Appendix F26. Guinea Long Term Monitoring of the Marine Turtles of Scott Reef: February 2010 field survey report

Appendix F26 Guinea 2010 Long Term Monitoring of the Marine Turtles of Scott Reef: February 2010 field survey report Browse FLNG Development Draft Environmental Impact Statement EPBC 2013/7079 November 2014

Long Term Monitoring of the Marine Turtles of Scott Reef February 2010 Field Survey Report Michael L. Guinea Charles Darwin University Darwin 0909 Northern Territory September 2010 1

2

3

Executive Summary Woodside Energy Ltd (Woodside) is proposing to develop the Torosa, Brecknock and Calliance gas fields located approximately 270 km off the Kimberley coast in the Browse Basin. The Torosa field lies partly under Scott Reef, which is listed on the Register of the National Estate, with its heritage value detailed as providing, among other values, "breeding habitat for the nationally vulnerable green turtle (Chelonia mydas). This document reports the findings of a field survey undertaken in January February 2010 to investigate nesting behaviour and movements of marine turtles at Scott Reef during their peak nesting period. A combination of techniques such as onshore monitoring, manta tows and satellite tracking of turtles were used. The survey extends information obtained from previous surveys of nesting marine turtles at Scott Reef in 2006, 2008 and 2009. Initial surveys in 2006 occurred opportunistically with other studies occurring at Scott Reef and therefore were limited in scope and access to Sandy Islet (Guinea 2006). Subsequent surveys in 2008/2009 were designed in order to delineate the peak nesting period and form the basis of the 2010 survey (Guinea 2009). The 2010 survey was timed to occur during the peak nesting period from late January/early February. Therefore, the most appropriate data for comparison with the 2010 survey data is the January 2009 data set as it also occurred during the defined peak nesting period. The findings of this survey are being used to support the environmental impact assessment of upstream aspects of the Browse LNG Development, which includes proposed future infrastructure over the offshore Torosa gas field near Scott Reef. Population Abundance A total of 56 green turtles were tagged at Sandy Islet, Scott Reef and 16 green turtles were tagged at Browse Island during the 2010 survey. Fewer turtles were tagged on Browse Island due to survey limitations (only one night was spent on Browse Island compared with 13 nights on Sandy Islet). There was no recapture of any turtle that had been tagged in previous years. Based on the survey findings, an indicative estimate (the Petersen estimate) of the size of the Sandy Islet breeding population in 2009/2010 nesting season was 79 ± 25 individuals, which represents approximately 10% of the 2008/2009 total breeding population. There are possible violations of the inherent assumptions of the Petersen estimates and therefore should be used as a preliminary estimate only. Possible reasons for the fluctuation in the Sandy Islet breeding population are likely to be linked to their trophic status and the fluctuations in food supply in the foraging grounds. No manta tow surveys were conducted at Browse Island therefore no population estimates could be calculated. Morphometric Analyses Nesting green turtles on Sandy Islet were on average 3 cm smaller in size than other green turtle populations in northern Australia, with curved carapace lengths ranging from 83 to 110 cm and averaging 98.5 ± 5.8 cm (n = 50). Nesting green turtles on Browse Island varied in curved carapace length from 94 to 105 cm with an average of 99.7 ± 3.5 cm (n = 15). 4

Nesting Success and Hatching and Emergence Success of Clutches Sandy Islet was patrolled for thirteen nights in Jan/Feb 2010 during which time 46 nests and 171 tracks were recorded. Rough seas prevented a lengthy survey of Browse Island. A survey of just over two hours duration during daylight hours established that 45 nests from 104 tracks had been laid on Browse Island since the last set of spring tides (approximately 7 days given the wind on the Island). On Sandy Islet, successful nests varied from 1 to 10 per night with an average of 3.8 ± 2.5 clutches laid. The density dependent mortality of nests and eggs that occurred during the 2008/2009 breeding season at Sandy Islet was not observed in the current 2010 survey, possibly due to lower numbers of nesting females observed. Nesting success rate for 2010 at Sandy Islet was lower than 2008/2009 with more than 3.4 tracks recorded for each successful nest. The reason for this result is not fully understood at this time. Clutch sizes were determined from the remains of the egg shells after the hatchlings had emerged. Those recorded during the 2010 survey at Sandy Islet ranged from 64 to 151 eggs (average of 105.6 ± 26.9 eggs; n = 8) whilst the clutch sizes of the three hatched nests examined at Browse Island ranged from 62 to 135 eggs (average clutch size of 105.7 ± 38.5 eggs; n = 3). The clutch sizes recorded are similar to those recorded from Ashmore Reef but are smaller than those from other rookeries in northern Australia. Hatching success of the emerged clutches varied from 39.6% to 100% at Sandy Islet and 77.0% to 87.1% at Browse Island. The success of hatchlings emerging from these nests ranged from 28.6% to 96.2% at Sandy Islet and 72.6% to 87.1% at Browse Island. These data are generally consistent with that obtained from other rookeries located within northern Australia and similar to hatching and emergence success recorded at Sandy Islet, Scott Reef in December 2008 and January 2009. Inter nesting Habitat Previous observations (during surveys in 2008/2009) indicated adult male and nesting female green turtles aggregated at a sand patch at the southern end of Sandy Islet Reef. This area appeared to function as an inter nesting area during the 2009 survey. Visual observations from manta tows of tagged, painted and satellite tagged nesting females in the 2010 survey supported these earlier observations that this was an inter nesting area of some significance. Twelve satellite tags attached to nesting green turtles demonstrated that the inter nesting habitat also included the reef around Sandy Islet, and the inner lagoon of the southern reef located approximately 15 km to the south of the islet. Satellite transmitters revealed re nesting intervals varied from 8 to 22 days. One satellite tagged individual nested on four occasions indicating that breeding green turtles at Sandy Islet may nest up to at least 4 times per season. Migration to Foraging Grounds By 31 March 2010, nine satellite tagged females had commenced their migration to the Australian mainland coast by travelling either north east to the Bonaparte Archipelago and then following the coast to the Northern Territory, or south to Cape Leveque and along the coast to the mouth of the 5

De Grey River in the Pilbara. The latter is significant as it is the first time post nesting migration of Scott Reef turtles to the Pilbara Coast has been recorded. Although the migration routes passed through known foraging areas, there was no aggregation of satellite location points within a defined area which would indicate the end of the migration route. Locations recorded after March 31 2010 may confirm whether any of these turtles have reached their foraging grounds. Turtles migrating from Scott Reef had average speeds of 1.9 km/h. These speeds are similar to the migration speeds of previous green turtles from Scott Reef and to the migration speeds of other green turtle populations. 6

Table of Contents 1 Introduction... 12 1.1 Background... 12 1.2 Background on Green Turtle Breeding Cycle... 13 1.3 Scope of Works... 15 2 Methods... 17 2.1 Survey Dates... 17 2.2 Tagging Programme... 18 2.3 Nest/Hatchling Programme... 21 2.4 Satellite Transmitter Programme... 21 2.5 Habitat Identification... 23 2.6 Environmental Data... 23 3 Results... 25 3.1 Tagging Programme... 25 3.2 Nest/Hatchling Programme... 35 3.3 Biopsy Analysis... 39 3.4 Marine Turtle Habitats... 39 3.5 Sky Glow... 40 3.6 Satellite Tracking... 43 4 Discussion... 53 4.1 Population Abundance... 53 4.2 Nesting Location, Nesting Success and Hatching Success... 54 4.3 Morphometrics of Green Turtles... 55 4.4 Inter nesting Habitat... 55 4.5 Migration to Foraging Grounds... 55 4.6 Sky Glow... 57 4.7 Sandy Islet and Browse Island... 57 5 Conclusion... 58 7

List of Figures Figure 1 1: Location of Scott Reef, off the Kimberly Coast, north western Western Australia. 12 Figure 1 2: Generalised life cycle of marine turtles showing the female cycle of 2 to 8 years (green), the inter nesting cycle of 8 to 16 days, the activities on Scott Reef (red) and the pelagic and neritic foraging stages (blue). 14 Figure 2 1: Location of numbered beach sectors at Sandy Islet. Each sector was approximately 100 m in length. 19 Figure 2 2: The locations of the manta tow surveys of Sandy Islet reef crest and southern reef are shown for each day. 24 Figure 3 1: Nesting area at Sandy Islet was identified by body pits on arrival (shaded area). Nests laid during the survey, 29 January to 11 February 2010, are indicated in green. 26 Figure 3 2: Numbers of turtle tracks, nests and hatched nests in each of the sectors of Sandy Islet, Scott Reef during the survey period. 27 Figure 3 3: Numbers of green turtle tracks and nests for each night of monitoring and the height of the nocturnal high tide. The full moon occurred on 30 January 2010. 27 Figure 3 4: Change in estimated nesting population of green turtles at Scott Reef in December 2008, January 2009 and February 2010. The data presented are the mean for the days sampled with standard error and the minimum and maximum estimates. 31 Figure 3 5: Relationship between curved carapace length and curved carapace width for 488 individual female green turtles at Sandy Islet, Scott Reef. 33 Figure 3 6: Comparison of the curved carapace length of green turtles nesting at Lacepede Island (n = 22), Scott Reef (n = 538) and Browse Island (n = 19). The data presented are the mean with standard deviation and the maximum and minimum lengths. 34 Figure 3 7: Comparison of the curved carapace length of green turtles nesting at Scott Reef (n = 538) with those nesting at other localities in northern Australia: Ashmore Reef (n = 102), Heron I (n = 1942), Rain I (n = 20,947), Bountiful I (n = 219) and Coral Sea islands (n = 100) (Limpus 2007). The data presented are the mean with standard deviation and the maximum and minimum lengths. 34 Figure 3 8: Summary of the contents of hatched nests from Sandy Islet, Scott Reef. 35 Figure 3 9: Summary of the contents of hatched nests from Browse Island. 36 Figure 3 10: The maximum luminance (A) and the height of the tide (B) during green turtle nesting activity with the respective numbers of tracks and successful nests (C) for each night of monitoring on Sandy Islet, Scott Reef. 42 Figure 3 11: Inter nesting movements of four green turtles tracked from Sandy Islet, Scott Reef (using Fastloc data). A SN49666; B SN49669; C SN49675; D SN49676. 46 8

Figure 3 12: Inter nesting movements of four green turtles tracked from Sandy Islet, Scott Reef (using Fastloc data). A SN49686; B SN50060; C SN50062; D SN50065. 47 Figure 3 13: Inter nesting movements of three green turtles tracked from Sandy Islet, Scott Reef (using Fastloc data). A SN50066; B SN50070; C SN50075. 48 Figure 3 14: Migration route of SN49666, SN49669, SN49675, SN49676, SN50062 and SN50065 following nesting (Foraging locations from Pendoley 2005). 51 Figure 3 15: Migration route of SN49686 and SN50075 following nesting (Foraging locations from Pendoley 2005). 52 9

List of Tables Table 2 1: Timing and duration of nightly surveys to Scott Reef and Browse Island 18 Table 2 2: The combinations of letters and colours used as temporary marks for each day of the survey on Sandy Islet. 21 Table 3 1: Petersen estimates with Bailey s correction and standard error (SE) of estimates for eight recapture dates at Sandy Islet, Scott Reef. 29 Table 3 2: Breeding seasonality of green turtles on Sandy Islet, Scott Reef, as demonstrated by the mean (± standard deviation, sd), number of nests and tracks during the survey periods that spanned most seasons (Guinea 2009). The results from the 2010 survey are highlighted in grey. 30 Table 3 3: Biometrics for nesting green turtles at Sandy Islet, Scott Reef and Browse Island during the 2010 survey. 31 Table 3 4: Sizes of nesting female green turtles at Sandy Islet, Scott Reef from 2006 to 2010. 32 Table 3 5: Clutch Size, Hatching Success and Emergence Success for Scott Reef and Browse Island from 2008 to 2010. 37 Table 3 6: Biometrics for green turtles hatchlings (samples of ten hatchlings from each individual nest) from Sandy Islet, Scott Reef in the 2008/2009 and 2009/2010 nesting seasons and from Browse Island during the 2010 survey. 38 Table 3 7: Daily summary of manta tow surveys of Sandy Islet reef crest waters and the outer reef crest waters in the southern lagoon, to 10 m in depth. Each manta tow lasted 15 minutes with two in water observers (combined results of both observers provided). 40 Table 3 8: Sky quality meter (SQM) values (mag/arcsec 2 ) for the nights ashore at Sandy Islet. Provided are the dates and the percentage of the lunar disk illuminated and the corresponding SQM values. The period for which the moon was visible during the night to dawn is shaded. The SQM values with asterisk (*) represent the hour segment in which tide peaked for that evening. (lunar data from http://aa.usno.navy.mil/cgi bin/aa_rstablew.pl). 41 Table 3 9: Processed Fastloc transmission information of the green turtles that were satellite tracked following nesting at Sandy Islet, Scott Reef. 44 Table 3 10: Inter nesting intervals of satellite tagged green turtles from Sandy Islet, including the maximum distance turtles moved away from the nesting beach during this period. Date of nesting is the nesting date and date returned is the first time after nesting that the turtle returned to Sandy Islet. 45 Table 3 11: Dive characteristics of six inter nesting green turtles at Scott Reef 49 Table 3 12 Summary of migration of green turtles tracked following nesting at Sandy Islet 50 Table 3 13: Dive characteristics of three green turtles while migrating to their foraging grounds 50 10

11

1 Introduction 1.1 Background Woodside Energy Ltd (Woodside) is proposing to develop the Torosa, Brecknock and Calliance gas fields located approximately 270 km off the Kimberley coast in the Browse Basin (Figure 1 1). While Brecknock and Calliance are located in deep water (400 m to 700 m), the Torosa field lies partly under Scott Reef. 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. 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 north south direction (with approximately 400 m remaining above water at extreme high tide) and approximately 60 m at its widest part. Figure 1 1: Location of Scott Reef, off the Kimberly Coast, north western Western Australia. Scott Reef is listed on the Register of the National Estate, with its heritage value detailed as providing, among other values, 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. 12

Five 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 (Woodside 2008). However, only green turtles and hawksbill turtles (Eretmochelys imbricata) have been observed breeding and nesting on Sandy Islet (Guinea 2006). 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. This document reports the findings of a field survey undertaken in January February 2010 to investigate nesting behaviour and movements of marine turtles at Scott Reef during their peak nesting period. The survey extends information obtained from previous surveys of nesting marine turtles at Scott Reef in 2006, 2008 and 2009 (Guinea 2006; Guinea 2009). The findings of this survey are being used to support the environmental impact assessment of upstream aspects of the Browse LNG Development, which includes proposed future infrastructure over the offshore Torosa gas field near Scott Reef. 1.2 Background on Green Turtle Breeding Cycle Marine turtle species have complex life cycles from which generalisations can be drawn for green turtles (Figure 1 2). The current understanding of the life cycle begins with hatchlings moving from the beach to oceanic and coastal habitats before returning some decades later to the natal beaches to breed. In the case of Scott Reef turtles, this life cycle may involve most reefs in the tropical Indian Ocean and northern Australia. During the breeding season the turtles from the Scott Reef management unit, a genetically isolated population, mate, nest and spend the inter nesting period associated with the reef. Non breeding turtles that forage and reside on the reef are thought to belong to other management units, although this has not been tested. 13

Source Miller 1997 Figure 1 2: Generalised life cycle of marine turtles showing the female cycle of 2 to 8 years (green), the inter nesting cycle of 8 to 16 days, the activities on Scott Reef (red) and the pelagic and neritic foraging stages (blue). Within the generalised life cycle, there are at least two cycles that involve breeding females using the reefs, lagoons and beaches at Scott Reef: the female nesting cycle every 2 to 8 years female green turtles migrate from their foraging grounds to Scott Reef to breed (remigration interval), the inter nesting cycle females spend 2 to 3 months within the vicinity of Scott Reef moving only short distances of some tens of kilometres or less and nesting from 3 to 6 times every 8 to 16 days. Female Nesting Cycle The female nesting cycle involves an individual female attaining sufficient body fat and reserves in the foraging grounds to become reproductive (Hamann et al. 2003b). An unknown trigger initiates her migration, up to several thousand kilometres, from foraging grounds to a mating area close to her natal beach (Miller 1997). This migration coincides with similar movements by males (Hamann et al. 2003a). After mating, the female turtle becomes unreceptive to further courtship (Booth & Peters 1972) and moves closer to the nesting beach over the next few weeks. At this point, she enters the inter nesting cycle(hamann et al. 2003b). After laying several clutches of eggs, and having depleted 14

the fat reserves and ripe ovarian follicles, the females undertake post breeding migration to the foraging grounds (Hamann et al. 2003a). Being herbivorous, the time until the next breeding migration depends on the richness of the algal beds and seagrass pastures on the foraging grounds(miller 1997). Typically, at least three years or more is required to replenish the fat reserves and to develop mature ovarian follicles (Miller 1997, Arthur et al. 2009). The time between successive reproductive periods is called the remigration interval. Neophytes (first time nesters) are thought to undertake a breeding migration when they have sufficient fat reserves which have accumulated over many years (Hamann et al. 2003b, Miller 1997). However, they respond to the same triggers that initiate the reproductive migration of seasoned nesters and make the journey to their natal beaches (Plotkin 2003, Miller & Limpus 2003). In nesting seasons following a period of low productivity on the foraging grounds, the neophytes remain capable of making the nesting migration where as recently returned females may lack body condition (Miller & Limpus 2003). Small turtles may therefore have a disproportionately higher representation amongst the nesting population when fewer turtles make the breeding migration (Hamann et al. 2003b, Limpus et al. 2003). The remainder of the nesting population in sparse years may comprise individuals whose foraging grounds are closer and require fewer body reserves, individuals that have not nested for some time and individuals that are residents of very productive foraging grounds (Miller 1997, Hamann et al. 2003a, Miller & Limpus 2003). Inter nesting Cycle Once nesting starts, the inter nesting cycle is approximately two weeks in duration (Hamann et al. 2003a, Miller 1997, Hirth 1993). The female occupies an inter nesting habitat after each nesting event. For green turtles these inter nesting habitats are defined by their function rather than by location. Satellite tracking has demonstrated that during the inter nesting period, female green turtles visit different substrates, occupy different depths, and may move up to tens of kilometres from the nesting beach (Hays et al. 2002b, Alvarado & Murphy 1999, Bell et al. 2009, Limpus et al. 2003). Turtles that remain close to the nesting beach also appear to move through several near shore habitats during the inter nesting interval (Limpus & Reed 1985, Limpus 2008). 1.3 Scope of Works Woodside engaged the CES Partnership with Charles Darwin University (CDU), to develop and undertake Long Term Monitoring of Marine Turtles at Sandy Islet Scott Reef. The long term marine turtle monitoring programme at Scott Reef has involved yearly field trips to Scott Reef over a two week period in January/February. The scope of the surveys includes programmes for tagging, quantifying nest and hatchling success, satellite tagging of nesting females and an assessment of habitat usage and environmental data. This report presents the results from the 2010 survey. The objectives of the survey in 2010 were to: record the number of marine turtles nesting on Sandy Islet, Scott Reef and Browse Island; provide long term individual turtle identification using numbered titanium tags; record the length and width of green turtle carapaces for morphometric analyses; collect biopsy samples from untagged nesting marine turtles and dead hatchlings; 15

evaluate the hatching and emergence success of clutches; attach a Platform Transmitter Terminal (PTT) to the carapaces of twelve nesting green turtles to allow satellite tracking of their movements; conduct manta tow surveys to identify inter nesting habitat; and record GPS locations of the extent of nesting areas on Sandy Islet. The survey of January and February 2010 was conducted under authorisation, No. A02033, from the Animal Experimentation Ethics Committee of Charles Darwin University and permit, SF007200, to take Fauna for Scientific Purposes and permit, CE002687, to enter CALM land or Waters from the Department of Conservation, Western Australia. 16

2 Methods The methods used in this survey were consistent with those used in previous surveys at Scott Reef (Guinea 2009). An outline of the programme is shown below, with a summary of the methods provided in the subsequent sections. A full description of the methods can be found in Guinea (2010). The tagging programme (Section 2.2): recorded the number of marine turtles nesting on Sandy Islet (over thirteen nights) and Browse Island (over one night); provided long term individual turtle identification using numbered titanium tags; recorded the length and width of green turtle carapaces for morphometric analyses; and collected biopsy samples from untagged nesting marine turtles. The nest/hatchling programme (Section 2.3): collected biopsy samples from dead hatchlings for DNA analysis; and evaluated the hatching and emergence success of clutches. The satellite transmitter programme (Section 2.4): attached a Platform Transmitter Terminal (PTT) to the carapaces of twelve nesting green turtles at Sandy Islet. The identification and use of habitat within Scott Reef programme (Section 2.5): used manta tow surveys to identify inter nesting habitat; and recorded GPS locations of the extent of potential nesting areas on Sandy Islet. The collection of environmental data relevant to the nesting biology of marine turtles (Section 2.6): established a data logging station for sand temperatures at nest depth ( 50 cm); and recorded the brightness of the night sky during marine turtle nesting at Sandy islet over a lunar cycle. All field data was entered into an Excel spreadsheet of nightly summaries, tag registers and register of DNA samples are detailed in Appendix A. 2.1 Survey Dates The survey dates, 29 January to 12 February 2010, coincided with a period of full to new moon, when spring tides reached their maximal ranges (Table 2 1). 17

Table 2 1: Timing and duration of nightly surveys to Scott Reef and Browse Island Date Moon Phase Maximum Tide Height (m) Arrival Time Departure Time Time Ashore (hr) Scott Reef 29/01/2010 3.5 1800 0030 6.5 30/01/2010 Full Moon 3.7 1930 0145 6.25 31/01/2010 3.8 2115 0315 6.00 1/02/2010 3.8 2200 0400 6.00 2/02/2010 3.6 2215 0400 5.75 3/02/2010 3.7 2015 0400 7.75 4/02/2010 3.7 Mechanical issues with the tender prevented access to Sandy Islet. 5/02/2010 3.6 0030 0430 4.00 6/02/2010 3.4 2345 0515 5.50 7/02/2010 Third Quarter 3.1 0030 0540 5.16 8/02/2010 2.8 1900 0400 2230 0545 3.50 1.75 9/02/2010 2.6 1630 2300 6.50 10/02/2010 2.8 1815 2145 3.50 11/02/2010 3.0 1800 0130 7.50 Total number of Hours Ashore at Sandy Islet 75.66 Browse Island 12/02/2010 4.2 1050 2030 1300 2400 2.16 3.50 Total Number of Hours Ashore 5.56 Previous surveys at Scott Reef (e.g. Guinea 2009) indicated the last week of January and the first week of February was the peak of nesting when approximately 70% of the season s breeding population would nest during the two week period. During the peak nesting period, nests laid earlier in the season were hatching and the success of incubation could be gauged. 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. On the night of 4 February 2010 the field team was unable to access Sandy Islet due to technical difficulties related to the tender vessel. 2.2 Tagging Programme Sandy Islet and Browse Island were patrolled on foot using ambient moonlight and starlight to gain a clear view of the beach. When ambient light was insufficient, a lighting protocol was followed to reduce extraneous light spill that could disturb nesting turtles. The beach at Sandy Islet was divided into 13 sectors; each approximately 100 m in length (Figure 2 1). 18

Figure 2 1: Location of numbered beach sectors at Sandy Islet. Each sector was approximately 100 m in length. 19

The marker at the junction of sector 2 and sector 3 coincided with the south western leg of the derelict weather station for future replication. In sector 13 only tracks and nests were recorded due to the influence of human disturbances as it served as the base camp for the survey team. Tracks on the sand spit to the south of the sand cay were considered false crawls when there was no attempt to nest and the tracks lead into the sea on the opposite side of the spit. Newly observed turtle tracks in each sector were identified as either ascending or descending the beach and recorded within each sector. Tracks were delineated to avoid duplication in the dataset (Schroeder & Murphy 1999). Turtle tracks were followed and recorded as: having laid if a nest was identified; having not laid if the tracks were followed back to the ocean with no identification of a nest; or unknown if the track was lost and therefore it was unknown if the turtle nested. Nesting turtles were tagged with individually numbered titanium tags (WA series). A tag was applied to the innermost scale (axillary scale) on the trailing edge of the left and right front flippers. Carapace dimensions and any individual distinguishing marks or injuries were recorded for each tagged turtle. The curved carapace length (CCL) and the curved carapace width (CCW) were measured with a fibreglass tape (Guinea 2010). Univariate analysis (t tests) was undertaken to compare the data gathered in 2010 at Sandy Islet to that recorded during the 2008/2009 season as well as data available from other northern Australian green turtle rookeries. A skin biopsy (4.5 mm by 2.5 mm) was collected for the trailing edge of the hind flipper with a sterile leather punch and stored in a Dimethyl Sulfoxide (DMSO) and Sodium Chloride (NaCl) solution (Guinea 2010). Vials were individually labelled and sent to Dr Nancy FitzSimmons (University of Canberra) for sequencing. The carapace was marked with spray paint with a large letter indicating the day of the survey and a numeral indicating the order in which the turtle was encountered on that day (Table 2 2). Those turtles that nested successfully were marked on the parietal scale 1 with a non toxic coloured acrylic aerosol. A different colour and letter were used for each day of the survey. Individuals could be identified on the beach at a distance as having laid or having not laid yet. This technique minimised human disturbance to nesting turtles. Marking the parietal scale distinguished successful nesters from unsuccessful nesters in inter nesting habitats. The capture, release and recapture of individuals (identified during manta tow surveys, see Section 2.5 for details) is used to calculate a population estimate using the Petersen estimate (N) of the population and Bailey s correction for improving the accuracy of the estimate of an incomplete census (Caughley 1977, Gerrodette & Taylor 1999, Limpus et al. 2003) as follows: 1 Scales on the top of the turtles head 20

N = M ( n + 1) /( m + 1) [ M ² ( n + 1)( n m) ( + 1 ) ² ( m + 2) ] SE = m 1/2 N = estimate of the population M = number of sprayed individuals marked prior to the day of observation m = sample of sprayed individuals n = adult short tailed individuals observed (adult female turtles or immature males) Table 2 2: The combinations of letters and colours used as temporary marks for each day of the survey on Sandy Islet. Date Letter Marking Carapace Colour Parietal Colour 29 Jan 2010 A Orange Yellow 30 Jan 2010 B Yellow Pink 31 Jan 2010 C Pink Yellow 1 Feb 2010 D Green Pink 2 Feb 2010 E White Green 3 Feb 2010 F Blue White 4 Feb 2010 G 5 Feb 2010 H Red Blue 6 Feb 2010 I Red Yellow 7 Feb 2010 J Yellow Green 8 Feb 2010 K White Orange 9 Feb 2010 L Green White 10 Feb 2010 M Pink Green 11 Feb 2010 N Any colour Any colour 2.3 Nest/Hatchling Programme Marine turtle nest data were collected to determine nesting success at Sandy Islet. Identified hatched nests were excavated in order to determine the nest contents. The nest contents were recorded using the categorisation system of Miller (1999). The data was then used to calculate clutch size, hatching success and emergence success. A sample of ten hatchlings from each hatched nest were weighed, measured (straight carapace length and width) and the major head and carapace scutes counted. Univariate analysis (t tests) was undertaken to compare the data gathered in 2010 at Sandy Islet to that recorded during the 2008/2009 season as well as data available from other northern Australian green turtle rookeries. 2.4 Satellite Transmitter Programme Satellite transmitters (6 Sirtrack transmitters and 6 St Andrews Sea Mammal Research Unit SMRU transmitters) were attached to 12 green turtles following nesting at Sandy Islet. The procedures followed the general guidelines for attaching satellite tags (Mitchell 1998, Coyne et al. 2008). 2.4.1 Data Processing and Analysis The Sirtrack and SMRU transmitters produce two types of data; ARGOS LC data and Fastloc GPS data. Only Fastloc data were used in the analysis as it has a higher level of spatial accuracy compared to ARGOS LC data. The accuracy of ARGOS LC data ranges from ± 150 m to >1000 m while the Fastloc 21

systems derives position fixes using six or more satellites and have an error of less than 100 m (Bryant 2007). The Fastloc data were downloaded from the Argos satellite system and processed using a specially developed SirTrack/SMRU software program. The data was then filtered to remove inaccurate location points. The filter removed locations on land (based on the GeoScience Australia Coastline 100k), speeds greater than 5 km/h, locations with turning angle greater than 70 or turning angle less than 70, and locations with step length greater than 150 km. Filtering for the SMRU data was automated within the software program and locations were eliminated based on speeds over 5 km/h and residuals greater than 100. The following definitions were used during data processing: Nesting attempt where a fix was located on Sandy Islet Successful nesting attempt where a fix was located on Sandy Islet and was not recorded on Sandy Islet for the following seven days. If the renesting interval was more than 12 days (based on 12 days renesting interval for green turtles in the northern Great Barrier Reef; Limpus 2007) fixes within 200m were examined and the last recorded fix within 200m was used as the nesting date. Nesting interval time in days from one successful nesting attempt to the next nesting attempt (whether successful or not) (Whiting et al. 2007). Migration starting point the first day of movement away from Scott Reef after the last successful nesting. Migration end point when the tracks began to traverse within a defined area for an extended period of time. As the SMRU transmitters record and transmit dive and out of water data, the haul out/out of water data allowed confirmation of nesting attempts as defined below: Nesting attempt where a fix was located on Sandy Islet which coincided with a haul out event. Successful nesting attempt where a fix was located on Sandy Islet which coincided with a haul out event and was not recorded on Sandy Islet for the following seven days. The SMRU satellite transmitters also recorded and transmitted a summary of dive data (including maximum depth, average dive depth, average dive duration, % surface time, % dive time and % haul out/out of water time) over 6 hour blocks. While data on individual dives are available, there are limits on the number of dives uploaded to the satellite (due the limitations in the satellite link i.e. time and size of data) therefore, the summary provides a more robust dataset. There are limitations to the summary data as well, as only a proportion of the recorded 6 hour blocks are transmitted. Subsequently there will be missing data for both entire days as well as blocks of days. 22

2.5 Habitat Identification 2.5.1 Nesting Habitat An outline of the suitable nesting habitat and its usage was determined by the locations of nests or body pits on the outer edge of Sandy Islet. GPS locations were recorded for all nests that were laid during the survey. 2.5.2 Inter nesting Habitat Visual in water transects for inter nesting marine turtles were conducted using paired manta boards on a daily basis during the survey. Manta tows were conducted in the reef crest waters of Sandy Islet providing visibility to a depth of 10 m (Figure 2 2). These surveys were conducted in accordance with agreed protocols to record the time of observations, stage of maturity and the sex of the turtles. Sex was recorded by identifying long tailed turtles (males) and short tailed turtles (females and immature males). Turtles were examined for carapace paint, tags and satellite tags. The numbers of turtles observed on the port and starboard sides of the tow were recorded separately and duplicate sightings were removed after discussions with the observers and reference to the time of sighting. Each manta tow survey covered approximately 1 km. Approximately 8 to 10 km were surveyed each day resulting in a total of 87 km covered throughout the 2010 survey. 2.6 Environmental Data 2.6.1 Sand temperatures Two temperature recording data loggers (Hobo Pro water temperature) were positioned together at 50 cm depth on Sandy Islet at one location on 3 February 2010 (14º 03.318 S, 121º 46.636 E). The temperature of the sand during the mid third of incubation is closely linked with the sex of the hatchlings produced (Guinea 2010). The data loggers were set to record temperature every 30 minutes and will record continuously for approximately 600 days. 2.6.2 Sky Glow The brightness of the night sky affects the behaviour of nesting turtles. Females may be more unsettled by the light and prone to abandon nesting attempts during the full moon. Sky glow is the wide scale illumination by visible light in the night sky. This was measured using the Sky Quality Meter (SQM Unihedron) only while the research team was on Sandy Islet. This approach was used to avoid extraneous light from the charter vessel interfering with sky glow measurements. Each hour while ashore, three Sky Quality measurements were taken by holding the meter at arms length vertically towards the zenith. The modal value was used in calculations. The values recorded in mag/arcsec 2 were converted to luminance units (candela per square metre) to produce a reading more meaningfully related to human vision. The luminance values (cd/m 2 ) = 10.8 104 10( 0.4*[SQM value in mag/arcsec 2 ]) (http://www.unihedron.com/projects/darksky/magconv.php). During the survey period, the full moon occurred on 30 January and the new moon on 14 February 2010. 23

Figure 2 2: The locations of the manta tow surveys of Sandy Islet reef crest and southern reef are shown for each day. 24

3 Results 3.1 Tagging Programme As there was no evidence (tracks, turtles or hatchlings) of hawksbill turtle nesting on Sandy Islet during January to February 2010, the results and discussion refer to green turtles only, unless otherwise stated. 3.1.1 Turtle Nesting Activity Sandy Islet Sandy Islet was patrolled for thirteen nights during which time 56 nesting turtles were tagged. During this period, 171 tracks were observed and resulted in identifying 46 nests on the islet. Most nests were laid above the spring high tide water mark, but during the neap tides nests were laid on the southern sand spit as well as on the more elevated part of the islet (Figure 3 1). Sectors 3, 4, 5, and 6 on the northern and western side of the islet were the most popular sectors for turtles emerging from the water, nesting and hatching (Figure 2 1; Figure 3 2). Turtles which emerged in those sectors of the beach then moved across the islet. The beach rock along sectors 1 and 2 (on the eastern side of the islet) deterred females from coming ashore except on the nights of the spring tides when the water level was above the rocks. Monitoring began on the evening prior to the full moon and continued until three days before the new moon. Nesting activity was greatest during the spring tides associated with the new and full moons (Figure 3 3). There appeared to be an association between the numbers of turtles attempting to nest, and successfully nesting with the height of the nocturnal high tide. The highest high tide occurred on the two nights following the full moon, but the greatest number of tracks and nests did not occur until 9 and 10 February 2010 as turtles favoured nesting on the early evening high tide before the moon was visible. 25

Figure 3 1: Nesting area at Sandy Islet was identified by body pits on arrival (shaded area). Nests laid during the survey, 29 January to 11 February 2010, are indicated in green. 26

Nesting Activity by Sector Numbers of Tracks, Nests and Hatched Nests per Night 40 35 30 25 20 15 10 5 0 1 2 3 4 5 6 7 8 9 10 11 12 13 Beach Sector Tracks Nests Hatched Nests Figure 3 2: Numbers of turtle tracks, nests and hatched nests in each of the sectors of Sandy Islet, Scott Reef during the survey period. Nightly Nesting Numbers 30 4 Numbers of Tracks and Nests 25 20 15 10 5 3.5 3 2.5 2 1.5 1 0.5 Height of Night Tide (m) 0 0 29/01/2010 30/01/2010 31/01/2010 1/02/2010 2/02/2010 3/02/2010 4/02/2010 5/02/2010 6/02/2010 7/02/2010 8/02/2010 9/02/2010 10/02/2010 Date Nests Tracks Height of high water Figure 3 3: Numbers of green turtle tracks and nests for each night of monitoring and the height of the nocturnal high tide. The full moon occurred on 30 January 2010. 27

Browse Island Rough seas prevented a lengthy survey of Browse Island. A survey of just over two hours during daylight established that 45 nests from 104 tracks had been laid in the last few nights since the last set of spring tides (approximately 7 days given the wind on the Island). Nine sets of hatchling tracks were visible amongst the coral rubble. Another two hours were spent ashore at night during which time 16 nesting green turtles were tagged. 3.1.2 Inter Season and Within Season Recaptures Sandy Islet The re nesting interval ranged from 6 to 12 days (n=6) with a mean of 9.7 ± 2.1 days. There was no recapture of any turtle that had been tagged in previous years. The average number of nesting events within a season for turtles on Sandy Islet and the remigration interval remain unknown. Browse Island Browse Island was surveyed on only one night, therefore inter nesting intervals could not be determined. 3.1.3 Estimates of Population Size Sandy Islet Between 29 January and 11 February 2010, 56 female turtles were tagged while ashore at Sandy Islet. Therefore estimates of the size of the nesting population should be above this value, as at least 56 female turtles nested on Sandy Islet during the 2009/2010 nesting season. Petersen estimates, with Bailey s correction and standard errors, were calculated for the eight days for which recaptures were reported (Table 3 1). Estimates of turtle abundance varied from 45 to 107 individuals (mean ± standard deviation = 79 ± 25.0; n = 8) with individual standard errors reaching 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 (Caughley 1977, Gerrodette & Taylor 1999, Limpus et al. 2003) which are unverified in this study. Browse Island No manta tow surveys were conducted at Browse Island therefore no population estimates could be calculated. 28

Table 3 1: Petersen estimates with Bailey s correction and standard error (SE) of estimates for eight recapture dates at Sandy Islet, Scott Reef. Date # Marked 1 #Recapture 2 Sample Size 3 Petersen Standard Estimate Error (SE) 3/02/2010 29 3 9 72.5 25.11 5/02/2010 34 2 3 45.3 11.33 6/02/2010 37 5 11 74.0 19.77 7/02/2010 38 1 3 76.0 31.02 8/02/2010 46 5 13 107.3 30.66 9/02/2010 50 0 1 100.0 50.00 10/02/2010 51 6 7 58.2 7.28 11/02/2010 54 5 10 99.0 25.22 1 Total number of turtles marked (cumulative total) 2 Number of identified turtles during manta tows that were marked 3 Total number of identified turtles during manta tows 3.1.4 Inter annual Variability in Nesting Population and Intensity Sandy Islet Comparing the results from the 2010 survey with previous summer nesting seasons highlights the natural inter annual variability in turtle nesting intensity and population size (Guinea 2009; Table 3 2; Figure 3 4). The decrease in the number of nesting turtles in 2010 (56) compared with January 2009 (193) is obvious, regardless of the increased effort provided in the 2010 survey (Figure 3 4). This reduced nesting intensity was also illustrated in the number of tracks per night. During the current 2010 survey, the number of tracks varied from 6 to 26 with a mean of only 13.2 while in January 2009 the mean number of tracks was higher by more than a factor of five (71.9). The estimate of the nesting population size for the 2010 nesting season (79 ± 25.0) also demonstrated a large decrease from the estimates calculated for the nesting population in 2008/2009 (779 ± 383). While there were large differences in nesting activity and estimates of population size between different nesting seasons the results from previous surveys demonstrated no trends in the tracks to nest ratio. Nesting females make several trips ashore before successfully nesting regardless of the season (Table 3 2). 29

Table 3 2: Breeding seasonality of green turtles on Sandy Islet, Scott Reef, as demonstrated by the mean (± standard deviation, sd), number of nests and tracks during the survey periods that spanned most seasons (Guinea 2009). The results from the 2010 survey are highlighted in grey. Season Month Year #Green Turtles tagged Mean # of tracks / night ± sd Mean # of nests/ night ± sd # hatched Nests Tracks: Nest Ratio Number of Nights Surveyed Time Ashore (hr) by Survey Team Summer Feb 2006 29 16.2 ± 4.87 6.0 ± 4.82 3 2.70 3 24.0 Dec.2008 107 53.1 ± 16.87 26.9 ± 9.32 10 1.98 7 46.5 Jan 2009 193 71.9 ± 6.98 29.1 ± 4.52 13+ 2.50 7 67.5 Feb 2010 56 13.2 ± 5.43 3.8 ± 2.52 21 3.43 13 75.7 Winter Aug 2008 10 9.3 ± 1.15 2.0 ± 1.0 1 4.67 2 26.0 Spring Sep 2006 0 0 0 0 1 9.0 Nov 2006 92 23.6 ± 7.57 4.0 ± 0.71 1 5.90 5 45.0 Sep 2008 4 7.8± 5.89 2.4 ± 1.14 1 3.25 6 35.0 Initial surveys, prior to 2008, occurred opportunistically with other studies occurring at Scott Reef. Subsequent surveys in 2008/2009 were designed in order to delineate the peak nesting period. The 2010 survey was timed to occur during the peak nesting period therefore the most appropriate data for comparison is the January 2009 data set as it also occurred during the defined peak nesting period. 30

2500 Scott Reef Green Turtle Nesting Population (range, n +/- s.e.) Estimate of Nesting Population 2000 1500 1000 500 0 14-16 Dec 2008 26-31 Jan 2009 3-11 Feb 2010 Dates of Survey Source Guinea 2009 Figure 3 4: Change in estimated nesting population of green turtles at Scott Reef in December 2008, January 2009 and February 2010. The data presented are the mean for the days sampled with standard error and the minimum and maximum estimates. 3.1.5 Biometrics of Nesting Females During the 2010 survey, the curved carapace length (CCL) for nesting turtles at Sandy Islet ranged from 82.8 cm to 110.0 cm, while at Browse Island the CCL ranged from 94.3 cm to 105.1 cm. The curved carapace width (CCW) for nesting turtles at Sandy Islet ranged from 77.6 cm to 102.1 cm, while at Browse Island the CCW ranged from 83.2 cm to 97.4 cm (Table 3 3). There was no significant difference in CCL (t = 1.06, p > 0.05, df = 63) and CCW (t = 0.44, p > 0.05, df = 63) of green turtles from the two localities. Overall, 10% of nesting females at Sandy Islet had a CCL below 90 cm. This is a 5% increase from the 2008/2009 survey where 19 turtles (5% of nesting turtles) at Sandy Islet had a CCL below 90 cm (Guinea 2009). Table 3 3: Biometrics for nesting green turtles at Sandy Islet, Scott Reef and Browse Island during the 2010 survey. Location (sample size) Scott Reef (n = 50) Browse Island (n= 15) Curved Carapace Length (cm) Curved Carapace Width (cm) Mean ± sd Minimum Maximum Mean ± sd Minimum Maximum 98.5 ± 5.8 82.8 110.0 89.6 ± 5.6 77.6 102.1 99.8 ± 3.5 94.3 105.1 90.2 ± 4.3 83.2 97.4 31

These measurements are within the ranges previously recorded at Sandy Islet. A total of 561 individual female turtles have been measured since 2006 at Sandy Islet. Damage to the carapace prevented the measurements of CCL and CCW for some individual nesting turtles and CCLs have been attained for 538 nesting turtles and CCWs attained for 549 nesting turtles (Table 3 4).The CCL of these nesting females varied from 82.8 cm to 117.3 cm (mean ± standard deviation = 97.7 ± 5.36 cm; n = 538). CCW varied from 73.0 cm to 106.0 cm (mean ± standard deviation = 88.5 ± 5.56 cm; n = 549). A linear relationship exists between CCL and CCW (Figure 3 5). Table 3 4: Sizes of nesting female green turtles at Sandy Islet, Scott Reef from 2006 to 2010. Date Mean s. d. 1 Sample size Minimum Maximum Curved Carapace Length (cm) Feb 06 97.70 5.50 27 86.3 107.7 Sep 06 Nov 06 97.77 6.23 78 88.0 117.3 Aug 08 95.50 3.92 10 90.5 101.5 Sep 08 100.22 3.25 5 94.8 102.9 Dec 08 97.78 4.99 123 82.8 110.0 Jan 09 97.16 5.12 204 87.1 110.3 Feb 10 98.52 5.76 51 82.8 110.0 Combined 2006 10 97.7 5.36 538 82.8 117.3 Curved Carapace Width (cm) Feb 06 88.40 6.30 28 77.5 100.0 Sep 06 Nov 06 88.43 6.03 80 79.7 106.0 Aug 08 87.36 3.69 10 83.0 92.0 Sep 08 91.60 4.70 5 83.6 95.1 Dec 08 88.46 5.54 123 73.0 105.4 Jan 09 88.09 5.47 212 77.4 105.8 Feb 10 89.6 5.57 51 77.6 102.1 Combined 2006 10 88.5 5.56 549 73.0 106.0 1 Standard deviation Source Guinea 2009 32

Curved Carapace Width (cm) 110 105 100 95 90 85 80 75 70 65 60 y = 0.8544x + 4.8371 R 2 = 0.4365 80 85 90 95 100 105 110 115 120 Curved Carapace Length (cm) Figure 3 5: Relationship between curved carapace length and curved carapace width for 488 individual female green turtles at Sandy Islet, Scott Reef. The present study provides a robust data set of nesting turtle morphometrics. Based on available data, the nesting turtles of Sandy Islet and Browse Island are slightly larger in average CCL than those from the Lacepede Island nesting populations (Figure 3 6) but not significantly (t = 1.69, p < 0.05, df = 39). When compared with other nesting populations in northern Australia (Limpus 2007), the nesting turtles of Scott Reef are significantly smaller (Figure 3 7), on average, by about 3 cm in CCL (t = 4.19, p < 0.05, df = 547). The differences become statistically significant because of the similar variance and the large sample sizes from the rookeries in Queensland. Few comparisons of CCW can be made as few data are provided from other studies. However, the Scott Reef green turtles are significantly smaller (t = 3.35, p < 0.05, df = 647) in CCW than those nesting on Ashmore Reef (Guinea et al. 2005). 33

Green Turtles in Western Australia Curved Carapace Length (cm) 120 115 110 105 100 95 90 85 80 Lacepede Island Scott Reef Browse Island Locality Figure 3 6: Comparison of the curved carapace length of green turtles nesting at Lacepede Island (n = 22), Scott Reef (n = 538) and Browse Island (n = 19). The data presented are the mean with standard deviation and the maximum and minimum lengths. Green Turtle Curved Carapace Length (cm) 140 130 120 110 100 90 80 Scott Rf Ashmore Rf Heron I Raine I Bountiful I Coral Sea Locality Figure 3 7: Comparison of the curved carapace length of green turtles nesting at Scott Reef (n = 538) with those nesting at other localities in northern Australia: Ashmore Reef (n = 102), Heron I (n = 1942), Rain I (n = 20,947), Bountiful I (n = 219) and Coral Sea islands (n = 100) (Limpus 2007). The data presented are the mean with standard deviation and the maximum and minimum lengths. 34

3.2 Nest/Hatchling Programme 3.2.1 Clutch Size, Hatching Success and Emergence Success Sandy Islet Only eight of the 21 recorded hatched nests were examined as the location of the other nests could not be determined either due to the soft sand, hatchling tracks being obliterated by ghost crabs or by nesting turtles. Clutch sizes varied from 64 to 151 eggs. Hatching success varied from 39.6% to 100% and emergence success varied from 28.6% to 96.2% (Figure 3 8). Nest depths were recorded for all excavated nests. The depth to the top of the nest (first eggshell) ranged between 54 to 87 cm (mean ± standard deviation = 69.0 ± 12.15 cm) while depth to the bottom of the nest (last eggshell) ranged from 72 to 111 cm (mean ± standard deviation = 83.6 ± 13.33 cm). Sand temperatures at the bottom of these nests varied from 29.0 C to 29.9 C (mean ± standard deviation = 29.15 ± 0.33 C). Early 4% Dead 1% Late 4% Undeveloped 3% Live 12% Shells 76% Figure 3 8: Summary of the contents of hatched nests from Sandy Islet, Scott Reef. Shells = empty shells; live = live hatchlings; dead = dead hatchlings, early = unhatched egg containing dead embryo in early stages of development; late = unhatched egg containing dead embryo in late stages of development; undeveloped = unhatched egg with no embryo development. 35

Browse Island Hatchling tracks amongst the coral rubble characteristic of Browse Island indicated at least nine nests had recently hatched. Of these, only three were found and their nest contents examined. Clutch sizes varied from 62 to 135 eggs (mean ± sd = 105.7 ± 38.5 eggs). Hatching success varied from 77.0% to 87.1% and emergence success varied from 72.6% to 87.1% (Figure 3 9). The depth to the top of the nest ranged from 62 to 76 cm (mean ± standard deviation = 68.0 ± 7.2 cm) while that to the bottom of the nest ranged from 74 to 80 cm (mean ± standard deviation = 76.3 ± 3.2 cm). Sand temperatures at the bottom of these nests varied from 28.8 C to 30.3 C (mean ± standard deviation = 29.6 ± 0.75 C). Early 6% Late 4% Undeveloped 7% Dead 3% Live 3% Shells 77% Figure 3 9: Summary of the contents of hatched nests from Browse Island. Shells = empty shells; live = live hatchlings; dead = dead hatchlings, early = unhatched egg containing dead embryo in early stages of development; late = unhatched egg containing dead embryo in late stages of development; undeveloped = unhatched egg with no embryo development. Inter annual variations The emergence and hatching success between surveys in 2008 and 2009 at Sandy Islet were relatively similar to the results from the 2010 survey (Table 3 5). There were differences in the mean clutch sizes from the surveys in 2008/2009 season and those in the 2010 survey but these were not significant (t = 1.82, p < 0.05, df = 20). Significantly larger clutch sizes (t = 2.87, p < 0.05, df = 62) are reported for green turtles nesting at Heron Island, Raine Island and Bramble Cay. There was no significant difference between nest depths in 2008/2009 season and the 2010 survey at Sandy Islet (t = 1.57, p > 0.05, df = 8). Similarly, there was no significant difference between the maximum nest depths between Sandy Islet in the 2008/2009 season and the nests excavated on Browse Island (t = 1.87, p > 0.05, df = 3). Although the nests excavated on Sandy Islet in the 2010 survey were significantly shallower than the nests on Browse Island (t = 2.48, p < 0.05, df = 9). Maximum nest depths on Sandy Islet in the 2008/2009 season were shallower that than those 36

reported on Raine Island (t = 3.51, p < 0.05, df = 138) but deeper than those on Heron Island (t = 2.21, p < 0.05, df = 17). Table 3 5: Clutch Size, Hatching Success and Emergence Success for Scott Reef and Browse Island from 2008 to 2010. Location Scott Reef Scott Reef Scott Reef Browse Island Survey Dec 08 Jan 09 Jan/Feb 2010 Jan/Feb 2010 Sample Size (n) 4 10 8 3 Clutch Sizes (Eggs) Hatching Success (%) Minimum 50 66 64 62 Maximum 111 132 151 135 Mean 83 86.3 105.6 105.7 s.d. 25 21.5 26.9 38.5 Minimum 65.8 77.1 39.6 77 Maximum 89.9 95.4 100 87.1 Mean 80.2 88.5 88.2 83.3 s.d. 10.9 5.9 19.9 5.48 Emergence Success (%) Minimum 45.1 54.1 28.6 72.6 Maximum 84.9 95.4 96.2 87.1 Mean 70.3 80.9 74.3 81.8 s.d. 17.4 12.1 22.7 8 3.2.2 Biometrics of Hatchlings Samples of 10 hatchlings from each of the six hatched nests from Sandy Islet, Scott Reef (n = 60) and samples of 10 hatchlings from each of the three hatched nests from Browse Island (n = 30) were weighed and measured. Findings are shown in Table 3 6; in addition to results from Sandy Islet in 2008/2009. There were no significant differences in the SCL (t = 0.49, p < 0.05, df = 12), SCW (t = 0, p > 0.05, df = 12) between the 2008/2009 season and 2010 survey at Scott Reef and the 2010 survey of Browse Island. However, the mean mass of hatchling samples differed significantly from the 2008/2009 season and the 2010 survey (t = 3.16, p < 0.05, df = 15). The average mass of hatchling green turtles from Scott Reef in the 2008/2009 season were significantly larger (t = 2.53, p< 0.05, df = 12) than those in the samples from Scott Reef and Browse Island in 2010 and from Ashmore Reef (Guinea et al. 2005). There was no significant difference in SCW between samples of hatchlings from Scott Reef in 2009/2009 and 2010 seasons, and hatchlings form Ashmore Reef (Guinea et al. 2005) (t = 1.24, p > 0.05, df = 14). The Scott Reef hatchlings from this and previous years were significantly smaller (t = 2.67, p < 0.05, df = 33) in SCL than those reported from Heron Island, Raine Island and Bramble Cay. 37

Table 3 6: Biometrics for green turtles hatchlings (samples of ten hatchlings from each individual nest) from Sandy Islet, Scott Reef in the 2008/2009 and 2009/2010 nesting seasons and from Browse Island during the 2010 survey. Nesting Season Location (n = nests) 2008/2009 Scott Reef (n = 12) 2009/2010 Scott Reef (n = 6) 2009/2010 Browse Island (n = 3) Straight Carapace Length (cm) Straight Carapace Width (cm) Mass (g) Mean ± sd Minimum Maximum Mean ± sd Minimum Maximum Mean ± sd Minimum 4.7 ± 0.14 4.3 5.2 3.7 ± 0.1 3.5 3.9 24.1 ± 2.0 20.0 29.5 4.7 ± 0.14 4.5 4.8 3.7 ± 0.1 3.6 3.8 21.7 ± 1.1 20.1 22.8 4.6 ± 1.0 4.5 4.8 3.7 ± 0.6 3.7 3.8 21.9 ± 0.7 20.5 23.0 Maximum 38

3.3 Biopsy Analysis A total of 47 biopsy samples were collected from nesting females and dead hatchlings for DNA analysis and identification of genetically differentiated populations. These samples were forward to Dr. Nancy FitzSimmons at the University of Canberra for analysis and results are not included within this report. 3.4 Marine Turtle Habitats 3.4.1 Nesting Habitats The highest part of the sand cay that accommodated consistent nesting was approximately 60 m wide by 400 m in length in 2010. This trend is consistent with the findings from previous surveys in 2006, 2008 and 2009. The southern and northern sand spits had again changed shape and position since previous surveys (Figure 3 1). Turtle nests and attempted turtle nests extended along the entire length of Sandy Islet and the sand spits. The only part of the islet that was not freshly disturbed by nesting turtles in the 2010 surveys was the small stretch on the northern beach that was behind an expanse of beach rock. The paths taken by the turtles moving to and from the nest localities extended across the sand cay leaving very little of the rest of the islet untouched by nesting turtles. 3.4.2 Inter nesting Habitats In total, 145 sightings of turtles occurred during the manta board surveys at Scott Reef. Immature turtles comprised 42% of sightings although many of these were probably multiple sightings of the same individuals (Table 3 7). There were only nine sightings of male turtles, which was equivalent to 6% of total sightings and again multiple sightings were suspected. Adult turtles with short tails (either female turtles or immature males) were recorded in nearly every tow and comprised 51.7% of the counts. Short tailed turtles carrying tags (females) comprised 18.6% of the sightings. Several individual females were sighted more than once with one (WA78253/WA78254; C02; SN49675) sighted three times in two different locations, on the eastern and western side, on three separate surveys. These positions on the eastern and western side of Sandy Islet reef coincided with satellite locations (refer to Figure 3 11C in Section 3.6). Satellite fixes from turtles in the southern lagoon prompted manta tows at the southern reef. There was no visual contact with any adult turtles in that part of Scott Reef. However manta tows were conducted along the outer line of reef while the satellite track locations have indicated that the turtles were located further south on the reef flat. 39

Table 3 7: Daily summary of manta tow surveys of Sandy Islet reef crest waters and the outer reef crest waters in the southern lagoon, to 10 m in depth. Each manta tow lasted 15 minutes with two in water observers (combined results of both observers provided). Date Number of Manta Tows Kilometres Surveyed (km) Juvenile Adult Male Unmarked Adult Short tail 31/01/10 4 4 5 9 1/02/10 6 6 7 6 2/02/10 7 7 24 2 3 3/02/10 8 8 1 1 6 3 5/02/10 8 8 4 3 1 2 6/02/10 10 10 6 2 6 5 7/02/10 10 10 3 1 2 1 8/02/10 8 8 6 8 5 9/02/10 8 8 1 10/02/10 10 10 3 1 6 11/02/10 8 8 2 5 5 Total 87 87 61 9 48 27 Marked Female 3.5 Sky Glow During the survey period, the nights were mostly cloudless with only a few scattered clouds passing from the West. The apparent distance from Sandy Islet to the moon ranged from 356,813 km on 30 January to 406,065 km on 12 February (Heavens above 2010). During that time the illuminated lunar disc decreased from 100% to 4% with non linear increases in SQM (mag/arcsec 2 ) values, as shown in Table 3 8. In the period from full moon (30 January 2010) to third quarter moon (7 February 2010) the high tide used by nesting turtles coincided with a waning moon. Nesting activity was unhindered by the bright moonlight illuminating the beach. From the third quarter to new moon (14 February 2010), the turtles started to come ashore on the early evening high tide. They nested under a moonless night sky as the beach was illuminated only by starlight with corresponding low luminance values (Figure 3 10). 40

Table 3 8: Sky quality meter (SQM) values (mag/arcsec 2 ) for the nights ashore at Sandy Islet. Provided are the dates and the percentage of the lunar disk illuminated 2 and the corresponding SQM values. The period for which the moon was visible during the night to dawn is shaded. The SQM values with asterisk (*) represent the hour segment in which tide peaked for that evening. (lunar data from http://aa.usno.navy.mil/cgi bin/aa_rstablew.pl). Date Lunar Disk Time % 1900 2000 2100 2200 2300 0000 0100 0200 0300 0400 30/01/2010 100 16.81 16.14* 14.65 14.63 31/01/2010 99 18 16.31 15.51* 15.23 14.94 1/02/2010 96 18.75 18* 16.1 15.6 15.42 15.99 2/02/2010 90 19.16 16.59* 16.14 15.81 15.75 3/02/2010 82 19.61 18.86 18.51* 17.88 16.19 4/02/2010** 72 ** 5/02/2010 62 20.4 18.75* 17.63 17.18 6/02/2010 51 20.83 20.62 18.24* 7/02/2010 41 20.44 19.47* 8/02/2010 32 17.03 21.67 20.33 9/02/2010 23 ** 21.74 21.73 21.73 10/02/2010 15 19.83 21.63* 21.67 21.65 21.68 21.66 11/02/2010 9 16.66 ** 12/02/2010 4 17.82 21.71 21.64 21.61* 21.59 ** no trip ashore due to boat difficulties 2 Considering the moon as a circular disk, the ratio of the area illuminated by direct sunlight to its total area is the fraction of the moon's surface illuminated, which multiplied by 100 represent the percent illuminated 41

A Luminance (cd/m2) 0.18 0.16 0.14 0.12 0.1 0.08 0.06 0.04 0.02 0 B 4 Height of Night Tide (m) 3.5 3 2.5 2 1.5 1 0.5 C Numbers of Tracks and Nests 30 25 20 15 10 5 0 29/01/2010 30/01/2010 31/01/2010 1/02/2010 2/02/2010 3/02/2010 4/02/2010 5/02/2010 6/02/2010 7/02/2010 8/02/2010 9/02/2010 10/02/2010 Date Nests Tracks Figure 3 10: The maximum luminance (A) and the height of the tide (B) during green turtle nesting activity with the respective numbers of tracks and successful nests (C) for each night of monitoring on Sandy Islet, Scott Reef. 42

3.6 Satellite Tracking 3.6.1 Transmitters The twelve satellite transmitters were still active when most recently downloaded, 31 March 2010 and all available data was post processed for accurate GPS data (Table 3 9). The difference in last transmission dates between the Sirtrack and SMRU tags is due to the way that SMRU tags keep the data that it has collected in internal buffers, which means that it can transmit data up to 3 months after the event. The intention is to smooth out variations in the opportunity to transmit data, caused by weather conditions or turtle behaviour. The GPS track will therefore gradually fill in over several weeks. The days at large range from 54 to 59 (mean = 52; Table 3 9). One transmitter, SN49683, was active but the data could not be processed to determine Fastloc locations. Sirtrack are currently trying to resolve this issue, however there is no data currently available from SN49683 for this report. If Sirtrack cannot find a solution, the ARGOS data will be used to analyse its inter nesting and migration/foraging behaviour in subsequent reports. 3.6.2 Inter nesting A total of 11 of 12 turtles returned to re nest at Sandy Islet, following the fitting of the tags to the last transmissions received. One turtle returned to re nest once and one turtle returned to re nest four times, the rest of the turtles re nested either two or three times (Table 3 10). Inter nesting intervals ranged from 8 to 22 days, with a mean of 10 days (Table 3 10). During the inter nesting period, turtles remained within 3 km of Sandy Islet, with the exception of two turtles that travelled into South Scott Reef Lagoon, over 12 km south (Figure 3 11; Figure 3 12; Figure 3 13). The majority of the location fixes were within an area south of Sandy Islet, where the water depth ranged from 5 to 10 m. However, the mean recorded maximum water depth was 35.3 m, which was attributed to the marked change in bathymetry, that is, within 1 km of Sandy Islet water depths drop in excess of 40 m. This data was supported by the SMRU tags which also recorded dive depths. The average dive depths were recorded in 6 hour blocks and ranged from 1 m to 24 m, while the average dive depths (based on the average of 6 hour blocks) ranged from 5.8 m to 9.6 m. The maximum dive depth ranged from 35 to 45 m. During the inter nesting period turtles spent the majority of their time in dives which accounted for 83.2% of their activities, while 15.2% was spent at the surface and only 1.6% was spent on the Sandy Islet. On average dive duration ranged from 15 to 25 minutes, and the maximum dive duration was 55 minutes (Table 3 11). Three turtles (SN50060, SN50066, and SN50070) still remain in the vicinity of Scott Reef as of March 31 st 2010. 43

Table 3 9: Processed Fastloc transmission information of the green turtles that were satellite tracked following nesting at Sandy Islet, Scott Reef. Turtle ID Date Released Last Transmission Days at Large # Fixes 1 Type of Satellite Tag SN49666 6/02/2010 31/03/2010 53 373 Sirtrack SN49669 1/02/2010 31/03/2010 58 298 Sirtrack SN49675 31/01/2010 29/03/2010 57 297 Sirtrack SN49676 3/02/2010 31/03/2010 56 290 Sirtrack SN49683 30/01/2010 Sirtrack SN49686 2/02/2010 20/03/2010 49 264 Sirtrack SN50060 3/02/2010 27/03/2010 52 156 SMRU SN50062 4/02/2010 11/03/2010 35 94 SMRU SN50065 2/02/2010 27/03/2010 53 134 SMRU SN50066 6/02/2010 24/03/2010 46 110 SMRU SN50070 2/02/2010 21/03/2010 47 88 SMRU SN50075 29/01/2010 21/03/2010 51 65 SMRU Mean (SE) 51 (4) 197 (1) 1 Number of locations received by the ARGOS satellites 44

Table 3 10: Inter nesting intervals of satellite tagged green turtles from Sandy Islet, including the maximum distance turtles moved away from the nesting beach during this period. Date of nesting is the nesting date and date returned is the first time after nesting that the turtle returned to Sandy Islet. Turtle ID Date of Nesting 49666 10/02/2010 22/02/2010 08/03/2010 18/03/2010 Date Returned 19/02/2010 03/03/2010 17/03/2010 Re nesting Interval (days) 9 9 9 Max km from beach Max Water Depth (m) 2.0 39 49669 01/02/2010 49675 31/01/2010 10/02/2010 20/02/2010 28/02/2010 09/02/2010 20/02/2010 28/02/2010 9 10 8 3.0 43 49676 04/02/2010 15/02/2010 26/02/2010 13/02/2010 23/02/2010 49683 Fastloc data not available 49686 01/02/2010 12/02/2010 23/02/2010 11/02/2010 23/02/2010 10 11 50060 02/02/2010 16/02/2010 11/03/2010 24/03/2010 50062 03/02/2010 13/02/2010 23/02/2010 50065 01/02/2010 12/02/2010 50066 05/02/2010 19/02/2010 04/03/2010 12/03/2010 24/03/2010 50070 13/02/2010 25/02/2010 10/03/2010 50075 29/01/2010 10/02/2010 24/02/2010 12/02/2010 10/03/2010 22/03/2010 12/02/2010 23/02/2010 9 8 10 22 11 9 10 2.7 37 13.6 56 12.7 43 2.4 15 10/02/2010 8 1.9 30 15/02/2010 28/02/2010 12/03/2010 23/03/2010 25/02/2010 08/03/2010 10/02/2010 24/02/2010 10 9 8 11 12 11 12 14 3.5 21 2.6 39 2.1 30 Mean (SE) 10.4 (0.6) 4.7 (1.4) 35.3 (3.7) 45

A B C D Figure 3 11: Inter nesting movements of four green turtles tracked from Sandy Islet, Scott Reef (using Fastloc data). A SN49666; B SN49669; C SN49675; D SN49676. 46

A B C D Figure 3 12: Inter nesting movements of four green turtles tracked from Sandy Islet, Scott Reef (using Fastloc data). A SN49686; B SN50060; C SN50062; D SN50065. 47

A B Figure 3 13: Inter nesting movements of three green turtles tracked from Sandy Islet, Scott Reef (using Fastloc data). A SN50066; B SN50070; C SN50075. C 48

Table 3 11: Dive characteristics of six inter nesting green turtles at Scott Reef Turtle ID Depth of dives (over a 6 hour block)(m) Range of depths Average depth of dives Max depth (m) Dive durations (over a 6 hour block) (hh:mm:ss) Range of dive durations average duration % dive time % surface time % out of water time 50060 1 20 8.1 40 00:00:20 00:23:00 81.5 16.3 2.2 00:55:00 50062 1 10 6.11 40 00:01:20 00:24:38 84.2 14.7 1.1 00:45:00 50065 1 24 5.8 35 00:01:20 00:15:16 80.8 16.7 2.5 00:30:00 50066 2 20 6.6 45 00:01:40 00:19:23 82.3 15.4 2.3 00:45:00 50070 1 20 7.5 40 00:03:00 00:25:52 82.9 17.1 0 00:45:00 50075 4 24 9.6 35 00:05:00 00:19:46 87.2 11.3 1.5 00:40:00 Mean 39 83.2 15.3 1.6 Note: the data presented only represents a proportion of the data recorded due to the limitations in satellite transmittal time and capacity. 3.6.3 Migration and Foraging Grounds Eight out of the 11 turtles had commenced migration prior to 31 March 2010. Following their final nesting, turtles quickly left Scott Reef and followed one of two routes: one to the South and one to the north east. Two individuals moved south to the Kimberley coast near the Lacepede Islands and Cape Leveque. One individual (SN50075) remained close to the coast while the track for the other (SN49686) may have been similar but was not recorded. Both turtles passed into the Eighty Mile Beach area (Figure 3 14). By the end of the recorded period (31 March 2010) one turtle (SN49686) was close to the mouth of the De Grey River in the Pilbara. The turtles that tracked north east headed towards the Bonaparte Archipelago on the Kimberley coast. Again, there was a preference for swimming close to the coast, although SN50062 preferred to remain some distance from the coast but still moved towards the east. Five tracks extended into the Joseph Bonaparte Gulf and four crossed to the Northern Territory reaching the Australian mainland coast between Docherty Island near Port Keats and Point Blaze at the southern end of Fog Bay. The path of SN49676 did not continue past this point as of 31 March 2010. The paths of the other three continued along the chain of islands and reefs past Grose Island and into Beagle Gulf (Figure 3 15). Three turtles passed either through Beagle Gulf or around Bathurst and Melville islands to pass Cape Don at Cobourg Peninsula to continue their eastward migration. They continued to move eastwards remaining close to the coast. Turtle SN49669 had reached Elcho Island by 31 March 2010 and had averaged 1.86 km/h for the last 56 days. Turtle migration to date has ranged from 255 km to 1378 km and only one turtle (SN50075) may have reached its foraging grounds as it has remained off Eighty Mile Beach for four days (Figure 3 15). To date, turtles SN49669 and SN50065 have travelled over 1000 km towards the Gulf of Carpentaria. 49

Average migration time is currently 28 ± 6 days; however, the majority of the marine turtles had not finished their migration and therefore migration times are likely to increase. On average, turtles travelled 1.99 ± 0.13 km/hr. SN49666 had the highest average of 2.38 km/hr, however, this turtle had the shortest migration displacement to date and therefore its average speed is likely to change (Table 3 12). Only three of the turtles with SMRU tags had begun their migration from 31 March 2010 whilst the other three turtles with SMRU tags remained at Scott Reef (Figure 3 13). During the migration period turtles spent the majority of their time in dives (72%), while 28% of time was spent at the surface. SN50062 did spend some time out of the water. This is assumed to be basking time as it was during the middle of the day and lasted 37 minutes (Figure 3 13) Table 3 12 Summary of migration of green turtles tracked following nesting at Sandy Islet Turtle ID Migration Days Average Migration Speed (km/h) Max Migration(km) 49666 3 2.38 (0.26) 255 49669 56 1.86 (0.06) 1378 49675 31 2.30 (0.09) 1142 49676 30 1.94 (0.10) 966 49686 24 2.00 (0.09) 700 50062 15 2.24 (0.14) 850 50065 43 1.97 (0.10) 1321 50075 23 1.19 (0.20) 625 Mean 28 (6) 1.99 (0.13) 905 (134) Table 3 13: Dive characteristics of three green turtles while migrating to their foraging grounds Turtle ID Depth of dives (over a 6 hour block)(m) Range of depths Average depth of dives Max depth (m) Dive durations (over a 6 hour block) (hh:mm:ss) Range of dive durations average duration % dive time % surface time % out of water time 50062 2 45 16.7 80 00:01:20 0:14:40 52.9 46.6 0.5 00:30:00 50065 3 40 10.2 45 0:03:00 0:09:52 72.5 27.5 0.0 0:28:00 50075 10 26 17.1 35 0:06:00 0:19:45 90.6 9.4 0.0 0:40:00 Mean 53 72 28 0 Note: the data presented only represents a proportion of the data recorded due to the limitations in satellite transmittal time and capacity. 50

Figure 3 14: Migration route of SN49666, SN49669, SN49675, SN49676, SN50062 and SN50065 following nesting (Foraging locations from Pendoley 2005). 51

Figure 3 15: Migration route of SN49686 and SN50075 following nesting (Foraging locations from Pendoley 2005). 52