal of Applied Ecology 36: 218–232. Pons, J.M. & Migot, P. 1995. Life-history strategy of the herring gull, changes in survival and fecundity in a population subjected to various feeding conditions. Journal of Animal Ecology 64: 592– 599. Prieto, J., Jover, L. & Ruiz, X. 2003. Effect of refuge supplement on Andouin’s Gull chick survival. Scientia Marina 67: 103–108. Regehr, H.M. & Montevecchi, W.A. 1997. Interactive effects of food shortage and predation on breeding failure of black-legged kittiwakes: indirect effects of fisheries activities and implications for indicator species. Marine Ecology Progress Series 155: 249–260. Siegel-Causey, D. & Kharitonov, S.P. 1990. The evolution of coloniality. In: D.M. Power (ed.) Current Ornithology, vol. 7. Plenum Press, New York, pp 285–330 Sol, D., Arcos, J.M. & Senar, J.C. 1994. Do Yellow-legged Gulls (Larus cachinans) use refuse tips whenever they need to? Miscellània zoològica 17: 199–203. Stearns, S. 1976. Life-history tactics: a review of the ideas. Quarterly Review of Biology 51: 3–47. Sutherland, W.J. 1998. The importance of behavioural studies in conservation biology. Animal Behaviour 56: 801–809. Sutherland, W.J. 2006. Predicting the ecological consequences of environmental change: a review of the methods. Journal of Applied Ecology 43: 599–616. Vidal, E., Medail, F. & Tatoni, T. 1998. Is the Yellow-legged Gull a superabundant bird species in the Mediterranean? Impact on fauna and flora, conservation measures and research priorities. Biodiversity and Conservation 7: 1013–1026. Votier, S.C., Furness, R.W., Bearhop, S., Crane, J.E., Caldow, R.W.G., Catry, P., Ensor, K., Hamer, K.C., Hudson, A.V., Kalmbach, E., Klomp, N.I., Pfeiffer, S., Phillips, R.A., Prieto, I. & Thompson, D.R. 2004. Changes in fisheries discard rates and seabird communities. Nature 427: 727–730. Votier, S.C., Bearhop, S., Crane, J.E., Arcos, J.M. & Furness, R.W. 2007. Seabird predation by Great Skuas Stercorarius skua – intraspecific competition for food? Journal of Avian Biology 38: 234–246. Vucetich, J.A. & Nelson, M.P. (in press). What are 60 warblers worth? Killing in the name of conservation. Oikos. Weidinger, K. 1998. Effect of predation by skuas on breeding success of the Cape petrel Daption capense at Nelson Island, Antarctica. Polar Biology 20: 170–177.


141

Western Cape Nature Conservation Board. 1995. Predation of Kelp Gull and Cape Cormorant chicks by Great White Pelicans on Dassen Island. Unpublished report, Cape Town, South Africa. Williams, A.J. & W.D. Borello. 1997. White Pelican Pelecanus onocrotalus. In: Harrison, J.A., D.G. Allan, L.G. Underhill, M. Herremans, A.J. Tree, V. Parker & C.J. Brown (eds.) The atlas of southern African birds vol 1: Non-Passerines. BirdLife South Africa, Johannesburg, pp 736–737.

142


Wittenberger & Hunt. 1985. The adaptive significance of coloniality in birds. In: Famer, King Parker (eds.) Avian Biology, vol. 8. Academic Press, pp 1–78. Yorio, P. & Quintana, F. 1997. Predation by Kelp Gulls Larus dominicanus at a mixed-species colony of royal terns Sterna maxima and Cayenne Terns Sterna eurygnatha in Patagonia. Ibis 139 (3): 536–541.

Cape Gannet

169

Trends in numbers of Cape gannets (Morus capensis), 1956/1957 – 2005/2006, with a consideration of the influence of food and other factors Robert J. M. Crawford, Benedict L. Dundee, Bruce M. Dyer, Norbert T. W. Klages, Michael A. Meyër, and Leshia Upfold Crawford, R. J. M., Dundee, B. L., Dyer, B. M., Klages, N. T., Meyër, M. A., and Upfold, L. 2007. Trends in numbers of Cape gannets (Morus capensis), 1956/57– 2005/06, with a consideration of the influence of food and other factors – ICES Journal of Marine Science, 64, 169 – 177.

Cape gannets (Morus capensis) breed at six colonies in Namibia and South Africa. Population size averaged about 250 000 pairs over the period 1956/1957 – 1968/1969 and about 150 000 pairs from 1978/1979 to 2005/2006. Over the whole 50-y period, numbers at the three Namibian colonies fell by 85 –98%, with greater proportional decreases in the south. There were increases at two South African colonies between 1956/1957 and 2005/2006. The colony at Lambert’s Bay increased between 1956/1957 and 2003/2004, but attacks by Cape fur seals (Arctocephalus pusillus) on birds at nests caused abandonment of the entire colony in 2005/2006. Longterm changes at colonies are thought to be largely attributable to an altered abundance and distribution of prey, especially sardine (Sardinops sagax) and anchovy (Engraulis encrasicolus). In both Namibia and South Africa, the numbers of Cape gannets breeding were significantly related to the biomass of epipelagic fish prey. Over the 50-y period, there was also a marked similarity in the proportions of gannets and epipelagic fish in the Benguela system, which were present in Namibia and South Africa. In the 2000s, there was an eastward shift in the distribution of sardine off South Africa and a large increase in the number of gannets breeding at South Africa’s easternmost colony. When sardine were scarce off South Africa, gannets fed on anchovy, but off Namibia anchovy only temporarily and partially replaced sardine. Ecosystem management measures that might improve the conservation status of Cape gannets are considered. Keywords: anchovy, Arctocephalus pusillus, Cape fur seal, Cape gannet, distribution, food, long-term change, Morus capensis, sardine. Received 23 May 2006; accepted 7 September 2006; advance access publication 2 November 2006. R. J. M. Crawford: Department of Environmental Affairs and Tourism, Marine and Coastal Management, Private Bag X2, Rogge Bay 8012, South Africa, and. R. J. M. Crawford: Avian Demography Unit, Department of Statistical Sciences, University of Cape Town, Rondebosch 7701, South Africa.. B. L. Dundee: Ministry of Fisheries and Marine Resources, PO Box 394, Lu¨deritz, Namibia.. B. M. Dyer, M. A. Meyër and L. Upfold: Department of Environmental Affairs and Tourism, Marine and Coastal Management, Private Bag X2, Rogge Bay 8012, South Africa.. N. T. W. Klages: Environmental and Coastal Management, PO Box 77000, Nelson Mandela Metropolitan University 6031, South Africa.. Correspondence to R. J. M. Crawford: tel: þ27 21 4023140; fax: þ27-21-4217406; e-mail: [email protected].

Introduction The Cape gannet (Morus capensis) is one of three gannet species worldwide, the others being the North Atlantic gannet (M. bassanus) and the Australasian gannet (M. serrator) (Nelson, 2002). The Cape gannet, regarded as vulnerable (BirdLife International, 2004), has bred at 10 localities off the coasts of Namibia and South Africa, but at only six of these since 1956: Mercury, Ichaboe, and Possession Islands in Namibia; Bird (Lambert’s Bay), Malgas (both in the Western Cape), and Bird (Algoa Bay, Eastern Cape) Islands in South Africa (Figure 1) (Crawford et al., 1983b). Information on the numbers of gannets at these six localities is available for the 50-y period 1956/1957–2005/2006 and is reported in this paper. Cape gannets feed mainly on sardine (Sardinops sagax) and anchovy (Engraulis encrasicolus) (references in Hockey et al., 2005), species that are also exploited by the purse-seine fisheries of Namibia and South Africa. Industrial fisheries can affect predator populations adversely through competition for shared prey

(Frederiksen et al., 2004). Here we examine the influence of changes in the abundance and distribution of prey, and other factors, on the numbers of Cape gannets and consider ecosystem management measures that might improve the conservation status of the species. Recent declarations by international summits have emphasized the need to account for ecosystem issues in fisheries management, including the requirements of predators that are dependent on species targeted by a fishery. Such policy has been incorporated into law by South Africa (Crawford, 2004).

Material and methods The numbers of breeding pairs of Cape gannets at colonies were estimated from measurements of the area occupied by breeding birds on aerial photographs that were taken vertically, combined with measures of the densities of nests (Klages et al., 1992). Photographs were taken in November or December, when most birds are incubating or brooding, using methods described by Shelton et al. (1982). The extent of the area occupied by

# 2006

International Council for the Exploration of the Sea. Published by Oxford Journals. All rights reserved. For Permissions, please email: [email protected] Top Predators of the Benguela System

145

170

R. J. M. Crawford et al.

Figure 1. Southern Africa, showing the six extant colonies of Cape gannets and other localities mentioned in the text. breeding birds was measured using an Ibas interactive image-analysis system. The photographs were scaled from ground measurements of straight edges along walkways, walls, or buildings near the colonies. In 2004/2005, the outer limit of the gannet colony at Lambert’s Bay was delimited using a Global Positioning System that had a circular error probability of 4 m for the horizontal position. Estimates of the area occupied by breeding Cape gannets were obtained for 152 of the 300 possible locality/season combinations, with coverage sporadic before the 1980s, but better subsequently. Measurements of the densities of nests at colonies were undertaken during breeding seasons by placing four poles, each 2 m long, on the surface of the ground, so as to form a square of 4 m2. The numbers of whole nests and part nests within the square were counted. The overall number of nests in the square was taken to be the number of whole nests plus half the number of part nests. The number of measurements made in any season ranged from two at Possession Island in 2002/2003 and 2005/2006, when it was desired to minimize disturbance, to 30 at Algoa Bay in 2005/ 2006 (mean 17 per locality per season; n ¼ 33). Estimates of the density of nests were obtained for 11 seasons at Malgas Island, 10 at Lambert’s Bay, 4 at Possession Island, 3 each at Mercury and Ichaboe Islands, and 2 at Algoa Bay. Additionally, three published estimates were available for Algoa Bay (Randall and Ross, 1979; Batchelor 1982). However, information on density was available for just 23% of the locality/season combinations for which the 146


area occupied by breeding birds was measured. Therefore, for each locality, a mean density was obtained, by giving equal weight to each season for which information existed, and applied throughout the study period. Trends in the numbers of Cape gannets nesting in Namibia and South Africa were compared with trends in the biomass of sardine and anchovy in these regions using correlation analysis. The Namibian and South African stocks of sardine and anchovy are thought to be relatively discrete (Crawford et al., 1987). Estimates of biomass for Namibia were obtained from virtual population analysis (VPA) for the period 1956/1957–1988/1989. From 1990/1991, they were obtained by hydroacoustic surveys. For 1956/1957–1982/1983, estimates for South Africa were obtained by VPA. From 1984/1985, estimates were obtained from hydroacoustic surveys (sources in Schwartzlose et al., 1999). No estimates of the abundance of anchovy are available prior to the mid-1960s, when fisheries for this species commenced. However, anecdotal information suggests that anchovy were not abundant off southern Africa in the 1950s and early 1960s (Crawford et al., 1987), so biomass of the species was assumed to be negligible, as it was also off Namibia from 1990 on. Because the estimates of biomass derived from VPA and hydroacoustic surveys are not strictly comparable, and because of data gaps in the time-series of information, especially of gannet numbers and also of fish biomass, it was not possible to prewhiten the time-series prior to their being cross-correlated.

171

50-year trends in Cape gannet numbers However, the long gaps between observations up until 1980 preclude the likelihood of serial correlation in the time-series then. In addition to the original data values, decadal averages of values were also correlated. Further, to gauge the influence of the relative abundance of prey on the distribution of gannets, the proportions of gannets nesting in Namibia and in South Africa were compared with the proportional contribution of these two countries to the overall biomass of sardine and anchovy off southern Africa.

Results Numbers of breeding birds Between 1956/1957 and 2005/2006, there were large decreases in the numbers of breeding Cape gannets at all three Namibian colonies (Table 1). The largest proportional decrease (98%) was at the southernmost colony of Possession Island. The decrease at Ichaboe Island was 95%, and that at Mercury Island, the northernmost extant colony of the species, was 85%. All three colonies showed steep decreases between 1956/1957 and the early 1980s. They then stabilized, before decreasing again in the 1990s and early 2000s. In South Africa, there were increases in the numbers of birds breeding at all three colonies between 1956/1957 and 2003/2004 (Table 1). The number at Lambert’s Bay decreased from 1956/ 1957 to 1969/1970, before increasing to attain a peak of about 14 000 pairs in 1987/1988. From 1988/1989 to 2003/2004, the colony fluctuated between 9 000 and 12 000 pairs. It decreased to half this level in 2004/2005, and there were no birds at the colony on 16 December 2005. The colony at Malgas Island more than doubled in size between 1956/1957 and 1996/1997, with substantial fluctuations during the 1980s and 1990s, but it then decreased again. However, in 2005/2006, there were 40% more birds at the colony than in 1956/1957. At Bird Island in Algoa Bay, the easternmost colony, numbers breeding increased fivefold between 1956/1957 and 2005/2006. In Namibia, the overall number of gannets breeding fell by 95% from 204 000 pairs in 1956/1957 to 10 000 pairs in 2005/ 2006 (Table 1). The number breeding in South Africa increased from 50 000 pairs in 1956/1957 to a peak of 145 000 pairs in 2001/2002, and was 135 000 pairs in 2005/2006. The overall number was about 250 000 pairs in 1956/1957 and in the late 1960s. It averaged 151 000 pairs from 1978/1979 to 2005/2006 (s.d. ¼ 15 000 pairs; n ¼ 13). In 1956/1957, Namibia was home to 80% of the breeding Cape gannets. This fell to 50% in 1978/1979 and to 7% in 2005/2006. The proportion breeding in the Western Cape rose from 12% in 1956/1957 to 21% in 1978/1979, then to 25% in 2005/2006. Algoa Bay had 7% of breeding birds in 1956/1957, 28% in 1978/ 1979, and 68% in 2005/2006. Hence, there has been a shift to the south and east in the centre of the distribution of Cape gannets. The largest number breeding at an individual locality was 175 000 pairs at Ichaboe Island in 1956/1957; the smallest was 351 pairs at Possession Island in 2005/2006 (Table 1).

Density of nests Average numbers of nests counted per square metre (weighting seasons equally) were: Mercury Island 3.73 (s.d. 0.58; n ¼ 3), Ichaboe Island 3.56 (s.d. 0.67; n ¼ 3), Possession Island 4.08 (s.d. 1.11; n ¼ 4), Lambert’s Bay 3.22 (s.d. 0.43; n ¼ 10), Malgas Island 2.84 (s.d. 0.19; n ¼ 11), and Algoa Bay 2.71 (s.d. 0.22; n ¼ 2). The highest average density recorded for any colony in a

season was 5.60 nests m22 at Possession Island in 1978/1979 and the lowest 2.54 nests m22 at Malgas Island in 1999/2000 (Figure 2). There was a suggestion of a decrease in densities of nests at the three Namibian colonies between 1978/1979 and 2002/2003, although there was a large coefficient of variation (42%) for Mercury Island in 1978/1979, and measurements were conducted in few seasons (Figure 2). If there was a long-term decrease, the decrease in the number of birds breeding in Namibia will have been underestimated. Average densities measured at the three South African colonies fluctuated, but showed no trend.

Relationship between numbers breeding and biomass of fish Numbers of gannets breeding in Namibia were significantly correlated with the biomass there of sardine (r ¼ 0.900; n ¼ 18; p , 0.001) and of sardine and anchovy combined (r ¼ 0.928; n ¼ 18; p , 0.001). Similarly, in South Africa, the numbers of gannets breeding were significantly correlated with the biomass of sardine (r ¼ 0.535; n ¼ 22; p , 0.02) and of sardine and anchovy combined (r ¼ 0.661; n ¼ 22; p , 0.001). In Namibia, the decadal averages of numbers of gannets breeding and fish biomass both decreased steeply after the 1960s (Figure 3a). The two series were significantly correlated (r ¼ 0.945; n ¼ 6; p , 0.005). In South Africa, the decadal average of numbers of gannets breeding increased almost linearly over the study period, whereas the average biomass of fish increased only after the 1970s (Figure 3b). The two series were significantly correlated (r ¼ 0.823; n ¼ 6; p , 0.05). There were large changes in the biomass of epipelagic fish off Namibia in the 1960s and off South Africa since 2000, as indicated by the large standard deviations (Figure 3). The proportions of Cape gannets breeding in Namibia and in South Africa showed marked similarity to the proportional contribution of these two countries to the overall biomass of sardine and anchovy in the southern African region (Figure 4). Until 1968/1969, Namibia held most of the gannets (80%) and most of the epipelagic fish (85%). In 1978/1979, the numbers of gannets and the biomass of epipelagic fish in the two countries were almost equivalent. This was approximately the case until 1982/ 1983, after which the proportions of gannets and fish in Namibia and South Africa decreased and increased, respectively. In 2005/ 2006, 93% of the gannets and 97% of the combined biomass of sardine and anchovy were in South Africa.

Discussion Population size Whereas numbers of Cape gannets decreased in the 20th century, numbers of North Atlantic and Australasian gannets both increased (Montevecchi and Myers, 1997; Nelson, 2002; Wanless et al., 2005). North Atlantic gannets increased at a rate of about 3% through the 20th century, and now number some 343 000 pairs, making it the world’s most abundant gannet (Nelson, 2002). The population of Australasian gannets in New Zealand increased from 21 000 pairs in 1946/1947 to 46 000 pairs in 1980/1981 (Wodzicki et al., 1984), and numbers have continued to grow (Bunce et al., 2002). There were 6 600 pairs breeding in Australia in 1980/81, which increased to about 20 000 pairs at the end of the 20th century, a rate of increase of 6% per annum (Bunce et al., 2002). The total population of Australasian gannets was about 53 000 pairs in 1990 and about 66 000 pairs now, making it the least numerous of the world’s three gannets (Nelson, 2002). Top Predators of the Benguela System

147

172


Table 1. Pairs of Cape gannets estimated to be breeding at the six extant colonies, 1956/57 –2005/06. Season

Mercury

Ichaboe

Possession

Total for Lambert’s Malgas Algoa Total for Total for Namibia Namibia Bay Bay South Africa and South Africa 1956/57 9 396 175 116 19 258 203 770 5 915 25 040 19 092 50 047 253 817 . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1957/58 – – – – – – – – – . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1958/59 – – – – – – – – – . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1959/60 – – – – – – 24 683 – – . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1960/61 – – – – – – – – – . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1961/62 – – – – – – – – – . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1962/63 – – – – – – – – – . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1963/64 – – – – – – – – – . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1964/65 – – – – – – – – – . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1965/66 – – – – – – – – – . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1966/67 – – – – – – – – – . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1967/68 5 147 166 601 16 296 188 044 3 711 30 801 – – – . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1968/69 – – – – – – – – 249 426* . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1969/70 2 835 – – – 3 149 27 442 30 791 61 382 – . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1970/71 – – – – – – – – – . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1971/72 – – – – – – 28 726 – – . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1972/73 – – – – – – – – – . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1973/74 – – – – – – – – – . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1974/75 – – – – – – 36 791 – – . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1975/76 – – – – – – – – – . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1976/77 – – – – – – – – – . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1977/78 – – – – – – 40 593 – – . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1978/79 5 696 69 908 4 357 79 961 5 595 28 168 45 032 78 796 158 757 . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1979/80 – – – – – – 36 796 – – . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1980/81 – 54 328 3 260 – 7 345 29 546 45 720 82 612 – . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1981/82 2 063 49 683 4 100 55 846 7 570 – – – – . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1982/83 3 364 – 2 942 – 6 888 32 261 49 458 88 606 – . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1983/84 2 380 46 758 3 537 52 675 6 102 19 826 – – – . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1984/85 3 603 56 208 2 383 62 194 10 024 27 980 44 688 82 691 144 885 . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1985/86 2 286 40 704 3 533 46 524 8 832 27 619 45 745 82 196 128 720 . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1986/87 2 480 – – – 11 161 38 895 51 195 101 251 – . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1987/88 2 361 49 261 2 742 54 363 14 265 40 720 51 219 106 204 160 567 . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1988/89 3 279 55 614 3 097 61 990 11 125 41 324 56 111 108 560 170 550 . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1989/90 2 932 58 239 3 040 64 210 10 793 53 194 50 921 114 908 179 118 . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1990/91 – – – – 9 630 29 075 53 376 92 081 – . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1991/92 – – – – 11 834 43 715 56 869 112 418 – . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1992/93 2 760 25 525 3 158 31 443 10 550 42 380 55 945 108 875 140 318 . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1993/94 2 723 31 484 1 399 35 607 10 462 32 320 65 173 107 955 143 562 . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1994/95 1 619 – 922 – 9 705 48 043 65 040 122 788 – . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1995/96 1 992 31 183 767 33 942 9 560 50 109 65 040 124 709 158 651 . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1996/97 1 794 32 657 726 35 177 10 826 56 376 – – – . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1997/98 1 434 23 702 – – 11 065 50 425 63 353 124 844 – . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1998/99 1 358 19 972 339 21 668 10 027 50 960 55 075 116 063 137 731 . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1999/00 – – – – – – – – – . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2000/01 – – – – 11 631 48 058 – – – . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2001/02 1 694 13 268 400 15 362 9 525 48 680 87 257 145 462 160 823 . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2002/03 – – – – 9 364 – 76 571 – – . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2003/04 1 503 11 837 437 13 777 10 529 31 070 78 929 120 528 134 305 . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2004/05 1 910 10 947 477 13 334 4 962 – 87 569 – – . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2005/06 1 414 8 669 351 10 433 0 36 156 98 419 134 575 145 008 For seasons for which there were sufficient data, the number of pairs breeding in Namibia, South Africa, and in total are also shown.

*Assumed to be the sum of those in Namibia in 1967/1968 and those in South Africa in 1969/1970.

148


173

50-year trends in Cape gannet numbers

Figure 2. Densities of nests of Cape gannets at the six extant colonies, for seasons in which these were measured. Where available, the standard deviations of measurements are shown as dotted lines. Information for Mercury, Ichaboe and Possession Islands in 1978/1979 is from Crawford et al. (1983b) and data on which that paper was based. Information for Algoa Bay from 1974/1975 to 1980/1981 is from Randall and Ross (1979) and Batchelor (1982).

The largest present colony of North Atlantic gannets is about 60 000 pairs at St Kilda, Outer Hebrides, Scotland (Wanless et al., 2005). At Bird Rocks, Canada, there were 100 000–125 000 pairs in 1833 (Nelson, 2002). The largest colony of Australasian gannets is 8 000 pairs at Gannet Island, New Zealand (Wodzicki et al., 1984). Therefore, Ichaboe Island (175 000 pairs of Cape gannets in 1956/1957) held the largest known gannetry. The colony of Cape gannets at Bird Island in Algoa Bay (98 000 pairs) is at present the largest gannetry in the world.

Influence of food The significant correlations between the numbers of Cape gannets and the biomass of epipelagic fish in Namibia and South Africa suggest that numbers of gannets are strongly influenced by the abundance of food. Especially in Namibia, the large decrease in abundance of epipelagic prey appears to have had a major impact on the gannet colonies. Changes in prey abundance accounted for

some 86 –89% of the variability in numbers of gannets nesting in Namibia, compared with 44 –68% of the variability in South Africa. Gannets increased in South Africa throughout the study period, whereas estimates of fish biomass increased towards the end of the period. Until the mid-1980s, assessments of fish stocks in South Africa were based on information for the region west of Cape Agulhas only. As the Namibian sardine collapsed, its range contracted to the north, placing the remaining shoals increasingly distant from Cape gannet colonies. Most anchovy also were located well north of the gannet colonies (Crawford et al., 1987). In accord with this altered distribution of prey, Possession Island, the southernmost colony in Namibia, had the greatest decrease after 1956/1957 and Mercury Island, the northernmost, the least decrease. When sardine were decreasing in Namibia, anchovy, because it had replaced sardine in South Africa’s catches, were fished intensively in a deliberate management effort to minimize its possible Top Predators of the Benguela System

149

174


Figure 3. Trends in the average area of Cape gannet breeding colonies and the average biomass of epipelagic fish (sardine and anchovy) in each decade from the 1950s to the 2000s: (a) off Namibia and (b) off South Africa. The standard deviations of the means are shown, where available.

competition with sardine (Crawford et al., 1987). It had a shortlived period of abundance in Namibia, where it never fully replaced sardine. The largest catch of anchovy in Namibia, ,0.4 million tonnes in 1987, was just 26% of the peak catch of

Figure 4. Trends in the proportions of (a) Cape gannets and (b) epipelagic fish (sardine and anchovy) off southern Africa found off Namibia and South Africa, 1956/1957– 2005/2006. 150


1.4 million tonnes of sardine in 1968. Before 1967 and after 1988, Namibia’s annual catch of anchovy never exceeded 0.1 million tonnes. From 1968 to 1988, it averaged 0.18 million tonnes, 32% of Namibia’s average catch of sardine of 0.57 million tonnes from 1956 to 1967 (Schwartzlose et al., 1999; updated). Changes in the numbers of gannets in Namibia accord with this partial and only temporary replacement of sardine by anchovy. From 1981/1982 to 1989/1990, the number of gannets in Namibia fluctuated about a level of 58 000 pairs, 30% of the average of 196 000 pairs up to 1967/1968. In the 1990s, the numbers of gannets breeding in Namibia fell further to an average of 30 000 pairs, and then in the early part of the 21st century to about 13 000 pairs. In Namibia, from 1957 to 1959, sardine contributed 90% by mass of the diet of Cape gannets, whereas anchovy were not recorded. From 1978 to 1982, anchovy contributed 53% by number of the prey items eaten (sources in Hockey et al., 2005). In the same period, anchovy contributed 34% by mass of the diet, whereas sardine contributed ,1% (RJMC, unpublished). From 1989 to 2004, anchovy contributed 7% by mass of the diet and sardine 11% (BLD, unpublished), emphasizing the continuing decrease in availability of these epipelagic fish species to gannets in Namibia. In South Africa, maintenance of the numbers of Cape gannets, in spite of the collapse of South Africa’s stock of sardine, is attributable to a ready availability then of anchovy. In the period 1953–1956 in the Western Cape, sardine contributed 60% by mass of the diet of Cape gannets and anchovy 13%. From 1978 to 1989, sardine formed 18% of the diet and anchovy 44% (sources in Hockey et al., 2005). In every year from 1978 to 1984, sardine contributed ,10% of the diet and anchovy .44% (up to 64%) (Schwartzlose et al., 1999). Anchovy replaced sardine as the dominant species in South Africa’s purse-seine catch from 1966 to 1995, and there is evidence that it increased in the Western Cape in the 1960s following the decrease of the sardine (Schwartzlose et al., 1999). During the 1980s, sardine biomass recovered in South Africa, and it was high by the end of the 20th century. However, after 2000, there was a marked eastward shift in the distribution of the species (Van der Lingen et al., 2005), which led to a decreased availability of this prey species to gannets at Lambert’s Bay and Malgas Island, but an increased availability to birds in Algoa Bay. Numbers of gannets breeding at Malgas Island and Lambert’s Bay decreased, whereas numbers in Algoa Bay showed a sharp increase. In Algoa Bay from 1979 to 1990, sardine contributed 31% by mass of the diet of Cape gannets and anchovy 24% (Klages et al., 1992). From 1993 to 2000, sardine contributed 45% and anchovy 18%. From 2001 to 2006, the values were 72% and 15%, respectively (RJMC and NTW, unpublished). Therefore, unlike the situation in Namibia, the combined contribution of these epipelagic fish species to the diet of gannets has steadily increased in Algoa Bay. It seems evident from the similar trends in the proportions of gannets and epipelagic fish in Namibia and South Africa that large changes in the availability of fish have had a major influence on the numbers of gannets breeding in different areas. The changes are likely to have been primarily responsible for the shift from Namibia to the Eastern Cape in the location of the majority of breeding gannets. The extent to which the shift in the distribution of Cape gannets has been influenced by differences in local rates of

175

50-year trends in Cape gannet numbers production and survival of gannets or by movements of birds is uncertain. Both the production of young birds and the emigration of first-time breeders away from localities where food was scarce to those where conditions were more favourable are likely to have had an influence. In 1970, following the collapse of the Namibian sardine, there was much starvation of chicks at all three Namibian colonies (Crawford et al., 1983a). Off Australia, when a mass mortality of sardine in 1998 led to a sudden reduction in the availability of prey, there was poorer breeding success by younger than by older Australasian gannets, unlike years when food was not limited (Bunce et al., 2005). In 2003/2004, Cape gannets at Namibian colonies had longer foraging trips of longer duration, spent less time at the colony, and left chicks unattended more frequently than those at colonies in the Western Cape (Lewis et al., 2006). Off South Africa, reproductive success of African penguins (Spheniscus demersus) is significantly related to food availability (Crawford et al., 2006a). Some growth of colonies of Cape gannets is thought too rapid to attribute to local breeding alone (Crawford et al., 1983b). Movements of young gannets between colonies off western southern Africa, including from Namibia to South Africa, have been noted (Crawford, 1999). Similarly, the rapid growth of at least some North Atlantic gannetries is attributed partially to the immigration of young breeders from elsewhere (Nelson, 2002). The formation of new colonies of Australasian and North Atlantic gannets indicates that there must be some emigration of birds (Norman, et al. 1998; Nelson, 2002). Emigration of young African penguins to colonies where feeding conditions are favourable at the time has been demonstrated and is thought an important mechanism whereby the species copes with long-term changes in the distribution of prey (Crawford, 1998). The altered distribution of prey available to Cape gannets is emphasized by the fact that, whereas many birds previously undertook extensive movements northwards along the West African coast after breeding (Broekhuysen et al., 1961; Crawford et al., 1983b), the proportion of birds making such movements is now greatly reduced (Oatley, 1988; Klages, 1994).

Influence of other factors Other factors have also influenced the numbers of Cape gannets breeding. Until the 1980s, guano was collected at each of the six extant gannet colonies, usually annually. Thereafter, as revised policies to conserve seabirds at islands were implemented, the collection of guano was discontinued (Best et al., 1997). The last collections were made at Mercury Island in 1984, Possession and Malgas Islands in 1985, Algoa Bay in 1988, Ichaboe Island in 1989, and Lambert’s Bay in 1991. Subsequently, there were no large increases in numbers of gannets nesting at the three Namibian colonies and at Lambert’s Bay, northernmost of the South African colonies. There also was no immediate increase in the colony at Algoa Bay. However, there was a substantial increase in numbers nesting at Malgas Island in 1986/1987, the second season after termination of guano scraping, and in several of the following seasons (Table 1). Cape gannets construct their nests almost entirely from guano (Hockey et al., 2005). Hence, the removal of guano leads to a loss of nesting material. In the 1989/1990 season, nests at Malgas Island, 4.5 y after the last guano scrape there, had a mean height of 106 mm, compared with 31 mm at Lambert’s Bay, where scraping was still being undertaken (Crawford and Cochrane, 1990).

In 2005/2006, Cape fur seals killed about 200 gannets in the colony at Lambert’s Bay and caused abandonment of breeding by the entire colony (Wolfaardt and Williams, 2006). Such disturbance may have caused partial abandonment in the previous season, when the colony decreased in size by about 50%. Seals also killed about 20 gannets in the colony at Malgas Island in 2005/ 2006 (L. Pichegru, pers. comm.). Prior to these observations, there had only been one record of a seal killing a Cape gannet ashore (Crawford and Cooper, 1996), although seals regularly kill fledglings around islands (David et al., 2003). Clearly, attacks by seals on birds ashore may have a major impact on colonies of Cape gannets, unless controlled. The recent shift of sardine away from western South Africa (Van der Lingen et al., 2005) may have influenced this new behaviour by Cape fur seals, by causing them to seek alternative sources of food.

Ecosystem considerations An altered abundance or distribution of prey may have a substantial influence on predators, particularly those such as Cape gannets that are constrained to be central-place foragers when rearing young and that generally show a strong fidelity to breeding localities (Hockey et al., 2005). In ecosystems supporting sardine and anchovy, regime changes in abundance have been reported (Schwartzlose et al., 1999). It is important to develop an understanding of ecosystem changes during periods of regime shifts so that appropriate management of resources can be applied at such times. For example, the decision to fish anchovy intensively in Namibia as sardine there was decreasing may have influenced the subsequent prolonged scarcity of epipelagic fish in the northern Benguela, which resulted in large decreases in numbers of African penguins (Crawford, 1998) and Cape gannets in Namibia. Fishing has the potential to shorten and decrease peaks in abundance of fish stocks and to deepen and lengthen troughs. Such changes may present a challenge even to long-lived predators, which are able to cope with natural interannual variability in resources (Crawford et al., 2006a). Environmental change, as well as fishing, may influence the availability of prey to central-place foragers (e.g. Frederiksen et al., 2004; Crawford et al., 2006b). In such instances, it may be appropriate to incorporate spatial considerations into the management of fish resources. For example, under the present scenario where most sardine in the Benguela system is located east of Cape Agulhas, the continued extraction of most of the allowable catch along the west coast, where most canning and reduction plants as well as most seabird breeding localities are located, will further exacerbate the shortage of food available to seabirds in the area. However, commercial fishing quotas are set frequently on the assumption that prey abundance is the only important factor for multispecies management (Camphuysen, 2005). It is often the production of seabirds that is first influenced by prey scarcity (Cairns, 1987). However, in averting decreases in populations, it is important also to minimize mortality (Crawford et al., 2006a). It may be necessary to cull animals, such as individual seals, that are inflicting excessive mortality on, or causing extensive disturbance to, threatened species (David et al., 2003). Again, an understanding of processes is likely to assist management interventions. For example, increased predation by predators on seabirds has been noted when alternative food resources for predators have diminished (Votier et al., 2004), and the individual seals damaging seabird colonies in southern Africa are almost all young males (David et al., 2003). Top Predators of the Benguela System

151

176


Acknowledgements We thank our research institutes and the National Research Foundation for supporting this research, as well as all who assisted with surveys of Cape gannets. CapeNature, Department of Environmental Affairs and Tourism (South Africa), Ministry of Fisheries and Marine Resources (Namibia), South African National Parks, and South African Navy provided logistical support for the surveys. We thank R. Cloete, J. C. Coetzee, and G. Dalmeida for providing updated information on the biomass of sardine and anchovy off Namibia and South Africa, and two anonymous reviewers for valuable comments. This paper is a contribution to the project LMR/EAF/03/02 of the Benguela Current Large Marine Ecosystem (BCLME) Programme.

References Batchelor, A. L. 1982. The diet of the Cape gannet Sula capensis breeding on Bird Island, Algoa Bay. MSc thesis, University Port Elizabeth. Best, P. B., Crawford, R. J. M., and Van der Elst, R. P. 1997. Top predators in southern Africa’s marine ecosystems. Transactions of the Royal Society of South Africa, 52: 177 –225. BirdLife International. 2004. Threatened Birds of the World 2004. CD Rom version. BirdLife International, Cambridge, UK. Broekhuysen, G. J., Liversidge, R., and Rand, R. W. 1961. The South African gannet Morus capensis. 1. Distribution and movements. Ostrich, 32: 1– 19. Bunce, A., Norman, F. I., Brothers, N., and Gales, R. 2002. Long-term trends in the Australasian gannet (Morus serrator) population in Australia: the effect of climate change and commercial fisheries. Marine Biology, 141: 263 – 269. Bunce, A., Ward, S. J., and Norman, F. I. 2005. Are age-related variations in breeding performance greatest when food availability is limited? Journal of Zoology, London, 266: 163– 169. Cairns, D. K., 1987. Seabirds as indicators of marine food supplies. Biological Oceanography, 5: 261 – 271. Camphuysen, C. J. (Ed.) 2005. Understanding marine foodweb processes: an ecosystem approach to sustainable sandeel fisheries in the North Sea. IMPRESS Final Report, Royal Netherlands Institute for Sea Research, Texel. 240 pp. Crawford, R. J. M. 1998. Responses of African penguins to regime changes of sardine and anchovy in the Benguela system. South African Journal of Marine Science, 19: 355 – 364. Crawford, R. J. M. 1999. Seabird responses to long-term changes of prey resources off southern Africa. In Proceedings of the 22nd International Ornithological Congress, Durban, pp. 688– 705. Ed. by N. J. Adams and R. H. Slotow. Birdlife South Africa, Johannesburg. Crawford, R. J. M. 2004. Accounting for food requirements of seabirds in fisheries management—the case of the South African purse-seine fishery. African Journal of Marine Science, 26: 197– 203. Crawford, R. J. M., Barham, P. J., Underhill, L. G., Shannon, L. J., Coetzee, J. C., Dyer, B. M., Leshoro, T. M., et al. 2006a. The influence of food availability on breeding success of African penguins Spheniscus demersus at Robben Island, South Africa. Biological Conservation, 132: 119– 125. Crawford, R. J. M., and Cochrane, K. L. 1990. Onset of breeding by Cape gannets Morus capensis influenced by availability of nesting material. Ostrich, 61: 147– 149. Crawford, R. J. M., and Cooper, J. 1996. Cape fur seal Arctocephalus pusillus catches Cape gannet Morus capensis ashore at Malgas Island. Marine Ornithology, 24: 53 – 54. Crawford, R. J. M., Dyer, B. M., Cooper, J., and Underhill, L. G. 2006b. Breeding numbers and success of Eudyptes penguins at 152


Marion Island, and the influence of mass and time of arrival of adults. CCAMLR Science, 13: 175– 190. Crawford, R. J. M., Shannon, L. V., and Pollock, D. E. 1987. The Benguela ecosystem. 4. The major fish and invertebrate resources. Oceanography and Marine Biology. An Annual Review, 25: 353– 505. Crawford, R. J. M., Shelton, P. A., and Berruti, A. 1983a. Cape cormorants as potential indicators of pelagic fish stocks off southern Africa. South African Journal of Science, 79: 466 – 468. Crawford, R. J. M., Shelton, P. A., Cooper, J., and Brooke, R. K. 1983b. Distribution, population size and conservation of the Cape gannet Morus capensis. South African Journal of Marine Science, 1: 153 – 174. David, J. H. M., Cury, P., Crawford, R. J. M., Randall, R. M., Underhill, L. G., and Meyer, M. A. 2003. Assessing conservation priorities in the Benguela ecosystem: analysing predation by seals on threatened seabirds. Biological Conservation, 114: 289– 292. Frederiksen, M., Wanless, S., Harris, M. P., Rothery, P., and Wilson, L. J. 2004. The role of industrial fisheries and oceanographic change in the decline of North Sea black-legged kittiwakes. Journal of Applied Ecology, 41: 1129– 1139. Hockey, P. A. R., Dean, W. R. J., and Ryan, P. G. (eds) 2005. Roberts Birds of Southern Africa, 7th edn. John Voelcker Bird Book Fund, Cape Town. Klages, N. T. W. 1994. Dispersal and site fidelity of Cape gannets Morus capensis. Ostrich, 65: 218– 224. Klages, N. T. W., Willis, A. B., and Ross, G. J. B. 1992. Variability in the diet of the Cape gannet at Bird Island, Algoa Bay, South Africa. South African Journal of Marine Science, 12: 761 – 771. Lewis, S., Gre´milet, D., Daunt, F., Ryan, P. G., Crawford, R. J. M., and Wanless, S. 2006. Using behavioural and state variables to identify proximate causes of population change in a seabird. Oecologia, 147: 606– 614. Montevecchi, W. A., and Myers, R. A. 1997. Centurial and decadal oceanographic influences on changes in northern gannet populations and diets in the north-west Atlantic: implications for climate change. ICES Journal of Marine Science, 54: 608– 614. Nelson, B. 2002. The Atlantic Gannet, 2nd edn. Fenix Books Ltd, Great Yarmouth. 396 pp. Norman, F. I., Minton, C. D. T., Bunce, A., and Govanstone, A. P. 1998. Recent changes in the status of Australasian gannets Morus serrator in Victoria. Emu, 98: 147 – 150. Oatley, T. B. 1988. Change in winter movements of Cape gannets. In Long-Term Data Series Relating to Southern Africa’s Renewable Natural Resources, pp. 40 – 42. Ed. by I. A. W. Macdonald and R. J. M. Crawford. South African National Scientific Programmes Report, 157. CSIR, Pretoria. Randall, R. M., and Ross, G. J. B. 1979. Increasing population of Cape gannets on Bird Island, Algoa Bay, and observations on breeding success. Ostrich, 50: 168– 175. Schwartzlose, R. H., Alheit, J., Bakun, A., Baumgartner, T. R., Cloete, R., Crawford, R. J. M., Fletcher, W. J., et al. 1999. Worldwide large-scale fluctuations of sardine and anchovy populations. South African Journal of Marine Science, 21: 349– 366. Shelton, P. A., Crawford, R. J. M., Kriel, F., and Cooper, J. 1982. Methods used to census three species of southern African seabirds, 1978– 1981. Fisheries Bulletin South Africa, 16: 115– 120. Van der Lingen, C. D., Coetzee, J. C., Demarcq, H., Drapeau, L., Fairweather, T. P., and Hutchings, L. 2005. An eastward shift in the distribution of southern Benguela sardine. Globec International Newsletter, 11: 17– 22. Votier, S. C., Furness, R. W., Bearhop, S., Crane, J. C., Caldow, R. W. G., Catry, P., Ensor, K., et al. 2004. Changes in fisheries discard rates and seabird communities. Nature, 427: 727 – 730.

177

50-year trends in Cape gannet numbers Wanless, S., Murray, S., and Harris, M. P. 2005. The status of northern gannet in Britain and Ireland in 2003/04. British Birds, 98: 280– 294. Wodzicki, K., Robertson, C. J. R., Thompson, H. R., and Alderton, C. J. T. 1984. The distribution and numbers of gannets in New Zealand. Notornis, 31: 232– 261.

Wolfaardt, A. C., and Williams, A. J. 2006. Sealed off – predation threatens seabirds and tourism. Africa—Birds and Birding, 11: 60 – 67.

doi:10.1093/icesjms/fsl011


153

154


Chapter 18 Breeding in a dynamic system: intra- and inter-seasonal variability in foraging behaviour and chick growth of Cape Gannets Ralf H.E. Mullers1, René A. Navarro2, Les G. Underhill2 and †G. Henk Visser1 1

University of Groningen, Department of Behavioural Biology, Kerklaan 30, 9751 NN Haren, the Netherlands [email protected] 2 Avian Demography Unit, Department of Statistical Sciences, University of Cape Town, Rondebosch 7701, South Africa Foraging in a highly dynamic system requires flexibility of breeding animals. Intra-seasonal fluctuations in food abundance and distribution during the chick rearing period will affect the foraging behaviour of parents and consequently breeding performance within the breeding season. Foraging behaviour and chick growth of Cape Gannets (Morus capensis) breeding at Malgas Island (South Africa) in the Benguela upwelling ecosystem was investigated during the breeding seasons of 2003/04 and 2004/05. Trip duration varied by as much as 6–8 hrs between the beginning and the end of the same season, without any apparent seasonal trend. In 2003/04 feeding trips were longer later in the season, whereas the opposite occurred in 2004/2005. The time spent flying during foraging trips differed significantly within the second season, decreasing by 37% from 10.7 hrs to 6.8 hrs be-

tween the first and second half of the season. Large intra-seasonal variation in the growth of the chicks was found in both seasons and growth was 28.6% faster in the first season compared to the second. The slower growth in the second season was associated with a decrease in the proportion of two prey species with high calorific contents, anchovies (Engraulis encrasicolus) and sardines (Sardinops sagax), in the diet (from 58.7% to 23.8% wet mass). The decreased growth rate of the chicks was also associated with a decrease in fledging success (36.6% compared to 54.5%). Breeding in a dynamic system like the Benguela upwelling ecosystem generates large variation in breeding performances, indicating the importance of studying intra-seasonal variation besides inter-seasonal comparisons.

Keywords: foraging effort, growth performance, diet, variability, Morus capensis, Benguela upwelling system

Introduction Reproductive parameters such as the timing of breeding and breeding success are strongly linked to food availability (Lack 1968, Martin 1995). In small passerine birds the timing of breeding is crucial because of the strong relationship between the peak in food supply and fledging success (e.g. Verboven et al. 2001, Tremblay et al. 2003). In seasonal environments, both food availability per se and the predictability of food abundance seem to be crucial for breeding (Siikamäki 1998). Seabirds however, frequently forage in a highly patchy and dynamic habitat characterised by fluctuations in prey distribution, availability and abundance throughout the breeding season (Crawford 1999, Suryan et al. 2002, Weeks et al. 2006). The effects of such variability in food supply on breeding success may be amplified by the prolonged chick rearing period which is typical of seabirds (Lack 1968), resulting in a long breeding season in a dynamic environment. Rather then expecting a particular seasonal direction in breeding performance, the fluctuating food resources of seabirds will affect foraging behaviour and breeding parameters on a more short-term scale (Shea & Ricklefs 1985; Le Corre et al. 2003). Most seabird species are central place foragers, commuting considerable distances from their breeding colony to feeding areas and returning with food for their offspring (Orians & Pearson 1979). Adults have to support the increasing energy demand of their chicks, and at the same time

cope with any sudden changes in the distribution, abundance or composition of prey. To negotiate such changes successfully, they need to alter their foraging behaviour in space and time. Seabirds are able to adjust their time spent away from the colony (Charrassin et al. 1999; Dall’Antonia et al. 2001), make different allocation decisions (Takahashi et al. 2003) or catch different prey species (Berrow & Croxall 2001). Flexibility in foraging behaviour is an essential adaptation for when feeding conditions change within the breeding season (Granadeiro et al. 1998). Additionally parents have to meet the increasing chick’s energy requirements, which put large energetic demands on the parents (e.g. Birt-Friesen et al. 1989; Adams et al. 1991), especially when food is scarce. Chick rearing parents are restricted by their maximum working capacity (Drent & Daan 1980) above which they do not increase their foraging effort, therefore boundaries are set to energetic flexibility. Because periods with low food abundance will negatively effect chick growth, this parameter can be a useful indicator of local food availability (Ricklefs et al. 1984, Adams et al. 1992). Seabird breeding numbers are closely associated with prey availability (Monaghan et al. 1989) and therefore islands in the vicinity of upwelling systems are characterised by large numbers of breeding seabirds. Prey species are particularly abundant in these systems (Brown & Gaskin 1986; Wolanski & Hamner 1988) on account of the high rates of primary production sustained by the upwelling of nutrient-rich waters. However, fluctuations in prey abundance associated with Top Predators of the Benguela System

155

variation in the location and intensity of the upwelling cells within and between years (Crawford 1999; Demarcq et al. 2003) can have considerable effects on breeding conditions for seabirds (Montevecchi & Myers 1995; Suryan et al. 2002). For example, the Benguela Current upwelling system on the west coast of southern Africa is wind driven and therefore upwelling can stop when the wind drops. Furthermore, in the southern Benguela, there have been large-scale distributional shifts in the populations of sardines (Sardinops sagax) and anchovies (Engraulis encrasicolus) (van der Lingen et al. 2005). These species are high in calorific content (Batchelor & Ross 1984) and are favoured prey of several seabird species, including the Cape Gannet (Morus capensis). The Cape Gannet is one of the most abundant locally breeding seabirds in the Benguela ecosystem. At Malgas Island (Saldanha Bay, South Africa) the breeding season of the gannet colony lasts eight to nine months (pers. obs.); although individual breeding attempts are 4–5 months long if successful (c. 42 days of incubation and c. 110 days to raise a chick), there is inter-individual variation of up to several months in the onset of breeding. Together, the long breeding season of the Cape Gannet and the dynamic nature of the Benguela system (Weeks et al. 2006) are ideal circumstances to study the responses by a breeding seabird population to temporal changes in the environment. Therefore, using measures of chick growth and adult foraging behaviour, the aim of this study was to investigate how Cape Gannets cope with the challenges of breeding in a fluctuating environment. Material and Methods During the 2003/2004 and 2004/2005 breeding seasons, the gannet colony at Malgas Island (33°02'S, 17°55'E) was visited every alternate week between early October and late February, for a week at a time. The same study protocol was followed during each week of fieldwork. Foraging behaviour of adults The foraging behaviour of the gannets was studied using GPS-dataloggers (Newbehavior; Zürich, Switzerland), which recorded the speed and geographic position of a deployed bird at 10 sec intervals (c. 10 m resolution). The loggers were sealed in two waterproof bags; total weight was about 50 g, less than 2% of the adult body mass (2605 g in this study). Birds with different feeding objectives (i.e. chicks of different age and size) were selected as study animals. The devices were deployed when the birds left the nest for a foraging trip, which always occurred once the chick was attended by the other parent to correct for different trip durations when chicks are left unattended (Lewis et al. 2004). The departing bird was caught with a hooked pole, measured and weighed. A logger was attached to its tail feathers with waterproof Tesa®-tape (Beiersdorf AG, Hamburg), which does not damage the tail feathers, and the bird was released near its nest. The procedure took c. 5 min and most birds left the colony within minutes after deployment. The same devices were deployed on Cape Gannets during a previous research project, with no obvious adverse effects on the behaviour of the birds (Gremillet et al. 2004). A couple of hours after deployment (to allow for digestion of the food obtained from the returned parent), the chick was measured for bill length (to 0.1 mm), wing length (to 1 mm) and body mass (to 5 g below 1 kg; to 25 g above 1 kg). The nest was then monitored hourly until the parent with the logger returned, whereupon the logger was retrieved by re-capturing the bird and taking off the tape and logger. 156


The following behaviour types were recognised from the GPS data recovered from the loggers: out-flight, searchflight, time spent drifting, fishing events, return-flight, total distance covered and duration of the foraging trip. From these data we calculated flying time (total duration minus time spent drifting) and foraging time (search flight plus fishing events). Flying time was taken as an indicator for the foraging effort, affecting parent’s allocation decisions, while the foraging trip duration determined the feeding frequency of chicks, affecting chick growth. Chick growth Access to the centre of the colony involved unacceptable levels of disturbance; therefore only chicks near the periphery of the colony, at four different sites, were measured for growth. Although predation pressure from kelp gulls (Larus dominicanus) increases towards the periphery of the colony (pers. obs.), this would not affect growth rates. The chicks were taken from the nest, measured and returned to the nest within three minutes. Bill length, wing length and body mass were measured (as above for the chicks of instrumented birds) and the number of ticks (Ornithodorus capensis) on each chick was counted. In order to standardize the measuring protocol, chicks were measured in the same sequence and at the same time of the day by the same person (RAN). Newly hatched chicks at the four sites were included in the measuring protocol to have data on the growth of young chicks throughout the season. The measurements continued until either the chick died or was completely feathered and ready to fledge. The same four sites were used in 2003/2004 and 2004/2005. Growth index Because none of the standard parametric growth models fitted the data adequately, the growth rates were analysed using a non-parametric approach developed by le Roux & Underhill (in prep), (see appendix for full description of method). This growth index represents a common currency to measure departures from “mean” growth, and is independent of whether growth measures are taken shortly after hatching, when the absolute growth rates (gd–1) tend to be small, at the maximum growth spurt, when growth rates tend to be large, or late in growth, when growth rates tend to decrease. The scores are assumed to be normally distributed (which to a first approximation is probably reasonable), so that the magnitudes of z-values can be expected to be within the standard normal distribution. Diet sampling The diet of the gannets was sampled during 1–3 consecutive days each month by Marine and Coastal Management. Birds were captured with a hooked pole upon arrival from a foraging trip and inverted over a bucket, into which they regurgitated. Fifty samples were collected per month, each of which was analysed independently. The analyses involved identifying each species represented in the sample and then, for each prey species separately, determining its weight in the sample, the number of individuals (by counting the numbers of head and tails) and size of whole fish. From these data the relative contribution to the total diet in percentage wet mass was calculated for each species. Data analysis The foraging behaviour of the adults was analysed using multiple regression in which the potential explanatory effect

Table 1: Morus capensis. Summary statistics for foraging data obtained from GPS data-loggers for the 2003/2004 and 2004/2005 breeding seasons. Chick age is the mean age of the chicks of all deployed birds and flying time is obtained by subtracting time spent drifting on the sea-surface from total trip duration Chick age (days) 2003–2004

2004–2005

mean N stdev min–max mean N stdev min–max

33.0 74 21.7 0–100 32.7 82 21.3 0–93

Duration (hrs) 22.5 78 11.8 3.1–50.9 23.8 85 13.0 3.1–54.0

of variables was tested using a backwards deletion method. The residuals of significant models were tested for normal distribution. Growth indices were calculated using GenStat 8 and statistical analyses were done with SPSS 13.0 statistical package. Results Characteristics of feeding trips GPS loggers were deployed on 224 birds during this study (99 in 2003/04 and 125 in 2004/05). One hundred and sixty three complete tracks were obtained (72.8%); the main reasons for incomplete tracks were insufficient battery power during longer foraging trips and gaps in the tracks due to communication problems between the logger and the GPS satellites. Three GPS loggers were lost, either because the bird returned without the logger, or did not return to the nest at all. Only complete tracks were used for analyses. The summary statistics of the foraging parameters are given in Table 1. All the foraging parameters were inter-related and correlations between them were positive (Table 2). Variation in foraging behaviour To investigate intra-annual variation in foraging behaviour, weekly means for the foraging parameters were calculated and compared. Deployment occurred during 3–4 days each week the colony was visited and the data retrieved represented foraging behaviour over 5–7 days. For each of these periods, data for a minimum of 7–19 complete foraging trips were obtained. Trip duration had a tri-modal distribution due to the effect of birds spending either zero, one or two nights at sea. Total time spent flying during the trip was normally distributed (Kolmogorov-Smirnov Z = 1.1, P = 0.14). There was no difference in flying time between seasons (One-way ANOVA: F1,161 = 0.2, P = 0.634), but there was a difference between periods within 2004/2005 (One-way ANOVA: 2003/ 04 F6,71 = 1.1, P = 0.358; 2004/05 F5,79 = 2.4, P = 0.029, Fig. 1). Trip duration did not differ between seasons (KruskalWallis Test: χ 2 = 10.2, df = 12, P = 0.596, Fig. 1) or between periods within seasons (Kruskal-Wallis Test: 2003/04 χ 2 = 6.7, df = 6, P = 0.347; 2004/05 χ 2 = 3.2, df = 5, P = 0.671). In 2003/2004, the mean trip duration decreased gradually between September and mid December, from 23.3 hrs to 18.9 hrs (overall mean = 20.5 hrs). This increased to 28.7 hrs at the end of December and the beginning of January. The trend was the opposite in 2004/2005; long foraging trips were made between the beginning of the season and the end of November (mean duration = 26.3 hrs), after which the duration decreased (mean duration = 20.0 hrs). From these results it was possible to broadly divide each season into two periods, each with different characteristics for the foraging

Distance (km) 428.4 78 198.8 77.8–933.1 455.8 85 253.1 101.3–1220.7

Flying time (hrs)

Drifting time (hrs)

8.8 78 4.2 1.8–18.9 9.1 85 5.2 2.2–26.8

13.7 78 8.3 0.5–35.5 14.7 85 8.8 0.9–36.7

parameters. The division in 2003/2004 was between the first five and the two last periods (Fig. 1), whereas in 2003/2004, the six periods (one less than 2003/2004) were evenly split in the middle (Fig. 1). Parameters were then compared between the resulting four periods. Trip duration was significantly higher at the end of 2003/2004 than at the beginning, and varied significantly between 2003/2004 and 2004/2005 (Mann-Whitney: 2003/04 Z = –2.6, P = 0.011; 2004/05 Z = –1.7, P = 0.096; Kruskal-Wallis: both χ 2 = 9.6, df = 3, P = 0.023). Flying time in 2004/2005 was significantly longer during the first period compared to the second period, and also differed significantly from flying time in 2003/2004 (Oneway ANOVA: 2003/04 F1,76 = 3.8, P = 0.054; 2004/05 F1,83 = 13.8, P < 0.0001; both F3,159 = 1.4, P < 0.0001). Growth of chicks Growth measurements were available for 279 chicks (152 in 2003/04, 127 in 2004/05). The mean number of measurements per chick was 5.2 and the mean interval between measurements was 7.6 days (range; 5 – 42 in 2003/04, 4 – 32 in 2004/05). From these measurements 808 growth index scores were derived for 2003/2004 and 470 for 2004/2005. The mean growth (gd–1) differed between the two breeding seasons (One-way ANOVA: F1,1276 = 10.4, P = 0.001); being 28.6% higher in 2003/2004 (mean = 32.32 gd–1 ±32.50) than in 2004/2005 (mean = 25.14 gd–1 ±46.70). After correcting for chick age, growth was still significantly different between the seasons (One way ANOVA: F1, 1276 = 5.0, P = 0.027). Similarly, the mean growth index was significantly different between the two seasons (One-way ANOVA: F1,1276 = 5.2, P = 0.023), and between periods within each season (One-way ANOVA: 2003/04, F17,790 = 3.7, P < 0.0001; 2004/ 05, F13,456 = 7.9, P < 0.0001). The growth indices were modelled using a multiple regression with year, age, number of ticks and the hatching date as explanatory variables (Table 3). The model explained 3% of the variation found. The residuals still varied between periods in both seasons (Oneway ANOVA: F31,1246 = 5.8, P < 0.001, Fig. 2), suggesting

Table 2: Morus capensis. Cross-correlations between foraging parameters of adult Cape Gannets for the 2003/04 and 2004/05 breeding seasons. For all correlations the sample size is 162 and the P-value < 0.0001

Duration (hrs) Flying time (hrs) Foraging (hrs) Drifting (hrs)

Distance (km)

Duration (hrs)

Flying (hrs)

Foraging (hrs)

0.883 0.937 0.901 0.768

0.884 0.883 0.966

0.924 0.734

0.774


157

Figure 1: Morus capensis. (a) Box plots for trip duration (in hrs) per period for two breeding seasons (2003/04 and 2004/05). The black line indicates the median, the box the upper and lower quartiles and the error bars the extreme values. Outliers are shown by dots above the error bars. The means per period are shown by the black dots in the boxes and sample sizes are stated in top of Fig. 1(b) Error graphs for flying time (in hrs) per period for two breeding seasons. The dots indicate the mean and the error bars the standard error

Table 3: Morus capensis. Results of the multiple regression with the growth index of chicks as dependent variable and year (breeding season), age (age of the chicks in days), ticks (number of ticks counted on chick) and hatching date of chick as explanatory variables df Corrected model 4 Intercept 1 Age 1 Ticks 1 Hatching date 1 Year 1 Error 1273

158

Mean square

F

P

7.959 6.259 22.258 8.384 6.219 4.580 0.933

8.534 6.711 23.865 8.990 6.669 4.911

100 pairs in some seasons) subsequently bred at Schaapen and Meeuw islands, Top Predators of the Benguela System

175

Figure 4: Trends in the counts of White-breasted Cormorants breeding at eight islands in the Western Cape, and in the overall number estimated to breed at the ten islands that were monitored, 1978–2006. The overall number may have been underestimated in the earlier part of this period, when fewer counts were undertaken (see text). At individual localities, gaps for any given season indicate that no counts were undertaken

where the counts had large fluctuations that were often out of phase with each other. The maximum number recorded at Vondeling Island was nine pairs. Counts at Dassen Island had long-term stability with large fluctuations. There was a large increase in the number breeding at Dyer Island between 1994 and 1998, after which the number fluctuated above 70 pairs. Discussion Counts of the three species at all the islands were less regularly conducted during the period 1978–1989 than subsequently. This may mean that numbers breeding at some islands were underestimated in the earlier period, for example Bank Cormorants at Lambert’s Bay in 1988 and at Dassen Island in 1979, 1980, 1985 and 1988 (Figure 2). Hence the overall populations at the ten islands also may have been underestimated. At Lambert’s Bay from 1997– 2002, counts of pairs of White-breasted Cormorants were made at approximately weekly intervals by Ward and Williams (in press). Their highest single count was of 29 pairs in 2000. Over the same period, the highest count obtained during monthly surveys for this study was similar: 28 pairs in 2000 (Fig. 4). This suggests that monthly counts are adequate to gauge population trends. However, Ward and Williams (in press) believed that because seven pairs had completed breeding earlier in 2000, the actual population at Lambert’s Bay in 2000 was 37 pairs. Therefore, even in the period of more frequent counting the actual population may be underestimated. For birds that move between breeding sites, as all three species may (Crawford et al. 1994, 1999a), there is the possibility that the same birds may be counted at two localities in any given season, which would overestimate abundance. For Bank Cormorants, the number breeding at the ten islands was probably stable at 550–650 pairs from 1978– 176


1991 (Crawford et al. 1999a), thereafter halving to about 300 pairs. This was as a result of decreases at Lambert’s Bay, Malgas and Dassen islands. Numbers were little changed from 1978–2006 at the other localities. The extinction of the colony at Lambert’s Bay was partly attributable to the displacement of birds from breeding sites by Cape Fur Seals Arctocephalus pusillus pusillus (Crawford et al. 1999a), which earlier caused a decrease in the colony at Mercury Island (Crawford et al. 1989). The reasons for the declines at Malgas and Dassen islands are not well understood, especially since the numbers breeding at three islands (Marcus, Jutten and Vondeling) between these localities remained stable over the period investigated. Some birds from Malgas Island may have moved to nearby Jutten Island, where numbers increased as those at Malgas Island began to decrease. Prior to their demises, the colonies at Malgas and Dassen islands were the largest at the ten islands that were investigated. It is possible that food resources became limiting in their vicinity, a scenario that is supported by a halving in the catch of rock lobsters Jasus lalandii off South Africa between the 1980s and the 1990s (Van der Lingen et al. 2006). There was a large decrease in the growth rate of rock lobsters off South Africa between 1985 and 1991, after which growth rate remained low (Cruywagen 1997) suggesting reduced productivity of rock lobsters. At Dassen Island, the annual growth increment of rock lobsters of size 70–80 mm averaged 5.1 mm from 1969–1971 and 3 mm from 1992–1995, whereas for lobsters of size 80–90 mm it was 4.4 mm from 1969–1988 and 2.1 mm from 1989–1995 (Pollock et al. 1997). Rock lobsters are a major component of the diet of Bank Cormorants in South Africa (Rand 1960, Avery 1983) and in the 1990s were their most important prey item between Saldanha and Table Bay (Hockey et al. 2005), where both Malgas and Dassen islands are located. The decreased commercial harvest of rock lobsters was not alleviated by a

reduced minimum size limit for catches of lobsters, from 89 to 75 mm carapace length in 1993 (Pollock et al. 1997), but the exploitation of smaller lobsters probably further reduced their availability to Bank Cormorants. It is thought that in the late 1970s and early 1980s, about 1 000 pairs of Crowned Cormorants bred at the 10 islands investigated (Crawford et al. 1982), and this remained the case until 2002/03. A subsequent increase was driven by a large increase at Dassen Island. Opposite trends at Dassen and Robben islands after 1996/97 (Fig. 3) indicate there may have been movement of birds between these localities. Crowned Cormorants are known to shift their nest sites around islands (Crawford et al. 1994, Crawford and Dyer 1996, Underhill et al. in press), and may do so between adjacent breeding localities. Crowned Cormorants have moved distances up to 560 km from the site at which they were banded (Underhill et al. 1999). There were substantial numbers of feral cats Felis catus at Robben Island in the late 1990s (Crawford and Dyer 2000), when the number of Crowned Cormorants there decreased. There was no discernible long-term trend in the population of White-breasted Cormorants breeding at the islands, but large fluctuations were apparent (Fig. 4). White-breasted Cormorants stopped breeding at Jutten and Marcus islands in the 1980s, when they colonised Schaapen, Meeuw and Vondeling islands (Crawford et al. 1994). The reason for this shift in their location is unknown, although it is believed that disturbance at breeding colonies may cause birds to move to other localities (Ward and Williams in press). Two pairs of White-breasted Cormorants commenced breeding at Robben Island, when a platform was provided offshore, at which they could not be disturbed. The trough in the overall number of pairs breeding in the mid 1980s suggests that some birds may skip breeding seasons when looking for a new locality at which to breed. In the 1990s and 2000s, it appears that some birds alternated their breeding sites between Schaapen and Meeuw islands (Fig. 4), which are in close proximity to each other. Again, troughs in the overall counts (1994, 1999, 2003) precede years at which the alternation took place (Fig. 4). The increase in the number breeding at Dyer Island followed a decrease in the number breeding at adjacent Geyser Rock, which supported some 50–60 pairs from 1979 until 1993 (Brooke et al. 1982, unpublished information). The birds at Geyser Rock were displaced from their breeding sites by Cape Fur Seals (pers. obs.). In summary, the populations of Crowned and Whitebreasted Cormorants have remained more or less stable, or even increased, over the 29 years that they have been monitored at the ten islands in the Western Cape, although there have been marked shifts in numbers breeding at different localities. By contrast, the numbers of Bank Cormorants breeding at these islands have halved. It seems probable that food has not been limiting for Crowned and Whitebreasted Cormorants, but it may have been for Bank Cormorants. Factors such as disturbance by humans, displacement from breeding sites by seals and introduced predators have influenced trends at particular localities (e.g. Crawford et al. 1999a). Acknowledgements – I am grateful to South Africa’s Department of Environmental Affairs and Tourism for supporting this work. Financial support was provided by Earthwatch Institute and the Marine Living Resources Fund. CapeNature, Robben Island Museum, South African National Parks and South African Navy provided logistical support. I thank all who assisted with counts of the three seabirds that are considered in this paper, especially D.A.E. Crawford, P.B. Crawford, P.J.M. Crawford, B.M. Dyer, T.M. Leshoro and L. Upfold. I am grateful to C. Boucher for preparing the artwork. The paper is a contribution to the project LMR/EAF/03/02 of the Benguela Current Large Marine Ecosystem (BCLME) Programme.

References Avery G 1983. Bank Cormorants Phalacrocorax neglectus taking Cape rock lobster Jasus lalandii. Cormorant 11(1/2): 45–48. Brooke RK, Cooper J, Shelton PA and Crawford RJM 1982. Taxonomy, distribution, population size, breeding and conservation of the Whitebreasted Cormorant, Phalacrocorax carbo, on the southern African coast. Gerfaut 72:188–220. Cooper J 1981. Biology of the Bank Cormorant, Part 1: Distribution, population size, movements and conservation. Ostrich 52(4): 208–215. Cooper J 1987. Biology of the Bank Cormorant, Part 5: Clutch size, eggs and incubation. Ostrich 58(1): 1–8. Crawford RJM, Cruickshank RA, Shelton PA and Kruger I 1985. Partitioning of a goby resource amongst four avian predators and evidence for altered trophic flow in the pelagic community of an intense, perennial upwelling system. South African Journal of Marine Science 3: 215–228. Crawford RJM, David JHM, Williams AJ and Dyer BM 1989. Competition for space: recolonising seals displace endangered, endemic seabirds off Namibia. Biological Conservation 48(1): 59– 72. Crawford RJM and Dyer BM 1996. Age at breeding of Crowned Cormorants Phalacrocorax coronatus. South African Journal of Marine Science 17: 315–318. Crawford RJM and Dyer BM 2000. Wildlife of Robben Island. Avian Demography Unit, Cape Town: 28 pp. Crawford RJM, Dyer BM and Brooke RK 1994. Breeding nomadism in southern African seabirds – constraints, causes and conservation. Ostrich 65(2): 231–246. Crawford RJM, Dyer BM, Cordes I and Williams AJ 1999a. Seasonal pattern of breeding, population trend and conservation status of bank cormorants Phalacrocorax neglectus off south western Africa. Biological Conservation 87(1): 49–58. Crawford RJM, Dyer BM and Upfold L 1999b. Seasonal pattern of breeding by Cape and Crowned Cormorants off western South Africa. Ostrich 70(3/4): 193–195. Crawford RJM, Shelton PA, Brooke RK and Cooper J 1982. Taxonomy, distribution, population size and conservation of the Crowned Cormorant, Phalacrocorax coronatus. Gerfaut 72: 3–30. Cruywagen GC 1997. The use of generalized linear modelling to determine inter-annual and inter-area variation of growth rates: the Cape rock lobster as example. Fisheries Research 29: 119– 131. du Toit M, Boere GC, Cooper J, de Villiers MS, Kemper J, Lenten B, Simmons RE, Underhill LG and Whittington PA (eds). 2003. Conservation Assessment and Management Plan for southern African coastal seabirds. Avian Demography Unit, Cape Town and Conservation Breeding Specialist Group, Apple Valley. Hockey PAR, Dean WRJ and Ryan PG (eds) 2005. Roberts’ Birds of Southern Africa. 7th edn. John Voelcker Bird Book Fund, Cape Town. Olver MD and Kuyper MA 1978. Breeding biology of the Whitebreasted Cormorant in Natal. Ostrich 49(1): 25–30. Pollock DE, Cockroft AC and Goosen PC 1997. A note on reduced rock lobster growth rates and related environmental anomalies in the southern Benguela, 1988–1995. South African Journal of Marine Science 18: 287–293. Rand RW 1960. The biology of guano-producing seabirds. 3. The distribution, abundance and feeding habits of the cormorants Phalacrocoracidae off the south-western coast of the Cape Province. Investigational Report Sea Fisheries Research Institute South Africa 42: 1–32. Randall RM, Tregoning C, Randall BM and Martin AP 2002. Adaptability of Great Cormorants Phalacrocorax carbo in a coastal environment demonstrated by their exploitation of introduced prey species and use of artificial breeding sites. South African Journal of Marine Science 24: 317–321. Underhill LG, Crawford RJM, Harebottle DM and Tjørve KMC in press. The development of the heronry on Robben Island, Western Cape, South Africa, 1980–2005. Ostrich. Underhill LG, Tree AJ, Oschadleus HD and Parker V 1999. Review of Ring Recoveries of Waterbirds in Southern Africa. Avian Demography Unit, University of Cape Town. Van der Lingen CD, Shannon LJ, Cury P, Kreiner A, Moloney Top Predators of the Benguela System

177

CL, Roux J-P. and Vaz-Velho F 2006. Resource and ecosystem variability, including regime shifts, in the Benguela Current System. In: Shannon V, Hempel G, Malanotte-Rizzoli P, Moloney C and Woods J (eds) Benguela: Predicting a Large Marine Ecosystem. Elsevier, Amsterdam: 147–184. Ward, VL and Williams AJ in press. Seasonal pattern of breeding, fledging success and an irruption of White-breasted Cormorants

178


at Penguin Island, South Africa. Ostrich. Whitfield AK 1986. Predation by Whitebreasted Cormorants on fishes in a southern Cape estuarine system. Ostrich 57(4): 248– 249. Williams AJ and Cooper J 1983. The Crowned Cormorant: breeding biology, diet and offspring-reduction strategy. Ostrich 54(4): 213–219.

Kelp Gull

Chapter 21 The influence of culling, predation and food on Kelp Gulls Larus dominicanus off western South Africa Robert JM Crawford1,2, Les G Underhill2, Res Altwegg2, Bruce M Dyer1 and Leshia Upfold1 1Department

of Environmental Affairs and Tourism, Marine and Coastal Management, Private Bag X2, Rogge Bay 8012, South Africa 2Avian Demography Unit, Department of Statistical Sciences, University of Cape Town, Rondebosch 7701, South Africa The number of Kelp Gulls Larus dominicanus breeding at 11 islands in South Africa’s Western Cape Province increased from 6 500 pairs in 1978 to more than 16 000 pairs in 1999 and then decreased to 13 000 pairs in 2005. The increase came after removal of controls on gulls and was associated with supplementary food provided at rubbish tips and fish factories. The decrease resulted from predation of up to 85% of gull chicks hatched at some colonies by Great White Pelicans Pelecanus onocrotalus. Pelicans also were held at a low level in the early part of the 20th century by control measures and were later provided with additional food, in the form of agricultural offal and fish in impoundments. Disturbance

by humans influenced trends at some smaller gull colonies and it is likely there was inter-colony transfer of first breeders. Decreases at the northern gull colonies and increases in numbers to the south and east correspond to a similarly altered distribution of several species of seabird and fish off South Africa, thought attributable to environmental change. At Dassen Island, the density of gull nests remained constant as the colony doubled, but decreased by 50% as the colony decreased, suggesting that Kelp Gulls spaced out their nests more when subjected to predation. At this locality the clutch size increased following initiation of predation.

Keywords: culling, environmental change, food, Great White Pelican, human disturbance, Kelp Gull, Larus dominicanus, nest density, Pelecanus onocrotalus, predation

INTRODUCTION The abundance of a species is regulated by recruitment into and mortality from its breeding population. Hence its numbers may be controlled by modifying either recruitment or mortality. In the management of gull populations, both methods have been applied (e.g. Coulson et al. 1983, Wanless et al. 1996, Bosch et al. 2000). Natural predators, by affecting recruitment or mortality, can similarly influence prey populations. Humans can directly, or indirectly, e.g. through the removal or provision of food (e.g. Pons & Migot 1995, Tasker et al. 2000), alter the sizes of predator populations and thereby also influence their prey resources. For example, decreasing the amount of fish discarded may cause opportunistic predators, such as Great Skuas Stercorarius skua, to increase their predation on other birds, such as Black-legged Kittiwakes Rissa tridactyla, leading to decreases in populations of the latter (Votier et al. 2004). When several factors influence a population, their outcome is not always easy to predict. In this paper we examine the impact of culling, predation by Great White Pelicans Pelecanus onocrotalus, food and disturbance on the numbers of Kelp Gulls Larus dominicanus breeding at 11 islands in the Benguela upwelling system off South Africa’s Western Cape Province. Kelp Gulls Larus dominicanus have a wide distribution in the southern hemisphere (Croxall 1984, Croxall 1991, Higgins & Davies 1996, Hockey et al. 2005). The race L. d. vetula is endemic to southern Africa (Brooke & Cooper 1979), except for a few birds that bred north of the equator

in Senegal in 1998 (Keijl et al. 2001). This race is sometimes regarded as a separate species, the Cape Gull or Khoisan Gull (Williams 2004). In 1976–1981, the 11 islands studied supported 58% of the overall population of L. d. vetula, 73% the South African population and 85% of numbers in the Western Cape (Crawford et al. 1982). METHODS Numbers of nests of Kelp Gulls were counted at 11 islands off South Africa’s Western Cape Province: Lambert’s Bay, Malgas, Marcus, Jutten, Meeuw, Schaapen, Vondeling, Caspian, Dassen, Robben and Dyer islands. The six localities situated between Lambert’s Bay and Dassen Island are in close proximity to each other (Fig. 1). Counts were made at all localities in 1978, 1985 or 1986, and annually from 1992– 2005, except for Caspian Island from 1997–1999, Vondeling Island in 1992, 1995 and 1999 and Dyer Island in 1992. In some other years, counts were conducted at some of the localities. Breeding was first recorded at Caspian Island in 1993. This locality was not surveyed in earlier years but it is unlikely that substantial numbers bred without being noticed. Breeding commenced at Robben Island in 2000 (Calf et al. 2003). Of the 271 possible counts, 97 were not made (Table 1). These missing values were filled in by linear interpolation between the actual counts made at each island. Counts were undertaken during the main breeding season of Kelp Gulls in the Western Cape, usually during late October or early November when most clutches have been initiated and birds at most nests are incubating eggs Top Predators of the Benguela System

181

For the colony at Dassen Island, and for all other colonies combined (the most numerous of which are grouped in close proximity to each other), the breeding success required to achieve the observed growths of the populations in the absence of immigration was estimated as follows. The number of adult birds dying in year k, Dk, was calculated as 2Bk(1–Sa), where Sa is the annual survival rate of adults. This assumes all adults were breeding in year k. The number of adults recruiting to the breeding population in year k, Rk, was taken to be 2(Bk–Bk–1)+Dk–1 for 2(Bk–Bk–1)+Dk–1 > 0; 0 for 2(Bk–Bk–1)+Dk–1 ≤ 0. The number of chicks fledged per pair, Fk, was computed as (Rk+4(Si)–1(Sa)–3)/Bk, Figure 1: South Africa’s Western Cape Province showing the 11 localities where Kelp Gull colonies were monitored, 1978–2005

(Crawford et al. 1982). Counts were made from vantage points using binoculars after gulls had settled at nests, or by walking tightly spaced grids and marking nests. Methods used in counting are described more fully in Crawford et al. (1982). At each locality, the counts and interpolated counts were smoothed using an algorithm described by Underhill et al. (2006). Let x1, x2, …, x28 be the counts of breeding pairs at a single locality over the 28 year study period. For year k, the weight attached to the count in year j was wj = exp(–((k–j)/σ)2) where s, the smoothing parameter, was taken to be 2. The smoothed estimate of the number of breeding pairs was Bk = Σwjxj /Σwj, where the summations are over all the years j. The choice of σ = 2 ensures that the weights attached to years decrease rapidly on either side of the target year, which has weight 1. The years on each side of the target year both have weight 0.78, the years two years from the target year have weights 0.37, and at three years distant, the weights are 0.11. At four years distant, the weights are a negligible 0.018, and become vanishingly small thereafter. The weighted estimate is therefore based almost entirely on the data from the target year and three years on either side of it, with the third years having only 11% of the weight of the target year. For 15 of the 19 years in which Kelp Gulls were counted at Dassen Island, the areas occupied by breeding gulls were mapped as described by Crawford et al. (1994). The extents of the areas used for breeding were measured on these maps using an Ibas interactive image-analysis system. The maps, based on an aerial photograph, were scaled from ground measurements of straight edges of walls and buildings at the island. For each year for which maps were available, except 1991 when the count was made early in October before all breeders had settled (Crawford et al. 1994), the mean density of nests was estimated from the count and the area of occupation. At Schaapen Island, observations were made on the extents of the distributions of nests in all years when nests were counted. 182


where Si is the survival rate of immature birds from fledging until they are one year old, and Sa is the annual survival rate of birds older than one year. The possible excess/deficit in home-grown recruitment at Dassen Island was investigated, by holding Fk constant at 1.63 and 2.0 chicks per pair until 1991. After this date, Kelp Gull chicks were preyed on by pelicans (Crawford et al. 1997). The number of birds hatched at Dassen Island and aged 4 years, D4k, was taken to be Bk–4Fk(Si)(Sa)3. The excess/deficit in the number of recruits fledged at Dassen Island was computed as D4k–Rk. Similarly, the excess/deficit in the number of recruits fledged at other colonies was estimated when Fk was held at 1.63 and 2.0 chicks per pair over the entire period. The age at first breeding of Kelp Gulls in the Western Cape is generally four years, which is also when most birds acquire adult plumage (Crawford et al. 2000). Breeding success was capped at 1.63 chicks per pair based on a mean clutch of 2.2 eggs and a chick survival of 0.74 (Altwegg et al. in press), or at 2.0 chicks per pair based on the observation that 15 pairs fledged 29 chicks (1.93 per pair) at Robben Island in 2001 (Calf et al. 2003). The mean brood size at nests in the Western Cape during 1976–1981 was 1.7 chicks (Crawford et al. 1982). Also for the Western Cape, the reported range in mean brood size at colonies was 1.33–2.0, which has a median of 1.67 (Burger & Gochfield 1981). In accord with mortality between brooding and fledging, 1.63 is smaller than these values. Annual survival of Kelp Gulls in southern Africa was 0.44 (95% confidence interval 0.35– 0.54) for birds in their first year and, from two independent data sets, 0.84 (95% confidence intervals 0.77–0.89 and 0.78–0.89) for birds older than one year (Altwegg et al. in press). Between 1990 and 2000, in most seasons the number of eggs in nests was recorded for a sample of nests at the four larger colonies (Crawford et al. 1982). We used a generalised linear mixed model with a log link function and Poisson errors to estimate the spatial and temporal variance in clutch size. The random effects were colony, year, and their interaction, and the intercept was treated as a fixed effect. We considered adding day in the season as a covariate to account for the possibility that surveys during the beginning of the season were affected by incomplete clutches, but there was no detectable trend in clutch size over the season (P > 0.5), and we therefore did not include this effect.

RESULTS The total number of pairs of Kelp Gulls counted at the 11 islands in the Western Cape was about 6 500 pairs in 1987, 8 000 pairs in 1985/86, 18 000 pairs in 2000 and 13 000 pairs in 2005 (Table 1). The largest colonies were at Schaapen and Dassen islands, at both of which counts exceeded 6 000 pairs in some years. The colonies at Jutten and Meeuw islands also often exceeded 1 000 pairs. The sum of the smoothed estimates of the number of breeding pairs increased steadily from 6 500 pairs in 1978 to more than 16 000 pairs in 1999 and 2000 and then decreased to 13 000 pairs in 2005 (Fig. 2). At Lambert’s Bay, the northernmost of the 11 islands, there were about 10 pairs breeding from 1978–1988. The smoothed estimate then increased to a plateau of 80 pairs or more from 1996–2002, before decreasing (Fig. 3). The colony at Malgas Island also began to increase after 1988, reaching an asymptote in 2004. At Marcus Island, there was a slight decrease from 80 pairs in 1978 to 65 pairs in 1990, and then a rapid fall. The colony was extinct in 1993 (Table 1). Conversely, breeding was first recorded at Caspian Island in 1993, when 98 pairs were counted. This colony increased until 1999. The colony at Vondeling Island increased after 1991, attaining a plateau from 1996–2003 (Fig. 3). The colonies at Jutten, Meeuw and Schaapen islands all increased substantially between 1978 and 2000 and then decreased. There was little change in the size of the colony at Dassen Island between 1978 and 1985, but it then grew rapidly attaining a peak in about 1998, before decreasing again (Fig. 3). The colony at Robben Island grew after its initiation in 2000 (Table 1). At Dyer Island, easternmost of the 11 localities considered, the colony was relatively stable until 1996 and then increased (Fig. 3). At Dassen Island (222 ha), as the colony grew, the area occupied by breeding Kelp Gulls increased from 45 ha in 1978 to more than 120 ha in 2005 (Figs 3 & 4). The density of nests remained between about 60 and 80 ha–1 from 1978– 2000 and then decreased to about 30 ha –1 in 2005. At

Figure 2: The smoothed trend in the overall number of Kelp Gulls breeding at 11 localities in South Africa’s Western Cape Province, 1978–2005

Schaapen Island (41 ha), Kelp Gulls bred throughout the island during the entire period of observations. The density of nests increased from 60 ha–1 in 1978 to almost 150 ha–1 in 2000 and then fell to 125 ha–1 in 2005. In the absence of immigration, a breeding success of 1.66–3.0 chicks per pair would have been required from 1985–1994 to sustain the growth of the colony at Dassen Island during 1989–1998. Thereafter, as the colony decreased, the required breeding success fell to 0.3 chicks per pair in 2001 (Fig. 5). For the other colonies, treated together, a breeding success of more than two chicks per pair would have been required from 1980–1986 and 1991–1992. From 1997–1999, the required success would have been less than one chick per pair. When the maximum breeding success was capped at 1.63 or 2.0 chicks per pair, there were periods when the recruitment of birds aged four years to the breeding population would have been insufficient, in the absence of immigration, to achieve the observed growth, both for Dassen Island and for the other colonies treated together (Fig. 5). There were also periods when the recruitment of first breeders would have exceeded the required amount. Generally periods of

Table 1: Numbers of breeding pairs of Kelp Gulls counted at 11 localities in South Africa’s Western Cape Province, 1978–2005. The absence of values indicates that no count was undertaken. Year

Lambert’s Bay

Malgas

Marcus

1978

14

26

80

714

42

58

1009

1985 1986 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005

7 18 31 50 15 27 52 68 100 88 95 47 86 110 73 92 62 40

7 21 20 56 46 42 61 63 79 59 99 84 114 71 78 125 109 101

65 8 4 0 0 0 0 0 0 0 0 0 0 0 0 0

Jutten

528 1197 406 1457 1020 1348 1826 1239 1775 2067 2582 1960 1821 2463 1710 1728

Meeuw

Schaapen

Caspian Vondeling Dassen

185

2342

145

2892

261

2424

530

3613

5294 4944 4081 5168 5378 4808 5161 6074 6055 6178 6225 5686 5196 5124 4993

Dyer

88

126

143 1231 861 1271 1233 1738 1786 1337 1675 1562 2215 2200 1209 1398 1395 991

Robben

3291 4814 4457 150

98 109 74 94

188 237

171 95 111 124 87 128

363 114 377 366 219 259

356 276 270

4541 5019 3916 6406 5983 5865 6157 5983 6179 5088 3941 3843 3242 4389

4 15 50 80 63 167


131 110 121 110 151 141 193 198 283 157 181 320 465 183

Figure 3: The smoothed trends in the numbers of Kelp Gulls breeding at each of 11 localities in South Africa’s Western Cape Province, 1978–2005. For Dassen Island (d), the area occupied by breeding gulls (ha) and the density of nests (ha–1) are also shown for years with this information available

insufficient recruitment at Dassen Island coincided with excess recruitment at other colonies and vice versa. The overall mean clutch size was 2.04 eggs. However, there was considerable variation in clutch size, between both colonies (var = 0.011) and years (var = 0.004), but most of the variance was due to the interaction between colonies and years (var = 0.046). While clutch size increased at Dassen Island and after 1995 at Meeuw and Schaapen islands, it was variable at Jutten Island (Fig. 6). DISCUSSION Up until 1978 Kelp Gulls in the Western Cape were held at low numbers by measures implemented to control their abundance. Before the 1960s, it was the policy of the then Guano Islands Division of South Africa’s Department of Industries to destroy their eggs and chicks and to poison or shoot adult 184


birds. These practices were implemented as a result of concerns that Kelp Gulls were having an adverse influence on populations of guano-producing seabirds. For example, Hewitt (1938, p. 589) noted “Although efforts are being made to overcome the depredations of seagulls – the natural enemies of the guano-producing birds – serious inroads are still being made on bird life. Steps taken to prevent gulls from settling and breeding on islands do not appear sufficiently effective, for notwithstanding the continuous campaign against them, it seems they are present in greater numbers than previously.” The persecution caused many Kelp Gulls to nest at Meeuw and Schaapen islands in Langebaan Lagoon, where few guano-producing seabirds bred and guano was seldom harvested, after the whaling station at Langebaan was closed down (Hewitt 1938). However, an entry in the diary of D.B. Price, former inspector of the Guano Islands Division, relating to Schaapen Island, dated 4 November 1940, reads “very few duikers [cormorants] nesting on the rocks, quite a lot of sea gulls, destroyed about 50 eggs”. An entry in his diary for Dassen Island on 9 December 1940 states “About 1,800 young sea gulls killed in last few days.” For 1 August 1941 at Dyer Island, he noted “poisoning sea gulls (about 10)” and for 2 August “catching fish for use in poisoning gulls, a number poisoned (about 25).” The practice of shooting gulls was discontinued in the early 1960s but forms of control persisted until at least 1978 (Crawford et al. 1982). The numbers of Great White Pelicans in the Western Cape were also controlled in the early part of the 20th century, by direct measures and by disturbance. From 1869 or earlier until at least 1919, pelicans bred at Dyer Island, being sometimes persecuted by island staff to decrease their predation on guano-producing birds. Between 1894 and 1904, some bred at Quoin Rock, but they were eventually displaced from this locality by Cape Fur Seals Arctocephalus pusillus. From 1930 until 1954, they bred at Seal Island in False Bay, where human disturbance included riflemen shooting at seals, the commercial harvesting of seals and the use of the island for naval target practice (Brooke 1984, Crawford et al. 1995). An entry in the diary of D.B. Price relating to this island, dated 17 October 1939, reads “destroyed sea gull and pelican eggs”. From 1956, the pelicans bred at Dassen Island (Crawford et al. 1995). After 1978, numbers of Kelp Gulls remained relatively stable at most localities in the Western Cape until the next census in the mid 1980s, suggesting a low recruitment of young birds to colonies, as might be expected following a period in which reproductive output was controlled. After 1985, there was strong growth at most localities until the late 1990s. Similarly, for Herring Gulls L. argentatus the breeding population decreased for four years after termination of the practice of repeated destruction of clutches of eggs, before then increasing rapidly (Wanless et al. 1996). Kelp Gulls in the Western Cape were provided with several supplementary sources of food by humans during the latter part of the 20th century, including offal discarded by fish factories and fishing vessels and food at rubbish tips (Steele & Hockey 1990). Decreased post-fledging mortality, attributable to supplementary food, was thought the most likely reason for the increase in the gull population between 1978 and 1985. Many young birds dispersed from breeding colonies to mainland rubbish tips (Steele & Hockey 1990, Steele 1992). Great White Pelicans also were provided with additional sources of food at this time, including agricultural offal and fish at newly-constructed freshwater impoundments (Crawford et al. 1995, da Ponte Machado & Hofmeyr 2004). In the Western Cape, the numbers of pelicans increased from 20–30 pairs during 1931–1954 to 174 pairs in 1978, 434 pairs in 1991, 504 pairs in 1993 and 650 pairs in 2001

Figure 4: The area occupied by breeding Kelp Gulls at Dassen Island in 15 years between 1978 and 2005

(Cooper 1980, Crawford et al. 1995, Hockey et al. 2005). An exception to the growth of Kelp Gull colonies after the mid 1980s, was the demise of the colony at Marcus Island, which was connected to the mainland by a causeway in 1978 in order to create safe anchorage in Saldanha Bay. Gulls bred at the island and on the causeway. However, the causeway provided access to the island for mainland predators and brought increased human disturbance (Crawford et al. 1994). There was initially a gradual decrease in the size of the colony and then a precipitous decrease. It is likely that during 1991–1993 gulls from Marcus Island moved else-

where to breed. The colony at Caspian Island may have been founded as a result of this displacement. At the same time there was also an increase in the numbers breeding at Vondeling Island. Colonies at eight of the nine northern localities began to decrease between 2000 and 2004. The exception was the colony at Malgas Island, which stabilized. At Lambert’s Bay, the decrease began in 2002, four years after the construction of a bird hide at the island in 1998 (Underhill et al. 2006). In 1998, breeding success of Kelp Gulls at the island was 0.16 chicks per pair but this improved to 1.33 chicks per pair Top Predators of the Benguela System

185

Figure 6: Temporal variation in clutch size at the four larger Kelp Gull colonies in the Western Cape Province. Plotted are the best linear unbiased predictors from a generalised linear mixed model with colony, year and their interaction as random effects and the intercept as a fixed effect. The colonies are: Dassen (Da), Jutten (J), Meeuw (M) and Schaapen (S) islands. Smoothed trend lines are provided for all colonies except Jutten Island

Figure 5: The breeding success (chicks fledged per pair) necessary to produce the observed growth of the Kelp Gull colonies at Dassen Island and at other localities (considered together) in the Western Cape Province, in the absence of immigration, 1978–2005 (a); and the surplus and deficit in the recruitment of first breeders (aged four years) to the Kelp Gull colonies at Dassen Island and at other localities (considered together), if the breeding success is capped at 1.63 (b) and 2.0 (c) chicks per pair

in 2000 (Hockey et al. 2005). However, increased human activity, mainly as a result of tourism, led to the displacement of Kelp Gulls to the roofs of nearby buildings (RJMC pers. obs.). Apart from a nest on a roof of a disused building at Sinclair Island, Namibia, in 1978 (Crawford et al. 1982), this is the only instance to date of vetula nesting on roofs. Therefore, at the two relatively small colonies of Marcus Island and Lambert’s Bay, human disturbance had a major influence on trends. A decrease in the number of gulls was first evident at Dassen Island, which is where predation of gull chicks by pelicans was first reported (Crawford et al. 1997). Smoothed estimates of the size of this colony peaked during 1996– 1999. The reduced recruitment of first breeders to the colony from 1996 onwards resulted from predation of gull chicks by pelicans. The estimated breeding success of Kelp Gulls fell to 0.3 chicks per pair in 2001, suggesting that at this time pelicans were eating some 80–85% of gull chicks hatched at the island. Pelicans later began to feed on gull chicks at Jutten, Meeuw, Schaapen and Vondeling islands (Hockey et al. 2005), where decreases in gull colonies followed that at Dassen Island. Only after the pelican population had reached a level of more than 400 pairs and the Kelp Gull colony at Dassen Island had grown to more than 4 500 pairs was a predator-prey interaction between these two species first observed. Cooperative hunting was employed by pelicans, which spread out in lines across sections of the island to locate gull chicks (LU pers. obs.). The stable density of gull nests at Dassen Island as the colony expanded suggests that gulls there preferred not to increase their density of nests. Kelp Gulls seize unguarded eggs and chicks from the nests and territories of other Kelp Gulls (Hockey et al. 2005). Such cannibalism may increase at high densities of nests. As pelicans began to feed on gull chicks, the density of gull nests reduced to a lower value than 186


previously recorded. Given the feeding strategy of the pelicans, it would not have been advantageous to Kelp Gulls to maintain high densities of chicks. Densities of nests also decreased for Yellow-legged Gulls L. cachinnans after they were culled (Bosch et al. 2000). At Schaapen Island, where space was restricted, the maximum density of Kelp Gull nests was about double that recorded at Dassen Island. At Dassen Island, the clutch size of Kelp Gulls increased from about 1.9 eggs in 1990, before predation of gull chicks by pelicans, to 2.6 eggs in 2000. Clutch sizes at Meeuw and Schaapen islands also increased after the mid 1990s, when pelicans started to feed on gull chicks at these localities. Therefore, gulls responded to predation by decreasing the density of their nests, where possible, and increasing the size of clutches. Great White Pelicans also feed on other seabirds off southern Africa (Crawford et al. 1995). The Cape Fur Seal Arctocephalus pusillus pusillus, another opportunistic predator, is also having a major impact on seabirds in the Benguela ecosystem. In 2000, predation at Lambert’s Bay displaced the entire breeding colony of Cape Gannets Morus capensis from that locality (Wolfaardt & Williams 2006). From 2000– 2006, seals killed 29–83% of gannet chicks fledged at Malgas Island, leading to a decrease in the gannet colony there (Makhado et al. 2006). Unlike the situation between Lambert’s Bay and Dassen Island, there was rapid growth after 2000 of the gull colonies at Robben and Dyer islands, the southernmost colonies monitored in this study. Numbers of Kelp Gulls breeding east of 21°E increased by 71% between 1982 and 2003 (Whittington et al. 2006). Contrasting the situations at Lambert’s Bay and Marcus Island, initiation of the colony at Robben Island in 2000 was due to reduced human disturbance at the island (Calf et al. 2003). To the east of 21°E, the increase was attributed to reduced control of gulls at islands and supplementary food provided by fishing boats and at rubbish tips (Whittington et al. 2006). It is of interest to note that the recent decreases of Kelp Gulls between Lambert’s Bay and Dassen Island, and the increases farther south and east, followed a substantial eastward shift in the distribution of sardine Sardinops sagax off South Africa after 1997 (Fairweather et al. 2006). There similarly was an eastward displacement of the population of Cape gannets (Crawford et al. 2007). There were also large eastward expansions in the breeding ranges of Hartlaub’s Gull L. hartlaubii from 1995–2001 (Hockey et al. 2005) and Crowned Cormorant at some time between 1981 and 2003 (Whittington 2004). The recent altered distributions of these birds and fish prey suggest some overall influence of environmental change, and a changed distribution of at least some food items of Kelp Gulls.

At Dassen Island, from 1990–2000 the growth of the colony exceeded the estimated production there of birds aged four years, even when breeding success was assumed to be 2.0 chicks per pair. It is unlikely that a higher breeding success could have been achieved. For the other colonies grouped together, there was a similar deficit of first breeders for 1984–1990. Possible explanations for this include the prior presence of non-breeders at colonies, a higher survival rate and a decrease in the age at first breeding. Non-breeders have been observed at larid colonies and their survival may be higher than that of breeders (Pugesek & Diem 1990). In New Zealand, the nominate race dominicanus of the species may breed when aged three years, although most breed for the first time at four years (Fordham 1964). However, the out-of-phase nature between the estimated surplus and deficit of recruits at Dassen Island and the other colonies suggests the most likely provision of the home-grown deficits was the movement of first breeders between colonies. Other models also suggest that dispersal is an important determinant of the dynamics of Kelp Gull colonies in the Western Cape (Altwegg et al. in press). For Audouin’s Gull L. audouinii, food scarcity caused high emigration of young breeders away from natal colonies (Oro et al. 2004). A variety of factors have influenced the histories of the Kelp Gull and Great White Pelican populations in the Western Cape Province, illustrating the wide range of human activities that can affect seabird populations. Whereas at the start of the 20th century, management of South Africa’s islands aimed to maximise products obtained from seabirds, such as guano and eggs, by the commencement of the 21st century the emphasis had shifted to the conservation of threatened seabirds (Best et al. 1997). The Benguela ecosystem has been perturbed through the removal of the prey of specialist feeders, such as penguins (e.g. Crawford et al. 2006), and the provision of supplementary food for opportunistic predators, such as seals, gulls and pelicans. As a consequence, management interventions may be needed to maintain the diversity and conservation of the seabird assemblage (David et al. 2003). For Kelp Gulls in eastern South Africa, it is thought that improved waste management and reduction of fisheries bycatch may suffice to prevent the species attaining pest status (Whittington et al. 2006). In the west, pelicans have been effective in reducing large colonies of Kelp Gulls through limiting recruitment. Should it again prove necessary to implement controls for Kelp Gulls, it has been suggested that, although targeting adults would have a higher impact on populations than targeting other life stages, limitation of recruitment through the destruction of eggs would probably be the most efficient method because it requires much less effort than killing adults (Altwegg et al. in press). This would assert an effect similar to the natural control of pelicans. It has been shown that limiting reproductive output is an effective method for the control of Herring Gulls, though not of Lesser Black-backed Gulls L. fuscus (Wanless et al. 1996). Acknowledgements – We thank our research institutes, the National Research Foundation and the Earthwatch Institute for supporting this research. RA was supported by the Swiss National Science Foundation. CapeNature, Department of Environmental Affairs and Tourism (South Africa), Robben Island Museum and South African National Parks provided logistical support. We thank all who assisted with the counts of Kelp Gulls. This paper is a contribution to the project LMR/ EAF/03/02 of the Benguela Current Large Marine Ecosystem (BCLME) Programme.

REFERENCES Altwegg, R., Crawford, R.J.M., Underhill, L.G., Martin, A.P. & Whittington, P.A. in press. Geographic variation in reproduction and survival of Kelp Gulls Larus dominicanus vetula in southern Africa. Journal of Avian Biology. Best, P.B., Crawford, R.J.M. & Van der Elst, R.P. 1997. Top predators in southern Africa’s marine ecosystems. Transactions of the Royal Society of South Africa 52(1): 177–225. Bosch, M., Oro, D., Cantos, F.J. & Zabala, M. 2000. Short-term effects of culling on the ecology and population dynamics of the Yellow-legged Gull. Journal of Applied Ecology 37: 369–385. Brooke, R.K. 1984. South African Red Data Book – Birds. South African National Scientific Programmes Report 97: 1–213. Brooke, R.K. & Cooper, J. 1979. The distinctiveness of southern African Larus dominicanus (Aves: Laridae). Durban Museum Novitates 12: 27–37. Burger, J. & Gochfeld, M. 1981. Colony and habitat selection of six Kelp Gull Larus dominicanus colonies in southern Africa. Ibis 123: 298–310. Calf, K.M., Cooper, J. & Underhill, L.G. 2003. First breeding records of Kelp Gulls Larus dominicanus vetula at Robben Island, Western Cape, South Africa. African Journal of Marine Science 25: 391–393. Cooper, J. 1980. Fatal sibling aggression in pelicans – a review. Ostrich 51: 183–186. Coulson, J.C., Monaghan, P., Butterfield, J., Duncan, N., Thomas, C. & Shedden, C. 1983. Seasonal changes in the Herring Gull in Britain: weight, moult and mortality. Ardea 71: 235–244. Crawford, R.J.M., Cooper, J. & Shelton, P. A. 1982. Distribution, population size, breeding and conservation of the Kelp Gull in southern Africa. Ostrich 53: 164–177. Crawford, R.J.M., Dyer, B.M. & Brooke, R.K. 1994. Breeding nomadism in southern African seabirds – constraints, causes and conservation. Ostrich 65: 231–246. Crawford, R.J.M., Cooper, J. & Dyer, B.M. 1995. Conservation of an increasing population of Great White Pelicans Pelecanus onocrotalus in South Africa’s Western Cape. South African Journal of Marine Science 15: 33–42. Crawford, R.J.M., Nel, D.C., Williams, A.J. & Scott, A. 1997. Seasonal patterns of abundance of Kelp Gulls Larus dominicanus at breeding and non-breeding localities in southern Africa. Ostrich 68: 37–41. Crawford, R.J.M., Dyer, B.M. & Upfold, L. 2000. Age at first breeding and change in plumage of Kelp Gulls Larus dominicanus in South Africa. South African Journal of Marine Science 22: 27– 32. Crawford, R.J.M., Barham, P.J., Underhill, L.G., Shannon, L.J., Coetzee, J.C., Dyer, B.M., Leshoro, T.M. & Upfold, L. 2006. The influence of food availability on breeding success of African Penguins Spheniscus demersus at Robben Island, South Africa. Biological Conservation 132: 119–125. Crawford, R.J.M., Dundee, B.L., Dyer, B.M., Klages, N.T.W., Meÿer, M.A. & Upfold, L. 2007. Trends in numbers of Cape Gannets (Morus capensis), 1956/57–2005/06, with a consideration of the influence of food and other factors. ICES Journal of Marine Science 63: 169–177. Croxall, J.P. 1984. Status and Conservation of the World’s Seabirds. Cambridge; ICBP. Croxall, J.P. 1991. Seabird Status and Conservation: a Supplement. Cambridge; ICBP. Da Ponte Machado, M. & Hofmeyr, J. 2004. Great White Pelicans Pelecanus onocrotalus: waterbirds or farm birds? Bird Numbers 13(1): 11–13. David, J.H.M., Cury, P., Crawford, R.J.M., Randall, R.M., Underhill, L.G. & Meyer, M.A. 2003. Assessing conservation priorities in the Benguela ecosystem: analysing predation by seals on threatened seabirds. Biological Conservation 114: 289–292. Fairweather, T.P., van der Lingen, C.D., Booth, A.J., Drapeau, L. & van der Westhuizen, J.J. 2006. Indicators of sustainable fishing for South African sardine (Sardinops sagax) and anchovy


187

(Engraulis encrasicolus). African Journal of Marine Science 28: 661–680. Fordham, R.A. 1964. Breeding biology of the Southern Black-backed Gull: II: incubation and the chick stage. Notornis 11: 110–126. Hewitt, J.C.H. 1938. The Government Guano Islands. Annual Report. Farming in South Africa 13: 589–590. Higgins, P.J. and Davies, S.J.J.F. 1996. Handbook of Australian, New Zealand & Antarctic Birds. Volume 3 Snipe to pigeons. Melbourne; Oxford University Press. Hockey, P.A.R., Dean, W.R.J. & Ryan, P.G. 2005. Roberts Birds of Southern Africa, 7 th edn. Cape Town; The Trustees of the John Voelcker Bird Book Fund. Keijl, G.O., Brenninkmeijer, A., Schepers, F.J., Stienen, E.W.M., Veen, J. & Ndiaye, A. 2001. Breeding gulls and terns in Senegal in 1998, and proposals for new population estimates of gulls and terns in north-west Africa. Atlantic Seabirds 3: 59–74. Makhado, A.B., Crawford, R.J.M. & Underhill, L.G. 2006. Impact of predation by Cape Fur Seals Arctocephalus pusillus pusillus on Cape Gannets Morus capensis at Malgas Island, Western Cape, South Africa. African Journal of Marine Science 28: 681–687. Oro, D., Cam, E., Pradel, R. & Martínez-Abraín, A. 2004. Influence of food availability on demography and local population dynamics in a long-lived seabird. Proceedings of the Royal Society, London B 271: 387–396. Pons, J-M. & Migot, P. 1995. Life-history strategy of the Herring Gull: changes to survival and fecundity in a population subjected to various feeding conditions. Journal of Animal Ecology 64: 592– 599. Pugesek, B.H. & Diem, K.L. 1990. The relationship between reproduction and survival in known-aged California Gulls. Ecology 71: 811–817. Steele, W.K. 1992. Diet of Hartlaub’s Gull Larus hartlaubii and the Kelp Gull L. dominicanus in the southwestern Cape province,

188


South Africa. Ostrich 63: 68–82. Steele, W.K. & Hockey, P.A.R. 1990. Population size, distribution and dispersal of Kelp Gulls in the southwestern Cape, South Africa. Ostrich 61: 97–106. Tasker, M.L., Camphuysen, C.J., Cooper, J., Garthe, S., Montevecchi, W.A. & Blaber, S.J.M. 2000. The impacts of fishing on marine birds. ICES Journal of Marine Science 57: 531– 547. Underhill, L.G., Crawford, R.J.M., Wolfaardt, A.C., Whittington, P.A., Dyer, B.M., Leshoro, T.M., Ruthenberg, M., Upfold, L. & Visagie, J. 2006. Regionally coherent trends in colonies of African Penguins Spheniscus demersus in the Western Cape, South Africa, 1987–2005. African Journal of Marine Science 28: 697–704. Votier, S.C., Furness, R.W., Bearhop, S., Crane, J.C., Caldow, R.W.G., Catry, P., Ensor, K., Hamer, K.C., Hudson, A.V., Kalmbach, E., Klomp, N.I., Pfeiffer, S., Phillips, R.A., Prieto, I. & Thompson, D.R. 2004. Changes in fisheries discard rates and seabird communities. Nature 427: 727–730. Wanless, S., Harris, M.P., Calladine, J. & Rothery, P. 1996. Modelling responses of Herring Gull and Lesser Black-backed Gull to reduction of reproductive output: implications for control measures. Journal of Applied Ecology 33: 1420–1432. Whittington, P.A. 2004. New breeding locality for Crowned Cormorant. Koedoe 47: 125–126. Whittington, P.A., Martin, A.P. & Klages, N.T.W. 2006. Status, distribution and conservation implications of the Kelp Gull (Larus dominicanus vetula) within the Eastern Cape Region of South Africa. Emu 106: 127–139. Williams, A.J. 2004. Is that a Kelp, Cape or Khoisan Gull? Bird Numbers 13: 21–23. Wolfaardt, A.[C.] & Williams, [A.J.] 2006. Sealed off – predation threatens seabirds and tourism. Africa – Birds and Birding 11(2): 60–67.

Chapter 22 Geographic variation in reproduction and survival of Kelp Gulls Larus dominicanus vetula in southern Africa Res Altwegg1*, Robert J.M. Crawford1, 2, Les G. Underhill1, A. Paul Martin3 and Philip A. Whittington4 1

Avian Demography Unit, Department of Statistical Sciences, University of Cape Town, Rondebosch 7701, South Africa Marine & Coastal Management, Department of Environmental Affairs & Tourism, Private Bag X2, Rogge Bay 8012, South Africa 3 30 Himeville Drive, Bluewater Bay 6210, South Africa 4 Department of Zoology, P.O. Box 77000, Nelson Mandela Metropolitan University, Port Elizabeth 6031, South Africa * Corresponding author, e-mail: [email protected] 2

Different populations of a species tend to vary in survival and reproduction, but the extent and scale of such spatial variation are poorly known. We estimated survival and clutch size of Kelp Gulls (Larus dominicanus vetula) across their entire African range. At this large geographic scale, we found no evidence for spatial variation in survival, and there was no variation in clutch size. However, there was considerable variation in clutch size among colonies within regions. Over the whole study, mean annual survival of juvenile and adult birds was 0.44 and 0.84, and mean clutch size was 2.2 eggs. A matrix

population model showed that population growth was least sensitive to variation in clutch size, and the observed variation in clutch size could not fully account for the observed variation in population growth among colonies and regions. Our results thus suggest that dispersal and/or variation in survival (including egg/nestling survival) at a small spatial scale are also important for the spatial pattern of Kelp Gull population dynamics. These results are consistent with a metapopulation approach to spatial population dynamics.

Keywords: capture–mark–recapture, demographic variation, matrix population model, metapopulation dynamics, spatial fitness variation

Introduction Survival and reproduction can vary among populations as much as among species (Dhondt 2001). At the scale of metapopulations, such demographic variation can result in net increases in some populations and decreases in others (Ringsby et al. 1999, 2002, Sæther et al. 1999). Demographic variation among populations at a larger scale, across the geographic distribution of a taxon probably is also significant, but has rarely been investigated (e.g. Blondel et al. 1992, Harris et al. 2005). Recently, Frederiksen et al. (2005a) found considerable variation in demography and dynamics among populations of kittiwakes Rissa tridactyla across their range in the northern Atlantic and Pacific. It is currently unknown, whether such large-scale variation in demography is the rule or the exception among birds and other animals. Here we examine survival and reproduction of Kelp Gulls Larus dominicanus vetula across their African range. Kelp gulls Larus dominicanus have a wide distribution in the southern hemisphere, including southern Africa (Croxall 1984, 1991, Higgins and Davies 1996, Hockey et al. 2005). The race, L. d. vetula is endemic to southern Africa (Brooke and Cooper 1979), except for a few birds that bred north of the equator in Senegal in 1998 (Keijl et al. 2001). Other than that, L. d. vetula breeds from Isla dos Tigres, southern Angola to Hamburg in South Africa’s Eastern Cape Province (Dean et al. 2002, Hockey et al. 2005, A.J. Tree in litt.).

South Africa is home to the largest part of the L. d. vetula population: in 1976–1981, there were 7650 pairs in the Western Cape province, and 1100 in the Eastern Cape province (Crawford et al. 1982). Namibia had an additional 2300 pairs at that time (Crawford et al. 1982). Since then, the overall population of L. d. vetula increased from about 11 200 pairs to 22 500 pairs or more in 2000 (Hockey et al. 2005, who also include a distribution map). The mean annual population growth rate was thus ~3% (λ = 1.033). These population increases were accompanied by a probable range expansion into southern Angola (Dean et al. 2002). Over the same time, populations increased at nine colonies in the Western Cape from 6 500 to 17 900 pairs (Steele and Hockey 1990, Hockey et al. 2005), which corresponds to a mean annual population growth rate λ = 1.049. Likewise, the Kelp Gull population in the Eastern Cape as a whole increased at λ = 1.022 since 1982, but individual colonies increased, decreased or stayed constant (Whittington et al. 2006). These observations suggest spatial variation in Kelp Gull population dynamics at several scales. Kelp gulls in southern Africa thus offer an excellent opportunity to study spatial variation in demography and to examine to what extent such variation may have led to differences in local population dynamics. At the largest scale, we distinguish three geographic regions, Namibia, Western Cape and Eastern Cape. The Namibian colonies are separated from those in the Western Cape by ca. 500 km of coastline where only few Top Predators of the Benguela System

189

Kelp Gulls breed (150 pairs, Crawford et al. 1982). The colonies in Western and Eastern Cape are closer to each other, but ring recoveries suggest that there is nevertheless little movement between these two regions (Underhill et al. 1999). Within these regions, we examined variation among colonies. Ringing recoveries suggest that there is little movement of fledglings away from their natal colonies, as 75% of birds ringed as chicks and recovered at ages five years and older were within 18 km of their place of ringing (Underhill et al. 1999). The rapid increase in Kelp Gull numbers over the last decades indicates that the populations were below their carrying capacities at least until the recent past and we can use simple density-independent population models to describe their dynamics. Kelp Gull populations were controlled at guano islands until the early 1960s, mainly through egg and chick destruction (Crawford et al. 1982, Randall et al. 1981). At some localities, control continued until 1978 (Crawford et al. 1982). Kelp Gull populations may thus still be recovering from these control measures, but supplementary food made available by human activities is probably also an important factor in the increases (Steele 1992, Steele and Hockey 1990). The population densities are now so high that control measures were recently reintroduced at two Namibian islands (Hockey et al. 2005). A second goal of this study is to provide critical information for management plans. Both conservation and control efforts tend to be most effective if they target a life stage that has the greatest effect on population dynamics (Caswell 2001, Govindarajulu et al. 2005). Yet, culling programmes implemented by management agencies have not always taken account of all factors responsible for the population dynamics of colonies and the effects of culls (Bosch et al. 2000). The construction of demographic models can assist in the consideration of such factors and hence in better decisions. For example, models have suggested that limiting reproductive effort can be an effective method of control for certain species of gull but not for others (Wanless et al. 1996). Methods Adult Kelp Gulls were mostly caught in walk-in traps placed over the nest, and incidentally in mist nets targeting other species. Juveniles were caught and ringed prior to fledging. Regular checks were kept for breeding at known breeding sites, and nests were counted and their content recorded as close as possible to the period when breeding activity was at its peak. We estimated survival of Kelp Gulls along the southern African coast using two independent data sets. The first data set included 10 059 individuals ringed at ca. 60 sites along the coast of Namibia and South Africa between 1990 and 2004; the second data set included 75 adult individuals ringed and observed at Lambert’s Bay in the Western Cape between October 1999 and January 2006. We treated the data sets separately because they differed in their quality. The birds in the first data set were ringed with a numbered metal ring, whereas the birds in the second data set additionally received a plastic ring on which a unique combination of two letters or numbers could be read from a distance with the use of binoculars. The first data set therefore consisted of recaptures of live birds and recoveries of dead birds, whereas the second consisted of resightings and recaptures of live birds (391 resightings and recaptures of 75 individuals in total). We were interested in finding possible large-scale geographic variation in survival and therefore divided the first data set along the coastline into three regions, motivated by 190


the near-discrete breeding populations: Namibia, Western Cape, and Eastern Cape. In Namibia, the data comprised 2270 individuals ringed as juveniles and 14 ringed as adults; 10 were later recaptured alive at least once and 51 were found dead. In the Western Cape, 4265 individuals were ringed as juveniles and 246 as adults; 37 were recaptured at least once and 123 were found dead. In Eastern Cape, 3240 juveniles and 38 adults were ringed; 17 were recaptured at least once and 169 were found dead. The individuals that were either recaptured or recovered had moved 0 to 1169 km between encounters, but only 24 individuals moved from one region to another. First-year birds usually disperse away from breeding colonies. In South Africa, 60% of birds may not return to natal islands until they are aged two years. Three-year-old birds have visited non-natal colonies, but breeders have all nested at natal colonies (Crawford et al. 2000). However, the recent establishment of several new colonies means some birds must settle at non-natal colonies or breed at a different colony from the one where they first bred (e.g. Calf et al. 2003, Whittington et al. 2006). We analysed the first data set using combined live-recapture and dead-recovery models (Brownie et al. 1985, Burnham 1993, Lebreton et al. 1992), implemented in program MARK (White and Burnham 1999). These models allow separating true survival from the probability of recapturing a live individual (referred to as recapture rate) and the probability that a dead individual is found and reported (referred to as recovery rate). These models also estimate fidelity, the probability of staying at the study site, because recoveries can be made outside the study area. In our case, ringing activity was not confined to a restricted area, and we interpret the fidelity rates therefore as movement away from the ringing locations. Different causes of death could have determined whether a bird dies at the ringing locations or somewhere else, and would thus have affected the fidelity rates, as would heterogeneity in recapture probability. We examined differences in survival, recapture, recovery and fidelity between young gulls (first year of their life) and older ones. We expected the age effects to be largest between first year gulls and older ones, but we also examined a model allowing for full age dependence in survival up to four years when Kelp Gulls start breeding (Crawford et al. 2000). A third model held survival constant for 1- to 3-yearold immature birds. We examined differences in all model components among the three regions, but the data were too sparse to examine temporal variation. Our most general model therefore included interactive effects of age and region on all four model components. Then we simplified each component in turn and evaluated the models using the sample size adjusted Akaike’s Information Criterion (AICc, Burnham and Anderson 2002). The methods we used assume that all individuals within a group have the same survival, recapture, recovery, and fidelity rates. We examined how closely the data met these assumptions for our most general model (Model 5, Table 1) using the median c-hat approach implemented in program MARK (White and Burnham 1999). Using 20 replicated simulations at 12 levels of overdispersion between 1 and 2, this procedure estimated c-hat at 1.52 (se = 0.01). A second method to estimate goodness-of-fit, the parametric bootstrap procedure in program MARK, yielded a slightly higher estimate of c-hat (1.64, based on 200 simulations, comparing deviances), but is known to overestimate c-hat when the data are sparse (White 2002). These estimates of c-hat indicate that the model structurally fits the data, but that there is some heterogeneity among individuals. We therefore used the chat obtained by the first approach and based model selection on Quasi-likelihood Akaike’s Information Criterion,

QAICc (Anderson et al. 1994). Our results do not depend on the choice between these two estimates of c-hat. The second data set consisted of resightings and recaptures of live gulls, and we used ordinary capture–mark– recapture models to estimate survival and recapture rates (Lebreton et al. 1992). This data set allowed us to examine variation in survival and recapture over the years, in addition accounting for differences in recapture between breeding seasons and non-breeding seasons. Using identical procedures for estimating c-hat as above, we found no evidence of overdispersion in the second data set and therefore made no adjustments (c-hat = 0.896, se = 0.076). Eleven colonies in Namibia, eight colonies in the Western Cape, and six colonies in the Eastern Cape were visited once a year during the peak breeding season (10 October to 13 December) for one to nineteen years between 1985 and 2004. During these surveys, all nests were counted and their contents recorded. Here we concentrate on the nests that contained at least one egg. The number of eggs (or eggs plus downy chicks) was taken as a measure of brood size. Clutch sizes ranged from one to five eggs, or one egg and one to two downy chicks. The clutches with more than three eggs could possibly have been laid by two females (Hockey et al. 2005), but excluding those nests (n = 43 of n = 14995 in total) did not change our results. We partitioned the variance in clutch size into three components, within colony, among colonies, and among the three geographic regions. For this we used a generalized linear mixed model with Poisson errors and a log link function implemented in procedure glmmPQL in program R 2.2.1 (Venables and Ripley 2002, R Development Core Team 2003). We added year as a fixed effect, and time of the season (days since 1 October) as a covariate to the model to account for potential variation in clutch sizes caused by these factors. Laying occurs from late September to January, with replacement eggs laid into February (Hockey et al. 2005). We illustrate the effect of variation in demography on population growth rates using a simple matrix population model. Assuming a post-breeding census and a one-year projection interval, we project the number of individuals in year t + 1 from the number of individuals in year t:

nt +1 = Ant ,

eqn. (1)

where n is a vector with the number of individuals in each age-class, and A is the projection matrix:

0   sy A= 0  0 

0

0

0

0

si 0

0 si

c * sn * sa * 0.5   0   0   sa 

eqn. (2)

The matrix elements are survival of first-year birds (sy ), immature birds (si ), and adult birds (sa ), and reproduction, which is the product of clutch size (c), nestling survival (sn ), adult survival and sex ratio (assumed to be 0.5). For L. d. vetula, the age at first breeding is four years (Crawford et al. 2000), and we assumed that adult individuals breed every year. Our study yielded estimates for all of these variables except nestling survival (survival from the time the nests were censused until the young gulls reached a size at which they could be ringed). Therefore, we chose a value of nestling survival that produced the mean observed population growth rate of Kelp Gulls in southern Africa over the last 30 years. Assuming a stable age distribution, the dominant eigenvalue of A gives the population growth rate (λ), and the sensitivity of λ to changes in each matrix element shows how

much population growth is affected by variation in each fitness component (Caswell 2001). We conducted all matrix analyses in program R 2.2.1 (R Development Core Team 2003). Results For the first capture–recapture data set, model selection favoured a model allowing for differences in recapture, recovery and fidelity rates among the areas (Model 1, Table 1). This model further implied that survival and fidelity differed between the first-year birds and older ones (Figure 1). The best model was almost twice as likely as the second best model, which allowed for variation in survival over three ageclasses (Model 2, Table 1; ratio of QAICc weights: 0.46/0.25 = 1.84). According to the best model, juvenile survival was 0.44 (95% confidence interval: 0.35 to 0.54) and adult survival 0.84 (0.77 to 0.89) over the whole area of our study (i.e. nearly the whole range in Africa). The probability of recapturing a live individual was 0.234 (0.085 to 0.501) in Namibia, 0.013 (0.006 to 0.031) in the Western Cape, and 0.003 (0.001 to 0.006) in the Eastern Cape. The probabilities of finding and reporting a dead individual were 0.027 (0.019 to 0.038) in Namibia, 0.039 (0.030 to 0.050) in the Western Cape, and 0.064 (0.051 to 0.077) in the Eastern Cape. The fidelity rates were 0.023 (0.008 to 0.063) and 0.740 (0.397 to 0.925) for Namibian juveniles and adults, and 0.608 (0.130 to 0.942) and 0.995 (0.920 to 0.999) for juveniles and adults in the Western Cape. In the Eastern Cape, there was hardly any emigration even though we could not obtain a reliable estimate for juveniles (1, se = 0.00003 for adults). For the second data set, model selection favoured a model with constant survival and time dependent recapture probabilities (best model: Deviance = 331.08, number of parameters, K, = 13, AICc = 715.03; model with time dependent survival: ∆ AICc = 4.67; model with constant recapture: ∆AICc = 10.21). Annual adult survival at Lambert’s Bay between 1999 and 2006 was 0.84 (95% CI: 0.78 to 0.89) and thus identical to the mean estimate for rest of the range. The recapture rates at Lambert’s Bay varied over time from 0.19 (0.09 to 0.37) to 0.64 (0.48 to 0.78). Clutch size varied significantly from year to year (F18,14951 = 25.72, P < 0.001), and increased slightly (slope on log scale: 0.001; i.e. ca. 0.1% per day) through the season (F1,14951 = 4.509, P = 0.034), and we therefore left these fixed effects in the model when estimating the spatial variance. Of the remaining variance, most was due to variation among nests within colonies (var = 0.221), less due to variation among colonies nested within region (var = 0.0061), and almost nothing due to variation among regions (var = 1.6 × 10–9). Our data were insufficient to examine whether the temporal pattern was similar among colonies or regions (sensu Frederiksen et al. 2005b). We compared the colonies in 1995, when 15 of the 25 colonies were counted, based on the best linear unbiased predictors (BLUP) of the random effects (colony nested within region). In that year, estimated mean clutch size varied between 1.87 at Meeuw Island in the Western Cape and 2.37 at Penguin Island in Namibia, around a mean of 2.12 (the mean over the whole study period was 2.20). There was no significant correlation between mean clutch size and colony size (r = 0.20, N = 25, P = 0.33). We used a simple matrix population model (equations 1 and 2) to examine population level consequences of the estimated survival rates and clutch size. This model requires also an estimate of egg/nestling survival between the time of the nest censuses and the age at which young gulls generally were ringed, immediately prior to fledging. We could not get such an estimate directly. However, with a mean Top Predators of the Benguela System

191

Table 1: Summary of model selection for Kelp Gull survival in Southern Africa. The models are combined mark–recapture and dead-recovery models. They consist of four parts modelling survival (S), recapture (P), recovery (r), and fidelity (F) rates. The subscripts denote the factors affecting a particular rate. We considered the effect of age (age2: first-year birds vs older; age3: juveniles, immatures, adults; age5: full age dependence up to age 4 years) and area (Namibia, Western Cape, and Eastern Cape) on each of the model components. Additive effects are denoted by ‘+’, interactive effects by ‘*’. Constant model components are denoted by ‘.’. The models are sorted according to their Quasi-likelihood Akaike’s Information Criterion (QAICc), where a lower value indicates a better model. Delta QAICc is the difference in QAICc between the current and the best model; QAICc weights give the relative support each model has compared to the other ones in the set; K is the number of parameters; and QDeviance is the model deviance divided by the variance inflation factor c-hat, which was 1.52 Model 1) S age2 P area r area F 2) S age3 P area r area F

QAICc

Delta QAICc

QAICc weights

K

QDeviance

age2+area

3098.092

0.000

0.463

12

511.102

age2+area

3099.288

1.196

0.254

13

508.973

3100.205

2.113

0.161

14

509.204

3101.727

3.634

0.075

12

514.736

3) S age2+area P

area

r area F

4) S age2+area P

area

r.F

5) S age5 P area r area F

age2+area

age2+area

age2+area

3102.736

4.643

0.045

15

508.408

6) S age2+area P

area r area F age2

3110.020

11.928

0.001

12

523.029

7) S age2*area P

age2*area

3115.430

17.338

0.000

24

504.350

8) S age2+area P

area

3119.232

21.139

0.000

13

530.236

3124.561

26.469

0.000

12

537.574

9) S age2+area P . r

r age2*area F

r area F

area

F

age2*area

area

age2+area

clutch size of 2.20 and the best estimates for juveniles, immatures and adults, nestling survival would have had to be 0.74 for the model to produce the observed mean population growth rate λ = 1.033. At these parameter values, the model suggests that Kelp Gull population dynamics are most sensitive to changes in adult survival (sensitivity: 0.64), less sensitive to changes in juvenile survival (0.28) and reproduction (0.18), and least sensitive to changes in immature survival (0.15). This model also allowed us to explore the potential effect of the observed variation in clutch size among colonies on population growth. The observed variation could have led to variation in the population growth rate λ ranging from 1.018 to 1.048. While this is a considerable amount of variation (corresponding to population doubling times between 38 and 15 years), variation in clutch size alone cannot account for all the observed variation in population growth among colonies (some colonies decreased, Whittington et al. 2006).

Figure 1: Age-specific survival of Kelp Gulls in Southern Africa. Dots and vertical lines show the maximum likelihood estimates and their 95% confidence intervals taken from the model with full age dependence up to age 4 years (Model 5, Table 1). The solid line shows the best model (Model 1, Table 1), and the dashed line shows the second best model (Model 2, Table 1) 192


Discussion We estimated survival and clutch size of Kelp Gulls across their entire range in southern Africa, and examined spatial variation in these demographic rates. We found no evidence for spatial variance in survival, and there was no variance among regions in clutch size. However, there was considerable variance in clutch size among colonies within regions. Kelp gull populations increased in southern Africa over the last 30 years, and the rate of increase varied among regions and colonies, with a few colonies even decreasing (Whittington et al. 2006). Using a simple matrix population model, we found that the observed variation in clutch size could not account for all of this variation in population growth. There may have been variation in survival (including nestling survival) among colonies, but we did not have sufficient data to examine variation at this scale. Our results suggest that most of the spatial variance in demography was at the among-colony level, and very little at a larger inter-regional level. Our results are consistent with the metapopulation paradigm focusing on variation among local populations that are connected through dispersal, (Hanski 1998, Hanski and Gilpin 1997). Our results are also consistent with Harris et al. (2005), who found no spatial variation in survival of Atlantic puffins (Fratercula artica), but they are in contrast to a recent study showing large-scale variation in demography of kittiwakes Rissa tridactyla across the north Atlantic and Pacific oceans (Frederiksen et al. 2005a). Our survival estimate of 0.84 for adult Kelp Gulls is the first such estimate for this species, and is somewhat lower than comparable estimates for the closely related gulls Larus fuscus and L. argentatus, for which survival rates around 0.9 were reported (Wanless et al. 1996). Our mean estimate of clutch size (2.20) is close to earlier estimates for Kelp Gulls in South Africa (2.1, 2.1, and 2.2, Calf et al. 2003, Crawford et al. 1982, Williams et al. 1984). A Kelp Gull population in Wellington, New Zealand (L. d. dominicanus subspecies), had a mean clutch size of 2.4, and 1.3 young per breeding pair reached the flying stage (Fordham 1964). Nestling survival was therefore lower in that study (0.54) compared to our indirect estimate of 0.74. On the other hand, in a newly established colony in South Africa nestling survival was 0.88

(Calf et al. 2003). In Golfo San Jorge, Patagonia, mean clutch sizes were 2.3 and 2.5 in two different years (Yorio and Garcia Borboroglu 2002). Further south, Kelp Gull clutch sizes were 1.9 on the subantarctic Marion Island (Williams et al. 1984), and 2.6 near Palmer Station on the Antarctic Peninsula (Parmelee et al. 1977). Even at this very large spatial scale and across two subspecies, clutch size thus seems to vary little more than among colonies in South Africa. We probably slightly underestimated clutch size because some of the clutches may have been incomplete at the time of the census. However, the laying interval in Kelp Gulls is short (2.5 days) compared to the incubation period (26–27 days, Williams et al. 1984) so that this bias would have been small. Although there is substantial dispersal of adult Kelp Gulls away from some breeding colonies in the Western Cape (Crawford et al. 1997, but see Underhill et al. 1999), the high fidelity rate estimated for these birds suggests they remain mainly within the regions considered (Namibia, Western Cape, Eastern Cape). First-year Kelp Gulls had much lower fidelity rates than older birds, implying that they undertake longer movements, as also indicated by recoveries and resightings of banded birds (Underhill et al. 1999). Heterogeneity in recapture rates may also have lowered the fidelity estimates (Frederiksen, pers. comm.). Kelp gull populations in Namibia have increased to a point where population control is being resumed again. In addition to the discontinued control efforts (Crawford et al. 1982) and supplementary food from offal (Steele and Hockey 1990), which probably increased survival, birds may also breed at a younger age now than in earlier years. In New Zealand, birds of the nominate race dominicanus may breed when aged three years, but most breed for the first time when four years old (Fordham 1964, Higgins and Davies 1996). In South Africa, one of 11 known-aged birds may have bred three years old, but 50–80% of birds aged four years and all older birds bred (Crawford et al. 2000). In the Eastern Cape, however, some individuals acquire full adult plumage at 2– 3 years and one know aged bird bred at the age of three (PAW pers. obs., SAFRING, unpublished data). Even though we cannot identify the cause of the observed population increases, our matrix model showed that Kelp Gull population growth was most sensitive to changes in adult survival. Conservation and management efforts for this species would thus potentially have the highest impact if they targeted the adult life stage. Acknowledgements – Data collection was coordinated by SAFRING. RA and PAW were supported by the South African National Research Foundation (NRF). LGU acknowledges support from the Sea and Shore 2 programme of the NRF. We also acknowledge our respective institutions for support. We thank M.G. Boorman, B.L. Dundee and staff of Ministry of Fisheries and Marine Resources, Namibia for banding Kelp Gulls in Namibia. We are grateful to B.M. Dyer, L. Upfold, V.L. Ward and others who assisted with the banding and resighting of Kelp Gulls in South Africa, and to Morten Frederiksen, Theunis Piersma and an anonymous reviewer for comments on the manuscript. This paper is a contribution of Project LMR/EAF/03/02 of the Benguela Current Large Marine Ecosystem (BCLME) Programme.

References Anderson, D.R., Burnham, K.P. and White, G.C. 1994. AIC model selection in overdispersed capture–recapture data. Ecology 75: 1780–1793. Blondel, J., Pradel, R. and Lebreton, J-D. 1992. Low fecundity insular blue tits do not survive better as adults than high fecundity mainland ones. J. Anim. Ecol. 61: 205–213. Bosch, M., Oro, D., Cantos, F.J. and Zabala, M. 2000. Short-term effects of culling on the ecology and population dynamics of the yellow-legged gull. J. Appl. Ecol. 37: 369–385.

Brooke, R.K. and Cooper, J. 1979. The distinctiveness of southern African Larus dominicanus (Aves: Laridae). Durban Museum Novitates 12: 27–37. Brownie, C., Anderson, D.R., Burnham, K.P. and Robson, D.S. 1985. Statistical inference from band recovery data – a handbook. US Fish and Wildlife Service, Resource publication 156. 2nd edition. Burnham, K.P. 1993. A theory for combined analysis of ring recovery and recapture data. In: Lebreton, J-D. and North, P.M. (eds). Marked individuals in the study of bird populations. Birkhäuser Verlag, Basel. pp. 199–213. Burnham, K.P. and Anderson, D.R. 2002. Model selection and multimodel inference: a practical information–theoretic approach. Springer. Calf, K.M., Cooper, J. and Underhill, L.G. 2003. First breeding records of kelp gulls Larus dominicanus vetula at Robben Island, Western Cape, South Africa. Afr. J. Marine Sci. 25: 391–393. Caswell, H. 2001. Matrix population models. Sinauer. Crawford, R.J.M., Cooper, J. and Shelton, P.A. 1982. Distribution, population size, breeding and conservation of the Kelp Gull in southern Africa. Ostrich 53: 164–177. Crawford, R.J.M., Dyer, B.M. and Upfold, L. 2000. Age at first breeding and change in plumage of kelp gulls Larus dominicanus in South Africa. S. Afr. J. Marine Sci. 22: 27–32. Crawford, R.J.M., Nel, D.C., Williams, A.J. and Scott, A. 1997. Seasonal patterns of abundance of Kelp Gulls Larus dominicanus at breeding and non-breeding localities in southern Africa. Ostrich 68: 37–41. Croxall, J.P. 1984. Status and conservation of the world’s seabirds. ICBP. Croxall, J.P. 1991. Seabird status and conservation: a supplement. ICBP. Dean, W.R.J., Dowsett, R.J., Sakko, A. and Simmons, R.E. 2002. New records and amendments to the birds of Angola. Bull. Brit. Ornithol. Club 122: 180–185. Dhondt, A.A. 2001. Trade-offs between reproduction and survival in tits. Ardea 89: 155–166. Fordham, R.A. 1964. Breeding biology of the southern black-backed gull: II: incubation and the chick stage. Notornis 11: 110–126. Frederiksen, M., Harris, M.P. and Wanless, S. 2005a. Inter-population variation in demographic parameters: a neglected subject? Oikos 111: 209–214. Frederiksen, M., Wright, P. J., Harris, M. P., Mavor, R. A., Heubeck, M. and Wanless, S. 2005b. Regional patterns of kittiwake Rissa tridactyla breeding success are related to variability in sandeel recruitment. Mar. Ecol. Progr. Ser. 300: 201–211. Govindarajulu, P., Altwegg, R. and Anholt, B.R. 2005. Matrix model investigation of invasive species control: bullfrogs on Vancouver Island. Ecol. Appl. 15: 2161–2170. Hanski, I. 1998. Metapopulation dynamics. Nature 396: 41–49. Hanski, I. and Gilpin, M.E. 1997. Metapopulation dynamics: ecology, genetics, and evolution. Academic Press. Harris, M.P., Anker-Nilssen, T., McCleery, R.H., Erikstad, K.E., Shaw, D.N. and Grosbois, V. 2005. Effect of wintering area and climate on the survival of adult Atlantic puffins Fratercula arctica in the eastern Atlantic. Mar. Ecol. Progr. Ser. 297: 283–296. Higgins, P.J. and Davies, S.J.J.F. 1996. Handbook of Australian, New Zealand & Antarctic birds. Volume 3 Snipe to pigeons. Oxford University Press. Hockey, P.A.R., Dean, W.R.J. and Ryan, P. G. 2005. Roberts – Birds of Southern Africa. The Trustees of the John Voelcker Bird Book Fund. Keijl, G.O., Brenninkmeijer, A., Schepers, F. J., Stienen, E.W.M., Veen, J. and Ndiaye, A. 2001. Breeding gulls and terns in Senegal in 1998, and proposals for new population estimates of gulls and terns in north-west Africa. Atlantic Seabirds 3: 59–74. Lebreton, J.D., Burnham, K.P., Clobert, J. and Anderson, D.R. 1992. Modeling survival and testing biological hypotheses using marked animals: a unified approach with case studies. Ecol. Monogr. 62: 67–118. Parmelee, D.F., Fraser, W.R. and Neilson, D.R. 1977. Birds of the Palmer Station area. Antarc. J. United States 12: 14–21. R Development Core Team. 2003. R: A language and environment for statistical computing. R Foundation for Statistical Computing. Randall, R.M., Randall, B.M., Batchelor, A.L. and Ross, G.J.B. 1981. The status of seabirds associated with islands in Algoa Bay,


193

South Africa, 1973–1981. Cormorant 9: 85–104. Ringsby, T.H., Sæther, B.E., Altwegg, R. and Solberg, E.J. 1999. Temporal and spatial variation in survival rates of a house sparrow, Passer domesticus, metapopulation. Oikos 85: 419–425. Ringsby, T.H., Sæther, B.-E., Tufto, J., Jensen, H. and Solberg, E.J. 2002. Asynchronous spatiotemporal demography of a house sparrow metapopulation in a correlated environment. Ecology 83: 561–569. Sæther, B-E., Ringsby, T.H., Bakke, Ø. and Solberg, E.J. 1999. Spatial and temporal variation in demography of a house sparrow metapopulation. J. Anim. Ecol. 68: 628–637. Steele, W.K. 1992. Diet of Hartlaub’s Gull Larus hartlaubii and the Kelp Gull L. dominicanus in the southwestern Cape province, South Africa. Ostrich 63: 68–82. Steele, W.K. and Hockey, P.A.R. 1990. Population size, distribution and dispersal of Kelp Gulls in the southwestern Cape, South Africa. Ostrich 61: 97–106. Underhill, L.G., Tree, A.J., Oschadleus, H.D. and Parker, V. 1999. Review of ring recoveries of waterbirds in southern Africa. Avian Demography Unit, University of Cape Town. Venables, W.N. and Ripley, B.D. 2002. Modern applied statistics

194


with S. Springer-Verlag. Wanless, S., Harris, M.P., Calladine, J. and Rothery, P. 1996. Modelling responses of herring gull and lesser black-backed gull populations to reduction of reproductive output: Implications for control measures. J. Appl. Ecol. 33: 1420–1432. White, G.C. 2002. Discussion comments on: the use of auxiliary variables in capture–recapture modelling. An overview. J. Appl. Stat. 29: 103–106. White, G.C. and Burnham, K.P. 1999. Program MARK: Survival estimation from populations of marked animals. Bird Study 46: S120–139. Whittington, P.A., Martin, A.P. and Klages, N.T.W. 2006. Status, distribution and conservation implications of the Kelp Gull (Larus dominicanus vetula) within the Eastern Cape region of South Africa. Emu 106: 127–139. Williams, A.J., Cooper, J. and Hockey, P.A.R. 1984. Aspects of the breeding biology of the Kelp Gull at Marion Island and in South Africa. Ostrich 55: 147–157. Yorio, P. and Garcia Borboroglu, P. 2002. Breeding biology of Kelp Gulls (Larus dominicanus) at Golfo San Jorge, Patagonia, Argentina. Emu 102: 257–263.

Chapter 23 Prefledging energetics of Kelp Gull (Larus dominicanus vetula) chicks in a warm environment †G. Henk Visser1, Trineke Bakker1, Kathy M.C. Tjørve2 z and Les G. Underhill 2 * 1

University of Groningen, Behavioural Biology, P.O. Box 14, 9750 AA Haren, The Netherlands Avian Demography Unit, University of Cape Town, Rondebosch 7701, South Africa * Corresponding author: e-mail: [email protected]

2

Several studies have measured prefledging energy budgets of gull and tern chicks living in a cold environment. They showed that thermoregulation costs are the largest part of their energy expenditure. This is one of the first studies which determine, with the use of DLW, the prefledging energy budget of a gull (Kelp Gull; Larus dominicanus vetula) in a warm environment. The research took place on Robben Island in South Africa where the average ambient temperature during the field season was about 20°C. Behavioural observations showed that the chicks are mainly inactive without being brooded, 92% of the time during the day. This is con-

sistent with the semi precocial mode of development, the chicks are fed by their parents and don’t show much activity. Of the total MEI (metabolisable energy intake) 22% was used for tissue gain and the remaining energy is used for RMR (resting metabolic rate), activity, syntheses and thermoregulation. The predictions of the energy expenditure are close to the measured values in this research. When we compare residuals with the other tern and gull studies at different latitudes we find a significant effect of latitude; an increase of about 1% in peak and total MEI per degree latitude.

Keywords: Kelp Gull, Larus dominicanus vetula, Robben Island, energetics, doubly-labelled water

Introduction Gull and tern (Laridae) chicks exhibit the semi-precocial mode of development, whereby they leave the nest within a few days after hatching but are still fed by their parents at least until fledging (Starck and Ricklefs 1998). Time budget analysis of semi-precocial tern chicks has indicated that they can be at rest for up to 90% of their time (Klaassen et al. 1989). Therefore, chicks with this developmental mode exhibit low energetic costs for locomotion, potentially resulting in a high growth efficiency compared with precocial shorebird chicks which forage for themselves (Schekkerman and Visser 2001). However, because of their exposed life-style, semi-precocial chicks may need to allocate a substantial amount of energy to temperature regulation in order to compensate for the heat lost through their downy plumage. In young chicks, these costs are reduced by parental brooding (Klaassen 1994), whereas older chicks rely on shivering thermogenesis to maintain their body temperature (Visser and Ricklefs 1993, Hohtola and Visser 1998). At high latitudes these costs can amount to about 25% of the entire prefledging energy budget of growing chicks. It is thought that at low latitudes, chicks expend much less energy on temperature regulation (Klaassen 1994), possibly resulting in a lower level of energy requirement between hatching and

z Present

fledging, and therefore in a higher growth efficiency. The construction of energy budgets represents an important tool for understanding the effects of developmental mode and climate on postnatal development. Of the total metabolisable energy intake (i.e., the gross energy intake minus fecal and urinary energy loss) of larid chicks, only about 25% is used for the synthesis of tissue (Drent et al. 1992, Klaassen et al. 1992, Klaassen 1994, Weathers 1996). The remainder is dissipated as heat, as a by-product of tissue synthesis or due to the maintenance of physiological functions when at rest at thermoneutrality, heat increment of feeding, locomotion or temperature regulation (Ricklefs 1974). In the past, the question of latitudinal adaptation in larid postnatal development has mainly been addressed by constructing energy budgets for six species in temperate-zone, arctic, and high arctic habitats. Until now, an energy budget for growth has been constructed for only two (sub) tropical species (Common Tern [Sterna hirundo] and Sooty Tern [Sterna fuscata], Ricklefs and White 1981). Clearly, species with (sub)tropical breeding distributions are underrepresented in the existing dataset. Therefore, using the Doubly Labelled Water (DLW) method and growth measurements, we constructed the energy budget of the Kelp Gull Larus dominicanus vetula in South Africa, where the

address: Lista Bird Observatory, Research Group, Fyrveien 6, N-4563 Borhaug, Norway Top Predators of the Benguela System

195

temperatures are high relative to other localities where the energetics of gulls and terns have been investigated. In this way, comparisons can be made to provide more insight the adaptations and differences between birds breeding different latitudes. Materials and Methods Study species The Kelp Gull inhabits temperate-zone and circumpolar areas in the southern hemisphere. The vetula subspecies is resident and common in southern Africa (Crawford et al. (1982), Crawford (1997), Underhill et al. (2001) and Calf et al. (2003)). Adults primarily forage at sea where they feed on fish and fish offal, but they also feed along the coast on beaches and sandy shores (Crawford 1997). Kelp gulls breed in colonies, and complete clutches typically contain 2– 3 eggs which are incubated for about 28 days. Chicks are fed by both parents. After hatching chicks stay in the nest for the first few days; thereafter they are mobile. Study area and weather measurements Measurements were obtained at two colonies on Robben Island, South Africa (33°47'S, 18°21'E), every 3–4 days between October 2003 and January 2004. Both colonies were situated at a distance of about 100 m from the coastline. Ambient temperatures were obtained from a weather station close to the coastline, and were stored in a data logger at 10min intervals. Growth rate Upon discovery, nests and their eggs were marked and subsequently inspected every 2–3 days, with increased frequency prior to the expected hatching date. After hatching each chick was weighed with a Pesola spring balance (to the nearest 0.1 g) and given unique yellow markings with picric acid. The hatching day was designated as Day 0. Each chick was weighed repeatedly between hatching and fledging, or until death. We only used growth data of chicks which fledged successfully. For each individual, different calculation procedures (SPSS 14.0) were tested to describe the growth rate: (1) the logistic growth curve with a fitted asymptote, (2) the Gompertz growth curve with a fitted asymptote (Ricklefs 1973). The model with the best fit was used to construct the general growth curve, which was also used to construct the energy budget for successful growth. Behavioural observations Behavioural observations of chicks were performed between November 10 and January 15. Observations took place at all times of day between 5:45 am and 20:00 pm in bouts of approximately two hours. We categorized behavioural activity following Klaassen et al. (1989) and Klaassen (1994): parental brooding, standing, sitting and lying in the vegetation, preening, begging, eating, locomotion, social aggressive interactions. Additionally, we scored the duration of the time that the chicks were protected from solar radiation by the parents. When exhibiting this behaviour, parents spread their wings above the chicks when standing near them without making any other physical contact. No rain fell during the observations. One to four chicks from different nests were observed simultaneously.

196


DLW measurements We applied the Doubly-labelled water (DLW) method to measure the daily energy expenditure in 16 free-living chicks. On capture, chicks were weighed with a Pesola spring balance (see above) and subsequently injected intraperitoneally with a doubly labeled water mixture ( 2H enrichment 32%, and 18O enrichment 63%) applying a dose (D, g) using the equation: D = 0.13 + 0.0008 M

(1)

where M represents the chick’s body mass (g). From each chick, after an equilibration period of 1h, 4–6 capillaries were filled with 10–15µl blood (henceforth referred to as “initial blood sample”) from the brachial vein. Capillaries were flame-sealed immediately. Chicks were returned to the place where they had been captured. Chicks were recaptured 23.83h later, on average (range 23.12–24.50h), to be reweighed and to collect a second blood sample (“final sample”) using the same procedure. Blood samples were also collected from four randomly chosen chicks to assess the natural abundances of 2H and 18O in their body water pools. For further details regarding the analytical procedures applied, see Visser and Schekkerman (1999). The average enrichments of the natural abundance samples were observed to be 159.1 (SD = 3.43) ppm, and 2004.1 (SD = 2.70) ppm, respectively. The amount of body water was calculated for each chick using the “plateau method”, based on the population specific background level, the dose, and the individual-specific 18O enrichment of the initial blood sample (Speakman 1997, Visser et al. 2000). Rates of CO2 production were calculated with Speakman’s equation 7.17 (1997). Validation studies on growing precocial and semi-precocial chicks have revealed this to be the most appropriate model (Klaassen et al. 1989, Visser and Schekkerman 1999, Visser et al. 2000). As the last step, daily rates of CO2 production were converted to daily energy expenditure (DEE, kJ/d) using a factor of 27.3 kJ/l based on a protein-rich diet (Gessaman and Nagy 1988). Energy budget The energy budget of growing chicks consists of two components: the amount of energy spent (as measured with the DLW method), and the amount of tissue energy accumulated. Daily energy expenditure comprises all routes of energy loss to fuel different types of behaviour (including locomotion), temperature regulation, heat increment of feeding, and the bio-synthesis related heat loss (Drent et al. 1992). Of the gross energy intake of a bird, only a fraction is metabolisable energy (MEI, kJ/d). MEI can be used for both anabolic (energy accumulated in body tissue, Etis, kJ/d) and catabolic purposes (DEE, Klaassen et al. 1989, Gabrielsen et al. 1992): MEI = DEE + Etis

(2)

The energy used for tissue growth Etis can be calculated with the growth rate of the chicks and the mass-specific energy content of the tissue. The DEE consists of the resting metabolism (RMR), the biosynthesis-related heat loss (Esyn, kJ/ d), costs of thermoregulation (Etr, kJ/d) and activity costs (Eact, kJ/d): DEE = RMR + Esyn + Etr + Eact

(3)

Figure 1: Development of maximum, average, and minimum operative temperatures as measured with a black sphere in the Kelp Gull colony

None of these components are calculated or determined separately. Nonetheless the equation provides insight in different avenues of heat loss, which affect the chicks’ energy requirements. Results Weather In Fig. 1 we display the maximum, average and minimum operative temperatures in the colony. Average maximum and minimum temperatures were 31.3 °C (SD = 3.95), and 12.8 °C (SD = 2.44), respectively. Reproductive success A total of 88 nests were found. The average laying date of the first egg of each nest was 2 November (N = 88, SD = 20.0) and the average hatching date was 21 November (N = 56, SD = 18.7). The duration of the incubation period was about 28 days. The fact that the difference between the average laying and hatching dates was less than this (19 days) is explained by the increased level of egg predation towards the end of the incubation period. The successfully developing chicks (N = 19) fledged after 38 days.

Figure 2: Development of body mass of chicks which fledged successfully (N = 19). The curve represents the Gompertz growth curve which gave the best fit: M = 1084 exp (–exp(–0.081 (t – 13.7))), where M and t represent the body mass (g), and age (d), respectively

Growth rate A total 104 chicks hatched from 65 successful nests (i.e., 1.6 chicks per nest on average). Nineteen of these chicks could repeatedly be captured until fledging, enabling us to construct individual-specific growth curves for chicks which fledged successfully (Fig. 2). Of the different models applied, the Gompertz growth curve with a fitted asymptote yielded the best fits (individual R2 values ranged between 0.98 and 1.00, N = 19). The average asymptotic body mass was found to be 1 084g (SD = 246.9g), the average growth rate constant was 0.081d–1 (SD = 0.02 d–1), and the average time to the point of inflection was 13.7d (SD = 2.7d). At day 38 the mass at fledging (Mfl) was calculated to be 970g.

Figure 3: Behavioural development of Kelp Gull chicks as a function of age


197

Energy budget In Fig. 5 the different components of the energy budget are depicted. The upper line is the MEI, which represents the sum of the DEE and tissue energy (equation 2), and equals the total daily amount of energy the chicks need to receive to successfully fledge after 38 days. The DEE component of the budget has been derived from equation 5, and the growth curve. The total gain in tissue energy (Etis, kJ/d) is calculated from the mass-specific energy density (ED, kJ/g), based on Drent et al. (1992) for Laridae: ED = 4.18 + 4.61 · (M / A)

Figure 4: The relationship between body mass and daily energy expenditure in Kelp Gull chicks. The line represents the linear regression relating log(DEE) to log(mass)

(6)

where M is the mass (g) of the chick, and A the asymptotic body mass (1 084g). The total energy content of the body at a specific time point is the product of ED and M. The daily change in the total energy content represents Etis. The highest MEI (Peak-DME, kJ/d) value was calculated to be 1 289kJ/d at 38 days. The sum of the daily MEI values (Total-MEI, kJ) between hatching and fledging was calculated to be 35 375kJ, which represents the total amount of metabolisable energy needed for successfully raising a chick. The total gain of tissue energy during growth was calculated to be 7 781kJ, which represented 22% of the TotalMEI. Thus, in Kelp Gull chicks a total of 78% of the Total-MEI was dissipated as heat. Discussion Behavioural observations

Figure 5: The energy budget for growing Kelp Gull chicks. The metabolisable energy intake is the sum of the energy used for gain in tissue energy and the daily energy expenditure, and is the daily amount of energy chicks require to fledge successfully after 38 days

Behavioural observations Total observation time of the chicks was about 80 h (Fig. 3). Parental shading was particularly observed in 1–3d chicks, which on average were shaded 55% of the time. This percentage decreased rapidly with age (Fig. 3). Chicks spent on average 42% of their time sitting in the vegetation, and only 1.1% of their time standing. Feeding time comprised on average only 1.7%. Preening behaviour showed a peak between 18 and 26 days of age, when they spent about 10% of their time engaged in this activity. This coincided with the rapid growth of feather sheaths. On average, chicks spent 92% of their time being inactive (parental shading, standing and sitting), which is consistent with the expectations for semi-precocial development. Daily energy expenditure The relationship between log DEE (kJ/d) and log body mass (M, g) is depicted in Fig. 4, and can be described by: log10 (DEE) = 0.832 · log10(M) + 0.585 (R 2 = 0.96, standard error of the slope 0.048, P < 0.01) (4) which can be rewritten as: DEE = 3.846 M0.832 198

(5)


Kelp Gull chicks were inactive for about 92% of their time. Parental shading was frequently observed in young chicks up to 10 days of age. In older chicks, this behaviour was no longer observed. From this age chicks may be capable of regulating their own body temperature at ambient temperatures above 30 °C. Chicks of both Arctic (Sterna paradisaea) and Common Terns (Sterna hirundo) were reported to be inactive for about 87% of their time (being brooded by their parents to prevent cooling of the body, and at rest; Klaassen 1994). The brooding component of the time budgets for chicks of both of these species were observed up to the age of 10 days. Growth rate and energy budget The observed Gompertz growth rate constant of 0.081/d, was 60% higher than predicted for a general bird species, but Laridae as a group are reported to exhibit high growth rate residuals (Starck and Ricklefs 1998). The Peak-DME, and Total-MEI values for the Kelp Gull were found to be 1 289kJ/d and 35 375kJ, respectively. Weathers (1992) reviewed the literature regarding the energy budgets during growth, and he derived predictive equations based on 30 mainly altricial bird species with an asymptotic body mass range of 9.7–3700g: Peak-DME = 11.69 Mfl0.9082 Tfl–0.428

(7)

where Mfl represents the mass at fledging and Tfl. The fledging period (for the Kelp Gull 970g and 38d, respectively). The estimated Peak-DME for Kelp Gull is 1 279kJ/d. Our measured value of 1 289kJ/d with is 0.8% above prediction. The Total-MEI (kJ) represents total amount of food the parents have to provide to their chicks to enable them to fledge successfully. According to Weathers (1992) this value can be predicted for a bird species by the equation:

Table 1: Energy budget parameters for larid species in relation to latitude. Sources: 1 C. Eising and G.H. Visser unpub. data; 2 Gabrielsen et al. (1992); 3 this study; 4, 5, 6, and 7 Klaassen (1994); 8 and 9: Ricklefs and White (1981). Species

Latitude(°N/S)

Fledging time (days)

Fledging mass (g)

Peak MEI (kJ/day)

Total MEI (kJ)

Rpeak %

Rtotal %

1. Black-headed gull 2. Kittiwake 3. Kelp Gull

53N 79N 34S

30 35 38

225 399 970

395 852 1 289

9 120 18 400 35 375

5.9 45.0 0.8

21.4 34.8 16.0

4. 5. 6. 7. 8. 9.

62S 53N 79N 53N 25N 25N

27 22 22 25 30 60

131 105 112 112 110 198

398 233 277 239 199 135

7 150 3 996 4 628 4 852 4 412 6 882

66.6 9.3 22.5 11.6 2.1 –45.3

62.7 27.0 39.1 33.2 8.1 –37.5

Antarctic tern Arctic tern Arctic tern Common tern Common tern Sooty tern

Total-MEI = 6.65 Mfl0.852 Tfl0.71

(8)

which yields a predicted value of 30 425kJ. Our observed value is 16% above the predicted value. We found that 22% of the Total-MEI was used for the gain in tissue energy. This is slightly below the average value of 27% reported by Drent et al. (1992), based on a review of 14 gull and tern species. To better interpret our energy budget data for the Kelp Gull, we assembled data from the literature for different larid species on Tfl, Mfl, Peak-DME, Total-MEI, and latitude (Table 1). Residuals for Peak-DME and Total-MEI (Rpeak, and Rtotal, respectively, %) were calculated as: Rpeak = 100 ·(observed Peak-DME – predicted Peak-DME)/ predicted Peak-DME (9) Rtotal = 100 ·(observed Total-MEI – predicted Total-MEI)/ predicted Total-MEI (10) where the predicted values were derived from Weathers (1992). The relationship between Rpeak and latitude L (degrees N or S) is displayed in Fig. 6, and can be described by:

Rpeak = –44.3 + 1.12 (SE = 0.40) · L

(11)

(R2 = 0.47, P = 0.024), which means that Rpeak increases by 1.12% per degree latitude. The relationship for Rtotal and latitude is also displayed in Fig. 6, and can be described by: Rtotal = –28.4 + 0.99 (SE = 0.34) · L

(12)

(R2 = 0.48, P = 0.022), which means that Rtotal increases by 0.99% per degree latitude. This confirms earlier predictions of Drent et al. (1992) and Klaassen (1994) that in larid species the costs of rearing offspring become higher at higher latitudes. As chick activity does not increase with latitude (see preceding paragraph), it is most likely that these increased costs at high latitudes are caused by elevated thermoregulary costs. Acknowledgements – This paper is a contribution to the project LMR/ EAF/03/02 of the Benguela Current Large Marine Ecosystem (BCLME) Programme. The research was conducted under permit from CapeNature. The travel expenses for T.B. have been covered by the University of Groningen Marco Polo Fund. Isotope analyses have been skillfully performed by Berthe Verstappen.

Figure 6: Relationships between latitude and the Relative Peak-DME (Rpeak , %, open symbols), and Relative Total-MEI (Rtotal , %, solid symbols) for different larid species. See Table 1 Top Predators of the Benguela System

199

Literature cited Calf, K. M., Cooper, J. and Underhill, L.G. (2003) First breeding records of Kelp Gulls Larus dominicanus vetula at Robben Island, Western Cape, South Africa. Afr. J. Mar. Sci. 25: 391–393. Crawford, R. J. M. (1997) Kelp Gull Larus dominicanus. Pp. 462– 463 in Harrison, J. A., Allan, D. G., Underhill, L. G., Herremans, M., Tree, A. J., Parker, V. & Brown, C. J. (eds) The atlas of southern African birds Vol 1. Non-passerines. BirdLife South Africa, Johannesburg. Crawford R. J. M., Cooper, J. and Shelton, P.A. (1982) Distribution, population size, breeding and conservation of the Kelp Gull in Southern Africa. Ostrich 53: 164–177. Drent R. H., Klaassen, M. and Zwaan, B. (1992) Predictive growth budgets in terns and gulls. Ardea 80: 5–17. Gabrielsen G. W., Klaassen, M. and Mehlum, F. (1992) Energetics of Black-legged Kittiwake Rissa tridactya chicks. Ardea 80: 29– 40. Gessaman, J. A. and Nagy, K. A.. (1988) Energy metabolism: errors in gas exchange conversion factors. Physiol. Zool. 61: 507–513. Hohtola, E. and Visser, G. H. (1998) Development of locomotion and endothermy in altricial and precocial birds. Pp. 157–173 in J. M. Starck and R. E. Ricklefs, eds. Avian growth and development: evolution within the altricial–precocial spectrum. Oxford University Press, Oxford. Klaassen M., Bech, C., Masman, D. and Slagsvold, G. (1989) Growth and energetics of Arctic Tern chicks (Sterna paradisaea). Auk 106: 240–248. Klaassen M., Zwaan, B., Heslenfeld, P., Lucas, P. and Luijckx, B. (1992) Growth rate associated changes in the energy requirements of tern chicks. Ardea 80: 19–28. Klaassen, M. (1994) Growth and energetics of tern chicks from temperate and polar environments. Auk 111: 525–544. Ricklefs, R. E. (1973) Patterns of growth in birds. 2. Growth rate and

200


mode of development. Ibis 115: 175–201. Ricklefs, R. E. (1974) Energetics of reproduction in birds. Pp. 152– 292 in R. A. Paynter jr. ed. Avian Energetics 15. Publications of the Nuttall Ornithology Club, Cambridge, Mass. Ricklefs, R. E. and White, S. C. (1981) Growth and energetics of the Sooty Tern (Sterna fuscata) and the Common Tern (S. hirundo). Auk 98: 361–378. Schekkerman, H. and Visser, G. H. (2001) Prefledging energy requirements in shorebirds: energetic implications of self-feeding precocial development. Auk 118: 944–957. Speakman, J. R. (1997) Doubly labelled water: theory and practice. Chapman & Hall, London. Starck, J.M. and Ricklefs, R. E. (eds) (1998) Avian growth and development: evolution within the altricial–precocial spectrum. Oxford University Press, Oxford. Underhill L. G., Whittington, P. A. and Calf, K. M. (2001) Shoreline birds of Robben Island, Western Cape, South Africa. Wader Study Group Bull. 96: 37–39. Visser, G. H. and Ricklefs, R. E. (1993) Temperature regulation in neonates of shorebirds. Auk 110: 445–457. Visser G. H. and Schekkerman, H. (1999) Validation of the double labeled water method in growing precocial birds: The importance of assumptions concerning evaporative waterloss. Physiological and Biochemical Zoology 72: 740–749. Visser, G. H., Boon, P. E. and Meijer, H. A. J. (2000) Validation of the doubly labeled water method in Japanese Quail Coturnix c. japonica chicks: is there an effect of growth rate? Journal of Comparative Physiology B, 170: 365–372. Weathers, W. W. (1992) Scaling nestling energy requirements. Ibis 134: 142–153. Weathers W. W. (1996) Energetics of postnatal growth. Pp. 461–496 in C. Carey ed. Avian energetics and nutritional ecology. Chapman and Hall, New York.

Chapter 24 Primary moult of the Kelp Gull Larus dominicanus vetula in the Western Cape, South Africa Vincent L. Ward1,2, H. Dieter Oschadleus1 and Leslie G. Underhill1* 1Avian

Demography Unit, Department of Statistical Sciences, University of Cape Town, Rondebosch 7701, South Africa Private Bag X9086, Cape Town 8000, South Africa *Corresponding author: e-mail: [email protected] 2CapeNature,

We transformed the primary moult protocols of 147 adult Kelp Gulls to Proportion Feather Mass Grown (PFMG), and used the Underhill–Zucchini moult model to estimate the parameters of moult: mean starting date, 29 January; duration 168 days; mean completion date 16 July, with 95% of birds estimated to start and complete moult within 43 days of these dates. The average number

of feathers moulting simultaneously was 1.9. The paper introduces the concept of Proportion Feather Mass Missing (PFMM); although there was no trend for the number of moulting feathers to change through moult, the PFMM increased, because large gaps in the primaries occurred towards the completion of moult.

Keywords: Kelp Gull, Larus dominicanus vetula, primary moult, South Africa, Underhill–Zucchini moult model

Introduction The nominate race of the Kelp Gull Larus dominicanus has a wide distribution in the southern hemisphere, including southern South America, the Antarctic Peninsula, the subantarctic islands of the Southern Ocean, Australia and New Zealand. In contrast, the subspecies vetula occurs only in southern Africa. Several aspects of the biology of L. d. vetula have been studied, including its distribution (Crawford 1997), population size and conservation (Crawford et al. 1982, Steele & Hockey 1990), movements (Underhill et al. 1999), survival (Altwegg et al. in press), breeding biology (Williams et al. 1984), chick growth and energetics (Visser et al. in prep.) and plumage development and age at first breeding (Crawford et al. 2000). In this study we focus on an aspect that has received little attention as yet, namely the moult. We estimated the primary moult parameters for L. d. vetula, using records of moult and an appropriate model for avian primary moult. Material and Methods We collected primary moult protocols for 142 adult Kelp Gulls, from our own observations (mainly during bird ringing operations), moult cards curated by SAFRING, and moult protocols submitted to the electronic database of SAFRING. The records were distributed throughout the year and consisted of actively moulting birds as well as birds which were not in moult. The data therefore correspond with data type 2 of Underhill & Zucchini (1988), and the moult parameters were estimated using the model developed by these authors. A lack of records prevented us from analyzing the moult protocols of immature birds. The Underhill–Zucchini moult model (Underhill & Zucchini

1988) requires an index of primary moult that increases linearly through time. For many species, this can be achieved, to a good approximation, by converting the moult protocols into Proportion Feather Mass Grown (PFMG), using transformations devised by Summers et al. (1980, 1983). The conversion requires that the ratios between the mass of each primary feather and the combined mass of the primary feathers (relative feather masses) are known. Underhill & Joubert (1995) demonstrated that small samples are adequate for estimating relative feather mass of primary feathers because there is little intra-specific variation. We therefore processed and weighed complete sets of primary feathers taken from two freshly dead adult Kelp Gulls which had moulted shortly before death. One of the specimens was collected at Robben Island (33°47'S, 18°21'E), the other at Dyer Island (34°40'S, 19°25'E). Both had died of natural causes. The extracted feathers were cleaned and dried in an oven at 60°C for 48 hours. Each feather, from the inner primary (P1) to the outer primary (P10), of each wing was then weighed individually on an Ohaus GA200D balance (precision: 0.0001 g). The average feather mass of each primary (estimated across the four wings of the two specimens) were used to calculate PFMG for moult protocols, following Underhill & Summers (1993) and Underhill & Joubert (1995). We also estimated the Proportion Feather Mass Missing (PFMM) using the obvious approach based on the calculation of PFMG; for feathers with moult scores 1 (missing or pin), 2 (small brush), 3 (about half grown) and 4 (about three-quarters grown), we computed 0.875, 0.625, 0.375 and 0.125 of the relative feather masses, respectively, and added these values. PFMM provides a measure for the size of the “gap” in the primary feathers which takes into account the relative sizes of the missing feathers.


201

Figure 1: Proportion Feather Mass Grown (PFMG) plotted against date for 142 adult Kelp Gulls in South Africa. The timing of primary moult of the average bird is shown by the solid line, and 95% confidence limits by the dashed lines (Table 1)

Results Relative feather mass ranged from 4.2% of total primary mass for P1 to 16.5% for P10 (Table 1). Thus P10, the outer primary was nearly four times as heavy as P1, the innermost. Our sample of adult moult protocols included 40 birds prior to moult (all primaries old), 79 birds in moult and 23 birds which had completed moult (all primaries new) (Fig. 1). The Underhill & Zucchini (1988) moult model identified four birds as outliers: moult protocols 5555551000 on 31 January, 5555555300 on 23 February, 5521000000 on 20 May and 5531000000 on 22 May (Fig. 1, the four markers outside the 95% confidence intervals); the first two of these were unusually early and the second two unusually late. These four records were removed and the moult parameters re-estimated. The duration of moult was estimated to be 168 days or 5.6 months, with 29 January the mean starting date and 16 July the mean date of completion (Table 2). The standard deviation of the mean starting date was 22 days, so that the estimated period during which 95% of birds started to moult was 43 days (= 1.96 × 22) on either side of 29 January (Fig. 1, Table 2). Of the birds in moult, the majority (60%) were moulting two primaries simultaneously; 25% were moulting one primary and 15% were moulting three. The average number of feathers in moult at one time was 1.9. Birds moulting three primaries were scattered across all stages of moult; there was no correlation between the number of feathers moulting and PFMG for actively moulting birds (r71= –0.056, P = 0.64). On the other hand, the percentage feather mass missing

(PFMM) was positively correlated with PFMG (r71 = 0.35, P = 0.003; linear regression PFMM = 0.0695 + 0.0223 × PFMG) (Fig. 2). There was a large scatter of values for PFMM towards the end of moult. The largest value for PFMM was 26% for a bird growing its three outer primaries with protocol 5555555421 on 23 April. Discussion The Kelp Gull is the third larid for which relative primary feather masses have been determined (Table 1). The results were very similar to those for the Hartlaub’s Gull Larus hartlaubii (Crawford & Underhill 2003); but differed slightly from the Grey-headed Gull L. cirrocephalus which has a slightly more pointed wing shape, as measured by the P10 to P1 ratio (McInnes 2006) (Table 1). The relative primary feather mass of gulls are similar to those of waders (Charadrii), for which P1 and P10 average 4.1% and 17.4% of total primary feather mass, respectively. However, the wing shapes of gulls are not as pointed as those of the Common Tern (Sterna hirundo) and Arctic Tern (S. paradisea), for which P1 and P10 comprise 2.8% and 21.3% of primary feather mass, respectively; a ratio of 7.6, compared with c. 4 for the waders and gulls (Underhill & Summers 1993; Underhill & Joubert 1995, Table 1). Underhill & Joubert (1995) suggested that species with large P10 to P1 ratios were likely to show extensive seasonal movements. The ratios of the three common gull species in southern Africa are 3.9 for the Kelp and Hartlaub’s Gulls, and 4.4 for the Greyheaded Gull, which is by far the most nomadic of the three

Table 1: Mass (mg) of each primary feathers (P) of both wings of two adult Kelp Gulls (KG) from the Western Cape, South Africa. Also given are the estimates of relative feather mass (%) of each primary, and comparative estimates for Hartlaub’s Gulls (HG) and Grey-headed Gulls (GHG) P1

P2

P3

P4

P5

P6

P7

P8

P9

P10

P Total

Right 1 (g) Left 1 (g) Right 2 (g) Left 2 (g)

247 230 245 246

293 276 295 293

343 324 342 343

414 410 414 415

520 516 521 522

618 582 617 616

697 677 695 695

820 720 819 821

905 899 903 904

951 942 952 953

5808 5576 5803 5808

KG (%) HG (%)1 GHG (%) 2

4.2 4.1 3.8

5.0 4.9 4.7

5.9 5.8 5.7

7.2 7.2 7.0

9.0 9.1 9.0

10.6 10.6 10.7

12.0 12.3 12.1

13.8 14.2 14.3

15.7 15.5 15.8

16.5 16.2 16.8

1 2

(Crawford & Underhill 2003) (McInnes 2006)

202


Figure 2: Proportion Feather Mass Missing (PFMM) (see text) plotted against Proportion Feather Mass Grown (PFMG) for moulting adult Kelp Gulls in South Africa. The pattern of lines is a consequence of the way in which moult protocols are recorded

species (Underhill et al. 1999). Egg-laying in the subspecies vetula in Southern Africa is from October to December and most checks fledge in December and January, with stragglers fledging until early March (Williams et al. 1984). The wide 95% confidence intervals surrounding the estimated mean start date (86 days) indicates that primary moult was not synchronized (Table 2). This lack of synchronization of moult probably reflects the lack of synchronization of breeding. Primary moult started with the smaller inner primaries, which, being small, cost less energy to replace than the outer larger primaries. Because the number of primaries moulted at any one time was consistent throughout the moult, the gap in the primaries (measured as PFMM) was smaller at the start of moult than towards the end. Because the outermost primaries are large (P8–P10 account for 46% of the total primary feather mass, Table 1), it is inevitable that PFMM is larger towards the end of moult than the beginning (Fig. 2). This would be true even if fewer feathers were moulted simultaneously towards the end of moult. This is the first species for which PFMM has been computed and plotted, and more examples are needed to evaluate the insights provided by this concept. Kinsky (1963) described in qualitative detail the primary moult of the nominate subspecies of Kelp Gull in New Zealand; his results are outlined in Higgins & Davies (1996), suggesting that little new information accumulated over four decades. The description was based largely on small samples of museum specimens. In Kinsky’s paper, it is difficult

to distinguish results concerning primary moult from the moult of other feather tracts; he explicitly suggested that primary moult lasts on average four months. But his sample sizes were small (12 adult birds between February and June, the main period of primary moult in New Zealand). Qualitatively, he stated that primary moult was first noted in January, but that the majority of individuals commenced the moult February and early March. He stated that primary moult “is completed by July”. Curiously, he noted that “Wellington birds” did not complete primary moult “before mid-August, and often somewhat later.” His sample sizes in July, August and September were 17 (all in moult), 28 (86% in moult) and 17 (41% in moult), respectively. Presumably, the large samples for these months were mostly from Wellington. Kinsky (1963) noted, in agreement with Dwight (1901), that the “full moult” is completed at the same time as the completion of primary moult. It is possible that he was trying to work as closely as possible to the moult pattern for large southern hemisphere gulls devised by Dwight (1925), the authority on the topic, who considered that primary moult lasted 2.5 months. However, if the Wellington birds are typical, rather than exceptional, then the timing of primary moult of Kelp Gulls in New Zealand appears to be closely similar, both in timing and duration, as primary moult in South Africa. The only other gull species for which moult parameters have been estimated using the Underhill & Zucchini (1988) moult model are Hartlaub’s Gull (Crawford & Underhill 2003) and Grey-headed Gull, for which moult duration were 115 days and 136 days respectively. In order to make a com-

Table 2: Estimates of primary moult parameters for adult Kelp Gulls (KG) in the Western Cape, South Africa. Comparative estimates are given for Hartlaub’s Gulls (HG) and Grey-headed Gulls (GHG). Standard deviations (days) are given in parentheses, CI denotes confidence intervals Species KG HG1 GHG2 1 2

Mean starting date

Duration

Mean completion date

Std Dev

Starting date 95% CIs

29 Jan (4) 11 Oct (6) 12 Oct (4)

168 (8) 115 (10) 136 (9)

16 Jul (6) 3 Feb (6) 25 Feb (8)

22 (2) 32 (3) 25 (2)

17 Dec–13 Mar 9 Aug–7 Dec 24 Aug–30 Nov

Completion date 95% CIs 3 Jun–28 Aug 2 Dec–7 Apr 7 Feb–15 May

(Crawford & Underhill 2003) (McInnes 2006)


203

prehensive comparison of moult parameters for gulls, a larger sample of gull species, including species from the northern hemisphere, is necessary. For other gull species, moult parameters have been estimated by a variety of methods, both formal and informal, and only qualitative comparisons are possible. The only available data for another species of large gull in the southern hemisphere appears to be for the Pacific Gull L. pacificus in Australia; primary moult of adults is said to span four months within the period January to July (Higgins & Davies 1996). In the northern hemisphere, the moult duration of large gulls is described as c. 4.5 months for Common Gull L. canus, between four and six months for Herring Gull L. argentatus, 6.7 months for Glaucous Gull L. hyperboreus, and 6.2 months for Great Black-backed Gull L. marinus (Ginn & Melville 1983). Most of these moult duration estimates are comparable with the 5.6 month period for the Kelp Gull estimated here. Further investigation of gull moult, using the Underhill–Zucchini (1988) moult model, would demonstrate whether there is a latitudinal patterning to duration of moult in large gulls. Acknowledgements – Besides the authors, most of the moult data were collected by D.M. Harebottle and staff of Marine and Coastal Management. R.K. Brooke and W.K. Steele compiled most of the moult cards. A.J. Williams commented on a draft. Lauren Waller (CapeNature) supplied the primary feathers of a Kelp Gull from Dyer Island. This paper is a contribution to the project LMR/EAF/03/02 of the Benguela Current Large Marine Ecosystem (BCLME) Programme.

References Altwegg R., Crawford R. J. M., Underhill L. G., Martin A. P., Whittington P. A. in press. Geographic variation in reproduction and survival of Kelp Gulls Larus dominicanus vetula in southern Africa. Oikos. Crawford R. J. M. 1997. Kelp Gull Larus dominicanus. In: Harrison J. A., Allan D. G., Underhill L. G., Herremans M., Tree A. J., Parker V., Brown C. J. (eds). The atlas of southern African birds Vol 1. Non-passerines. BirdLife South Africa, Johannesburg, pp. 462–463. Crawford R. J. M., Cooper J., Shelton P. A. 1982. Distribution, population size, breeding and conservation of the Kelp Gull in southern Africa. Ostrich 53: 164–177.

204


Crawford R. J. M., Dyer B. M., Upfold L. 2000. Age at first breeding and change of plumage of Kelp Gulls Larus dominicanus in South Africa. South African Journal of Marine Science 22: 27– 32. Crawford R. J. M., Underhill L. G. 2003. Aspects of breeding, population trend, measurements and moult of Hartlaub’s Gull (Larus hartlaubii) in Western Cape, South Africa. Waterbirds 26: 139– 149. Dwight J. 1901 The sequence of moults and plumages of the Laridae (gulls and terns). Auk 18: 49–63. Dwight J. 1925. The gulls (Laridae) of the world: their plumages, moults, variations, relationships and distributions. Bulletin of the American Museum of Natural History 52: 63–336. Ginn H. B., Melville D. S. 1983. Moult in birds. BTO Guide 19. British Trust for Ornithology, Tring. Higgins P. I., Davies S. J. J. F. 1996. Handbook of Australian, New Zealand and Antarctic birds. Vol. 3. Snipes to Pigeons. Oxford University Press, Melbourne. Kinsky F. C. 1963. The Southern Black-backed Gull (Larus dominicanus) Lichtenstein: measurements, plumage colour, and moult cycle. Records of the Dominion Museum. 12: 149–219. McInnes A. M. 2006. Biology of the Grey-headed Gull Larus cirrocephalus in South Africa. Unpubl. MSc thesis, University of KwaZulu-Natal. Steele W. K., Hockey P. A. R. 1990. Population size, distribution and dispersal of Kelp Gulls in the southwestern Cape, South Africa. Ostrich 61: 97–106. Summers R. W., Swann R. L., Nicoll M. 1980. Unbending moult data. Wader Study Group Bulletin 30: 12–13. Summers R. W., Swann R. L., Nicoll M. 1983. The effects of methods on estimates of primary moult duration in the Redshank Tringa totanus. Bird Study 30: 149–156. Underhill L. G., Joubert A. 1995. Relative masses of primary feathers. Ringing & Migration 16: 109–116. Underhill L. G., Summers R. W. 1993. Relative masses of primary feathers of waders. Wader Study Group Bulletin 71: 29–31. Underhill L. G., Tree A. J., Oschadleus H. D., Parker V. 1999. Review of ring recoveries of waterbirds in southern Africa. Avian Demography Unit, Cape Town. Underhill L. G., Zucchini W. 1988. A model for avian primary moult. Ibis 130: 358–372. Visser G. H., Bakker T., Tjørve K. M. C., Underhill L. G. in prep. Prefledging energetics of Kelp Gull Larus dominicanus vetula chicks in a warm environment. Williams A. J., Cooper J., Hockey P. A. R. 1984. Aspects of the breeding biology of the Kelp Gull at Marion Island and in South Africa. Ostrich 55: 147–157.

Other seabirds

Chapter 25 Population estimates and trends of seabird species breeding in Namibia J. Kemper1,2 1 Avian

Demography Unit, Department of Statistical Sciences, University of Cape Town, Rondebosch 7701, South Africa Penguin Conservation Project, PO Box 583, Lüderitz, Namibia [email protected]; [email protected]

2 African

Introduction Of the sixteen seabird species (including the African Black Oystercatcher Haematopus moquini ) breeding in the Benguela Ecosystem, fourteen breed in Namibia, mostly on islands and rocks along the Namibian coast but also on coastal cliffs, in dune fields, salt pans and estuaries and at localities further inland. Eight seabird species and two subspecies breeding in Namibia are endemic to the Benguela Ecosystem. Coastal seabirds in Namibia face a number of threats mainly because of human activity and its consequences. Although many of these species breed at relatively protected sites such as islands, away from the direct effects of human development, they are not immune to these pressures and a number of them are in serious need of better conservation measures. A number of seabirds in Namibia are considered threatened, following major population declines in recent decades (du Toit et al. 2003). Other species appear to be benefiting from an increased food supply near towns (e.g. Kelp Gulls Larus dominicanus vetula) or through the spread of an alien invasive mussel species (Haemotopus moquini) (Hockey et al. 2005). The main risks facing seabirds in Namibia include: a lack of prey availability, because of competition with commercial fisheries, other seabirds and seals (e.g. Crawford et al. 2001, 2006) past and current habitat alteration and loss, for example from guano scraping on islands, displacement by other species, diamond mining on land (breeding habitat) and at sea (feeding habitat), housing and harbour developments (e.g. Crawford et al. 1989, Braby et al. 2001, Simmons 2005, Kemper 2006)

other pollution; this includes entanglement in fishing gear, particularly during demersal trawls or longline fishing activities (e.g. Ryan and Boix-Hinzen 1998), but also from discarded fishing tackle at recreational beaches (JK pers. obs.), entanglement in lobster traps (Cooper 1985) and in aquaculture structures (MFMR unpubl.data) predation by other birds, by seals and by land-based predators such as Brown Hyena Parahyaena brunnea and Black-backed Jackal Canis mesomelas (e.g. Marks et al. 1997, Simmons and Kemper 2003) displacement of breeding colonies by other birds or seals (e.g. Shaughnessy 1984, Crawford et al. 1989) Climate change, including any changes in upwelling intensity, are likely to exacerbate threats faced by seabirds in Namibia, particularly with regard to prey and breeding habitat availability (Roux 2003, Simmons et al. 2004). This report summarises latest population size estimates for the fourteen seabird species breeding in Namibia and provides recent population trends (where feasible). Census information mainly stems from data obtained from island staff of the seabird section of the Ministry of Fisheries and Marine Resources, and was augmented with information from other sources, e.g. from counts done by staff of the Ministry of Environment and Tourism, independent individuals and from counts reported in the literature. Population trends were calculated from estimates of the number of breeding pairs based on nest counts (and, in the case of the African Penguin, from counts of moulting individuals in adult plumage) using exponential curves fitted by least-squares regression. Population estimates and trends of seabirds breeding in Namibia 1. African Penguin Spheniscus demersus

human disturbance, from guano scraping, uncontrolled tourism or recreation (e.g. Braby et al. 2001, Crawford et al. 2006) small-scale chronic oil pollution from ships discharging waste oil and sunken wrecks leaking oil. This mainly affects flightless African Penguins. An oil spill between Mercury and Ichaboe Islands would put 80% of the Namibian penguin population at risk (Kemper 2006). Fish oil pollution from factories and fishing fleets mainly affect Cape Gannets and gulls (du Toit and Bartlett 2001)

African Penguins breed on ten islands between Hollamsbird Island and Sinclair Island as well as on two mainland sites in Namibia. In addition, one breeding attempt was noted at Penguin Island in January 2006, where breeding had ceased more than a century ago (Kemper 2006). Between 1996 and 2006, the number of individuals in adult plumage decreased at a rate of 3.1% per year from c. 30 000 individuals in 1996 to c. 21 000 individuals in 2006. Between 2004 and 2006, nearly 4 000 individuals were lost from the Namibian population; emigration by these individuTop Predators of the Benguela System

207

als to South African localities is unlikely, because movement between the two regions has been shown to be infrequent (Whittington 2005a, b). The breeding population, estimated from counts of active nests, has decreased from 7 580 pairs in 1992 to 3 288 pairs in 2006 at a rate of 4.0% per year. 2. Great White Pelican Pelecanus onocrotalus Approximately 75 000 pairs of Great White Pelicans breed in Africa. In Namibia the species breeds at the artificial Bird Rock guano platform, in Etosha National Park, at Lake Oponomo and at the Hardap dam. Great White Pelicans may not breed if conditions are unsuitable. The number of breeding pairs at Bird Rock and Hardap dam, where they tend to breed annually, totals c. 500 pairs, although up to 1 376 pairs have been recorded at Hardap dam (van Zyl et al. 1995). Recent breeding population estimates from Etosha National Park are lacking, but numbers of individuals average 622, with a maximum of 3 000 individuals (Williams and Borello 1997). Population numbers appear to be stable in Namibia. 3. Cape Gannet Morus capensis The Cape Gannet breeds at six islands globally, of which three are located in Namibia. In 1956, the population in Namibia was estimated at nearly 204 000 breeding pairs. By 1978 the population had decreased to 80 000 pairs and by 2004 to 10 500 pairs (Crawford et al. 2007a). This is an annual population decline of 6.3% since 1956, 7.4% since 1978, and 10.5% since 1992. The decline is mainly attributable to losses from Ichaboe Island. The breeding population at Mercury Island has remained relatively stable over the last decade. The colony at Possession Island numbered c. 200 pairs in 2006 and is likely to become extinct in the near future. 4. Cape Cormorant Phalacrocorax capensis This cormorant breeds at numerous sites in Namibia. It is particularly sensitive to fluctuating environmental conditions and may not breed in some years or may abandon breeding activities. Estimates of numbers of breeding pairs therefore tend to fluctuate. The breeding population increased in Namibia with the erection of artificial breeding platforms (Crawford et al. 2007b). At most localities, estimates stem from aerial censuses as ground counts of large colonies tend to be inaccurate. The largest number of breeding pairs since comprehensive censuses were first done in 1956 was recorded in 1978/79 (143 000 pairs). Numbers have remained stable during the last decade with an estimated 60 000 to 70 000 pairs breeding in Namibia.

6. Crowned Cormorant Phalacrocorax coronatus In Namibia, Crowned Cormorants breed at 16 localities between Bird Rock guano platform to Sinclair Island. The erection of the Walvis Bay guano platform and the subsequent use of its supports has extended its breeding range (Crawford et al. 1982a). A relatively poorly defined breeding season, as well as low colony fidelity (and probably low locality fidelity) make it difficult to obtain accurate population size estimates. Between 1977 and 1981 an estimated 977 pairs bred in Namibia (Crawford et al. 1982a). In 2005/2006 there were roughly 1 010 pairs, with a maximum of 1 070 pairs recorded during 2000/2001. Between 1996/1997 and 2005/2006, the population increased at 2.8% per year. The population in Namibia appears to be stable or slightly increasing. 7. Whitebreasted Cormorant Phalacrocorax carbo Whitebreasted Cormorants breed at a number of localities along Namibia’s coast from Moewe Bay to the Orange River mouth, as well as at inland localities. The breeding population was estimated to number 1 400 pairs in Namibia between 1977 and 1981, with the majority of birds breeding at Bird Rock guano platform (Crawford et al. 1991). Recent counts from Hottentot’s Bay, Ichaboe Island and Penguin Island estimate a total of 77 pairs at these three localities, but no recent counts are available from localities north of Walvis Bay, where the majority of the Namibian coastal population breeds. It is therefore not possible to evaluate trends for the Namibian population of Whitebreasted Cormorants. 8. African Black Oystercatcher Haematopus moquini Between 1996 and 2006, the number of individuals occurring along the coast of Namibia averaged 1 726 individuals between Walvis Bay Lagoon / Saltworks and Sinclair Island. Counts may include subadults. This estimate is likely to be an underestimate as a number of sites which could potentially support individuals, such as stretches of rocky shore, could not be counted because of their inaccessibility. An estimate of 323 breeding pairs (based on nest counts) is also likely to be too low, because of the difficulty of finding nests. The population of African Black Oystercatchers was estimated at 1 200 individuals in the early 1980s (Hockey 1983). Unless the population in the 1980s were also underestimated, the apparent population increase since then is likely to be related to an increased availability of the alien invasive Mediterranean mussel Mytilus galloprovincialis throughout the species’ range (Hockey and Van Erkom Schurink 1992). 9. Kelp Gull Larus dominicanus vetula

5. Bank Cormorant Phalacrocorax neglectus Of the three cormorants species endemic to the Benguela Ecosystem, this species is the most threatened. In Namibia, Bank Cormorants breed at 17 localities (Roux and Kemper in press). In 1978, the breeding population in Namibia was estimated to consist of 7 144 pairs. The most recent estimate of the Namibian breeding population is 2 196 breeding pairs in 2006, comprising 87% of the global population. Mercury and Ichaboe Islands support more than 80% of the Namibian breeding population. Between 1993 and 2006 the Namibian breeding population is estimated to have declined by 65%. This loss is mainly attributable to the population crash at Ichaboe Island after 1994/95. Since then, numbers at Ichaboe Island have continued to decline and although numbers at Mercury Island are increasing, the population has not shown any recovery to pre-1993 levels. 208


Kelp Gulls breed at a number of localities along the entire coast of Namibia. The number of breeding pairs has almost doubled from 2 300 pairs in 1978/1979 (Crawford et al. 1982b) to approximately 4 000 pairs in 2006. This trend is mainly due to an almost seven-fold increase in the numbers of breeding pairs at Possession Island from 297 pairs in 1978/1979 to 1 992 pairs in 2005/2006. An additional 1 800 pairs breed at the three main islands near Lüderitz (i.e. Seal, Penguin and Halifax Islands). Between 1999/2000 and 2006/ 2007, the population of Kelp Gulls has increased at a rate of 2.1% per year. 10. Hartlaub’s Gull Larus hartlaubii In Namibia, Hartlaub’s Gulls breed between the Swakopmund Saltworks and Possession Island, although it is pos-

sible that the breeding range includes Plumpudding and / or Sinclair Islands. Breeding activities at these two localities need to be verified. The breeding population from 1976 to 1983 numbered between 1 260 and 1 425 pairs (Duffy et al. 1987); from 1984 to 1989 it numbered between 1 200 and 1 400 pairs (Williams et al. 1990). Breeding in Hartlaub’s Gulls tends to be opportunistic and therefore is not strongly seasonal (JK pers. obs). Locality fidelity is poor, e.g. at Shark Island, where the development of Lüderitz harbour led to the extinction of the Shark Island colony. It is likely that these birds now breed at Penguin or Seal Island, but annual counts at these two harbour islands take place when Hartlaub’s Gulls do not breed. In addition, recent breeding population estimates for localities north of Mercury Island are lacking. Between 2000 and 2006 the Namibian breeding population numbered between 420 and 600 pairs; this is likely to be an underestimate. 11. Grey-headed Gull Larus cirrocephalus This species is widely distributed throughout southern Africa. Relatively few pairs breed along the coast of Namibia. Scant information on numbers breeding along the coast in Namibia exist for the period 1973 to 1984, from the areas around Walvis Bay and Swakopmund (with two breeding pairs reported for Shark Island in 1977) (du Toit et al. 2003). Counts of breeding pairs are lacking since 1984. It is therefore not possible to provide estimates for the coastal population in Namibia. 12. Caspian Tern Sterna caspia The southern African population of this widely distributed species numbers between 1 000 to 1 500 individuals (Hockey et al. 2005). The breeding population in Namibia is limited to c. 40 pairs, which breed at the Swakopmund Salt Works, the Walvisbay Sewerage Works and Sandwich Harbour (Cooper et al. 1992). No recent counts of breeding pairs in Namibia are available. 13. Swift Tern Sterna bergii bergii In Namibia, Swift Terns mainly breeds around Lüderitz, from Ichaboe Island in the north to Possession Island in the south. The number of pairs breeding as well as the breeding localities vary between years. The Swift Tern colony on Shark Island supported up to 800 breeding pairs in 1986, the largest number was reported from Possession Island, where 1 476 pairs bred in 1989 (Cooper et al. 1990). Swift Terns stopped breeding at Shark Island after 2000, when harbour and housing developments (accompanied by an influx of feral cats and dogs to the area) adjacent to their breeding site caused excessive disturbance. It is likely that breeding activities moved to nearby Penguin or Seal Islands, but no counts are done there during the Swift Tern breeding season. In May 2007, roughly 1 000 pairs of Swift Terns bred at Halifax Island, ca. 10 km west of Shark Island. Whether these are the same birds which previously bred at Shark Island is unclear. An additional 250 pairs bred at Ichaboe and Possession Islands during May 2007. 14. Damara Tern Sterna balaenarum The Namibian population of Damara Terns numbers c. 800 breeding pairs, comprising 87% percent of the global breeding population of the species. The population is presumed to be stable. Breeding occurs along the coast in suitable habitat between the northern Skeleton coast and Elizabeth Bay as well as on Possession Island. The highest density of

breeding pairs occurs along the central Namibian coast. The status of Damara Terns in Namibia has recently been revised from “Endangered” to “Near Threatened” (Hockey et al. 2005). Acknowledgements – Thanks are due to staff from he seabird section of the Ministry of Fisheries and Marine Resources, Namibia, the Department of Environmental Affairs and Tourism, Marine and Coastal Management, South Africa, and the Ministry of Environment and Tourism, Namibia, particularly R. Simmons, H. Kolberg, and R. Braby for making counts available.

References Braby, R.J., Shapira, A. and Simmons, R.E. 2001. Successful conservation measures and new breeding records for Damara terns Sterna balaenarum in Namibia. Marine Ornithology 29: 81–84. Cooper, J. 1985. Biology of the Bank Cormorant, Part 3: Foraging behaviour. Ostrich 56: 86–95. Cooper, J., Crawford, R.J.M., Suter, W. and Williams, A.J. 1990. Distribution, population size and conservation of the Swift Tern Sterna bergii in southern Africa. Ostrich 61: 56–65. Cooper, J., Brooke, R.K., Cyprus, D.P., Martin, A.P., Taylor, R.H. and Williams, A.J. 1992. Distribution, population size and conservation of the Caspian Tern Sterna caspia in southern Africa. Cormorant 5: 15–16. Crawford, R.J.M., Shelton, P.A., Brooke, R.K. and Cooper, J. 1982a. Taxonomy, distribution, population size and conservation of the Crowned Cormorant Phalacrocorax coronatus. Gerfaut 72: 3–30. Crawford, R.J.M., Cooper, J. and Shelton, P.A. 1982b. Distribution, population size, breeding and conservation of the Kelp Gull in southern Africa. Ostrich 53: 164–177. Crawford, R.J.M., David, J.H.M., Williams, A.J. and Dyer, B.M. 1989. Competition for space: recolonising seals displace endangered, endemic seabirds off Namibia. Biological Conservation 48: 59– 72. Crawford, R.J.M., Ryan, P.G. and Williams, A.J. 1991. Seabird consumption and production in the Benguela and western Agulhas ecosystems. South African Journal of Marine Science 11: 357– 375. Crawford, R.J.M., David, J.H.M., Shannon, L.J., Kemper, J., Klages, N.T.W., Roux, J–P., Underhill, L.G., Ward, V.L., Williams, A.J. and Wolfaardt, A.C. 2001. African Penguins as predators and prey – coping (or not) with change. South African Journal of Marine Science 23: 435–447. Crawford, R.J.M., Barham, P.J., Underhill, L.G., Shannon, L.J., Coetzee, J.C., Dyer, B.M., Leshoro, T.M. and Upfold, L. 2006. The influence of food availability on breeding success of African Penguins Spheniscus demersus at Robben Island, South Africa. Biological Conservation 132: 119–125. Crawford, R.J.M., Dyer, B.M., Kemper, J., Simmons, R.E., Upfold, L. and vaz Velho F. 2007a. Trends in numbers of Cape Cormorants (Phalacrocorax capensis) over a 50-year period, 1956/57– 2006/07. In S.P. Kirkman (ed.) Final Report of the BCLME (Benguela Current Large Marine Ecosystem) Project on Top Predators as Biological Indicators of Ecosystem Change in the BCLME. Avian Demography Unit, Cape Town. Crawford, R.J.M., Dundee, B.L., Dyer, B.M., Klages, N.T.W., Meÿer, M.A. and Upfold, L. 2007b. Trends in numbers of Cape Gannets (Morus capensis), 1956/1957–2005/2006, with a consideration of the influence of food and other factors. ICES Journal of Marine Science: 64: 169–177. Duffy, D.C., Siegfried, W.R. and Jackson, S. 1987. Seabirds as consumers in the southern Benguela region. South African Journal of Marine Science 17: 771–790. du Toit, M. and Bartlett, P.A. 2001. ‘Soaked’ Cape Gannets at Ichaboe Island, Namibia. Bird Numbers 10(2): 8–9. du Toit, M., Boere, G.C., Cooper, J., de Villiers, M.S., Kemper, J., Lenten, B., Simmons, R.E., Underhill, L.G., Whittington, P.A. and Byers, O. (eds). 2003. Conservation assessment and management plan for southern African coastal birds. Avian Demography Unit, Cape Town and IUCN/SSC Conservation Breeding Specialist Group, Apple Valley, MN. Hockey, P.A.R. 1983. The distribution, population size, movements and conservation of the African Black Oystercatcher Haematopus


209

moquini. Biological Conservation 25: 233–262. Hockey, P.A.R. and Van Erkom Schurink, C. 1992. The invasive biology of the mussel Mytilus galloprovincialis in southern Africa. Transactions of the Royal Society of South Africa 48: 123–139. Hockey, P.A.R., Dean, W.R.J., and Ryan, P.G. (eds) 2005. Roberts Birds of Southern Africa, 7th edn. John Voelcker Bird Book Fund, Cape Town. Kemper, J. 2006. Heading towards extinction? Demography of the African Penguin in Namibia. PhD thesis, University of Cape Town, South Africa. Marks, M.A., Brooke, R.K., and Gildenhuys, A.M. 1997. Cape Fur Seal Arctocephalus pusillus predation on Cape Cormorants Phalacrocorax capensis and other birds at Dyer Island, South Africa. Marine Ornithology 25: 9–12. Roux, J-P. 2003. Risks. In: Molloy, F. and Reinikainen, T. (eds), Namibia’s marine environment. Directorate of Environmental Affairs of the Ministry of Environment and Tourism, Windhoek, Namibia. pp. 137–152. Roux, J-P. & Kemper, J. in press. Bank Cormorant. In: Simmons, R.E. & Brown, C.J. (eds), Birds to watch in Namibia: red, rare and peripheral species. National Biodiversity Programme, Windhoek, Namibia. Ryan, P.G. and Boix-Hinzen, C. 1998. Tuna longline Fisheries off Southern Africa: the need to limit seabird bycatch. South African Journal of Science 94: 179–182. Shaughnessy, P.D. 1984. Historical population levels of seals and seabirds on islands off southern Africa, with special reference to Seal Island, False Bay. Sea Fisheries Research Institute Investigational Report 127: 1–61. Simmons, R.E. 2005. Declining coastal avifauna at a diamond-

210


mining site in Namibia: comparisons and causes. Ostrich 76: 97– 103. Simmons, R.E. and Kemper, J. 2003. Cave breeding by African Penguins near the northern extreme of their range: Sylvia Hill, Namibia. Ostrich 74: 217–221. Simmons, R.E., Barnard, P., Dean, W.R.J., Midgley, G.F., Thuiller, W. and Hughes, G. 2004. Climate change and birds: perspectives and prospects from southern Africa. Ostrich 76: 295–308. Van Zyl, B.J., Hay, C.J. and Steyn, G.J. 1995. Some aspects of the reproductive biology of Labeo capensis (Smith, 1941) (Pices, Cyprinidae) in relation to exploitation and extreme environmental conditions in Hardap Dam, Namibia. South African Journal of Aquatic Science 21: 88. Whittington, P.A., Randall, R.M., Randall, B.M., Wolfaardt, A.C., Crawford, R.J.M., Klages, N.T.W., Bartlett, P.A., Chesselet, Y.J. and Jones, R. 2005a. Patterns of movements of the African Penguin in South Africa and Namibia. African Journal of Marine Science 27: 215–229. Whittington, P.A., Randall, R.M., Crawford, R.J.M., Wolfaardt, A.C., Klages, N.T.W., Randall, B.M., Bartlett, P.A., Chesselet, Y.J. and Jones, R. 2005b. Patterns of immigration to and emigration from breeding colonies by African Penguins. African Journal of Marine Science 27: 205–213. Williams, A.J. and Borello, W.D. 1997. White Pelican. In: Harrison, J.A., Allan, D.J., Underhill, L.G., Herremans, M., Tree, A.J., Parker, V. and Brown, C.J. (eds.). The Atlas of Southern African Birds. Vol. 1: 24–25. BirdLife South Africa, Johannesburg. Williams, A.J., Steele, W.K., Cooper, J. and Crawford, R.J.M. 1990. Distribution, population size and conservation of Hartlaub’s Gull Larus hartlaubii. Ostrich 61: 66–75.

Chapter 26 Trends in numbers of Leach’s Storm Petrel, Hartlaub’s Gull and Swift and Roseate Terns breeding in South Africa RJM Crawford1,2, PA Whittington3, BM Dyer1 and L Upfold1 1

Marine and Coastal Management, Department of Environmental Affairs and Tourism, Private Bag X2, Rogge Bay 8012, South Africa 2 Avian Demography Unit, Department of Statistical Sciences, University of Cape Town, Rondebosch 7701, South Africa 3 Department of Zoology, Nelson Mandela Metropolitan University, PO Box 77000, Port Elizabeth 6031, South Africa Leach’s Storm Petrel Oceanodroma leucorhoa has bred in South Africa since 1995/96, or earlier, with breeding recorded at three islands. The overall number of active nest sites fell from an average of 18 during 1995/96– 1999/00 to an average of six from 2002/03–2006/07. The numbers of Hartlaub’s Gulls Larus hartlaubii breeding in the Western Cape fluctuated between 1 450 and 7 578 pairs from 1976–2006, with no long-term trend. Small numbers commenced breeding in the Eastern Cape in 2000. The number of Swift Terns Sterna bergii breeding in the Western Cape fluctuated around a level of about

3 600 pairs from 1984–1999 and then increased to more than 7 000 pairs from 2004–2006. It was significantly related to the biomass of epipelagic fish in the region. The number breeding in the Eastern Cape is less than 1 000 pairs. Roseate Terns S. dougallii bred at Dyer Island in the Western Cape until 1971, were absent from this locality between 1972 and 1982, and bred there again from 1991. From 1996–2006, an average of 11 pairs bred at Dyer Island. The number breeding in the Eastern Cape increased from about 140 pairs in 1977 and 1986 to more than 240 pairs in 2000 and was 170 pairs in 2006.

Keywords: food, Hartlaub’s Gull, Larus hartlaubii, Leach’s Storm Petrel, Oceanodroma leucorhoa, population trend, Roseate Tern, South Africa, Sterna spp., Swift Tern

Introduction

Leach’s Storm Petrel

Fifteen species of seabird breed in southern Africa, including a penguin (Spheniscidae), a storm petrel (Hydrobatidae), a pelican (Pelicanidae), a gannet (Sulidae), four cormorants (Phalacrocoracidae), three gulls (Larinae) and four terns (Sterninae). Four of these birds (Great White Pelican Pelecanus onocrotalus, White-breasted Cormorant Phalacrocorax lucidus, Grey-headed Gull Larus cirrocephalus and Caspian Tern Sterna caspia) breed in inland water bodies, as well as around the coast (Crawford et al. 2006). Most of southern Africa’s Great White Pelicans, Grey-headed Gulls and Caspian Terns breed outside the Benguela ecosystem (Cooper et al. 1992, Brooke et al. 1999, du Toit et al. 2003, Hockey et al. 2005, McInnes 2006). The bulk of the Damara Tern Sterna dougallii population breeds north of South Africa. Trends in the South African populations of African Penguin Spheniscus demersus, Cape Gannet Morus capensis, the four cormorants and Kelp Gull Larus dominicanus have been described by Underhill et al. (2006), Whittington et al. (2006), Crawford (2007), and Crawford et al. (2007a, 2007b). This paper reviews trends in numbers breeding in South Africa for the other four seabirds: Leach’s Storm Petrel Oceanodroma leucorhoa, Hartlaub’s Gull L. hartlaubii, Swift Tern Sterna bergii and Roseate Tern S. dougallii.

Leach’s Storm Petrel was first confirmed breeding in southern Africa (and the Southern Hemisphere) at Dyer Island in November 1996. However, 17 birds occupied sites at this island in November 1995, when it is likely that they bred (Whittington et al. 1999). Subsequently, breeding was also recorded at Dassen and Jutten islands (Whittington et al. 2001). The numbers of sites at these three localities since 1995/96, that were considered active, are shown in Table 1. Breeding appears to have ceased at Jutten Island in 2003/ 04 and at Dyer Island in 2005/06, possibly because of large numbers of cormorants Phalacrocorax spp. nesting on top of the dry stone walls that were used for breeding by Leach’s Storm Petrels (Underhill et al. 2002). The overall number of active sites that were recorded fell from an average of 18 during 1995/96–1999/00 to an average of six from 2002/03– 2006/07. Hartlaub’s Gull Hartlaub’s Gulls breed at a large number of localities in South Africa’s Western Cape (Crawford and Underhill 2003). Birds move between different breeding localities (Crawford et al. 1994). In 22 years between 1978 and 2006, estimates are Top Predators of the Benguela System

211

Table 1: Numbers of active breeding sites for Leach’s Storm Petrels at islands off South Africa, 1995/96–2006/07 Season

Jutten

1995/96 1996/97 1997/98 1998/99 1999/00 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07

4 6 3 0 0 0 0

Dassen

Dyer

Total

1 1 2 1 1 2 4 4 4 5

17 19 11 20 20 7 3 2 5 1 0 0

17 19 12 21 22 12 10 7 9 5 4 5

Figure 1: Numbers of Hartlaub’s Gulls estimated to breed in South Africa’s Western Cape, 1978–2006

Information for Dyer Island for 1995/96–1996/97 from Whittington et al. (1999); for 1997/98–2002/03 from Underhill et al. (2002). Information for Jutten and Dassen islands for 2001/02–2002/03 from Underhill et al. (2002).

available of the number breeding in this province, based on methods described by Crawford and Underhill (2003). The numbers breeding fluctuated between 1 450 pairs in 1993 and 7 578 pairs in 1989, with no long-term trend (mean = 3 800 pairs, SD = 1 500 pairs, n = 22, Figure 1). One pair of Hartlaub’s Gulls bred in the Eastern Cape in 2000, and 16 birds in 2001 (Anon. 2001a, 2001b).

than 7 000 pairs from 2004–2006 (Figure 2). The number breeding was significantly correlated with the combined biomass of anchovy Engraulis capensis and sardine Sardinops sagax measured by acoustic survey in November of the previous year (n = 20, r = 0.745, P < 0.001). Some 100–125 pairs of Swift Terns bred at Stag Island in the Eastern Cape in 1978 and in 1984 (Cooper et al. 1990). At this island and nearby Seal Island, 101 pairs bred in 1992, 220 pairs in 1999, 668 pairs in 2000 and 660 pairs in 2001 (NTW Klages in lit., Marine and Coastal Management unpublished records). In 2006, 47 pairs nested at Bird Island (AJ Tree in lit.).

Swift Tern

Roseate Tern

The number of Swift Terns breeding in South Africa’s Western Cape has been estimated for 21 years between 1984 and 2006, using methods described by Crawford (2003). Birds move between different breeding localities (Crawford et al. 2002). The overall number breeding in the Western Cape fluctuated around a level of about 3 600 pairs from 1984– 1999 (SD = 1 100 pairs, n = 14) and then increased to more

A small number of Roseate Terns bred at Dyer Island, Western Cape, until 1971, but the species was not thereafter recorded at the island until 1982 (Randall et al. 1991, Harrison et al. 1997). In 1991, one pair bred at Dyer Island. From 1996–2006, an average of 11 pairs (maximum 18 pairs) bred at this locality (Harrison et al. 1997, Table 2). The number of Roseate Terns breeding in Nelson

Table 2: Numbers of pairs of Roseate Terns breeding at islands off South Africa, 1977–2006 Year

Dyer Island

Jahleel Island St Croix Island

Seal Island

Stag Island

Bird Island

Eastern Cape

1977

118–139

1986

134

1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006

1

10 5 3 11 6 18 17 15 3 16 12

0

9

0

0

112

121

0 0 0 0 0

0 44 0 0 0

0 0 0 0 0

0 0 0 0 0

144 178 180 210–220 240–250

152 222 180 210–220 240–250

0

30

0

0

70–75

100–105

0 0

58 0

0 0

0 0

75 170

133 170

Information for Eastern Cape for 1977 and 1986 from Randall et al. (1991). Information for Eastern Cape for 1996 and for Bird Island and Eastern Cape from 1998–2000 from Tree and Klages (2003). Information for St Croix and Bird islands and Eastern Cape for 2003 from Tree (in lit.).

212


Mandela (Algoa) Bay was estimated to be 118–139 pairs in 1977 and 134 pairs in 1986 (Randall et al. 1991). It was 152 pairs in 1996, 210–220 pairs in 1999, 240–250 pairs in 2000 (Tree and Klages 2003) and 100–170 pairs in the early 2000s (Table 2). Since 1978, Roseate Terns have only bred at Jahleel, St Croix and Bird islands in Nelson Mandela Bay (Randall et al. 1991). Between 1996 and 2006, the number of Roseate Terns breeding in South Africa was between about 100 and 250 pairs. Acknowledgements – Financial support for this study was provided by the Marine Living Resources Fund. CapeNature, Department of Environmental Affairs and Tourism (South Africa), Robben Island Museum, South African National Parks and South African Navy provided logistical support. The paper is a contribution to the project LMR/EAF/03/02 of the Benguela Current Large Marine Ecosystem (BCLME) Programme.

Figure 2: Numbers of Swift Terns estimated to breed in South Africa’s Western Cape, 1984–2006. The combined biomass of anchovy and sardine off South Africa, estimated by acoustic survey in November of the preceding year, is also shown

References Anon. 2001a Recent reports. Bee-eater 52(1): 14–16. Anon. 2001b Recent reports. Bee-eater 52(3): 56–58. Brooke RK, Allan DG, Cooper J, Cyrus DP, Dean WRJ, Dyer BM, Martin AP, Taylor RH 1999. Breeding distribution, population size and conservation of the Greyheaded Gull Larus cirrocephalus in southern Africa. Ostrich 70: 157–163. Cooper J, Brooke RK, Cyrus DP, Martin AP, Taylor RH, Williams AJ 1992. Distribution, population size and conservation of the Caspian Tern Sterna caspia in southern Africa. Ostrich 63: 58–67. Cooper J, Crawford RJM, Suter W, Williams AJ 1990. Distribution, population size and conservation of the Swift Tern Sterna bergii in southern Africa. Ostrich 61: 56–65. Crawford RJM 2003. Influence of food on numbers breeding, colony size and fidelity to localities of Swift Terns in South Africa’s Western Cape, 1987–2000. Waterbirds 26(1): 44–53. Crawford RJM 2007. Trends in numbers of three cormorants Phalacrocorax spp. breeding in South Africa’s Western Cape Province. In: Kirkman SP (ed.) Final Report of BCLME (Benguela Current Large Marine Ecosystem) Project on Top Predators as Biological Indicators of Ecosystem Change in the BCLME. Avian Demography Unit, Cape Town. Crawford RJM, Cooper J, Dyer BM, Upfold L, Venter AD, Whittington PA, Williams AJ, Wolfaardt AC 2002. Longevity, inter-colony movements and breeding of Crested Terns in South Africa. Emu 102: 1–9. Crawford RJM, Dundee BL, Dyer BM, Klages NTW, Meÿer MA, Upfold L 2007a. Trends in numbers of Cape Gannets (Morus capensis), 1956/57–2005/06, with a consideration of the influence of food and other factors. ICES Journal of Marine Science 64: 169–177. Crawford RJM, Dyer BM and Brooke RK 1994. Breeding nomadism in southern African seabirds – constraints, causes and conservation. Ostrich 65(2): 231–246. Crawford RJM, Goya E, Roux J-P, Zavalaga CB 2006. Comparison of assemblages and some life-history traits of seabirds in the Humboldt and Benguela systems. African Journal of Marine Science 28: 553–560. Crawford RJM, Underhill LG 2003. Aspects of breeding, molt, measurements and population trend of Hartlaub’s Gull in Western Cape, South Africa. Waterbirds 26(2): 139–149. Crawford RJM, Underhill LG, Altwegg R, Dyer BM, Upfold L 2007b. The influence of culling, predation and food on Kelp Gulls Larus dominicanus off western South Africa. In: Kirkman SP (ed.) Final

Report of BCLME (Benguela Current Large Marine Ecosystem) Project on Top Predators as Biological Indicators of Ecosystem Change in the BCLME. Avian Demography Unit, Cape Town. du Toit M, Boere GC, Cooper J, de Villiers MS, Kemper J, Lenten B, Simmons RE, Underhill LG, Whittington PA (eds) 2003. Conservation Assessment and Management Plan for southern African coastal seabirds. Avian Demography Unit, Cape Town and Conservation Breeding Specialist Group, Apple Valley. Harrison JA, Allan DG, Underhill LG, Herremans M, Tree AJ, Parker V, Brown CJ (Eds) 1999. Southern African Bird Atlas. BirdLife South Africa, Johannesburg. Hockey PAR, Dean WRJ, Ryan PG 2005. Roberts Birds of Southern Africa. 7th ed. John Voelcker Bird Book Fund, Cape Town. McInnes AM 2006. Biology of the Grey-headed Gull Larus cirrocephalus in South Africa. MSc thesis, University of KwazuluNatal, Pietermaritzburg. Randall RM, Randall BM, Ralfe M 1991. Roseate terns in South Africa: population size, revision of previous estimate and conservation. Bontebok 7: 1–6. Tree AJ, Klages NTW 2003. Status, biometrics moult and possible relationships of the South African population of Roseate Tern. Ostrich 74: 74–80. Underhill LG, Crawford RJM, Camphuysen CJ 2002. Leach’s Storm Petrels Oceanodroma leucorhoa off southern Africa: breeding and migratory status and measurements and mass of the breeding population. Transactions of the Royal Society of South Africa 57: 42–46. Underhill LG, Crawford RJM, Wolfaardt AC, Whittington PA, Dyer BM, Leshoro TM, Ruthenberg M, Upfold L, Visagie J 2006. Regionally coherent trends in colonies of African Penguins Spheniscus demersus in the Western Cape, South Africa, 1987– 2005. African Journal of Marine Science 28: 697–704. Whittington PA, Dyer BM, Crawford RJM, Williams AJ 1999. First recorded breeding of Leach’s Storm Petrel Oceanodroma leucorhoa in the Southern Hemisphere at Dyer Island, Souith Africa. Ibis 141: 327–320. Whittington PA, Dyer BM, Underhill LG 2001. Leach’s Storm Petrel Oceanodroma leucorhoa breeding in South Africa. Bulletin of the African Bird Club 8: 134. Whittington PA, Martin AP, Klages NTW 2006. Status, distribution and conservation implications of the Kelp Gull (Larus dominicanus vetula) within the Eastern Cape region of South Africa. Emu 106: 127–131.


213

214


Colonial waterbirds

Chapter 27 The development of the heronry on Robben Island, Western Cape, South Africa, 1980–2005 z LG Underhill1*, RJM Crawford1,2, DM Harebottle1 & KMC Tjørve1z

1

Avian Demography Unit, University of Cape Town, Rondebosch 7701, South Africa. Marine and Coastal Management, Department of Environmental Affairs and Tourism, Private Bag X11, Roggebaai 8012, South Africa * Corresponding author: e-mail: [email protected]

2

A heronry was first discovered on Robben Island in 1980. Since 1987 it has been monitored annually, and breeding was recorded each year, in spring and summer. In most years, all species nested in a single heronry usually sited in alien trees; occasionally, some or all Crowned Cormorants and African Sacred Ibises bred separately from the remaining species. The location of

heronry moved within the area of alien trees which covers most of the southern and eastern parts of the island. The site chosen for the heronry avoided the area close to the settlement. Cattle Egrets were the most numerous species. African Sacred Ibises were first recorded breeding in 1991.

Keywords: Herons, Egrets, Crowned Cormorant, African Sacred Ibis, Robben Island, alian vegetation, Phalocrocorax coronatus, Threskiornis aethiopicus

Introduction Europeans first settled on the shores of Table Bay in 1652, and initiated what ultimately became the City of Cape Town. Robben Island, close to this settlement, is consequently the most altered of the South African offshore islands (Smith 1997). One of the largest changes has been to the vegetation. Areas previously covered by low scrub and grassland have been replaced by dense stands of alien vegetation, so that approximately one-third of the island has been transformed into woodland containing rooikrans Acacia cyclops, cluster pine Pinus pinaster, manatoka Myoporum serratum and several species of Eucalyptus (Brooke 1983, Barnes 1998). Likewise, the bird species found on Robben Island have changed since the 1600s. Ostrich Struthio camelus, Chukar Partridge Alectoris chukar, Cape Francolin Francolinus capensis, Helmeted Guineafowl Numida meleagris and Indian Peafowl Pavo cristatus were introduced to the island (Barnes 1998, Crawford and Dyer 2000). Other species arrived naturally, and opportunistically made use of new habitats. Some of the species that form the heronry described in this paper are such arrivals, exploiting the breeding niche created by the stands of alien trees. As the City of Cape Town expanded, demands for space for development have resulted in the loss of wetlands, especially on the Cape Flats. Suitable sites for heronries have been destroyed; at some potential sites, large numbers of people in close proximity prevent heronries from forming or disturb them if they do. The establishment of a heronry on Robben Island may be an indicator of a shortage of nesting opportunities on the mainland, especially because we observed that most of the birds in the heronry commute to the z

mainland to bring food for their chicks. The existence of a heronry on the island was discovered during a survey of the shorebirds of the island undertaken on 9 December 1980. This survey was part of a census of the waders and other waterbirds of the Western Cape conducted in the austral summer 1980/81 (Ryan et al. 1988). This paper describes the subsequent development of the heronry. Methods Robben Island (33°49'S 18°22'E) is the largest (507 ha) of the islands which lie immediately offshore of the south-western coast of Africa (Fig. 1). The island is 7 km from the closest point on the mainland and 11 km from the port of Cape Town. Descriptions of the island and its birdlife are provided by Barnes (1998), Crawford and Dyer (2000) and Underhill et al. (2001). Between January 1987 and the 2000 breeding season, the heronry was searched for on 42 occasions; if an active heronry was found, its position was noted and the numbers of breeding pairs of each species was counted, with no record kept of stage of breeding. These surveys were opportunistic, with one to seven taking place per year. Maximum counts for each species during surveys for each breeding season were extracted, and reflect the minimum numbers of nests for each species. In the breeding seasons from 2001 to 2005, the island was monitored regularly and the heronry, once found, was visited at least monthly while active. From 2002, African Sacred Ibis chicks were ringed prior to fledging with SAFRING 12.5 mm stainless steel rings, and cohort specific colour rings, with the colours indicating the year and site. We report resightings of birds from the 2002 cohort.

present address: Lista Bird Observatory, Fyrveien 6, 4563 Borhaug, Norway Top Predators of the Benguela System

217

Figure 1: Location of Robben Island in Table Bay, showing localities mentioned in the text

Results When the heronry on the island was discovered in December 1980, it was situated in the mixed pine and eucalyptus “forest” near the northeastern corner of the island (Fig. 2), in an area which now falls within the African Penguin Spheniscus demersus breeding colony, which started in 1983/84 (Crawford and Dyer 2000). There were c. 30 nests

of Little Egret Egretta garzetta, most with large clamouring young, c. 300 nests of Cattle Egret Bubulcus ibis, most with young able to leave their nests either by climbing into branches or flying short distances, and c. 25 nests of Blackcrowned Night Herons Nycticorax nycticorax, with large young (Table 1). Eighty-six nests of Crowned Cormorant Phalocrocorax coronatus were counted, with young of varying sizes, from downy chicks to large young able to leave the nests and scramble into the trees. During 1987–2000, breeding was recorded on 20 occasions, and was never observed during the months March– July, the austral autumn and winter. Breeding was observed in each year, and the location of the heronry moved between most successive years (Fig. 2). In 1991, a single pair of Yellow-billed Egrets was recorded in the heronry; in 1997, only Crowned Cormorants bred on the island (Table 1). In 2001, the site of the heronry was first determined on 12 August, when a flock of African Sacred Ibises Threskiornis aethiopicus congregated on the roof of an abandoned building, c. 200 m west of the position in 2000 (Fig. 2). By early September, the ibises were building nests in Acacia trees adjacent to the building, numbers were increasing and they had been joined by the other species. By October, there were c. 600 African Sacred Ibis nests, at least four Black-crowned Night Heron nests, c. 100 Crowned Cormorant nests, c. 600 Cattle Egret nests and c. 50 Little Egret nests. The ibis chicks were the last to leave the heronry. All had fledged by mid January 2002; many fledglings remained close to the heronry and others foraged and roosted at Van Riebeck’s Quarry. Between 2002 and 2005, the heronry was located immediately inland of the beach just south of Murrays Bay Harbour. In 2002, it was located in September, more than a month later than the previous year. The heronry was in a dense and vigorous stand of rooikrans, 4 m tall. The heronry moved to some extent within this area, but there was a core area which was occupied in all four years. Old maps of the island show this area to have been a wetland; in winter 2002 this area was marshy with standing water up to 20 cm deep; subsequent years were not as wet. The configuration of the

Table 1: Maximum counts of nests in the heronries on Robben Island, 1980–2003. Each breeding season extends into January or February the following year Year

Crowned Cormorant

Black-crowned Night Heron

Little Egret

Cattle Egret

1980

86

25

30

300

41 20 11 >40 >100 “bred” 20 3 20 1 3

2 >10 30 >20 20 “bred” 10 3 2 1 “some”

872 >500 >50 “100s” 800 “bred” >500 >10 45 60 123

8 10 >4 6 25 9 6

2 10 50 10 8 4 12

1987 1988 1989 1990 1991* 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005

319 405 355 323 206 300 317 261 268 290 253 176 87 50 100 100 100 >200 40

* In 1991, one pair of Yellow-billed Egrets bred 218


95 530 600 600 700 564 >150

African Sacred Ibis

4 7 8

4 49 120 600 166 100 62 216

Observers, date of single visit

LGU, G.D. Underhill, H.G. Robertson, 9 December 1980; Ryan et al. 1988 MCM MCM MCM MCM MCM MCM MCM MCM MCM MCM MCM MCM MCM JA Harrison, S Kuyper, 30 November 2000 MCM, KMCT, LGU MCM, DMH, KMCT, LGU DMH, KMCT MCM, DMH, M Wheeler O Noels

Figure 2: Year-by-year positions of the heronry on Robben Island, 1980 and 1987–2005 Top Predators of the Benguela System

219

heronry made counting difficult; the counts of nests of Blackcrowned Night Herons, Crowned Cormorants, Cattle Egrets and Little Egrets over this four-year period are approximate (Table 1). In 2002, a few African Sacred Ibis were observed in the trees in the colony on several occasions. They did not settle to breed there and ultimately nested in a separate colony on the northeastern corner of the island, on the ground in dry grassland. This colony was first observed in late September, and grew to 166 nests. The last young fledged during December. In 2003 and 2004, the ibises also bred on ground outside the main heronry (Fig. 2). In 2003, the ibises bred close to the shore on the northwestern corner of the island (Fig. 2). On 8 September, courtship behaviour was observed. By 15 October there were 20 nests, and by 20 October this had increased to 100 nests, all containing eggs. All nests were washed away in a high tide event on 27 October. These ibises did not re-nest on the island and breeding failed completely in this year. In 2004, ibises bred in two places, both successfully: some at the same site close to the shore as in 2003, and some in the main heronry. In 2005, the ibises bred only in the main heronry. In 2003, Crowned Cormorants formed a second heronry at the southwestern extremity of the island, in an area of alien acacias close to the position of the 2001 heronry. By early November it contained 50 cormorant nests. Most nests produced fledglings. In the 2002 breeding season, 78 African Sacred Ibis fledglings were ringed and colour-ringed, 28 on 13 November and 50 on 4 December. Nine resightings of these birds were made: six on Robben Island (28 November 2002, 15 December 2003, 31 May 2004, 11 August 2004, 22 October 2005 and 2 December 2005) and three on the mainland of the Cape Peninsula, two at Strandfontein Sewage Works, on 24 and 31 May 2003, and one in a roost at Die Oog Bird Sanctuary, Bergvliet, on 22 October 2004 (Fig. 1). Four African Sacred Ibis fledglings, ringed at Rondevlei Nature Reserve (Fig. 1) during the 2002 breeding season, were subsequently resighted on Robben Island; on 8 September 2003, 15 December 2003, 22 October 2004 and 2 December 2005. The Rondevlei bird observed on Robben Island on 22 October 2004 was in the heronry, but not in full breeding plumage and the bird was not at an active nest. Both resightings made on 2 December 2005 were in the Robben Island heronry, and were of birds in full breeding plumage which were presumably breeding. Discussion In most years, all species nested in a single heronry; occasionally, some or all Crowned Cormorants and African Sacred Ibises bred separately from the remaining species. The heronry was usually sited in alien trees, either near the northwestern or southeastern corners of the island (Fig. 1). Between years, the heronry moved either short distances, or to the diagonally opposite corner of the island. The alien trees cover most of the southern and eastern parts of the island, and until 2002, the sites for the heronry were far from areas close to concentrations of people (Fig. 1). In the four years 2002–2005, the main heronry was adjacent to the tourist route between the harbour and the former high security prison, which is the focus of tourist activities. Cattle Egrets were the most numerous species in every breeding season, with a peak count of 872 breeding pairs in the 1987 breeding season. A pair of Yellow-billed Egret Garzetta intermedia bred in 1991 (Table 1). The recorded peak laying period for Cattle Egrets (Martin 1997b) and Little Egrets (Martin 1997a) in the Western

220


Cape is September–October and October–December respectively. Cattle Egrets breeding on Robben Island laid mostly in September resulting in many chicks fledging in early December. By mid January there were still some nests with large chicks. Little Egrets on Robben Island laid later than most of the other species resulting in most chicks starting to hatch in November and fledging in January. African Sacred Ibises were first recorded breeding on Robben Island in the 1991 breeding season, with fewer than 10 pairs breeding in most subsequent seasons until 1998. Then numbers increased rapidly to c. 600 pairs in 2001, but subsequently decreased (Table 1). The peak breeding season for African Sacred Ibises in the Western Cape is from August to January (Anderson 1997). They bred in trees until the 2002 breeding season, when they bred successfully in grassland, far from water, for the first time. This is surprising, given the presence of feral cats Felis catus on the island. They bred between the perimeter road and the sea in 2003 and 2004. In the first of these two years, the colony was washed away in a flood tide event; in spite of this loss, the birds chose to breed at the same site the following year, and bred successfully. Black-crowned Night Herons mainly start breeding in the Western Cape in September with few records of breeding later than January (Anderson 1997). On Robben Island three chicks had fledged by 29 November 2001; one pair, near the edge of the colony, was seen to be incubating eggs at this time. The chick from this nest fledged in mid-February 2002. The first record of Crowned Cormorants breeding at Robben Island was in 1949 (Kriel et al. 1980). Crowned Cormorants rarely nest in trees, which are absent from most of southern Africa’s marine islands, but those breeding on Robben Island often do so (Kriel et al. 1980, Crawford and Dyer 2000). Most laying in the Western Cape is from December to March, although breeding may take place throughout the year (Rand 1960, Crawford et al. 1999). The conservation status of the Crowned Cormorant was changed from “Near-threatened” (Barnes 2000, BirdLife International 2000) to “Least Concern” in 2002 (du Toit et al. 2003). In many years Robben Island supports the world’s largest colony of Crowned Cormorants (Crawford and Dyer 2000); in 1988 it held 15% of the global population of the species (information in Table 1 and Crawford et al. 1982, du Toit et al. 2003). However, the population at Robben Island has decreased since then (Table 1). Both Cattle Egrets and African Sacred Ibises commuted to the mainland to feed and gather food for their young (pers. obs). Most birds made the shortest sea crossing to the mainland flying from the island in the direction of Bloubergstrand, 7 km away (Fig. 1). The distance farther inland at which they gathered food is unknown, but the extensive dairy farming area of the southern Swartland starts about 8 km farther inland, an area where both Cattle Egrets and African Sacred Ibises are abundant (Hockey et al. 1989). The round trip distance is therefore likely to exceed 30 km. Both species were observed leaving the island from first light, often flying against the strong southeasterly winds that prevail during the breeding season, and they returned to the island until darkness fell. They flew singly or in small single-species flocks, seldom exceeding 10 birds. The energetics of breeding on an offshore island need to be investigated. Between most breeding seasons, the position of the heronry shifted. This is probably because the deposited guano kills leaves and branches of the trees in the heronry, and trees may take a couple of years to recover. The exception was the use of adjacent areas between the 2002 and 2005 breeding seasons; the trees in this area were sufficiently vigorous that they had been able to recover from the deposition

of guano of the previous year. The area of alien trees required to maintain the heronry on the island is considerably larger than the actual area occupied by the heronry in a single year. The size of this area is unknown, but is likely to be of the order of several hectares. The presence of the heronry adds to the dilemma associated with the clearance of alien vegetation from Robben Island. The fact that the birds breed on an island 7 km offshore, and c. 15 km from the nearest feeding areas, points to a shortage of suitable breeding sites on the mainland. The largest known heronry in the Greater Cape Town area, at Blouvlei (Fig. 1), where nearly 2000 breeding pairs of 12 species were recorded, was destroyed during the “Century City” development in 1996. An artificial heronry, subsequently constructed, is space-limited and supports about 150 breeding pairs of eight species (Harrison & Underhill 2000, JA Harrison in litt.). If all alien trees were removed from Robben Island, the heronry there would be lost. Besides the loss of bird biodiversity on the island, there is little doubt that this would have a negative impact on populations of these species in the Greater Cape Town area. Robben Island was declared a World Heritage Site in 1999, motivated primarily by its cultural history. The alien vegetation also represents this cultural history, with various groves and avenues of trees having been planted by soldiers, lepers and prisoners. In the past, the island was used to exploit seabirds and seals, to quarry slate and lime, to isolate lepers, to guard Table Bay as a naval base during the Second World War, and, most infamously, to imprison political dissidents (Fish 1924, de Villiers 1971, Smith 1997). Currently, the entire island forms the Robben Island Museum, and welcomes 300 000 tourists per year. In spite of all these anthropogenic influences, Robben Island remains a wildlife sanctuary of outstanding value, and is an Important Bird Area (Barnes 1998, Crawford and Dyer 2000). In terms of the Ramsar Convention, it qualifies as a Wetland of International Importance (Underhill et al. 2001). The only other information on the age of first breeding for the African Sacred Ibis is of three records of birds breeding at age three years; these chicks had been cohort colourringed at the Rondevlei Nature Reserve between September and December 2002, and recorded breeding at the same locality in November 2005 (Harebottle & Gibbs 2006). The two birds resighted on 2 December 2005 in the Robben Island heronry both confirm three years as the age of first breeding, and provided the first record of breeding away from the natal site. Acknowledgements – We acknowledge support from the National Research Foundation, Earthwatch Institute, Darwin Initiative, Benguela Current Large Marine Ecosystem “Top Predators” project and the University of Cape Town Research Committee. Robben Island Museum and Marine and Coastal Management Branch (MCM), Department of Environmental Affairs and Tourism provided logistical support. Surveys were conducted by BM Dyer, JA Harrison, DM Harebottle, S Kuyper, O Noels, HG Robertson, GD Underhill and L Upfold. JA Harrison and AJ Williams supplied additional information and commented on drafts. Cathy Boucher prepared the figures.

References Anderson MA 1997. Sacred Ibis Threskiornis aethiopicus. In: Harrison JA, Allan DG, Underhill LG, Herremans M, Tree AJ, Parker V and CJ Brown (Eds). The Atlas of Southern African Birds, Vol 1. pp 102–103. BirdLife South Africa, Johannesburg Barnes KN (ed.) 1998. The Important Bird Areas of Southern Africa. BirdLife South Africa, Johannesburg Barnes KN (ed.) 2000. The Eskom Red Data Book of Birds of South Africa, Lesotho and Swaziland. BirdLife South Africa, Johannesburg BirdLife International 2000. Threatened Birds of the World. Lynx Edicions and BirdLife International, Barcelona and Cambridge Brooke RK 1983. On the 17th century avifauna of Robben Island, South Africa. Cormorant 11: 15–20 Crawford RJM and Dyer BM 2000. Wildlife of Robben Island. Bright Continent Guide 1. Avian Demography Unit, University of Cape Town, Cape Town Crawford RJM, Dyer BM and Upfold L 1999. Seasonal pattern of breeding by Cape and Crowned Cormorants off western South Africa. Ostrich 70: 193–195 Crawford RJM, Shelton PA, Brooke RK and Cooper J 1982. Taxonomy, distribution, population size and conservation of the Crowned Cormorant Phalacrocorax coronatus. Gerfaut 72: 3–30 De Villiers SA 1971. Robben Island: Out of Reach, out of Mind. Cape Town; C. Struik: 169 pp Du Toit M, Boere GC, Cooper J, de Villiers M, Kemper J, Lenten B, Simmons RS, Underhill LG and Whittington PA 2003. Conservation Assessment and Management Plan for Southern African Coastal Seabirds. Cape Town: Avian Demography Unit and Apple Valley: Conservation Breeding Specialist Group Fish JW 1924. Robben Island: An Account of Thirty-four Years’ Gospel Work amongst Lepers of South Africa. John Ritchie, Kilmarnock Harebottle DM and Gibbs D 2006. At what age do African Sacred Ibis breed? Promerops 265: 13 Harrison JA and Underhill LG 2000. Blouvlei: development is for the birds. Africa – Birds & Birding 5(1): 42–47 Hockey PAR, Underhill LG, Neatherway M and Ryan PG 1989. Atlas of the birds of the southwestern Cape. Cape Bird Club, Cape Town Kriel F, Crawford RJM and Shelton PA 1980. Seabirds breeding at Robben Island between 1949 and 1980. Cormorant 8: 87–96 Martin AP 1997a. Little Egret Egretta garzetta. In: Harrison JA, Allan DG, Underhill LG, Herremans M, Tree AJ, Parker V and CJ Brown (Eds). The Atlas of Southern African Birds, Vol 1. pp 54– 56. BirdLife South Africa, Johannesburg Martin AP 1997b. Cattle Egret Bubulcus ibis. In: Harrison JA, Allan DG, Underhill LG, Herremans M, Tree AJ, Parker V and CJ Brown (Eds). The Atlas of Southern African Birds, Vol 1. pp 61– 63. BirdLife South Africa, Johannesburg Rand RW 1960 The biology of guano-producing seabirds. 3. The distribution, abundance and feeding habits of the cormorants Phalacrocoracidae off the south-western coast of the Cape Province. Investigational Report of the Sea Fisheries Research Institute of South Africa 42: 1–32 Ryan PG, Underhill LG, Cooper J and Waltner M 1988. Waders (Charadrii) and other waterbirds on the coast, adjacent wetlands and offshore islands of the southwestern Cape Province. Bontebok 6: 1–19 Smith C 1997. Robben Island. Struik, Cape Town Underhill LG, Whittington PA and Calf KM 2001. Shoreline birds of Robben Island, Western Cape, South Africa. Wader Study Group Bulletin 96: 37–39


221

222


Turtles

Chapter 28 Brief report on the status of turtles in Angola Miguel Morais Universidade Agostinho Neto, Angola

Sporadic reports from the 1970s and 1980s indicated that extensive nesting and non-nesting marine turtle populations occurred on the coast of Angola, from the northern province of Cabinda south to the Namibian border at the Cunene River. Generally however, the distribution and status of marine turtles along this ~1650 km coastline is not well known. More recently, between 2000 and 2006, several surveys were conducted in Angolan waters or along the coast, in search of marine turtles. These included dedicated nesting surveys at Palmeirinhas, pelagic surveys off northern Angola, casual beach surveys along most of the Angolan coastline and interviews at fishing communities and markets. Of the turtle species that were recorded, the Green Turtle Chelonia mydas is perhaps the most abundant. Although their distribution is uneven, nesting of this species has been widely documented in many regions, including Cabinda and Quicombo. Nesting populations of the Green Turtle are also known to occur in the waters of neighbouring countries, including the Congo and the Democratic Republic of Congo. It is thought that Mussulo Bay and the mouth of the Cunene River may comprise important year-round feeding areas for of all age classes of this species. Angola is one of the most important nesting areas, worldwide, for the Olive Ridley Turtle Lepidochelys olivacea, which has been recorded at locations around Cabinda, Ambriz, Luanda, Rio Longa, Quicombo and Lobito. At-sea sightings of this species have been reported along the entire coast, particularly within the Bay of Bengo and the Bay of Cabinda. The Leatherback Turtle Dermochelys coriacea appears to

be rare in Angola; those that occur here are likely a continuum of the large nesting population in Gabon. This species nests widely and has been documented in Cabinda and throughout the 200 km stretch of coast south of Luanda including Quicombo and Rio Longa. Both adults and subadults of this species have been observed at sea, quite close to the coast. While the Hawksbill Turtle Eretmochelys imbricata has been reported to occur in Angola, their status is unknown and there are no confirmed instances of nesting. The status of the Loggerhead Turtle Caretta caretta in the region is also unclear. This species is reported to nest along the Skeleton Coast in Namibia and may therefore be expected to also nest in southern Angola. In summary, nesting of the Green, Olive Ridley and Leatherback Turtles in Angola, have been confirmed. The breeding status of the other two species which have been recorded in Angola, the Loggerhead and Hawksbill Turtles, are uncertain. All the above are classified as either endangered (Green, Olive Ridley and Loggerhead) or critically endangered (Leatherback and Hawksbill) by the IUCN, and are included on Appendix I of CITES which prohibits international trade to or from the signatory countries. Disturbingly, there is some evidence of a decline in turtle numbers in Angola since the 1980s, when the density of turtles along the coast in the vicinity of Luanda was estimated at 75 nests per km. During more recent surveys, conducted from 2003– 2006, the density of nests in this region was estimated to have declined to below 40 per km.


225

226


The influence of the environment and fish stocks on trends in top predators

Chapter 29 Food, fishing and seabirds in the Benguela upwelling system Robert J.M. Crawford1,2 1 2

Marine and Coastal Management, Private Bag X2, Rogge Bay 8012, South Africa. e-mail: [email protected] Avian Demography Unit, Department of Statistical Sciences, University of Cape Town, Rondebosch 7701, South Africa

The Benguela upwelling system off south-western Africa supports sardine Sardinops sagax and anchovy Engraulis encrasicolus that are harvested by purseseine fisheries and are the main prey of three endemic seabirds: African Penguin Spheniscus demersus, Cape Gannet Morus capensis and Cape Cormorant Phalacrocorax capensis. There have been large, long-term changes in the abundance and distribution of the fish resources that have influenced the seabird populations. After 1956/57, the numbers of penguins and gannets breeding in Namibia decreased by 90% and 95%, respectively. After 1978/79, the number of Cape Cormorants breeding in Namibia decreased by 76%. These decreases were significantly related to the biomass of sardine and anchovy in Namibia and are thought to result mainly from a greatly reduced availability of prey. In South

Africa, when sardine collapsed it was replaced by anchovy. In the Western Cape, numbers of Cape Gannets and Cape Cormorants were stable after the collapse of the sardine but African Penguins decreased. The sardine resource recovered in the 1980s and 1990s but at the turn of the century was displaced to the east, leading to increases and then decreases in numbers of penguins and gannets in the Western Cape. In the Eastern Cape, there were long-term increases in numbers of penguins and gannets, until a recent decrease in penguins. In South Africa, the models used to advise allowable catches for sardine and anchovy are being modified to incorporate a model of African Penguins and functional relationships linking penguins and fish stocks. Consideration is also being given to precluding fishing in areas around seabird breeding colonies.

Keywords: Benguela, food, Morus capensis, Phalacrocorax capensis, Spheniscus demersus

Introduction The Benguela upwelling system off south-western Africa supports a high abundance of epipelagic fish, notably sardine Sardinops sagax and anchovy Engraulis encrasicolus (Schwartzlose et al. 1999). These fish species are the most important prey of three seabirds that are endemic to the region: African Penguin Spheniscus demersus, Cape Gannet Morus capensis and Cape Cormorant Phalacrocorax capensis (Hockey et al. 2005 and references therein). The main seabird breeding localities are grouped in three areas: the Namibian coastline from Cape Cross southwards; South Africa’s Western Cape; and islands near Port Elizabeth in the Eastern Cape (Fig. 1). The breeding localities in the Western Cape are separated from those in Namibia and in the Eastern Cape by distances of about 400 km and 550 km, respectively. Commercial purse-seine fisheries for small pelagic fish commenced off Namibia and South Africa after World War II. They initially targeted sardine and later also anchovy. Other species were caught, when available. There were collapses of the sardine stocks off South Africa in the early 1960s and off Namibia in the late 1960s (Crawford et al. 1987). The South African sardine began a recovery in the 1980s and was again abundant at the start of the 21st century (Fairweather et al. 2006), but prey remained scarce off Namibia, with severe repercussions for seabirds (Crawford 1998, 1999). In the late 1990s, the distribution of sardine off South Africa changed (Fairweather et al. 2006), influencing its availability to seabirds.

This paper reviews trends in the catches and biomass of sardine and anchovy and in the abundance of African Penguins, Cape Gannets and Cape Cormorants in the Benguela system and considers means of accounting for the food requirements of seabirds in management of the purse-seine fisheries. Trends in fish stocks Catches Trends in the catches of sardine and anchovy off Namibia, South Africa and in the Benguela system as a whole are shown in Fig. 2. In Namibia, catches of sardine peaked at 1.4 million t in 1968 but after 1997 were less than 0.2 million t in all years (Fig. 2a). Anchovy was first caught by purse-seine boats in 1964. Catches were small until 1968, when 0.16 million t was landed. In 1978, anchovy became the most important contributor to the purse-seine fishery. This situation continued until 1984 when < 0.02 million t was caught. Low catches continued in 1985 and 1986. In 1987, the anchovy catch (0.38 million t) was the highest yet recorded. Thereafter, catches plummeted and by 1996 just 0.001 million t was caught (Schwartzlose et al. 1999). Off South Africa, catches of sardine fluctuated around 0.1 million t during the early and mid 1950s and then increased to a maximum of around 0.4 million t during 1961– 1963 (Fig. 2b). Catches fell to less than 0.1 million t in 1967. Except in 1968, 1972 and 1976, they remained below this level until 1995. Catches above 0.2 million t were recorded Top Predators of the Benguela System

229

Figure 1: Southern Africa showing the locations of the main seabird breeding colonies around the coastline

from 2002–2005. Anchovy catches increased steadily from 1964 onwards, reaching a peak of approximately 0.6 million t in 1987 and 1988. Subsequently, anchovy catches decreased, with some variability, to a minimum of 0.04 million t in 1996, and then increased sharply again (Fig. 2b). In the Benguela system as a whole, sardine dominated the purse-seine catch in the 1950s and 1960s, and anchovy from 1977–1993. At other times, the two species contributed approximately the same catches (Fig. 2c). Biomass In Namibia, estimates of the biomass of sardine and anchovy were obtained from Virtual Population Analysis (VPA) for the periods 1952–1988 and 1972–1985, respectively (Le Clus 1986; Thomas 1986; Kreiner et al. 2001). From 1990, they were obtained by acoustic surveys (Boyer and Hampton 2001). In South Africa, VPA was used for the period 1950– 1982 (Armstrong et al. 1983) and acoustic surveys from 1984 onwards (Hampton 1987; Fairweather et al. 2006). Estimates obtained by the two methods are not comparable, but they provide an indication of the relative stock sizes of the two fish species for the periods when they were applied. In Namibia, the biomass of sardine was estimated to be above 10 million t from 1963–1965 (Fig. 3a). It decreased rapidly to less than 2 million t by 1970. There was a partial recovery from 1972 to 1974, after which biomass fell to a very low level in 1979. Thereafter, the biomass remained low. Anchovy was thought to be scarce in 1963 (Newman 1970; Thomas 1985). VPA estimates of biomass decreased with fluctuations between 1972 and 1985 (Le Clus 1986; Fig. 3a). Catch rates of anchovy also decreased during this period (Le 230


Clus 1985). In South Africa, the biomass of sardine peaked at 1.7 million t in 1959, after which it decreased sharply until 1966. From 1985 onwards, the stock size increased rapidly. It exceeded 4 million t in 2002, but fell to less than 1 million t in 2005 (Fig. 3b). From 1997–2005, the centre of gravity of purse-seine catches of sardine was displaced some 400 km to the east (Fairweather et al. 2006), and the species became considerably less available to seabirds breeding in the Western Cape. Anchovy may have increased in abundance in 1973, although VPA estimates of biomass possibly reflect catches. After 1985, the anchovy stock showed large fluctuations. In the late 1990s it grew rapidly. The biomass peaked at almost 8 million t in 2001 and then decreased (Fig. 3b). In the Benguela system as a whole, sardine was abundant and anchovy scarce in the 1950s and early 1960s. From the mid 1970s to the mid 1990s, the combined biomass of these two prey species was low, but it improved at the turn of the century as both species increased off South Africa (Fig. 3c). Trends in numbers of seabirds Estimates of the numbers of African Penguins breeding in Namibia, the Western Cape and the Eastern Cape were obtained from Rand (1963a; 1963b), Shelton et al. (1984), Crawford et al. (1995; 2001; updated), Kemper (2006) and Underhill et al. (2006). Estimates of the numbers of Cape Gannets breeding in these regions were obtained from Crawford et al. (2007). Estimates of the numbers of Cape Cormorants breeding in Namibia and the Western Cape were obtained from Rand (1963a; 1963b), Cooper et al. (1982),

Figure 2: Catches of sardine and anchovy made a) off Namibia, b) off South Africa and c) in the Benguela system, 1950–2005

Crawford and Dyer (1995) and Hockey et al. (2005), and were updated from unpublished information. Cape Cormorants breed only in low numbers (several hundred pairs) in the Eastern Cape (Cooper et al. 1982). In Namibia, the number of African Penguins that bred decreased by about 50% between 1956/57 and 1967/68, and by 90% between 1956/57 and 2004/05 (Fig. 4a). Numbers of Cape Gannets decreased by 8% between 1956/57 and 1967/68, by a further 31% by 1978/79, and by 95% between 1956/57 and 2005/06 (Fig. 4a). Numbers of Cape Cormorants fell by 76% between 1978/79 and 2005/06 (Fig. 4a). The number estimated to be breeding in 1956/57 is likely to be too low because birds may have initiated breeding after aerial photographs, on which they were counted, were taken in November. However, it is likely that there was a real increase in the number breeding between 1956/57 and the 1970s because extra space for nesting was provided by the building of a platform north of Walvis Bay and by extension of platforms near Cape Cross (Cooper et al. 1982). Additionally, breeding space became available at Ichaboe Island, as the number of gannets there decreased (Crawford 1991). In the Western Cape, the number of African Penguins breeding decreased by 42% between 1956/57 and 1979/80, and by 64% between 1956/57 and 1993/94 (Fig. 4b). The number breeding increased in the late 1990s and early 2000s as the South African stocks of sardine and anchovy attained high levels, but in 2006/07 was just 33% of that in 1956/57. Between 1956/57 and the mid 1980s, the number of Cape Gannets breeding in the Western Cape was relatively stable (Fig. 4b). The number approximately doubled by 1996/97, but then decreased. The number of Cape Cormorants in the Western Cape was stable from 1956/57–1991/ 92 (Fig. 4b), although the proportion breeding decreased in periods of food scarcity, such as 1990/91 (Crawford and Dyer 1995). However, the number fell by 60% between 1991/92 and remained at a low level thereafter. In the Eastern Cape, the number of gannets increased

Figure 3: Estimates of the biomass of sardine and anchovy a) off Namibia, b) off South Africa and c) in the Benguela system, 1950– 2005

throughout the period of observations (see also Randall and Ross 1979; Klages et al. 1992). There were five times as many gannets breeding in 2005/06 as in 1956/57 (Fig 4c). The number of penguins breeding in this region increased three-fold between 1956/57 and 1993/94, remained at a high level until 2001/02 and then decreased by about 50% (Fig. 4c). Taking the Benguela system as a whole, and therefore the species populations, the number of African Penguins fell by 60% from 141 000 pairs in 1956/57 to 57 000 pairs in 2004/05. The number of Cape Gannets decreased by 45% from 253 000 pairs in 1956/57 to 140 000 pairs in 2005/06. The number of Cape Cormorants fell by 19% from 111 000 pairs in 1956/57 to 91 000 pairs in 2005/06. However the 1956/57 population was probably underestimated (see above). Influence of food on seabird trends The influence of food on trends in seabird populations is most clear for Namibia. Large decreases in numbers of African Penguins and Cape Gannets occurred as the fishery for Namibian sardine expanded and as the stock later collapsed. The decrease of penguins, which have a foraging range of up to about 40 km when breeding (Heath and Randall 1989; Petersen et al. 2006), preceded that of gannets, which are able to forage up to 240 km from colonies (Grémillet et al. 2004). As sardine collapsed, its range contracted to the north making it increasingly less available to the penguin and gannet colonies, which occur in the vicinity of Lüderitz. The fish processing plant at Lüderitz was closed in 1974 because boats had to travel too far to the north before encountering fish shoals (Crawford et al. 1987). The Cape Cormorant may have benefited from an increase in breeding space at the guano platforms and at Ichaboe Island and from an increased abundance of pelagic goby Top Predators of the Benguela System

231

Figure 5: A comparison of trends in the combined biomass of anchovy and sardine and numbers of African Penguins, Cape Gannets and Cape Cormorants breeding in Namibia. Because of the irregular estimates of seabird abundance, the comparisons are of means for five-year periods: 1956–1960 to 2001–2005. All means are expressed as proportions of the maximum observed. The estimate for Cape Cormorants for 1956/57 has been omitted because it was based on aerial photographs taken before the main breeding season of Cape Cormorants in Namibia

Figure 4: Trends in the breeding populations of African Penguin, Cape Gannet and Cape Cormorant a) in Namibia, b) in South Africa’s Western Cape, c) in South Africa’s Eastern Cape (Cape Cormorants are omitted as very few breed in this region), and d) in the Benguela system, 1956/57–2005/06

Sufflogobius bibarbatus, which partially replaced sardine off central Namibia (Crawford et al. 1985). In addition, the platforms north of Walvis Bay gave Cape Cormorants access to the diminishing resources of sardine. However, cormorant numbers also fell substantially after the 1970s. For the periods for which reliable information is available, the sizes of seabird populations breeding in Namibia were significantly related to the combined biomass of anchovy and sardine (Fig. 5). Because of the irregular estimates of seabird abundance, the comparisons were made using the means for five-year periods: African Penguin (r = 0.916, n = 8, P < 0.002); Cape Gannet (r = 0.973, n = 8, P < 0.001); Cape Cormorant (r = 0.899, n = 6, P < 0.01). In each instance fish biomass accounted for more than 80% of the variation in the numbers of seabirds breeding. The abundance of Cape Gannets (r = 0.745, n = 8, P < 0.05) and Cape Cormorants (r = 0.932, n = 6, P < 0.01) was significantly related to the combined catch of anchovy and sardine, suggesting that the availability of fish to these seabirds and the purse-seine fishery was similar. In the Western Cape, there were sometimes contrasting trends for the three seabird species. Following the collapse of the South African sardine, numbers of African Penguins decreased, whereas numbers of Cape Gannets and Cape Cormorants were stable. In the 1960s and 1970s, anchovy replaced sardine in the system (Crawford et al. 1987) and 232


was available to Cape Gannets and Cape Cormorants (Berruti et al. 1993; Crawford and Dyer 1995). However, the southern distribution of most of the parent stock of anchovy placed it beyond the limited foraging range of penguins at western colonies. These colonies decreased but the southern penguin colony of Dyer Island increased (Crawford 1998). When sardine increased in the 1980s, there was concomitant growth of several penguin colonies in the Western Cape (Crawford et al. 2001). Numbers of Cape Gannets similarly increased (Crawford et al. 2007). The overall number of African Penguins breeding in the Western Cape increased at the end of the 20th century as pelagic fish abundance increased off South Africa, and then decreased as sardine became less available. At this time numbers of gannets also decreased. However, the responses of penguins and gannets were not synchronous. The Cape Cormorant decreased in the early 1990s and did not take advantage of the increased abundance of pelagic fish towards the end of that decade. Although the effect of food on trends in seabird numbers is not as clear cut for the Western Cape as for Namibia, in instances seabird abundance has been influenced by food. The numbers of Cape Gannets breeding in South Africa were significantly related to the abundance of sardine and anchovy (Crawford et al. 2007). Numbers of African Penguins breeding at certain colonies in the region were strongly related to pelagic fish biomass (e.g. Cury et al. 2000; Crawford et al. 2001). Other factors have also influenced trends of seabirds in the Western Cape. Repeated outbreaks of avian cholera Pasteurella multocida caused extensive mortality of Cape Cormorants from 1991–2006, especially at Lambert’s Bay, Dassen and Dyer islands. More than 14 000 Cape Cormorants were killed at Dassen Island in 1991, 7 000 at Lambert’s Bay in 2002 and more than 27 000 at Dyer Island between 2002/03 and 2005/06 (Crawford et al. 1992; Williams and Ward 2002; Waller and Underhill 2007). Predation by seals Arctocephalus pusillus around breeding colonies was thought unsustainable for African Penguins at Dyer Island and Lambert’s Bay and for Cape Gannets at Malgas Island (Marks et al.1997; Crawford et al. 2001; Makhado et al. 2006). Information on pelagic fish abundance in the Eastern Cape is not readily available before the mid 1980s so that its influence on long-term trends in seabirds in that province

cannot be gauged. The importance of changes in the distribution of prey species was recently highlighted when the eastward displacement of sardine off South Africa placed the bulk of this forage resource between the breeding localities for seabirds that are in the Western Cape and in the Eastern Cape, and consequentially beyond the foraging ranges of many seabirds in both these regions. By 2005, the centre of distribution of sardine catches was at about 22°E off the South African south coast (Fairweather et al. 2006). In an apparent attempt by penguins to adapt to this shift in the distribution of prey, a penguin breeding colony was initiated at De Hoop on the mainland in 2003. Decreases in the number of African Penguins breeding were observed at those colonies in the Western Cape that are west of Cape Town (unpublished information). However, the eastward displacement of sardine placed more of it within the foraging range of Cape Gannets in the Eastern Cape, and led to a sharp increase in its contribution to their diet (Crawford et al. 2007). The wide foraging range of Cape Gannets buffers them to a greater extent than African Penguins against changes in the distribution of fish prey. From 1988–2004, the coefficient of variation for breeding success of Cape Gannets at Malgas Island was 31%, compared to 37% for African Penguins at Robben Island from 1989–2005 (unpublished information). Accounting for the food requirements of seabirds From the foregoing, it is apparent that both the abundance and the distribution of prey have influenced trends in seabird numbers in the Benguela system. The recent decreases in seabird populations suggest a need to account for their food requirements in the management of the purse-seine fisheries, with which they compete for prey. These fisheries take a substantially greater proportion of the anchovy and sardine stocks than do the birds. In the 1980s, the average annual catch of anchovy and sardine was 580 000 t (Crawford et al. 1987), whereas it was estimated that seabirds in the Benguela ecosystem consumed about 155 000 t of these species each year (Crawford et al. 1991). At Robben Island, the proportion of adult African Penguins that bred and their breeding success were both significantly related to fish biomass (Crawford et al. 1999; Crawford et al. 2006), indicating the possibility of coupling models of seabirds and fish stocks to gauge levels of escapement of fish that are necessary to maintain seabird populations (Crawford 2004). Such reproductive parameters are likely to respond more rapidly to variations in food availability than are population sizes (e.g. Cairns 1987). The models that are used to advise total allowable catches (TACs) for anchovy and sardine in South Africa are at present being modified to incorporate a model of African Penguins and functional relationships linking penguins and the fish stocks. For example, the breeding success of penguins, and its variation, can be estimated for a certain level of fish abundance and used to predict, in Monte Carlo style, the probability of the African Penguin population falling below specified values over a certain time frame. Hence, the risk for the penguin population of alternative management strategies for fish stocks may be ascertained. Unlike the earlier situation in Namibia, when the range of sardine contracted to the north as the stock collapsed (Crawford et al. 1987), the eastward displacement of sardine off South Africa occurred when sardine was at a high level of abundance. High TACs continued to be allocated but, because most fish factories were in the Western Cape, substantial catches continued to be made in the west (Fairweather et al. 2006), further depleting the abundance of prey around seabird breeding colonies in the Western Cape.

The desirability of closing fishing areas around seabird breeding colonies, and means of measuring the success of such interventions, are at present being discussed by scientists charged with advising on management of the South African purse-seine fishery. Another issue that is likely to be addressed in the South African context is the establishment of target levels for seabird populations (e.g. Underhill and Crawford 2005). Targets will be chosen so as to reduce the risk of extinction of species (Crawford et al. 2001; Crawford 2004). A final matter that merits consideration is the definition of exceptional ecosystem circumstances. In the operational management procedure (OMP) used to advise TACs for sardine and anchovy in South Africa, circumstances are defined when a departure from the procedure may be justified, e.g. when a resource falls to a very low level. Similarly, it will be advisable to define ecosystem circumstances that will allow deviation from the OMP. Acknowledgements – I am grateful to South Africa’s Department of Environmental Affairs and Tourism for supporting this work. The study was conducted under permit issued by this Department. Financial support was provided by Earthwatch Institute and the Marine Living Resources Fund. CapeNature, Ministry of Fisheries and Marine Resources (Namibia), Robben Island Museum, South African National Parks and South African Navy provided logistical support. I thank all who assisted with counts of the three seabirds that are considered in this paper, especially D.A.E. Crawford, P.B. Crawford, P.J.M. Crawford, J.H.M. David, B.L. Dundee, B.M. Dyer, P.G.H. Kotze, M.A. Meÿer and L. Upfold. I am grateful to R. Cloete, J.C. Coetzee, G. Dalmeida, T. Fairweather and J. van der Westhuizen for making available information on sardine and anchovy, and those colleagues who have through discussion assisted with the formulation of ideas. I am grateful to Professors Ens and Furness for arranging the symposium Responses of birds to (over)fishing, at which this paper was presented. The paper is a contribution to the project LMR/ EAF/03/02 of the Benguela Current Large Marine Ecosystem (BCLME) Programme.

References Armstrong MJ, Shelton PA, Prosch RM, Grant WS (1983) Stock assessment and population dynamics of anchovy and pilchard in ICSEAF Division 1.6 in 1982. Collection of Scientific Papers International Commission for the Southeast Atlantic Fisheries 10(1):7–25 Berruti A, Underhill LG, Shelton PA, Moloney C, Crawford RJM (1993) Seasonal and interannual variation in the diet of two colonies of the Cape Gannet (Morus capensis) between 1977–78 and 1989. Colonial Waterbirds 16:158–175 Boyer DC, Hampton I (2001) An overview of the living marine resources of Namibia. South African Journal of Marine Science 23:5–35 Cairns DK (1987) Seabirds as indicators of marine food supplies. Biological Oceanography 5:261–271 Cooper J, Brooke RK, Shelton PA, Crawford RJM (1982) Distribution, population size and conservation of the Cape Cormorant Phalacrocorax capensis. Fisheries Bulletin South Africa 16:121– 143 Crawford RJM (1991) Factors influencing population trends of some abundant vertebrates in sardine-rich coastal ecosystems. South African Journal of Marine Science 10:365–381 Crawford RJM (1998) Responses of African Penguins to regime changes of sardine and anchovy in the Benguela system. South African Journal of Marine Science 19:355–364 Crawford RJM (1999) Seabird responses to long-term changes of prey resources off southern Africa. In: Adams NJ, Slotow RH (eds). Proceedings of 22nd International Ornithological Congress, Durban, 1998. BirdLife South Africa, Johannesburg, pp 688–705 Crawford RJM (2004) Accounting for food requirements of seabirds in fisheries management – the case of the South African purseseine fishery. African Journal of Marine Science 26:197–203 Crawford RJM, Allwright DM, Heyl CW (1992) High mortality of Cape Top Predators of the Benguela System

233

Cormorants (Phalacrocorax capensis) off western South Africa in 1991 caused by Pasteurella multocida. Colonial Waterbirds 15(2):236–238 Crawford RJM, Barham PJ, Underhill LG, Shannon LJ, Coetzee JC, Dyer BM, Leshoro TM, Upfold L (2006) The influence of food availability on breeding success of African Penguins Spheniscus demersus at Robben Island, South Africa. Biological Conservation 132(1):119–125 Crawford RJM, Cruickshank RA, Shelton PA, Kruger I (1985) Partitioning of a goby resource amongst four avian predators and evidence for altered trophic flow in the pelagic community of an intense, perennial upwelling system. South African Journal of Marine Science 3:215–228 Crawford RJM, David JHM, Shannon LJ, Kemper J, Klages NTW, Roux J-P, Underhill LG, Ward VL, Williams AJ, Wolfaardt AC (2001) African Penguins as predators and prey – coping (or not) with change. South African Journal of Marine Science 23:435–447 Crawford RJM, Dundee BL, Dyer BM, Klages NTW, Meÿer MA, Upfold L (2007) Trends in numbers of Cape Gannets (Morus capensis), 1956/57–2005/06, with a consideration of the influence of food and other factors. ICES Journal of Marine Science 64:169–177 Crawford RJM, Dyer BM (1995) Responses by four seabirds to a fluctuating availability of Cape anchovy Engraulis capensis off South Africa. Ibis 137:329–339 Crawford RJM, Ryan PG, Williams AJ (1991) Seabird consumption and production in the Benguela and western Agulhas ecosystems. South African Journal of Marine Science 11:357–375 Crawford RJM, Shannon LJ, Whittington PA (1999) Population dynamics of the African Penguin at Robben Island. Marine Ornithology 27:135–143 Crawford RJM, Shannon LV, Pollock DE (1987) The Benguela ecosystem. Part IV. The major fish and invertebrate resources. Oceanography and Marine Biology Annual Review 25:353–505 Crawford RJM, Williams AJ, Hofmeyr JH, Klages NTW, Randall RM, Cooper J, Dyer BM, Chesselet Y (1995) Trends of African Penguin Spheniscus demersus populations in the 20th century. South African Journal of Marine Science 16:101–118 Cury P, Bakun A, Crawford RJM, Jarre A, Quiñones RA, Shannon LJ, Verheye HM (2000) Small pelagics in upwelling systems: patterns of interaction and structural changes in “wasp-waist” ecosystems. ICES Journal of Marine Science Symposium Edition 57(3):603–618 Fairweather TP, van der Lingen CD, Booth AJ, Drapeau L, van der Westhuizen JJ (2006) Indicators of sustainable fishing for South African sardine (Sardinops sagax) and anchovy (Engraulis encrasicolus). African Journal of Marine Science 28:661–680 Grémillet D, Dell’Omo G, Ryan PG, Peters G, Ropert-Coudert Y, Weeks S (2004) Offshore diplomacy, or how seabirds mitigate intra-specific competition: a case study based on GPS tracking of Cape Gannets from neighbouring colonies. Marine Ecology Progress Series 268:265–279. Hampton I (1987) Acoustic study on the abundance and distribution of anchovy spawners and recruits in South African waters. South African Journal of Marine Science 5:901–917 Heath RGM, Randall RM (1989) Foraging ranges and movements of Jackass Penguins (Spheniscus demersus) established through radio telemetry. Journal of Zoology 217:367–379 Hockey PAR, Dean WRJ, Ryan PG (2005) Roberts Birds of Southern Africa. VII th ed. John Voelcker Bird Book Fund, Cape Town. Namibia. PhD Thesis, University of Cape Town Kemper J (2006) Heading towards extinction? Demography of the African Penguin in Namibia. PhD Thesis, University of Cape Town Klages NTW, Willis AB, Ross GJB (1992) Variability in the diet of the Cape Gannet at Bird Island, Algoa Bay, South Africa. South African Journal of Marine Science 12:761–771 Kreiner A, van der Lingen CD, Fréon P (2001) A comparison of condition factor and gonadosomatic index of sardine Sardinops

234


sagax stocks in the northern and southern Benguela upwelling ecosystems, 1984–1999. South African Journal of Marine Science 23:123–134 Le Clus F (1985) Effect of a warm water intrusion on the anchovy fishery off Namibia: 1984. Collection of Scientific Papers International Commission for the Southeast Atlantic Fisheries 12(1):99–106 Le Clus F (1986) Aftermath of environmental perturbations off Namibia relative to the anchovy stock, and comparative pilchard spawning. Collection of Scientific Papers International Commission for the Southeast Atlantic Fisheries 13(2):19–26 Makhado AB, Crawford RJM, Underhill LG (2006) Impact of predation by Cape Fur Seals Arctocephalus pusillus pusillus on Cape Gannets Morus capensis at Malgas Island, Western Cape, South Africa. African Journal of Marine Science 28:681–687 Marks MA, Brooke RK, Gildenhuys AM (1997) Cape Fur Seal Arctocephalus pusillus predation on Cape Cormorants Phalacrocorax capensis and other birds at Dyer Island, South Africa. Marine Ornithology 25: 9–12 Newman GG (1970) Stock assessment of the pilchard Sardinops ocellata at Walvis Bay, South West Africa. Investigational Report, Sea Fisheries Research Institute, South Africa 85:1–13 Petersen SL, Ryan PG, Gremillet D (2006) Is food availability limiting African Penguins at Boulders?: a comparison of foraging effort at mainland and island colonies. Ibis 147:14–26. Rand RW (1963a). The biology of guano-producing seabirds. 4. Composition of colonies on the Cape islands. Investigational Report, Sea Fisheries Research Institute, South Africa 43:1–32 Rand RW (1963b). The biology of guano-producing seabirds. 5. Composition of colonies on the South West African islands. Investigational Report, Sea Fisheries Research Institute, South Africa 46:1–26 Randall RM, Ross GJB (1979) Increasing population of Cape Gannets on Bird Island, Algoa Bay, and observations on breeding success. Ostrich 50:168–175 Schwartzlose RA, Alheit J, Bakun A, Baumgartner TR, Cloete R, Crawford RJM, Fletcher WJ, Green-Ruiz Y, Hagen E, Kawasaki T, Lluch-Belda D, Lluch-Cota SE, MacCall AD, Matsuura Y, Nevarez-Martinez MO, Parrish RH, Roy C, Serra R, Shust KV, Ward MN, Zuzunaga JZ (1999) Worldwide large-scale fluctuations of sardine and anchovy populations. South African Journal of Marine Science 21:289–347 Shelton PA, Crawford RJM, Cooper J, Brooke RK (1984) Distribution, population size and conservation of the Jackass Penguin Spheniscus demersus. South African Journal of Marine Science 2:217–257 Thomas RM (1985) Age studies on pelagic fish in the South-East Atlantic, with particular reference to the South West African pilchard, Sardinops ocellata. PhD thesis, University of Cape Town Thomas RM (1986) The Namibian pilchard: the 1985 season, assessment for 1952–1985 and recommendations for 1986. Collection of Scientific Papers International Commission for the Southeast Atlantic Fisheries 13(2):243–269 Underhill LG, Crawford RJM (2005) Indexing the health of the environment for breeding seabirds in the Benguela system. ICES Journal of Marine Science 62:360–365 Underhill LG, Crawford RJM, Wolfaardt AC, Whittington PA, Dyer BM, Leshoro TM, Ruthenberg M, Upfold L, Visagie J (2006) Regionally coherent trends in colonies of African Penguins Spheniscus demersus in the Western Cape, South Africa, 1987– 2005. African Journal of Marine Science 28:697–704 Waller LJ, Underhill LG (2007). The management of avian cholera Pasteurella multocida outbreaks on Dyer Island, South Africa, 2002–2005. African Journal of Marine Science 29:105–111 Williams AJ, Ward VL (2002). Catastrophic cholera: coverage, causes, context, conservation and concern. Bird Numbers 11(2):2–6

Chapter 30 Influences of the abundance and distribution of prey on African Penguins Spheniscus demersus off western South Africa Robert J.M. Crawford1, 2, *, Les G. Underhill2, Janet C. Coetzee1, Tracey Fairweather2, Lynne J. Shannon2 and Anton C. Wolfaardt2, 3 1

Department of Environmental Affairs and Tourism, Marine and Coastal Management, Private Bag X2, Rogge Bay 8012, South Africa 2 Avian Demography Unit, Department of Statistical Sciences, University of Cape Town, Rondebosch 7701, South Africa 3 CapeNature, Private Bag X5014, Stellenbosch 7599, South Africa * Corresponding author: [email protected]

Off South Africa, anchovy Engraulis encrasicolus and sardine Sardinops sagax are the main prey of African penguins Spheniscus demersus . The combined spawner biomass of these fish species increased from less than one million t in 1996 to more than nine million t in 2001 and then decreased to four million t in 2005. The combined biomass of young-of-the-year of these species increased from 0.2 million t in 1996 to 3.2 million t in 1991 before falling to 0.4 million t in 2005. There was a large eastward shift in the distribution of sardine between 1999 and 2005. The number of African penguins breeding in the Western Cape Province increased from 18 000 pairs in 1996 to more than 30 000 pairs from 2001–2005 before falling to 21 000 pairs in 2006, as the availability of fish decreased near breeding localities. Numbers of penguins breeding and numbers of birds in adult plumage moulting were significantly correlated with the

young-of the-year biomass of anchovy and sardine and with the available biomass of older sardine. The increase in the number of penguins breeding was mainly attributable to a greater proportion of mature birds breeding and improved breeding success. The decrease probably resulted from high mortality. Delayed breeding and abstinence from breeding during periods of food shortage may both increase survivorship when food is scarce and enable seabirds rapidly to take advantage of improved feeding conditions. Although long-lived seabirds are buffered against short-term variability in food supplies, environmental change that influences the abundance and availability of prey can have severe consequences for central-place foragers, such as penguins, if prey is displaced to regions where no suitable breeding localities occur.

Keywords: African penguin; breeding proportion; breeding success; food; mortality; Spheniscus demersus

Introduction Seabirds often have life-history characteristics, such as low fecundity, high survivorship and extended longevity, which buffer their populations against inter-annual fluctuations in their food sources (Hunt et al., 1996). However, there are longer-term trends in the supply of food, which pose a much greater challenge. For instance, around the world, the abundance of anchovies Engraulis spp. and sardine (pilchard) Sardinops spp. has often remained at high or low levels over extended periods (Barange et al., 1999, Schwartzlose et al., 1999). Such long-term fluctuations in fish populations have been termed regimes (Lluch-Belda et al., 1989) and it is of interest to consider how seabirds respond to regime changes in the abundance of their prey. Off South Africa, opportunity to examine such responses arose when there was a large increase in the abundance of anchovy E. encrasicolus and sardine S. sagax in the late 1990s (Fairweather et al., 2006b), followed by a rapid decrease after 2002. These two fish species dominate the diet of African penguins Spheniscus demersus in South Africa (Hockey et al., 2005). There was also an eastward shift in the distribution of sardine-directed fishing (Fairweather et al.,

2006b), which implies altered availability to penguins off South Africa’s west coast. Fishing effort is often used as an index of relative abundance (King, 1995) and thus it can be argued that the distribution of fishing effort is an index of availability. African penguins are endemic to southern Africa, where they breed in three regions: southern Namibia, South Africa’s Western Cape Province and Nelson Mandela (formerly Algoa) Bay in the Eastern Cape Province (Shelton et al., 1984). Localities where penguins breed in the Western Cape lie about 500 km south of those in Namibia and 600 km west of those in the Eastern Cape. In this paper we consider the responses of penguins in the Western Cape to an altered availability of their prey in this region. Methods From 1987–2006, African penguins bred at 13 localities in South Africa’s Western Cape Province. The colony at De Hoop was established in 2003 (Underhill et al., 2006; Fig. 1). Annual counts were made of the number of active nest sites of penguins at these localities during the main breeding season, which is February–September (Crawford et al., 1995a; Top Predators of the Benguela System

235

At Robben Island, there is a peak in the numbers moulting between November and early January, with small numbers moulting at other times of the year (Underhill and Crawford, 1999). The feather-shedding phase of moult, from the time the first feathers stand out until the last loose feathers fall away, has a mean duration of 12.7 d (standard deviation 1.4 d, n = 45, Randall et al. 1986). Counts were interpolated linearly to estimate numbers in moult for each day between actual counts (Underhill and Crawford, 1999). These interpolated counts were summed for the split year 1 July–30 June and divided by 12.7 to estimate the number of birds at Robben Island moulting in each split year: ARIt, where t = the second of the two years in the split year. An index of the proportion of adults breeding at the island (Pt) was calculated as: Pt = 2NRI t /ARI t,

Figure 1: The location of breeding colonies of African penguins in South Africa’s Western Cape Province. The inset shows the locations of Namibia, the Eastern Cape Province and Nelson Mandela Bay

1995b). Of the 244 possible counts, 46 were not made, most at smaller localities that are difficult to access. These data gaps were filled by linear interpolation between previous and subsequent counts at the same locality. Numbers that were obtained from interpolation of information contributed 14% of the overall number estimated to be breeding in the 20-year period (Underhill et al., 2006). In 1991 and 2005, all the extant colonies were counted. A nest site was considered active if it contained eggs or chicks, or if it was defended by an adult bird (Crawford et al., 1990). Numbers of chicks in crèches were divided by two to estimate the number of nest sites they represented, with remainders taken to signify an additional site – for example crèches of five and six chicks were both taken to represent three nests (Shelton et al., 1984). Counts from 1987–2005 have been reported by Underhill et al. (2006). At Robben Island, one of the breeding localities in the Western Cape, counts were made along the coast of penguins in adult plumage in the feather-shedding phase of moult at approximately two-weekly intervals, commencing October 1988. At this locality, most penguins moult along the coastline but some, which were not counted, moult inland at breeding sites. It was assumed that the proportion of birds moulting along the coast remained constant during 1988– 2006. African penguins moult annually (Randall et al., 1986).

Table 1: Models used to obtain prewhitened residuals of time series that were cross correlated – ar(x) and ma(x) indicate autoregressive and moving average models of order x Time series African penguin breeders in the Western Cape (pairs) African penguin breeders at Robben Island (pairs) African penguin adults moulting at Robben Island Anchovy recruits (biomass) Anchovy spawners (biomass) Sardine recruits (biomass) Sardine spawners (biomass) Anchovy and sardine recruits (biomass) Anchovy and sardine spawners (biomass) Sardine availability (biomass) 236


Model ar(1) ma(1) ar(1) ma(1) ar(1) ar(2) ar(1) ma(1) ar(1) ar(1) ar(2) ar(1) ma(1) ar(1) ar(2) ar(1) ma(1) ar(1) ar(2)

where NRI = the number of pairs breeding at Robben Island. Pt is an index and not the actual proportion for two reasons. Firstly, counts of moulting birds were only undertaken around the coast. Secondly, African penguins moult to adult plumage when about 18 months (Randall, 1989) but many breed for the first time when aged four years, although occasional breeding has been reported at two years old (Whittington et al., 2005a). Hydro-acoustic surveys to estimate the biomass of anchovy and sardine spawning off South Africa were conducted each year in November from 1984–2005. Similar surveys to estimate the abundance of young-of-the-year of these fish species were conducted in May. The surveys covered the known areas of spawning and recruitment. The methods used in the surveys and sampling procedures have been described by Hampton (1987) and Barange et al. (1999). After 1998 there was a large eastward shift in the distribution of catches of sardine off South Africa (Fairweather et al., 2006b). By 2005 the majority of sardine-directed fishing (and by supposition sardine) was placed to the east of breeding localities of African penguins in the Western Cape, changing their availability to penguins. The longitudinal coordinate for the centre of gravity of the distribution of commercial catches from 1987–2005 (Fairweather et al., 2006b) was standardized to fall between 1 (the most western distribution observed) and 0 (the most eastern distribution) in order to produce an index of W–E distribution. The biomass of spawning sardine in the previous November was multiplied by this standardized index to obtain an estimate of the spawner biomass of sardine (ASS, million t) available to penguins in each year. The foraging range of African penguins while breeding in the Western Cape is about 20–40 km (Hockey et al., 2005; Petersen et al., 2005). After breeding is complete and after moult, penguins at Western Cape breeding localities may travel considerably farther to fatten up, including towards the easternmost distribution observed for sardine catches (Barham et al., 2006). When sardine catches were at their westernmost distribution, they were centred in the vicinity of the main African penguin colonies off the Western Cape. Because anchovy catches maintained their western distribution throughout the study period (Fairweather et al., 2006b), its entire biomass was deemed to be available to penguins in the Western Cape in all years. The relationships between the numbers of penguins breeding or moulting in the Western Cape and the biomass of spawning and recruiting anchovy and sardine and the availability of sardine were investigated using correlation analysis. Cross correlation was undertaken with “prewhitened residuals” of time series (Box and Jenkins, 1970), from which autocorrelation was removed using the software programme EViews (Quantitative Micro Software, 2000). The

Table 2: Models use to investigate trends in the number of African penguins breeding in the Western Cape, 1987–2006 (t = year; NWC = number of pairs breeding in the Western Cape; RC = combined recruit biomass of anchovy and sardine; ASS = available spawner biomass of sardine) Model NWC t = NWC t = NWC t = NWC t =

Durbin–Watson statistic 8.21 8.43 7.09 5.31

+ + + +

0.39NWC t–1 + 2.53RC t–1 + 5.31ASS t–1 0.45NWC t–1 + 4.66RC t–1 0.45NWC t–1 + 7.83ASS t–1 0.77NWC t–1

2.021 1.911 2.123 1.540

Akaike Information Criterion 5.106 5.439 5.270 6.088

Adjusted r2 0.852 0.785 0.818 0.570

models employed to obtain the prewhitened residuals are shown in Table 1. Because most African penguins in the Western Cape commence breeding early in the year, with some pairs establishing territories in January, and stop breeding before November (Crawford et al., 1995a; 1995b), when the surveys to estimate the biomass of spawning fish are undertaken, the numbers of breeders or adults were compared to the biomass or availability of fish in the preceding year. Based on the results of the correlation analysis, the number (thousand of pairs) of penguins breeding in the Western Cape in year t (NWCt) was modelled using equations of the form: NWCt = a + bNWCt–1 + cRCt–1 + dASS t–1, where a, b, c and d are constants and RC is the combined biomass (million t) of young-of-the-year anchovy and sardine. The models used are listed in Table 2. For each model, the Durbin–Watson statistic, the Akaike Information Criterion (AIC) and the adjusted r2 value were calculated. The Durbin– Watson statistic measures the serial correlation in the residuals. If there is no serial correlation, it will have a value close to two. The smaller values of the AIC are preferred in model selection. The r 2 value indicates the proportion of variation in Nt accounted for by the model. Results The number of African penguins estimated to be breeding in South Africa’s Western Cape Province decreased from about 23 000 pairs in 1987 and 1988 to 13 000 pairs in 1993. It then increased to 38 000 pairs in 2004 before falling to 21 000 pairs in 2006 (Fig. 2). There were large increases from 18 000 pairs in 1998 to 24 000 pairs in 1999 and from 26 000 pairs in 2000 to 34 000 pairs in 2001. At Robben Island, the number of penguins breeding increased from about 500 pairs in 1987 to 8 500 pairs in 2004 and then decreased to 3 700 pairs in 2006 (Fig. 2). The number of birds in adult plumage that moulted along the coast was about 3 500 in 1987, 17 500 in 2004 and 7 800 in 2006. The index of the proportion of adult penguins breeding increased after 1995 (Fig. 2). It averaged 0.68 (standard deviation 0.12) from 1989–1995 and 0.97 (standard deviation 0.11) from 1996–2006. The means of the index in these two periods were significantly different from each other (t16 = 4.54, P < 0.01). The spawner biomass of anchovy off South Africa fell from 2.5 million t in 1986 to less than one million t in 1989, 1990 and from 1994–1996. It rose rapidly to almost seven million t in 2001 before decreasing to three million t in 2005 (Fig. 2). The spawner biomass of sardine increased from 0.2 million t in 1986 to more than four million t in 2002 and then decreased to one million t in 2005. The biomass of young-ofthe-year anchovy and sardine increased from less than one million t during 1986–1999 to three million t in 2001, thereafter falling to 0.4 million t in 2005. The biomass of mature

Figure 2: Trends in (a) numbers of African penguins breeding in the Western Cape Province, (b) numbers of penguins breeding, penguins in adult plumage moulting and an index of the proportion of mature penguins breeding, (c) the spawner biomass of anchovy, sardine and these two species combined, (d) the recruit biomass of anchovy, sardine and these two species combined and (e) the biomass of mature sardine available to penguins in the Western Cape

sardine gauged to be available to penguins in the Western Cape increased from 0.2 million t in 1986 to 1–2 million t from 1998–2004, but was negligible in 2005 (Fig. 2). The results of cross correlation of residuals of prewhitened time series are shown in Table 3. The number of Top Predators of the Benguela System

237

penguins breeding at Robben Island was significantly correlated with the number breeding in the Western Cape and with the number of adults moulting at Robben Island immediately before the breeding season. In all 21 comparisons made, the number of penguins breeding in the Western Cape or at Robben Island, or the number of adults moulting at Robben Island, was positively related to the biomass of fish in the previous year. In only one of nine comparisons was the relationship between penguins and biomass of spawning fish significant. However, in seven of nine comparisons the relationship between penguins and biomass of recruiting fish was significant. All three comparisons between numbers of penguins and the available biomass of mature sardine were significant. The number of penguins breeding in the Western Cape was best modelled by an equation that incorporated estimates for the previous year of the number of penguins breeding, the combined abundance of young-of-the-year anchovy and sardine and the biomass of mature sardine available to penguins (Table 2). This model accounted for 85% of the observed variation in the number of penguins breeding. The Durbin–Watson statistic was close to two and the value of the AIC was less than for any other model investigated. Discussion Responses to an altered availability of food From 1996–2001, an increase in the biomass of anchovy and sardine off South Africa from less than one million t to nine million t was followed by a large increase in the numbers of

Table 3: Results of cross correlation between pairs of residuals from pre-whitened time series (t = year; NWC = number of pairs breeding in the Western Cape; NRI = number of pairs breeding at Robben Island; ARI = number of adults moulting at Robben Island; SA = spawner biomass of anchovy; SS = spawner biomass of sardine; SC = combined spawner biomass of anchovy and sardine; RA = recruit biomass of anchovy; RS = recruit biomass of sardine; RC = combined recruit biomass of anchovy and sardine; ASS = available spawner biomass of sardine; Ns = not significant) Series 1

Series 2

n

r

P

NWCt NRI t

NRI t ARIt

19 16

0.804 0.810