Student Intern Investigates Oceanic Habitat of Loggerhead Sea Turtles in the South Pacific

Alice Ren
Alice Ren

Alice Ren, one of the Center's three student interns during the summer of 2010, has completed her research project on sea turtle biology under the guidance of Donald Kobayashi, a scientist in the PIFSC Ecosystems and Oceanography Division (EOD). Alice is a senior at Duke University with interests in physical oceanography and animal migration. In mentoring Alice, Don introduced her to the fascinating topic of sea turtle migration, and in particular, explained the method he had used to estimate the oceanic habitat of North Pacific loggerhead turtles. This approach involves analyzing data on the location and movements of free-swimming turtles outfitted with satellite-transmitting sensor tags together with data on a suite of environmental characteristics. In her internship study, Alice applied this approach to a set of loggerhead turtle movement data from the South Pacific. Her research is part of the ongoing collaboration on sea turtle pelagic ecology between EOD and the Marine Turtle Research Program of the Center's Protected Species Division.

Knowledge of the location and migration behavior of loggerheads is important to efforts to protect the threatened species from incidental capture (bycatch) in fishing gear, particularly on the high seas. In the North Pacific, research on loggerhead sea turtle migrations aimed at minimizing bycatch led to discovery of the Transition Zone Chlorophyll Front and its role as a loggerhead migration corridor and foraging area. Similar work has yet to be done in the South Pacific, which supports a separate population of loggerheads.

Ren's internship study represents a preliminary investigation of loggerhead sea turtle movements in the open ocean of the western South Pacific based on 42 loggerheads reared in captivity, tagged and released as juveniles (19 months of age) off the coast of New Caledonia. Movements of the turtles were tracked using SPOT-5 ARGOS satellite tags. The satellite position data were processed using a state space model (SSM) to produce corrected, daily tracks; the SSM primarily served to standardize the data temporally. Description of the loggerhead habitat was accomplished using satellite data on various environmental variables. The variables studied are presented in Table 1. Most of the satellite data were obtained from the NOAA OceanWatch database. The magnetic data were generated from the GEOMAG software program and the Globcolour chlorophyll-a (ocean color) dataset which was obtained from the GlobColour Project of the European Space Agency.

The analysis of habitat characteristics was primarily a resource selection experiment in which the various environmental variables were viewed as possible turtle resources. In analyzing satellite images of environmental variables, the value of a given variable in the image pixel closest to the turtle's location was assumed to be the value or level of the resource selected and utilized by the turtle. The available levels of the resource were estimated from a dynamic circle of habitat centered on the turtle's release point with a radius expanding at a rate determined by the swimming behavior of the turtle. For a given variable, all pixels of data within the habitat circle were collected at the image resolution of the variable. The relative percentage of time the turtles spent in locations with a particular value of a habitat resource was estimated, as was the relative frequency of occurrence (availability) of that resource value within the habitat circle, spatially and temporally. Cumulative frequency distributions of habitat utilization and habitat availability (each over the full range of habitat resource values) were compared using the non-parametric Kolmogorov-Smirnov (K-S) test. Of the 16 variables tested, only 7 resulted in significant values of the K-S test D statistic, as shown in Table 1 below.

Table 1. Kolmogorov-Smirnov tests indicated significant associations between the loggerhead turtles and 7 out of the 16 habitat descriptor variables studied (asterisks denote significance, ns means non-significant).
SST *** 37.729
SST gradient ns 14.162
SST gradient zonal ns 8.2576
SST gradient meridional ns 4.9835
Altimetry ** 31.4494
Geostrophic zonal ns 4.3997
Geostrophic meridional ns 7.6934
Magnetic force *** 42.81422
Magnetic declination *** 51.23715
Magnetic inclination *** 39.6088
Ocean color ns 19.4223
Bathymetry ** 30.2119
Bathymetric gradient ns 17.868
Bathymetric gradient meridional ns 8.7557
Bathymetric gradient zonal ns 8.4427
Wind Stress Curl ** 33.4422

For the variables whose cumulative distributions differed significantly according to the K-S test, the difference between the percentage of time the turtle spent at a given level of resource and the percentage availability of that level, temporally and spatially, was analyzed using a Linear Index proposed by Strauss (1979). The resulting selection curves (see Figure 1 for an example using the SST variable) are being used to generate potential loggerhead habitat maps.

Figure 1. Loggerhead turtle selectivity index for sea surface temperature.
Figure 1. Loggerhead turtle selectivity index for sea surface temperature.

An analysis of movement data showed that turtles spent more time in the Exclusive Economic Zones of neighboring countries than on the high seas (Figure 2), underscoring the importance of understanding pelagic movement of loggerheads in this complex region.

An interesting aspect of the results is the apparent significance of geomagnetic variables in the K-S tests and the possible importance of these variables for loggerhead movement. Work by Kenneth Lohmann and others in the Atlantic has suggested that loggerhead turtles there use magnetic declination, inclination, and force as navigational cues. From animations of the loggerhead movement (satellite tracks) over contour lines of magnetic force, magnetic declination, and magnetic inclination, it appears that the loggerheads in the South Pacific use the contours to navigate, as they seem to travel southward at an angle very closely parallel to the contour lines. While evidence from the statistical tests and animations is not conclusive, it suggests that loggerhead turtles could be using magnetic cues to navigate in pelagic waters of the South Pacific.

Although the habitat definition was not strong, the loggerhead turtles appeared to travel to three distinct habitat areas: the Tasman Sea between Australia and New Zealand; north and east towards Fiji, Kiribati, and the Cook Islands; and off the east coast of New Zealand between 40° and 45° S latitude. There is evidence that turtles were using the East Australian Current to forage, with some tracks circling along the coast of Australia. Sea surface height gradients should be examined further to determine whether there is a relationship between turtle tracks and eddies.

The loggerhead research in the South Pacific is ongoing. While the expanding circle habitat approach is an improvement over the static habitat box concept used in the North Pacific, future studies should be sensitive to the distinction between a turtle traveling to a given habitat and residing in the habitat; in this study, they were taken to be the same. Future work might also study foraging areas versus lanes of travel and should examine other habitat variables and try to identify an instrumental variable linked to loggerhead forage.

Figure 2. Tagged loggerhead sea turtles released in the South Pacific spent most of their time in EEZs 
            of nearby countries, rather than on the high seas.
Figure 2. Tagged loggerhead sea turtles released in the South Pacific spent most of their time in EEZs of nearby countries, rather than on the high seas.