Migration Paths of the Adult Female and Male Loggerhead Turtles Caretta caretta Determined through Satellite Telemetry

Fisheries Science 63(4), 547-552 (1997) Migration Paths of the Adult Female and Male Loggerhead Turtles Caretta caretta Determined through Satellite Telemetry Wataru Sakamoto,*1 Takeharu Bando,*1,*3 Nobuaki Arai,*1 and Norihisa Baba*2 *1Faculty of Agriculture, Kyoto University, Sakyo, Kyoto 606, Japan *2National Research Institute of Far Seas Fisheries, Shimizu, Sizuoka 424, Japan *3The Institute of Cetacean Research, Toyomi, Chuo, Tokyo 104, Japan (Received August 19, 1996) Migration paths of the loggerhead turtles were determined through satellite telemetry. The post nes ting females returned to the East China Sea, southwestern side of Japan. As these paths were bounded on the northeastern flowing Kuroshio, two different paths were found; F-1 migrated southern off side of the Kuroshio and F-2 migrated along the northern side lying between the Kuroshio and Japan Islands. The adult male was released on January 10, 1996, from Kushimoto, southern tip of the Kii peninsula. It returned again in nesting season near the Kii peninsula and stayed until August. The migration path and time were referred to the sunrise and sundown hour to determine the sun compass orientation. Key words: loggerhead turtle, telemetry, migration, Kuroshio, artificial satellite Adult female loggerhead turtles land on sandy beaches to lay eggs during early summer through autumn. The nesting beach is the same or near areas where they themselves hatched.1) The geographic range of nesting beaches in the main island of Japan, is distributed in the temperate zone from 33 N to 36 N latitude. Turtles return to the feeding ground again after the nesting season. The results of plastic tag return suggest that they will go back to the East China Sea within two or three months after nesting. The East China Sea is thought to be their feeding ground. Recently, the diving behavior of adult females during internesting periods has been investigated by time series data analysis based on the micro data tags.2-5) However the accurate post nesting migration paths of adult females have not yet been investigated in Japan. Neither the adult male turtle nor the new hatching turtle migration path in the oce an is known since they never land throughout the year. The long term tracking of sea turtles has been tried by many researchers by means of the satellite tag method, i.e., Argos system.6-10) Renaud and Carpenter (1994)8) tracked young and adult 4 loggerhead turtles for periods of 5.0 to 10.5 months in the Gulf of Mexico. They meas ured swimming speed, submergence interval and the corre lation between seasonal front shift in the estuary and swim ming areas. Their tracking results suggested the similar direct movement of turtles into estuaries and to feeding grounds as shown by Stoneburner (1982).6) In this paper, we investigated post nesting migration paths of two adult female loggerhead turtles. To compare with their migration paths, we tried to measure a migra tion path of an adult male before nesting season. This com parison is useful to understand when and where log gerhead turtles mate with each other. The migration paths are referred to the geomagnetic field, sea surface tempera ture, and the periods from the sunrise to the sundown along the migration path. Methodology for the attach ment of satellite tag (platform transmitter terminal; PTT) has improved such that PTT may remain on the turtle's carapace for long term. The attaching method is very simi lar to the micro data tag attachment as reported in our previous papers. 11,12) Methods Turtles Used.for the Experiment PTTs were attached on the carapace of two female log gerhead turtles on July 28, 1995 and on July 30, 1996 at the Senri coast, Wakayama prefecture, Japan (Fig. 1). They were named Female-1 (F-1) and Female-2 (F-2) re spectively. The former F-1 turtle weighed 100.0 kg, with 96.2 cm in straight carapace length (SCL). This turtle landed on the beach on June 1, 1995 at first and left the beach on July 30, after it deposited a total of 409 eggs in four time nests. We tracked this individual from July 30 to Sept. 30. The latter F-2 weighed 92.0 kg with 89.0 cm in SCL. It was found on the beach on June 11 and left on July 30, 1996. Since its first landing date was unknown, we could not count the total deposited egg number. It was tracked from July 31 to Sept. 10. A male turtle was captured by the set net in Kushimoto, Wakayama prefecture on January 9, 1996, and was released again on Jan. 10 after attaching the PTT. This tur tle weighed 72.5 kg, with 78.4 cm in SCL and above CLA class location was obtained until Aug. 1, 1996. The release point was shown in Fig. 1. PTTs Used for Tracking The PTT (Toyokomu, T-2038, Toyokomu Inc.) weighed 170 g in water by 35 mm in diameter with 195 mm length and an antenna, 189 mm in length, was located near an anterior portion on the PTT. The location (latitude and longitude) was informed us through CLS (Collect, Locali sation, Satellites, Argos and Service Argos Inc.,) as well as ID number, class location index, GMT, and the number of transmission. The model class location (CL), presented 6

548 Sakamoto et al. levels with circular confidence areas; 150 m, 350 m, 1 km for CL3, CL2, CLI, set no limits of accuracy for CLO, three locations for CLA and two locations for CLB. In these experiments, we adopted above CLA levels as a loca tion to plot the migration path. The swimming distance was defined to the length between anterior and posterior locations and their sum total was defined to the total traveling distance. Three environmental conditions were collected; the half monthly report of fluctuation of the Kuroshio and sea surface temperature (SST),*4 and earth's geomagnetic field.*5 In addition to these data, we calculated the daily time difference of the sunrise, sundown and the day length period as solar information along the migration path by means of the celestial navigation program. All post nesting turtles were identified by plastic tags at tached to a fore-left flipper when they landed on the beach. This work has been routinely carried out since 1990 to almost all landing turtles on the Senri coast. The loca tions by plastic tag return amount to 21 before satellite tracking experiment beginning. We plotted the recovery locations on the chart to compare with the females' satellite tracking paths (Fig. 1). Results Migration Paths of the Post Nesting Females A total of 30 locations were obtained by F-1 turtle above CLA for a period of 62 days from July 30 to Sept. 30. In Table 1, we summarized JST, CL, location, solar informa tion, and daily mean swimming speed. Female-1 crossed the Kuroshio on the day between August 1 and 3 (Fig.1). After passing through the Kuroshio, it swam toward the south direction with 3.45 km/h mean speed until Aug. 7 when F-1 changed its swimming direction toward southwest. This swimming direction was kept until Sept. 1. The swimming speed was estimated at 1.5 km/h during these southwest going periods. Female-1 seemed to cross the Kuroshio again near the Ryukyu Islands and arrived in the East China Sea until Sept. 19. It did not migrate such a long distance each day after arriv ing in the East China Sea (Table 1). Total traveling dis tance for a period of 51 days (July 30-Sept. 19) was 1667 km, therefore hourly mean speed within 51 days was about 1.4 km/h before arriving in the feeding ground. A total of 5 locations were obtained by F-2 for a period of 40 days from July 31 to Sept. 10, 1996. In Table 2, we summarized details. The mean swimming speed of F-2 was estimated about 0.8 km/h and a total traveling distance was accounted 800 km. These two post nesting turtles migrated similarly toward southwestern direction soon af ter nesting. However each migration path was separated by the Kuroshio; F-1 migrated southern side of the Kuroshio and F-2 passed through the northern boundary zone between the Kuroshio and Janan Islands. Migration Path of the Male Turtle A total of 9 locations were obtained by the satellite tag above CLA for a period of 115 days, from January 10 to May 3, 1996. Since the turtle was transported toward the northeastern direction by the Kuroshio soon after release, the migration path was nearly equal to the Kuroshio mean dering until it left the Kuroshio on Jan. 29 as shown in Fig. 2. On the 29th Jan. (JST), it began to swim toward the south direction until March 5 (29ß04.4 LN, 143ß12.5 LE). Af ter arriving on the Tori Shima island on Mar. 20, the tur- Fig. 1. Migration paths of post nesting female loggerhead turtles F-1 (solid circle and line) and F-2 (half solid circle and dotted line), tag recovery lo cations (solid square), longitudinal daily shift of same sundown hour (open triangle with broken line) and the Kuroshio meander in 1995 (broken thin line). *4 Hydrographical data report, published by the Maritime Safety Agencies. *5 WDC -C2 for geomagnetism, published by Data Analysis Center for Geomagnetism and space Magnetism, Faculty of Science, Kyoto University.

Migration Paths of the Loggerhead Turtles 549 Table 1. Tracking record of F-1 loggerhead turtle from July 30 to Sept. 30 and its solar information; sunrise, sundown and day length periods, in 1995 Fig. 2. Migration path of an adult male loggerhead turtle (solid circle and line), open circles mean the locations during nesting period around beach, and thin lines mean the Kuroshio meander in each period. Table 3. Tracking record of the male loggerhead turtle from Jan. 10 to May 3, the solar information and locations after arriving at nesting area, in 1996 Table 2. Tracking record of F-2 turtle from July 30 to Sep. 10, in 1996 tle began to swim toward the Japan archipelago and ap peared again off the Kii-Channel on the 3d May (33ß 43.3 ŒN, 134ß37.9 ŒE), where was a neighboring area where it had been captured by the set net (33ß28.0 ŒN, 134ß 37.0 ŒE). Total traveling distance was accounted about 2144 km for 115 day (Jan. 10-May 3), i.e., hourly mean speed was 0.8 km/h. We summarized JST, CL, location, solar information and daily mean swimming speed of the male in Table 3. Usually, the nesting season begins in mid May in the Kiipeninsula. According to the newspapers, the first landing of an adult female was reported on the 14th May in 1996. This adult male turtle arrived in the nesting ground about 10 days earlier than the first female's landing at the latest. The male swam around the eastern side along the coastal zone of the peninsula and the locations above CLA were obtained until Aug. 1 (Fig. 2 and Table 3). Sea Surface Temperature We compared migration path of F-1 with the sea surface water temperature (SST) after passing through the Kuroshio on Aug. 7 and found that areas along with the migration were covered with waters higher than 28ßC in SST. There were no characteristic features in the sea surface such as the existence of the cold water mass. The turtle moved directly toward the southwest direction without progressive fluctuation. Swimming speed was shown in Fig. 3 with the solar information. The male was transported by the Kuroshio soon after release. The temperature in the Kuroshio was higher 2 to 3ßC than the neighboring area. After passing through the Kuroshio, the male always swam in 20ßC area. Finally it ar-

550 Sakamoto et al. rived in 22 Ž area off the Kyushu Island on April 22 and entered into the Kuroshio zone again. During periods, the Kuroshio was flowing toward northeast as if running into Kii-peninsula from Kyushu Island. The turtle could arrive in the Kii-Channel easily by the Kuroshio. This area was covered 18 to 20 Ž in SST. Correlation between Long Range Migration and Solar Information We think underwater solar information may be an im portant factor for long range migration of the sea turtles. The navigation factors induced solar information were cal culated such as; the sunrise and sundown hours, the day length periods along the migration paths. They are shown in Figs. 3 and 4. Figure 3 shows the data of F-1 turtle from the released point. The constant sundown hour, 18:33 in mean JST time is shown during Aug. 10 through Sept. 1, when the turtle migrated toward southwestern direction. We could not find any correlation between solar informa tion and F-2 migration path as shown in Table 2. The sunrise or sundown hour seems to be correlated with the long range migration in the open ocean. We searched the same sundown hour position along the migration path of F-1 as following process. At first, we cal- Fig. 4. Solar condition along with the migration path of the adult male loggerhead turtle, upper is the day length period (solid circle) and dai ly mean swimming speed between anterior and posterior positions (bar). Change of the sunrise hour (middle) and the sundown hour (bot tom), culated total latitudinal migration length from Aug. 7 to Sept. 1 and divided into 25 (days) to define to daily latitudi nal migration component. Next, we searched the longitudi nal point where was shown the same sundown hour, 18:33 in mean JST in each day from Aug. 7 to Sept. 1. They are shown in Fig.l in every 2 days interval. Figure 4 shows the data of the male. The similar charac teristic pattern is shown in sunrise hour, 05:50 in mean JST, during March 20 through April 22, when the male be gan long range migration toward Japan archipelago. Two contrast patterns are shown, the similar sundown hour by the southwestern going F-1 and the similar sunrise hour by the northwestern going male. We tried to compare the migration paths with earth's ge omagnetic field as well as solar information but we could not find an accurate correlation between the path and total geomagnetic field value. Fig. 3. Solar condition along with the migration path of the post nest ing loggerhead turtle, upper is the day length period (solid circle) and daily mean swimming speed between anterior and posterior posi tions (bar). Change of the sunrise hour (middle) and the sundown hour (bottom) indicated in JST (Japan Standard Time). Discussion Migration Path of Post Nesting Females Adult female loggerhead turtles visit the Japan Islands in intervals of every 2 or 3 year.11) The Kuroshio is flowing toward northeastern direction along the Japan Islands.

Migration Paths of the Loggerhead Turtles 551 The current speed is estimated about 4 km/h off the Kii peninsula.*4 While mean swimming speed of the turtles were estimated about 1.5 to 0.8 km/h as shown in Tables 1 to 3. The flow will be available when they visit to the nesting beach in early summer from the East China Sea since the flow direction is the same as their migration course. On the other hand, the post nesting turtles have to avoid the flow since energy consumption is too much to swim against the flow. Two different paths are found on both sides of the Kuroshio, one is southern off side. Another is northern side, where is the boundary zone between the Kuroshio and Japan Islands. The satellite tracking result in 1995 indicated that F-1 returned to the East China Sea along southern side. In this case, turtles will have to cross the strong current twice, once was soon after leaving the nesting beach, another was just before arriving in the East China Sea. Typical effect was shown in the deviation on the positions between Aug. 1 and Aug. 3 in Fig. 1. It suggests that F-I was transported toward eastern direction until passing through the Kuroshio during 2 days. However we could not find the other deviation pattern induced by the Kuroshio in the satellite data. The plastic tag recovery locations in Fig. 1 are distribut ed along the F-2 migration path. These plastic tags are reco vered by coastal fishermen since the turtes passing ttthrough northern side are sometimes captured by coastal fishing gear such as set nets or trawl nets. The fishermen detach these tags from turtles' fore-left flipper before they release the turtles again in the ocean. Long Range Migration of Sea Turtles The migration paths indicate neither a zigzagging progress nor random directional fluctuation. Furthermore the adult male returned again near the Kii peninsula as if being in time for nesting season. It somewhat suggests that the turtles are able to know seasonal progress as well as to detect their position in the sea. Lohmann and Lohmann (1993)13) reported that new hatching leatherback turtles have an ability to detect the earth's geomagnetic field and they pointed out in the same paper that hatchlies were in dependent against light in their orientation. They also dis cussed in the other report (1996)141 that sea turtles possess the minimal sensory abilities necessary to approximate global position using a bicoordinate magnetic map. Simi lar geomagnetic orientation of hammerhead shark has been reported by Klimley.15) On the other hand, the polarotactic behavior is known for the orientation as well as geomagnetic orientation in some marine organisms.16,17) We think the earth's geomag netic compass may be effective to detect the swimming direction in the long range navigation, but it is useless to know date. The solar information may be effective to know seasonal and temporal progress depending on the sun orbital. This mechanism seems to be somewhat differ ent from the polarotactic because specific visual sensor may be necessary to detect polarized light vector compo nent, so called e-vector, in the sea for polarotactic. In the case of solar information, a turtle merely needs visual sen sor to detect the temporal progress to be dark and light in the sea. And the animal must possess a biological clock to know the diurnal sun time and the season. The temporal information may be important for the adult male to arrive on the nesting ground again to mate before nesting season beginning. If the turtle begins to migrate with sunrise hour until noon against the sun's orbital, it may be possible to mi grate toward the northwest direction. If the turtle begins to migrate along the sun from noon to sundown, it will be able to go toward the southwest direction. The best fit course is presented in Fig. 1 based on the sundown hour in each day for F-1. This hypothesis was already presented by Balchen (1977).18) However, typical surfacing time could not be found by the PTT location data as shown in Tables 1 and 2. If we try to understand the hypothesis, we need the other information by means of micro data tags capable to record the magnetic components, or the swimming speed since we need the evidence whether the turtle active ly swims in accordance with geomagnetic field or the sun rise or sundown hour. Correlation between SST and Migration Path of the Male The male migrated toward northeast direction soon af ter releasing. It was transported by the Kuroshio until Jan. 29. After leaving the Kuroshio, it migrated toward southern area where was always covered higher than 20 Ž in SST throughout the migration periods. The migration course seemed to be closely correlated with SST distribu tion pattern. In the winter season, SST in the East China Sea becomes lower than 15 Ž."This low temperature is a critical thermal condition for turtles since they have to maintain higher body temperature than the ambient water temperature. It is said that the bigger sea turtle is higher in body tem perature and capable to keep constant against the environmental thermal fluctuation.19-21) We do not know the corre lation between body mass and critical ambient water tem perature in loggerhead turtles, but it seemed to be a critical temperature for this male in winter season and therefore began to migrate toward southern warmer than 20 Ž areas in winter season. Acknowledgments We would like to express our gratitude to Mr. Tsuneo Miyawaki, of the Kushimoto Marine Park for the providing us the male loggerhead turtle. We give our thanks to the many students in the Division of Fisheries, Kyoto University for their kind assistance and cooperation in patrolling during the night. This research was partly supported by Grant-in-Aid for Scientific Research (A) and (B), The Ministry of Education, Science, Sports and Culture (08556031) and (08456097). References 1) B. W. Bowen, F. A. Abreu-Grobois, G. H. Balazs, N. Kamezaki, C. L. Limpus, and R. L. Ferl: Trans-Pacific migrations of the log gerhead turtles (Caretta caretta) demonstrated with mitochondrial DNA markers. Proc. Natl. Acad. Sci. USA, 92, pp. 3731-3734 (1995). 2) S. A. Eckert, D. W. Nellis, K. L. Eckert, and G. L. Kooyman: Div ing patterns of two leatherback sea turtles (Dermochelys coriacea) during internesting intervals at Sundy point, St. Crox, U. S. Virgin Islands. Herpetologica, 42, 381-388 (1986). 3) S. A. Eckert, K. L. Eckert, P. Ponganis, and G. L. Kooyman: Div ing and foraging behavior of leatherback sea turtles (Dermochelys coriacea). Can. J. Zool., 67, 2834-2840 (1989). 4) Y. Naito, W. Sakamoto, I. Uchida, K. Kureha, and T. Ebisawa: Es timation of migration route of the loggerhead turtle Caretta caretta around the nesting ground. Nippon Suisan Gakkaishi, 56, 255-262

552 Sakamoto et al. (1990). 5) W. Sakamoto, K. Sato, H. Tanaka, and Y. Naito: Diving patterns and swimming environment of two loggerhead turtles during internesting. Nippon Suisan Gakkaishi, 59, 1129-1137 (1993). 6) D. L. Stoneburner: Satellite telemetry of loggerhead sea turtle move ment in the Georgia Bight. Copeia, 2, 400-408 (1982). 7) R. A. Byles and C. K. Dodd: Satellite biotelemetry of a loggerhead turtle (Caretta caretta) from the east coast of Florida, in "S. A. Eck ert, K. L. Eckert and T. H. Richardson (compilers)" Proc. Ninth Ann. Workshop Sea turtle Cons. Bio., NOAA Tech. Memo. NMFS-SEFC-232, 1989, pp. 215-217. 8) M. L. Renaud and J. A. Carpenter: Movements and submergence patterns of loggerhead turtles (Caretta caretta) in the Gulf of Mexi co determined through satellite telemetry. Bull. Mar. Sci., 55, 1-15 (1994). 9) H. C. Liew, E. H. Chan, P. Luschi, and F. Papi: Satellite tracking data on Malaysian green turtle migration. Rend. Fis. Ace. Lincei, Series 9, 6, 239-246 (1995). 10) S. J. Morreale, E. D. Standra, J. R. Spotila, and F. V. Paladino: Migration corridor for sea turtles. Nature, 384, 319-320 (1996). 11) K. Sato, W. Sakamoto, Y. Matsuzawa, H. Tanaka, and Y. Naito: Correlation between stomach temperatures and ambient water tem peratures in free-ranging loggerhead turtles, Caretta caretta. Ma rine Biology, 118, 343-351 (1994). 12) W. Sakamoto, 1. Uchida, Y. Naito, K. Kureha, M. Tujimura, and K. Sato: Deep diving behavior of the loggerhead turtle near the frontal zone. Nippon Suisan Gakkaishi, 56, 1435-1443 (1990). 13) K. J. Lohmann and C. M. Lohmann: A light-independence magnet ic compass in the leatherback sea turtle. Biol. Bull., 185, 149-151 (1993). 14) K. J. Lohmann and C. M. Lohmann: Detection of magnetic field in tensity by sea turtles. Nature, 380, 59-61 (1996). 15) A. P. Klimley: Highly directional swimming by scalloped ham merhead sharks, Sphyrna lewini, and subsurface irradiance, temper ature, bathymetry, and geomagnetic field. Marine Biology, 117, 1-22(1993). 16) A. D. Hasler: Orientation and fish migration, in "Fish physiology VI" ed. by W. S. Hoar and D. J. Randall, Academic Press, New York and London, 1971, pp. 429-510. 17) T. H. Waterman and R. B. J. Forward: Field evidence for polarized light sensitivity in the fish Zenarchopterus. Nature, 228, 85-87 (1970). 18) J. G. Balchen: Modeling, prediction and control of fish behavior, SINTEF Rep., SINTEF, Trondheim, 1977, p. 49. 19) N. Mrosovsky and P. C. Pritchard: Body temperatures of Der mochelys coriacea and other sea turtles. Copeia, 4, 624-631 (1971). 20) C. H. Ernst: Temperature-activity relationships in the painted tur tle, Chrysemys pitta. Copeia, 2, 217-222, (1972). 21) W. Frair, R. G. Ackman, and N. Mrosovsky: Body temperature of Dermochelys coriacea warm turtle from cold water. Science, 177, 791-793 (1972).