Post-release wide-ranging movements of sea turtles after prolonged captivity

1 1 2 Post-release wide-ranging movements of sea turtles after prolonged captivity 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 Mestre, Frederico 1, 2 ; Bragança, Marco P. 1, 3 ; Nunes, Antonieta 1 ; dos Santos, Manuel E. 4 1 Zoomarine - Mundo Aquático S.A., E.N. 125, km 65, Guia, 8201-864 Albufeira, Portugal. 2 Current affiliation: CIBIO-UE - Centro de Investigação em Biodiversidade e Recursos Genéticos/Universidade de Évora, Departamento de Biologia, Núcleo da Mitra, 7002-554-Évora, Portugal. 3 Corresponding author. Email: marco.braganca@zoomarine.pt, Telf.: +351289560300. 4 Eco-Ethology Research Unit, ISPA - Instituto Universitário, Rua Jardim do Tabaco, nº34, 1149-041 Lisboa, Portugal. This is a preprint version of the following paper: Mestre, F., Bragança, M. P., Nunes, A., & dos Santos, M. E. (2014). Satellite tracking of sea turtles released after prolonged captivity periods. Marine Biology Research, 10(10), 996-1006. URL: http://dx.doi.org/10.1080/17451000.2013.872801. 18 19 20 21 22

2 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 Post-release wide-ranging movements of sea turtles after prolonged captivity Rehabilitation is an important part of sea turtle conservation efforts, and tracking the animals is both a way of obtaining information on their movements and behaviour, and to monitor, at least temporarily, the success of the rehabilitation process. Two green turtles, Chelonia mydas (Linnaeus, 1758), and one loggerhead turtle, Caretta caretta (Linnaeus, 1758), were rehabilitated after long periods in captivity, and their movements were followed by satellite tracking after their release near the South of Portugal. Location data were obtained for an average of 688 days after release. All three animals showed a directional movement and the last observed locations coincided with known feeding areas for both species, near the coasts of Mauritania and the state of Ceará, Brazil, for the green turtles, and near Cuba, for the loggerhead. Bathymetry, surface currents, wind, sea surface temperature, chlorophyll-a concentration and geomagnetic field were analysed as spatial and environmental variables potentially affecting movement. Only bathymetry, sea surface temperature and geomagnetic variables showed significant association with the path choice. This project supports the notion that sea turtles have the ability to survive in the wild after long periods in captivity and to return to known feeding areas of the species, justifying the effort in their rehabilitation and post-release tracking. Keywords: Sea turtle rehabilitation, satellite tracking, Chelonia mydas, Caretta caretta. 45 46 47 48 49 50 51 52 53 54 Introduction Sea turtles are threatened throughout their life cycle by several anthropogenic factors such as bycatch, poaching, modification of nesting beaches, pollution and marine debris (Bolten et al. 2010; Donlan et al. 2010). Although not every sea turtle that is ill, injured or entangled strands ashore (Epperly et al. 1996), beached individuals of such charismatic fauna normally attract more public attention (Feck & Hamann 2013), which may lead to a rehabilitation response. The reasons listed by Moore et al. (2007) for the rehabilitation of marine mammals can also be used to justify sea turtle rehabilitation attempts: conservation of endangered species; care of animals harmed by human

3 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 activities; mitigation of human use of sea turtle nesting beaches; research focused on rehabilitation or wildlife medicine; post-release tracking, as a way to improve knowledge on wild populations and migration; public education campaigns about marine ecosystem health and its importance to sea turtle conservation. Considering the low survival probability of each turtle (from egg to adulthood, e.g. Frazer 1986) but also its potential reproductive output (Broderick et al. 2003), the rehabilitation of a single, or a small number, of individuals becomes demographically relevant. Studies on the post release behaviour of rehabilitated sea turtles are limited and have mostly been restricted to sea turtles accidentally captured by long-liners and released after hook removal on-board the fishing boats (Swimmer et al. 2006 and Sasso & Epperly 2007) or to individuals entangled in trammel nets and released after a short period of time (e.g. Snoddy & Southwood 2010). Some sea turtles have been released carrying satellite transmitters, which allowed the study of their subsequent movements and to monitor their rehabilitation success (see Godley et al. 2008). Post-release adaptation to the wild is critical to the rehabilitation process. With a few exceptions (e.g. Bentivegna 2001; Cardona et al. 2012) this has not been studied in depth, particularly for sea turtles that experienced long captivity periods. This study provides new data on sea turtles post-release movements after long periods in captivity, through the satellite tracking of two green turtles (Chelonia mydas) and one loggerhead turtle (Caretta caretta). A comprehensive set of potential variables was analysed, building on other studies that assessed only one or a low number of parameters, such as sea surface temperature (Hays et al. 2001 and Hawkes et al. 2007), geomagnetic variables (Benhamou et al. 2011), surface currents (Luschi et al. 2003; Bentivegna et al. 2007 and Hawkes et al. 2007) or chlorophyll-a concentration (Polovina et al. 2001). The main objective of this paper is to obtain information on sea turtle movement and behaviour, assessing the success of the rehabilitation process after prolonged periods of captivity.

4 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 Materials and Methods Specimen handling and release Two female green turtles and one female loggerhead turtle were released in the Atlantic Ocean, about 20 miles south of Portimão, Portugal (approximately 36 46'48.57"N; 8 32'16.54"W) on 30 September 2009, after being rehabilitated at Zoomarine s Rehabilitation Centre, Porto d Abrigo (PdAZ). The three turtles tracked in this study came from the following backgrounds: CM1 (green turtle, nicknamed Tartaruga) was transferred from the Funchal Municipal Museum (Madeira, Portugal) on the 22 nd of April 2004, where it was kept for about 30 years. On arrival at the rehabilitation center: weight = 32.0 kg, straight carapace length (SCL) = 60.6 cm; On release: weight = 80.4 kg, SCL = 78 cm. CM2 (green turtle, nicknamed Cat) was seized by the Portuguese airport customs on the 27 th of August 2001 (on arrival at the rehabilitation center: weight = 4.8 kg, SCL = 33.9 cm). It was sent to the PdAZ by the national authority for nature conservation (ICNF Institute for Nature and Forest Conservation). On March of 2006 the left fore flipper was surgically amputated due to a severe infection but its subsequent swimming ability was apparently not affected. On release: weight = 60.5 kg, SCL = 77.5 cm. CC (loggerhead turtle, nicknamed Calantha) was transferred from the Vasco da Gama Aquarium (Oeiras, Portugal) on the 13 th of October 2005, where it was exhibited for about 25 years (on arrival at the rehabilitation center: weight = 115 kg, SCL = 84 cm; on release: weighted = 125.8 kg, SCL = 84 cm). All specimens had a long history of captivity. During the rehabilitation time at PdAZ, human interaction with these individuals was minimized. Each turtle was considered fit for release when it met certain predefined criteria (Bluvias 2008): 1) no medication for at least two weeks; 2) the ability to capture live prey; 3) good body condition; 4) no diseases or lesions (wounds, tumours, skin irritation, epibionts or endoparasites); 5) normal digestive function; 6) the ability for active movement; 7) the capacity to lift its head and breathe normally; 8) attempts to move when on dry substrate; 9) normal blood parameters for 2 to 4 weeks. The three turtles were fitted with Argos System satellite tags on top of the carapace following the procedure described by Coyne et al. (2008). The KiwiSat 101 transmitters (Sirtrack Limited, New Zealand) operated with a 40-second repetition rate

5 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 and a salt water switch (transmitting only when the animal surfaced). The location data time span was from September 2009 to September 2011. Environmental data General Bathymetric Chart of the Oceans (GEBCO) bathymetry (in meters) was retrieved through the British Oceanographic Data Centre website (BODC 2012). This dataset has a resolution of 30 arc-seconds, and was generated by combining qualitycontrolled ship depth soundings with interpolation between sounding points, which is estimated by satellite-derived gravity data (GEBCO 2012). Bathymetric data might be an important variable to understand the movement of sea turtles, since these species have continental (nesting) and neritic (feeding) stages in their life history (Kobayashi et al. 2008). Data on surface currents (in ms -1 ) were downloaded from the Ocean Surface Current Analysis (OSCAR) website (Bonjean & Lagerloef 2002; OSCAR 2012). These data consisted of information about the East-West (u) and North-South (v) components of the surface currents which were transformed to speed ( S flow direction (D = arctan(u/v); in degrees). u 2 v 2, in ms -1 ) and water Chlorophyll-a data (in mg m -3 ) and Sea Surface Temperature (SST), (in ºC) were downloaded from the NASA Earth Observations website (http://neo.sci.gsfc.nasa.gov) and were obtained by the Moderate Resolution Imaging Spectroradiometer (MODIS) instrument in the NASA's satellite Aqua, with a resolution of 1 km. The chlorophyll-a data provide information on the concentration of surface phytoplankton, and the SST is the temperature of the top millimetre of the ocean, accurate to a half a degree Celsius. Both chlorophyll-a and SST were considered by previous studies (e.g., Kobayashi et al. 2008) as important variables in loggerhead movement. Geomagnetic variables were based on the World Magnetic Model for 2010/2015 and were downloaded from NOAA s National Geophysical Data Center website (ftp://ftp.ngdc.noaa.gov/geomag/wmm/wmm2010/shapefiles/). Three geomagnetic variables were considered for analysis: declination (in degrees), inclination (in degrees) and intensity (in nanoteslas). The spatial representation of the geomagnetic isolines will depend on the orientation strategy followed by the turtles in relation to these variables, as discussed by Lohmann et al. (2007). Here, we interpolated the isolines in

6 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 order to obtain continuous surfaces, as these gradients might be the best way to identify any influence of the magnetic field. Geomagnetic variables are important navigation cues at intermediate scales to green turtles (Benhamou et al. 2011) and also to the movements of loggerheads (Lohmann et al. 2004 and Kobayashi et al. 2008). Surface winds were also considered as a relevant variable. East-West (u) and North- South (v) components were obtained from NOAA s Multiple-Satellite Blended Sea Surface Winds, downloaded from the NOAA-NOMADS Live Access Server (http://nomads.ncdc.noaa.gov/las/), and speed (in ms -1 ) and direction (in degrees) were subsequently computed. Collard & Ogren (1990) found that surface winds are important to dispersal patterns of post-hatching Kemp s ridley turtles (Lepidochelys kempii); Hochscheid et al. (2010), on the other hand, reported that loggerheads may remain at the surface for extended periods to absorb solar radiation or to recover from anaerobic activity, and therefore may be significantly affected by winds; the effect of the drag caused by the satellite transmitter on the movements of the sea turtles has been noted by Watson & Granger (1998). All these factors highlight the potential influence of surface winds on sea turtle paths. Shoreline data were downloaded from the GSHHS (Global Self-consistent Hierarchical High-resolution Shoreline) website (GSHHS 2012) (Wessel & Smith 1996). The highest resolution data available were used. All the ecogeographical variables (with the exception of bathymetry geomagnetic variables) were averaged monthly in the considered time period. The wind direction and current flow values follow the oceanographic convention (angles are measured clockwise from North). All GIS analysis and plotting was carried out using Quantum GIS, version Lisboa 1.8.0 (Quantum GIS Development Team, 2013) and two R packages: raster (Hijmans & van Etten 2012) and maptools (Lewin-Koh et al. 2012). The software R version was x64 2.15.1 (R Core Team 2012). 178 179 180 181 Track analysis Data were received from the Argos system through the STAT (Satellite Tracking and Analysis Tool) platform, which is a web-based tool that facilitates the reception and

7 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 filtration of Argos data (Coyne & Godley 2005). The specimen s location, in near realtime display, may be sent to a website, allowing the sharing of this information with the general public. The Argos data are classified according to the location accuracy of each point, which mainly depends on the geometrical conditions of the satellite pass at the time it receives the uplinks, and on the stability of the transmitter frequency (Argos, 2011). In the particular case of marine species, the limited time spent at the surface restricts the number of uplinks received by the satellite, so these data are particularly prone to location error (Freitas et al. 2008). The Argos data locations are classified according to an error estimate, and only those in the error classes 1(<1500 m), 2 (<500 m) and 3 (<250 m) were retained for further analysis. Locations on land, as well as those that imply an unrealistically high swimming speed of more than 6 km.h -1 for the loggerhead turtle (Bentivegna et al. 2007) and 5 km.h -1 for the green turtles (Luschi et al. 1998) were also discarded. Defining resource use and availability is a fundamental stage in resource selection studies. Resource use can be defined in several ways, such as time spent and distance travelled within the used area (Buskirk & Millspaugh 2006). The availability of resources was defined for each turtle by considering a kernel 99% to avoid the exclusion of data from paths actually used. This was calculated using the fixed kernel with the smoothing parameter computed by the reference bandwidth method (Powell 2000), using the R package adehabitathr (Calenge 2006). The used resources are those within a buffer with a width of 60 km (considering that the maximum daily average distance travelled was just less than 30 km). To identify variables accounting for the turtle path choice (departing from resource availability) a Kolmogorov-Smirnov test (KS) was performed highlighting deviations between availability and actual use of space. Subsequently, for each of the variables for which significant differences were found, a Strauss Linear Index (SLI) (Strauss 1979) was computed: 210 SLI ˆ (1) O i i 211

8 212 213 214 215 216 217 218 219 Where Oi ui / u (sample proportion of used units in category i) and ˆ i m i / m (sample proportion of available units in category i). The analysis of current and surface winds significance (speed and direction in both cases) was performed using a different strategy, pairing movement direction and speed with the underlying variables. For wind and current direction, the circular correlation statistic was used, running the R package CircStats (Agostinelli 2012), while for wind and current speed a linear model was adopted. The analyses were performed separately for each of the turtles. 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 Results Location data were obtained for an average of 688 days after release (CC=653 d; CM1=675 d; CM2=736 d). The tracking data from the three turtles are plotted on Figure 1, showing that all three animals followed a targeted movement (rather than random) and, most importantly, that the last observed locations coincided with known feeding areas for each species (Marcovaldi & Marcovaldi 1999; Fretey 2001; Dodd & Byles 2003; Santos et al. 2011). In the beginning of this study, both of the C. mydas turtles followed a path along the coast, with CM2 crossing the Atlantic to Brazil after reaching Western Sahara, and the CM1 staying near the African coast. The third turtle, the C. caretta, crossed the Atlantic shortly after being released, passing between the Madeira and Canary Islands until it reached the coastal waters of Cuba. FIGURE 1 In studies of this nature, data quality can be assessed by evaluating the percentage of locations within each of the error classes, as seen in Table 1. CC had more accurate locations, probably due to longer or more frequent surfacing periods. CM1 s satellite data were the least accurate, with more than 80% of the locations having no error estimation, although it is the specimen with the largest time span of data (Figure 2). For unknown reasons, locations from specimen CM2 were interrupted for over a year, during which the animal stayed in the same general area (see Hays et al. 2007 for a discussion of transmission problems).

9 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 TABLE 1 FIGURE 2 The last recorded locations for each of the turtles were: East coast of Cuba (CC); Banc d Arguin National Park, off the coast of Mauritania (CM1) and north-eastern coast of Brazil (CM2). The average daily speed of each specimen was: 0.191 km.h -1 (CM1), 1.15 km.h -1 (CC) and 0.995 km.h -1 (CM2). CM1 has a lower average speed because the animal s movement was reduced upon arriving at the coastal region of the Banc d Arguin National Park. Ignoring these last locations, speed values were similar to those of the other turtles. Track analysis Tables 2 to 4 show the main results of the statistical analysis on the influence of environmental variables in the paths chosen by each sea turtle. TABLE 2 TABLE 3 TABLE 4 The following analysis, using the Strauss Linear Index, evaluates the relation between each turtle and the environmental variables. The curves representing the SLI values are presented in Figure 3. FIGURE 3 These results highlight the importance of the geomagnetic variables in the sea turtles navigation. In the case of CM1, all three geomagnetic variables were significant to the path followed. Nevertheless there seems to be no particular preference within each variable. That is explained by the isoline layout, which is fairly perpendicular to CM1 s path (Figure 4). Bathymetry was also important to this specimen s movement, which showed a strong preference for shallow areas. Surface currents and winds had no significant impact on space use. To CM2, inclination and declination had a more important effect. Regarding geomagnetic field inclination there is a strong selection of values between -12º and -

10 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 7.89º, explained by the final part of the path where the turtle moves parallel to the isolines. Geomagnetic field declination is basically perpendicular to this turtle s path, explaining the variation in the SLI. The coastal path followed indicates a strong selection of shallow areas (and a strong avoidance of depths over 3830 m). Wind direction was also significant. However, it had a residual importance in path choice, considering that the correlation coefficient was very low (r = 0.134). CC was also influenced by the geomagnetic variables. It followed a path parallel to some inclination isolines with a selection of values between 33.1º and 35.3º. Geomagnetic field intensity also shows a degree of selection between 38760 nt and 39369 nt, for the same reason. CC moves in a perpendicular path to the declination isolines. In what regards field declination, there was an avoidance of values between 20.26º and 19.88º, with no particular selection for values above, as shown by the SLI. Bathymetry was an important variable for this turtle, with a positive selection of depths between 6140 m and 5620 m, and between 5010 m and 3980 m. Contrary to the green turtles, this specimen approached and departed the coast perpendicularly, which is expressed in a negative selection of shallower areas. Finally, SST was also a significant variable to CC. This specimen selected values under 22.4ºC, according to the SLI. Surface currents and wind direction were also significant. FIGURE 4 292 293 294 295 296 297 298 299 300 301 302 Discussion Three turtles, of various and unknown origins, spent a long time in captivity, and were finally brought to a rehabilitation centre in the South of Portugal, where their condition was much improved. Considered fit to return to the ocean, they were released after being tagged with satellite transmitters for an opportunistic study of their movements. It was hoped that this effort would produce some insights concerning their preferences and a notion about the survival chances of animals in such circumstances. All three turtles were tracked for a significant time and showed non-random displacements that interestingly brought the animals to traditional feeding regions of both species. The extension of the post-release movements is remarkable (CC = 9709 km, CM1 = 3053

11 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 km and CM2 = 6806 km), but see Hays & Scott (2013) for context. It can be compared to those in a previous study carried out in the Pacific (Luschi et al. 2003). Interestingly, the final locations determined are close to known feeding grounds for both species. Banc d Arguin National Park (Fretey 2001; Godley et al. 2010) and the coast of Ceará, Brazil (Marcovaldi & Marcovaldi 1999; Santos et al. 2011) for green turtles and the Great Bahamas Bank, near the north coast of Cuba, for the loggerheads (Dodd & Byles 2003). Having no detailed information on the pre-capture background of these specimens, it is impossible to know to what extent any previous migration experience was relevant (CC and CM1 were probably captured as adults, while CM2 was likely collected as a juvenile). The observed effect of bathymetry, which was significant to both C. mydas turtles, might be explained by the extension of their path following the coast of Africa. These coastal movements have been reported in previous works (such as Godley et al. 2002 and Troëng et al. 2005) and were considered an optimum strategy regarding feeding. So, it can also be hypothesized that this option might maximize post-release food intake, by moving along the coast and thus increasing the chance of finding food. Geomagnetic variables (intensity, inclination and declination) were the most significant variables explaining the use of space by the turtles. Studies with magnetically disturbed sea turtles and with displaced animals (e.g. Luschi et al. 2007 and Benhamou et al. 2011) have supported the notion that sea turtles perceive the magnetic field and navigate using this information. The ability of sea turtles to use geomagnetic inclination as an approximation of latitude (Lohmann & Lohmann 1994) combined with the detection of geomagnetic intensity (Lohmann & Lohmann 1996) might allow navigation with a bicoordinate magnetic map as has been previously proposed and is considered to be possible in some regions of the world (Putman et al. 2011 and Boström et al. 2012). Additionally, as mentioned by Monzón-Argüello et al. (2009), juvenile loggerheads are most frequently found in geomagnetic inclination isolines common to their natal rookeries (between 34º and 46º). Given that CC selected a mean inclination of 37.013º, we might presume that this was an important factor in path choice, although it is an adult. Overall, our data support the notion of geomagnetic navigation abilities for these animals. SST was only significant to CC, which selected values between 20ºC and 22.4º C. The

12 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 initial portion of CC s path is consistent with an avoidance of lower SST on the northernmost part of the available area. Following this, CC s path reveals a preference for intermediate SST values, avoiding the warmer southern temperatures. This might be due to other factors, namely the magnetic inclination. The values of the SST selected are comparable to those found in previous works, such as 14.45 C to 19.95 C for loggerheads in the Pacific (Kobayashi et al. 2008), 10.21ºC to 28.4ºC for juvenile loggerheads in the South-western Atlantic (Barceló 2011) and 20.20ºC to 29.50ºC for green turtles in the Gulf of California (McDermott et al. 2003). However it has been reported that, in some areas, SST is a variable of minor importance in regard to the path choice by green turtles (Hays et al. 2001). Chlorophyll concentration did not have any significant effect on the path chosen by any of the turtles. Nevertheless, the values of chlorophyll in the area used are similar to those from previous studies, such as 0.458 ± 1.012 mg/m 3 for juvenile loggerheads in the South-western Atlantic (Barceló 2011) or between 0.11 mg/m 3 and 0.31 mg/m 3 for loggerheads, in the Pacific Ocean (Kobayashi et al. 2008). The lower mean chlorophyll concentration in the area used by CC was probably due to the open ocean path chosen by this specimen. The higher concentration in the area used by CM1 was due to the fact that this turtle spent a considerable proportion of the time in an area with high chlorophyll concentration, at the Banc d Arguin National Park, Mauritania. Current speed had no significant effect on the turtles paths, suggesting that neither of the turtles used the current to assist their movement. This is expected since flow direction assessment can be very difficult in the ocean (Chapman et al. 2011). Only the specimen CC (which spent more time at the surface) was significantly affected by surface current direction. Previous works had confirmed the relevance of ocean currents to loggerheads (Bentivegna et al. 2007) and green turtles (Girard et al. 2006). A recent review on the behaviour of animals in flows, whether air or water (Chapman et al. 2011) has suggested that sea turtles follow a strategy of Full Drift Constant Compass Orientation, which means that sea turtles keep constantly heading towards their goal regardless of the flow, resulting in some lateral displacement. In fact, at least for green turtles, currents have a negative effect on navigation ability, as the turtles cannot compensate the deflecting action of currents (Girard et al. 2006). Surface wind speed and current speed had no significant effect on the paths.

13 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 Nevertheless, the direction of surface winds was considered significantly correlated with the directions followed by CC and CM2. Given that sea turtles are air breathing animals, and spend a portion of their time at the water surface, they might be subjected to wind driven circulation of the water column, as well as wind drag on their carapace, as also suggested by Watson & Granger (1998). In the case of CC, the influence of surface currents and wind direction might be explained by more frequent or extended surfacing periods, therefore with higher exposure to surface winds and currents. This is supported by two sources of information: our own observations (when at the rehabilitation centre this specimen spent more time at the surface than the other two) and the number and quality of locations received. CC had more location data (until transmissions stopped) and the locations were more accurate, which was due to the more frequent presence of the transmitter at the surface. Rehabilitation of stranded or confiscated specimens is an important component of conservation efforts, particularly in species that are facing conservation challenges, such as climate change (Witt et al. 2010; Wallace et al. 2011 and Pike 2013), poaching (eggs), fisheries bycatch, pollution and coastal urban development (Wallace et al. 2011). Additionally, each adult is particularly valuable since few embryos survive to adulthood, (approximately 1/1000 in loggerheads (Frazer 1986) and given that sexual maturity is reached relatively late in life (42 to 44 years in green turtles and 23.5 to 29.3 years in loggerheads) (Goshe et al. 2010 and Casale et al. 2011). So, every reproductively active adult is of major importance to the species conservation. The main purpose of this study was to evaluate the ability of these sea turtles to survive in the wild after long periods in captivity. Overall, we note that their return to the wild was apparently successful, considering that these animals swam to known feeding areas of their species. The rehabilitation process does not end with the release of the individuals to their natural habitat. Tracking the animals is both a way of obtaining information on their movements and behaviour, and to monitor, at least temporarily, the success of the rehabilitation effort. Acknowledgements The administration board of Zoomarine Portugal funded the spatial analysis, and Élio

14 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 Vicente was instrumental to this study in many ways. We are grateful to the Portuguese Navy for the continuous collaboration in turtle release operations. We thank Michael Coyne (seaturtle.org) for help in data collection, Paulo Relvas (CCMAR/FCT, University of Algarve) for assistance in the interpretation and treatment of some variables, and José Matos (Instituto Nacional de Investigação Agrária e Veterinária, I.P.) for the very useful comments on the manuscript. References Agostinelli C. 2012. CircStats: Circular Statistics, from Topics in circular Statistics. 2001. R package version 0.2-4. http://cran.r-project.org/package=circstats (accessed 28 March 2013). Computer program. Argos 2011. Argos User s Manual. http://www.argos-system.org (accessed 1 March 2013). Barceló C. 2011. Movement Patterns and Marine Habitat Associations of Juvenile Loggerhead Sea Turtles (Caretta caretta) in the Southwestern Atlantic Ocean. Master of Science Thesis. Oregon State University, USA. 76 pages. Benhamou S, Sudre J, Bourjea J, Ciccione S, De Santis A, Luschi P. 2011. The Role of Geomagnetic Cues in Green Turtle Open Sea Navigation. PLoS ONE 6(10): e26672. 11 pages. Bentivegna F. 2001. Possibilities of reintegrating sea turtles kept in captivity to the wild. Bulletin - Institut Oceanographique Monaco, numero special (1):373-378. Bentivegna F, Valentino F, Falco P, Zambianchi E, Hoschscheid S. 2007. The relationship between loggerhead turtle (Caretta caretta) movement patterns and Mediterranean currents. Marine Biology 151:1606-1614 Bluvias JE. 2008. Marine Turtle Trauma Response Procedures: A Husbandry Manual.

15 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 Duke University, USA: Nicholas School of the Environment and Earth Sciences. 75 pages. BODC, British Oceanographic Data Centre (2012). www.bodc.ac.uk (accessed 3 November 2012). Bolten AB, Crowder LB, Dodd MG, MacPherson SL, Musick JA, Schroeder BA, et al. 2010. Quantifying multiple threats to endangered species: an example from loggerhead sea turtles. Frontiers in Ecology and the Environment 9(5):295-301. Bonjean F, Lagerloef GSE. 2002. Diagnostic Model and Analysis of the Surface Currents in the Tropical Pacific Ocean, Journal of Physical Oceanography 32:2938 2954. Boström JE, Åkesson S, Alerstam T. 2012. Where on earth can animals use a geomagnetic bi-coordinate map for navigation? Ecography 35:1039-1047. Broderick AC, Glen F, Godley BJ, Hays GC. 2003. Variation in reproductive output of marine turtles. Journal of Experimental Marine Biology and Ecology 288(1):95-109. Buskirk SW, Millspaugh JJ. 2006. Metrics for Studies of Resource Selection. Journal of Wildlife Management 70(2):358-366. Calenge C. 2006. The package adehabitat for the R software: a tool for the analysis of space and habitat use by animals. Ecological Modelling 197: 516-519. Cardona L, Fernández G, Revelles M, Aguilar A. 2012. Readaptation to the wild of rehabilitated loggerhead sea turtles (Caretta caretta) assessed by satellite telemetry. Aquatic Conservation: Marine and Freshwater Ecosystems 22:104 112. Casale P, Conte N, Freggi D, Cioni C, Argano R. 2011. Age and growth determination by skeletochronology in loggerhead sea turtles (Caretta caretta) from the Mediterranean

16 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 Sea. Scientia Marina 75(1):197-203. Chapman JW, Klaassen RH, Drake VA, Fossette S, Hays GC, Metcalfe JD, et al. 2011. Animal orientation strategies for movement in flows. Current Biology 21(20):R861- R870. Collard SB, Ogren LH. 1990. Dispersal scenarios for pelagic post-hatchling sea turtle. Bulletin of Marine Science 4701: 233-243 Coyne M, Godfrey M, Godley B, Lay K. 2008. Hard Shell Sea Turtle PTT Attachment Protocol. p.8. http://www.seaturtle.org/documents/ptt_attachment_protocol.pdf (accessed 26 March 2013). Coyne MS, Goldley BJ. 2005. Satellite Tracking and Analysis Tool (STAT): an integrated system for archiving, analyzing and mapping animal tracking data. Marine Ecology Progress Series 301:1-7. Dodd CK, Byles R. 2003. Post-nesting movements and behavior of loggerhead sea turtles (Caretta caretta) departing from East-Central Florida nesting beaches. Chelonian Conservation and Biology 4:530 536. Donlan C, Wingfield DK, Crowder LB, Wilcox C. 2010. Using expert opinion surveys to rank threats to endangered species: a case study with sea turtles. Conservation Biology 24(6):1586-1595. Epperly SP, Braun J, Chester AJ, Cross FA, Merriner JV, Tester PA, et al. 1996. Beach strandings as an indicator of at-sea mortality of sea turtles. Bulletin of Marine Science 59(2):289-297. Feck AD, Hamann M. 2003. Effect of sea turtle rehabilitation centres in Queensland, Australia, on people s perceptions of conservation. Endangered Species Research. 20:153-165.

17 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 Frazer NB. 1986. Survival from Egg to Adulthood in a Declining Population of Loggerhead Turtles, Caretta caretta. Herpetologica 42(1):47-55. Freitas C, Lydersen C, Fedak MA, Kovacs KM. 2008. A simple new algorithm to filter marine mammal Argos locations. Marine Mammal Science 24(2):315-325. Fretey J (2001) Biogeography and Conservation of Marine Turtles of the Atlantic Coast of Africa/Biogéographie et conservation des tortues marines de la côte Atlantique de l'afrique. CMS Technical Series Publication No. 6, UNEP/CMS Secretariat, Bonn, Germany. GEBCO (2012). General Bathymetric Chart of the Oceans, Gridded bathymetry data. http://www.gebco.net/data_and_products/gridded_bathymetry_data/ (accessed 1 February 2013). Girard C, Sudre J, Benhamou S, Roos D, Luschi P. 2006. Homing in green turtles Chelonia mydas: oceanic currents act as a constraint rather than as an information source. Marine Ecology Progress Series 322:281-289. Godley BJ, Barbosa C, Bruford M, Broderick AC, Catry P, Coyne MS, et al. 2010. Unravelling migratory connectivity in marine turtles using multiple methods. Journal of Applied Ecology 47:769-778. Godley BJ, Blumenthal JM, Broderick AC, Coyne MS, Godfrey MH, Hawkes LA, et al. 2008. Satellite tracking of sea turtles: Where have we been and where do we go next. Endangered Species Research 4(1-2):3-22. Godley BJ, Richardson S, Broderick AC, Coyne MS, Glen F, Hays GC. 2002. Longterm satellite telemetry of the movements and habitat utilisation by green turtles in the Mediterranean. Ecography 25:352-362. Goshe LR, Avens L, Scharf FS, Southwood, A. 2010. Estimation of age at maturation

18 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 and growth of Atlantic green turtles (Chelonia mydas) using skeletochronology. Marine Biology 157:1725-1740. GSHHS, Global Self-consistent Hierarchical High-resolution Shoreline (2012). http://www.ngdc.noaa.gov/mgg /shorelines/gshhs.html. (accessed 14 January 2013) Hawkes LA, Broderick AC, Coyne MS, Godfrey MH, Godley BJ. 2007. Only some like it hot - quantifying the environmental niche of the loggerhead sea turtle. Diversity and Distributions 13:447 457. Hays GC, Bradshaw CJA, James MC, Lovell P, Sims DW. 2007. Why do Argos satellite tags deployed on marine animals stop transmitting? Journal of Experimental Marine Biology and Ecology 349(1):52-60. Hays GC, Dray M, Quaife, T, Smyth TJ, Mironnet NC, Luschi P, et al. 2001. Movements of migrating green turtles in relation to AVHRR derived sea surface temperature. International Journal of Remote Sensing 22 (8):1403-1411. Hays GC, Scott R. 2013. Global patterns for upper ceilings on migration distance in sea turtles and comparisons with fish, birds and mammals. Functional Ecology 27(3):748 756. Hijmans RJ, van Etten J. 2012. raster: Geographic analysis and modeling with raster data. R package version 2.0-12. http://cran.r-project.org/package=raster (accessed 28 March 2013). Computer program. Hochscheid S, Bentivegna F, Hamza A, Hays GC. 2010. When surfacers do not dive: multiple significance of extended surface times in marine turtle. Journal of Experimental Biology 213:1328-1337. Kobayashi DR, Polovina JJ, Parker DM, Kamezaki N, Cheng I, Uchida I, et al. 2008. Pelagic habitat characterization of loggerhead sea turtles, Caretta caretta, in the North

19 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 Pacific Ocean (1997-2006): Insights from satellite tag tracking and remotely sensed data. Journal of Experimental Marine Biology and Ecology 356: 96-114. Lewin-Koh NJ, Bivand R, Pebesma EJ, Archer E, Baddeley A, Bibiko H, et al. 2012. maptools: Tools for reading and handling spatial objects. R package version 0.8-20. http://cran.r-project.org/package=maptools (accessed 28 March 2013) Computer program. Lohmann KJ, Lohmann CMF. 1994. Detection of magnetic inclination angle by sea turtles: a possible mechanism for determining latitude. Journal of Experimental Biology 194:23-32. Lohmann KJ, Lohmann CMF. 1996. Detection of magnetic field intensity by sea turtles. Nature 380(7):59-61. Lohmann KJ, Lohmann CMF, Ehrhart LM, Bagley DA, Swing T. 2004. Geomagnetic map used in sea-turtle navigation. Nature 428:909-910. Lohmann KJ, Lohmann CMF, Putman NF. 2007. Magnetic maps in animals: nature s GPS. The Journal of Experimental Biology 210:3697-3705. Luschi P, Benhamou S, Girard C, Ciccione S, Roos D, Sudre J, et al. 2007. Marine Turtles Use Geomagnetic Cues during Open-Sea Homing. Current Biology 17(2):126-133. Luschi P, Hays GC, Del Seppia C, Marsh, R, Papi F. 1998. The navigational feats of green sea turtles migrating from Ascension Island investigated by satellite telemetry. Proceedings of the Royal Society of London B 265:2279-2284. Luschi P, Hays GC, Papi F. 2003. A review of long-distance movements by marine turtles, and the possible role of ocean current. Oikos 103:293 302.

20 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 Marcovaldi MA, Dei Marcovaldi GG. 1999. Marine turtles of Brazil: the history and structure of Projeto TAMAR-IBAMA. Biological Conservation 91(1),35-41. McDermott AJ, Seminoff J.A, Jones TT, Resendiz A. 2003. Food intake and retention time in green turtles (Chelonia mydas) from the Gulf of California: preliminary development of a digestive model. Proceedings of the Twenty-third Annual Symposium on Sea Turtle Biology and Conservation. 17 to 21 March, 2003, Kuala Lumpur, Malaysia: 17-20 Monzón-Argüello C, Rico C, Carreras C, Calabuig P, Marco A, López-Jurado LF. 2009. Variation in spatial distribution of juvenile loggerhead turtles in the eastern Atlantic and western Mediterranean Sea. Journal of Experimental Marine Biology and Ecology 373(2):79-86. Moore M, Early G, Touhey K, Barco S, Gulland F, Wells R. 2007. Rehabilitation and release of marine mammals in the United States: risks and benefits. Marine Mammal Science 23:731 750. OSCAR, Ocean Surface Current Analysis (2012). www.oscar.noaa.gov (accessed 5 November 2012). Pike DA. 2013. Climate influences the global distribution of sea turtle nesting. Global Ecology and Biogeography 22:555 566. Polovina JJ, Howell E, Kobayashi, DR, Seki MP. 2001. The transition zone chlorophyll front, a dynamic global feature defining migration and forage habitat for marine resources. Progress in Oceanography 49:469 483. Powell RA. 2000. Animal home range and territories and home range estimators. In: Boitani L, Fuller TK, editors. Research Techniques in Animal Ecology. New York: Columbia University Press, p 65-110.

21 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 Putman NF, Endres CS, Lohmann CMF, Lohmann KJ. 2011. Longitude perception and bicoordinate magnetic maps in sea turtles. Current Biology 21(6):463-466. Quantum GIS Development Team. 2013. Quantum GIS Geographic Information System. Open Source Geospatial Foundation Project. http://qgis.osgeo.org. Computer program. R Core Team. 2012. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0, http://www.r-project.org. Computer program. Santos AS, Almeida AP, Santos AJB, Gallo B; Giffoni B, Baptistotte C, et al. 2011. Plano de Ação Nacional para Conservação das Tartarugas Marinhas. Série Espécies Ameaçadas nº 25. Instituto Chico Mendes de Conservação da Biodiversidade, ICMbio. (in Portuguese). Sasso CE, Epperly SP. 2007. Survival of pelagic loggerhead turtles in the open ocean. Journal of Wildlife Management 71:1830 1835. Snoddy JE, Southwood A. 2010. Movements and post-release mortality of juvenile sea turtles released from gillnets in the lower Cape Fear River, North Carolina, USA. Endangered Species Research 12:235 247. Strauss RE. 1979. Reliability estimates for Ivlev's electivity index, the forage ratio, and a proposed linear index of food selection. Transactions of the American Fisheries Society 108:344 352. Swimmer Y, Arauz R, McCracken M, McNaughton L, Ballestero J, Musyl M, et al. 2006. Diving behavior and delayed mortality of olive ridley sea turtles Lepidochelys olivacea after their release from longline fishing gear. Marine Ecology Progress Series 323:253 261.

22 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 Troëng S, Evans DR, Harrison E, Lagueux CJ. 2005. Migration of green turtles Chelonia mydas from Tortuguero, Costa Rica. Marine Biology 148(2):435-447. Wallace BP, DiMatteo AD, Bolten AB, Chaloupka MW, Hutchinson BJ, Abreu- Grobois FA, et al. 2011. Global Conservation Priorities for Marine Turtles. PLOS One, 6(9): e24510. 14 pages. Watson KP, Granger RA. 1998. Hydrodynamic effect of a satellite transmitter on a juvenile green turtle (Chelonia mydas). Journal of Experimental Biology 201:2497 2505. Wessel P, Smith WHF. 1996. A global, self-consistent, hierarchical, high-resolution shoreline database. Journal of Geophysical Research 101(B4):8741 8743. Witt MJ, Hawkes LA, Godfrey MH, Godley BJ, Broderick AC. 2010. Predicting the impacts of climate change on a globally distributed species: the case of the loggerhead turtle. Journal of Experimental Biology 213:901-911. 672 673 674 675 676 677 678 679 680

23 681 682 683 684 685 Table 1 - Number of locations for each animal (% percentage) (CC Caretta caretta; CM1 and CM2 - Chelonia mydas) and error class: 3 (<250 m); 2 (250-500 m); 1 (500-1500 m); 0 (>1500 m); A (no error estimation, 3 uplinks); B (no error estimation, 1-2 uplinks) and Z (invalid locations). Error Class CC (%) CM1 (%) CM2 (%) 3 659 (11,69) 57 (2,52) 295 (11,20) 2 1313 (23,28) 112 (4,96) 419 (15,91) 1 887 (15,73) 125 (5,53) 360 (13,67) 0 509 (9,03) 87 (3,85) 190 (7,21) A 856 (15,18) 461 (20,41) 562 (21,34) B 1367 (24,24) 1386 (61,35) 783 (29,73) Z 48 (0,85) 31 (1,37) 25 (0,95) 686 687 688 689 690 691 692 693 694 695 696 697 698

24 699 700 701 702 Table 2 Influence of environmental variables in the paths followed by CM1; Mean (±SD); Range range of values in the used area; coefficient of the Linear Model (LM); Kolmogorov-Smirnov's D (KS-D); Circular correlation coefficient (corr - r); NS - non significant relation; * - significant relation (p<0.05). Variable Mean (±SD) Range Bathymetry -986.919 m (± 1328.181) SST 21.691ºC (± 1.482) Chlorophyll 2.498 mg m -3 (± 4.303) -4388.0 m to - 1.0 m 17.61ºC to 27.05ºC 0.0 mg m -3 to 56.110 mg m - 3 Statistic Significance LM KS-D corr r - 0,588 - * - 0,500 - NS - 0,194 - NS 703 704 705 706 707 708 709 710 711 712 Inclination Declination Intensity 31.671º (± 6.775) -20.372º (± 0.081) 39757.960 nt (± 1898.371) 16.93º to 46.12º -20.54º to - 20.26º 36440 nt to 42440 nt Current speed 0.100 ms -1 0.027 to 0.204-0,628 - * - 0,929 - * - 0,714 - * ms -1 0.025 - - NS Current direction - - - - 0.061 NS Surface wind speed Surface wind direction 5.267 ms -1 1.400 to 7.657 ms -1-0.021 - - NS - - - - 0.133 NS

25 713 714 715 716 Table 3 Influence of environmental variables in the paths followed by CM2; Mean (±SD); Range range of values in the used area; coefficient of the Linear Model (LM); Kolmogorov-Smirnov's D (KS-D); Circular correlation coefficient (corr - r); NS - non significant relation; * - significant relation (p<0.05). 717 718 Variable Mean (±SD) Range Bathymetry SST Chlorophyll Inclination Declination Intensity -2988.544 m (± 1656.854) 25.185ºC (± 2.894) 0.455 mg m -3 (± 1.060) 8.829º (± 17.960) -20.586º (± 0.205) 34741.95 nt (± 3939.630) -5612 m to -2 m 19.46ºC to 30.20ºC 0.0 mg m -3 to 12.850 mg Statistic LM KS-D corr r Significance - 0,424 - * - 0,200 - NS m -3-0,095 - NS -15.770º to 46.280º -21.24º to - 20.05º 27300 nt to 42980 nt Current speed 0.114 ms -1 0.026 to - 0,293 - * - 0,462 - NS - 0,293 - * 0.392 ms -1-0.207 - - NS Current direction - - - - 0.012 NS Surface wind speed Surface wind direction 6.170 ms -1 4.394 to 8.118 ms -1-0.045 - - NS - - - - 0.134 * 719 720 721 722 723 724

26 725 726 727 728 Table 4 Influence of environmental variables in the paths followed by CC; Mean (±SD); Range range of values in the used area; coefficient of the Linear Model (LM); Kolmogorov-Smirnov's D (KS-D); Circular correlation coefficient (corr - r); NS - non significant relation; * - significant relation (p<0.05). 729 Variable Mean (±SD) Range Bathymetry SST Chlorophyll Inclination Declination Intensity -4679.162 m (± 1052.146) 24.141ºC (± 3.267) 0.063 mg m -3 (± 0.037) 37.013º (± 3.331) -19.674º (± 0.912) 40087.49 nt (± 1146.686) -6539 m to -2 m 19.09ºC to 30.75ºC 0.0 mg m -3 to Statistic LM KS-D corr r Significance - 0,382 - * - 0,391 - * 0.427 mg m -3-0,333 - NS 31.35º to 48.69º -20.36º to - 16.84º 38090 nt to 42790 nt Current speed 0.080 ms -1 0.006 to 0.328-0,775 - * - 0,325 - * - 0,436 - * ms -1-0.138 - - NS Current direction - - - - 0.328 * Surface wind speed Surface wind direction 6.748 ms -1 4.564 to 8.605 ms -1 0.011 - - NS - - - - 0.094 * 730 731 732 733 734 735 736

27 737 Figure 1 - Track of the three specimens followed in this study. 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754

28 755 Figure 2 Monthly number of locations for each turtle. CC, CM1 and CM2. 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772

29 773 Figure 3 - Strauss' Linear Index for the significant variables. 774 775

30 776 777 Figure 4 - Geomagnetic Field (A) Inclination, (B) Intensity and (C) Declination (World Magnetic Model for 2010/2015). 778