African penguins are amongst the most threatened of seabird species globally and an African Penguin Biodiversity Management Plan (BMP) published in 2013 guides conservation strategies to prevent their extinction. To counter the impact of past guano exploitation which reduced nesting habitat of these burrowing seabirds, rendering them vulnerable to predators and extreme weather events, a suite of artificial nests was deployed in various colonies over the past decades with varying success. The BMP called for new nest designs to maximise breeding output. • This study assessed African penguin breeding success over the past 14 years in four types of artificial nests and in natural nests at all major colonies in South Africa. 

Overall, artificial nests outperformed natural nests and successfully increased breeding output of African penguins by 16.5% (95% CI: 6.7–26.2%). No design was, however, consistently most effective across locations. For example, fibreglass nests had the poorest hatching success of all nest types at one colony (Bird Island), but out-performed cement nests at another (Boulders Beach). The latest design of double-layered ceramic nests tended to perform better than other artificial nests, although not consistently at all colonies and occupancy rates were low in some areas. 

Practical implication: The deployment of artificial nests proved to be a successful conservation measure for African penguins although the variety of habitats in the different colonies seems to have prevented a one solution-fits-all approach. The consistent poor breeding output in some areas in all nest types (<30% at the mainland colony in Simons’ Town and the islands in Algoa Bay) further calls for colony-specific conservation strategies beyond artificial nests, while the continuous decline in African penguin numbers globally calls for broader measures urgently, primarily increasing access to their forage fish prey.

Breeding success monitoring

All research was conducted under permits approved by the Department of Forestry, Fisheries, and the Environment, SANParks, and/or CapeNature. Methods were approved by the Ethics Review Bodies at the University of Bristol, the University of Cape Town, the University of Exeter, and Nelson Mandela University at different times during the lifetime of the project.

African penguins typically lay a clutch of two eggs (Hockey et al., 2005). Most nests were selected during incubation, although some were monitored continuously each year, and others selected at a later stage. In most colonies, nests were selected randomly in various areas of the colony, but at Robben Island all nests in a selected area were monitored, with nests being added when eggs were laid throughout the season. The nests’ contents (number of adults, eggs, chicks and size of chicks) were monitored every 7-10 days. If the eggs disappeared between successive checks, the nest was assumed to have failed at the incubation stage.  After 6-8 weeks, the chicks may leave their nests to join crèches (Seddon and van Heezik, 1993) and are not reliably associated with individual nests. Hence, we assumed that a nest was successful (up to the post-guard stage) if the chicks were known to have reached 8 weeks of age and the nest bowl was unoccupied. Reason for mortality could generally not be determined and, even when this was possible, it was not systematically noted across years and colonies. It was, therefore, not included in this study. Each breeding attempt in a marked nest was considered independently, as most birds were not individually marked, and more than one pair may occupy a nest site in a season in some colonies (Barham, 2017). While nest desertion due to disturbance by researchers was not systematically noted, African penguins breeding in the monitored areas used in this study are generally habituated to the presence of researchers (Scheun et al., 2021) and only very rarely desert their nests temporarily after human disturbance (Pichegru et al. 2016).

 Study site description

African penguin breeding success was monitored across eight colonies in South Africa between 2008 and 2022. Dassen Island (32°25'S, 18°05'E) is a large (220 ha), predominantly sandy and flat island, with the majority of penguins nesting in burrows excavated in sand, or between boulders and under shrubs (Wolfaardt, 2000). Robben Island (33°48'S, 18°22'E) is the largest of the islands off the coast of South Africa, with an area of 507 ha, and is covered by exotic vegetation such as rooikrans Acacia Cyclops, manatoka Myoporum serratum and various Eucalyptus that penguins use as nest sites (Tol, 2015). The mainland colony of Boulders Beach (34°11′S, 18°26′E) is located in the urban area of Simon’s Town, with birds breeding on the beach and under vegetation in a National Park and an area under the management of the City of Cape Town, but also extends across a busy road as well as in private residential gardens. Stony Point (34°22’S, 18°53’E) is also a mainland colony, located in a vegetated area around a rocky point and consisted of two municipal properties about 8.39 ha in size (Hugo 2021). Dyer Island (34°24’S, 19°15’E) is a flat island, 20 ha in size ca 6 m above sea level, with mixed vegetation over a pebble-covered surface. St Croix (33°48’ S, 25°46’ E) is a 12 ha rocky, steep island rising 58 m above sea level, with sparse vegetation. Finally, Bird Island (33°50’ S, 26°17’ E) is relatively flat, 19 ha in size about 9 m above sea level, covered with sparse vegetation that provide little cover for penguins.

 Artificial nests description

Wooden A-framed boxes are free-standing structures constructed from 3 mm thick pine plywood, varnished on the outer surface. The boxes are constructed to form a triangular prism (790 (l) x 430 (h) mm on each side and 380 (h) x 380 (w) mm at the back of the box). The front of the box is reinforced with a piece of wood 380 mm long (across the base of the entrance), but otherwise left open. An opening (c. 70 (h) x 90 (w) mm) is left near the top of the back of the box for ventilation. These wooden boxes were deployed on Robben Island in 2001, to compensate for the removal of alien vegetation used by penguins for breeding. They have been monitored since 2001 and are part of the longest existing monitoring programme on African penguins.

Cement nests made of a mixture of fibreglass and cement, approximately 30 cm high and 50 cm deep, were deployed at Dyer Island, Boulders and Stony Point between 2003 and 2005. A different type of cement nest, made of cement pipes cut longitudinally in half ca 30 cm high, 70 cm long, and 10 mm thick, and opened on both ends, were deployed on Bird Island in 2011, although no numbers were tracked. Most of them are now covered with vegetation and difficult to monitor (L. Pichegru, pers. obs.).

Fibreglass burrows were manufactured from fibreglass mesh and resin by the Dyer Island Conservation Trust (DICT, http://www.dict.org.za/) from 2005 onwards and aimed at mimicking the shape of a natural African penguin burrow. They are 60 cm long, 40 cm wide and 30 cm high, with an opening 25 cm wide and 28 cm high. Most were covered with rocks or soil upon deployment from 2006 until 2016, and some were overgrown with vegetation within a year or two. However, a number of fibreglass nests remained exposed with no coverage, while some filled up with sand and became unusable.

Ceramic nests are the latest design of artificial nests for African penguins, hand-manufactured by applying several layers of slurry-soaked geotextile over a mould. The slurry is a concoction of a ceramic powder, water, and a waterproofing agent. They have been manufactured by DICT since 2016 and installed at most colonies from 2018 onward. They are double layered with a 5 cm air gap and have a roof slanted at 27°. The height at the entrance is 20 cm and the height at the highest point is 50 cm, with an overall length of 60 cm and a width of 40 cm. The back side is closed except for seven holes for ventilation.

Various natural nest types were monitored at different colonies: surface open nests, nests under vegetation and natural burrows.

Analysis of breeding data: the Mayfield method

To account for biases induced by the monitoring protocol (e.g., it being unclear in some cases if chicks fledged or disappeared), an extension of the Mayfield (1961; 1975) method was used to determine nesting success. For each egg and each chick, we calculated the number of days exposed to potential mortality (egg days or chick days) by taking the mid-points between visits to nests (Mayfield, 1961) and recorded whether mortality occurred (= 1) or not (= 0). Exact visit times were not recorded, so nest-days were calculated to the nearest half-day. In cases where the hatching date could not be reliably estimated because of unusually long intervals between visits, or birds appeared to incubate long-dead eggs for many days, the estimate of nest-days was constrained to be 40 days (the maximum interval between laying and hatching is 41 days, e.g. Rand, 1960). We considered chicks to have died if they disappeared after <40 nestling days and no fate was clearly recorded (even if no carcass was observed). If chicks disappeared after ≥40 days, but had not been recorded in fledging plumage, we considered the monitoring to have been truncated (the observation was considered right censored) at the last time the chick was seen (Sherley et al., 2013). Accordingly, for censored observations, the time to death was imputed by the model using the interval distribution in JAGS (Plummer, 2003) as a random value greater than a specified censoring limit (Kruschke, 2015) based on the assumption that the individual egg or chick survived for at least one more day after they were last observed in the nest (Sherley et al., 2018, 2021).