seaturtle.org : MTN : ARCHIVES : Sign In

Despite being listed as endangered or critically endangered by the IUCN, marine turtle species are included only in the Brazilian (IBAMA 2005) and Canadian (DFO 2005) mitigation guidelines and there has been no airgun-related research on free-ranging turtles. Controlled exposure experiments on captive turtles found an increase in swim speed and erratic behaviour indicative of avoidance, at received airgun sound levels of 166–176 dB re 1 μPa (rms) (O’Hara & Wilcox 1990; McCauley et al. 2000). The limited available data on marine turtle hearing suggest highest auditory sensitivity at frequencies of 250–700 Hz, and some sensitivity to frequencies at least as low as 60 Hz (Ridgway et al. 1969; O’Hara & Wilcox, 1990; Moein-Bartol et al. 1999), overlapping with the higher frequencies produced by airguns.

This paper describes observations of marine turtles during a ten-month seismic survey off Angola on the west coast of Africa, with discussion of turtle responses to airgun sound and recommendations for future surveys. The study area and data collection methods are described by Weir et al. (In press). In summary, two consecutive 3D seismic surveys were carried out by BP Exploration (Angola) Ltd and partners over a 288-day period between 1 August 2004 and 15 May 2005 in a deep-water (1000–3000m) area off northern Angola (5–11°S latitude and 9–13°E longitude). Two airgun arrays fired alternately at approximately 10 sec intervals. Each array comprised 24 Bolt airguns of 30–290 cu. in., producing total volume of either 5085 cu. in. or 3147 cu. in. (Table 1).

Table 1. Source parameters utilised during each survey.

Concurrently with marine mammal observations, a single observer located on the ship’s helideck (18 m eye height) searched for turtles 360° around the vessel with the naked eye and with 10x42 binoculars. A total of 2769 h was spent on-effort during daylight hours, of which 676.4 h occurred in Beaufort sea state ≤2. Effort logs (position, water depth and environmental data including Beaufort sea state) were recorded for every watch. Airgun status was defined as full-array, partial-array (soft start or test firing) or guns-off.

Two hundred turtle sightings (240 animals) were recorded, including 33 olive ridley Lepidochelys olivacea, three leatherback Dermochelys coriacea, four loggerhead Caretta caretta and 160 unidentified turtle (hard-shell species) sightings. For each sighting the position, species, number, distance from the airguns (using a range-finding stick based on Heinemann 1981), behaviour, and airgun status were recorded. Behaviour was recorded initially (at the time of first sighting), and subsequently (noting any responses) using descriptive categories including basking (lying motionless at surface with carapace exposed), swimming (surface or subsurface), mating, breathing and diving (normal or ‘startle’ dive). Where apparent responsive behaviour was observed, it was noted where this occurred in relation to the vessel and towed equipment (i.e. bow, side, stern or astern of the vessel).

The UK marine mammal mitigation guidelines (which do not include turtles) (JNCC 2004) were in use throughout. BP voluntarily included marine turtles in the mitigation measures, such that the airguns were not allowed to start-up for 20 min if a turtle had been observed within 500 m of the airgun array during the 30-min pre-shoot watch.

Turtle data were collected concurrently with marine mammal surveys and were not gathered solely to assess turtle responses to airgun sound. Data were therefore analysed using broad criteria including sighting rate (turtles/h), distance and behaviour, which were compared in different airgun status categories using a combined dataset of all turtle species. Only data collected in Beaufort sea state ≤2 were analysed to ensure a reasonable likelihood of turtle detection. Since detection rate decreases with increasing distance from the vessel, distance categories rather than absolute distances were used for analysis.

Table 2. Survey effort and turtle sighting rate (all species combined) according to airgun status (Beaufort sea state ≤2 only).

Out of 639 airgun uses, airgun start-up was delayed on five instances due to sightings of marine turtles within the 500 m exclusion zone around the airguns. Three occasions involved single sightings of olive ridley turtles and incurred only short delays. However, delays of over one hour were caused by numerous basking turtles encountered on 10 and 11 August 2004. These lengthy delays resulted from a combination of factors including: (1) peak offshore turtle abundance during August (Weir et al. In press); (2) glassy (Beaufort sea state 0) seas from 8–12 August 2004; and (3) time of day. There was a significant difference in the frequency of turtle numbers observed according to Beaufort sea state (X²=5830.4, d.f.=4, P=<0.001), with significantly more turtles than expected observed in Beaufort sea states 0 and 1, and fewer in sea states 2 to 4. Glassy sea conditions both increased turtle detection and also appeared to trigger the onset of prolonged basking at the surface (further increasing detection) (Figure 1). Most turtles were observed between 10:00 and 14:00 h (UTC) (95% of which were basking) during the hottest part of the day (Figure 2).

Figure 1. Initial behaviour of turtles according to Beaufort sea state, showing number of turtles and survey effort (hr).

The turtle sighting rate during guns-off (0.43 turtles/h) was double that of full-array seismic (0.20/h) (Table 2). Although suggestive of avoidance of airguns by turtles these data should be treated with caution since 118 of the sightings occurred from 8–12 August 2004 during unusually calm conditions. On each of these dates the airguns were inactive between 11:00 to 14:00 h during peak diurnal abundance of turtles (Figure 2), and it is likely that this biased the analysis. Nevertheless, glassy seas also occurred during the midday period on 25 August during full-array seismic, with only seven turtles recorded.

Figure 2. Diurnal trend in turtle relative abundance (Beaufort sea state ≤2) (solid bars), and mean air temperature in Luanda, 8–12 August 2004.

There was indication that turtles occurred closer to the source during guns-off than full-array, with double the sighting rate during guns-off in all distance bands within 1000 m of the array (Figure 3). However, there was no significant difference in the median distance of turtle sightings from the airguns during full-array (mean= 779 m, SD=464, N=57) or guns-off (mean= 743 m, SD=449, N=112) (Mann-Whitney U = 3035, N= 169, P = 0.6). While this result apparently indicates a lack of movement away from active airguns, it is possible that turtles only detect airguns at close range or are not sufficiently mobile to move away from approaching airgun arrays (particularly if basking for metabolic purposes when they may be slow to react).

It was difficult to assess turtle behaviour in the field since animals were usually either distant (61% were ≥500 m away) or seen briefly. Both initial and subsequent behaviour were recorded for 180 turtles. Basking comprised the predominant initial behaviour during both full-array seismic (94%) and guns-off (96%). Although a slightly higher proportion of turtles dived during full-array seismic (12.5%) than guns-off (11%), in most cases (full-array=77% ; guns-off=83%) turtles continued to bask at the surface as the vessel and towed equipment moved past them and remained visible at the surface well astern of the ship. Apparent responsive dives were noted for 20 turtles, six during full-array seismic and 14 during guns-off. Thirteen turtles dived in apparent response to the vessel, nine of which startle dived at the bow (full-array=2; guns-off=7). Seven turtles startle dived in apparent response to seismic equipment, including six in response to towed surface floats (full-array=1; guns-off=5) and one in apparent response to the inactive airgun array. An assessment of turtle behaviour in relation to seismic status was therefore hindered by apparent reaction of individuals to the ship and towed equipment rather than specifically to airgun sound. These reactions occurred at close range (usually <10 m) to approaching objects and appeared to be based principally on visual detection.

Collisions between turtles and vessels are clearly not limited to seismic ships, which are slow (survey speed of 4–5 knots) compared with other vessel types. However, the large amount of equipment towed astern of seismic ships does increase the potential for collision. Basking turtles were particularly slow to react; for example one animal was washed away in the bow-wave while others had ‘near misses’ with towed surface floats. While little can be done to avoid outright collision, turtles can also become entrapped within some seismic equipment leading to suffocation. For example, during seismic surveys off West Africa in 2003, turtles became fatally entrapped within gaps in the tail-buoys (seismic personnel, pers. comms.). Modifying equipment (e.g. with ‘turtle guard’ bars placed over such gaps to exclude turtles) can prevent these scenarios and should be implemented on all seismic vessels operating in turtle-inhabited areas.

It was not possible to draw conclusions on the impact upon turtles of seismic airgun sound during this study. There was some indication that fewer turtles were seen during full-array seismic, although there was no obvious behavioural avoidance (e.g. swimming away) of the airgun array. However, basking turtles may not be able to move rapidly away from approaching airguns even if motivated to do so, since their responses to approaching objects are slow. The main limitations in assessing the reaction of turtles to airgun sound were: (a) difficulties in detecting animals in Beaufort sea states >1; (b) difficulties in determining at-sea turtle behaviour (notably for subsurface animals); and (c) difficulties in distinguishing responses to airgun sound from responses to the vessel and towed equipment. The data indicate that visual detection of turtles will not be effective in Beaufort sea states >1, with implications for mitigation during seismic surveys. Turtles were only observable at the surface (where received sound levels are lower due to the ‘Lloyd mirror’ effect (Urick 1983)), and it could not be ascertained whether subsurface turtles reacted to airgun sound. Most responsive dives occurred in clear reaction to the ship or towed equipment rather than airguns, and since the airguns were located at least 300 m astern of the ship’s bow, turtles were more likely to initially encounter, and respond to, the ship than the airgun array.

The use of measures to avoid turtle entanglement in seismic apparatus should become mandatory throughout industry. Future surveys should aim to collect detailed turtle behavioural information, and investigate the reaction of different turtle species in varying water depths and with different airgun arrays. Controlled experiments on turtle responses to airgun sound are also required, together with detailed studies of turtle hearing and the acoustic properties of airgun arrays.

Acknowledgements: This work was carried out under the sponsorship of BP Exploration (Angola) Ltd and their partners in Block 31. I am grateful to Nathan Gricks for collecting some of the field data, and to the crews of the Geco Triton for their hospitality. Richard Seaborne & Bill Streever provided useful comments on a draft, and Nick Rose-Innes provided source information. Comments from two anonymous reviewers improved the paper

Department of Fisheries and Oceans, Canada (DFO). 2005. Statement of Canadian Practice: Mitigation of Seismic Noise in the Marine Environment. April 2005.

ENGÅS, A., S. LØKKEBORG, E. ONA & A.V. SOLDAL. 1996. Effects of seismic shooting on local abundance and catch rates of cod (Gadus morhua) and haddock (Melanogrammus aeglefinus). Canadian Journal of Fisheries and Aquatic Sciences 53: 2238-2249.

IBAMA. 2005. Brazilian Environmental Licensing Guide. Atividades de Sísmica Marítima na Costa Brasileira. Viewed May 2006 online at <http://www.anp.gov.br/ibamasismica/>

GORDON, J., D.GILLESPIE, J. POTTER, A. FRANTZIS, M.P. SIMMONDS, R. SWIFT & D. THOMPSON. 2004. A review of the effects of seismic surveys on marine mammals. Marine Technology Society Journal 37: 16-34.

GULLAND, J.A. & C.D.T. WALKER. 2001. Marine seismic overview. In: M.L. Tasker & C.R. Weir (Eds). Proceedings of the Seismic and Marine Mammals Workshop, London, 23-25 June 1998.

HEINEMAN, D. 1981. A range finder for pelagic bird censusing. Journal of Wildlife Management 45: 489-493.

JNCC. 2004. Guidelines for minimising acoustic disturbance to marine mammals from seismic surveys. Joint Nature Conservation Committee, Peterborough, UK.

MCCAULEY , R.D., J. FEWTREL , A.J. DUNCAN , C. JENER, M.N. JENER, J.D. PENROSE, R.I.T. PRINCE, A. ADHITYA , J. MURDOCH & K. MCCABE. 2000. Marine seismic surveys: analysis of airgun signals, and effects of airgun exposure on humpback whales, sea turtles, fishes and squid. Report to APPEA by the Centre for Marine Science and Technology, Curtin University of Technology, Australia.

MOEIN-BARTOL, S., J.A. MUSICK & M.L. LENHARDT. 1999. Auditory evoked potentials of the loggerhead sea turtle (Caretta caretta). Copeia 1999: 836-840.

O’HARA, J. & J.R. WILCOX. 1990. Avoidance responses of loggerhead turtles, Caretta caretta, to low frequency sound. Copeia 1990: 564-567.

RIDGWAY, S.H., E.G. WEVER, J.G. MCCORMICK, J. PALIN & J.H. ANDERSON. 1969. Hearing in the giant sea turtle, Chelonia mydas. Proceedings of the National Academy of Sciences USA 64: 884-890.

URICK, R.J. 1983. Principles of underwater sound. McGraw-Hill, New York. 423pp.

WEIR, C.R. & S.J. DOLMAN. In press. Comparative review of the regional marine mammal mitigation guidelines implemented during industrial seismic surveys, and guidance towards a worldwide standard. Journal of International Wildlife Law and Policy.

WEIR, C.R., T. RON, M. MORAIS. & A.D.C. DUARTE . In press. Nesting and pelagic distribution of marine turtles in Angola, West Africa, 2000–2006: occurrence, threats and conservation implications. Oryx.