Foraging habitat locations of flatback (Natator depressus) and olive ridley turtles (Lepidochelys olivacea) in northern Australia

Introduction

Understanding the spatial ecology of threatened migratory species is a priority for conservation and management (Hyrenbach et al. 2000; Shuter et al. 2011). The spatial distribution of migratory species varies extensively, within and among species, particularly throughout foraging periods (Hussey et al. 2015). Foraging areas can support several genetic stocks of multiple taxa, occurring across administrative and legislative boundaries (Blumenthal et al. 2006; Whiting et al. 2007; Whittock et al. 2016) or representing strong fidelity with highly localised, well-defined feeding locations (Hawkes et al. 2006; Schofield et al. 2010).

It is challenging to manage threats to species or populations that undertake large-scale movements or use large home ranges, such as marine turtles, because they can occur across jurisdictional boundaries and often overlap with multiple anthropogenic stressors (Hays et al. 2019; Miller et al. 2019). Further, travel between various habitats, such as those used for breeding and foraging, could mean that turtles move between areas with different levels of threat exposure, such as protected breeding areas and foraging areas with high shipping traffic or fishery pressures (Blumenthal et al. 2006; Pendoley et al. 2014; Thums et al. 2017). As such, understanding spatial movements of individuals or species can provide robust information for conservation by detailing specific habitats and migration corridors, while highlighting potential threats and population implications (Lucchetti et al. 2016; Fuentes et al. 2020).

Satellite telemetry provides an advanced understanding of spatial distribution and movement behaviours of marine species (Hays and Hawkes 2018). Individuals can be tracked over extended periods, but briefly within the overall life span of the turtles, and their fine-scale movements can be monitored with high position accuracy (Godley et al. 2008). Satellite telemetry, such as fastloc GPS, coupled with data-filtering measures and application techniques, allows finer-scale detail of movement characteristics, indicating specific spatial and temporal attributes of movement behaviour, such as home ranges and movement within them (Godley et al. 2002; Polovina et al. 2004). Consequently, satellite-tracking is increasingly able to provide the information necessary to inform management and conservation policies (Hays et al. 2019; Armstrong et al. 2020; Santos et al. 2021). Increases in available data can aid implementation and inform management initiatives at local and regional scales. They may indicate specific threat implications and species interactions at remote localities within complex environments such as the broad expanse of neritic waters of northern Australia–New Guinea continental shelf (Shimada et al. 2017).

Olive ridley (Lepidochelys olivacea) and flatback (Natator depressus) marine turtles are generally less understood than are other marine turtle species breeding within Australia, with knowledge of their foraging ecology in northern Australia particularly lacking (Limpus 2009; Department of Environment and Science 2021). The paucity of information is likely to be due to remote nesting locations, foraging areas and non-specific migratory paths (Limpus 2009). In general, both species typically inhabit the broad expanse of subtidal neritic waters of the continental shelf of remote, tropical, northern Australia and southern New Guinea (Robins et al. 2002; Whittock et al. 2016), restricting available survey methods (Whiting et al. 2007), except via fisheries bycatch data (Poiner et al. 1990). Efforts to determine spatial coverage and abundance in northern Australia have been made for both flatback and olive ridley turtles, and, although limited, these data have provided a generic understanding of species distribution, foraging areas and areas that are exposed to threats (Robins 1995; Dryden et al. 2008; Wilcox et al. 2013; Riskas et al. 2016). However, ecological information at a smaller habitat-use scale and migration routes are lacking, hindering targeted protection/management efforts or designation of critical habitat.

Consequently, this study aims to determine and describe the foraging locations of flatback turtles from the Arafura Sea (AS) genetic stock and olive ridley turtles of the eastern Gulf of Carpentaria (GoC) genetic stock (FitzSimmons and Limpus 2014) by using satellite telemetry. To achieve these aims, we attached satellite-tracking tags to post-nesting female turtles to discover the spatial and temporal characteristics of their post-nesting foraging activity. This will provide insight into the spatial ecology of these species that can be used to develop and implement management strategies that mitigate adverse anthropogenic pressures on marine turtle populations in northern Australia.

Materials and methods

Results

Foraging home range

Flatback turtles were tracked for a mean of 264 ± 61.39 days (range: 43–871), and olive ridley turtles were tracked for a mean of 71 ± 5.40 days (range: 43–92) (Table 1). Most individuals of both species showed comprehensive movement (i.e. 95% UD) around a core (50% UD) foraging areas, which ranged from 1 to 1065 km² (mean value 173 km², n = 16) (Fig. 1) and from 2 to11 km² (mean value 33 km², n = 8) (Fig. 2) for flatback and olive ridley turtles respectively (Table 2).

Fig. 1.

Distribution of flatback turtle migration and foraging areas (50% and 95% utilisation distributions) for individuals tracked from northern Australian rookeries. Tagging sites L–R; Greenhill Island, Waral Kawa, Jardine River.

Fig. 2.

Distribution of olive ridley turtle migration and foraging areas (50% and 95% utilisation distributions) for individuals tracked from north-eastern Gulf of Carpentaria (western Cape York Peninsula) rookeries. Inserted boxes represent foraging areas identified by prior studies: (a) McMahon et al. (2007) and Whiting et al. (2007); (b) Whiting et al. (2007); (c) Whiting et al. (2007); Fukuoka et al. (2022); (d) McMahon et al. (2007); Fukuoka et al. (2022); and (e) McMahon et al. (2007). Tagging site is Mapoon.

Table 2.The foraging time (days), core foraging area (50% UD), foraging area (95% UD) and average depth of foraging (m).

Species	Turtle ID	Foraging time (days)	Core foraging area 50% UD (km2)	Foraging area 95% UD (km2)	Average depth used (m)
N. depressus	108468	215	257	5811	−27.73
N. depressus	54530	58	212	1375	−28.75
N. depressus	64208	162	257	2795	−62.39
N. depressus	95894	521	1065	12410	−69.97
N. depressus	95896	567	532	11369	−56.73
N. depressus	133406	32	52	253	−14.24
N. depressus	133407	63	31	162	−57.87
N. depressus	133410	80	125	1103	−49.31
N. depressus	140117	103	67	530	−3.66
N. depressus	140119	47	32	306	−6.18
N. depressus	111064	269	47	250	−34.26
N. depressus	129555	232	31	232	−15.17
N. depressus	88239	808	1	74	−20.7
N. depressus	133409	17	30	218	−17.91
N. depressus	133405	50	25	114	−2.28
N. depressus	133408	8	10	77	−53.71
L. olivacea	K65252_48845	13	2	11	−14.54
L. olivacea	K90229_133765	10	3	20	−16.61
L. olivacea	QA39899_140118	77	12	328	−2.81
L. olivacea	K90108_48850	52	19	600	−60.61
L. olivacea	QA14005_48850	30	113	623	−51.74
L. olivacea	QA14046_48845	61	28	1134	−12.94
L. olivacea	QA23520_48850	22	16	71	−14.25
L. olivacea	QA57043_48845	32	76	1099	−110.90

The tracking data identified foraging areas for flatbacks in regions of the northern Torres Strait, south-eastern GoC, central GoC, Kakadu coast, Northern Territory, Joseph Bonaparte Gulf, Western Australia, and Indonesia (Fig. 1). Home-range analyses for the olive ridleys identified four separate foraging areas, namely, two in Indonesian waters, one in eastern GoC and one in the central GoC (Fig. 2). Of the four olive ridley foraging areas identified in this study, two were within a previously identified foraging area (Fig. 2).

All individuals demonstrated multiple interconnected core foraging areas (Figs 1, 2). Seven individuals had foraging areas in Indonesian waters of the Arafura Sea (flatback: 64208, 95896, 108468, 54530, 133406; olive ridley: QA14005_48850, QA57043_48845) and two had foraging areas in the Papua New Guinean region of the coral sea (133405, 140119); five of these individuals also recorded foraging areas in Australian waters. Overall, we identified foraging areas in Indonesian waters for flatback (n = 7) and olive ridley (n = 2) turtles from breeding sites in northern Australia.

There was an overlap in the overall home range between the two species (95% UD) at three separate locations. First, there is an area of 54.8 km² in Indonesian waters approximately 180 km west of Indonesia and Papua New Guinea, shared by a flatback individual from Waral Kawa (64208) and an olive ridley individual from Flinders Beach (QA14005_48850) (Figs 1, 2). Second, there is an area in the central GoC, 320 km north of the Northern Territory coastline and 250 km off the Queensland coastline, which supports foraging by both species. Flatback 95896 and olive ridley (K90108_48850) overlapped 21.7 km² in this area. The third area utilised by both species is located in the eastern GoC, approximately 10 km from the Queensland coast, covering 56.7 km². Here, there was an overlap between a flatback tracked from Crab Island (140117) and an olive ridley tracked from Mapoon (QA14046_48845). The area of overlap specified in Indonesian waters has also been used by turtles tracked in previous studies (Fig. 2; Whiting et al. 2007; Fukuoka et al. 2022).

Seasonal foraging

Two flatback individuals (95894 and 95896) used large foraging areas over the study duration (Table 3). Analysis of seasonal changes in foraging activity indicated that turtles 95894 (Fig. 3) and 95896 (Fig. 4) show distinct seasonal shifts in the location of the 50% UD area, with most consecutive seasons showing more than 90% of season-to-season shifts of core foraging to have changed (Table 4). However, the overlap of 95% UD is more prevalent for both individuals between seasons and a lower percentage, but still >60%, the percentage of area occupied has changed (Table 5). For both individuals, there was little consistency in the use of space across corresponding seasons when examining core foraging (50%UD) activity (Figs 3, 4).

Table 3.Total area used by flatback individuals 94894 and 95896 over 50% and 95% utilisation distributions (km²).

Season	95894 UD 95 (UD 50) (km2)	95896 UD 95 (UD 50) (km2)
Autumn 2013	2168 (211)	1757 (88)
Winter 2013	3184 (191)	2906 (125)
Spring 2013	3296 (321)	1988 (207)
Summer 2013/2014	1724 (258)	1372 (66)
Autumn 2014	2927 (473)	2964 (150)
Winter 2014	1711 (191)	1199 (211)
Spring 2014	1587 (222)	1333 (211)

Fig. 3.

Individual foraging core area (50% UD) for flatback individual 95894 in the Joseph Bonaparte Gulf, highlighting space use among and within seasons.

Fig. 4.

Individual foraging core area (50% UD) for individual 95896 in the Gulf of Carpentaria among seasons, highlighting space use among and within seasons.

Table 4.Proportion of new core foraging (50% UD) area relative to the area used between consecutive seasons for flatback individuals 95894 and 98596.

Consecutive seasons	95894 UD 50 (%)	95894 UD 50 (km2)	95896 UD 50 (%)	95896 UD 50 (km2)
Autumn 2013 – winter 2013	99	189	100	125
Winter 2013 – spring 2013	99	318	98	203
Spring 2013 – summer 2013/2014	90	233	96	63
Summer 2013/2014 – autumn 2014	89	422	100	105
Autumn 2014 – winter 2014	70	136	100	211
Winter 2014 – spring 2014	100	222	91	190

Table 5.Proportion of new 95% UD foraging relative to the area between consecutive seasons for individuals 95894 and 95896.

Consecutive seasons	95894 UD 95 (%)	95894 UD 95 (km2)	95896 UD 95 (%)	95896 UD 95 (km2)
Autumn 2013 – winter 2013	89	2822	96	2776
Winter 2013 – spring 2013	90	2959	88	1758
Spring 2013 – summer 2013/2014	81	1391	89	1220
Summer 2013/2014 – autumn 2014	79	2341	99	2946
Autumn 2014 – winter 2014	61	1037	98	1169
Winter 2014 – spring 2014	99	1578	75	996

Discussion

Post-nesting flatback and olive ridley turtles in northern Australia have large spatial distributions and widely dispersed foraging grounds across different jurisdictions. In the longest flatback dispersal, one turtle moved 1660 km over 55 days from Waral Kawa in Torres Strait to a foraging area in north-western Western Australia. The longest olive ridley migration was 844 km over 50 days from Mapoon, Queensland, into Indonesian waters. Our data are similar to other data from the species elsewhere in Australia, which collectively demonstrates that both species move long distances between nesting and foraging grounds and variable-sized foraging areas are distributed exclusively within neritic continental shelf waters of northern Australia–southern New Guinea (McMahon et al. 2007).

Home-range analysis for flatbacks showed high inter-individual variation in the size of the core (50% UD) foraging area, ranging from 1 to 1065 km² (mean = 173 km²). It consequently demonstrated that flatback turtles in this region can move extensively during foraging across large areas, and, in general, more than any other marine turtle species tracked for similar durations in Australia (Block et al. 2011; Gredzens et al. 2014; Shimada et al. 2016, 2020; Barr et al. 2021). Similarly, analysis of olive ridley home ranges showed a large variation in core (50% UD) areas of 2 to 113 km² (mean = 34 km²) for four individuals throughout the GoC (n = 2) and Indonesian waters (n = 2). The upper limit of the home-range area is similar to that found for individuals tracked from the Northern Territory, where Whiting et al. (2007) found olive ridley home ranges to vary between 138 and 1260 km². Together, these results demonstrate that olive ridley turtles can forage over large areas. Still, the results of the present study also indicated that individual olive ridley and flatback turtles can constrict their foraging activity to a small specific region, presumably if resources allow.

When comparing the foraging areas identified in this study to results from other studies, the partial overlap of foraging areas between females from genetically distinct populations is evident (Whittock et al. 2016). In particular, the collective research conducted on the spatial ecology of flatback and olive ridley turtles in northern Australia has indicated that the central and northern waters of the Arafura Sea, including those in Indonesian waters, are important foraging sites for both species, with nine individuals using multiple foraging core areas (50% UD). In addition, it has given additional strength to justify a regionally significant foraging area in the Joseph Bonaparte Gulf region of Northern Territory, Western Australia, and the adjacent Commonwealth waters. This strengthens the need for cooperation in the management or development of regional-focused strategies for managing pressures on the marine environment across the northern Australian and southern Indonesia region (Miller et al. 2020; Pilcher and Welly 2021).

Two flatback individuals, tracked for over a year, displayed seasonal-based variations in their foraging home ranges. In both cases, turtles used separate 50% UD core foraging areas across seasons, highlighting the potential for temporal variation in foraging areas. Previous evidence of seasonal influence on adult marine turtle foraging habitat use was recorded for green and loggerhead turtles in southern Queensland and loggerhead turtles in the Mediterranean (Shimada et al. 2016; Dujon et al. 2018). Although the spatial scales are different, and the green turtles moved only short distances, in this study, the general pattern of our results is similar in that the tracked turtles used different foraging areas across seasons. Flatback turtles seem to require large areas relative to some other hard-shelled turtles (e.g. greens, loggerheads, hawksbills), presumably because their primary diet is sparsely distributed or less predictable benthic and planktonic soft-bodied invertebrates (Limpus 2009). That said, while there are no data from any study to indicate whether flatback turtles have the capacity to react to changes in food accessibility, if they are like other hard-shelled species, they will tend to remain in their familiar home habitats, possibly because shifting locations to unknown areas could be too risky, for example, because of poor knowledge of food locations or shelter from predators in the new habitat. Therefore, despite the occupancy of large areas, flatback turtles are likely to be vulnerable to aperiodic changes in habitat quality or availability of food.

From bathymetry projections, we can infer that flatback and olive ridleys typically reside in habitats deeper than 10 m, which supports knowledge from previous research and studies of bycatch (Poiner et al. 1990; Robins 1995; Whittock et al. 2016; Wildermann et al. 2017). All home ranges indicated a foraging depth range that extended from 2 to 70 m for flatbacks and from 3 to 111 m for olive ridley turtles. However, correlating spatial-use data with spatial bathymetric data does not provide information on an individual’s diving behaviour and depth use. Future studies could combine telemetry with depth recorders to obtain specific data on dive behaviour and foraging (e.g. Hounslow et al. 2023). However, our data concur with those of studies by Whiting et al. (2007), Whittock et al. (2016) and Thums et al. (2017), which indicated that flatbacks from the Pilbara populations use mean water depths of 36.5 ± 22.5 m, whereas olive ridley populations from Northern Territory rookeries are reported to forage in water depths between 20 and 150 m, also on the basis of bathymetry information rather than diving data.

Foraging data show that flatbacks and olive ridleys use coastal and open-water habitats from nearshore areas and up to 225 km offshore. From previous tracking overlap, we can confirm that the foraging areas identified within the Joseph Bonaparte Gulf are remote, difficult to access and comprise various habitat types (Whiting et al. 2007; Whittock et al. 2016). Given the remoteness of the area, little information is available on bathymetry and biophysical features of this region. Inference of resource quality and availability can be made only in certain areas concerning fisheries’ productivity or satellite-derived chlorophyll layers, and this will not illuminate the trophic structure and complete available food source. Flatbacks feed primarily on benthic and planktonic soft-bodied invertebrates (Limpus 2009), reacting to specific environmental predictors influencing prey distribution (Wildermann et al. 2017). Benthic-community variability may significantly affect species distribution and foraging behaviours, because individuals must constantly move to follow mobile or unreliable food sources or seek specific habitats. Information on the triggers of community variability in particular regions may be used to indicate potential population movement and behaviour. Further dietary-analysis details will help highlight and determine the significant resources of both species, providing possible insight into population distribution patterns and may aid in management initiatives of specific populations and habitat areas.

Further research into the links between space use and foraging ecology may explain and determine resource use and availability for marine turtles in the northern Australian region. Identifying the predictors that drive responses can help determine species and population distribution patterns. However, telemetry-tracking of a small number of turtles from restricted breeding sites within a stock will almost certainly underestimate the actual foraging range of marine turtles (Shimada et al. 2021; Perez et al. 2022). Flipper-tag recoveries (n = 3) from flatback turtles tagged while nesting at eastern Gulf of Carpentaria nesting beaches for the AS genetic stock have been recorded within Papua New Guinea coastal waters up to 500 km east of Torres Strait (C. J. Limpus, unpubl. data). Overall, this study has shown that flatbacks and olive ridleys of northern Australia are capable of large-scale movement patterns, reside in large foraging habitats and exhibit intra-annual shifts in their core habitats. Future research should concentrate on elucidating the factors influencing these behaviours. Focusing on spatial population ecology, resource type, and availability will help develop a management protocol to aid species recovery by outlining potential threats from known spatial and temporal movement characteristics, identifying the resources that may influence specific distribution and setting boundaries of areas such as biologically important areas, or as a basis for other spatial-based threat mitigation.