Mar Biol (2007) 152:1039–1049 DOI 10.1007/s00227-007-0752-7

R ES EA R C H A R TI CLE

A model of area Wdelity, nomadism, and distribution patterns of loggerhead sea turtles (Caretta caretta) in the Mediterranean Sea Paolo Casale · Daniela Freggi · Roberto Basso · Carola Vallini · Roberto Argano

Received: 14 April 2007 / Accepted: 14 June 2007 / Published online: 10 July 2007 © Springer-Verlag 2007

Abstract Sea turtle tagging carried out in Italy in the period 1981–2006 resulted in 125 re-encounters of loggerhead turtles (Caretta caretta) after a mean of 2.5 years, from diVerent marine areas in the Mediterranean. At Wrst Wnding, turtles ranged 25–83 cm of curved carapace length. Data were analyzed according to size, area, habitat type, season, in order to provide indication of movement patterns. When integrated with other information, results indicate that: (1) a part of turtles in the oceanic stage show a nomad behavior with movements among diVerent oceanic areas; (2) another part show Wdelity to an oceanic area; (3) turtles in the neritic stage show Wdelity to neritic areas, and once settled to one area, change to other neritic areas is unlikely; (4) nomad oceanic turtles are signiWcantly larger than sedentary ones, and also larger than turtles found in neritic areas; it is hypothesized that these could be Atlantic turtles that eventually leave the Mediterranean; (5) ecological transition from oceanic to neritic habitats occurs at a wide range of sizes, and some turtles may have a very brief oceanic stage; (6) turtles in the oceanic stage are more Communicated by R. Cattaneo-Vietti. P. Casale (&) · R. Argano Dipartimento di Biologia Animale e dell’Uomo, Università “La Sapienza”, 00185 Rome, Italy e-mail: [email protected] D. Freggi Sea Turtle Rescue Centre WWF Italy, 92010 Lampedusa, Italy R. Basso Museo Civico di Storia Naturale di Jesolo, Via BaWle 172, 30017 Jesolo Lido, Italy C. Vallini A.R.C.H.E’, via Mulinetto, 40/A, 44100 Ferrara, Italy

likely to recruit to neritic areas close to their oceanic areas than to distant ones; (7) part of turtles from some Mediterranean nesting beaches might frequent a relatively limited area range, including both oceanic and neritic areas; (8) in most of the Mediterranean, latitudinal seasonal migrations are unlikely. A general model of movement patterns of loggerhead turtles in the Mediterranean is proposed.

Introduction Most sea turtle species move great distances and analyzing these movement is of interest for a variety of reasons. It may show key areas for conservation (Margaritoulis et al. 2003); give clues on habitat use (Eckert 2006); reveal the mechanisms underlying navigation (Luschi et al. 2001); help understanding potential eVects of climate change (McMahon and Hays 2006). Loggerhead sea turtles (Caretta caretta) undergo two ecological stages, frequenting diVerent habitats and covering vast areas during their life. First they prey upon epipelagic animals in the oceanic zone (Bjorndal 1997; Bolten 2003) and in this stage they can disperse through a wide area, crossing entire oceans like the Atlantic and the PaciWc, where their distribution is aVected by oceanographic features like gyre systems and oceanic fronts (Musick and Limpus 1997; Polovina et al. 2000), so that their dispersal can be broadly inferred from oceanographic particle tracking models and buoy trajectories (Hays and Marsh 1997). In some cases, they may remain in a certain area, like the one around the Azores (Bolten 2003). Then, as larger juveniles, they recruit to neritic habitats where they mainly prey upon benthic animals (Bjorndal 1997). There are indications from diVerent populations that these neritic juveniles show Wdelity to speciWc areas (Musick and Limpus 1997; Avens and Lohman 2000; Limpus 1994). The adults show Wdelity to their

123

1040

neritic feeding grounds too (Papi et al. 1997; Schroeder et al. 2003), which are probably the same ones to which they recruited as juveniles (Limpus and Limpus 2001; Limpus 1994). Adults also show strong Wdelity to the nesting areas, which are the same where they were born (natal homing) (Bowen et al. 1993). However, there can be a great variability in this general pattern; for instance the oceanic stage can have diVerent duration depending on the population or habitat characteristics (Bolten 2003) and in some areas, like the north-western Atlantic, there is evidence of seasonal migrations (Musick and Limpus 1997). The Mediterranean basin is rather small if compared with oceans, but it hosts oceanic and neritic habitats for the Mediterranean loggerhead turtle populations and at least oceanic habitats for individuals from the Atlantic too (see Margaritoulis et al. 2003, for a review). Given the very low genetic contribution to Mediterranean populations (Laurent et al. 1998; Carreras et al. 2006a), these Atlantic specimens probably return to the Atlantic to breed. This suggests a complex situation of movements due to several factors such as dispersal, ecological transition, seasonal migration, homing, and reproductive migration. Identifying these patterns of movements of turtles at sea is fundamental for reconstructing how these animals use the Mediterranean basin. This is not only important for a better understanding of sea turtle ecology, but also for a desirable conservation management of marine areas. Although in recent times satellite tracking technology improved greatly and is providing more and more information on habitat use and migratory pathways, it still has two limits that prevent to obtain clear movement patterns, such as area Wdelity and ecological transition. One limit is the time a transmitter can remain attached to the carapace, especially in young and fast-growing turtles. For instance, in a sample of 42 green and loggerhead turtles the average duration was 117 days (Hays et al. 2003), although individual cases of long tracking are possible (e.g., in the Mediterranean 420– 450 days; Bentivegna 2002; Cardona et al. 2005). The other one is the limited number of turtles that can be tracked, due to the high cost of the transmitters. Another way to follow individual turtles is the classic Capture-Mark-Recapture approach (CMR) with Xipper tags, used since the 1950s (Hendrickson 1958). However, also in this case it diYcult to obtain the necessary sample size useful to give indications about patterns, for the low rates of re-encounter. Moreover, tagging programs have been traditionally carried out on adult females tagged while nesting, while fewer have tagged juveniles, that require a more logistically diYcult sampling at sea, and even less have studied juveniles in the oceanic stage. Another important factor is the chance to receive tag returns from distant areas, which depends on several variables such as human presence, language, attitude, etc.

123

Mar Biol (2007) 152:1039–1049

The Wrst tagging program on turtles at sea in the Mediterranean was launched in Italy in 1981, and was particularly focused on the oceanic stage. This program provided Wrst insights on juvenile movements and trophic areas (Argano et al. 1992), to be compared with the results of a tagging program on adult females nesting in Greece (Margaritoulis 1988; Margaritoulis et al. 2003). Italy is particularly suitable for CMR studies because it is a large peninsula protruding in the center of the Mediterranean and so it borders diVerent marine areas and habitats. This alone enhances greatly both tagging opportunities and chances of tag returns. This study aims to improve the present knowledge of loggerhead turtle distribution in the Mediterranean, through the identiWcation of movement patterns, thanks to a large data set of tag returns of juveniles found at sea.

Materials and methods To identify long-term movement patterns in sea turtles is a very diYcult task, because although Xipper tags can last for long, they can only provide information about two moments: release and re-encounter. For this reason, the approach used in this study was to set arbitrary classes of conditions (e.g., habitats; Wdelity/non-Wdelity) and to assign each individual turtle record to the most probable of the possible classes. As a consequence, it was assumed that the majority, but not necessarily the total, of turtles included in a certain class were really in the condition described by that class. So, possible patterns were investigated by a statistical comparison among classes, in spite of the intrinsic uncertainty at the individual level. The Mediterranean was divided in the following areas (Fig. 1): North (NT) and South Tyrrhenian (ST), Central Mediterranean (CM), North (NI) and South Ionian (SI), North (NA) and South Adriatic (SA), West (WM), and East Mediterranean (EM). The year was divided into two seasons: warm (April–September) and Cold (October–March). Because of lack of information on date of Wnding in tag returns, it was not possible to assign nine turtles to a season. Loggerhead turtles can swim faster than 20 km/h (e.g., Renaud and Carpenter 1994) and an average speed higher than 70 km/day (2.9 km/h) for a period of 11 days was observed (Meylan et al. 1983), while 1 km/h can be considered as a conservative cruise speed (e.g., in the Mediterranean: Bentivegna 2002; Cardona et al. 2005; Godley et al. 2003a). So, in order to investigate possible movement patterns, and particularly to recognize area Wdelity versus change of area, only turtles re-encountered more than 30 days after release were considered, assuming that this is enough time for a non-sedentary turtle to move to an area diVerent of that of release (Fig. 1). In other words, if after 30 days a turtle was found in the same area, the turtle was

Mar Biol (2007) 152:1039–1049

1041

Fig. 1 Study areas. NA, SA north and south Adriatic, NI, SI north and south Ionian, NT, ST north and south Tyrrhenian, WM, CM, EM western, central and eastern Mediterranean

considered more likely to have Wdelity to the area than the opposite. The aim of this study is to recognize diVerent movement patterns on the Mediterranean scale: movement versus apparent lack of movement (even though within the capability of the turtle). Therefore, the term “Wdelity” should be considered in association with the areas deWned in Fig. 1, and not as Wdelity to very small marine spots, that could even be the case but was not the target of this study. A total of 125 loggerhead turtles tagged and released from diVerent sites, basically around Italy, and eventually reencountered after more then 30 days in Italy and other Mediterranean countries in the period 1981–2006 were considered. Most of these turtles were tagged by three Italian sea turtle tagging programs (University of Roma 1 “La Sapienza”; WWF Italy; ARCHE’), including 34 records already reported by Argano et al. (1992). Four additional records were included from literature (Gramentz 1991; Tomas et al. 2001). Turtles were originally captured by Wshing gears, found stranded, gathered while Xoating at sea, or found while nesting. The exact spot in the sea that these turtles frequented was unknown, and the nesting beach, stranding beach, or port where turtles were found or reported, was considered the place on land which was closest to their marine area. Curved carapace length notch-to-tip (CCLn-t; Bolten 1999) was measured at release while just in some cases it was measured at re-encounter too, because who found them was often not part of a research program (e.g., Wshermen, tourist, etc.). Turtles were tagged on the front Xippers with monel or inconel tags, styles 49 or 681 (National Band and Tag, Newport, KY, USA), stamped with an alphanumeric code and a return address. It is assumed that turtles may frequent a speciWc area for the following reasons: favorable habitat characteristics (trophic or abiotic), migratory pathway, just visit by chance, reproduction. Since the aim of this study is to detect patterns of possible Wdelity versus possible change of area, the approach used was to conservatively consider any alternative explanation as more likely than Wdelity or change. Particular

attention was given to the confounding eVect of reproduction. Specimens in the adult size range and found close to a nesting area were considered potentially in reproduction (category R or RM, see below). Greece, Libya, and the Central Mediterranean (Tunisia and Italy) were considered possible nesting areas (although some of them host just few nests per year), while all the other areas were considered as non-nesting areas. Turtles smaller than 70 cm CCLn-t were assumed to be immature. This is the minimum size recorded for a nesting female in the nearest nesting sites (Greece and Libya), which are also the most important ones in the Mediterranean (Margaritoulis et al. 2003). On the average, males mature at a larger size than 70 cm too (Casale et al. 2005a). When size at re-encounter was not available, turtles were considered to be smaller than 70 cm on the basis of the size at release and a growth rate of 2.1 cm/year (Casale et al. 2007a). Turtles re-encountered in the same area (Fig. 1) of release were suspected to show Wdelity, and the possible patterns, ranked in a conservative order against strong Wdelity, were: R (possible Reproduction), SF (Seasonal Fidelity, i.e., Wdelity to the area only during the warm or cold season), YRF (Year-Round Fidelity). Those re-encountered in a diVerent area were suspected to have changed area permanently, and the possible patterns, ranked in a conservative order against permanent change of area, were: RM (possible Reproductive Migration), SM (Seasonal Migration), C (Change of area). Criteria to assign these patters are described below and also shown in Table 1. Turtles considered potentially adult and in a nesting area were classiWed as R/SF/YRF (i.e., any of these patterns are possible) if they were re-encountered in the same area, and as RM/SM/C if in a diVerent area. The other turtles (not considered potentially adult and not in a nesting area) were classiWed according to the area and season. Those reencountered in the same area were classiWed as SF/YRF (i.e., the turtle frequents that area at least in a season but possibly also all year round) if re-encountered in the same or unknown season and as YRF if in a diVerent season. Those re-encountered in a diVerent area were classiWed as

123

1042

Mar Biol (2007) 152:1039–1049

Table 1 Criteria, movement patterns, and habitat type at Wrst and second Wnding (see text) Criteria

Movement pattern

Habitat type O–O

O–N

Total O–?

N–N

N–O

N–?

?–O

?–N

?–?

Adult size and possible nesting area Same area

R

SF

YRF

DiVerent area

RM

SM

C

Same or unknown season

SF

YRF

DiVerent season

YRF

3

1

1

2

20

3

1

8

2

8

8

1

3

14

13

3

1

4 3

Not adult size or not nesting area Same area 15

1

1

11

2

43

1

22

DiVerent area DiVerent or unknown season

SM

Same season

C

C

1

1

22 31

Total

125

Fidelity

SF + YRF

28

5

1

26

0

1

1

0

3

65

Change

SM + C

22

21

4

0

4

1

0

1

0

53

Total

118

R possible reproduction, SF at least seasonal Wdelity, YRF year-round Wdelity, RM possible reproductive migration, SM possible seasonal migration, C change of area, O oceanic, N neritic, ? uncertain

SM/C (i.e., the turtle at least moved from the Wrst area temporarily as a seasonal migration but possibly even permanently) if re-encountered in a diVerent or unknown season and as C if in the same season. Since loggerhead turtles pass through two diVerent ecological stages (oceanic and neritic) and this may determine their choice of areas/habitats, encounter events were assigned to oceanic/neritic habitats according to the following criteria. A turtle was considered more likely to frequent an oceanic than a neritic habitat if found either in the oceanic zone of the Central Mediterranean, or captured by drifting longline (which Wsh in oceanic zones), or found in the following areas where the oceanic zone is predominant: Tyrrhenian, Ionian, South Adriatic, West Mediterranean. However, if in these areas the turtle was found on the continental shelf or it was captured by bottom trawlers that Wsh on the continental shelf, the turtle was considered to frequent a neritic habitat. A turtle was considered more likely to frequent a neritic than an oceanic habitat if captured by bottom trawlers or by bottom longline (both Wshing on the continental shelf) or if found in the North Adriatic (continental shelf) and in the continental shelves of the Central and East Mediterranean. All the other cases (e.g., found Xoating or stranded in the central Mediterranean which has both oceanic and neritic zones) were considered as uncertain.

Results The 125 turtles considered were re-encountered after a mean of 2.5 years, ranging from 31 days to 14.9 years,

123

however 89% of them were re-encountered after more than 90 days. At release, they ranged between 25 and 83 cm CCLn-t (mean 53.9; SD 12.7; n = 114). All turtles were assigned to a pattern of movement (n = 125; Table 1). However, it should be borne in mind that such a classiWcation has an intrinsic degree of uncertainty, so that attention should be focused more on the emerging general patterns rather than on individual turtles. Seven turtles were not included in further analysis because the eVect of reproductive behavior could not be excluded (R/SF/YRF; RM/SM/C) (Table 1). Of the remaining 118 turtles, 65 were re-encountered in the same area of release and showed either a year-round Wdelity (YRF; n = 22) or at least a seasonal Wdelity (SF/ YRF; n = 43) (Tables 1, 2). The 53 turtles re-encountered in a diVerent area, showed either a change of area (C; n = 31) or at least a seasonal change (SM/C; n = 22) (Table 1; Figs. 2, 3, 4). However, in Table 2 Areas where Wdelity patterns (see Table 1 and text) were observed Area

SF/YRF

YRF

TOT

N. Adriatic

10

3

13

S. Adriatic/N Ionian

15

3

18

C. Mediterranean

14

12

26

N Tyrrhenian

1

2

3

S. Tyrrhenian

1

1

2

W. Mediterranean

2

1

3

43

22

65

Total

Mar Biol (2007) 152:1039–1049

1043

Fig. 2 Release (circles) and re-encounter places (triangles) of turtles moving between diVerent oceanic areas (n = 22). Grades of shades represent bathymetry

Fig. 3 Release (circles) and re-encounter places (triangles) of turtles moving from an oceanic to a neritic area (n = 26), including Wve doing so within the central Mediterranean

Fig. 4 Release (circles) and re-encounter places (triangles) of turtles moving between diVerent areas, with patterns O–?, N–O, N–?, ?–N (see Table 1; n = 10)

most individuals of the latter group the change of area was unlikely due to seasonal migration for climatic reasons, because no evident patterns South/Cold–North/Warm was evident (Fig. 5); so that the SM/C group is likely to represent a Change of area actually. Although all combinations of release-re-encounter habitat types (Oceanic and Neritic) were observed (O–O; O–N; N–O; N–N), they were not equally represented between the two groups of Fidelity and Change of area, and N–O was particularly underrepresented (Table 1). Both Fidelity and

Change groups included specimens released and re-encountered in oceanic habitats (O–O) but the N–N case was observed in the Fidelity group only, while more turtles of the Change group showed the O–N or O–? pattern. In other words, while about the same number of turtles showing Wdelity were Wrst encountered in an Oceanic (n = 33) and Neritic habitats (n = 27), most turtles showing a possible change of area were Wrst encountered in an Oceanic habitat (n = 47), while just Wve in a Neritic one (Table 1; Fisher exact test; P < 0.001; n = 112). The O–N pattern has an

123

1044

Mar Biol (2007) 152:1039–1049

Fig. 5 Release (circles) and re-encounter places (triangles) of turtles released in the cold season and re-encountered in the warm one (open symbols; n = 5) and the opposite (black symbols; n = 14)

important biological meaning because it may represent the shift between the oceanic and neritic ecological phases of loggerhead turtles, and the speciWc areas involved are shown in Fig. 3 and Table 3. The mean size at Wrst encounter of the Fidelity and Change groups was 51.4 cm (SD 12.6; range 25–76.5; n = 60) and 55.1 (SD 11.6; range 27–83; n = 47) respectively, with no signiWcant diVerence (Mann–Whitney Utest; n = 107). No signiWcant diVerence in size was observed among O–O, O–N, and N–N groups (Kruskall– Wallis test; n = 93) (Fig. 6a), while within the O–O group, turtles showing Fidelity were smaller than those showing Change (Mann–Whitney U-test; P < 0.05; n = 44) (Fig. 6b).

Discussion Sea turtle species diVer in diet and as a consequence they frequent diVerent habitats at diVerent periods of their life. For instance, the leatherback turtle (Dermochelys coriacea) moves across pelagig habitats searching for gelatinous zooplankton (Eckert 2006); the green turtle (Chelonia mydas) is basically herbivorous and its range is limited to coastal foraging pastures (Godley et al. 2003b). Loggerhead turtles feed upon a broad variety of animals that they found in a wide range of habitats, and they pass through two ecological phases (Bjorndal 1997; Bolten 2003): Wrst an oceanic,

Table 3 Areas where movement from an oceanic to a neritic area was observed

Oceanic habitat

Neritic habitat N. Adriatic

S. Adriatic

2

N. Ionian

5

S. Ionian

1

C. Mediterranean

1

S. Tyrrhenian Total

123

then a neritic one. Accordingly, three main patterns were observed in this study: O–O (turtles assumed to be in the oceanic phase), O–N (turtles assumed to have changed phase, from Oceanic to Neritic), and N–N (turtles assumed to be in the neritic phase). Just few turtles showed a N–O pattern and no one showed a N–N pattern of change of area (Table 1). This strongly suggests that, in the Mediterranean, turtles switch to the neritic phase in a certain area and then use it as foraging ground for the rest of the life, without moving to other neritic areas. This ecological transition probably occurs at a wide range of sizes, as suggested by the lack of any signiWcant diVerence among the O–O, O–N, and N–N groups and by the small size of specimens probably in the neritic phase observed in this study (min: 28 cm; Fig. 6a) as well as in other studies in the North Adriatic (min: 29.5 cm; n = 61, Casale et al. 2004) and in the Central Mediterranean (min: 22 cm; n = 368; Casale et al. 2007b). Turtles in the oceanic phase (O–O) seem to display two diVerent behaviors, Wdelity to an area and change of area, and the two groups diVer in size signiWcantly (Fig. 6b). Two scenarios are proposed to explain these results. In the Wrst one, turtles pass through Wve stages in the Mediterranean: (a) a relatively brief post-hatching dispersal from nesting beaches to oceanic habitats, (b) Wdelity to an oceanic area, (c) “nomad stage” visiting diVerent oceanic areas, (d) settlement to a neritic area, (e) Wdelity to this neritic

Total C. Mediterranean

E. Mediterranean

W. Mediterranean

5

1

1

2 2

3

5

6

3 9

13

12

3 3

1

26

Mar Biol (2007) 152:1039–1049

Fig. 6 Size distribution of turtles released and re-encountered a in oceanic–oceanic (O–O), oceanic–neritic (O–N), and neritic–neritic (N–N) areas (n = 93); b in the same (O–O Fidelity) or a diVerent (O–O Change) oceanic area (n = 44)

area. The second scenario hypothesizes two groups of turtles having diVerent behavioral patterns, one with the four stages (a), (b), (d), and (e), while the other group would be “nomad” (c). Three considerations support the second scenario. First, turtles are assumed, also on the basis of the above results, to become neritic after the oceanic stage and not the opposite. For this reason, the large size of “nomad” turtles, with just few individuals smaller than 50 cm (O–O Change; Fig. 6b) does not Wt well with the size of turtles showing ecological transition or Wdelity to neritic areas, which have more individuals smaller than 50 cm (O–N and N–N; Fig. 6a; Fisher Exact test; P < 0.05; n = 68). This diVerence is the opposite of the expected one and does not support a scenario in which turtles go through a “nomad” stage before settling to neritic areas. Second, the Mediterranean is known to be frequented by at least two distant populations: the Mediterranean and the Atlantic one, the latter occurring in large numbers at least in the Western and Central Mediteranean (Laurent et al. 1998; Carreras et al. 2006b). Since genetic Xow from the Atlantic to the Mediterranean is estimated to be very low (Carreras et al. 2006a), these Atlantic turtles are supposed to return to the Atlantic to breed. Third, Atlantic turtles mature at a larger size (range of means of nesting females: 98.9–105.1 cm CCL; Dodd Jr 1988) than Mediterranean ones (range of means of nesting females: 66.5–84.7 cm CCL; Margaritoulis et al. 2003), and Bjorndal et al. (2000) estimated that most loggerhead

1045

turtles recruit to the demersal habitats in the southeastern USA between 46 and 64 cm CCL, which is similar to the range of the “nomad” turtles observed in this study. Accordingly, the maximum size recorded in the Mediterranean of a specimen with a mtDNA haplotype endemic of the Atlantic (Laurent et al. 1998) was 65 cm CCL (P. Casale, unpublished data). It can be hypothesized that the “nomad” turtles (O–O Change) are Atlantic ones which continuously move through diVerent oceanic areas across the Atlantic and the Mediterranean, and eventually leave the Mediterranean to reach their Wnal neritic areas in the Atlantic. However, although it is likely that most Atlantic specimens in the Mediterranean are in the oceanic stage (Laurent et al. 1998) it is possible that some of them frequent also neritic areas, as suggested by few specimens with a haplotype endemic of the Atlantic and caught by demersal Wshing gear in the Central Mediterranean (Laurent et al. 1998). This could explain the pattern N–O, and especially the turtle which moved from the North Adriatic to the Western Mediterranean (Fig. 4). DiVerently from Atlantic turtles, under this scenario Mediterranean turtles would show a general Wdelity pattern in both oceanic and neritic stages. They would show an early Wdelity to oceanic areas and they would also settle early in neritic areas. The size range of turtles found in a neritic area (N–N and O–N) (Fig. 6a) suggests that in some cases the oceanic stage can be very short, while in others it would be longer. Accordingly, dietary analyses found benthic prey in small specimens in neritic areas in the central Mediterranean (Abbate et al. 2007) and in the north Adriatic (Lazar et al. 2007). This indicates that in the Mediterranean loggerhead turtles recruit to neritic areas at a smaller size than in the north-western Atlantic (Bjorndal et al. 2000), the south-western (Limpus and Limpus 2003), and the north-eastern PaciWc (SeminoV et al. 2004). The observed Wdelity to oceanic areas can in part be due to diVerent small oceanic zones that can be found in the Mediterranean in comparison with open oceans. However, the diVerent patterns observed within the oceanic stage (O–O Fidelity and O–O Change; Fig. 6b) indicate that in addition to oceanographic factors, a behavioral component is involved in determining Wdelity or nomadism. An implication of this model is that large turtles found in oceanic areas not along reproductive migration pathways are more likely to be of Atlantic than Mediterranean origin. Seasonal migrations (north/south) probably due to temperature change are known from the north-western Atlantic (Musick and Limpus 1997), but do not seem to be a general pattern for all populations (e.g., Limpus and Limpus 2001). No evident north/south seasonal patters are

123

1046

shown by this study, with even counter-intuitive movements observed (Fig. 5): among four Cold-to-Warm movements, three are toward lower latitudes, while among 14 Warm-to-Cold movements, Wve are toward higher latitudes. In the western Mediterranean no clear seasonal patterns were observed by satellite tracking and aerial surveys (Cardona et al. 2005; Gomez de Segura et al. 2006), suggesting that previous reports based on incidental captures (Caminas and De La Serna 1995) could be aVected by Wshing eVort (Gomez de Segura et al. 2006). Although Bentivegna (2002) suggested a seasonal migration pattern for four satellite-tracked turtles, actually longitudinal movements were more evident than latitudinal ones, the direction of some movements were counter-intuitive, and it cannot be excluded that factors other than temperature aVected the observed movements. Moreover, it is interesting that high numbers of turtles are found in winter in the north Adriatic (Casale et al. 2004), which is the coldest area of the Mediterranean. The only evidence of a seasonal emigration is from a small area in the northernmost part of the North Adriatic (northern to 45°N), where temperatures in winter drop below 11–12°C (Lazar et al. 2003). All this considered, it seems likely that seasonal temperature variation alone does not determine strong seasonal migrations in loggerhead turtles in the Mediterranean, with the excepFig. 7 Main surface currents in the Mediterranean (modiWed from Malanotte-Rizzoli et al. 1997; Millot and Taupier-Letage 2004)

Fig. 8 A hypothetical pattern of frequented areas and movements. Dashed lines link natal sites and oceanic habitats (the bold dashed line considers a hypothetical oceanic habitat in the eastern Mediterranean). Continuous lines link oceanic and neritic habitats. Lines just link diVerent areas and should not be necessarily considered as speciWc routes

123

Mar Biol (2007) 152:1039–1049

tion of the northernmost part of the Adriatic. Accordingly, Hochscheid et al. (2005, 2007) recently observed that loggerhead turtles in the Mediterranean can undergo a state of dormancy to overcome the cold season, without the need of migrating to warmer areas. When integrated with the other information available, the above results help to understand how loggerhead turtles use some Mediterranean areas during their two ecological phases. Although oceanographic features, including main currents (Fig. 7), may aVect the distribution of turtles, and especially of small ones, and suggest routes, many other factors and small-scale oceanographic features are involved. Thus the proposed model (Fig. 8) concerns the general movement pattern among areas, rather than speciWc routes. Nesting sites are restricted to the eastern basin (Margaritoulis et al. 2003), but turtles born there spend the Wrst phase of their life (oceanic) not only in the eastern but also in the western basin, were they share oceanic habitats with comparable numbers of specimens of Atlantic origin (Laurent et al. 1998; Carreras et al. 2006b). The observed Wdelity pattern in the south Adriatic/north Ionian area, the central Mediterranean and the Tyrrhenian (Table 2) suggests that these areas are permanent oceanic habitats where turtles Wnd adequate trophic resources.

Mar Biol (2007) 152:1039–1049

The south Adriatic/north Ionian area was suggested to be a particularly important developing oceanic habitat on the basis of the very small specimens found in this area in comparison to other areas (Casale et al. 2005b). In this area, genetic studies found a high occurrence of turtles from the Greek nesting sites (Carreras et al. 2006b). In particular, Zakynthos gives a very important contribution in terms of nests at the Mediterranean scale (about 1,300 nests/year on a total of about 5,000 documented in the basin; Margaritoulis et al. 2003) and is in the eastern Ionian. So, it is possible that part of the turtles born in Zakynthos remain in the south Adriatic/north Ionian area for all their oceanic phase. This area is also probably frequented by turtles from the more distant Turkish nesting sites, and to a less extent also by turtles from the Atlantic (Carreras et al. 2006b; MaVucci et al. 2006). The Tyrrhenian is probably frequented by turtles from Greece, Atlantic, and Turkey, in order of relative contribution (Carreras et al. 2006b). Turtles from other nesting areas not adequately genetically assessed, like Libya, could frequent these areas as well. The situation in the oceanic part of the Central Mediterranean is still unclear due to methodological problems (Carreras et al. 2006b) but is certainly frequented by both Mediterranean and Atlantic turtles (Laurent et al. 1998). Another area which appears to be an important oceanic habitat, on the basis of the high interaction with drifting longline, is the westernmost part of the Mediterranean (Caminas et al. 2006), frequented by Atlantic and Mediterranean turtles (Carreras et al. 2006b). Other areas representing oceanic habitats for turtles can be in the eastern basin, but little information is available at present. After the oceanic stage, Mediterranean turtles frequenting the oceanic habitats in the south Adriatic/north Ionian area, the central Mediterranean and the Tyrrhenian eventually recruit to the North Adriatic and the Central Mediterranean neritic areas, and possibly also to other neritic areas in the eastern Mediterranean where tag returns could be underreported. Although sample size is small to be conclusive, results (Table 3; Fig. 3) suggest that turtles undertaking their ecological transition may have more chance to settle in neritic areas closer to their oceanic areas than in more distant ones. This could be favored by currents approaching or entering the neritic areas, as proposed by Bolten (2003). For instance, from the Tyrrhenian and Central Mediterranean oceanic area to the Central Mediterranean neritic area, and from the south Adriatic/north Ionian oceanic area to the North Adriatic neritic area (Table 3; Fig. 3). This is further supported by the small size of turtles captured by bottom trawlers (Casale et al. 2004) and various Wshing gear (Lazar and Tvrtkovic 2003) in the neritic areas of North Adriatic and by bottom trawlers in the neritic areas of the Central Mediterranean (Casale et al. 2007b) suggesting an early transition and so a short oceanic phase.

1047

On the basis of present results, once settled in these neritic areas, juvenile turtles would not move to other areas. As far as adults are concerned, it is interesting that of the females tagged while nesting in Greece and re-encountered in other circumstances, only few (2%) were found in the western basin, while most of them were found in the North Adriatic, the Central Mediterranean, and the Aegean (Margaritoulis et al. 2003). Although in mark-recapture studies some areas may be overlooked because of less probability of reports (lower Wshing eVort, lower human presence on the beaches, diVerent language, lower education or attitude in reporting such Wndings, etc.), the higher proportion of reencounters from the western basin observed in this study (20.5%) (Fisher Exact test; P < 0.01; n = 232) suggests that the small proportion reported by Margaritoulis et al. (2003) is not due to such a bias but to a real preference by Greek adult females for the above areas in the eastern basin. Altogether, these Wndings suggest that almost all turtles born in Greece, after spending their oceanic phase in the western (e.g., Tyrrhenian; western Mediterranean) or eastern basin (e.g., Adriatic/north Ionian, Central Mediterranean; other), move into neritic habitats in the eastern basin, like the North Adriatic, the Central Mediterranean, and the Aegean, where they remain for the rest of their life, except for reproductive migrations. Moreover, it is even possible that a part of the Greek population frequents just a limited area of the Mediterranean: the oceanic habitats in the south Adriatic/north Ionian, the neritic habitats in the north/central Adriatic, the reproductive area in the Ionian (Zakynthos, Greece), and the migration pathway between this nesting site and the North Adriatic, as also suggested by reencounters of adult females (Lazar et al. 2004). Both genetic and tagging data from other Mediterranean populations are not adequate at present to reveal their movement patterns. However, it is possible that the central Mediterranean neritc area is also frequented by populations other than the Greek one. Results indicate that, in the Mediterranean, part of the turtles in the oceanic phase remain in the same oceanic area, and that all turtles in the neritic phase remain in the same neritic area. This area Wdelity could imply an higher vulnerability for the populations of this threatened species, because in case of intense disturbance in an area turtles might not simply shift among neritic areas or even among oceanic ones. Moreover, diVerent populations might prefer speciWc areas as oceanic and neritic habitats, being more or less vulnerable to threats occurring in the areas at diVerent degrees. Further and detailed genetic studies to assess this aspect are highly and urgently needed. Classic CMR studies of sea turtles require tagging large numbers of animals, suVer from possible bias due to low reporting rates from some areas, and certainly cannot provide information on turtle movements and behavior

123

1048

between the two encounters. Nonetheless, at present time Xipper tagging is the only way to make an individual turtle recognizable by almost everyone after many years, especially in a size class with a high growth rate that prevents many kinds of attachments to hold for long. The sample size that is possible to obtain with this method can help proposing models that can then be further evaluated by more speciWc studies using Wne scale methods such as satellite tracking or genetic markers. In the Mediterranean, Xipper tagging could provide important information on turtle distribution in underrepresented areas, such as the eastern part of the basin, especially if new approaches of intensive tagging at sea are developed. In this respect, collaboration with Wshermen has shown to be very helpful and can be realized in conjunction to awareness campaigns that are fundamental for conservation. Acknowledgments This work was possible thanks to the invaluable eVorts of many individuals. All the Wshermen who collaborated to the project; the students and volunteers who helped in the Weld work; the central and local authorities; and WWF-Italy. In particular, we thank: G. Gerosa (Chelon); M. Cocco, C. Jacomini, G. Cantoro, B. Cristo, C. Focarelli, G. Guerrasio, G. Paolillo, A. Piras, G. Rallo, D. Ricciardi, F. Sottile, V. Spica, M. Spoto, A. Vanadia (WWF-Italy); A. Di Natale (Acquario di Genova); P. Meschini (Acquario di Livorno); C. Froglia (CNR Ancona); A.L. Stanzani (Fondazione Cetacea); A. Dominici and S. Nannarelli (Hydrosphera); G. Marzano (Museo di Storia Naturale del Salento di Calimera); V. Olivieri (Museo Ittico di Pescara); F. Bentivegna and S. Bocchetti (Stazione Zoologica di Napoli); A. Donato (Università di Messina); M. Passarella. Maps were prepared with the Maptool program (SEATURTLE.ORG, www.seaturtle.org). An earlier version of the manuscript was greatly improved, thanks to valuable comments by two anonymous referees.

References Abbate G, Casale P, Freggi D, Rocco M, Argano R (2007) Diet and ecological transition of loggerhead sea turtles (Caretta caretta) in the central Mediterranean. In: Frick M, Panagopoulou A, Rees AF, Williams K (eds) Book of abstracts, 27th Annual Symposium on Sea Turtle Biology and Conservation, Miami Argano R, Basso R, Cocco M, Gerosa G (1992) New data on loggerhead (Caretta caretta) movements within the Mediterranean. Boll Mus Ist Biol Univ Genova 56–57:137–164 Avens L, Lohman K (2000) Homing behavior in juvenile loggerheads (Caretta caretta) from Core Sound, North Carolina, USA. In: Kalb HJ, Wibbels T (eds) Proceedings of the 19th Annual Symposium on Sea Turtle Biology and Conservation. U.S. Dept. Commerce. NOAA Tech. Memo. NMFS-SEFSC-443, pp 147–148 Bentivegna F (2002) Intra-Mediterranean migrations of loggerhead sea turtles (Caretta caretta) monitored by satellite telemetry. Mar Biol 141:795–800 Bjorndal KA (1997) Foraging ecology and nutrition of sea turtles. In: Lutz PL, Musick JA (eds) The biology of sea turtles. CRC, Boca Raton, FL, pp 199–232 Bjorndal K, Bolten AB, Martins HR (2000) Somatic growth model of juvenile loggerhead sea turtles Caretta caretta: duration of pelagic stage. Mar Ecol Prog Ser 202:265–272 Bolten AB (1999) Techniques for measuring sea turtles. In: Eckert KL, Bjorndal KA, Abreu-Grobois FA, Donnelly M (eds) Research and

123

Mar Biol (2007) 152:1039–1049 management techniques for the conservation of sea turtles. IUCN/ SSC Mar Turtle Spec Group Publication 4, pp 110–114 Bolten AB (2003) Active swimmers—passive drifters: the oceanic juvenile stage of loggerheads in the Atlantic system. In: Bolten AB, Witherington BE (eds) Loggerhead sea turtles. Smithsonian Books, Washington, DC, pp 63–78 Bowen BW, Avise JC, Richardson JI, Meylan AB, Margaritoulis D, Hopkins-Murphy S (1993) Population structure of the loggerhead turtle Caretta caretta in the northwest Atlantic Ocean and Mediterranean Sea. Conserv Biol 37(4):834–844 Caminas JA, De La Serna JM (1995) The loggerhead distribution in the western Mediterranean Sea as deduced from captures by the Spanish long line Wshery. In: Llorente GA, Montori A, Santos X, Carretero MA (eds) Scientia Herpetologica, Barcellona, pp 316– 323 Caminas JA, Baez JC, Valeiras X, Real R (2006) DiVerential loggerhead by-catch and direct mortality due to surface longlines according to boat strata and gear type. Sci Mar 70(4):661–665 Cardona L, Revelles M, Carreras C, Sanfelix M, Gazo M, Aguilar A (2005) Western Mediterranean immature loggerhead turtles: habitat use in spring and summer assessed through satellite tracking and aerial surveys. Mar Biol 147:583–591 Carreras C, Cardona L, Marco A, Pascual M, Aguilar A, Margaritoulis D, Rees A, Castillo JJ, Tomás J, Raga JA, Fernández G, SanFelix M, Turkozan O, Levy Y, Gasith A, Aureggi M, Khalil M, Rico C, Roques S (2006a) Degree of isolation of the Mediterranean nesting populations of the loggerhead sea turtle using nuclear markers. In: Frick M, Panagopoulou A, Rees AF, Williams K (eds) Book of abstracts, 26th Annual Symposium on Sea Turtle Biology and Conservation, Athens Carreras C, Pont S, MaVucci F, Pascual M, Barcelo A, Bentivegna F, Cardona L, Alegre F, SanFelix M, Fernandez G, Aguilar A (2006b) Genetic structuring of immature loggerhead sea turtles (Caretta caretta) in the Mediterranean Sea reXects water circulation patterns. Mar Biol 149(5):1269–1279 Casale P, Laurent L, De Metrio G (2004) Incidental capture of marine turtles by the Italian trawl Wshery in the north Adriatic Sea. Biol Conserv 119:287–295 Casale P, Freggi D, Basso R, Argano R (2005a) Size at male maturity, sexing methods and adult sex ratio in loggerhead turtles (Caretta caretta) from Italian waters investigated through tail measurements. Herpetol J 15:145–148 Casale P, Freggi D, Basso R, Argano R (2005b) Oceanic habitats for loggerhead turtles (Caretta caretta) in the Mediterranean Sea. Mar Turtle Newsl 107:10–11 Casale P, Mazaris AD, Freggi D, Basso R, Pino d’Astore P, Argano R (2007a) Growth rates and age-at-size of juvenile loggerhead sea turtles (Caretta caretta) in the Mediterranean Sea, estimated through length frequency data. In: Frick M, Panagopoulou A, Rees AF, Williams K (eds) Book of abstracts, 27th Annual Symposium on Sea Turtle Biology and Conservation, Miami Casale P, Cattarino L, Freggi D, Rocco M, Argano R (2007b) Incidental catch of marine turtles by Italian trawlers and longliners in the central Mediterranean. Aquat Conserv (in press) Dodd CK Jr (1988) Synopsis of the biological data on the loggerhead sea turtle Caretta caretta (Linnaeus 1758). U.S. Fish and Wildlife Service Biological Report, 88(14):110 Eckert SA (2006) High-use oceanic areas for Atlantic leatherback sea turtles (Dermochelys coriacea) as identiWed using satellite telemetered location and dive information. Mar Biol 149(5):1257–1267 Godley BJ, Broderick AC, Glen F, Hays GC (2003a) Post-nesting movements and submergence patterns of loggerhead marine turtles in the Mediterranean assessed by satellite tracking. J Exp Mar Biol Ecol 287:119–134 Godley BJ, Lima EHSM, Åkesson S, Broderick AC, Glen F, Godfrey MH, Luschi P, Hays GC (2003b) Movement patterns of green

Mar Biol (2007) 152:1039–1049 turtles in Brazilian coastal waters described by satellite tracking and Xipper tagging. Mar Ecol Prog Ser 253:279–288 Gomez de Segura A, Tomas J, Pedraza SN, Crespo EA, Raga JA (2006) Abundance and distribution of the endangered loggerhead turtle in Spanish Mediterranean waters and the conservation implications. Anim Conserv 9:199–206 Gramentz D (1991) Migrationen von Caretta caretta im Mittelmeer. Herpetofauna 13(72):20–25 Hays GC, Broderick AC, Godley BJ, Luschi P, Nichols WJ (2003) Satellite telemetry suggests high levels of Wshing-induced mortality in marine turtles. Mar Ecol Prog Ser 262:305–309 Hays GC, Marsh R (1997) Estimating the age of juvenile loggerhead sea turtles in the North Atlantic. Can J Zool 75(1):40–46 Hendrickson JR (1958) The green sea turtle, Chelonia mydas (Linn.) in Malaya and Sarawak. Proc Zool Soc Lond 130:455–535 Hochscheid S, Bentivegna F, Hays GC (2005) First records of dive durations for a hibernating sea turtle. Biol Lett 1:82–86 Hochscheid S, Bentivegna F, Bradai MN, Hays GC (2007) Overwintering behaviour in marine turtles: dormancy is optional. Mar Ecol Prog Ser (in press) Laurent L, Casale P, Bradai MN, Godley BJ, Gerosa G, Broderick AC, Schroth W, Schierwater B, Levy AM, Freggi D, Abd El-Mawla EM, Hadoud DA, Gomati HE, Domingo M, Hadjichristophorou M, Kornaraky L, Demirayak F, Gautier Ch (1998) Molecular resolution of marine turtle stock composition in Wshery bycatch: a case study in the Mediterranean. Mol Ecol 7:1529–1542 Lazar B, Borboroglu PG, Tvrtkovic N, Ziza V (2003) Temporal and spatial distribution of the loggerhead sea turtle, Caretta caretta, in the eastern Adriatic Sea: a seasonal migration pathway? In: SeminoV JA (eds) Proceedings of the 22nd Annual Symposium on Sea Turtle Biology and Conservation. NOAA Technical Memorandum NMFS-SEFSC-503, pp 283–284 Lazar B, Tvrtkovic N (2003) Corroboration of the critical habitat hypothesis for the loggerhead sea turtle Caretta caretta in the eastern Adriatic Sea. In: Margaritoulis D, Demetropoulos A (eds) Proceedings of the Wrst Mediterranean conference on marine turtles. Barcelona Convention—Bern Convention—Bonn Convention (CMS). Nicosia, Cyprus, pp 165–169 Lazar B, Margaritoulis D, Tvrtkovic N (2004) Tag recoveries of the loggerhead sea turtle Caretta caretta in the eastern Adriatic Sea: implications for conservation. J Mar Biol Ass UK 84:475–480 Lazar B, Gracan R, Zavodnik D, Tvrtkovic N (2007) Feeding ecology of “pelagic” loggerhead turtles, Caretta caretta, in the northern Adriatic Sea: proof of an early ontogenetic habitat shift. In: Proceedings of the 25th Annual Symposium on Sea Turtle Biology and Conservation (in press) Limpus CJ (1994) The loggerhead turtle, Caretta caretta, in Queensland: feeding ground selection following her Wrst nesting season. In: Bjorndal KA, Bolten AB, Johnson DA, Eliazar PJ (eds) Proceedings of the 14th Annual Symposium on Sea Turtle Biology and Conservation. NOAA Technical Memorandum NMFS-SEFSC-351, pp 78–81 Limpus CJ, Limpus DJ (2001) The loggerhead turtle, Caretta caretta, in Queensland: breeding migrations and Wdelity to a warm temperate feeding area. Chel Conserv Biol 4(1):142–153 Limpus CJ, Limpus DJ (2003) Biology of the loggerhead turtle in Western South PaciWc Ocean foraging areas. In: Bolten AB,

1049 Witherington BE (eds) Loggerhead sea turtles. Smithsonian Books, Washington, DC, pp 93–113 Luschi P, Akesson S, Broderick AC, Glen F, Godley BJ, Papi F, Hays GC (2001) Testing the navigational abilities of ocean migrants: displacement experiments on green sea turtles (Chelonia mydas). Behav Ecol Sociobiol 50(6):528–534 MaVucci F, Kooistra WHCF, Bentivegna F (2006) Natal origin of loggerhead turtles, Caretta caretta, in the neritic habitat oV the Italian coasts, Central Mediterranean. Biol Conserv 127:183–189 Malanotte-Rizzoli P, Manca BB, Ribera D’Alcala M, Theocharis A, Bergamasco A, Bregant D, Budillon G, Civitarese G, Georgopoulos D, Michelato A, Sansone E, Scarazzato P, Souvermezoglou E (1997) A synthesis of the Ionian Sea hydrography, circulation and water mass pathways during POEM Phase I. Prog Oceanogr 39(3):153–204 Margaritoulis D (1988) Post-nesting movements of loggerhead sea turtles tagged in Greece. Rapp Comm int Mer Médit 31(2):284 Margaritoulis D, Argano R, Baran I, Bentivegna F, Bradai MN, Caminas JA, Casale P, De Metrio G, Demetropoulos A, Gerosa G, Godley B, Houghton J, Laurent L, Lazar B (2003) Loggerhead turtles in the Mediterranean Sea: present knowledge and conservation perspectives. In: Bolten AB, Witherington B (eds) Loggerhead sea turtles. Smithsonian Institution Press, Washington, DC, USA, pp 175–198 McMahon CR, Hays GC (2006) Thermal niche, large-scale movements and implications of climate change for a critically endangered marine vertebrate. Global Change Biol 12:1–9 Meylan AB, Bjorndal KA, Turner BJ (1983) Sea turtle nesting at Melbourne Beach, Florida. II. Post-nesting movements of Caretta caretta. Biol Conserv 26:79–90 Millot C, Taupier-Letage I (2004) Circulation in the Mediterranean Sea. The Handbook of Environmental Chemistry, Vol 1 (The Natural Environment and the Biological Cycles), Springer-Verlag, Berlin Musick JA, Limpus CJ (1997) Habitat utilization and migration in juvenile sea turtles. In: Lutz PL, Musick JA (eds) The biology of sea turtles. CRC, Boca Raton, FL, pp 137–163 Papi F, Luschi P, Crosio E, Hughes GR (1997) Satellite tracking experiments on the navigational ability and migratory behavior of the loggerhead turtle Caretta caretta. Mar Biol 129(2):215–220 Polovina JJ, Kobayashi DR, Parker DM, Seki MP, Balazs GH (2000) Turtles on the edge: movement of loggerhead turtles (Caretta caretta) along oceanic fronts, spanning longline Wshing grounds in the central North PaciWc, 1997–1998. Fish Oceanogr 9(1):71–82 Renaud ML, Carpenter JA (1994) Movements and submergence patterns of loggerhead turtles (Caretta caretta) in the Gulf of Mexico determined through satellite telemetry. Bull Mar Sci 55(1):1–15 Schroeder BA, Foley AM, Bagley DA (2003) Nesting patterns, reproductive migrations, and adult foraging areas of loggerhead turtles. In: Bolten AB, Witherington BE (eds) Loggerhead sea turtles. Smithsonian Books, Washington, DC, pp 114–124 SeminoV JA, Resendiz A, Resendiz B, Nichols WJ (2004) Occurrence of loggerhead sea turtles (Caretta caretta) in the Gulf of California, Mexico: evidence of life- history variation in the PaciWc Ocean. Herp Rev 35(1):24–27 Tomas J, Dominici A, Nannarelli S, Forni L, Badillo FJ, Raga JA (2001) From hook to hook: the odyssey of a loggerhead sea turtle in the Mediterranean. Mar Turtle Newsl 92:13–14

123