Methods

The nesting season at Jamursba-Medi, Papua, extends from about April to September, peaking in July (Hitipeuw et al. 2007). To capture post-nesting movements, transmitters were deployed during the latter part of the nesting peak, between 20 and 25 July 2003. Transmitters were attached to nesting females using a flexible harness constructed of soft nylon webbing with flexible polyvinyl tubing over the shoulder straps (following methods of Eckert and Eckert 1997). The design included a corrodible pin to allow release of the harness after about 18–24 months. Satellite-linked transmitters (Sirtrack ‘Kiwisat 101') were programmed to transmit geographic location to the Argos system on a duty cycle of 6 hours on and 19 hours off. Minimum distances traveled by tracked leatherback turtles were estimated based on great circle distances between subsequent positions along track. Only the best position for each day with an Argos location quality code of ‘LC0' or better (between 150 m and 1 km accuracy) was included in the tracks.

Results

Deployment durations ranged from 111 to 695 days (Table 1), and individual turtles traveled an estimated 4784 to 20,558 km. Migration routes varied among individuals (Figs. 1, 2), but most animals either traveled northeastward across the tropical Pacific towards temperate waters of the eastern North Pacific (n = 5), or westward through the Sulawesi and Sulu Seas (n = 2) and into the South China Sea (n = 1). One turtle moved northward to the Sea of Japan. The speed of migration through the equatorial Pacific appeared related to the turtle's track relative to equatorial currents. Two individuals (05397 and 27957) made rapid eastward progress within the eastward-flowing Equatorial Counter Current. In contrast, 3 others exhibited slower eastward progress within the westward-flowing North Equatorial Current. Thus, the turtles did not exhibit any consistent association with major equatorial currents. Of the 5 eastward-migrating turtles, 1 (27957) completed a trans-Pacific migration to foraging areas located about 50–100 km off the coasts of Oregon and Washington, USA. This individual spent about 62 days in this region before turning southwest during late October 2004 towards tropical waters southeast of Hawaii. In March 2005, about 6 weeks before transmissions ceased, the turtle turned back northeastward towards the coast of North America, suggesting this individual spent 2 consecutive years foraging in the temperate eastern Pacific following nesting at Jamursba-Medi in 2003.

Table 1. Details for 9 transmitters deployed on leatherback turtles (Dermochelys coriacea) at Jamursba-Medi, Indonesia, during 2003 (*movement to Balabac Island only). CCL = Curved carapace length, CCW = curved carapace width, n/a = not available.

View Fullscreen

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

The 3 westward-migrating turtles (Fig. 2) traveled rapidly to coastal waters of the Philippines and Malaysia, where they spent prolonged periods of time (at least 6–8 months) in productive nearshore waters, suggesting they may have been foraging. The turtle with the longest transmission record (27960) may have also foraged for 2 consecutive seasons, however, evidence suggests that it might have been captured at sea sometime after February 2004. At that time, the turtle was located near Balabac Island, Philippines, after traveling 10,352 km from the nesting beach. Following a one-month lapse, the subsequent transmission indicated it was 887 km across the South China Sea near Vietnam. From that point, it traveled quickly (> 3 km/hr) across the South China Sea to an area between Kepulauan Anambas, Indonesia, and the Malaysian Peninsula by June 2004. Transmissions ceased again until October 2004 when patterns became suspicious, with an increase in the quantity and quality of transmissions and prolonged periods (10 days or more) near seaports in Thailand and Terengganu, Malaysia, and nearby Malaysian islands, followed by rapid movements (> 6.5 km/hr) between these locations. This pattern suggests that the turtle may have been captured and the transmitter retained aboard a vessel. Between January 2005 and the final transmission in June 2005, positions on land were more frequent, including a two-month period near Kampong Cherang Ruku, Malaysia, and 10 days near Pattani, Thailand's second largest fishing port.

The ninth turtle, which moved northward through the East China Sea and into the Sea of Japan, remained in these productive coastal waters for about 2 months before it moved into deeper waters of the northern Philippine Sea. Although this also may suggest foraging behavior, the record is less clear. The transmitter for this turtle ceased transmissions after about 195 days.

Discussion

The movement patterns of leatherback turtles from Jamursba-Medi we report here provide the first record of a trans-Pacific migration by a leatherback. The track between New Guinea and shelf waters off Oregon, USA, may represent the longest known migration between breeding and foraging areas of any marine vertebrate. The direct linkage between nesting sites in the western Pacific and foraging grounds in the northeastern Pacific confirms previous conclusions regarding stock structure by Dutton et al. (2000) based on genetic analysis. Migration routes were varied and displayed use of temperate and tropical foraging regions, but were markedly different from post-nesting migrations of leatherback turtles from Kamiali, Papua New Guinea (Benson et al. 2007b). Both rookeries are part of the same genetic stock, which likely represents a regional metapopulation (Dutton 2006; Dutton et al. 2007). Combined results of post-nesting leatherback turtles from Jamursba-Medi and Kamiali indicate that the western Pacific metapopulation uses multiple foraging areas, similar to Atlantic leatherback turtles (Ferraroli et al. 2004; Hays et al. 2004; Eckert 2006) but contrasting with the relatively uniform pattern of eastern Pacific leatherback turtles, which have only been documented to migrate southward post-nesting (Eckert and Sarti 1997; Dutton 2006).

The causes of variability in migration patterns of western Pacific leatherbacks are unknown, however, the time of year of nesting appears to plays a role in determining whether leatherbacks that forage in temperate waters migrate into the Northern or Southern Hemisphere. Peak nesting of leatherbacks at Kamiali occurs during the austral summer (late December), and post-nesting animals moved southeastward (Benson et al. 2007b) towards temperate waters of the Southern Hemisphere. In contrast, peak nesting at Jamursba-Medi occurs during the boreal summer (July) (Hitipeuw et al. 2007), and the leatherbacks tracked in this study either remained in tropical waters or moved into temperate waters of the Northern Hemisphere. The turtle that traveled across the Pacific to waters off Oregon arrived during August, when aggregations of large jellyfish are common (Shenker 1984). Thus, migratory destinations might be related to the timing of prey availability in different regions. Further analyses of the tracks in relation to oceanographic processes and prey life history will be required to elucidate these linkages for western Pacific leatherback turtles.

The inference of likely foraging areas in this study was based on track characteristics (i.e., transmissions within a relatively small geographic area for periods of weeks or months) and independent studies of jellyfish occurrence along the US West Coast (Shenker 1984; Graham et al. 2001; Suchman and Brodeur 2005) and in waters of the Philippines and Malaysia (Omori and Nakano 2001). Foraging of leatherbacks in areas of jellyfish aggregations has been well documented in the Pacific and Atlantic Oceans (e.g., Starbird et al. 1993; James and Hermann 2001, Ferraroli et al. 2004; Houghton et al. 2006; Benson et al. 2007a). Further work is underway to analyze fine-scale dive behavior of foraging leatherbacks in conjunction with habitat characteristics (Benson et al. 2006). Along the Pacific coast of North America, previous studies have also documented the presence of foraging leatherbacks during the summer and fall months, when dense aggregations of jellyfish are present (Bowlby 1994; Starbird et al. 1993; Benson et al. 2007a).

Migrations of this magnitude expose animals to a multitude of risks from fisheries on the high seas and in territorial waters of many countries. Leatherback turtles are susceptible to incidental take in pelagic long-line and gillnet fisheries that operate throughout much of the Pacific Ocean (Chan et al. 1988; Eckert and Sarti 1997; Lewison et al. 2004; Carretta et al. 2005; Alfaro-Shigueto et al. 2007). The extent to which small-scale coastal fisheries and high seas commercial fisheries operating around the Philippines and South China Sea interact with leatherbacks remains unknown, but our findings suggest a possible interaction with a vessel in the South China Sea. Incidental take of leatherbacks off the coast of Malaysia is likely one of the factors preventing the recovery of the near-extinct nesting population at Terengganu (Chan et al. 1988). Our track of a post-nesting leatherback from Jamursba-Medi into waters off Terengganu raises an intriguing possibility that turtles from Jamursba-Medi may frequent waters close to the former Malaysian rookery (Chan and Liew 1996). The potential for movement between Jamursba-Medi and the Malaysian rookeries has conservation implications and warrants further investigation.

Our findings illustrate the importance of international cooperation in addressing leatherback turtle conservation in the marine environment for this regional metapopulation. This study represents a crucial first step towards recognizing and mitigating adverse impacts at sea for western Pacific leatherbacks. Given the diversity of tracks observed in this study, however, a greater sample size will be required to evaluate the importance of different foraging areas and thoroughly assess potential anthropogenic threats.