- Top of page
- Supporting Information
Many animal populations invest in long-distance migration to reduce foraging competition when food resources are scarce at breeding sites (e.g. Greenwood, 1980; Alerstam et al., 2003; Fryxell et al., 2004). Hence, while a population may aggregate at one site to breed, many disparate foraging sites may be used depending on resource availability and the pattern of dispersal which may be influenced by factors such as wind direction, sea currents or land barriers in avian, marine and terrestrial migrants, respectively (e.g. Åkesson & Hedenstrom, 2007). Consequently, important habitats for protection may be distributed across wide regions, often traversing international borders or economic exclusion zones (e.g. Hannah et al., 2002; Bradshaw et al., 2010). At present, of the 150,000 protected areas across the world, just 3% are in the oceans (c. 5000 marine protected areas) (Bradshaw et al., 2010). Hence, as countries introduce measures to improve protection along their coastlines (Bradshaw et al., 2010), it is important to ensure that governments and environmental agencies select optimal sites for regulation (Ferrier, 2002).
Remote tracking systems of animals, using, for example, GPS or ARGOS, may generate extensive datasets that directly show the placement of individuals within ecosystems, portraying habitat use and migratory patterns. This technology provides an opportunity to collect fine-scale spatio-temporal information about species that would otherwise be difficult to study (i.e. tuna, Block et al., 2005; shearwaters, Shaffer et al., 2006; leopards, Gavashelishvili & Lukarevskiy, 2008). For instance, tracking studies of sea turtles, which are of world-wide conservation concern, have gained momentum over the last 20 years (Godley et al., 2008), with datasets of > 30 (max. 186) turtles from individual populations gradually emerging; however, these works are primarily focused on females and juveniles (e.g. Polovina et al., 2006; Hawkes et al., 2011), rather than males (but see Arendt et al., 2012a, b). In general, tracking studies of sea turtles have inferred (1) high fidelity of adults to foraging sites (e.g. Broderick et al., 2007; Schofield et al., 2010a), (2) seasonal variability in dispersal to foraging sites (e.g. van Dam et al., 2008; Hawkes et al., 2011; Zbinden et al., 2011), (3) male and female differences in dispersal and foraging site use (e.g. van Dam et al., 2008; Arendt et al., 2012a) and (4) phenotypic variation in foraging habitat (coastal versus oceanic) use within populations (e.g. Hatase et al., 2002, 2010; Hawkes et al., 2006). Several theories have been proposed to explain these differences, including (1) hatchling drift influenced by sea currents (Hays et al., 2010b; Gaspar et al., 2012), (2) the evolutionary context of food abundance predictability (Drakare et al., 2006; van Dam et al., 2008) and (3) possible adult mortality risk of migratory routes and foraging sites (van Dam et al., 2008).
The Mediterranean Sea has been the focus of extensive sea turtle research, including satellite tracking and by-catch studies (e.g. Broderick et al., 2007; Zbinden et al., 2008; Schofield et al., 2009; Casale et al., 2010; Schofield et al., 2010a; Zbinden et al., 2011). This work has resulted in the delineation of two Regional Management Units (RMU) for loggerhead sea turtles, representing populations from the Mediterranean and north-west Atlantic (Wallace et al., 2010), with green turtle (Chelonia mydas) populations also being resident to the region. Interestingly, while juvenile loggerheads from Atlantic populations enter the western Mediterranean basin (via the Strait of Gibraltar) for development (Casale et al., 2008; Eckert et al., 2008), the breeding and foraging habitats of permanent Mediterranean loggerhead and green populations are primarily concentrated in the central and eastern basins, with Zakynthos Island (Greece) in the central basin hosting the largest known breeding site for loggerhead turtles (Fig. 1a; Margaritoulis et al., 2003). Initial stranding data of dead turtles from Zakynthos (based on external flipper tag returns) that were washed ashore or entangled in nets indicated potentially broad dispersal to foraging sites (Margaritoulis et al., 2003); yet, this observation could have been an artefact of oceanic drift by dead or injured animals. However, subsequent satellite tracking studies began to support the stranding data, but indicated differences in movement patterns. For instance, a tracking study of 18 female loggerheads from Zakynthos (Zbinden et al., 2011) showed a clear dispersal pattern to foraging sites, either north (to the Adriatic) or south (to the Gulf of Gabes). In contrast, a tracking study of 10 males from the same site (Schofield et al., 2010a) showed primarily northerly dispersal to the Adriatic and Aegean. This raised the question of whether males and females use different foraging sites, or whether this was simply an artefact of sample size.
Figure 1. (a) The location of the breeding area of Zakynthos in the Mediterranean compared with other loggerhead (stars) and green (triangles) turtle nesting areas in the Mediterranean. The main seas of the Mediterranean that are cited in this study are indicated. (b) Migration of male (black, n = 31) and female (red n = 29) loggerhead turtles from the breeding area of Zakynthos in Greece to the primary foraging sites. Excluded tracks: five retracked (repeat tracked, either through the device operating for more than one year, or a new transmitter being attached) males; one male and five females without locations between the breeding and foraging sites. White star = Zakynthos. (c) Schematic showing the patterns (as actual numbers of turtles) of initial dispersal by male (n = 27) and female (n = 33) turtles on departing the breeding area (excluding residents, n = 5). Note: diagram not to scale.
Download figure to PowerPoint
Hence, in this study, we assessed the dispersal and foraging destinations of 75 adult male and female loggerheads that were captured and tracked from the key Mediterranean breeding area of Zakynthos between 2004 and 2011. Based on our tracking data, we suggest where suitable neritic foraging habitat may lie in winter and summer months, enabling prioritization for conservation. Finally, we combine our data with published literature from other sea turtle (green and loggerhead) populations across the Mediterranean, to identify key foraging sites according to contributions of species, breeding population, sex and age class. This study supports the importance of using large datasets to correctly infer population level dispersal patterns and foraging habitat use, as well as to delineate important areas for the protection of sea turtles inhabiting the Mediterranean Sea.
- Top of page
- Supporting Information
The present study shows that turtles from a key loggerhead sea turtle population in the Mediterranean exhibit variable dispersal patterns, due to the large availability of neritic foraging habitat throughout the region. As we tracked over 30 males and 30 females in the current study, the observed dispersal patterns are considered to be representative for this population, allowing any differences between the sexes to be clearly defined. However, while the sample size of the current study is large compared with the majority of existing sea turtle studies (see Godley et al., 2008), the tracked individuals in this study still only represented about 9% of this breeding population (Schofield et al., 2010b). Hence, it was necessary to draw on available tracking, stranding and genetic information from the wider literature to delineate core foraging areas for protection.
It is possible that we may have missed foraging areas that are rarely used by the Zakynthos adult breeding population, despite the size of the tracking dataset. For instance, adult turtles foraging in the Bay of Naples have been tracked to foraging and breeding sites in the central Mediterranean (Bentivegna, 2002; Hochscheid et al., 2007). Furthermore, increased tracking is revealing increased behavioural plasticity (e.g. Hatase et al., 2007; Seminoff et al., 2008; Rees et al., 2010; Schofield et al., 2010a), which confers some resilience against change, despite complicating the identification of all core foraging areas.
Interestingly, while our previous work using n = 10 males (Schofield et al., 2010a) indicated a bias of male movement towards northerly foraging sites, once the sample size exceeded 20 individuals, similar numbers dispersed north as south, more closely reflecting that observed for females from the same population. Furthermore, the current study supported adult female tag returns from the stranding datasets (Margaritoulis et al., 2003), as well as hatchling drift scenarios (Hays et al., 2010b). However, this consistency across studies was stronger for tracked females compared with tracked males. Male dispersal deviated from that of females in that 24% of males remained resident or close (< 100 km) to the breeding area, supporting recently published studies (Shaver et al., 2005; van Dam et al., 2008; Arendt et al., 2012a). Yet, studies on Cyprus (loggerheads, Broderick et al., 2007) and the Galapagos (green turtles, Seminoff et al., 2008) have recorded females resident to the general area; hence, we may have just not yet tracked a resident female. The observed difference in migratory pattern between the sexes at Zakynthos, and other areas, may be due to males facing much lower reproductive costs/investment compared with females (Clutton-Brock & Vincent, 1991). Similar differences in movement patterns and habitat use between the sexes have also been recorded for other species (e.g. wandering albatrosses, Weimerskirch et al., 1997; fur seals, Boyd et al., 1998). This behaviour might facilitate more opportunistic foraging strategies by males, as well as allow annual reproductive activity and greater (earlier) access to returning females (for review see Morbey & Ydenberg, 2008).
In contrast to Hatase et al. (2010) in the Pacific, but similar to Hawkes et al. (2011) in the Atlantic, the majority of male and female loggerheads frequented neritic foraging sites of < 100 m seabed depth in the central Mediterranean, either to the north (Adriatic) or south (Gulf of Gabes), with only a few individuals using oceanic sites. This north versus south division is similar to that recorded by Hawkes et al. (2011) in the Atlantic. Furthermore, our habitat model showed that these two areas of the Mediterranean support the greatest areas of available foraging habitat in the Mediterranean. Due to sea temperatures dropping below 13 °C in winter (i.e. the minimum Mediterranean wintering temperatures recorded by Hochscheid et al., 2007) in the northern area, turtles were more likely to use more than one foraging/wintering site and move southward compared with those inhabiting the southern area.
High neritic habitat availability, multi foraging site use and temperature regulated shifts to wintering sites make establishing focal areas for protection in the Mediterranean extremely challenging. Despite this, tracking datasets from the current study combined with existing publications, based on tracking and other techniques (i.e. stranding, capture-recapture and genetics), produced baseline information identifying key foraging sites for protection at national and regional levels. At a regional scale, pressure to protect the Zakynthos population is extremely difficult due to the complexity of national boundaries and competition between governments and government agencies regarding jurisdictional responsibility (Agardy, 1994). Furthermore, detailed by-catch studies, from the Adriatic to the Gulf of Gabes (Casale et al., 2008, 2010; Echwikhi et al., 2010; Chaieb et al., 2012), clearly indicate the high risk of mortality to sea turtles as a result of intensive fisheries activities in these key foraging sites. However, in both regions, we identified key areas for protection that would also safeguard adult and juvenile turtles from at least 10 other breeding sites. At northern sites, which are primarily used by turtles in the summer, seasonal protection measures could be introduced, with different levels of fisheries activity being permitted in different seasons (Hannah et al., 2002). At a national scale, within the Ionian Sea of Greece, we reaffirm the need to establish a large multiple use protected area, with an integrated coordinated management system providing varying levels of protection. Within this framework of small MPAs (López Ornat, 2006), six foraging and/or breeding sites (including Zakynthos) would receive protection, which are frequented by year-round resident males, sexually active males shuttling between breeding sites (Schofield et al., 2010a, b) and reproductive females conducting forays among sites (Schofield et al., 2010a, b). Furthermore, we recommend that protective legislation and effective management should be strengthened in Amvrakikos Gulf, as the existing national park primarily focuses its protection actions in the northern section (transitional ecosystems/wetlands) of the marine area (see also Rees et al., 2013). To further refine important areas for conservation planning, it is important to assimilate all available environmental criteria (e.g. Edwards & Richardson, 2004; Coll et al., 2010), including areas of seasonal use and how the distribution of sea turtle foraging sites might change over time (Hannah et al., 2002; Araújo et al., 2004).
In conclusion, our study provides a good example of a sea turtle population with high variability in dispersal and foraging habitat use, demonstrating the importance of having large tracking sample sizes to infer core areas for protection at an ocean basin scale.