Introduction
Seabirds are considered excellent indicators of marine ecosystem health. They are among the most threatened groups of birds, affected by multiple issues, including plastic pollution, fisheries, invasive alien species at breeding sites and changes in prey availability driven by climate change (Dias et al., Reference Dias, Martin, Pearmain, Burfield, Small and Phillips2019). Their need to return to land to breed allows population trends to be detected with systematic, repeated counts of breeding colonies over time. Population trend monitoring can provide insights into the health of their foraging areas and inform decisions on marine protected area management (Ronconi et al., Reference Ronconi, Lascelles, Langham, Reid and Oro2012).
Aldabra Atoll, a UNESCO World Heritage site in Seychelles, is a marine protected area and an important breeding site for several seabird species (Diamond, Reference Diamond1971). Comprehensive studies on seabirds breeding at Aldabra were first conducted by the Royal Society in 1968 (Diamond, Reference Diamond1971, Reference Diamond1974, Reference Diamond1975; Hnatiuk, Reference Hnatiuk1980). The most significant populations are red-footed boobies Sula sula, great frigatebird Fregata minor and lesser frigatebird Fregata ariel, which breed in mixed colonies along Aldabra’s extensive mangrove lagoon shorelines. Red-footed boobies are categorized globally as Least Concern on the IUCN Red List (BirdLife International, 2021) despite historical local population collapses attributed to human exploitation in the 20th century (Diamond, Reference Diamond1975; Feare, Reference Feare1978; Reville, Reference Reville1983).
A survey of the Aldabra red-footed booby population during 1968–1969 resulted in the first atoll-wide estimate of 6,000–7,000 annual breeding pairs (Diamond, Reference Diamond1971, Reference Diamond1974). That study also identified that egg laying occurs in two defined breeding pulses: in the dry season during August–September and in the wet season during January–February (Diamond, Reference Diamond1974). A second survey in 2000 (Burger & Betts, Reference Burger and Betts2001) estimated 5,000 breeding pairs in the wet season but did not consider pairs breeding in the dry season. It was assumed that the number of pairs breeding in each season was similar and the population was thus assumed to be 10,000 breeding pairs annually (Seychelles Islands Foundation, unpubl. data, 2000). Aldabra’s breeding red-footed booby population is regarded as the region’s second largest after Cosmoledo Atoll, which was estimated to support c. 15,000 pairs in 2001 (Rocamora et al., Reference Rocamora, Feare, Skerrett, Athanase and Greig2003). The paucity of population counts at Aldabra is due to logistical challenges: breeding colonies are extensive, covering c. 50 km of shoreline, and access to these colonies is tidally constrained.
Despite the inferred increase in Aldabra’s red-footed booby population during 1969–2000, the current population status was unknown, as recent ocean warming events (e.g. Koester et al., Reference Koester, Migani, Bunbury, Ford, Sanchez and Wild2020), overfishing in the region (Danckwerts et al., Reference Danckwerts, McQuaid, Jaeger, McGregor, Dwight, Le Corre and Jaquemet2014) and the presence of invasive alien species could all have had cumulative negative impacts on numbers. Black rats Rattus rattus occur across the atoll, with densities highest in mangroves (Harper et al., Reference Harper, van Dinther, Russell and Bunbury2015), and feral cats Felis catus occur on Aldabra’s largest island of Grande Terre (Wanless et al., Reference Wanless, Cunningham, Hockey, Wanless, White and Wiseman2002). Red-footed boobies are also known to be kleptoparasitized by frigatebirds when returning to nesting grounds after foraging. However, it is not known whether the proximity of nesting frigatebirds affects red-footed booby breeding success.
This study aimed to (1) record the current breeding phenology of red-footed boobies at Aldabra to determine the most suitable time for a survey of the entire population in each breeding season; (2) produce an updated atoll-wide population estimate across both breeding peaks by repeating past boat-based survey methods; (3) trial the use of unmanned aerial vehicles (UAVs) to survey breeding red-footed boobies and compare the results with those from boat-based surveys to identify the most appropriate method; and (4) investigate the effect of breeding frigatebird proximity on red-footed booby breeding success. In addition, we modelled the observed changes in Aldabra’s breeding red-footed booby population to establish whether it could be explained by intrinsic growth or whether alternative factors such as immigration would be required. We discuss our results in the context of Aldabra as a significant regional breeding seabird site and provide conservation management recommendations.
Study area
Aldabra Atoll is a large (34 × 14 km), raised coral atoll in the western Indian Ocean (Fig. 1). The atoll comprises four main islands surrounding a large lagoon (c. 200 km2) with tidal ranges of 2–3 m. The climate is dictated by the north-west monsoon (warmer, wetter; November–April; hereafter ‘wet season’) and the south-east trade winds (cooler, drier; May–October; ‘dry season’). Aldabra was designated a Special Reserve in 1981 and a UNESCO World Heritage Site in 1982. Managed by the Seychelles Islands Foundation, it hosts a research station on Picard staffed by 10–20 personnel who carry out monitoring and research. Red-footed boobies breed in mangroves along the lagoon shorelines of Picard, Malabar, eastern Grande Terre and smaller islets (Diamond, Reference Diamond1971; Fig. 1). A smaller colony was recorded during 1968–1969 on Gros Ilot Cavalier (off south-western Grande Terre; Fig. 1), but no breeding has been reported along Grande Terre’s southern lagoon shoreline since then.
Aldabra Atoll showing shoreline (red) and inland (blue) red-footed booby Sula sula breeding colonies in 2022–2023. Subcolonies monitored monthly are indicated by squares. The location of Aldabra Atoll in the Indian Ocean with reference to other major red-footed booby colonies is shown inset. (Readers of the printed journal are referred to the online article for a colour version of this figure.)
Methods
Breeding phenology and breeding success
We selected five subcolonies for monitoring across the red-footed booby breeding area at Aldabra (Fig. 1), based on the following criteria: well-defined boundaries, easily accessible by boat with minimal potential disturbance to nesting seabirds, a range of sites with some containing breeding frigatebirds, and a high density of red-footed booby nests. We monitored these monthly from February 2022 to February 2023, except for two subcolonies that were not visited in October 2022 and one that was not visited in November 2022 because of logistical constraints.
During each visit, subcolonies were counted by boat, except the colony at Middle Camp, which was accessible by foot at low tide. We classified active nests as apparently occupied nests with an adult sitting prone on the nest, or nests with chicks visible: a small chick (C1) covered in down feathers, a medium-sized chick (C2) partially covered in down feathers or a large chick (C3) with < 10% down feathers remaining (Burger & Lawrence, Reference Burger, Lawrence, Schreiber and Lee2000). We did not count larger chicks without down feathers because they are no longer attributable to a specific nest once they become mobile. We also recorded the number of active frigatebird nests, identified by adults sitting prone or chicks with down feathers. We gauged the optimal times for the boat-based and UAV surveys based on our observations of the breeding phenology, timing them to coincide with the peak of active nest numbers within monitored colonies.
We calculated breeding success from the total number of large chicks 4 months after the peak nest count (when most chicks still had down feathers and were counted as either advanced C2 or C3 chicks) as a percentage of the total number of active nests counted.
Boat-based surveys
Based on the breeding phenology results, we surveyed all shoreline colonies when counts of apparently occupied nests peaked: in the dry season during 12 August–3 September 2022 and in the wet season during 18–24 February 2023. We surveyed the entire lagoon shoreline of Picard, Polymnie, Malabar, eastern Grande Terre and associated islets from small outboard boats (Supplementary Fig. 1). Gros Ilot Cavalier was surveyed opportunistically when transiting through the lagoon to sites on southern Grande Terre. We surveyed a small section of the lagoon shoreline on eastern Malabar on foot at low tide. Aldabra’s lagoon is only accessible by boat at high tide, and we timed surveys to coincide with spring tides to maximize survey time. When possible, two teams surveyed different areas of the shoreline simultaneously.
We recorded all active nests, noting whether they were apparently occupied or contained chicks with down feathers. We noted whether the chicks were mostly C1 or C3, but did not record exact counts per chick category. We marked the sections of shoreline with nesting boobies using a handheld GPS to record the extent of breeding colonies. For each section of nests recorded, we noted the presence or absence of breeding frigatebirds occurring within the same area. The survey boats kept a distance of c. 10–15 m from the lagoon shoreline, apart from areas with narrow channels. We recorded all active nests visible at the shoreline and did not calculate an incubation index (Burger & Betts, Reference Burger and Betts2001; Diamond, Reference Diamond1971). We noted the presence of any white-tailed brown morph individuals (breeding or non-breeding) because the predominant phenotype breeding at Aldabra is the white morph (Diamond, Reference Diamond1971). We estimated Aldabra’s annual breeding population by summing the number of active nests from the 2022 dry season and the 2023 wet season. We considered these seasonal groups to be independent for two reasons: (1) chicks from nests counted in the dry season in 2022 that may still be within the colony in the 2023 wet season would have lost all their down feathers by then and therefore would not be counted; (2) our counts were conducted 27 weeks apart, which is too short a time for pairs failing in the dry season to re-nest in the subsequent wet season (previously reported to be a minimum of 37 weeks; Nelson, Reference Nelson1969), and it is therefore unlikely that we would have counted any pairs that nested in both seasons.
UAV surveys
Aerial surveys were conducted during the wet season in 2023, to cover mangrove habitats inaccessible by boat or foot (Supplementary Fig. 1), using a Wingtra One GENII UAV mapping drone (Wingtra AG, Zurich, Switzerland) equipped with a Sony Cyber-shot DSC RX1RII camera (full-frame CMOS sensor, 42 MP, 35 mm F2 lens; Sony Group Corporation, Tokyo, Japan). From Middle Camp, 19 flights (19–23 February 2023) surveyed 21.6 km2 of eastern Malabar and north-eastern Grande Terre mangroves, and from the Aldabra Research Station, six flights (28 February 2023) covered 4.8 km2 of Picard mangroves (Supplementary Fig. 1; Supplementary Table 1). Because of UAV range limitations from available launch sites, we did not survey Polymnie, Passe Gionnet, and central or western Malabar.
Survey flights, planned and executed using WingtraHub and WingtraPilot 2.8.1 software (Wingtra AG, Zurich, Switzerland), were 104 m above ground, achieving a ground sampling distance of 1.4 cm/pixel. We manually delineated survey grids around mangrove areas using high-resolution satellite images, and UAVs flew overlapping strip transects with 55% side and 30% front overlap.
We visually analysed all UAV images, georeferenced using QGIS 3.16 (QGIS Development Team, 2021), and manually counted active red-footed booby nests by identifying adults in the prone posture with nesting material visible, or if chicks with down feathers were visible (chicks were not classified as C1, C2, etc.). Observers ensured no double-counting in image overlap margins and identified all inland nests. We subsampled 10% of boat-surveyed shorelines to compare UAV and boat-based surveys. We counted nests in the corresponding UAV images within 25 m of the subsampled shoreline sections, assuming this range approximated the maximum distance from the shoreline at which nests can be detected during boat-based surveys (Espíndola et al., Reference Espíndola, Cruz-Mendoza, Garrastazú, Nieves, Rivera-Milán and Carlo2023).
Population growth rate
Our survey found a much larger number of breeding red-footed boobies than previous surveys, so we investigated whether this increase could have resulted from intrinsic population growth or whether immigration from other colonies in the region (e.g. Cosmoledo, Chagos) would have been necessary. We used a female-based population model adapted from Cubaynes et al. (Reference Cubaynes, Doherty, Schreiber and Gimenez2011), which assumes that red-footed boobies start breeding at the age of 4 years, that a proportion of breeding birds attempt to reproduce every year (breeding propensity), that a proportion of reproducing pairs succeed in raising a chick (productivity), that 50% of chicks are female and that a proportion of juvenile and adult birds survive from one year to the next (survival probability). We calculated annual population growth rate, λ, based on a deterministic matrix model estimated with the R package popbio 2.8 (Stubben & Milligan, Reference Stubben and Milligan2007; R Core Team, 2023) and multiplied abundance in year t by that growth rate to calculate abundance in the following year t + 1. Once abundance exceeded half the carrying capacity (assumed to be 20,000–40,000 breeding pairs in the wet season, based on available nesting habitat), we used a Ricker equation to limit population growth rates (Fowler, Reference Fowler1981). We used the count data of apparently occupied nests from the wet season of 1969 (2,277 pairs; Diamond, Reference Diamond1971) as a starting point and projected subsequent population sizes for 55 years to replicate the time series 1969–2024 covered by our most recent survey.
The demographic parameters required to estimate intrinsic population growth rate were unavailable from Aldabra and had to be adapted from other populations, so we performed 1,000 simulated population projections and drew the demographic parameters for each year of these simulations from distributions bounded by realistic values from other tropical seabirds (Nelson, Reference Nelson1978; Doherty et al., Reference Doherty, Schreiber, Nichols, Hines, Link, Schenk and Schreiber2004; Colchero, Reference Colchero2008; Dunlop, Reference Dunlop2009; Oro et al., Reference Oro, Torres, Rodríguez and Drummond2010). Specifically, we used beta distributions to inform juvenile survival (α = 85 and β = 17, resulting in a mean proportion of 0.835 ± SD 0.036, assumed to apply only during the first year) and adult survival (α = 92 and β = 8, mean = 0.920 ± SD 0.028, which we assumed to apply to all birds after their first year), and a random uniform distribution of 0.56–1.09 for breeding propensity, thus allowing for potential sub-annual breeding cycles in this tropical species. We estimated productivity from data collected in 2022 and drew productivity for all years from a random normal distribution with a mean of 0.48 ± SD 0.11. We present the median and all simulated population projections graphically. The code replicating these analyses is available online (Oppel, Reference Oppel2024).
Results
Breeding phenology and breeding success
During the wet season the number of apparently occupied nests peaked in February 2022 in the eastern lagoon (Middle Camp, Bras Cinq Cases) and slightly later during February–March 2022 in the western lagoon (Grande Poche, Passe Gionnet, Camp Frigate; Fig. 2). During the dry season the number of apparently occupied nests began to increase in June 2022 in the eastern lagoon and in July 2022 in the western lagoon. The eastern and western counts peaked in July and August 2022, respectively. A higher nest density was recorded in the monitored subcolonies in the wet season for most of the atoll, apart from Bras Cinq Cases, which had a higher nesting density in the dry season. The nest count was higher at the peak of the 2023 wet season compared with the 2022 wet season peak (Fig. 2). The mean breeding success of the five subcolonies was 48.0 ± SD 10.6% in the wet season and 29.1 ± SD 20.1% in the dry season (Table 1).
The number of breeding pairs of Aldabra red-footed boobies at the five monitored subcolonies over 13 months (February 2022–February 2023) calculated from the number of active nests counted during boat-based surveys. The sum of all five colonies is also shown to identify peak nesting activity across the atoll.
Breeding success at five red-footed booby Sula sula subcolonies monitored at Aldabra Atoll during the 2022 breeding periods in the wet and dry seasons, calculated from the total number of large chicks recorded 4 months after the peak nest count as a percentage of the total number of occupied nests. Also shown are the maximum number of breeding frigatebirds Fregata spp. recorded during the surveys and whether feral cats were present. Seasons were defined by the north-west monsoon (wet season, November–April) and the south-east trade winds (dry season, May–October).
1 The number of nests at Bras Cinq Cases was not included in the breeding success calculation for the dry season of 2022 because no data were available on the number of large chicks.
Breeding success was consistently higher in the Passe Gionnet subcolony compared to other subcolonies. The subcolonies in Camp Frigate and Middle Camp showed substantially lower breeding success in the 2022 dry season (17.9% and 10.0%, respectively) compared to other subcolonies, which coincided with higher nesting frigatebird activity. The breeding success of the Bras Cinq Cases subcolony could not be calculated in the 2022 dry season because of a lack of data. Bras Cinq Cases and Passe Gionnet were the only two monitored subcolonies with no recorded frigatebird nests; however, feral cats are present on Bras Cinq Cases. The subcolonies in Grande Poche and Camp Frigate had less than half the number of nests in the dry vs the wet season.
Boat-based surveys
In the August 2022 (dry season) atoll-wide survey, we recorded 14,092 active red-footed booby nests, including 13,562 apparently occupied nests and 530 with chicks visible. Most chicks counted were C1 chicks from the east lagoon, with low numbers of C3 chicks recorded across the atoll. Approximately 80% of red-footed boobies nested in areas without breeding frigatebirds. We recorded 34 white-tailed brown morphs, nearly half of which occurred around the Camp Frigate area (Fig. 1).
In the February 2023 (wet season) survey, we recorded 22,628 active nests, including 22,607 apparently occupied nests and 21 with chicks. Approximately 65% of red-footed boobies were nesting in areas without frigatebirds. All colonies in the 2023 wet season had a higher density of nesting red-footed boobies than in the 2022 dry season survey, apart from Bras Takamaka and inner channels, which had a higher number of breeding birds in the 2022 dry season. A total of 43 white-tailed brown morphs were recorded, most occurring around Camp Frigate. The nesting distribution of red-footed boobies was consistent across the two seasons surveyed, and representative data are shown in Fig. 1.
We recorded a total of 36,720 red-footed booby pairs breeding at Aldabra during 2022–2023. Compared with historical counts (Table 2), the most significant increases in colony size were in the Picard and Passe Gionnet areas, both in the wet season. No nests were recorded on Gros Ilot Cavalier in the 2022 dry or 2023 wet seasons.
Numbers of apparently occupied nests of breeding red-footed boobies recorded in boat-based, atoll-wide counts at Aldabra Atoll. Seasons were defined by the north-west monsoon (wet season, November–April) and the south-east trade winds (dry season, May–October).
1 The count areas in the 1968–1969 surveys were delineated differently compared to later counts; the number represents the combined count from both the area from Ile Verte to Middle Camp colony and Middle Camp Frigatebird colony.
UAV surveys
We captured a total of 11,154 UAV images (e.g. Plate 1) in 23 flights spanning 645 minutes (Supplementary Table 1). The aerial nest counts from the subsampled 10% of shoreline were similar to the boat-based counts, with the UAV images capturing a mean of 3.4 ± SD 5.7% more nests (total n = 2,547) than boat-based counts (total n = 2,480; Table 3). We also identified several inland colonies from the UAV images in the 2023 wet season. They covered a cumulative area of c.85 ha and contained 5,574 occupied nests that could not be detected by boat (Fig. 1). These accounted for 19.8% of the total number of nests (22,628 nests from boat-based shoreline count + 5,574 nests from UAV inland count = 28,202 nests in total).
Image captured by an unmanned aerial vehicle (UAV) at 104 m above ground level (ground sampling distance: 1.4 cm/pixel) showing a red-footed booby Sula sula breeding colony on the shoreline of Aldabra’s lagoon (white line) and another colony c. 50 m inland from the shoreline that was undetectable by boat-based observers.
Comparison of the number of apparently occupied red-footed booby nests counted in February 2023 (wet season) during boat-based and unmanned aerial vehicle (UAV) surveys of a sub-sample of 10% of the shoreline colonies, and in UAV images inland of the shoreline.
1 NA, not applicable: this area was included in boat and drone surveys of the shoreline, but excluded from the 10% of the colonies subsampled for the comparison of drone vs boat surveys.
The most extensive inland colonies were on Picard (n = 3,013 nests), with nearly double the number of nests counted by boat along the shoreline (Tables 2 & 3). Assuming that a similar proportion of colonies occur inland during the dry season as in this partial aerial survey in the wet season, there would be at least another 3,523 inland nests during the dry season (not accounting for any inland nests on western Malabar, which was not surveyed by UAV). Therefore, including the UAV survey results, we estimate that Aldabra’s total red-footed booby population is at least 45,817 breeding pairs.
Population growth rate
We estimated a mean annual population growth rate of 1.037 (range 1.000–1.050, n = 54 years) from our population model, which was close to the observed rate of 1.033 from counts made during the wet season on Aldabra between 1969 and 2023, without immigration from other colonies (Fig. 3). However, this population growth would only be possible if the count in 2000 was an underestimate, because the projected population in 2000 was almost twice as high as the estimate from actual counts (Fig. 3), and intrinsic growth could not explain the increase between the observed counts in 2000 and 2023 (Supplementary Fig. 2). Because of the logistic growth curve assumed in the model, the mean annual population growth between 1969 and 2000 was 1.055 (range 1.048–1.058), while the annual population growth slowed to a mean of 1.014 (range 1.000–1.047) between 2000 and 2024 as the population approached the assumed carrying capacity of Aldabra.
Observed (red dots) and inferred mean (red line) breeding population size of red-footed boobies at Aldabra nesting in the wet season during 1969–2023 based on a population model with demographic parameters derived from Cubaynes et al. (Reference Cubaynes, Doherty, Schreiber and Gimenez2011). Grey lines represent 1,000 simulations to account for uncertainty in the demographic parameters. Note that although the count in 2023 can be explained by intrinsic population growth, the count in 2000 appears to have been an underestimate (see also Supplementary Fig. 2). (Readers of the printed journal are referred to the online article for a colour version of this figure.)
Discussion
We show that Aldabra’s breeding red-footed booby population has grown by an order of magnitude over the past 55 years, and the last survey in 2000 was probably an underestimate of the population at that time. In the 2023 wet season, the boat-based survey recorded 22,607 incubating birds, a significant increase compared to the previous wet season survey in 2000 (Burger & Betts, Reference Burger and Betts2001). The annual breeding population counted by boat during 2022–2023 (wet and dry seasons) was 36,720 pairs, with approximately one-third (14,092 pairs) breeding in the dry season and the remaining two-thirds (22,628 pairs) breeding in the wet season. We also found from our UAV surveys that several colonies extended inland into mangrove habitats that cannot be surveyed by boat. The annual breeding population was, therefore, likely to be closer to 45,817 pairs, not accounting for any inland pairs breeding on western Malabar, which was not surveyed by UAV. This result makes Aldabra the largest recorded red-footed booby breeding population in the Indian Ocean, based on available estimates of other key breeding sites (Rocamora et al., Reference Rocamora, Feare, Skerrett, Athanase and Greig2003; Le Corre et al., Reference Le Corre, Danckwerts, Ringler, Bastien, Orlowski and Rubio2015; Carr et al., Reference Carr, Votier, Koldewey, Godley, Wood and Nicoll2021; BirdLife International, 2021), and probably one of the largest populations in the world. However, recent and reliable population estimates are lacking for many major breeding areas, such as the Galápagos, Caribbean and Pacific Islands (BirdLife International, 2021).
Comparisons of the 1968–1969 and 2000 counts are problematic, as discussed in Burger & Betts (Reference Burger and Betts2001): surveys were conducted at different times of the breeding season, used different census methods (including correction factors in 1968–1969 but direct nest counts in 2000), applied different treatments of fledglings/immature individuals, and defined colony boundaries differently. By introducing breeding phenology monitoring, standardizing nest stage categories and defining clear survey maps, this study establishes a framework for repeatable and comparable boat-based surveys. However, our findings indicate that UAVs are the most effective method for surveying red-footed boobies at Aldabra. The nest counts from UAV images closely matched boat-based estimates (within a mean of 3.4 ± SD 5.7%, Table 3) while also identifying 5,574 previously undocumented inland nests. The use of UAVs also reduced logistical effort, requiring fewer person-hours and enabling more frequent monitoring. Moreover, UAV imagery provides an objective, permanent reference for future assessments (Hodgson et al., Reference Hodgson, Mott, Baylis, Pham, Wotherspoon and Kilpatrick2018). To ensure comprehensive population coverage, we recommend UAVs similar to the Wingtra GENII, with extra surveys from additional launch sites such as Anse Porche and Anse Malabar (Supplementary Fig. 1).
Wildlife surveyors increasingly favour UAVs because of their ability to deliver consistent, repeatable population counts (Chabot & Francis, Reference Chabot and Francis2016; Hodgson et al., Reference Hodgson, Mott, Baylis, Pham, Wotherspoon and Kilpatrick2018; Dickens et al., Reference Dickens, Hollyman, Hart, Clucas, Murphy and Poncet2021; Edney & Wood, Reference Edney and Wood2021). They are particularly effective for surveying large, conspicuous seabirds such as red-footed boobies (Corregidor-Castro et al., Reference Corregidor-Castro, Riddervold, Holm and Bregnballe2022; Espíndola et al., Reference Espíndola, Cruz-Mendoza, Garrastazú, Nieves, Rivera-Milán and Carlo2023). In this study, UAVs were used to explore extensive mangrove habitat to identify inland colonies and also to compare counts with those from a shoreline subsample of the boat-based surveys. However, logistical constraints prevented image capture with sufficient forward and side overlap (80–90%) to produce orthophoto mosaics, which could allow detailed spatial analyses such as nest density mapping (Rümmler et al., Reference Rümmler, Esefeld, Pfeifer and Mustafa2024). We recommend that future surveys configure UAV flights for orthophoto generation, which would enhance spatial analyses and facilitate robust comparisons over time.
Monitoring of other red-footed booby colonies in the region (e.g. Farquhar Atoll, Seychelles) has also recorded substantial increases in population size (Island Conservation Society, 2020), suggesting immigration from other colonies elsewhere. Changes in the proportion of different colour morphs in the population of red-footed boobies can provide inferences into whether immigration has occurred (Le Corre et al., Reference Le Corre, Danckwerts, Ringler, Bastien, Orlowski and Rubio2015). Our survey recorded more white-tailed brown morph individuals (n = 77) than previous studies (n = 0 by Burger & Betts, Reference Burger and Betts2001; n = 4 by Diamond, Reference Diamond1974). However, white-tailed brown morph individuals remain rare at Aldabra (0.2%), suggesting that immigrants are either numerically few or that source populations are not in the Mozambique Channel (e.g. Europa Island) where the white-tailed brown morph is the dominant colour morph (Le Corre, Reference Le Corre1999).
Our model shows that the population growth between 1969 and 2023 is similar to observed growth rates in other seabird populations where immigration was unlikely to have occurred (Saunier et al., Reference Saunier, Amy, Baumann, Bignon, Cartraud and d’Orchymont2024). However, although our model suggests that intrinsic growth is theoretically possible, we emphasize that we have no evidence to infer whether immigration occurred. Besides birds from elsewhere potentially immigrating to Aldabra, red-footed boobies from Aldabra may also emigrate to other breeding islands. Recent genetic research has shown that red-footed boobies from North-Keeling Island in the south-east Indian Ocean cluster with Aldabra individuals and are distinct from individuals from nearby Christmas Island (Morris-Pocock et al., Reference Morris-Pocock, Anderson and Friesen2016). One hypothesis is that the North-Keeling Island colony results from a relatively recent dispersal from Aldabra. The significant growth of red-footed booby populations in the Indian Ocean warrants further investigation through genetic analysis at all key breeding sites, as Aldabra may act as a source population, further enhancing its importance as a protected site in the region.
Although we cannot rule out a contribution of immigration to the population growth on Aldabra, our model shows that intrinsic population growth could explain the observed increase despite comparatively low breeding success. On average, red-footed booby breeding success on rat-free Tern Island in Hawaii was 62% (Dearborn et al., Reference Dearborn, Anders and Flint2001). By contrast, our study recorded an average breeding success of 29% (dry season) to 48% (wet season), with lower figures reported at sites with breeding frigatebirds (10–18%). In addition, breeding success in our study may have been overestimated as some adults may lay eggs after the peak incubation count, and some nests may fail during the late chick stage. Frigatebirds at Aldabra have been recorded both kleptoparasiting birds in flight and mobbing red-footed booby nests for nesting material (Diamond, Reference Diamond1974). It is therefore plausible that lower breeding success at sites such as Middle Camp and Camp Frigate is caused by the presence of breeding frigatebirds. Although breeding and roosting frigatebirds are absent from Bras Cinq Cases, feral cats on Grande Terre may be responsible for the relatively low breeding success in the wet season at this site. The impacts of cats on seabirds at Aldabra are largely unknown, but cats are known to cause extinctions of large seabirds on other islands (Dias et al., Reference Dias, Martin, Pearmain, Burfield, Small and Phillips2019). Rats occur at their highest densities in mangrove habitats (Harper et al., Reference Harper, van Dinther, Russell and Bunbury2015), but currently there is limited information about their potential impacts on seabirds at Aldabra (Fayet et al., Reference Fayet, Sanchez, Appoo, Constance, Clucas, Turnbull and Bunbury2023). After rats were eradicated from Tromelin Island, seabird colonies rebounded, numbers increased significantly, and some extirpated species recolonized the island without any further conservation action (Saunier et al., Reference Saunier, Amy, Baumann, Bignon, Cartraud and d’Orchymont2024). A number of confounding factors make it difficult to determine the reasons for the variation in breeding success recorded in different monitored colonies at Aldabra. However, despite the presence of invasive species and impacts from neighbouring frigatebirds, Aldabra’s red-footed booby population can be self-sustaining at the current levels of breeding success.
Red-footed boobies are categorized as Least Concern on the IUCN Red List, with a global population of 1.4 million individuals and a slow decline (BirdLife International, 2021). Although the species was reported to be decreasing in the western Indian Ocean 4 decades ago (Feare, Reference Feare1978), recent years have seen regional increases, with more roosting birds and the establishment of new breeding colonies (Carr et al., Reference Carr, Votier, Koldewey, Godley, Wood and Nicoll2021; Curd, Reference Curd2021; Feare, Reference Feare2021). This recovery is remarkable, given the region’s environmental pressures including ocean warming and overfishing. Following a history of wildlife harvesting at Aldabra (Stoddart, Reference Stoddart1971), the gradual development of protective legislation throughout the 20th century, culminating in the designation of the atoll as a UNESCO World Heritage Site in 1982, has most probably been pivotal in the recovery of Aldabra’s red-footed booby population. This hypothesis is supported by the significant growth recorded in colonies on Picard and Passe Gionnet (Table 2) in areas closest to the former settlement and most probably subjected to the highest harvesting pressure (Stoddart, Reference Stoddart1971). Additionally, it is possible that prey resources in the Indian Ocean are adapting favourably to ocean warming, potentially positioning red-footed boobies as one of the few beneficiaries of rapid climate change (Webster et al., Reference Webster, Colton, Darling, Armstrong, Pinsky, Knowlton and Schindler2017).
Our study highlights the global importance of Aldabra Atoll, a protected site that supports the largest red-footed booby colony in the Indian Ocean. We propose that research into population genetics, together with tracking studies in the region, could address the remaining questions around the dispersal of red-footed boobies and the connectivity of populations. Our study could not conclusively discern whether the current large population on Aldabra has resulted from rapid growth and immigration post-2000, or whether the population has been growing for >50 years and the previous survey in 2000 was an underestimate. To avoid such uncertainty in the future, we recommend the use of UAVs for rapid surveys of red-footed boobies and other large, conspicuous seabirds breeding in low-lying mangrove colonies so that their population status can be monitored regularly and reliably.
Author contributions
Study design: MMR, CWJ; data collection: MMR, CWJ, MOB, CAB; data analysis: MMR, CWJ, SO; writing: MMR, CWJ; revision: MMR, CWJ, SO, NB, MOB, CAB; project administration: NB, FFD.
Acknowledgements
We thank the Seychelles Islands Foundation Aldabra Research Station team who participated in this survey: Alvin Bonnelame, Alain Bouchary, Sebastian Cowin, Mickael Esparon, Emma Mederic, Guilly Mellie and Nikita Pothin; Francis Salomon and Nigel Corgat for logistical support at Aldabra; Ronny Rose, Julio Agricole and Christina Quanz for delivering the UAV equipment to the atoll; and Matthieu Le Corre for his valuable feedback on an earlier version of the text. This project was funded internally by the Seychelles Islands Foundation.
Conflicts of interest
None.
Ethical standards
The research abided by the Oryx guidelines on ethical standards. The Seychelles Islands Foundation coordinated and implemented the research and monitoring on Aldabra; no specific approval was required, and all surveys were conducted in accordance with the Aldabra Management Plan. This study did not require any animal handling or sample collection.
Data availability
Data pertinent to this study are available on request from the corresponding author.
Open access
