Appendix F25. Guinea Long Term Marine Turtle Monitoring at Scott Reef. Browse FLNG Development Draft Environmental Impact Statement

Transcription

1 Appendix F25 Guinea 2009 Long Term Marine Turtle Monitoring at Scott Reef Browse FLNG Development Draft Environmental Impact Statement EPBC 2013/7079 November 2014

2 Long Term Monitoring of the Marine Turtles of Scott Reef Michael L. Guinea Charles Darwin University Darwin 0909 Northern Territory 6 May 2009

3 Executive Summary Woodside Energy Ltd (Woodside), on behalf of the Browse Joint Venture, engaged Sinclair Knight Merz (SKM) in conjunction with Dr M. Guinea of Charles Darwin University (CDU) to undertake surveys in 2008 and 2009 at Sandy Islet, Scott Reef, to establish baseline marine turtle nesting activity at this location. This report presents the results of seven surveys of marine turtle nesting activity at Sandy Islet. Three surveys were undertaken in 2006 by Dr M. Guinea and URS, on behalf of Woodside, and four surveys were conducted in 2008/2009 by Dr M. Guinea and SKM, on behalf of Woodside. The dataset is summarised to indicate the current knowledge of marine turtle nesting activity at Scott Reef. The objectives of the seven surveys were to: describe the species and abundance of marine turtles present at Sandy Islet; collect nesting location, nesting success and hatching success information for the marine turtles at Sandy Islet; collect morphometric information (carapace length and width) for the marine turtle species at Sandy Islet; collect marine turtle skin tissue DNA for analysis; and undertake opportunistic random surveys to identify internesting habitat. A total of 435 nesting green turtles were tagged during the 2006, 2008 and 2009 surveys. Of those, 314 were female green turtles that had been tagged since August A preliminary estimate of the size of the population using sightings of painted female turtles was 779 ± 383 (± se) from eight days of survey over the 2008/2009 breeding season. However, these estimates are highly variable and may not accurately reflect the nesting population as none of the assumptions used to estimate the population size have been validated. There was considerable inter-annual variability in the maximum mean track count per night with a maximum mean (± sd) track count/night of 16.2 ± during the February 2006 survey compared to the mean track count/night of 71.9 ± 6.98 in January The most accurate method to measure the size of a turtle nesting population would be to undertake a whole-ofseason tagging census of the rookery. The mean number of nests/night ranged from zero (September 2006) to 29.1 ± 4.5 (January 2009). The mean size of clutches for nests laid at Sandy Islet were 85.3 ± 21.7 (n = 14), while hatching success was 84.4 % ± 8.4 and emergence success was 75.6 % ± These data are consistent with other rookeries within northern Australia. Density dependent mortality was evident in nesting success and hatching success. Such processes are natural when large numbers of marine turtles compete for limited beach space.

4 An area bordering the south of Sandy Islet, with a sandy substrate, was identified as an important internesting and within nesting habitat where breeding female green turtles congregate.

5 Table of Contents 1 Introduction Project Description Location Conservation Value Previous Studies Species Background Scope of Works Methods Survey Dates Population and Abundance Estimates Nesting Marine Turtle Survey Beach Patrol Marine Turtle Sightings Tagging and Marking Protocols Estimates of Population Size Flipper Tags Mark and Recapture Morphometrics of Green Turtles Nesting Females Hatchlings Skin Biopsy Collection Clutch Size and Nesting Success Environmental Data Hatched Nest Identification Nest Excavation Examination of clutch Marine Turtle Habitats Results Population Abundance and Estimates Green Turtle Nesting Activity Inter-Season and Within Season Recaptures Estimates of Population Size Morphometrics Biopsy Analysis Clutch Size, Hatching Success and Emergence Success i

6 3.4.1 Environmental Temperatures: Marine Turtle Habitats Nesting Habitats Internesting Habitat Discussion Population Abundance Nesting Location, Nesting Success and Hatching Success Morphometrics of Green Turtles Internesting Habitat Genetics and Connectivity with Foraging Areas Conclusion Acknowledgements: References Appendix A Appendix B ii

7 List of Tables Table 2-1: Timing and duration of the seven survey trips to Scott Reef between 2006 and Table 3-1: Breeding seasonality of green turtles on Sandy Islet, Scott Reef, as demonstrated by the mean (± standard deviation, sd) number of nests and tracks per night during the survey periods that spanned most seasons Table 3-2: Petersen estimates with Bailey s correction and standard error (SE) of estimates for eight recapture dates at Sandy Islet, Scott Reef Table 3-3: Sizes of nesting female green turtles at Sandy Islet, Scott Reef, in the 2005/2006, 2006/2007 and 2008/2009 breeding seasons and for all seasons combined Table 3-4: Numbers of marked female green turtles and unmarked green turtles along the reef crest and adjacent waters at Sandy Islet Table B 1: Clutch sizes recorded after excavation of the hatched nest from this study and from other green turtle nesting areas in northern Australia Table B 2: Hatching success recorded from this study and from other green turtle nesting areas in northern Australia Table B 3: Hatchling emergence success by green turtle hatchlings on Sandy Islet and those from other northern Australian green turtle rookeries Table B 4: Comparison of the Curved Carapace Length (CCL) of green turtles in northern Australian rookeries from this study and the literature Table B 5: Comparison of Nest Depth (cm) of green turtle hatched nests on Sandy Islet, Scott Reef and those from other rookeries in Australia Table B 6: Hatchling straight carapace length (cm) from Scott Reef green turtle hatchlings and green turtle hatchlings from other populations in northern Australia Table B 7: Hatchling mass (g) of green turtle hatchlings from Sandy Islet and those reported from other green turtle populations in northern Australia Table B 8: Scale counts of green turtle hatchlings from Sandy Islet and those from hatchlings from other northern Australian rookeries. The post marginal counts for Bramble Cay hatchlings are included in the count of the marginal scales iii

8 List of Figures Figure 1-1: Location of Sandy Islet within Scott Reef, off the Kimberly Coast, north-western Western Australia... 2 Figure 1-2: Generalised life cycle of marine turtles (after Miller 1997, in Lutz and Musick 1997)... 4 Figure 2-1: Marked green turtles on shore (A) and the internesting area (B)... 9 Figure 3-1: Sand temperatures at turtle nest depth (50 cm) at Sandy Islet, Scott Reef between 28 August 2008 and 25 January The black line indicates is the theoretical pivotal temperature for sex determination of 29.5 C for green turtles Figure 3-2: Changes in the area of Sandy Islet used by nesting marine turtles in February, September and November Figure 3-3: Changes in the areas of Sandy Islet used by nesting marine turtles in March 2007, August and December Figure 3-4: Marine turtle nesting activity on Sandy Islet, Scott Reef, in December Figure 3-5: Marine turtle nesting activity on Sandy Islet, Scott Reef, in January Figure 3-6: Internesting habitats on Sandy Islet Reef, Scott Reef iv

9 1 Introduction 1.1 Project Description Woodside, on behalf of the Browse Join Venture 1, is proposing to develop the Torosa, Brecknock and Calliance gas fields located approximately 290 km off the Kimberley coast in the Browse Basin. The proposed Browse Upstream Development includes a gas collection system and the necessary infrastructure to enable the production and transport of natural gas and associated products from the three separate gas fields. While Brecknock and Calliance are located in deep water (400 m to 700 m), the Torosa field lies in part within the vicinity of Scott Reef (Woodside 2008). 1.2 Location Scott Reef consists of two emergent shelf atolls, called North Reef and South Reef which occur on the edge of Australia s North West Shelf (Figure 1-1). Water depths within the lagoons of North Reef and South Reef vary from 1 m to 70 m. Within the South Reef lagoon there is a mobile sand cay, Sandy Islet. The northern and southern sand spits of Sandy Islet shift several metres over the reef flat with prevailing weather conditions. However, it is approximately 700 m long in a northsouth direction and approximately 60 m at its widest part (with approximately 400 m remaining above water at extreme high tide). 1 The participants in the Browse LNG Development are: BP Developments Australia Pty Ltd., BHP Billiton (North West Shelf) Pty Ltd., Chevron Australia Pty Ltd., Shell Development Australia Pty Ltd. and Woodside Energy Ltd 1

10 Figure 1-1: Location of Sandy Islet within Scott Reef, off the Kimberly Coast, northwestern Western Australia 2

11 1.3 Conservation Value Scott Reef is listed on the Register of the National Estate, with its heritage value detailed as providing breeding habitat for the nationally vulnerable green turtle (Chelonia mydas)'. Additionally, Sandy Islet, East Hook Island (off the eastern edge of South Reef) and the intertidal reef flat of South Reef are recognised as an area of reserved land (formally known as a C class nature reserve) vested in the Conservation Commission of Western Australia. The reserved land is designated for the purpose of conservation of flora and fauna. Four species of marine turtles, listed as Threatened under the Commonwealth Environment Protection and Biodiversity Conservation Act 1999 (EPBC Act), have been identified as potentially occurring within the Scott Reef area (DEWHA 2008). However, only green turtles and hawksbill turtles (Eretmochelys imbricata) have been observed breeding and nesting on Sandy Islet (Guinea 2006a). Under the EPBC Act green and hawksbill turtles are listed as migratory, marine species and are considered Vulnerable. Under the Western Australian Wildlife Conservation (Specially Protected Fauna) Notice 2008, green and hawksbill turtles are considered Rare or Likely to Become Extinct. 1.4 Previous Studies Infrequent marine turtle surveys of limited duration have been conducted at Scott Reef due, in part, to its remote location and logistical constraints (Guinea 1995; Pendoley 2005; Guinea 2006a; Guinea 2006b; Guinea 2006c; Guinea 2007). In November 1991, an estimated 40 to 50 green turtles were ashore each evening on Sandy Islet during a two night visit (WAMTP unpublished data; R. I. T. Prince pers. comm.). A one-day visit on 18 September 1994 reported seven green turtle tracks from the previous evening and 28 tracks from the last set of spring tides (Guinea 1995). These earlier studies, although opportunistic and limited in duration, indicated that green turtle nesting on Sandy Islet occurred in spring. Three dedicated trips to Sandy Islet to attach satellite transmitters on nesting green turtles took place in October 2002, September 2003 and January 2004 (Pendoley 2005). For these trips the maximum numbers of nesting green turtles per night were two, four and six respectively (Pendoley pers. comm. 2009). The low numbers of nesting turtles seen during these trips and evidence of Indonesian harvest of turtles were assumed to be linked and raised the likely prospect of local extinction of this population (Pendoley 2005). The lack of systematic surveys of the nesting green turtle population at Scott Reef resulted in the assumption that few turtles nested within the area. This information gap on nesting seasonality, population dynamics and genetics of marine turtles of the continental shelf edge islands of the Indian Ocean (Dethmers et al. 2006) has been addressed in part by more recent surveys. Three surveys were conducted in 2006 (Guinea 2006a, 2006b, 2006c) to increase our knowledge regarding the nesting population size and seasonality of green turtles at Scott Reef. Observations 3

12 were also made with respect to the species diversity and abundance of sea snakes during these surveys undertaken during February, September and November An opportunistic survey of Sandy Islet in March 2007 reported decreased nesting turtles and evidence of green turtle nesting during previous months, thereby supporting the November 2006 observations of increased nesting activity during the summer months, as was observed at Ashmore Reef 240 km to the north (Guinea 2007). 1.5 Species Background Marine turtle species have complex life cycles that may be generalised for the green turtle (Figure 1-2). The life cycle depicted is refined from earlier models and presents the current understanding of marine turtle breeding biology (Limpus 1979, Limpus et al. 1984, Carr 1987, Hirth 1993, Ackerman 1997, Limpus 2007). Figure 1-2: Generalised life cycle of marine turtles (after Miller 1997, in Lutz and Musick 1997) Marine turtles live most of their lives in areas that are often a great distance from their natal beach. Adult and immature turtles forage in the same habitats and have similar diets. The abundance and quality of the available food affects the success of breeding females. Females with access to food of suitable abundance and quality, store sufficient energy reserves to swim from the feeding areas to the vicinity of the nesting beaches where they mate with adult male turtles. This migration is timed to coincide with the breeding season. 4

13 During the receptive phase the females mate with several males, storing the sperm in the deepest regions of their oviducts. This mixing and storage of sperm promotes multiple paternities among the hatchlings of a clutch. As the females become unreceptive, some males leave the mating area to return to their foraging areas whilst others remain. For several weeks after mating the females rest in areas close to the nesting beach. They usually stop feeding, preferring to spend their time almost motionless except for the regular movements to the surface to breathe. When the females are ready to lay a clutch of eggs they come ashore at the nesting beach. Most green turtle nesting occurs at night. Reef flats often block access to the beach resulting in females using the high tide to reach the shore. Within a nesting season a female green turtle will lay 3 to 10 clutches at about fortnightly intervals. Most green turtles lay 5 or 6 clutches, with first time nesters (neophytes) laying fewer clutches than the seasoned nesters. The number of eggs laid in a clutch varies with the species. Green turtles lay medium-sized clutches of about 100 eggs (Miller, 1997). Most female green turtles usually return to the same beach to nest each fortnight (strong philopatry) but others may nest on beaches within 100 km of their early site. This strong philopatry helps keep the integrity of the nesting population between and within nesting seasons. Between nesting events the female occupies a shallow water habitat within some tens of kilometres of the nesting beach. In this internesting habitat the female adopts the pre-nesting, non-feeding and resting behaviours. After laying the last clutch of eggs the females return to their foraging areas from where they swam some months earlier (Limpus 2007). Eggs are laid at a depth of about 50 cm and take seven to twelve weeks to hatch. The warmth of the sand heated by the sun, with additional heat produced by metabolic activity within the clutches, provides a stable thermal environment throughout incubation. The sex of the hatchlings is determined by environmental factors of which temperature is the most easily measured and recorded. Warm temperatures, usually above 29.5 C, in the middle third of the incubation period produce more females. Cooler temperatures occurring at that time produce more males. Incubation progresses at temperatures above 24 C. Temperatures above 34 C are lethal to the embryos (Ackerman 1997; Miller 1997; Limpus 2007). Hatchlings take several days of group effort to dig from the bottom of the nest to the sand surface. They usually emerge en mass at night and move towards the lowest light horizon. On entering the sea they swim quickly, and helped by currents, reach deeper water beyond the shore. Hatchlings use the energy reserves in their remaining yolk to swim to open water. There they live in the surface waters for some years, possibly a decade or more, feeding on floating algae, soft-bodied animals and small crabs. At 40 cm in carapace size, juvenile green turtles move to reef habitats where they feed on benthic algae and seagrass. They may occupy several different reef habitats over the next few decades before they mature and begin breeding. The age to maturity in wild populations of green turtles is approximately 30 years. Populations that are genetically isolated from each other are referred to 5

14 as Management Units (MU). Members of different management units mix on foraging areas, but there is little interbreeding with members of neighbouring green turtle management units. 1.6 Scope of Works Woodside engaged Sinclair Knight Merz (SKM) with Charles Darwin University (CDU), to undertake turtle surveys at Sandy Islet, Scott Reef. The objective of this study The Long Term Monitoring of Marine Turtles Sandy Islet Scott Reef was to collect baseline data on turtle nesting activity and describe species richness and abundance at Scott Reef. The information is to provide a baseline against which to assess the potential impact that the proposed Browse LNG Development may have on marine turtles. It will also assist the development of any future environmental management plans that may be needed. The Long Term Monitoring of Marine Turtles at Scott Reef involved three trips to Scott Reef in 2008 and one in The scope of the survey trips were to: determine the population abundance through monitoring and recording the number of marine turtles nesting on Sandy Islet; provide long-term individual turtle identification using numbered titanium tags; record the length and width of the green turtle carapaces for morphometric analyses; collect biopsy samples from all nesting marine turtles and dead hatchlings for DNA analysis; evaluate the hatching success and the emergence success of clutches; undertake opportunistic random surveys to identify internesting habitat, and map the extent of potential nesting areas on Sandy Islet. 6

15 2 Methods The methods used in the 2008 and 2009 surveys were consistent with those used previously (Guinea 2006a, 2006b, 2006c). However, not all methods were used during each survey (for example, no hatching success data was collected in September 2008). Details of the following methods are provided in subsequent subsections: survey dates (Section 2.1); population and abundance estimates (Section 2.2); population morphometrics (Section 2.3); skin biopsy collection (Section 2.4); clutch size and nesting success (Section 2.5); and habitat mapping (Section 2.6). All field data was entered into an Excel spreadsheet of nightly summaries, tag registers and register of DNA samples (Appendix A). 2.1 Survey Dates The survey dates coincided with periods of full and new moon when spring tides reached their maximal ranges (Table 2-1). The number of nights spent ashore was determined by concurrent projects operating from the same vessel. The time spent ashore each night was determined by the tidal heights offering sufficient water depth to enable a medical evacuation should it be required. Table 2-1: Timing and duration of the seven survey trips to Scott Reef between 2006 and Date Moon Phase Maximum Tide Height (m) Number of nights Time Ashore (hr) Feb-06 New Sep-06 Third Quarter Nov-06 First Quarter Aug-08 New Sep- 3 0ct 08 New Dec-08 Full Jan 1 Feb 2009 New Total

16 2.2 Population and Abundance Estimates Nesting Marine Turtle Survey Nesting marine turtles were identified to species by the size and shape of the track left in the sand. The front flippers of the green turtle extend beyond the tracks of the hind flippers and plastron, leaving a conspicuous chevron pattern along the outside of the track. The hind flippers move simultaneously leaving a symmetrical track (Limpus 2007). The relatively shorter front flippers of the hawksbill turtle barely protrude beyond the plastron and the hind flippers. The hind flippers produce an asymmetrical track caused by the alternate gait of this species (Limpus 2004). Sand pushed backwards by the front and hind flippers indicates the direction of movement Beach Patrol Sandy Islet beach was patrolled on foot with ambient moonlight and starlight sufficient to gain a clear view of the beach. When ambient light was insufficient for monitoring, a red light emitting diode (LED) headlamp (Energizer HDL33A) was used to illuminate the path ahead of the survey team over the broken terrain of the island. The light use was limited to the smallest possible area so as to not influence the behaviour of nesting turtles. Turtle tracks from the previous 24 hours and older were identified and a foot dragged across the tack to avoid duplication in the dataset. This crossing of old tracks on the beach aided identifying new turtles coming onshore during the survey. When the field team departed the island before nesting had finished and when new turtles were likely to come ashore, a reconnaissance of nesting activity and success took place on either the following day or early the next evening. The down tracks crossing the beach were followed back to the last diggings made by the turtle and assessed as to whether the nest had been successful. The categories of nesting activity of the turtle were: having laid, having not laid and unknown Marine Turtle Sightings When a marine turtle was encountered, its activity was assessed by listening to the flipper strokes and observing its behaviour. If a turtle was heading towards the water, its escape was blocked by a team member with a white light and the turtle measured and tagged on the beach, otherwise turtles were left undisturbed to nest Tagging and Marking Protocols Turtles were tagged with a single, individually numbered, 4.1 gram titanium tag (WA series) attached through the axial scale (closest to the body) of the left front flipper. Aerosol paint (Signet Spot Marking Yellow) held close to the carapace, was used to produce a large number (approximately 30 cm in height and 5 cm in thickness). A large letter on the left side indicated the day of the survey and a number on the right indicated the order in which the turtle was encountered. 8

17 In 2009, the marking procedures were altered to using a fast drying yellow acrylic aerosol (Dy- Mark Line Marking Paint) for the letters and numbers painted on both sides of the carapace. Those turtles that nested successfully were marked on the parietal scale with a coloured acrylic aerosol (Dy-Mark Spray and Mark Marking Out Paint). A different colour was used for each day of the survey. The colours and the order of their use were: Fluorescent Pink 25 January Green 26 January White 27 January Fluorescent Orange 28 January Fluorescent Yellow 29 January Orange 30 January Yellow 31 January This colour combination and symmetrical numbering enabled the identification of green turtles while ashore as either having been marked that night, on a previous night or not during the survey. Marking the parietal scale enabled successful nesters to be recognised at possible internesting habitats away from the island (Figure 2-1). A B Figure 2-1: Marked green turtles on shore (A) and the internesting area (B) Estimates of Population Size The two methods used to estimate the abundance of the nesting turtles present at Sandy Islet in nesting season included: a count of flipper tags and an estimate from the marking and recapture of individuals, the Petersen estimate Flipper Tags Each nesting turtle received a titanium flipper tag. The numbers of tags applied each trip and their cumulative number for the summer nesting period provides the minimum number of turtles present for the season. 9

18 Mark and Recapture A Petersen estimate of the number of marine turtles in the vicinity of Scott Reef was gained by marking a sample of animals with paint while they were ashore combined with later identification of marked animals at sea. The Petersen estimate (N) of the population with Bailey s correction for improving the accuracy of the estimate of an incomplete census, was calculated using the number of sprayed individuals marked prior to the day of observation as the marked population (M), from the sample of sprayed individuals (m) seen amongst adult short tailed individuals (n) observed during surveys of potential internesting habitats. The estimate is calculated as: [ ] N = M ( n + 1) /( m + 1) SE = M ² ( n + 1)( n m) ( m + 1 ) ² ( m + 2) (Caughley 1977, Gerrodette et al. 1999, Limpus et al. 2003). The underlying assumptions of this method are: marked animals are uniformly distributed amongst the unmarked animals at Scott Reef; no marks are lost between marking and the time of recapture; marks are visible on a turtle in the water; marked and unmarked animals are equally visible; all adult turtles are part of the breeding stock and resident turtles are not included in the survey; marked and unmarked animals move out of the area at the same rates; and there is no difference in mortality and behaviour of marked animals (Caughley 1977; Gerrodette et al. 1999; Limpus et al. 2003). The assumptions outlined above demonstrate the limitations of this survey and statistical analysis method. While the method may be appropriate for preliminary surveys it is not an acceptable method for ongoing monitoring programmes as it lacks the power to detect trends and cannot show stock status. 2.3 Morphometrics of Green Turtles 1/ Nesting Females Carapace dimensions and any individual distinguishing marks or injuries were recorded for each adult turtle. Curved carapace lengths (CCL) and curved carapace widths (CCW) were measured with a flexible fibreglass tape (± 0.05 cm). The CCL was measured from the junction of the skin and the nuchal (most anterior carapace) scale to the notch between the post vertebral scutes at the posterior of the carapace (Limpus et al. 2001; Limpus et al. 2003; Guinea et al. 2005). The CCW was measured across the carapace at its widest point with the same tape. These procedures comply with standard measurement protocols (Limpus 2007). 10

19 2.3.2 Hatchlings A sample of ten hatchlings were collected from either the surface as hatchlings emerged or from a group of hatchlings positioned just below the sand surface ready to emerge. The sample of hatchlings from each nest was collected and held in clean cotton bags until processing which was within three hours of emergence. All hatchlings were released from the beach from which they were collected before daylight. Each hatchling was measured, weighed and scutes counted. Straight carapace length (SCL) was measured from the anterior edge of the nuchal to the notch between the post vertebral scales with digital callipers (± 0.01 cm). Straight carapace width (SCW) was measured across the carapace at its widest part to the outer edge of the marginal scales with the same digital callipers (± 0.01 cm). The mass was measured (± 0.5 g) by attaching the post vertebral scutes to a 100 g spring balance (Pesola). Scale counts included the following scutes: nuchal, vertebral, post vertebral, costal, marginal, infra marginal, prefrontal, post ocular, post parietal. Paired scales were recorded as the numbers occurring on the respective side such as Left/Right (Brongersma 1968, Limpus et al. 2001, Limpus et al. 2003). 2.4 Skin Biopsy Collection Collection of a skin biopsy involved removing a small piece of skin with an open blade. Skin biopsies (3 mm x 5 mm) from the trailing edge of the left hind flipper were taken using a sterile scalpel and forceps and stored in individually-numbered vials containing 20% dimethyl sulfoxide (DMSO) and sodium chloride (NaCl). The trailing edge of the left hind flipper was chosen as the biopsy location because this particular region incurs damage from mating and nest digging, is relatively insensitive, and does not bleed easily. To stop the turtle struggling and injuring either itself or the researcher, a member of the team applied finger pressured massage to the turtle s neck, thereby stimulating the vagus nerve and placing the turtle in a relaxed state (tonic immobility). The tonic immobilisation enabled the skin biopsy, tagging, measuring and spray marking to be completed within a minute and with minimal stress to the animal. A single tissue sample from each hatched nest was collected for DNA analysis. A piece of flipper from either a dead embryo or dead hatchling was removed and preserved in an individuallynumbered vial containing 20% DSMO and NaCl and stored with the samples from adult female turtles. 2.5 Clutch Size and Nesting Success Marine turtle nest data was collected to determine nesting success at Sandy Islet, details are provided in the following subsections. 11

20 2.5.1 Environmental Data Two temperature recording data loggers (Hobo Pro water temperature) were positioned together at 50 cm depth on Sandy Islet at one location on 28 August 2008 (14º S, 121º E). One of these loggers was retrieved and downloaded in December 2008 and reset in the same position for retrieval in January Instantaneous temperatures of nest were recorded with an infrared thermometer (Pro Exotics Temp gun PE-2) positioned within 2 cm of the surface to be measured. The average of three temperatures was recorded Hatched Nest Identification Nests ready to hatch or having hatched were identified from either a circular depression in the sand, observation of hatchlings emerging or hatchling tracks emanating from a central locality. Pulses of hatchlings on the beach were used as an indicator of the number of nests hatched per night Nest Excavation The nest was excavated using either bare hands or a stainless steel bowl, 20 cm in diameter with a smooth rounded rim to prevent damaging any hatchlings remaining in the nest. The soft dry coralline sand at Sandy Islet meant that completed nest excavation was typically 2 m wide at the top to a depth of approximately 1 m and tapering to only a few centimetres in width at the base. As a result any excavation required considerable effort. A successful excavation resulted in the discovery of the empty eggshells in the bottom of the egg chamber Examination of clutch The contents of the nest were recorded using the hatched nest datasheet. The dataset included numbers (#) of: shells (S), live hatchlings in nest (L), dead hatchlings in nest (D), undeveloped eggs (UD), unhatched eggs (UH), unhatched eggs with full-term embryos (UHT), emerged hatchlings (E), depredated eggs (P), and yolkless eggs (Y) after Miller s (1999) categorisation of nest contents. Nest depth to the top of the egg shells and sand temperature in the base of the nest were recorded when the nest held its shape during excavation. Hatching and emergence success were calculated for each nest. Equations used to determine clutch size, hatching and emergence successes (Miller 1999) were as follows: 1 Clutch Size = #shells+ #UD + #UH + #UHT + #P 2 Hatching Success (%) = 3 Emergence Success (%) = #shells X 100 #shells + # UD + # UH + # UHT + # P #shells - (# L + # D) X 100 #shells + # UD + # UH + # UHT + # P 12

21 Live hatchlings in the nest that were fit and active were released onto the beach during darkness to replicate natural emergence. Hatchlings in poor condition due to dehydration or deformation were transferred aboard the research vessel and given a day to recuperate and rehydrate. They were released the following night into deeper water beyond the reef edge. 2.6 Marine Turtle Habitats The outline of the area used by nesting turtles on Sandy Islet was recorded using a global positioning satellite (GPS) receiver (Garmin 72). This area along with sites of turtle nests, including both identified hatched nests and recently laid nests, were recorded and stored in a Geographic Information System (GIS). The marine habitat used by female marine turtles during their internesting period was investigated using two methods. In December 2008 the ship s tenders with at least three observers using a bathyscope and viewing tubes, motored at 2 km/hr along the reef crest recording the positions of marine turtle sightings with a GPS receiver. All marine turtles encountered were identified to species and stage of maturity. Short-tail turtles of adult size included immature males and adult females. Subadult and juvenile green turtles were identified by their small size and characteristic sunray pattern on the scutes on the carapace. Hawksbills were identified by their carapace shape, head shape and presence of fouling organisms (sponges and corals). Tender surveys of possible internesting habitats occurred during the afternoon of 12 and 13 December On 13 December the survey focused on the grid covering almost 4 square kilometres (km 2 ) between S and S and E and E that was indicated previously as a prominent internesting area (Pendoley 2005). Surveys of the possible internesting areas using the RV Kimberley Quest took place daily from 14 to 16 December 2008 and 26 to 31 January Observers with binoculars standing on the deck 4 m above the waterline reported the numbers and letters on the carapace and coloured marks on the heads of turtles visible along the reef edge. 13

22 3 Results 3.1 Population Abundance and Estimates This study has demonstrated that two species of marine turtle nest on Sandy Islet, Scott Reef. Nesting by the hawksbill turtle (Eretmochelys imbricata) was not reported before the February 2006 survey. One hawksbill turtle (WA53447/WA53464) was initially tagged on Sandy Islet on 18 February 2006 and was then recaptured on the island on 25 January 2009 after an interval of almost three years. Given that only one Hawksbill turtle was observed to nest on Sandy Islet, no further comment on its abundance can be made. The most common species of marine turtle nesting on Sandy Islet was the green turtle (Chelonia mydas) which is the focus of the remainder of this report Green Turtle Nesting Activity The results from the seven surveys conducted during 2006, 2008 and 2009 indicate that the summer months from late November to February are the preferred breeding season for green turtles (Table 3-1). Although nesting may occur throughout the year the winter months support a decreased amount of nesting activity on Sandy Islet. Of the 435 turtles tagged since February 2006, 106 (24%) were tagged in the four winter/spring surveys. The results from the 2005/2006, 2006/2007 and 2008/2009 nesting seasons highlight the natural interannual variability in green turtle nesting numbers. During the 2005/2006 nesting season the mean number of tracks/night was 16.2 during peak nesting periods which was noticeably fewer than the 71.9 tracks/night observed during peak nesting periods in 2008/2009. The ratio of the number of tracks to the number of nests was often several times greater than one regardless of the season. Hatched nests were more numerous in the summer months although some hatching occurred throughout the year (Table 3-1). 14

23 Table 3-1: Breeding seasonality of green turtles on Sandy Islet, Scott Reef, as demonstrated by the mean (± standard deviation, sd) number of nests and tracks per night during the survey periods that spanned most seasons. Season Month Year #Green Turtles tagged Mean # of tracks /night ± sd Mean # of nests/ night ± sd # hatched Nests Tracks: Nest Ratio Summer Feb ± ± Dec ± ± Jan ± ± Winter Aug ± ± Spring Sep Nov ± ± Oct ± ± Inter-Season and Within Season Recaptures The average number of nesting events within a season for green turtles on Sandy Islet and the remigration interval remains unknown. Recaptures within a season came from those that were unsuccessful in nesting on a previous night, or those that had nested earlier in the season and had returned to nest again. Previously unsuccessful nesting turtles that had been tagged were identified by the paint on their carapaces and were left undisturbed to complete their nesting. Within season recaptures were recorded during both September 2008 and January 2009 surveys. In the September 2008 survey only one within season recapture was recorded 33 days after it was initially tagged in August Twenty green turtles that were initially tagged in the December 2008 survey were recaptured during the January 2009 survey. The times between recaptures varied from 41 to 52 days (mean ± standard deviation = 46.2 ± 3.2 days; n = 20). This average period is approximately equal to three spring tidal cycles. Inter-season recaptures were from 3 green turtles that were initially tagged in the 1991/1992 season (WAMTP unpublished data, R. I. T. Prince, pers comm.) and then recaptured in the 2006 and 2008 surveys. WA11238 was recaptured on 18 February 2006 and WA15009 was recaptured on 23 November 2006 both after an interval of fourteen years. WA15027 was recaptured on 15 December 2008 after an interval of 16 years. WA11238 was again recaptured on 26 January 2009 while nesting on Sandy Islet after an interval of almost three years (1073 days) Estimates of Population Size Between 27 August 2008 and 31 January 2009, a total of 314 female green turtles were tagged while ashore at Sandy Islet. This represents the minimum number of nesting green turtles in the 2008/2009 summer nesting season. 15

24 Petersen estimates, with Bailey s correction and standard errors, of the number of female green turtles available (the difference between the cumulative arrivals and departures of nesting green turtles) in the nesting population at Sandy Islet were calculated for three days in December 2008 and five days in January 2009 (Table 3-2). Estimates of abundance varied from 389 to 1476 (mean ± standard deviation = 779 ± 383; n = 8) with individual standard errors exceeding 50% of the Petersen estimate in some cases. Consequently, these estimates must be considered indicative at best, especially given the assumptions associated with the Petersen estimates. Table 3-2: Petersen estimates with Bailey s correction and standard error (SE) of estimates for eight recapture dates at Sandy Islet, Scott Reef. Date # Marked # Recapture Sample Size Petersen Estimate SE 14/12/ /12/ /12/ /01/ /01/ /01/ /01/ /01/ Morphometrics During the seven trips to Scott Reef 460 individual female green turtles were encountered. Curved carapace lengths were attainable for 447 of these. The remainder had damage to the carapace that prevented their carapace lengths being analysed. Similarly, 458 females were encountered with undamaged carapaces that yielded reliable curved carapace widths (Table 3-3). The curved carapace lengths of these nesting females varied from 82.8 cm to cm (mean ± standard deviation = ± 5.26 cm; n = 447). Curved carapace widths varied from 73.0 cm to cm (mean ± standard deviation = cm ± 5.54; n = 458). 16

25 Table 3-3: Sizes of nesting female green turtles at Sandy Islet, Scott Reef, in the 2005/2006, 2006/2007 and 2008/2009 breeding seasons and for all seasons combined. Date Average s. d. n Minimum Maximum Curved Carapace Length (cm) Feb Sep Nov Aug Sep Dec Jan Combined Curved Carapace Width (cm) Feb Sep Nov Aug Sep Dec Jan Combined Samples of 10 hatchlings from 13 hatched nests were measured (n = 130), weighed and had their scalation recorded. Straight carapace lengths (SCL) varied from 4.29 to 5.24 cm (mean ± standard deviation = 4.74 ± 1.4 cm; n = 130). Straight carapace widths (SCW) varied from 3.28 to 4.44 cm (mean ± standard deviation = 3.7 ± 1.30 cm; n = 130). Hatchling mass varied from 20.0 to 29.5 g (mean ± standard deviation = ± 1.97 g; n = 130). The modal scute pattern for the Sandy Islet green hatchlings was typical of the species: 1 nuchal. 5 vertebrals, 1/1 post vertebrals, 4/4 costals, 11/11 marginals, 4/4 inframarginals, 4/4 postoculars, 2 prefrontals, 2 post parietals (Bustard 1972; Pritchard 1979; Hirth 1993; Limpus et al. 2003). 3.3 Biopsy Analysis A total of 150 biopsy samples were collected from nesting females and dead hatchlings for DNA analysis and identification of genetically differentiated populations. Populations that are genetically isolated from each other are referred to as Management Units (MU). A preliminary shipment has 17

26 been analysed (refer to FitzSimmons and Jensen 2008 for the results) and the remainder have been stored and await permits to forward them to Dr. Nancy FitzSimmons at the University of Canberra for analysis and reporting. 3.4 Clutch Size, Hatching Success and Emergence Success Hatched nests were only identified during the December 2008 and January 2009 surveys. Nest depths were recorded for all excavated nests, the top of the nest (where the eggs started) ranged between 53 to 80 cm (mean ± standard deviation = 64.0 ± cm) while the bottom of the nest ranged from 75 to 77 cm (mean ± standard deviation = 76.0 ± 1.41 cm). In the December 2008 survey only four of the 10 hatched nests were successfully located due to either the soft sand or the hatchling tracks being obliterated by either ghost crabs, or nesting turtles. The contents of the 4 nests were examined to ascertain the clutch sizes and the average hatching and emergence success. Clutch sizes varied from 50 to 111 eggs (mean ± standard deviation = 83 ± 25 eggs). Hatching success for these 4 nests varied from 65.8% to 89.9% (mean ± s standard deviation = 80.2 ± 10.9%). Emergence success for these nests varied from 45.1% to 84.9% (mean ± standard deviation = 70.3 ± 17.4%). In January 2009 a total of 10 hatched nests were excavated to determine the clutch sizes and the average hatching and emergence success at this time of the year. Clutch sizes varied from 66 to 132 eggs (mean ± standard deviation = 86.3 ± 21.5 eggs). Hatching success for these nests varied from 77.1% to 95.4% (mean ± standard deviation = 88.5 ± 5.9%). Emergence success for these nests varied from 54.1% to 95.4% (mean ± standard deviation = 80.9 ± 12.1%) Environmental Temperatures: Sand temperatures at a depth of 50 cm from 28 August 2008 to 25 January 2009 increased at a steady rate as summer months approached (Figure 3-1). The sand temperature passed the presumed pivotal temperature (29.5 C) for sex determination of hatchlings in late November Sand temperatures fluctuated above and below the pivotal temperature as summer progressed. Sand temperatures within the nests of variable depth were around 27.5 C in December 2008 and 29.3 C in January 2009, while egg temperatures at the time of laying varied from 30.0 C to 31.5 C. Water temperatures were consistently 30 C over the reef flat. 18

27 Sand Temperature at 50 cm Temperature (C) Aug-08 9-Sep Sep Oct-08 8-Nov Nov Dec-08 7-Jan Jan Feb-09 Date Figure 3-1: Sand temperatures at turtle nest depth (50 cm) at Sandy Islet, Scott Reef between 28 August 2008 and 25 January The black line indicates is the theoretical pivotal temperature for sex determination of 29.5 C for green turtles Marine Turtle Habitats Nesting Habitats The area of Sandy Islet used by nesting turtles was recorded for each trip. The highest part of the sand cay that accommodated consistent nesting was approximately 60 m wide by 400 m in length in 2006 (Figure 3-2) and 2008 (Figure 3-3). The southern and northern sand spits were dynamic and varied in length and width between surveys. Turtle nests and attempted turtle nests extended along the entire length of Sandy Islet and the sand spits. The only part of the island that was not freshly disturbed by nesting turtles in the 2008 surveys was the small stretch on the northern beach that was behind an expanse of beach rock (Figure 3-4). In January 2009 this area behind the beach rock was utilised by nesting turtles (Figure 3-5). The paths taken by the turtles moving to and from the nest localities extended across the sand cay leaving very little of the island untouched by nesting turtles. The evening high tide on 27 January 2009 washed over the southern sand spit immersing all existing turtle nests in salt water. 19

28 Figure 3-2: Changes in the area of Sandy Islet used by nesting marine turtles in February, September and November

29 Figure 3-3: Changes in the areas of Sandy Islet used by nesting marine turtles in March 2007, August and December

30 Figure 3-4: Marine turtle nesting activity on Sandy Islet, Scott Reef, in December Indicated is the region of the sand cay used by nesting green turtles (shaded), the locations of nests for each day of the survey (coloured circles) 22

31 Figure 3-5: Marine turtle nesting activity on Sandy Islet, Scott Reef, in January Indicated is the region of the sand cay used by nesting green turtles (shaded), the locations of nests for each day of the survey (coloured circles) and the locations of hatched nests (open circles) 23

32 Internesting Habitat The December 2008 surveys of the reef crest and adjacent waters, surrounding Sandy Islet, for female green turtles that had been marked with paint while nesting on Sandy Islet revealed 22 marked individuals over a five day period (Table 3-4). Table 3-4: Numbers of marked female green turtles and unmarked green turtles along the reef crest and adjacent waters at Sandy Islet. Green Turtle Date Habitat Marked Unmarked Adult Male Juvenile Female Adult Short-tail 12/12/08 Reef Crest /12/08 Sand Patch /12/08 Sand Patch /12/08 Sand Patch /12/08 Sand Patch /1/09 Western to southern Reef 27/1/09 Western and southern Reef 28/12/09 Western and southern Reef 30/1/09 Sand Patch Southern Reef 31/1/09 Sandy Islet Reef perimeter Surveys in January 2009 covered the entire perimeter of the reef crest. Three sand areas contained numbers of short tailed adult green turtles. The sand area to the south of Sandy Islet Reef ( A in Figure 3-6) was not only the largest in area, but contained the most short tailed adult green turtles. On 30 January 2009, 85 adult short tailed green turtles were recorded in 20 minutes of observation in the sand patch (Area A) at the southern end of Sandy Islet Reef (Figure 3-6). On 31 January 2009, Area A contained 75 short tailed adult green turtles, Area B contained 10 adult green turtles, and Area C contained 14 adult green turtles. Poor visibility and squalls were responsible for the low scores of adult short tailed green turtles on 28 January

33 Figure 3-6: Internesting habitats on Sandy Islet Reef, Scott Reef. Indicated are; the survey path of the tenders conducting the bathyscope survey on 12 and 13 December 2008 (red line), the area of more intense internesting localities (Pendoley 2005) (black square), and the major marine turtle aggregation area (shaded region) and the survey of the reef crest January 2009 (red dashed line) 25

34 4 Discussion 4.1 Population Abundance A count of flipper tags applied since August 2008 revealed a minimum of 314 female green turtles nested on Sandy Islet during the 2008/2009 summer breeding season. The abundance estimates derived from mark and recapture procedures were based on a small sample size (the mean from the eight estimates), varied greatly and had widely fluctuating standard errors. Consequently, even with Bailey s correction for incomplete census, the estimate of 779 individuals with a standard error of 383 should be used with caution and may not be a true measure the nesting population. Furthermore, none of the assumptions outlined by Caughley (1977) and Limpus et al. (2003) have been validated which limits the use of these estimates. For example, equal catchability between marked and unmarked female green turtles has not been demonstrated. Caughley (1977) reminds us that in natural populations, equal catchability is more the exception than the rule. By way of comparison, few estimates exist of the size of the breeding population at Sandy Islet. As part of the information database of marine turtle populations in the South East Asian region, the Indian Ocean South East Asia Sea Turtle Memorandum (IOSEA) places the breeding population at between 1000 and 5000 individuals annually ( For genetic analysis the estimated total number of breeding females has been placed at 1500 individuals (Dethmers et al. 2005). Both of these estimates are about the same as the higher values produced by the Petersen estimates and should be considered as coincidences pending further research. To obtain an accurate estimate of the nesting population, Limpus et al. (2003) recommend a whole-of-season tagging census of the rookery be undertaken as was conducted on Bramble Cay (Limpus et al. 2001). As this is logistically and financially unfeasible for Scott Reef, current literature (Limpus et al. 2001) recommends that a census completed over a two week period at the peak of nesting will provide an estimate that is approximately 80 percent of the number of green turtles that are available at that time (Limpus et al. 2003). The estimate of the nesting population is further complicated by the observation from the Bramble Cay study (Limpus et al. 2001) that at no time in the breeding season is the entire nesting population of green turtles present for counting (Limpus et al. 2003). Even if an accurate estimate of nesting population can be determined, the stability of these green turtle populations cannot be assessed by measuring the size of the annual nesting population over a few years or even over a decade (Limpus et al. 2003). Any change less than a catastrophic decline in numbers of nesting turtles may not be detectable for some decades (Limpus et al. 2003). However, given these limitations, the Petersen estimates produced by painting nesting turtles and conducting surveys of the internesting habitats is a cost effective means of estimating the preliminary annual population if further research addresses the validity of the inherent assumptions and the retention times for the paints are 26