Introduction

Organismal survival rates and reproductive performance often increase with age until the onset of senescence (Wunderle 1991; Forslund and Pärt 1995), the progressive deterioration of physiological functioning with age (Monaghan et al. 2008). This notably includes a decline in immune response (Palacios et al. 2011), muscular function (Hindle et al. 2009), and visual acuity over time (Schmolesky et al. 2000). Due to their exceptionally long lifespans and slow aging rates (Holmes and Austad 1995), birds, and in particular seabirds, are powerful study systems for researching age-related questions (Holmes and Austad 1995). Multiple studies have investigated the influence of seabird age on factors such as reproductive success (Anderson and Apanius 2003), survival (Ramírez et al. 2021), and physiology (Palacios et al. 2011). Of the few studies that have investigated behavioural changes in aging seabirds, conflicting results have been found (Pelletier et al. 2014; Elliott et al. 2015). For example, no foraging behavioural changes with age were found in Thick-billed murres (Uria lomvia), possibly due to their ability to make physiological changes (by reducing the rate of oxygen utilization with age following the reduced oxygen storage; Elliott et al. 2015). In contrast, spatial segregation was found between middle-aged and old Little penguins (Eudyptula minor) with older individuals foraging closer to shore in shallower waters as a result of diet preferences which differ among age classes (Pelletier et al. 2014). This calls for more research across different species especially focused on understanding the proximate causes of senescence, such as behaviour. In particular, foraging behaviour and success determine the resources an individual can allocate to maintenance and reproduction, and therefore play a key role in determining fitness. Previous studies have found that reduced physiological condition in older breeders affects trip duration, total distance travelled and habitat use (Catry et al. 2011; Jaeger et al. 2014). Thus, studying foraging performance may inform on mechanisms behind senescence. Behavioural adaptability to environmental change (which is crucial for survival; e.g. Jaeger et al. 2014) thus may differ between age classes as birds in each age class experience age-related foraging restrictions differently.

In addition to age, both sexually mono- and dimorphic seabird species can show sex-specific differences in aspects of their foraging behaviour e.g. at sea distribution, diving behaviour, or diet (Lewis et al. 2002; Phillips et al. 2004; Botha and Pistorius 2018). In sexually size-dimorphic species, differences in foraging behaviour may arise from size-related dominance at prey sites (González‐Solís et al. 2000) which ultimately reduces intersexual competition for food resources (De Pascalis et al. 2020). Within-sexes of dimorphic species, age can influence birds’ foraging. Foraging behaviour may change over the lifespans of specific sexes as seen in the sexually dimorphic Wandering albatrosses (Diomedea exulans) from the Crozet Islands (Weimerskirch et al. 2014) with older males increasing their foraging distance from the colony. In size-monomorphic species, it is unlikely for one sex to have a morphological advantage over another as a result of no or slight differences in size (Rishworth et al. 2014). Foraging differences between sexes in monomorphic species may instead be driven by divergent dietary requirements (Botha et al. 2017), intraspecific competition, or differential parental investment in chick provisioning (Elliot et al. 2010), with potential implications for energy expenditure, and in turn, aging trajectories.

The Cape gannet (Morus capensis) is a monomorphic seabird species, whose longevity makes it a useful species for investigating differences in aging between sexes of similar morphology. Endemic to Southern Africa (Crawford 2005), it is considered a useful indicator species of the ecological resilience of the Benguela upwelling ecosystem (Distiller et al. 2012). While age and sex both influence the foraging behaviour of various seabird species, the potential synergetic effects of these traits remain unclear (Fay et al. 2018), especially in monomorphic species. Therefore, in this study, we aimed to investigate age- and sex-related foraging behaviour in this monomorphic seabird.

Cape gannets mainly feed on sardines (Sardinops sagax) and anchovies (Engraulis capensis; Batchelor and Ross 1984; Adams and Klages 1999). In the southern Benguela upwelling system, the spatial distribution of these two small pelagic fish species has shifted since the early 1990s while their stock dwindled as a result of climate change and overfishing (Coetzee et al. 2008). This has led to a spatiotemporal mismatch between the traditional feeding areas of Cape gannets breeding on the west coast of South Africa and their eastward-shifting prey base, with detrimental effects for adult body condition, chick growth rates, individual fitness and ultimately population dynamics (Cohen et al. 2014; Grémillet et al. 2016).

Previous studies have demonstrated the impact of unprofitable foraging conditions on Cape gannet fitness, aligning with a regional population decline (Cohen et al. 2014; Grémillet et al. 2016). Foraging efficiency plays a key role in influencing the survival rates of gannets, and assessing how this behaviour varies across individuals of varying sex and age is necessary to estimate the vulnerability of Cape gannet populations to a changing environment. We use GPS and accelerometer data to determine age-related and sex-specific resource utilisation based on energy-related costs. We predict foraging performances to follow a bell-shaped curve (e.g. Monaghan et al. 2020; Fay et al. 2022; Saraux and Chiaradia 2022) with performance increasing to a certain age whereafter it decreases. We predict that middle-aged birds, having gained foraging experience (Fig. 1a) and have increased fitness, will forage more efficiently (Fig. 1c) than younger birds (with less experience) and older birds (potentially experiencing foraging-related senescence; Fig. 1b). With breeding females having generally longer foraging trips than males (Rishworth et al. 2014), we would also expect them to display earlier foraging senescence than males since life-history theory predicts that individuals which invest more energy into reproduction are likely to undergo faster and earlier senescence (Lemaître et al. 2015).

Fig. 1
figure 1

Theoretical representations of the non-linear relationships between Cape gannet foraging experience, physical fitness and foraging efficiency with increasing age

Full size image

Methodology

Study area

Fieldwork was conducted on Malgas Island (33.0528° S, 17.9254° E, Saldanha Bay, South Africa. The Cape gannet population on Malgas Island has decreased from 52,000 breeding pairs in 1996/1997 to 22,000 breeding pairs in 2018/2019 (Sherley et al. 2019), and continues to decline (BirdLife International 2018). At this site, gannet chicks have been fitted with metal rings since 1990 (Distiller et al. 2012), and resighting data are still being collected and collated by the South African Bird Ringing Unit (SAFRING). This allowed us to study the at sea movements and energetics of known age, breeding adult Cape gannets during October and November in 2017, 2019, and 2020. In total, we gathered data from 12 birds between 4 and 10 years old, 14 between 11 and 17, and 13 between 18 and 23. The Cape gannet generational length is estimated to be 18.3 years (Sherley et al. 2019). Although the Cape gannet’s life expectancy is unclear, that of their relatives, the Australasian gannet (Morus serrator) and Northern gannet (Morus bassanus), are 15.1 (Norman 2001) and 16.2 years (Nelson 1966), respectively. For this reason, we can assume that the sample used in this study, aged 4–23 years, appropriately captures the life expectancy of Cape gannets. These 39 birds comprised 11 females and 28 males (Table 1).

Table 1 The sample size (N) of chick-rearing male and female Cape gannets from different age categories (total = 39) on which a GPS and accelerometer were deployed on Malgas Island, South Africa, in 2017, 2019 and 2020. Young birds (4–10 years), middle-aged birds (11–17 years) and older birds (18–23 years)
Full size table

Foraging trip characteristics

Prior to being fitted with data loggers, 39 birds of known age rearing a one to six-week-old chick were caught on the nest and handled in the shade with their heads covered by a cotton cloth to minimize handling stress. Handling during deployment lasted less than 10 minutes. The birds were released as close to their nest as possible after logger deployment and retrieval. The birds were fitted with loggers for one foraging trip, lasting between 3 and 60 hours. A three-dimensional accelerometer recorder (Axy-4, TechnoSmArt, 9 × 15 × 2 mm, 16 g) was securely mounted on top of a GPS device (CatTrack1, Catnip Technologies, 4.7 × 3.0 × 1.3cm, 20 g) after which both were waterproofed using heat-shrink tubing. Loggers were attached to the lower back of the 39 birds using waterproof Tesa® tape (Wilson and Wilson 1989). This ensured that surging acceleration was measured along the longitudinal body axis of the birds, and heaving acceleration was measured dorsoventrally at 50 Hz (Fig. 2). The combined weight of the GPS device, accelerometer, and heat-shrink tubing was less than 2% of the average body mass of adult gannets (Grémillet et al. 2004). We monitored nests every second hour during daylight to retrieve the loggers soon after the birds returned to the colony. The Tesa® tape strips were removed completely. We used the software X Manager to configure the loggers and retrieve the data from the accelerometers, and @trip PC for the GPS devices. From the GPS data, we extracted information on the following foraging trip characteristics: trip duration (time between departure from and return to the colony, in hours), maximum distance from the nest (farthest point from the colony before returning, in km), and total distance travelled (distance covered between departure and return, in km; see Grémillet et al. 2004 for details). We mapped the core foraging ranges (50% fixed kernel density distributions in QGIS with HRef as a reference bandwidth; Kertson and Marzluff 2011; Signer and Balkenhol 2015) of birds of different age categories to visually determine if there was spatial segregation observed between sexes of different ages (Fig. 2).

Fig. 2
figure 2

A diagram showing the orientations of the X, Y and Z axes of an accelerometer attached to the lower back of a Cape gannet

Full size image

Energy expenditure

At-sea energy expenditure was estimated using acceleration data. We used Igor Pro 9.0 (WaveMetrics, OR, USA) to process data and classify behaviors (Grunst et al. 2023). In brief, acceleration in all dimensions was smoothed using a rolling algorithm (2 s sliding window) to extract static acceleration. Static acceleration was derived from the body angle with respect to gravity, and the values ranged from + 1 to − 1 g for a non-moving accelerometer. We calculated body pitch using the formula: atan(SX/(sqrt(SY^2 + SZ^2)) × (180/pi)), where SX, SY, and SZ are smoothed acceleration. Minimum specific acceleration (MSA) was calculated using the formula: abs(sqrt(X^2+Y^2+Z^2) − 1), where X, Y and Z are acceleration, as an indication of the bird’s activities (Simon et al. 2012). We reconstructed accurate Cape gannet time budgets (± 1 s) using activity specific acceleration (Fig. S1), whereby four activity categories and their associated energy requirements were determined (following Grémillet et al. 2016 and Green et al. 2009). Flying required 42.0 W.s−1, resting on the water surface required 26.5 W.s−1, diving required 55.2 W.s−1, taking off from the water surface required 85.9 W.s−1, as determined by Green et al. (2009) to determine the activity-specific rates of energy expenditure of Australasian gannets from their average heart rate per minute.

Genetic sexing

Chest feathers were collected from each of the 39 birds of known age on Malgas Island of which 11 were female and 28 were male. In the genetics laboratory, whole genomic DNA was isolated from the feathers using a Zymo Research Quick DNA extraction kit for feathers (Ryan et al. 2022). One chest feather per bird was submerged in a mixture of water, solid tissue buffer, dithiothreitol and Proteinase K, which was vortexed and incubated at 55 °C for 100 minutes in a dry block. The mixture was subsequently extensively washed, vortexed, and centrifuged (following Appendix D of the Zymo Research Quick DNA Plus extraction kit instructions) to extract the DNA. The supernatant was stored at − 20 ºC. The sex-linked CHD-1 genes were amplified using the forward primer 2550F (5′-GTTACTGATTC GTCTACGAGA-3′) and reverse primer 2718R (5′-ATTGAAA TGATCCAGTGCTTG-3′; Fridolfsson and Ellegren 1999). These primers detect males as ZZ (electrophoresis results display one band) and females as ZW (electrophoresis results displaying two bands; Rishworth et al. 2014). Polymerase chain reactions (PCRs) using 15 µl per sample (7.5 µl GoTaq Green Master Mix hot start, 0.6 µl of both primers, 4.3 µl nuclease-free water, and 2 µl of the DNA template) were performed in a C1000 Touch Thermal Cycler BioRad. Initial denaturation occurred at 94 °C for 2 minutes, followed by denaturation at 94 °C for 30 s, annealing for 30 s, and extension at 72 °C for 1 min (Rishworth et al. 2014). This process was repeated 42 times. A final extension at 72 °C for 5 min was subsequently initiated for the primer to form a double helix DNA strand (Connan et al. 2017), which was stored at 4 °C. PCR products were individually separated on a 1.8% agarose gel using 2.5 µl Prohasafe nucleic acid staining solution in 1X TAE buffer. Electrophoresis was then performed at 100 V for 30 minutes, and bands were visualized using ultraviolet radiation light (Connan et al. 2018).

Statistical analysis of foraging trip characteristics

To test for effects of age-related changes on the foraging trip characteristics (foraging trip duration, maximum distance from the nest, and total distance travelled) of tracked birds, we included age as well as age2 (representing a potential non-linear relationship, which was calculated by squaring the age of each bird; Lescroël et al. 2019; Debeffe et al. 2017) as predictor variables in Gaussian Generalized Linear Models fitted using the MASS Package (Venables and Ripley 2002) in R version 3.5.3 (R Core Team 2017). To normalize data spread we applied log10 transformation to maximum distance from the nest and total distance travelled, and log transformation to energy expenditure. Sex (and an interaction with age) was also included as a predictor variable to test for differences in aging trajectories between males and females. The full set of explanatory variables used in our models were age (continuous), age2 (continuous), and sex (categorical). An interaction between age and sex, and age2 and sex was run to test sexual differences in behavioural senescence. In each iteration, the variable with the highest p value was removed in a backward stepwise selection (Zhuang et al. 2018) to identify the best subset model (with and without interactions).

The year of study had no impact on the analysis of trip duration, maximum distance from the nest, total distance travelled or energy expended (year * age p > 0.05 in all models, Table S1) and was subsequently removed from further analysis.

Statistical analysis of activity-related energy expenditure

To determine the effect of age, sex, and their interaction on the amount of energy expended while flying, resting on the water, diving and taking off, we ran four separate General Linear Models with a Gaussian distribution using backwards stepwise selection in the MASS Package (Venables and Ripley 2002) in R version 3.5.3 (R Core Team 2017). The full set of explanatory variables used in our models were age (continuous), age2 (continuous), and sex (categorical), with an interaction between age and sex, and age2 and sex. A backward stepwise selection (Zhuang et al. 2018) was applied to identify the best subset model (with and without interactions).

We examined the effect of age and sex on (1) total energy expended, and (2) energy expenditure of each of the four at-sea activities. Total energy expenditure was calculated by summing energy expended per activity for a given foraging trip and dividing this number by foraging trip duration to calculate the average of energy expenditure rate during the trip (J/h). To normalize data spread, we log transformed energy spent diving and energy spent taking off from the water.

Results

Sex and age-related spatial distribution

In total, we analysed foraging trip characteristics data of 39 birds between the ages of 4 and 23 years with 11 females and 28 males (Table 1). There was a marginal degree of sex-related segregation at sea in relation with age in Cape gannet, as assessed visually (Fig. 2). Older birds tended to forage farther north than younger birds, and farther south than middle-aged birds (Fig. 3). Middle-aged birds generally tended to forage closest to the colony compared to other age classes. Older females travelled farther away from the colony than males irrespective of age, and young and middle-aged females.

Fig. 3
figure 3

Cape gannet core foraging areas (50% kernel density distributions) off Malgas Island, South Africa in 2017, 2019 and 2020 combined (a young birds (4–10, N = 12), b middle-aged birds (11–17, N = 14) and c old birds (18–23, N = 13)

Full size image

Sex and age-related differences in foraging trip characteristics

There was no significant effect of the variable ‘age2’ in any of our models, so it was removed from the subsequent analyses (Table S2–5). Therefore, the following results are based only on the best (linear) subset model (Table 2).

Table 2 The general linear model output (Gaussian distribution) with a different foraging characteristic as a response variable, and age (continuous variable) and sex (categorical variable with female birds as the reference) as explanatory variables before these explanatory variables were removed through the backward stepwise selection
Full size table

Neither age nor sex influenced trip duration, maximum distance from the nest, or total distance travelled (Table 2).

Sex and age-related differences in energy expenditure

There was no significant effect of the variable ‘age2’ in any of our models, so it was removed from the subsequent analyses (Table S5–9). Therefore, the following results are based only on the best (linear) subset model (Table 3).

Table 3 The general linear model outputs (Gaussian distribution) with energy expenditure activities as the response variable, with one or more explanatory variables varying from age (continuous variable), sex (categorical variable with female birds as the reference), and an interaction between them
Full size table

Females’ energy expenditure was positively related with age (p = 0.007; Fig. 4a) whereas overall, energy expenditure was not related to age in males. Females older than 17 years spent more energy per hour (mean = 423242 J/h; standard deviation = 611009) than males irrespective of age (mean = 120017 J/h; standard deviation = 7730; Fig. 4a; Table S5). Females, especially aging females (> 17 years), spent more energy in flight than males (p = 0.007; Table 3; Fig. 4b). Males spent similar amounts of energy resting irrespective of age, whereas females spent less energy resting with age (p = 0.005; Table S7; Fig. 4c). Males spent more energy on diving and taking off from the water than females (p = 0.005, Fig. 5a; p = 0.019, Fig 5b; Table S8–9).

Fig. 4
figure 4

Relation between age (in years) and both a total energy expenditure (flying, resting, diving and taking off from the water; J/h, log), b energy spent flying (J/h), and c energy spent resting on the water (J/h) of 11 female Cape gannets rearing chicks on Malgas Island, South Africa, between 2017 and 2020. The effect of age was not significant for males hence it is not included in the graph

Full size image
Fig. 5
figure 5

Difference in energy spent both a diving and b taking off from the water (J/h, log) while foraging at sea between female and male Cape gannets (N = 39) rearing chicks on Malgas Island, South Africa, between 2017 and 2020

Full size image

Discussion

While no differences were apparent in Cape gannets across age or sex in foraging trip characteristics, aging females (> 17 years) had higher energy expenditure than that of aging males, rested less and spent more energy in flight.

Despite there being no differences between age or sex in trip duration, maximum distance from the nest, or total distance travelled, older females (> 17 years) tended to spend more energy during a foraging trip than males of the same age, and generally expended increasingly more energy with age, unlike males. Aging females spent more time in flight than males, despite showing similar foraging trip durations and distances.

In a monomorphic species like the Cape gannet, there is no apparent difference in body size (Rishworth et al. 2014) and therefore no physical difference that might result in a competitive advantage of one sex over another while feeding at sea. We found that the energy expenditure of aging females (> 17 years) was higher than that of aging males, possibly as a result of the energetic cost of egg-laying (Lewis et al. 2002). Aging females also seemed to travel both farther north and south compared to young females (Fig. 3a) and old males (Fig. 3c). This could be a result of different dietary preferences (Lewis et al. 2002), and/or avoiding competition with younger, fitter individuals (Pettex et al. 2019). Young females expended the least amount of energy per hour, which suggests higher foraging efficiency (compared to middle-aged or old females) and/or different dietary preferences (compared to males; Lewis et al. 2002). However, younger gannets rested more than middle- or old-aged gannets, which may suggest an increased need to recuperate (Zango et al. 2020). Future research is required to determine which of the aforementioned explanations are more likely to reduce energy expenditure in young females. This could include the study of daily mass gains and/or foraging success (e.g., Fayet et al. 2015) to aid an understanding of the foraging behaviour and success of young gannets.

We also found that females, especially older females, expended more energy in flight than males irrespective of age, despite them having similar travelling duration and distance to males. This could be as a result of the early onset of senescence in females. Another explanation of this could be because adults of different sexes do not invest the same amount of time or energy into feeding the chick (Elliot et al. 2010). In our study, we speculate that aging females spent more time in flight, relative to males (marginal spatial segregation between sexes with age; Fig. 3), as a result of female nutritional requirements for egg production (Monaghan and Nager 1997). This could be driving them to search more actively for food than males (since females also rested less with age), and while doing so, spends energy more conservatively on diving and taking off from the water compared to males. Mortality and breeding costs may arise from the higher energetic expense of breeding in females (through egg production and laying), relative to males. This could lead to a decrease in population fitness through decreased female body condition and subsequent breeding attempts and success (Bijleveld and Mullers 2009; Froy et al. 2017) as foraging energetic constraints could lead to the onset of senescence and/or sex-biased mortality. With regards to males’ aging trends, we found that they expend the same amount of energy irrespective of their age. This could mean that they maintain a similar foraging strategy throughout their lifespan and might be experiencing senescence in a different trait that is not measured in our study e.g. foraging success, hence expending more energy on diving than females. Energy-related risks during foraging could influence foraging decisions, as individuals may either choose a food source that is less nutritious, but closer and more consistently available or a more nutritious, but less consistently available food source at greater distances. Central place foraging forms part of the life-history traits of Cape gannets whereby they are spatially constrained in the breeding season. A change in food distribution may result in longer foraging trips leading to irregular feeding of their chicks, and/or prey being beyond the foraging range leading to selection of poor-quality prey. As a life-history trait, gannets have slow-growing chicks which reduces the energetic risk of irregular food supply, however, during periods of low food availability, adults may prioritise their own survival above that of the chick (Bijleveld and Mullers 2009). Energy-related risks and trade-offs may influence foraging success (Wu and Giraldeau 2005), breeding success (Grémillet et al. 2016) as well as sex-biased mortality (Pichegru and Parsons 2014).

Conservation implications

The Cape gannet, having experienced a dramatic population decline since the late 1990s (Sherley et al. 2019), has already lost some genetic diversity (Reed and Frankham 2003), with consequences for population resilience to disruptions (Booy et al. 2000; Bradshaw and Holzapfel 2008; Vandewoestijne et al. 2008). Our results suggest that aging female gannets expend more energy than males to obtain food to sustain themselves and their chick while replenishing the resources lost from earlier investment in reproduction i.e. egg-laying. We speculate that older females might be physically incapable of further increasing energy they spend searching for food. If conditions of food availability deteriorate beyond a certain threshold, they may abandon the breeding attempt (Bijleveld and Mullers 2009). The already low breeding success on Malgas Island (Makhado et al. 2006) could have affected the population demography and could have increased the proportion of older birds in the population, which might further exacerbate the risk of increased breeding failure and lessen the resilience of this colony to environmental changes. Because of a shortage of food, the Cape gannet colony on Malgas Island appears to have entered an extinction vortex, defined as the reinforcement of processes that drive population extinction (Brook et al. 2008; Kovalenko 2019). Predation of Cape gannet offspring by both Cape fur seals (Arctocephalus pusillus pusillus) and Kelp gulls (Larus dominicanus vetula; Strydom et al. 2022a, b) further enhances pressure on that species. An additional stress factor could be that aging females are struggling in terms of energetic costs; therefore, Cape gannets might experience a further decrease in reproductive output. This could include a reduction in older females’ breeding attempts and food provision to chicks (Bijleveld and Mullers 2009), which influences population dynamics. It could be valuable in future to investigate body condition of females over time to establish if their body condition is deteriorating with age.

In conclusion, our study demonstrated the influence of age and sex on energy expenditure in foraging Cape gannets from Malgas Island. With the nature of our study being cross-sectional, future longitudinal research is required to better understand if either senescence or selective mortality is occurring in old Cape gannet females, as age-dependent trait variation can be caused by within-individual change (Zhang et al. 2015). This could further aid an understanding of intrinsic factors influencing the foraging behaviour of older females which is necessary to estimate the future population trends of this species more accurately. Our results underline the need to consider individual ages when investigating differences in foraging behaviour on a species level as these differences have potential consequences for the fitness of a threatened population of seabirds. Our study provided a new understanding of behavioural heterogeneity inherent to the Cape gannet population, which is necessary to estimate their vulnerability to a changing marine environment.