Flatback Turtle (Natator depressus) Nesting at Field Island, Kakadu National Park, Northern Territory, Australia, 1990–2001
ABSTRACT
Field Island in Kakadu National Park, Northern Territory, Australia, supports a nesting population of flatback turtles (Natator depressus) monitored since 1990. We summarize 9 years of survey data for this location and make recommendations for future research. At least 221 nesting flatbacks were tagged at Field Island between 1990 and 2001. Nightly emergence rates indicate that Field Island is a locally and regionally important rookery for flatbacks with no other species recorded nesting. The mean interseasonal remigration period was 2.82 years. Female flatbacks laid clutches of similar sizes to those reported for other Australian rookeries, but nests tended to be deeper than those on nearby mainland beaches. No evidence of size decline (curved carapace length) occurred over the years, and mean curved carapace length was similar to those observed at other nesting rookeries. The amount of biological information now available for the Field Island rookery and the number of nesting turtles already tagged makes it valuable as an index site for flatback turtles in northern Australia.
Flatback turtle (Natator depressus) nesting is widely distributed along the tropical and subtropical northern Australian coast from Mon Repos in southern Queensland to Barrow Island in Western Australia (Bustard et al. 1975; Limpus et al. 1981; Prince 1993; Guinea 1994a; Chatto 1998). Much of the published information on flatbacks has come from long-term studies in Queensland where several populations have been systematically monitored since the late 1960s (e.g., Bustard et al. 1975; Limpus et al. 1981, 1984, 1993; Robins 1995; Hewavisenthi and Parmenter 2000). Additionally, several introductory studies have been conducted in the Northern Territory (e.g., Guinea 1994b; Hope and Smit 1998; Winderlich 1998) and Western Australia (Prince 1994). However, significant gaps exist in knowledge regarding population sizes and trends for parts of its distribution, and the flatback is the only species of marine turtle listed globally as “Data Deficient” by the IUCN Redlist (International Union for the Conservation of Nature 2006). The urgent need to increase census and demographic data for flatback turtles is recognized by Australian conservation authorities. These authorities have recommended establishing long-term tagging studies at “index” beaches throughout northern and western Australia, and identifying and quantifying mortality sources for all sea turtle species (Environment Australia and Marine Turtle Recovery Team 1998).
Kakadu National Park (KNP) is in the northeast Northern Territory. Flatback, green (Chelonia mydas), olive ridley (Lepidochelys olivacea), loggerhead (Caretta caretta), and hawksbill (Eretmochelys imbricata) turtles occur in the waters around the park (Winderlich 1998). Nesting in the park is predominantly by flatbacks, but nesting by olive ridley, green, and hawksbill turtles also has been reported (Winderlich 1998). Preliminary surveys in 1990 (Winderlich 1998) and 1993–1994 (Vanderlely 1997) indicated that most flatback nesting in the park occurred on Field Island, with relatively little nesting occurring on mainland beaches. In 1994, KNP staff initiated a monitoring program for nesting turtles on Field Island and confirmed the presence of a sizeable nesting population of flatback turtles. As a result, and because of the park's designation as a protected World Heritage site, Field Island has been identified as a key monitoring site within a national framework (Environment Australia and Marine Turtle Recovery Team 1998). Given the limited number of flatback nesting beaches for which long-term data sets are available and the importance of such data sets for local and regional management, we provide a summary of tagging and associated biological data collected from flatback turtles nesting at Field Island between 1990 and 2001. We also make several recommendations regarding survey protocols and future research work on the island, many of which are already being addressed by KNP managers.
METHODS
Field Island (12°7.117′S, 132°25.19′E) is a small, uninhabited, offshore island forming part of the KNP in the Northern Territory, Australia (Fig. 1). Turtle nesting activity is almost entirely confined to a small (ca. 300 m) stretch of beach on the northeast coast of the island and mostly occurs between July and September, with occasional nesting throughout the year. Shallow waters offshore from the nesting beach with mud and rocky substrates coupled with a large tidal range (6–7 m) inhibits beach access by turtles from low to around mid-tide. Hence, tagging surveys of 8–15 days were timed to include a period of high or spring tides that favored beach access, and nightly beach patrols were usually conducted for 2–3 hours on either side of high tide.
Citation: Chelonian Conservation and Biology 5, 2; 10.2744/1071-8443(2006)5[188:FTNDNA]2.0.CO;2
Nesting turtles were measured (curved carapace length, CCL), weighed, and tagged with standard titanium flipper tags (Limpus 1985, 1992a). When a turtle was measured more than once per season we used the mean of all measurements. In 2001, all turtles also were implanted with Passive Integrated Transponder (PIT) tags in the right shoulder. PIT tags were used to overcome problems caused by the high rates of flipper tag loss that flatbacks typically experience (Balazs 1999; Parmenter 1993). Turtles were not systematically examined for flipper tag loss scars prior to 2001. In some years, and for some turtles, the following data were also recorded: duration of beach ascent and nesting phases, clutch size, and depth at the top and bottom of the nest. Emerged nests were excavated opportunistically to calculate clutch hatching and emergence success (% of eggs that result in hatchlings hatching from eggs and emerging from the sand surface, respectively) according to methods in Limpus (1985) and Miller (1999). However, it is impossible to estimate the total number of clutches that had zero success because such clutches would not have produced obvious visual clues to their locality, and staff would have been extremely unlikely to find them unless there was evidence of predator interference. Our statistical analyses were conducted in STATISTICA, v. 6 (StatSoft Inc 2001). All means are presented with standard deviations.
RESULTS AND DISCUSSION
Numbers of Nesting Females and Recapture Rates
Two hundred twenty-one individual female flatback turtles were first-time tagged at Field Island. Additionally, 14 tagged turtles were encountered for which there were no records of initial tagging. We assume from tag number sequences that these were first tagged either on Field Island, elsewhere in Kakadu National Park, or on Greenhill Island around 60 km to the north (Hope and Smit 1998). Eighteen turtles were recorded as having tag scars caused by the loss of previously applied titanium tags (Fig. 2). Flatback turtles were the only species encountered nesting on Field Island.
Citation: Chelonian Conservation and Biology 5, 2; 10.2744/1071-8443(2006)5[188:FTNDNA]2.0.CO;2
On average, 4–5 nesting emergences occurred per night, with a maximum of 10–16 (Table 1). This is greater than the 1–2 nesting turtles per night on mainland Kakadu beaches (C. Schäuble, unpubl. data, 2002), and similar to the nightly nesting numbers observed on nearby Greenhill Island (C. Schäuble, unpubl. data, 2002). Although likely to be locally and regionally important, the nightly emergence rate on Field Island falls below the rates observed at the largest known nesting populations in Australia. For example, estimates of 15–30 flatback turtles per night have been recorded for Western Australia's largest known rookery (Prince 1994). Limpus et al. (1989) observed up to 69 emerging flatback turtles per night at Deliverance Island in the Torres Strait, and Limpus et al. (1983, 1993) observed 6–235 flatback turtles per night on Crab Island in the eastern Gulf of Carpentaria (likely the largest flatback rookery in Australia).
There was no obvious evidence of a steep decline in nesting numbers based on emergences per night (Table 1). However, the utility of the 1990–2001 survey data from Field Island in assessing overall population trends is restricted because survey timing may not have always coincided with peak nesting, nights with zero emergences may not have been recorded in earlier years, and levels of tag loss are unknown (C. Schäuble, unpubl. data, 2002). Rates of tag loss could not be quantified because tag scars were not systematically recorded in all years; therefore, the data are not sufficient to support demographic modeling to estimate nesting population recruitment rates (Parmenter and Limpus 1995).
Tagging efforts on Field Island have been sufficiently intense to result in over half the nesting turtles encountered in 2000 and 2001 showing evidence of being previously tagged (Fig. 2), and substantial numbers of individual turtles have been recorded nesting in multiple years (Fig. 3). However, these recapture rates likely are underestimated because turtles were not systematically checked for tag loss prior to 2001. Individual turtles were recaptured in up to 4 different years (seasons) (Fig. 3), yielding a mean interseasonal remigration period (ISRP, the number of years between successive breeding seasons for individual females) of 2.82 ± 1.28 years and modal ISRP of 2 years (range 1–7 years) (Fig. 4). These estimates compare to those from Peak Island in Queensland (mean = 2.2 years, range = 1–5; Parmenter and Limpus 1995); and Mon Repos in Queensland (mean = 2.7 years; Limpus et al. 1984). The longer ISRPs (6–7 years) observed at Field Island may have resulted from tag loss, from tagged turtles missed by beach patrols in the intermediate years, or because the within-season renesting interval for some turtles may have been longer than the total survey period; i.e., a turtle may have nested just prior to, and just after, the survey period for that year. The use of PIT tags and longer survey periods after 2001 spanning the mean flatback renesting interval of 16 days (Limpus et al. 1984; Miller 1997) will assist in remedying these deficiencies in the dataset.
Citation: Chelonian Conservation and Biology 5, 2; 10.2744/1071-8443(2006)5[188:FTNDNA]2.0.CO;2
Citation: Chelonian Conservation and Biology 5, 2; 10.2744/1071-8443(2006)5[188:FTNDNA]2.0.CO;2
Nesting Behavior
The mean time taken for flatback turtles to complete the nesting process; i.e., emerging, body pitting, egg chambering, egg laying, and covering the nest, was 77 ± 20 minutes (n = 46). Mean times for climbing the beach to the high water mark (n = 24) and the slope/dune (n = 41) were both 3 ± 2 minutes. On average, females spent 19 ± 9 minutes body-pitting (n = 45), 19 ± 11 minutes digging an egg chamber (n = 44), 14 ± 5 minutes laying eggs (n = 45), and 23 ± 9 minutes covering the nest (n = 46). The mean duration of each nesting activity was similar to that recorded for five flatbacks in Queensland by Bustard et al. (1975); although Field Island flatbacks appear to have spent slightly more time body-pitting.
Morphology of Nesting Turtles
The mean CCL of Field Island flatback turtles (Table 2) fell within the range of sizes observed for flatback turtles at other nesting rookeries, but was more similar to mean CCL from other northern Australian locations (Fog Bay and Crab Island) than those from southeastern Queensland (Limpus et al. 1989, 1993; Guinea 1994b). Declines in CCL have been seen in green turtle nesting populations in the Great Barrier Reef and are interpreted as changes in the ratio of first-time nesters to experienced nesters and possible first signs of a population decline (Hamann 2002; Limpus et al. 2003). We found no evidence of a change in mean CCL across years for flatbacks nesting at Field Island (Kruskal-Wallis H = 8.619, df = 8, p = 0.375).
However, the absolute difference between any two CCL measurements on a single turtle within a season ranged between zero and 4.0 cm, with a mean of around 1.0 cm. Adult marine turtles have negligible growth (Limpus 1992b; Chaloupka and Limpus 1997), estimated at around 0.02 cm/y for flatback turtles (Parmenter and Limpus 1995), thus, significant measurement error exists in at least part of the Field Island data. Measurement error is an inherent part of the total error in morphological data sets, and tends to be larger when multiple people are performing the measuring (Frazier 1998). The amount of measurement error has important implications for the subsequent uses of that data (Shoop and Ruckdeschel 1986; Bjorndal and Bolten 1989; Frazier 1998; Bolten 1999). The level of measurement error in the Field Island CCL data for 1990–2001 limits its utility in calculating parameters such as long-term growth rates and minimum size at sexual maturity (Shoop and Ruckdeschel 1986; Bresette and Gorham 2001). Changes to training and data collection protocols post-2001 at Field Island should reduce the problem of measurement error (Shoop and Ruckdeschel 1986; Frazier 1998).
Clutch Size and Nest Morphology
The mean clutch size (Table 2) of flatback turtle nests laid at Field Island was similar to that observed on mainland beaches in KNP in 1993–1994 (Vanderlely 1997; C. Schäuble, unpubl. data, 2002) and at other Australian rookeries (Limpus 1971, 1981, 1989, 1993; Guinea 1994c; Parmenter and Limpus 1995; Hewavisenthi and Parmenter 2002). However, flatback nests at Field Island were deeper on average than those at mainland sites in the Kakadu region (Vanderlely 1997; C. Schäuble, unpubl. data, 2002). Similarly, data from previous studies also indicate there may be significant geographic variation in the depth of flatback turtle nests (Limpus et al., 1981, 1989, 1993; C. Schäuble, unpubl. data, 2002). However, the causes of geographic variation in nest depth remain untested.
Clutch Success
Hatching and emergence success were high at Field Island compared to those reported for mainland KNP beaches (Vanderlely 1997; C. Schäuble, unpubl. data, 2002). Mean clutch hatching and emergence successes for clutches producing emerged hatchlings were 88% ± 17% and 64% ± 32%, respectively (n = 16).
CONCLUSIONS AND RECOMMENDATIONS
This paper provides the first detailed study published on flatback turtle nesting demographics from the Northern Territory genetic stock, summarizing 9 years of tagging data from the Field Island rookery in KNP, Northern Territory. While the size of the Field Island rookery is comparatively small, other parameters such as turtle and clutch sizes and interseasonal remigration periods were consistent with those of other genetic stocks. Within the comparatively short study period (c.f. turtle generation times), there appears to have been no substantial change in either mean CCL or the annual index of nesting numbers (calculated as mean number of turtles emerging per night). However, problems arising from interannual variation in survey timing duration indicate that these conclusions should be regarded as tentative until additional years of data are available. Changes to data collection and management protocols initiated by KNP staff since 2001 will improve the utility of the data for monitoring long-term population trends.
The effort already invested in monitoring the Field Island rookery (over a decade of data and more than 50% of the nesting population now tagged), coupled with the accessibility of the island and its protected location in a World Heritage National Park, justifies maintaining Field Island as a key habitat and national index site under the Recovery Plan for Marine Turtles in Australia (Environment Australia and Marine Turtle Recovery Team 1998).
Based on our experience with the Field Island study and elsewhere in north Australia, we offer several recommendations regarding the monitoring of flatback turtles at Field Island and elsewhere. First, for a detailed analysis on long-term data sets, it is imperative that data collection methods are standardized and followed for all surveys. In the case of nesting beach census work, this must include the timing of the surveys because the numbers of nesting females that visit a beach/island per night varies significantly with both tidal cycles as well as the number of days since the start of the season (Dobbs et al. 1999; Limpus et al. 2001; Hamann 2002). Thus, field surveys should be conducted at the same time each year, and for at least the length of one renesting interval. To maximize data utility, nesting beach surveys should aim to coincide with peak nesting effort. This requires data on the annual distribution of nesting effort at a site so that the nesting numbers index for single renesting interval surveys can be placed in a whole-season context.
Second, to assist population monitoring and management at a local and regional scale there is a need to quantify the total number of clutches laid per season, hatchling production (specifically clutch failure and predation rates by feral and native animals; e.g., Engeman et al. 2003). A multiyear study to assess clutch success—including predation rates on nests—was initiated on Field Island in 2001, and results will be reported elsewhere.
Third, tagging efforts should continue to use both standard titanium flipper tags and PIT tags to reduce the tag loss problems typical of flatbacks. Changes to operating procedures at the Field Island site will ensure future assessments of the rate and consistency of tag loss and hence facilitate estimates of population size, recruitment rate, remigration interval, and survivorship (Limpus 1992a; Chaloupka and Musick 1997; Gerrodette and Taylor 1999). Effective, sufficient tagging is required to enable managers to detect serial changes in abundance and provide a better chance of distinguishing natural change from that caused by anthropogenically driven impacts.
Fourth, satellite telemetry and feeding ground studies should be initiated to locate the feeding grounds of flatback turtles that nest at Field Island, and to determine the species composition, demographic structure, and nesting rookeries of the feeding marine turtles inhabiting the waters surrounding Field Island. Effective management of these migratory inhabitants of KNP requires information on the status of the other habitats they utilize.
Location of Field Island, Northern Territory, Australia. Main map: Northern Territory shown in grey. Inset: Field Island shown in black, other areas of Kakadu National Park are shaded grey.
Relative frequency (across years) of first-time tagged turtles, interseasonal remigrant turtles with tags on, interseasonal remigrants without tags recognized from tag scars, and missed turtles as observed during Field Island surveys. Only 1 nesting event is included for any individual tagged turtle in any 1 year. Note that, except for 2001, data were not available on how consistently untagged turtles were examined for tag scars.
Recapture frequency for individual flatback turtles (Natator depressus) across years at Field Island. This graph includes females first tagged at Field Island (221), females with tags but for which primary tagging information was not available (14), and females that had no tags but had been tagged previously (tag scars) (18).
Frequency of different interseasonal remigration periods (ISRP) for flatback turtles (Natator depressus) tagged and resighted in subsequent years at Field Island. Field Island was first surveyed in 1990 and then every year between 1994 and 2001. Some individual females were observed repeatedly across many years, others only twice (Fig. 3). Therefore, ISRP values contained in this graph are not strictly independent, and the possibility of bias from particular females has not been assessed.
