The Influence of Gas Flares on the Orientation of Green Turtle Hatchlings at Thevenard Island, Western Australia

 Kellie Pendoley

 Abstract

Thevenard Island, situated 10 nm off the coast of north western Australia, is a known nesting site for Green turtles (Chelonia mydas), and also supports an oil production facility. A condition of the Government approval to develop the facility was to minimize light impact on sea turtles. The original gas flare was purpose-built to shield the flame from nearby nesting beaches. A second pit flare was installed in the early 1990 sfor short term use while the primary shieldedflare was undergoing maintenance. Post commissioning surveys and routine inspections indicated that both the flares and the facility lights were potential sources of impact on the sea finding success of C. mydas hatchlings. In September 1995, the spectral characteristics of the two flares were measured, and results suggested the flares emitted light in a spectral range outside of that visibk to C. mydas. Arena experiments were subsequently carried out to determine ~/ the light sources were disorienting hatchlings emerging in the vicinity of the ft ares and over what distance the influence might extend The results suggested that the flares caused disorientation of hatchlings during nights of new moon, however this impact was reduced with distance from the source and as the moon phase progressed towards full moon.

Introduction

Government approval to develop an oil and gas processing facility on Thevenard Island, 20 nm off the coast of north west Western Australia (Fig. 1), was contingent on the ability to minimise impacts on sea turtles breeding on the island. Specifically, the company was required to shield a proposed gas flare. A state-of-the-art flare tower was installed, but ongoing operational problems resulted in the construction of a backup pit flare adjacent to the flare tower.

Subsequent company and Government environmental inspections indicated the light from both the shielded flare and the pit flare had the potential to impact hatchling sea turtles. Flames were generally not visible from either flare at beach level, but both flares produced a highly visible sky glow. A two-stage study to characterise the flares and to test the impact of the flares on hatchling sea turtles was carried out between September 1995 and March 1996.

The objectives of the study were to determine whether there was a spectral difference between the tower and the pit flare, whether the spectral differences altered by changing gas flow rates, the magnitude of luminous loss with distance away from the flares, and what other sources of illumination occurred on the island. At the same time we investigated if the flares affected hatching orientation, over what distance, and did any other light sources on the island attract hatchlings.

Fig. 1: North westhelf location map.

Methods and Materials

Spectral characteristics of flares - A study of the spectral characteristics of the major illumination sources was undertaken by Peter Hick of the Commonwealth Scientific and Industrial Research Organization (CSIRO) using an Ocean Optics miniature fibreoptic spectrometer. This was coupled to a portable computer for real-time display and recording of spectra. Light readings were taken from the two flares, plant lighting, jetty lights and the fishing shacks along the southern beach. The results were disseminated in an unpublished report to WAPET (Hick 1995) and are summarised herein.

Hatchling orientation experiments - Hatching orientation experiments followed the methods used by Mrosovsky (1968, 1975). Circular orientation arenas were set up at four beach locations on Thevenard Island (Table I). Trials were run two weeks apart, under full moon and new moon conditions. The arena locations we chosen to be representative of possible views hatchlings would have when they emerged from their nests (Table II). Site 1 (Flares) was adjacent to the flares (30 m from the base of the pit flare), Site 2 (Mackerel Islands) was adjacent to the Mackerel Islands fishing WAPET camps, Site 3 (Crest Road) was 300 m west of the tower flare, and Site 4 (Crest Lease), which was used as a control site in March 1996, was 1.5 km west of the tower flare (Fig. 2). Arenas were 13 m in diameter and positioned on or just below the dune vegetation line. The arena was divided into 16 equal segments using wooden dividers.

Fig.2 : Study sites at Thevenard Island

Table I: Basic trial conditions

Table II: Test site light conditions

Hatchlings were removed from suitable nests on the day of each experimental run. Nests were identified by a saucer-shaped depression on the sand surface, or from hatchling tracks across the sand. Hatchlings were stored in a dark air- conditioned room until the evening, and were used in the trials within 12 hours of collection. In February, hatchlings were all from the same clutch, in March animals came from three different clutches. Hatchlings were placed in the centre of the arena under a cover that was lifted remotely from outside the arena. After removing the cover, hatchlings were given 10 minutes to reach the edge of the arena, after which the sector in which they were found was recoded.

Results and Discussion

Was there a spectral difference between the tower and pit flares? - There was no significant spectral difference between the two flares (Fig. 3). However, the intensity of the light above the tower flare appeared to be marginally higher than the pit for most wavelengths. This could be attributed to the design of the tower flare which both elevates the light source high into the sky and concentrates the light into a small area.

Did changing gas flow rates alter spectral differences? - Flow rates did not appear to change the spectra of the light above either flare. Furthermore, the amount of measured illumination did not increase, supporting anecdotal evidence from gas plant operators who had not observed any increase in sky glow when gas flow rates were increased. Verheijen (1985) concluded the major light problem associated with gas flares was glowing soot. On Thevenard Island the combustion process within each flare had been optimised to burn cleanly. It was believed that naturally occurring particles (marine aerosols) in the air, which scatter the light, were the limiting factor in this study. Varying the flow rates would therefore have no effect on the amount of light scattered.

Fig 3: Spectral signature of tower and pit flares.

Did magnitude of luminous decrease further away from the flares? - Light emissions from the atmosphere above the flares were very low and could not be reliably measured at distances greater than 200-300 m with the configuration of the instrument.

What other sources of illumination occurred on the island? - A number of light sources around Thevenard island were measured during this study, including white lights on the jetty, tennis court lights, a small fluorescent light within the plant, and a street light reflected from the side of a crude oil storage tank (Figs. 4-7). These lights were characterised by strong spectral peaks in the 400-660 nm range favoured by green turtles (Witherington 1991, Granda & O'Shea 1972). The illumination from the moon was also measured and compared to the street light reflecting from the nearby crude storage tanks (Fig. 8). Besides the flares, this reflectance was considered the most likely source of light that could cause disorientation of hatchlings, as the spectra from the full moon overwhelmed that of the reflected street light.

Fig. 4: Spectral signature of small white (pink) light on jetty

 Fig.5: Spectral signature of tennis court light

 Fig.6: Spectral signature of fluorescent light on electricity shed

 Fig.7: Spectral signature of light on the tower wall

 Fig.8: Spectral signature of the moon and the fllodlight tank side adjacent to the pit flare

Did the flares affect hatchling orientation? - Trials were conducted under moon and no moon conditions on the beach 30-100 m from the flares (Fig. 9a,b). On the night with no moon the hatchlings oriented between the ocean and the flare, with the mean orientation vector aligned more closely with the flares than the ocean. Mooring lights from a tanker anchored 17.5 km north of the island were not considered a sufficient cause for disorientation since they represented a small point source of light situated low on the horizon. To the naked eye they appeared to contribute nothing to the light field on the beach, and were not believed to have caused any undue attraction to segment 13. Hatchlings are thought to integrate light cues from a vertically low but horizontally broad field of view (Lohmann et al. 1997), and the amount of light (irradiance) reaching a turtle's 'cone of acceptance' was more important in a hatchling's assessment of the light field than the amount light emanating from the source (radiance).

The increase in ambient light over the ocean on the moonlit night appeared to cause a seaward shift in hatchling orientation direction. The light produced by the full moon appeared to override the influence of the flares with 80 % of hatchlings orienting towards the three most seaward segments. Similar results have also been reported by Salmon (1995), Mrosovsky (1966) and Irwin (1996).

Whilst the light spectra indicated that neither flare produced significant spectral peaks in the 400-600 nm visual range of green turtles, these results suggest that even the low levels of light emitted by the flares in this range do have the potential to disorient hatchlings, particularly on nights with no moonlight.

Over what distance could an influence be observed? - Trials were conducted under moon and no moon conditions at sites located 300 m from the pit flare and 400 m away from the pit flare (Fig. 9c,d). On the night with no moon the majority of the hatchlings oriented towards the seaward segments, with a single animal attracted to the tower flare. On the night of the full moon most hatchlings oriented in an eastward direction away, from the flares and almost parallel to the ocean. Control trials on the same night (Fig. 9f) also resulted in a mean vector east of a direct path to the ocean.

Fig.9 Hatchling orientation from trials in circular arenas

Do other light sources disorient hatchlings? - Trials were conducted under new moon conditions in the vicinity of vessels moored approximately 300 m offshore, a jetty and lights from tourist accommodation (Fig. 9e). Hatchlings appeared to orient between the vessels and the jetty lights, which were 500 m apart. The street light over the camp and lights from the tourist accommodation did not appear to affect the hatchlings. It is possible the anisotrophic effect of abnormally high light intensity from the jetty and moored vessels, relative to the tourist accommodation and street light, caused an the hatchlings to be blinded to both the secondary light sources and the ocean (see also Verheijen 1985, Salmon 1995).

Conclusions

• There was no significant spectral difference between the two flares;
• Flow rates did not appear to change the spectra in the atmosphere above the flares;
• The contribution of illumination sources other than the flares was significantly greater than the flares;
• Effects of the facility lights were insignificant when compared with the full moon;
• The flares disoriented hatchlings on nights near the new moon but had less of an influence on moonlit nights;
• The flares did not attract hatchlings from beaches 300 - 400 m away on either full or new moon nights;
• Other light sources on the island and the nearby jetty and moored vessels also disoriented hatchlings.

References

Granda, A.M. & P.J. O'Shea, 1972. Spectral sensitivity of the green turtle (Chelonia mydas) determined by electrical responses to heterochromatic light. Brain Behav. Evolution, 5, 143-154.

Hick, P.. 1995. Spectral measurement of illumination sources at Thevenard Island: a preliminary study of the probable effects of gas flares and oil facility lights on green turtles and a subsequent revisit to measure a range of gas-flow rates. CSIRO unpu. report.

Irwin, M.S., B.J. Godley & A.C. Broderick, 1996. The effect of anthropogenic lighting on marine turtles in Northern Cyprus, In Procs. of the Sixteenth Annual Symposium on Sea Turtle Biology and Conservation.

Lohmann, K.J., B.E. Witherington, C.M.F. Lohmann & M. Salmon, 1997. In The Biology of Sea Turtles (Lutz, P.L. & J.A. Musick, eds.) CRC Press, Boca Raton: 107-135.

Mrosovsky, N. & A. Carr, 1966. Preference for light of short wavelengths in hatchling sea turtles, Chelonia mydas, tested on their natural nesting beaches Behaviour 28 217-231.

Mrosovsky, N. & S.J. Shettleworth, 1968. Wavelength preference and brightness cues in the water finding behaviour of sea turtles. Behaviour 32: 211-257.

Mrosovsky, N. & S.J.S. Worth, 1975. On the orientation circle of the leatherback turtle, Dermochelys coriaeea. Animal Behaviour 23: 568-591.

Salmon, M. & B.E. Witherington, 1995. Artificial lighting and seafinding by loggerhead hatchlings: evidence for lunar modulation. Copeia 4: 931-938.

Verheijen, F.J., 1985. Photopollution: Artificial light optic spatial control systems fail to cope with: incidents, causations, remedies. Experimental Biology 44:1-18.

Witherington, B.E. & K.A. Bjorndal, 1991. Influence of artificial lighting on the seaward orientation of hatchling loggerhead turtles Caretta caretta. Biological Conservation 55: 139-149.