See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/248900198

Trends in numbers of Cape ( capensis ) over a 50-year period, 1956–57 to 2006–07

ARTICLE in THE EMU: OFFICIAL ORGAN OF THE AUSTRALASIAN ORNITHOLOGISTS' UNION · JANUARY 2007 Impact Factor: 1.11 · DOI: 10.1071/MU07015

CITATIONS READS 23 36

5 AUTHORS, INCLUDING:

Robert J. M. Crawford Jessica Kemper South Government University of Cape Town

193 PUBLICATIONS 4,431 CITATIONS 24 PUBLICATIONS 454 CITATIONS

SEE PROFILE SEE PROFILE

Robert Simmons University of Cape Town

103 PUBLICATIONS 925 CITATIONS

SEE PROFILE

All in-text references underlined in blue are linked to publications on ResearchGate, Available from: Robert Simmons letting you access and read them immediately. Retrieved on: 14 January 2016 CSIRO PUBLISHING www.publish.csiro.au/journals/emu Emu, 2007, 107, 253–261

Trends in numbers of Cape Cormorants (Phalacrocorax capensis) over a 50-year period, 1956–57 to 2006–07

Robert J. M. Crawford A,B,E, Bruce M. Dyer A, Jessica Kemper C, Robert E. Simmons D and Leshia Upfold A

ADepartment of Environmental Affairs and Tourism, Marine and Coastal Management, Private Bag X2, Rogge Bay 8012, South Africa. BAvian Demography Unit, Department of Statistical Sciences, University of Cape Town, Rondebosch 7701, South Africa. CAfrican Penguin Conservation Project, PO Box 586, Lüderitz, Namibia. DPercy FitzPatrick Institute, University of Cape Town, Rondebosch 7701, South Africa. ECorresponding author. Email: [email protected]

Abstract. The population trend of Cape Cormorants (Phalacrocorax capensis), a endemic to southern Africa and that feeds mainly on shoaling pelagic fish, is described for a 50-year period, from 1956–57 to 2006–07. The main breeding localities for the species are grouped in three regions in the Benguela upwelling ecosystem: guano platforms off central Namibia, islands off southern Namibia and islands off South Africa’s Western Cape Province. From 1956–57 to 1978–79, the numbers breeding off Namibia increased, as a result of increased availability of breeding space and adequate supplies of food. In the same period, numbers remained stable in the Western Cape. Numbers decreased off southern Namibia in the early 1980s and off central Namibia in the early 1990s, when environmental perturbations reduced the availability of food. Numbers decreased in the Western Cape in the early 1990s, following periods of scarcity of (Engraulis encrasicolus), an important prey item, and an outbreak of avian cholera caused by the bacterium Pasteurella multocida. They remained low as cholera outbreaks continued and some pelagic fish were displaced to the east beyond the foraging range of breeding Cormorants. The overall population of Cape Cormorants was of the of 100000 pairs in 1956–57, increased to ~250000 pairs in the 1970s, but reverted to ~100000 pairs in 2005–06.

Additional keywords: avian cholera, environmental perturbation, food, guano, population trend.

Introduction The Cape (Phalacrocorax capensis) is the most The Cape Cormorant is regarded as Near Threatened owing numerous of four cormorants that breed in the Benguela to a decrease in the breeding population from 277000 pairs in upwelling ecosystem off south-western Africa (Crawford et al. 1977–81 (Cooper et al. 1982) to ~72000 pairs in 1996 (Barnes 1991). It is endemic to the Benguela ecosystem, where its present 2000; Simmons and Brown 2007). In order to investigate breeding range extends from Ilha dos Tigres, southern Angola, to further trends in the abundance of the species, we collate infor- Nelson Mandela Bay, Eastern Cape Province, South Africa mation on numbers breeding over a 50-year period from (Cooper et al. 1982; Dean et al. 2002; Fig. 1). Its usual non-breed- 1956–57 to 2006–07. We consider various factors that may have ing range is from Lobito, Angola around southern Africa to influenced trends in the numbers of Cape Cormorants, includ- Delagoa Bay, Moçambique. Vagrant have been recorded as ing the availability of food, which may be altered by both far north as Gabon on theWestAfrican coast (Hockey et al. 2005). fishing and environmental change (Crawford et al. 2007), The Cape Cormorant forages in sub-surface marine waters availability of breeding space (Cooper et al. 1982), disease around the coast by pursuit diving, often in flocks (Rand 1960), (Crawford et al. 1992a) and predation (Marks et al. 1997), and though it also sometimes enters and forages in lagoons (Hockey also compare trends in the populations of Cape Cormorants and et al. 2005). Its main prey is anchovy (Engraulis encrasicolus), Cape . sardine (Sardinops sagax), horse mackerel (Trachurus trachu- rus) and pelagic goby (Sufflogobius bibarbatus). It also eats Methods several other fish species and crustaceans (Hockey et al. 2005). Estimating trends in the abundance of Cape Cormorants is dif- Together with the Cape (Morus capensis), it is the main ficult for several reasons. The species has bred at 70 localities producer of seabird guano in southern Africa (Crawford and and it breeds throughout the year (Hockey et al. 2005), although Shelton 1978). From 1930 to 1963, four platforms were con- seasons of intensive breeding may be distinguished (Berry structed in northern Namibia to facilitate the collection of guano 1976; Crawford et al. 1999b). Replacement laying may occur deposited by Cape Cormorants (Cooper et al. 1982). after nesting failure and successful breeders sometimes relay in

© Royal Australasian Ornithologists Union 2007 10.1071/MU07015 0158-4197/07/040253 254 Emu R. J. M. Crawford et al. the same season (Hockey et al. 2005). However, breeding may described in Hockey et al. (2005). The mean size of clutches is occur in waves, with different birds arriving at colonies at 2.82 eggs (Berry 1976). When unattended chicks were found in different times (Berry 1976). Therefore, we concentrate on 16 crèches away from nests, the number was divided by three to localities, which in 1977–81 supported more than 99% of the estimate the number of nesting sites they represented, as nests at total population (Cooper et al. 1982) and have been monitored which these chicks were reared would not have been counted. more consistently than smaller colonies: the four guano plat- Remainders were taken to represent further sites (Shelton et al. forms (Cape Cross North, Cape Cross Central, Swakopmund 1982). At breeding localities that were visited frequently, old and Rock) off central Namibia (74471 pairs, or 27% of the nests were not counted owing to the possibility that birds may population in 1977–81); Mercury, Ichaboe, Seal (Lüderitz), breed twice in a season (Hockey et al. 2005). At localities that Penguin (Lüderitz), North Reef and Possession Islands off were seldom visited, all nests were counted because old nests are southern Namibia (95654 pairs, 35%); and Lamberts Bay, unlikely to remain intact for more than a year owing to prevail- Malgas, Jutten, Vondeling, Dassen and Dyer Islands off South ing strong winds. At Dyer Island, repeat counts of two groups of Africa’s Western Cape (104741 pairs, 38%). We additionally breeding Cormorants were made in 1985 to estimate the coeffi- present information for some smaller colonies that were moni- cient of variation (CV) of the counts. tored sporadically during the 50-year period. Because breeders Photographs, taken vertically from an aeroplane, were often may have been absent at the time counts were conducted, where used to estimate numbers at localities where large numbers of feasible we have applied a smoothing algorithm (Underhill et al. Cape Cormorants breed, or which could not be easily accessed. 2006) to gauge trends from estimates of abundance. On the photographs, counts were made of Cormorant nests, Counts of active nesting sites of Cape Cormorants were which can be readily distinguished as black spots against a made on visits to breeding localities. Nesting sites were defined lighter background, whereas individual birds appear as smaller as paired birds defending a site, sites showing evidence of recent black specks. In general, photographs were taken as close to nest-construction, and sites with eggs or chicks. Nests were midday as possible to minimise the effect of shadows. On aerial identified as belonging to Cape Cormorants through identifica- photographs, the CV on a count of a colony of 3703 nests was tion of birds present at the nests, or using nest-characteristics as 3.6% (Shelton et al. 1982). On photographs of ,

S

ANGOLAANGOLA Lobito 15° Mercury Is. Ilha dos Tigres Cunen 26˚ e River

Ichaboe Is. NAMIBIA 20° Cape Cross North Seal Is. Penguin Is. Cape Cross Central Lüderitz North Reef Swakopmund 27˚ Bird Rock Possession WALVIS Is. BAY Pomona Is. 25° MOÇAMBIQUE Fig. 1. Southern Africa Plumpudding Is. Delgoa showing the breeding localities Sinclair Is. Bay of Cape Cormorants and other localities mentioned in the text. Or S an 15° E 33° g e River S Malgas Meeuw Is. 30° Is. Schaapen Jutten Is. SOUTH AFRICA Is. Vondeling Is. Lambert’s Bay 33° CAPE TOWN astern Cape e E stern Cap Dassen Is. We Nelson Mandela Bay Robben Is. 35° 18° E CAPE TOWN Cape Agulhas 34°

Dyer Is. 35° Breeding locality 18° 19° E

5° 10° 15° 20°25° 30° 35° E Trends in Cape Cormorants over 50 years Emu 255 where several species of cormorant nest in mixed colonies, platforms increased from 12000 pairs in 1956–57 to 74000 ground-truth information on the abundance of other species of pairs in 1978–79 (Fig. 2). It averaged 56000 pairs (s.d. = 12000, cormorant (e.g. Crawford et al. 1999a) was used to adjust n = 6) from 1988–89 to 1998–99 and 27000 pairs (s.d. = 5000, counts (Shelton et al. 1982). Counts for all localities up until n = 3) from 2002–03 to 2005–06. The Swakopmund platform 1981 have been reported by Rand (1963a, 1963b) and Cooper began operation in 1963 (Cooper et al. 1982). The first count of et al. (1982). Information for Lamberts Bay, Malgas, Jutten, the number of Cape Cormorants at this platform was made in Vondeling, Dassen and Dyer Islands for 1985–92 came from 1978–79, when ~44000 pairs bred there. Subsequently, except Crawford et al. (1992b) and Crawford and Dyer (1995). in 2002–03, more birds bred at this than at any other platform Cape Cormorants arrive at the Namibian platforms in July. (Table 1). Increasing daylength stimulates gonadal development from The number of Cape Cormorants breeding at the four plat- August and birds breed from October to February (Berry 1976). forms was significantly related to the production of guano at the In South Africa’s Western Cape, the peak of breeding is usually platforms in the same split year (r = 0.725, P < 0.02, n = 11). between September and February (Crawford et al. 1999b). When the outlying point for 1998–99 was excluded (Fig. 3), the Therefore, information was collated for split years beginning relationship improved (r = 0.944, P < 0.001, n = 10) and the 1 July and ending 30 June. If more than one count was available number of Cape Cormorants (N, thousand pairs) was related to for a locality in a split year, the highest count was taken as the the guano production (G, thousand t) by the equation: estimate of the number breeding in that split year. As far as pos- N = 24.165G – 11.867 sible, counts were conducted during the main breeding season of Cape Cormorants (September–February). Counts undertaken The low production of guano in 1998–99 may have resulted outside this period are indicated in Table 1. It is also indicated from early termination of breeding (e.g. Crawford et al. 1992b). whether counts were made by ground or aerial surveys. The smoothed production of guano at the platforms increased For 10 localities, counts were available in most split years for from 1394 t in 1956–57 to 3844 t in 1984–85 and then decreased the period since the late 1970s or mid 1980s. The missing counts to 1562 t in 2005–06, suggesting a similar trend in the number were filled by linear interpolation between actual counts made at of Cape Cormorants breeding at the platforms (Fig. 2). the locality. For these localities and periods, the counts and inter- polated counts were smoothed using an algorithm described by Southern Namibia Underhill et al. (2006). The weighted estimate is based almost At the six main breeding localities off southern Namibia, there entirely on the data from the target year and 3 years on either side were ~3000 pairs of Cape Cormorants in 1956–57 and ~70000 of it. The same algorithm was applied to the combined counts at pairs in the late 1970s (Fig. 2). Numbers at these six localities the six most important breeding localities for Cape Cormorants then decreased, and fluctuated around 30000 pairs from the in the Western Cape for the period 1985–86 to 2006–07. mid-1980s. Ichaboe Island was the most important breeding Information on the harvest of guano at the Namibian plat- locality for Cape Cormorants off southern Namibia – counts forms was updated from Schwartzlose et al. (1999). The com- exceeded 20000 pairs on 10 occasions, with a maximum of bined yield of the four platforms was smoothed using the same 46000 pairs. The smoothed trend for Ichaboe Island showed a algorithm that was applied to the time-series of counts, because decrease after 1983–84 (Fig. 2). Estimates of the number of in rare instances (12 of 190 possible cases) guano was not Cape Cormorants breeding at three small Namibian islands scraped in a season but allowed to accumulate until the next (Pomona, Plumpudding and Sinclair) to the south of Possession season. Banding has shown interchange of birds between the Island are available for eight seasons, for which the maximum Cape Cross platforms and the Swakopmund and Bird Rock plat- combined count was ~600 pairs in 1995–96, 2001–02 and forms (Rand 1963b; Berry 1976) and in 1974–75 birds aban- 2005–06. doned breeding at the Swakopmund platform and moved to the Western Cape platforms at Cape Cross (Crawford et al. 1980). Therefore, we considered the four guano platforms together. There were 11 At the six main localities for Cape Cormorants in the Western split years for which counts of breeding Cape Cormorants were Cape, ~100000 pairs bred in each of 1956–57, 1978–79 and made at all four platforms. For these split years, the number of 1988–89 (Fig. 2). The number decreased to just 8000 pairs in Cape Cormorants breeding at the platforms was regressed 1989–90 but increased to 90000 pairs in 1991–92. It then against the amount of guano collected at the platforms to inves- decreased again to 25000 pairs in 1994–95. From 1993–94 to tigate whether the guano yield could be used as a surrogate 2005–06, it fluctuated about a mean level of 30000 pairs index of the number of Cape Cormorants breeding. Because the (s.d. = 8000, n = 14). Smoothed trends at individual colonies are 11 complete counts extended over the 50-year period, with sub- shown in Fig. 4. At Robben Island, no Cape Cormorants were stantial gaps between most counts, no pre-whitening of residu- recorded breeding between 1956–57 and 1984–85 (Kriel et al. als was undertaken before cross correlation. 1980; R. J. M. Crawford, pers. obs.). Sporadic breeding then occurred until 2001–02, after which the colony increased (Table 1). Results At Dyer Island in 1985, for one group of breeding birds that Guano platforms was counted repeatedly the mean was 157 pairs (CV = 1.5%); Estimates of the number of Cape Cormorants breeding at selec- for a second group it was 1358 pairs (CV = 3.0%). It was esti- ted localities in split years for which information is available are mated that the overall count of the number of Cape Cormorants shown in Table 1. The overall number breeding at the four guano breeding at the island (48 000 pairs) had a CV of 2.6%. 256 Emu R. J. M. Crawford et al. 3.186 35.580 23.355 0.000 0.000 0.000 0.000 48.182 0.000 en Island in 1989–90 were en Island in 1989–90 were ing Dassen Robben Dyer 0.032 0.000 0.252 7.890 13.767 3.000 22.500 0.000 16.250 15.026 3.078 0.000 8.601 6–07 s Central and Bird Rock platforms 0.141 0.055 27.757 8.300 11.760 13.000 0.139 0.065 0.000 36.982 24.277 29.819 0.000 0.000 0.005 0.000 10.029 0.500 0.000 0.004 0.345 0.000 18.350 algas, Jutten, Dassen and Dyer Islands in 1956–57 assumed one bird to algas, Jutten, Dassen and Dyer 0.021 0.000 0.217 0.002 11.144 0.002 48.293 5.481 0.520 0.725 0.000 0.347 0.000 0.000 2.525 0.132 0.000 17.936 0.535 1.181 0.000 2.848 1.171 0.802 6.314 1.364 0.000 15.700 8.144 0.936 2.727 0.958 3.025 0.006 5.064 0.879 0.637 4.159 0.667 0.000 11.043 0.291 2.559 0.000 3.360 1.194 0.336 3.182 3.149 0.379 22.766 0.583 0.505 0.000 1.705 1.486 0.436 2.126 0.381 0.000 9.207 9.604 0.453 0.207 0.029 0.000 2.436 0.000 2.548 0.200 0.0020.107 0.000 2.206 3.255 0.000 2.742 0.557 0.326 1.449 0.173 0.616 25.964 1.446 0.000 0.050 0.000 0.214 0.264 0.107 10.717 0.008 0.000 14.556 13.275 11.573 0.577 3.381 3.932 0.000 7.252 1.577 1.406 5.018 2.062 0.003 18.436 0.198 0.005 30.977 14.547 0.594 islands & Sinclair 0.900 0.080 0.969 0.630 1.177 0.096 13.519 8.708 0.000 2.882 0.571 0.000 5.632 0.100 0.000 0.083 conducted outside the main breeding season 0.003 5.006 0.114 0.321 0.286 0.521 13.000 0.300 0.009 0.936 2.119 0.196 0.208 0.780 0.000 9.038 2.142 1.062 2.903 0.208 0.166 33.024 2.200 1.100 1.400 3.653 0.000 5.592 1.268 4.413 3.067 16.466 0.330 5.222 6.431 0.533 0.003 3.688 2.440 7.785 6.542 9.790 4.769 5.362 0.000 13.472 0.844 0.099 4.116 29.613 0.006 32.964 18.495 35.334 21.329 27.467 33.700 16.619 15.232 19.745 16.366 0.9851.1481.307 1.617 2.085 2.608 0.473 0.000 0.000 5.414 0.932 0.790 1.000 0.705 0.500 3.427 0.945 5.255 3.187 0.000 15.617 0.000 18.105 1.634 2.363 1.364 8.840 0.088 0.414 0.120 1.6650.810 22.698 14.318 1.5030.002 0.3102.834 3.014 18.2720.924 1.656 16.888 0.000 2.732 0.000 0.106 2.846 0.006 0.608 2.153 2.520 1.199 2.370 0.000 0.000 1.409 8.114 5.862 0.617 0.540 0.042 0.087 6.271 2.452 36.486 0.643 0.004 7.153 0.849 18.419 0.000 0.000 14.768 6.623 1.155 0.330 3.694 1.078 0.000 13.828 1.590 . (1982). Counts at Seal, Penguin, North Reef, Pomona, Plumpudding, Sinclair, Schaapen and Meeuw Islands in 1979–80 at Jutt Plumpudding, Sinclair, North Reef, Pomona, . (1982). Counts at Seal, Penguin, et al 0.924 ). The count at Vondeling Island in 1956–57 was mainly of disused nests. Counts of nests on photographs in 1956–57 of Cape Cros of disused nests. Counts nests on photographs mainly Island in 1956–57 was Vondeling The count at ). b , 1963 a 30.011 1.167 Table 1.Table at selected localities, breeding of Cape Cormorants (thousands of pairs) Counts of the number 1956–57 and 1978–79 to 200 7.058 12.173 5.274 Cape Cape Swakop- Bird Mercury Ichaboe Seal Penguin Possession North Pomona, Lamberts Malgas Marcus Jutten Schaapen Meeuw Vondel Meeuw Schaapen Jutten Marcus Malgas Lamberts Pomona, North Possession Penguin Seal Ichaboe Swakop-Cape Mercury Bird 0.029 12.500 6.250 5.865 1.276 0.400 1.234 0.250 Cross Cross mumdCross Rock Island Island Island Island Island Reef Plumpudding, Bay Island Island Island Island Island Island Island Island Island North Central 1999–00 2000–01 2001–02 2002–03 2004–05 1988–89 5.559 0.000 53.420 1998–99 6.908 0.000 56.208 3.942 2005–06 2003–04 5.577 27.649 Counts in ordinary font were made from aerial photographs; those in bold were ground counts. Counts at Ichaboe, Lamberts Bay, M ground counts. Counts at Ichaboe, Lamberts those in bold were made from aerial photographs; Bay, Counts in ordinary font were represent one pair (Rand 1963 1980–81 1987–88 1956–57 0.000 9.426 2.569 1.540 0.353 1.028 0.000 2006–07 1978–791979–80 4.944 25.6531981–82 43.542 0.332 5.890 21.695 21.457 18.292 1989–901990–91 45.805 3.071 38.636 26.781 1982–83 1983–84 1986–87 1984–851985–86 1991–92 1992–931993–94 0.0001994–95 0.000 0.0001995–96 0.000 0.0001996–97 50.871 0.0001997–98 55.927 0.000 3.000 1.855 0.438 0.000 0.131 32.041 31.920 8.677 24.201 1.554 0.000 13.000 2.081 and Mercury Island were from Cooper and Mercury Island were Trends in Cape Cormorants over 50 years Emu 257

Overall population of Cape Cormorants increased between 1956–57 and There are five split years in which counts of breeding pairs were 1978–79 and then decreased (Fig. 2). The increase resulted from made at each of the 16 most important localities for Cape more Cape Cormorants breeding at the platforms and in south- Cormorants. In these split years, the combined counts were ern Namibia, while the number in the Western Cape was not 111000 pairs in 1956–57, 247000 pairs in 1978–79, 157000 much changed. The overall number in Namibia increased from pairs in 1992–93, 98000 pairs in 2004–05 and 92000 pairs in 15000 pairs in 1956–57 to 143000 in 1978–79. There was a 2005–06. In 1985–86, counts were made at the 12 important decrease in numbers off southern Namibia between 1978–79 islands in southern Namibia and the Western Cape, but not at the and 1985–86, followed by decreases at the platforms and in the four platforms. If the relationship between number of birds Western Cape in the early 1990s. In 2005–06, the number breeding and guano production at the platforms is applied to the counted at the 16 main localities was 82% of that in 1956–57 guano produced at the platforms in that season (3945 t), there and 37% of that in 1978–79. would have been 83000 pairs at the platforms and 174000 pairs Discussion at the 16 main localities. Similarly, in 1995–96 there would have been 35000 pairs at the platforms (1952 t) giving a total of Reliability of estimates 92000 pairs for the 16 main localities. Therefore the breeding Cooper et al. (1982) estimated the overall population of Cape Cormorants in 1977–82 to be 277000 pairs, based on the highest count obtained at any locality in this 5-year period. In 1978–79, (a) Platforms some 247000 pairs bred at the 16 main localities, which repre- 80 4.0 Guano sented 90% of the estimate of Cooper et al. Birds may abstain 70 3.5 Count from breeding in years of unfavourable food supply (Crawford 60 3.0 50 2.5 and Dyer 1995) and not all birds breeding in a given year may 40 2.0 be present at the time of counting, so the numbers presented in 30 1.5 this paper are minimum estimates of the overall population. 20 1.0 The Namibian platforms have held large numbers of Cape

10 0.5 Guano (thousand t) Cormorants. Previous assessments have attempted to estimate the number based on the surface area of the platforms and den- 80 (b) Southern Namibia Ichaboe sities of birds (Rand 1963b; Berry 1976). As did Cooper et al. 70 S Namibia (1982), in this paper we use counts of nests to gauge the trend in 60 numbers breeding at the platforms. A reasonable correspon- 50 dence between nests counted and the guano yield at the plat- 40 30 forms (Fig. 3) suggests that trends in the number breeding at the 20 platforms may be gauged from the guano records. Similarly, 10 guano yield has been used as an index of the number of guano- producing seabirds in Peru (Crawford and Jahncke 1999). (c) Western Cape At Ichaboe Island, the most important breeding locality in 120 Smoothed southern Namibia, and at the six most important breeding local- Pairs (thousands) 100 Count ities in the Western Cape, counts of nests were undertaken in 80 many seasons since the late 1970s. Consequently, at these seven 60 islands, the smoothed values probably provide a reasonable 40 indication of trends in numbers of Cape Cormorants. These 20 islands and the guano platforms supported 224000 pairs in 1978–79, or 90% of the population in that season, so trends in (d) All main localities 250

200 70 1998–99 150 60 100 50 40 50 30 20

–01 –05 Pairs (thousands) 10 1956–571960–611964–651968–691972–731976–771980–811984–851988–891992–931996–9720002004 123 Fig. 2. Numbers of Cape Cormorants breeding, from 1956–57 to 2006–07, Guano (thousand t) at: (a) the Namibian platforms, (b) six islands off southern Namibia, (c) six islands in the Western Cape and (d) at these localities combined. Smoothed Fig. 3. The relationship between the harvest of guano at the Namibian trends are shown in (a) for the guano yield at the platforms, (b) the number platforms and the number of Cape Cormorants breeding at these platforms, of Cape Cormorants breeding at Ichaboe Island and (c) the number of Cape 1956–57 to 2006–07. The best-fitting regression line obtained after omitting Cormorants breeding at six islands in the Western Cape. the point for 1998–99 is shown. 258 Emu R. J. M. Crawford et al. the overall population are considered reliable for the latter half only at the north of the island in both the 1956–57 and of the period investigated. 1957–58 seasons (Rand 1963a) lends credence to the rather Almost all of Ichaboe Island was used by Cape Gannets from low estimate for this island in 1956–57. At Vondeling Island, 1956–57 to 1967–68 (Rand 1963b). Therefore, the low estimate the counts for 1956–57 and 1978–79 were similar, but for Dyer of the number of Cape Cormorants at that locality in 1956–57 is Island the count in 1956–57 was some 32000 pairs less than in thought to be accurate. The numbers at other localities in south- 1978–79. ern Namibia were gauged from aerial photographs taken on If, in 1956–57, the abundance in southern Namibia was 20 November 1956 (Rand 1963b). They may be underestimates, underestimated by 44000 pairs and that in the Western Cape by because in the 1990s and early 2000s many birds settled to breed 32000 pairs, the overall population would have been underesti- at these localities after November, with peak counts often in mated by 76000 pairs. This is more than 50000 pairs less than January or February (J. Kemper, unpubl. data). In 1978–79, the estimated increase in the population of 133000 pairs these localities supported 47000 pairs, compared with 3000 between the mid-1950s and the late 1970s, so it seems likely that estimated for 1956–57 (Table 1). there was a real increase in the population in this period. It is possible that numbers counted at some localities in the Western Cape in 1956–57 also underestimated the population. Factors influencing trends in Cape Cormorants This was probably not the case for Lamberts Bay, Malgas and The initial increase in the number of Cape Cormorants in Jutten Islands, where the counts made in 1956–57 were the Namibia can be attributed to two events that augmented the highest, or close to the highest, recorded in the 50-year period availability of breeding space, while still allowing an adequate (Table 1). At Dassen Island, the fact that Cape Cormorants bred supply of food. First, there was an increase in the number and

(a) Lambert’s Bay (e) Vondeling 12 8

9 6

6 4

3 2

(b) Malgas (f) Dassen 3 30

2 20

1 10

(c) Jutten (g) Robben 16 0.6 Pairs (thousands) Pairs (thousands) 12 0.4 8 0.2 4

(d) Schaapen & Meeuw (h) Dyer 3 40

30 2 20 1 10

1978–79 1982–83 1986–87 1990–91 1994–95 1998–992002–03 2006–07 1978–79 1982–83 1986–87 1990–91 1994–95 1998–992002–03 2006–07

Fig. 4. Smoothed trends in the number of Cape Cormorants breeding from 1978–79 to 2006–07 at: (a) Lamberts Bay, (b) Malgas Island, (c) Jutten Island, (d) Schaapen and Meeuw Islands, (e) Vondeling Island, (f) Dassen Island, (g) Robben Island, and (h) Dyer Island. Trends in Cape Cormorants over 50 years Emu 259 size of guano platforms, with the Swakopmund platform only (Simmons et al. 2006). On 25 November 2005, there were 2630 becoming operational in 1963 and some other platforms being pairs breeding on the island (B. M. Dyer, unpubl. data). extended after 1956–57 (Cooper et al. 1982). As the Namibian In the Western Cape, a long-term decrease in Cape sardine stock collapsed, its range contracted to the north, Cormorants at Lamberts Bay began in the late 1970s, whereas placing the remaining shoals increasingly distant from pronounced decreases at Malgas, Jutten, Vondeling and Dassen Namibia’s colonies. Most anchovy also were Islands occurred in the early to mid-1990s (Fig. 4). There was a located well north of the Gannet colonies (Crawford et al. large reduction in the overall number breeding in the Western 1987). However, these shoals were available to Cape Cape from the mid-1990s. Factors that influenced this decrease Cormorants at the guano platforms, where in 1973–74 sardine included mortality from avian cholera caused by the bacterium and anchovy contributed 84% and 15% by mass of their diet Pasteurella multocida (Crawford et al. 1992a; Williams and respectively (Berry 1976). Second, there was a decrease of Cape Ward 2002; Ward and Williams 2004; Waller and Underhill Gannets at Ichaboe Island, which increased the space available 2007), predation by Cape Fur Seals (Arctocephalus pusillus) for breeding by Cape Cormorants (Crawford 1991). Here, and at (Marks et al. 1997; Ward and Williams 2004) and Great White Mercury Island, Cape Cormorants were able to dive sufficiently (Pelecanus onocrotalus) (Crawford et al. 1995), and a deep to feed on pelagic gobies, which formed 84–100% by mass large eastward displacement of sardine off South Africa of their diet from 1978 to 1980 (Crawford et al. 1991) and 97% (Fairweather et al. 2006) decreasing its availability to seabirds by number of their diet in 1982 (Duffy et al. 1987). By contrast, breeding on islands west of Cape Point (Crawford et al. 2007). pelagic gobies were less available to the shallower diving Cape The increase in the number of Cape Cormorants breeding at Gannets and contributed only 10–28% by number of their diet Robben Island (Fig. 4) resulted from the erection of a platform at these localities during 1978–82, a period of food scarcity for adjacent to the island in late 2003. Gannets. In the 1970s, pelagic gobies may have partially Cape Cormorants do not breed in large numbers to the east replaced sardine off Namibia (Crawford et al. 1985). In the of Cape Agulhas (Cooper et al. 1982), where sardine have 1950s, when Gannets were abundant in Namibia, they fed recently been plentiful (Fairweather et al. 2006) and where there almost entirely on sardine (Hockey et al. 2005). has been a large increase in abundance of Cape Gannets Subsequently, the food sources available to Cape (Crawford et al. 2007). In 1977–81, 407 pairs of Cape Cormorants off Namibia diminished (Boyer and Hampton Cormorants bred to the east of Cape Agulhas (Cooper et al. 2001) and the number of birds breeding there decreased. The 1982). In 2003–05, 253 pairs bred there (B. M. Dyer, unpubl. biomass of anchovy in Namibia decreased from nearly 1 × 106 t data). in 1973 to <2 × 105 t in 1984 and 1985 (Le Clus et al. 1987). Clearly a variety of factors, including the abundance and By the end of the 20th century, the size of the Namibian distribution of food, the availability of suitable breeding habitat, anchovy resource was no longer estimated because it was so disease and predation, has influenced trends in the population of small (G. Dalmeida, Ministry of Fisheries and Marine Cape Cormorants. Available data suggest that the overall popu- Resources, Namibia, pers. comm.). In 1984, a Benguela Niño lation is at more or less the same level as it was when the first caused mortality and southward displacement of some fish survey was undertaken 50 years ago. In the intervening period, stocks in Namibia (Boyd et al. 1985; Le Clus 1985). The yield it expanded substantially before reverting to its initial level. Of of guano at the platforms decreased in the early 1990s (Fig. 2), particular concern for the future of the species has been a large suggesting that fewer Cape Cormorants bred there. In 1994, reduction in the abundance of food off Namibia that has per- low levels of dissolved oxygen off Namibia caused extensive sisted for more than a decade (Boyer and Hampton 2001), a mortalities of fish and seals and greatly reduced the availability decreased availability of prey in the Western Cape, brought of anchovy, sardine and horse mackerel in Namibia about more recently by an altered distribution of sardine (Hamukuaya et al. 1998; O’Toole and Bartholomae 1998). In (Fairweather et al. 2006), and an apparently increased frequency 1994–95, Cape Cormorants only initiated breeding at Ichaboe of outbreaks of avian cholera, which are generally associated Island in February (J. Kemper, unpubl. data). Then, in 1995, a with birds being under one or more forms of stress (Williams Benguela Niño advected anomalously warm water as far south and Ward 2002). as Lüderitz, displacing fish stocks to the south and causing mortality off northern Namibia of sardine, horse mackerel and Comparison with Cape Gannet other fish (Gammelsrød et al. 1998). It thereby further reduced Like the Cape Cormorant, the Cape Gannet feeds to a large the availability of food to Cape Cormorants. At the extent on anchovy and sardine, the two fish species that have Swakopmund platform, from 1994–95 to 1996–97 and in dominated the catch of purse-seine fishers in the Benguela 2004–05 Cape Cormorants had poor breeding success (J. Klein, ecosystem (Hockey et al. 2005). However, the Cape Gannet Salt Co., pers. comm.). Although a substantial number of Cape does not breed at the guano platforms and it breeds at only three Cormorants bred at the platforms in 1998–99 (Fig. 2), one of islands (Mercury, Ichaboe and Possession) off southern the owners of the guano platforms reported high mortality of Namibia and two (Lambert’s Bay and Malgas) in the Western cormorants at the end of the 20th century (W.Groenewald, Bird Cape (Crawford et al. 2007). It additionally breeds at Bird Rock Platform, pers. comm.). Island in South Africa’s Eastern Cape Province. From 1996, or earlier, some Cape Cormorants bred at Ilha There have been similarities and differences in trends of dos Tigres in southern Angola (Dean et al. 2002), probably Cape Cormorants and Cape Gannets. Numbers of Cape moving north there on account of the shortage of food in Cormorants breeding in southern Namibia may have increased Namibia. About 2000 pairs bred at this island in 2001–02 between 1956–57 and 1978–79. The number of Cape Gannets in 260 Emu R. J. M. Crawford et al.

Namibia was stable from 1956–57 to 1967–69 and then References decreased rapidly as a result of a greatly reduced abundance of Barnes, K. N. (Ed.) (2000). ‘The Eskom Red Data Book of Birds of South epipelagic fish prey that followed the collapse of the Namibian Africa, Lesotho and Swaziland.’ (BirdLife South Africa: Johannesburg.) stock of sardine (Crawford et al. 2007). Unlike the Cape Berry, H. H. (1976). Physiological and behavioural ecology of the Cape Gannet, the Cape Cormorant was initially able to feed substan- Cormorant Phalacrocorax capensis. Madoqua 9, 5–55. tially on pelagic gobies (Crawford et al. 1985). However, in the Boyd, A. J., Hewitson, J. D., Kruger, I., and Le Clus, F. (1985). Temperature 1980s, numbers of both Cape Gannets and Cape Cormorants and salinity trends off Namibia from August 1982 to August 1984, and decreased off southern Namibia. their relation to plankton abundance and the reproductive success of Off South Africa’s Western Cape, numbers of Cape Gannets pelagic fish. Collection of Scientific Papers International Commission remained fairly stable between the 1950s and the 1970s for the Southeast Atlantic Fisheries 12, 53–60. (Crawford et al. 2007), as did numbers of Cape Cormorants Boyer, D. C., and Hampton, I. (2001). An overview of the living marine (Fig. 2). The abundance of sardine in this region decreased in the resources of Namibia. South African Journal of Marine Science 23, 5–35. 1960s, but that of anchovy probably increased (Crawford et al. Cooper, J., Brooke, R. K., Shelton, P. A., and Crawford, R. J. M. (1982). 1987). In the Western Cape, numbers of Cape Gannets increased Distribution, population size and conservation of the Cape Cormorant, in the 1980s and early 1990s as South Africa’s sardine stock Phalacrocorax capensis. Fisheries Bulletin South Africa 16, 121–143. recovered (Crawford et al. 2007), but the population of Cape Crawford, R. J. M. (1991). Factors influencing population trends of some Cormorants was stable (Fig. 2), although a greatly reduced pro- abundant vertebrates in sardine-rich coastal ecosystems. South African portion of birds bred in 1989–90 and 1990–91 when there was a Journal of Marine Science 10, 365–381. low abundance of anchovy (Crawford and Dyer 1995). Then Crawford, R. J. M., and Dyer, B. M. (1995). Responses by four seabird Cape Cormorants decreased after 1992–93 (Fig. 2) and Cape species to a fluctuating availability of Cape anchovy Engraulis capensis Gannets in the early 2000s (Crawford et al. 2007). As fish were off South Africa. Ibis 137, 329–339. displaced to the east (Fairweather et al. 2006), they probably Crawford, R. J. M., and Jahncke, J. (1999). Comparison of trends in abun- remained available for a longer period to the Cape Gannet, dance of guano-producing seabirds in Peru and southern Africa. South which has a greater foraging range than the Cape Cormorant African Journal of Marine Science 21, 145–156. (Hockey et al. 2005). Crawford, R. J. M., and Shelton, P. A. (1978). Pelagic fish and seabird inter- relationships off the coasts of South West and South Africa. Biological In the Humboldt upwelling ecosystem off Peru and Chile, Conservation 14, 85–109. doi:10.1016/0006-3207(78)90026-5 where anchovy and sardine also are the main prey species, Crawford, R. J. M., Shelton, P. A., Batchelor, A. L., and Clinning, C. F. Guanay Cormorants (P. bougainvillii) and Peruvian (1980). Observations on the mortality of juvenile Cape Cormorants (Sula variegata) are considered to be the ecological equivalents Phalacrocorax capensis during 1975 and 1979. Fisheries Bulletin South of Cape Cormorants and Cape Gannets, respectively, in the Africa 13, 69–75. Benguela ecosystem (Crawford et al. 2006). The Guanay Crawford, R. J. M., Cruickshank, R. A., Shelton, P. A., and Kruger, I. (1985). Cormorant suffered a large decrease in abundance after the ini- Partitioning of a goby resource amongst four avian predators and evi- tiation of purse-seine fisheries, whereas there was long-term dence for altered trophic flow in the pelagic community of an intense, stability in the numbers of Peruvian Boobies, and the two perennial upwelling system. South African Journal of Marine Science 3, species presently have equivalent population sizes (Crawford 215–228. and Jahncke 1999). Once the most populous seabird in the Crawford, R. J. M., Shannon, L. V., and Pollock, D. E. (1987). The Benguela Benguela ecosystem (Crawford et al. 1991), the Cape ecosystem. Part IV. The major fish and invertebrate resources. Cormorant is now less numerous than the Cape Gannet Oceanography and Marine Biology: An Annual Review 25, 353–505. Crawford, R. J. M., Ryan, P. G., and Williams, A. J. (1991). Seabird con- (Crawford et al. 2007). This seems largely attributable to its sumption and production in the Benguela and western Agulhas ecosys- smaller foraging range and less flexible foraging strategy. The tems. South African Journal of Marine Science 11, 357–375. Cape Gannet, for example, is able to feed on fishery discards as Crawford, R. J. M., Allwright, D. M., and Heÿl, C. W. (1992a). High mor- well as catching its own prey (Hockey et al. 2005). The ability tality of Cape Cormorants (Phalacrocorax capensis) off western South of Cape Cormorants to colonise artificial habitat, such as plat- Africa in 1991 caused by Pasteurella multocida. Colonial Waterbirds 15, forms, provides opportunity to manage any long-term mismatch 236–238. doi:10.2307/1521458 between the distribution of the prey and breeding localities of Crawford, R. J. M., Underhill, L. G., Raubenheimer, C. M., Dyer, B. M., and Cape Cormorants. Märtin, J. (1992b). Top predators in the Benguela ecosystem – implica- tions of their trophic position. South African Journal of Marine Science Acknowledgements 12, 675–687. We thank our research institutes (listed under addresses) and the National Crawford, R. J. M., Cooper, J., and Dyer, B. M. (1995). Conservation of an Research Foundation for supporting this research. We are grateful to all who increasing population of Great White Pelicans Pelecanus onocrotalus in assisted with surveys of Cape Cormorants, to B. L. Dundee and F. V.Velho South Africa’s Western Cape. South African Journal of Marine Science for arranging surveys of the Namibian platforms and Ilha dos Tigres, and to 15, 33–42. W. Groenewald, D. Klein and J. Klein for providing information on the Crawford, R. J. M., Dyer, B. M., Cordes, I., and Williams, A. J. (1999a). guano platforms. CapeNature, Department of Environmental Affairs and Seasonal pattern of breeding, population trend and Tourism (South Africa), Ministry of Fisheries and Marine Resources of Bank Cormorants Phalacrocorax neglectus off south western Africa. (Namibia), Robben Island Museum, South African National Parks and Biological Conservation 87, 49–58. doi:10.1016/S0006-3207(98) South African Navy provided logistical support for the surveys. This paper 00042-1 is a contribution to the project LMR/EAF/03/02 of the Benguela Current Crawford, R. J. M., Dyer, B. M., and Upfold, L. (1999b). Seasonal pattern of Large Marine Ecosystem (BCLME) Program. We are grateful to three ref- breeding by Cape and Crowned Cormorants off western South Africa. erees, whose comments resulted in considerable improvement of the paper. Ostrich 70, 193–195. Trends in Cape Cormorants over 50 years Emu 261

Crawford, R. J. M., Goya, E., Roux, J.-P., and Zavalaga, C. B. (2006). Rand, R. W. (1960). The biology of guano-producing seabirds. 3. The distri- Comparison of assemblages and some life-history traits of seabirds in bution, abundance and feeding habits of the cormorants the Humboldt and Benguela systems. African Journal of Marine Science Phalacrocoracidae off the south-western coast of the Cape Province. 28, 553–560. Investigational Report Sea Fisheries Research Institute, South Africa 42, Crawford, R. J. M., Dundee, B. L., Dyer, B. M., Klages, N. T. W., 1–32. Meÿer, M. A., and Upfold, L. (2007). Trends in numbers of Cape Rand, R. W. (1963a). The biology of guano-producing seabirds. 4. Gannets (Morus capensis), 1956/57–2005/06, with a consideration of Composition of colonies on the Cape islands. Investigational Report Sea the influence of food and other factors. ICES Journal of Marine Science Fisheries Research Institute, South Africa 43, 1–32. 64, 169–177. doi:10.1093/icesjms/fsl011 Rand, R. W. (1963b). The biology of guano-producing seabirds. 5. Dean, W. R. J., Dowsett, R. J., Sakko, A., and Simmons, R. E. (2002). New Composition of colonies on the South West African islands. records and amendments to the birds of Angola. Bulletin of the British Investigational Report Sea Fisheries Research Institute, South Africa 46, Ornithologists’ Club 122, 180–185. 1–26. Duffy, D. C., Wilson, R. P., and Wilson, M. P. (1987). Spatial and temporal Schwartzlose, R. H., Alheit, J., Bakun, A., Baumgartner, T. R., and patterns of diet in the Cape Cormorant off southern Africa. Condor 89, Cloete, R., et al. (1999). Worldwide large-scale fluctuations of sardine 830–834. doi:10.2307/1368531 and anchovy populations. South African Journal of Marine Science 21, Fairweather, T. P., van der Lingen, C. D., Booth, A. J., Drapeau, L., and 349–366. van der Westhuizen, J. J. (2006). Indicators of sustainable fishing for Shelton, P. A., Crawford, R. J. M., Kriel, F., and Cooper, J. (1982). Methods South African sardine Sardinops sagax and anchovy Engraulis encrasi- used to census three species of southern African seabirds, 1978–1981. colus. African Journal of Marine Science 28, 661–680. Fisheries Bulletin South Africa 16, 115–120. Gammelsrød, T., Bartholomae, C. H., Boyer, D. C., Filipe, V. L. L., and Simmons, R. E., and Brown, C. J. (2007). ‘Birds to Watch in Namibia: Red, O’Toole, M. J. (1998). Intrusion of warm surface water along the Rare and Endemic Species.’ (National Biodiversity Programme, Angolan-Namibian coast in February–March 1995: the 1995 Benguela Ministry of Environment & Tourism and Namibian Nature Foundation: Niño. South African Journal of Marine Science 19, 41–56. Windhoek.) Hamukuaya, H., O’Toole, M. J., and Woodhead, P. M. J. (1998). Simmons, R. E., Sakko, A., Patterson, J., and Nzuzi, A. (2006). Birds and Observations of severe hypoxia and offshore displacement of Cape hake conservation significance of the Namib Desert’s least known coastal over the Namibian shelf in 1994. South African Journal of Marine wetlands: Baia and Ilha dos Tigres, Angola. African Journal of Marine Science 19, 57–59. Science 28, 713–717. Hockey, P. A. R., Dean, W. R. J., and Ryan, P. G. (Eds) (2005). ‘Roberts Underhill, L. G., Crawford, R. J. M., Wolfaardt, A. C., Whittington, P. A., Birds of Southern Africa.’ 7th edn. (John Voelcker Bird Book Fund: Dyer, B. M., Leshoro, T. M., Ruthenberg, M., Upfold, L., and Visagie, J. Cape Town.) (2006). Regionally coherent trends in colonies of African Penguins Kriel, F., Crawford, R. J. M., and Shelton, P. A. (1980). Seabirds breeding at Spheniscus demersus in the Western Cape, South Africa, 1987–2005. Robben Island between 1949 and 1980. Cormorant 8, 87–96. African Journal of Marine Science 28, 697–704. Le Clus, F. (1985). Effects of a warm water intrusion on the anchovy fishery Waller, L. J., and Underhill, L. G. (2007). The management of avian cholera in Namibia. Collection of Scientific Papers International Commission Pasteurella multocida outbreaks on Dyer Island, South Africa, for the Southeast Atlantic Fisheries 12, 99–106. 2002–2005. African Journal of Marine Science 29, 105–111. Le Clus, F., Hewitson, J. D., Melo, Y. C., Cooper, R. M., and Malan, P. E. Ward, V. L., and Williams, A. J. (2004). Coastal killers: causes of seabird (1987). The multispecies pelagic fishery of Namibia 1982–1986 and mortality. Bird Numbers 13, 14–17. stock assessments for pilchard and anchovy. Collection of Scientific Williams, A. J., and Ward, V. L. (2002). Catastrophic cholera: coverage, Papers International Commission for the Southeast Atlantic Fisheries 9, causes, context, conservation and concern. Bird Numbers 11, 2–6. 121–145. 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. O’Toole, M., and Bartholomae, C. (1998). The Benguela Niños. Namibia Brief 20, 109–114. Manuscript received 27 February 2007, accepted 22 August 2007

http://www.publish.csiro.au/journals/emu