Daily fine-scale habitat use by animals is often classified based on their activities, such as foraging, resting, socializing, and reproducing (Campbell and Tobler 1984; Häfker et al. 2017). For many aquatic animals, daily activity patterns are influenced by environmental factors such as food availability, predation risk, tidal movement, and light conditions, among others (Campbell and Tobler 1984; Siegel 2008; Petit et al. 2015). These daily activity patterns (i.e., circadian rhythms) are crucial to maintaining the good health of organisms by maintaining metabolic activities (e.g., metabolites peaks such as magnesium during high energy demand periods) (Dunlap and Loros 2016). For example, disturbing individuals' daily activity patterns might have negative consequences on their fitness and behavior (e.g., delaying feeding time to avoid contact with predators) (Tessmar-Raible et al. 2011; Häfker et al. 2017).

Disproportionate habitat use by humans is likely a significant factor influencing daily animal activities of shallow coastal marine fauna; first, through direct consumptive effects and, second, through indirect non-consumptive predatory effects such as inducing variations in the normal (undisturbed) behavior of the animals (e.g., habitat use of foraging and resting areas) (Pfister and Harrington 1992; Gaymer and Himmelman 2008; Hamann et al. 2010; Terborgh and Estes 2010). This human impact on habitats may be even more detrimental for highly migratory species, such as marine turtles, because daily activities such as foraging and resting might be especially vital for their survival after long migrations in terms of mass gain, reproductive success, and survival (Pfister and Harrington 1992; Dingle and Drake 2007; Sheehy et al. 2011; Rees et al. 2016). As such, determining habitat use of turtles in coastal foraging areas and pinpointing where turtles are most susceptible to human disturbances have been cited as major of research needs (Hays 2008; Hamann et al. 2010; Rees et al. 2016).

Green turtles (Chelonia mydas) mainly feed on seagrass and/or algae in tropical and subtropical habitats; however, they seem to vary their diet depending on the availability of resources and with ontogeny (Bjorndal 1980; Brand-Gardner et al. 1999; Vélez-Rubio et al. 2016). In addition, marine turtles may not graze in random places but instead use grazing plots or specific foraging areas (Brand-Gardner et al. 1999; González and Álvarez-Varas 2016; Christiansen et al. 2017). Those areas provide stable food sources, which supply their nutritional requirements (Kuiper-Linley 1994). Turtle foraging activity has been well studied over the last 3 decades, revealing activity peaks associated with diel times, light hours, moon phase, and tides (Bjorndal 1980; Ogden et al. 1983; Tighe 1981; Christiansen et al. 2017).

Resting is a central part of most organisms' lives for the function of metabolic activities, including food digestion (nutrient absorption), hormone-release modulation, cardiovascular functions, and glucose regulation (Campbell and Tobler 1984; Cauter et al. 2008). As a consequence, a reduction in resting behavior (e.g., by disturbance) may result in negative physiological and behavioral responses (Siegel 2008). In marine turtles, resting has been described based on dive times and underwater movements or steadiness, principally by the use of a variety of bio-telemetry devices and video records (see Hochscheid et al. 2005, 2007; Seminoff et al. 2006; Rice and Balazs 2008; Christiansen et al. 2017). Even though there are a few cases in which observations were made through direct underwater observations (see Von Brandis et al. 2010; Rincon-Díaz et al. 2011; Chassagneux et al. 2013) information concerning to this topic remains scarce.

Direct underwater observations by professional scuba divers have permitted the identification of more than 20 turtles that use small, specific areas of coral reefs as permanent resting sites in the waters surrounding Rapa Nui Island, Chile (also known as Easter Island). Those sites are individually used and can be identified according to features such as the plastron print upon coral structures and the eroded surfaces of small caves produced by the repeated contact with turtle's plastron, flippers, and carapace (C.F.G., pers. obs.). Likewise, a well-recognized turtle foraging site has been described in Hanga Roa Bay (Gaymer et al. 2011; González and Álvarez-Varas 2016), where several green turtles feed daily on green algae (i.e., Ulva sp.).

The first aim of this article is to describe the physical properties of the coral reef sites selected by turtles as resting sites (hereafter referred to as resting sites) as well as turtles' resting behavior in Hanga Roa Bay. Secondly, we describe their daily foraging behavior and its relation to tide cycles in Rapa Nui. Finally, this study provides a framework for managing and protecting turtles and their foraging and resting grounds within the Rapa Nui Multiple Uses Coastal Marine Protected Area. This management framework represents the synthesis and application of information gained from our applied research on one of the most iconic marine reptiles worldwide.

METHODS

Study Site

Rapa Nui is a remote Chilean oceanic island situated along the Salas y Gómez Ridge (Fig. 1) at the easternmost corner of the Polynesian Triangle. The island is located approximately 3700 km from the Chilean coast and represents the southeasternmost distribution of coral reefs and coral reef organisms in the Pacific. Indeed, Rapa Nui presents subtidal zones in which more than 50% of their area is covered by scleractinian (stony) corals of the genera Porites and Pocillopora (Friedlander et al. 2013). The coral reefs at the island can be large, reaching more than 10 m high and 40 m in length; they are largely present along the north and west faces of the island (Hubbard and García 2003).

View Figure

• Download as Powerpoint
• Download Figure

The dominant swell around Rapa Nui comes from the southeast, whereas the west and north faces tend to be less exposed to waves (Hubbard and García 2003). The main town, Hanga Roa, is situated on the west coast of Hanga Roa Bay (lat 27°08′50″S, long 109°26′01″W) and is the most populated zone of the island. The bay has the highest green turtle density around Rapa Nui (González and Álvarez-Varas 2016) and is where their feeding grounds and resting sites are found. Recently, the Chilean government and the Rapa Nui people established the biggest large-scale marine protected area in the South Pacific Ocean, the Rapa Nui Multiple Uses Coastal Marine Protected Area. The area of this conservation unit covers almost all the 728,000 km² of the maritime Exclusive Economic Zone of Rapa Nui.

The population of green turtles that inhabits the waters around the island has been identified and monitored by citizen scientists since 2010 through the use of photo-identification (photo-ID). Currently, 41 turtle individuals have been registered and identified around Rapa Nui through this technique (González and Álvarez-Varas 2016).

Anthropogenic Uses of Hanga Roa Bay

Hanga Roa Bay has multiple uses, and is highly frequented by tourists and locals. A small harbor (hereafter referred to as Caleta) anchors approximately 25 artisanal fishing boats. Ten diving centers use the bay to conduct their underwater activities. An artificial shallow intertidal pool in a small beach called “Pea” is frequented by swimmers. There are also 2 surf spots used by surfers year-round. During summer (January–February, Southern Hemisphere school holidays), the Caleta and its surroundings are used for recreational purposes, including the annual festival “Tapati”. In addition, sailboats and cruise ships use deeper areas of the bay for anchoring.

RESULTS

Resting Site Use

Fifteen green turtles were encountered at resting sites between 2013 and 2018 by citizen scientists. Among them, 12 were observed and photographed more than once using the same site during 5 yrs of observations. Turtle S was registered resting 7 times at the same resting sites (Resting Site ID: Ms_3; Table 1) in 2014 and 2017, whereas turtle F was registered 8 times at 2 different resting sites (Resting Site IDs: B2_2, B2_3; Table 1) during a period of 5 yrs. Moreover, 2 turtles (Turtles: D and L; Table 1) were registered using the same resting site twice (Resting Site ID: B2_4; Table 1), but during different periods of time.

Table 1 Number of green turtle underwater observations in specific resting sites in Hanga Roa Bay over 5 yrs (from 2013 to 2018), based on citizen-scientist observations. Only resting locations that have documented re-use are included.

View Fullscreen

Thirty-six observations of turtles in resting sites were collected from 2013 to 2018 at Hanga Roa Bay, with a significant increase in observations during low tide (χ² = 31.180, n = 33, p < 0.001). Among the 36 observations, 26 occurred during low tide, 6 during high tide, and 1 between high and low tides (Fig. 2); 3 observations had no time registered.

View Figure

• Download as Powerpoint
• Download Figure

Resting Site Characteristics

Nineteen turtle resting sites were used in our analyses. The spatial and structural similarities among them are shown in the nMDS ordination (Fig. 3). In general, there are no clear differences in resting site characteristics across the study areas. There was no significant difference between the proportions of resting sites in sheltered areas (n = 12) compared with exposed areas (n = 7). Most resting sites were oriented in a southeast cardinal direction (n = 8), followed by northwest (n = 3) and southwest (n = 3) (Fig. 4). Twelve resting sites occurred in small caves located within reef formations (shaded area), while the rest occurred on open, flat areas (Fig. 5). Although the proportion of resting sites in caves was not significantly higher than open sites (χ² = 1.3158, df = 1, p > 0.05), caves were more recurrent in exposed areas (5 of 7 resting sites). The type of substrate was predominantly characterized by sand or the coral Porites lobata, whereas a significantly lower proportion of resting sites presented rocky bottoms (χ² = 8.895, df = 1, p < 0.01). The depth of the resting sites ranged between 5.5 and 20 m with a median of 12.5 m (Fig. 6). Although most of the resting sites were observed between 5 and 10 m, no significant differences in resting sites frequency was observed across depth ranges (χ² = 2, df = 2, p > 0.05). The area of the resting sites ranged from 0.49 to 3.14 m² and it did not show any clear pattern across the study areas. We detected no significant relationship between the area of the resting sites and the local depth or orientation.

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

Foraging Sites

Marine turtles were found foraging throughout the entire area from Pea Beach (lat 27.14826°S, long 109.43205°W) to Caleta Hanga Roa (lat 27.14690°S, long 109.43185°W) throughout the whole year. The main foraged item was Ulva sp. Twenty-four different individuals were counted from a total of 36 observations during the sampling period. A significant difference was observed between foraging times (i.e., daily tides) (χ² = 7.111, n = 36, p = 0.008). During high tide, 3.7 individuals were counted on average compared to only 1.4 individuals during low tide (Fig. 7). From the total number of turtles recognized, 19 were identified as resident individuals that have been registered by citizen scientists on the island over approximately 5 yrs.

View Figure

• Download as Powerpoint
• Download Figure

DISCUSSION

Here we show that Hanga Roa Bay has key habitat features, such as specific resting sites and foraging areas, for C. mydas in Rapa Nui, Chile. Our results indicate a high level of site fidelity of turtles for specific resting sites, with several turtles (n = 12) being recorded using the same locations over 5 yrs of observations. Even though we did not have frequency data about resting site use, this apparent site fidelity suggests that those sites have specific attractive features for C. mydas in Hanga Roa Bay and that they were not chosen randomly. Furthermore, in some areas, different individuals have been registered using the same site at different periods of time.

The use of the same resting sites by multiple individuals may be the result of nonterritorial behavior and/or the disuse of specific sites by individuals. Fidelity for underwater resting sites has been well described by Christiansen et al. (2017) using tagged animals' movements in the Indian Ocean, where turtles used different geographic areas at different times of the day. Although our study did not include nighttime dives, turtles are rarely seen at night in the known foraging area but are commonly seen in their resting sites during nocturnal dives (C.F.G., pers. obs.). This nocturnal resting behavior might be related to predator avoidance (e.g., sharks) (Makowski et al. 2006). However, Rapa Nui is described as an overfished environment where predators are almost absent (Friedlander et al. 2013) and, therefore, their daily movements may be associated with other factors such as tidal cycles or the proximity to their feeding grounds.

Most observations of marine turtles at their resting sites occurred during low tide. This result is consistent with the findings from the daily foraging times, which showed a significantly higher number of turtles foraging during high tides. Thus, during high tides, C. mydas moved closer to the coast to forage, while during low tides their activity was focused on resting farther offshore.

We did not find statistical differences in resting site depth; however, the most common depths for the resting sites were between 8 and 11 m. These depths probably provide a balance between being protected against water movement, which increases closer to the surface, and the need to be close enough to the surface to breathe without spending too much energy with displacement (Seminoff et al. 2006).

Even though most resting locations occurred in the protected faces of reef structures (southeast orientation), some were also found along wave-exposed faces (northwest orientation). Nevertheless, most exposed resting sites were in caves, indicating that C. mydas might prefer sites with reduced water flow for resting. Vertical and horizontal distribution of resting sites showed a clear pattern, with almost all sites positioned at the bottom and middle of reef structures, which might provide the strongest protection against water flow related to the upper end and the corners of reef structures. Sites protected from tides and waves would require less energy expenditure to maintain immobility while resting (Seminoff et al. 2006).

As the largest proportion (42%) of resting sites were oriented to the southeast (protected from wave energy), the roughness of coral structures might also be a selection criterion as it has been shown that exposed reef faces present higher coral roughness compared with protected faces (Hearn 2011). The turtle's flat plastron may fit better on a flatter bottom, such as protected reef faces. Likewise, our findings showed that the most common bottom type for cave resting sites was sand, which is a flat bottom. In terms of temperature dissipation, which might be relevant for marine turtles, protected reefs are known to experience higher temperature variability, with higher overall temperatures, than exposed reefs because of the reduced water motion and consequent higher heat accumulation during daylight hours (Hearn 2011).

Another important variable might be the proximity to food sources, which is one of the key factors in determining resting places for a variety of vertebrates (Iwata and Ando 2007). For example, in the Virgin Islands, green turtles rest in areas within 1 km of their foraging sites (Ogden et al. 1983). However, home range sizes could be highly variable depending on factors such as predator presence, currents and tides, and maximizing energy intake, among others (Charnov 1976; Elner and Hughes 1978; Christiansen et al. 2017). In Rapa Nui, resting sites and the foraging area are within 0.5 km. It is important to note that the foraging area provides a stable food source (i.e., the green alga Ulva sp.) throughout the entire year. In addition, some of the resident green turtles are accustomed to the local artisanal fishing fleet, whose fishers feed them with fish remains, resulting in some individual turtles remaining in the area during the entire day.

The foraging behavior of green turtles in Hanga Roa Bay showed a clear pattern in relation to feeding times and tidal cycle, with a significantly higher number of turtles being present during high tides. The higher density of marine turtles during high tide might be related to the easier accessibility of food items, such as the green alga Ulva sp. that grow on intertidal boulders. During high tide, these rocks are completely submerged, while during low tide they are above the water surface, out of the turtles' reach. Moreover, wave height also seems to be a relevant issue determining the daily feeding times of turtles (I.J.P., pers. obs.), which warrants further investigation. Big waves breaking on the foraging area would make the feeding process difficult because of the strong water motion, which increases the risk of turtles colliding with rocks and thus decreases the stability necessary for them to obtain their food items.

During this research, one sample of C. mydas feces was collected from Manavai reef and then analyzed to identify its organic material. One type of algae, from the genus Ulva, and a sponge (Cribrochalina dura) were identified. These findings, in addition to direct observations frequently made by local divers, demonstrate that C. mydas are feeding upon both green algae and invertebrate resources in Hanga Roa Bay. This analysis provides the first evidence of marine turtles in Rapa Nui feeding upon sponges, such as Cribrochalina dura; however, more replicates are needed to determine whether sponges represent a common food item for green turtles at Rapa Nui.

Enhancing the research focused on fine-scale habitat use of fauna (e.g., marine turtles) is fundamental for species conservation. Knowledge about local movements, feeding areas, resting areas, and diel activity patterns are all vital for effective management of turtles in coastal habitats. This information could be used by the managers of the recently declared Rapa Nui Multiple Uses Coastal Marine Protected Area for effectively identifying critical areas for protection and scheduling uses of Hanga Roa Bay in order to diminish threats to marine turtles such as massive and invasive touristic diving activities, high-speed boat traffic, and ship anchoring, among others.

Some characteristics of the analyzed data set impose limitations on the kind of analytical protocols that could be applied. As with many data sets assembled from citizen observations, our data incorporate sporadic reports of recreational divers with no standardized sampling protocols. As the diving sites and routes might change for numerous reasons, the resulting observations were spatially and temporally unbalanced. Hence, multiple accounts of resting turtles at a site might be just a reflection of places that were more frequently visited by divers than actual preferences of the sea turtles for that site. The small number of observations (19 resting sites and 36 turtle observations) also reduces the power of our analyses to detect a real effect of environmental characteristics on the selection of resting sites by turtles. Even though the analyses applied in this study are limited by these constraints, they do reveal patterns about the behavior of sea turtles in Rapa Nui. Thus, the observations described here might lead the way for the development of new studies with standardized sampling protocols to test hypotheses about green turtle preferences for resting sites.

CONCLUSIONS

We show that C. mydas in Hanga Roa Bay appear to have fidelity to their resting sites, which are frequently used during low tide. The resting sites are probably selected based on their geo-morphologic features to reduce energy expenditure during resting. Thus, more research, protection, and management of these areas will consequently result in improved conservation of these turtles on Rapa Nui. In addition, we show that C. mydas regularly feed at specific places in Hanga Roa Bay during high tide, and that 19 resident individuals, including juveniles and adults, commonly forage in this area. Educating the community (locals and tourists) and regulating activities inside the foraging area would therefore likely help to ensure the normal development of C. mydas foraging activity.

MANAGEMENT RECOMMENDATIONS

1. Human–turtle encounters often occur during touristic scuba-diving activities, which constantly disturb animals while they are resting, thereby interrupting their normal routine. A simple management strategy would be the establishment of an appropriate distance limit for these encounters.

2. Ship anchors (e.g., sailboat anchors) can damage coral structures, destroying turtle resting sites and all the biodiversity associated with them (e.g., benthic invertebrates, fish refuge). Designated anchoring areas and specific buoys outside of critical turtle habitats would be a rapid solution against this threat.

3. Turtle fatalities occur when boats collide with turtles present at foraging and resting sites in Hanga Roa Bay. An appropriate management strategy is to reduce boat speed limits in specific areas of the bay and require the use of propeller protectors to decrease the number of fatal encounters between boats and turtles.

4. Tourist swimmers at the turtle foraging area might disturb their normal foraging behavior. A useful management strategy would be the establishment of an appropriate observation distance and a schedule that would limit tourist activities during the peak foraging times of the day.