Using the prey captured by breeding Crested Terns to assess the availability of forage fish for other coastal meso-predators

Introduction

The diets of seabirds have been used successfully as indicators of fish populations (Velarde et al. 1994; McLeay et al. 2009a) including the cohort strength of forage fishes (e.g. Sardines, Sardinops vagax) prior to recruitment into a fishery (Velarde et al. 2013). Seabirds are also indicators of the impact of changing forage fish stocks on meso-predators, such as Crested Terns (Thalasseus bergii), that consume them (McLeay et al. 2009b).

Seabird diets can be characterised using a variety of methods including the analysis of gut contents, spontaneous regurgitates, emetic or water off-loaded regurgitates, hard-parts in regurgitated pellets or faeces and the DNA analysis of any of these sources. All these methods involve entering and disturbing breeding colonies, and some involve the relatively invasive capture and handling of the birds. Observing the marine animals carried to colonies, roosts, or clubs by bill-loading seabird species provides an alternative method of assessing prey selection.

Observational methods have been advanced by the emergence of high resolution, telephoto digital photography (Edney and Wood 2021). Images can be examined in forensic detail, stored, verified by taxonomic experts and catalogued into voucher collections to underpin ongoing monitoring. Digital photography also facilitates participation by the citizen scientist as it usually does not require all the licences and permits required for using the more invasive methods (Forys and Hevesh 2017; Owen et al. 2024) and generally reserved for ‘professionals’.

Many methods used in seabird dietary analysis have biases and limitations (Barrett et al. 2007; Owen et al. 2024). Stomach contents and pellets can overestimate the prey-species that leave undigested hard-parts. Regurgitates (and off-loaded stomach contents) tend to contain the larger, most recently consumed prey items. The number of individual prey items can generally not be obtained from DNA analysis and the method depends on a comprehensive library of prey-genomes, DNA analysis is currently too expensive to be used as an ongoing monitoring method. With the exception of regurgitated pellets, dietary sampling is generally restricted to colony areas during the breeding season. Bill-loading generally occurs when adults are feeding young or mates during the breeding season. The adults may select and consume different prey that will not be detected using digital photography.

The ability to non-invasively assess prey selection and availability offers a means to investigate the impacts of changes in prey availability on small and declining seabird populations where more invasive methods are difficult or undesirable. The Little Penguin (Eudyptula minor) colony on Penguin Island, Western Australia (Fig. 1) has seen a 94% reduction in population size since 2008 and is now less than 75 pairs (Cannell 2024). This decline has been partly attributed to a severe marine heatwave in 2011 and the apparent disappearance of a key forage fish, Sandy Sprat (Hyperlophus vittatus), from the colony’s prey-field south of the Island at that time and in the subsequent year (Cannell et al. 2024). In 2011 and 2012 Tropical Sardines (Sardinella lemuru) and Sardines (Sardinops vagas) were the main prey species (Cannell et al. 2024).

Fig. 1.

Study area showing location of Penguin Island and Garden Island and the expected local foraging range of Little Penguins (indicated by vertical blue lines; Cannell 2016, 2017, 2018; Sutton 2022; Cannell et al. 2024) and Crested Terns (indicated by horizontal orange lines; Mcleay et al. 2010). Note: Little Penguins are also known to forage in waters approximately 150 km south of Penguin Island (not depicted here) during breeding when local food abundance is reduced.

Previous intermittent diet studies have been conducted on the Little Penguins on Penguin Island. Initially these were based on water-offloading stomach contents and latterly on DNA analysis of faeces (Cannell et al. 2024). Due to a lack of resources no dietary sampling was undertaken after 2012 that might inform the prospects of the colony recovering and the merits of ongoing land-based conservation actions. Sandy sprats are considered a critical food source during the chick-rearing phase of Little Penguins breeding on Penguin Island (Cannell et al. 2024). In south-western Australia, Sandy Sprats are a significant component of the region’s commercial fishery, primarily supplying the bait market, with annual catches of 150–200 tonnes reported in the early 2000s (Rogers et al. 2006). Annual monitoring of catch rates revealed a collapse in the fishery following the 2011 marine heatwave prompting a reassessment of sustainability. As a result, previous catch allowances were deemed unsustainable, and commercial catches have been restricted to 25 tonnes per year (Newman et al. 2023). The fishery for Sandy Sprat operating in the Little Penguin prey field was closed following the collapse and there has been limited monitoring of forage fish abundance in the region to provide information on changes in prey availability.

The diet and foraging range of the Little Penguin colony breeding on Penguin Island are likely to overlap with other central-place foraging species in the region, such as the Crested Tern breeding in the Perth region (Fig. 1). Crested Terns are widely distributed across coastal subtropical and tropical waters of the Indian Ocean, as well as the Indo-Pacific. In Australia, they are the most widely distributed tern species, occurring along the entire continental coastline (Dunlop 1987). Crested Terns nest in highly synchronous, densely packed colonies on coastal islands within range of aggregations of forage fishes (Dunlop 1987). They are a plunge diving, bill-loading species that forage at short range from breeding colonies (generally within 40 km; Mcleay et al. 2010). The availability of appropriately sized and nutritious forage fishes is critical for egg formation (Dunlop 1987) and for provisioning small nestlings shortly after hatching (Surman and Wooller 2006; Gaglio et al. 2018; Quiring et al. 2021). During incubation, Crested Terns may exhibit a high degree of flexibility in prey selection and feed their larger running chicks on a wide range of non-forage fish species, squid and crustaceans (Chiaradia et al. 2002; Surman and Wooller 2006; Gaglio et al. 2018; Quiring et al. 2021). However, certain prey species may be more profitable. A collapse in Sardine abundance following a herpesvirus epidemic was shown to reduce chick survival rates, and lead to stunted adults (McLeay et al. 2009b).

Early investigations of prey selection in Crested Terns were based on spontaneous regurgitations of food items collected during banding operations (Chiaradia et al. 2002; Surman and Wooller 2006; Mcleay et al. 2009a, 2009b). More recently less invasive methods have been employed using the digital photography of fish carried into breeding colonies at various stages in the nesting cycle (Gaglio et al. 2017, 2018; Quiring et al. 2021).

This project pilots the efficacy of using digital photography to monitor the prey carried by Crested Terns to a colony on Penguin Island as a means of assessing local forage fish availability for the penguins and other meso-predators. The engagement of citizen-scientists in photo-sampling is anticipated to reduce the cost of monitoring prey-availability relative to other methods.

Materials and methods

Results

Over the three Crested Tern breeding seasons, 1900 prey items, representing 62 species, were identified using 17,707 photographs. Of these species, 59 were prey species were fish, two were prawns; and one was squid. A complete list of prey species recorded in each of the three breeding seasons is in Table 1.

Overall, the dominant prey were clupeid forage fishes (64.17%), Blue Sprat (Spratelloides robustus) (21%), Sandy Sprat (Hyperlophus vittatus) (17%), Sardine (Sardinops vagax) (15%), Tropical Sardine (Sardinella lemuru) (1%), Anchovy (Engraulis australis) (9.16%), and Marray (Etrumeus jacksoniensis) (1.01%). Important secondary prey were mainly species inhabiting seagrass meadows including Brown-spotted Wrasse (Notolabrus parilus) (2.63%), Western Striped Grunter (Helotes octolineatus) (8.42%), Rough Leatherjacket (Scobinichthys granulatus) (2.68%), and Silverbelly (Parequula melbournensis) (2.37%).

The abundance of forage fish species varied over the three seasons (Fig. 2). Blue and Sandy Spats were the most abundant forage fishes in 2021 and 2022 but declined in 2023 as the numbers of Sardines and Anchovies carried by Crested Terns increased. In November during the peak in nesting activity, the abundance of Blue Sprat differed significantly between seasons (χ² = 16.64, P < 0.001) as did the numbers of Sardines (χ² = 10.9, P < 0.005). The differences in the number of Sandy Sprat and Anchovies recorded was not significant. At an overall seasonal level, there were significant differences in the numbers of Blue Sprat (χ² = 13.12, P < 0.005), Sandy Sprat (χ² = 4.88, P < 0.01), Sardines (χ² = 4.39, P < 0.05), and Anchovies (χ² = 6.73, P < 0.05).

Fig. 2.

Percentage frequency of forage fish species carried to Crested Tern colonies on Penguin Island in 2021, 2022, and 2023 breeding seasons.

Fig. 3 shows the length classes of Sandy Sprats recorded in each Crested Tern breeding season. In 2021, Sandy Sprats were relatively abundant with all fish being small and mostly below the adult size of around 60 mm, indicating Sandy Sprat recruitment following autumn winter/spawning (Gaughan et al. 1996) in 2021. In 2022, recruitment continued with some of the younger fish having moved into the adult age classes. In 2023, the Sandy Spat population was spread across both immature fish and those of spawning age.

Fig. 3.

The distribution of estimated lengths of Sandy Sprat (10 mm classes) in the 2021, 2022, and 2023 Crested Tern breeding seasons.

Discussion

Dietary studies from Crested Tern colonies in southern Australia and southern Africa have documented a diversity varying from 11 (Chiaradia et al. 2002) to 12 (Quiring et al. 2021), 36 (Mcleay et al. 2009b), and 53 (Gaglio et al. 2018) prey taxa. The 62 prey species taken from the Crested Tern colonies on Penguin Island was comparatively diverse. Of the 53 species captured at Robben Island in South Africa, 94% were classified as pelagic compared with 12.9% from the coastal waters around Penguin Island. The ‘non-pelagic’ taxa included species from shallow water habitats including seagrass meadows, sand flats and banks, macroalgae dominated limestone reefs, beach-wrack, and open water areas up to around 30 m in depth within Cockburn Sound, Shoalwater Bay, Warnbro Sound, and adjacent shelf waters. It is likely that this foraging habitat diversity contributed to the number of species recorded.

Generally, the diet of breeding Crested Terns is dominated by one or two forage fish species (at least in the prey fed to chicks). Anchovy has been an almost ubiquitous dominant forage fish in dietary investigations; e.g. making up 65% (Gaglio et al. 2018), 84.5% (Quiring et al. 2021), 63% (Chiaradia et al. 2002), and 36.3% (McLeay et al. 2009b) of the prey captured. Important (sometimes co-dominant) forage fish prey included Sardines and Jack Mackerel (Trachurus declivis). Anchovy and Sardine were the dominant prey off the western Cape of South Africa (Crawford 2003).

Between 1978 and 1987, Sardines were the dominant prey taken from Crested Tern colonies breeding in the Perth metropolitan region between April and June (Dunlop 1987; J. N. Dunlop pers. obs.). At Rottnest Island, the timing of the arrival of the Sardine shoals evidently influencing the onset of laying (Dunlop 1987). The autumn–winter breeding sub-season (Dunlop 1985) vanished from the local waters sometime after the Pilchard (Sardine) herpesvirus epidemics in 1995 and 1998 (Whittington et al. 2008) and has never re-established. Similar major shifts in colony persistence/location have been observed in southern Africa in response to changing forage fish (Sardine/Anchovy) distribution and abundance (Crawford 2003).

The diet of Crested Terns foraging from Penguin Island differs from all other colonies studied in that ‘sprats’, small coastal clupeids, tend to dominate the forage fish taken. This is also the case in comparison with Little Penguin colonies around the coastline of southern Australia (e.g. Chiaradia et al. 2010; Cannell et al. 2024). Sandy Sprats were a major contributor to the Crested Tern prey intake on Penguin Island in all three seasons. Blue Sprats were abundant in 2021 and early in the 2022 season but declined sharply in 2023. Coincidently, the later Crested Tern clusters (sub-colonies) abandoned their chicks soon after hatching in 2022 and this may have been related to the sudden disappearance of Blue Sprats at that time. Blue Sprats provide a prey-size that is suitable for newly hatched Crested Terns.

In 2021, Sandy Sprats were abundant in November with some Crested Terns returning to the Penguin Island colony with multiple bill loads (Fig. 4). In this season, the fish were mostly below 60 mm total length indicating the recruitment of new cohort (Fig. 3). The Year 0 fish were graduating into the adult (>60 mm) age classes in 2022 and in 2023, both Year 0 and older fish were represented indicating that the stock had become established. This suggests that Sandy Sprats, which are the most important prey type for breeding Little Penguins on Penguin Island, was available again.

Fig. 4.

Crested Tern with multiple bill load of Sandy Sprats. Photo: Brendan Kinsella.

The factors that contributed to the Sandy Sprat recovery are not known. However, 2021 was the second wettest year on record for the region (892 mm) in what is generally a drying climate (Bureau of Meteorology Report, 8 February 2022). The higher rainfall will have increased the outflow and the nutrient/organic matter plume from the Peel Harvey Estuary, elevating the productivity of the waters immediately south of Penguin Island within the Little Penguin foraging area. This in turn may have improved the growth and survival of Sandy Sprat larvae (Grimes and Kingsford 1996). The abundance of Sandy Sprat was demonstrated to be positively influenced by freshwater flows into the Murray River estuary (Bice et al. 2015).

Sandy Sprats disappeared from the Little Penguin diet on Penguin Island during and after the marine heatwave event in 2011 (Cannell et al. 2024). Tropical Sardine and Sardine dominated the penguin prey in 2011 and 2012 (Cannell et al. 2024). Between 2021 and 2023, tropical sardines made a negligible contribution to the prey taken by Crested Terns and the small number that were observed were in age/size classes above those preferred by the penguins (Cannell et al. 2024).

The potential return of Sprat within the foraging range of the Little Penguin may play a critical role in supporting the recovery of the Perth population, particularly the critically low Penguin Island colony. Further investigation into the relationship between Crested Tern diets and Little Penguin breeding productivity will be required to establish the value of this method in predicting periods of increased productivity. This could provide valuable guidance on management and indicate when intervention and conservation actions may be most beneficial.