Home Range and Movement Patterns of Subadult Hawksbill Sea Turtles in Southeast Florida
Subadult Hawksbill Turtles (Eretmochelys imbricata) reside on the coral reefs of Palm Beach County, Florida, but their movements and patterns of habitat use are poorly understood. In this study, six subadult Hawksbills were tracked with global positioning system (GPS)–linked satellite telemetry for a span of 102–429 days Total home ranges and within-range areas of “core” use were measured with minimum convex polygons (MCPs) and kernel density estimates (KDEs). Home-range estimates ranged 1.1–19.0 km2 (X = 10.1 km2) using MCP and 0.01–1.2 km2 (X = 0.49 km2) using the 95% KDE. Each turtle remained at or near the 15–25 m hard-bottom reef habitats of the area and exhibited strong site fidelity to centrally located core use areas (50% and 25% KDE >,0.03 km2); this was especially true at night, suggesting the repeated use of familiar refuges (shipwrecks/caves) for nocturnal shelter. Likely driven by predator avoidance, competition for a limited number of preferred refuges, or “roosts,” may restrict the extent of each turtle's home range and influence the abundance and distribution of the Hawksbill Turtles that occupy this site.Abstract
Hawksbill Turtles (Eretmochelys imbricate; Linnaeus, 1766) are globally distributed migratory marine turtles that, in the Western Hemisphere, are found throughout the Caribbean Sea, west to the Gulf of Mexico, and south to the coast of Central America and northern South America (Witzell, 1983). Although nesting is rare on the U.S. mainland, Hawksbill Turtles have been reported in waters near the Florida Keys northward to the reefs adjacent to Palm Beach County (Lund, 1985; Meylan and Redlow, 2006; Wood et al., 2013). As large spongivores, Hawksbills are considered important predators that may influence reef ecosystem structure and dynamics (Hill, 1998; Leon and Bjorndal, 2002). Because of long-term overexploitation, they are listed as Critically Endangered worldwide by the International Union for Conservation of Nature (IUCN) (Baillie and Groombridge, 1996).
Hawksbill distribution data from southeastern Florida remain largely unavailable, and a number of authors have recognized the need for expanded in-water research on this species there and elsewhere (Eaton et al., 2008; Hays, 2008; Hart et al., 2012; Hawkes et al., 2012). Several foraging areas have been identified in continental U.S. waters, including the Dry Tortugas and Florida Keys (Gorham et al., 2014; Hart et al., 2012) and recently Wood et al. (2013) found the abundance and growth rates among a Palm Beach County aggregation comparable to those measured elsewhere in the Caribbean. After originating from Caribbean rookeries, the dispersal of post-oceanic juvenile Hawksbills to southeastern Florida is likely facilitated by the Gulf Stream current (Blumenthal et al., 2009). Indeed, both the Palm Beach County and Florida Keys aggregations are represented by multiple regional haplotypes, the former being strongly represented by Mexico's Yucatan rookeries (Gorham et al., 2014; Wood et al., 2013).
Home ranges are sometimes maintained at some or all life stages by otherwise highly vagile species (Burt, 1943; Borger et al., 2008). Once established, the geographical extent of the home range typically is determined by a combination of resource availability, energy expenditure, competition, and predator–prey interactions (Rincón-Díaz et al., 2011; Barraquand and Murrell, 2012). Where studied in the Caribbean, subadult Hawksbills demonstrate fidelity to relatively small areas of reef (van Dam and Diez, 1998; Cuevas et al., 2007; Blumenthal et al., 2009; Hart et al., 2012; Hawkes et al., 2012, 2014). Likewise, in Palm Beach County, Florida, some may remain localized for extended periods of time, perhaps even throughout the majority of their juvenile and subadult life stages (Wood et al., 2013). Currently, no data exist on the spatial extent of Hawksbill movements in southeastern Florida. This basic information is fundamental to our understanding of the life history of this species and is critical for the effective management of its recovery (NMFS and USFWS, 1998).
Until recently, the use of satellite telemetry to study movement patterns in sea turtles has been conducted over relatively large geographic scales, typically addressing questions of migration and/or long-range dispersal. Because of short surface times and positional inaccuracies inherent in remote sensing technology, satellite telemetry has not been very suitable for fine-scale assessments of marine turtles (Bradshaw et al., 2007; Witt et al., 2010). The relatively small spatial scale over which typical subadult Hawksbill Turtles are suspected to move precludes the use of most standard, ARGOS-based transmitters. New technologies, however, such as global positioning system (GPS)–linked Fastloc® (Wildtrack Telemetry Systems, Ltd., Leeds, UK), have improved positioning accuracy by incorporating data from the GPS even when communication (surface) times are limited (Moen et al., 1996, 1997; Witt et al., 2010). This technology has been successfully used in studies of Green (Chelonia mydas; Hazel, 2009), Loggerhead (Caretta caretta; Schofield et al., 2010), and Leatherback (Dermochelys coriacea; Dodge et al., 2014) Turtles and is capable of providing the additional resolution required to explore intra-range patterns of habitat use.
Animal habitat use estimates can be determined with various spatial analyses such as the traditional minimum convex polygon (MCP) and more recent nonparametric kernel density (KDE) methods (Silverman, 1986; Worton, 1989; Seminoff et al., 2002; Cuevas et al., 2007). Because certain locations within the subject's home range may be visited more or less frequently, the concept of the habitat utilization distribution (UD) (Worton, 1987) was introduced to better describe areas of within-range core use that may be linked to important behavioral needs such as foraging, resting, or predator avoidance. Understandably, an animal's presence may not always be evenly distributed throughout its entire range, and it may change its UD patterns on various temporal and spatial scales. This is particularly applicable in cases where animals exhibit central-place foraging or repeatedly return to a focal point of activity, such as a den or nest (Orians and Pearson, 1979; Barraquand and Murrell, 2012). If present, these within-range subtleties are not detected by the MCP area estimates because they calculate the total polygonal area surrounding all observations, including the unusual, although potentially relevant, outliers (Burt, 1943). Alternatively, KDEs provide a more probabilistic estimate of the home range by scaling the boundaries to the area's most frequently visited by the subject, thereby creating a UD (reviews by Worton, 1987, 1989; White and Garrot, 1990). By excluding the outliers and recognizing clusters, UD frequency contours reveal areas of disproportionately high or low use within the overall home range, which can be used to link specific environmental features to important life-history traits. In this study, previously tagged subadult Hawksbills were tracked with satellite telemetry to determine both the extent of their home ranges and whether they exhibit within-range patterns of habitat use.
Materials and Methods
Study Site
This study was conducted on the northernmost section of the Southeast Florida Continental Reef Tract (Banks et al., 2007, 2008; Riegl et al., 2007), located in the waters of northern Palm Beach County, FL (Fig. 1). This nonaccreting, shore-parallel reef system is located ∼2 km offshore, forming a narrow ridge that features a 2–3 m tall landward ledge or cliff, a central plateau that rises to 16 m, and a gradual seaward slope that descends to the seafloor at ∼21 m. To the north and east of the reef tract's terminus, the hard bottom consists of what are thought to be cemented sand dunes known as the deep ridge complex (Riegl et al., 2007) that form a series of shore-parallel ridges featuring similar physical characteristics in slightly deeper water (21–27 m).
Citation: Journal of Herpetology 51, 1; 10.1670/15-133
In addition to the natural reefs of the area, various artificial reef structures (vessels, concrete pilings, limerock rubble, etc.) have been intentionally scuttled for ecological and recreational enhancement by the Palm Beach County Department of Environmental Resources Management. One of these sites, known locally as the “corridor,” consists of a series of ships and rock rubble that lie in close proximity to the deep ridge complex, in ∼24 m of water.
Collectively, Palm Beach County's reef-associated biota exists at a relatively high latitude and survive because of the warm waters of the Gulf Stream, which introduces comparatively wide water temperature variations seasonally (21–29°C) and episodically (temporary upwellings as low as 15°C). Persistent currents average 1.3 m s−1 but occasionally exceed 3 m s−1 (Banks et al., 2008). Additionally, the east coast of Florida is subject to periodic hurricanes and tropical storms, which can generate scouring wave surges and large-scale substrate redistribution. These conditions have resulted in a heterogeneous patchwork of benthic communities that are dominated by macroalgae, alcyonian corals, and poriferans and that are rounded out by zooanthids, tunicates, hydroids, and a relatively small (≈12%) proportion of scleractinian corals (Moyer et al., 2003; Jaap, 2006).
Sampling
Six subadult Hawksbill Turtles (49.1–70.6 cm; X = 59.0 ± 9.1 SD) were opportunistically hand captured during scuba diving surveys on the nearshore reefs of north–central Palm Beach County, Florida, between 6 January 2009 and 23 September 2013 (Table 1). Four of the turtles were captured on the Breakers reef tract adjacent to the Town of Palm Beach, and the other two came from an artificial reef site (“corridor” shipwrecks) ∼7 km to the north adjacent to the Town of Palm Beach Shores (Fig. 1). Recent tagging studies indicate that Hawksbills are abundant in both areas (Wood et al., 2013). To increase the likelihood of tracking long-term resident animals, previously tagged turtles from each site were exclusively selected to carry transmitters.
An Mk10-AF GPS-linked multisensor satellite transmitter (Wildlife Computers, Inc., Redmond, WA) was affixed to the apex of each turtle's carapace using a combination of T308 adhesive epoxy (Powers Fasteners, Inc., New Rochelle, NY) and Sonic-Weld® putty (Sonitek, Inc., Milford, CT). The turtles were retained on a boat for at least 45 min to allow the adhesives to dry and then released at the water's surface above the capture site. To reduce confusion during data retrieval and analysis, each transmitter's abbreviated manufacturer ID number (19–24) was assigned to each corresponding subject.
Each tag was programmed to calculate Argos and GPS coordinates at each surfacing event (∼20 per 24-h period) and to transmit data daily. Four tags were retrieved (21–24), and one was redeployed on a different turtle after manufacturer rebuild. Data were retrieved via the Argos system through CLS America, Inc. (Largo, MD) and/or directly downloaded (21–24) and decoded with DAP Processor version 3.0 software (Wildlife Computers, Inc.).
Only GPS coordinates were used for the homerange estimates. The DAP software assigns a quality indicator (a.k.a “residual” value) to each location, and as suggested by the manufacturer, residual values that exceeded a value of 20 were excluded from the analysis. The remaining locations were imported to Arcview 10.1 GIS (ESRI, Inc., Redlands, CA) software. Once projected, data points were visually inspected, and those deemed impossible or highly unlikely (on land, ≥ 10 km away from the study site) were eliminated (≤ 1% of total data). The locations were grouped by hour and classified as either “diurnal” (0800–1959 h EST) or “nocturnal” (2000–0759 h EST). The Spatial Analyst extension of Arcview 10.1 was used to create MCPs and KDEs (95%, 50%, and 25% contours) from the 24-h data set and the 12-h day/night data sets individually (per MacLeod, 2014), based on Silverman's (1986) quadratic kernel function. Smoothing parameters, or search radii (h-values), were generated by ArcView 10.1 based on the extent of each turtle's data set and are reported in Table 2.
Results
Fastloc® GPS location data were received from six subadult Hawksbill Turtles (straight carapace length [SCL] 50.0–70.6 cm; X = 59.0 ± 9.1; median = 58.2) between August 2010 and May 2014 (Table 1). Deployment duration ranged 102–429 days (X = 286 ± 121), providing between 33 and 2,116 (X = 1,005 ± 782) locations (Table 3). Turtle 20 ceased transmitting 102 days after deployment and provided only 33 locations that met the above standards, limiting its contribution to the study and the reliability of the home-range estimate (HRE) for this individual.
The home ranges estimated by MCP ranged 1.10–19.04 km2 (X = 10.1 ± 7.3). Diurnal HREs ranged 0.2–18.57 km2 (X = 7.1 ± 7.2), and nocturnal HREs ranged 0.98–8.99 km2 (X = 5.3 ± 3.0) (Table 3, Fig. 2). Although this method did not illustrate the details of within-range habitat use, it included what could have been occasional exploratory movements to nearshore or adjacent offshore areas and had enough resolution to detect a nearly 50% reduction in activity (movement) at night (Table 3).
Citation: Journal of Herpetology 51, 1; 10.1670/15-133
Alternatively, the KDE95 home-range size ranged 0.01–1.12 km2 (X = 0.49 ± 0.36) and largely confined the turtles' movements to 15–20 m offshore hard-bottom habitats of the study site (Table 3, Fig. 2). Turtles 22 and 24 had one continuous 95% isopleth (Fig. 2C,E), whereas turtles 19, 21, and 23 had four, three, and one, respectively, smaller additional 95% isopleth “islands” adjacent to the primary area of use (Fig. 2A,B,D). Turtle 20 is not represented because sufficient data were not acquired from this animal. When the locations were temporally segregated, the diurnal and nocturnal KDE95 HREs ranged 0.23–1.11 km2 (X = 0.61 ± 0.41) and 0.17–0.67 km2 (X = 0.33 ± 0.21), respectively (Table 3), again showing a nearly 50% reduction in the area occupied by the turtles at night.
The turtles tracked on the continental reef tract showed home ranges that were similarly shaped and tightly outlined the reef ridge (Fig. 2B,C,D). The northern half of turtle 21's home range overlapped the southern half of turtle 23's home range and also overlapped, although to a lesser degree, turtle 22's range to its south (Fig. 3). Conversely, the two turtles that were captured at the corridor shipwreck site to the north of the Lake Worth Inlet (turtles 19 and 24) showed differently shaped home ranges but extremely high overlap near the capture site itself (Fig. 3).
Citation: Journal of Herpetology 51, 1; 10.1670/15-133
Core use areas (KDE50 and KDE25) averaged 0.03 ± 0.03 and 0.01 ± 0.004 km2, respectively, and were roughly centered within each turtle's home range. (Table 4, Fig. 2). Turtle 21 had two KDE50 isopleths, the southern of which was considerably more populated (312 coordinates vs. 54) and was used for comparison with the other turtles (Fig. 2B). The areas within the KDE50 isopleths were populated by a roughly equal number of days versus night locations (46.1 ± 2.1% and 53.9 ± 2.1%, respectively); however, there was a bias toward daytime locations (68.1 ± 2.7%) in the areas contained between the 50% and 95% contour lines (Fig. 4). The KDE25 isopleths were centered within the KDE50 isopleths and were located very near sites that can provide shelter, such as shipwrecks and ledges (Fig. 3). Approximately 40% (41.0 ± 3.1) of the locations within these very small core use areas were diurnal (Fig. 4).
Citation: Journal of Herpetology 51, 1; 10.1670/15-133
Discussion
Over the years, various methodologies have been applied to estimate Hawksbill home ranges, but none have used data generated from Fastloc® GPS-quality satellite telemetry. The results of this study mirror previous estimates that subadult Hawksbills heavily use small areas of reef habitat (≈1 km2), indicating this to be a fairly ubiquitous natural history trait for this species (Table 5). Despite occasional forays in- and offshore of the main reef, the turtles studied in Palm Beach remained very closely associated with the study site's 15–25 m hard-bottom habitats, underscoring the value of supplementing MCPs with KDEs to acquire a more meaningful assessment of habitat use. The extent of subadults' daily movements should principally reflect food availability (Rincón-Díaz et al., 2011) rather than courtship and/or other reproductive behavior, although still unidentified gender-specific behavior may occur among immature turtles. Spongivory is common in Caribbean Hawksbills (reviewed by Bjorndal, 1997) and seems to dominate the diet of those in Palm Beach as well (LDW, pers. obs.). Known Hawksbill prey such as Loggerhead Sponges Spheciospongia vesparium, Chicken Liver Sponges Chondrilla caribensis, Giant Barrel Sponges Xestospongia muta, and Leathery Barrel Sponges Geodia sp. are commonly found at these depths (LDW, pers. obs.).
Hawksbills are known to become largely inactive at night (van Dam and Diez, 1998; Blumenthal et al., 2009; Hart et al., 2012) and to retreat to refuges (rock outcroppings, coral heads, etc.) during times of diurnal inactivity (Houghton et al., 2003; Blumenthal et al., 2009; Proietti et al., 2012). The cover of underwater structures creates low-light, multidimensional microhabitats; provides refuge from currents and predators; and potentially extends turtle dive times by counteracting mild positive buoyancy at depth, a.k.a. “assisted resting” (Houghton et al., 2003). This study site lacks coral heads, and relatively few sizable caves/caverns are found on the crest, patch reef, or seaward “spur-and-groove” slope of the reef. The Hawksbills tracked in this study consistently returned to structurally complex, cave-forming core use areas throughout both the day and night hours. The remarkable fidelity these turtles show to these particular sites may strongly influence their daily movement patterns and significantly frame the shape and extent of each turtle's foraging range.
Aside from the six turtles tracked in this study, dozens of additional individuals have been tagged and repeatedly encountered in the study site (Wood et al., 2013). What appears to be considerable range overlap among members of this aggregation suggests that foraging areas are not strongly defended. Still, aggression among individuals has been documented in both captive and wild settings, indicating some sort of mild territoriality (Sanches and Bellini, 1999; van Dam and Diez, 2000; Blumenthal et al., 2009; LDW, pers. obs.). For Hawksbills, the extent of the defensible territory may simply be the minimum distance from another turtle that does not invoke aggressive behavior that may change under varying circumstances. Resident turtles are much more likely to encounter one another at or near preferred refuge sites, particularly at night, leading to some degree of either cooperative or defensive action. Either way, as the space becomes more crowded, some individuals will not be able to find the desired cover and will likely be compelled to move on by more aggressive individuals. Over time, it may be advantageous to choose a reliable and familiar “home base” that is shared with other turtles (including those of other species) to the extent that physical space and competition allow.
In a number of species, including various birds (Eiserer, 1984; Beauchamp, 1999), bats (Lewis, 1995), insects (Pearson and Anderson, 1985; Finkbeiner et al., 2012), and primates (Anderson, 1998), individuals regularly congregate in a quiescent state known as communal roosting (DeVries et al., 1987). The drivers behind this behavior are hypothesized to include thermoregulation (Yom-Tov et al., 1977; du Plessis et al., 1994), information sharing (Ward and Zahavi, 1973), foraging efficiency (Caccamise and Morrison, 1988), and prey dilution/predator confusion (Turner, 1975; Gillette et al., 1979; Eiserer, 1984; Finkbeiner et al., 2012). In the case of ectothermic and otherwise nonsocial Hawksbill Turtles, predator avoidance/prey dilution emerges as the most likely benefit for individuals seeking nightly refuge in common areas, leading to some level of competition for the most protective locations. Southeastern Florida is known as a foraging ground and migratory corridor for numerous shark species, including Tiger Sharks (Galeocerdo cuvier), known predators of marine turtles. During times of inactivity, especially at night, the more quickly the turtle can return to refuge after surfacing for air (roughly hourly; LDW, pers. obs.) the presumably less exposed it would be to predation. Given the considerable drift from an often strong northerly current, a return to an established location could be advantageous over repeatedly seeking out unfamiliar, potentially already occupied, downstream refuges. Although Hawksbills can conceal themselves well in gorgonian pastures, these areas lack the protective cover of ledges and caves and may not be preferred resting areas for dominant, long-term residents.
Turtles extend their range during the day into the patch reef and spur-and-groove formations, likely reflecting foraging expeditions. However, the nearly equal proportion of diurnal and nocturnal coordinates that occur within the core use areas (both KDE50 and KDE25; Fig. 5) suggests that turtles make repeated diurnal visits to these areas, which may be for food (foraging is commonly observed at wrecks, rubble, and ledge walls) but may also serve to reinforce some level of dominance over each turtle's preferred roost. Consistent patterns of movement within a home range imply spatial learning and memory (Shettleworth, 2001), and repeated behaviors are likely to reflect prior experiences. Coincidentally, among the three turtles tracked on the Breakers reef (21–23), each turtle's KDE95 isopleth terminated at a conspecific's core use area, where a higher probability of interaction between them would be expected. Although perhaps for other reasons, there is a conspicuous lack of Hawksbill Turtles under 40-cm SCL in this study site (Wood et al., 2013) that could be a competitive disadvantage if alternative refuges are unavailable.
Citation: Journal of Herpetology 51, 1; 10.1670/15-133
Collectively, the home ranges increased with the number of coordinates acquired for each turtle, suggesting the deployment duration may have influenced the resulting home-range estimates (Fig. 6). This trend has occurred in other marine turtle studies as well (Seminoff et al., 2002; Hart et al., 2012). In this study, this relationship was strongly influenced by two turtles that provided very small and very large data sets (20 and 22, respectively). The other four turtles had very similar HREs, although one of them (24) collected at least twice the coordinates of the other two (21 and 23). Home-range estimates among all six turtles did not increase with carapace length, however, indicating that increased residency time does not coincide with continual spatial expansion; otherwise, larger turtles (presumably those that have resided longer at the site) would be expected to have proportionately larger areas of activity. Upon closer temporal inspection of turtle 22's extensive data set, it became apparent that an abrupt shift occurred during the deployment to a new center of activity on an adjacent reef to the south (Fig. 4), giving the illusion of a nearly doubled HRE. This move occurred on 12 July 2013, after which the turtle developed activity patterns that were nearly identical to its own previous patterns and to those of the other turtles (KDE95 = 0.25 km2 [346 days] and 0.35 km2 [83 days] in the north and south portions, respectively). Because of the timing of the move, insufficient data were accumulated in the second center of activity, and a second core use isopleth could not be created from the combined data set.
Citation: Journal of Herpetology 51, 1; 10.1670/15-133
Overall, the smoothing factor or search radius (h) generated by ArcView 10.1 resulted in one or few isopleth contours for each turtle that roughly outlined the contours of the hard-bottom habitat, where the turtles were by far most likely to be found. The h-value can have an effect on the home-range estimate calculations, based on the size and spatial arrangement of the data set (Hemson et al., 2005). If the chosen h-value is too low, the home range decreases through fragmentation into numerous discontinuous isopleth “islands,” whereas high h-values increase home-range estimates by strongly including more distant points. In this case, the 95% isopleth contours were continuous for two turtles and were only slightly fragmented in three others, suggesting that analyses captured realistic home-range estimates.
To ensure the reliability of their study's HREs based on deployment duration, Seminoff et al. (2002) determined that juvenile Green Turtles (N = 12) took less than 100 days to occupy 100% of each corresponding home range, based on between 16 and 61 (X = 37) coordinates per turtle. Therefore, home ranges reported for the deployments exceeding 200 days (≥ 600 positions) in this study likely are reasonably accurate (turtles 19, 21, 23, and 24) and are overestimated because of turtle 22's late-deployment move (429 days; 2,116 positions). This trend was not seen in the nocturnal KDE50 estimates; all turtles (except turtle 20) occupied areas measuring less than 0.03 km2 at night (Table 4), regardless of the number of positions in the analysis (Fig. 6). Moreover, the KDE25 isopleth for each turtle was nearly exactly centered within the KDE50 contours, further shrinking the size of the core use areas to less than 0.01 km2. This consistency reflects the high fidelity of these turtles to specific sites within their observed ranges, but these may be temporary. Turtle 22's apparent move was captured only by a deployment that exceeded 12 months, revealing that home ranges may shift and move over time, perhaps reflecting continuing competition for preferred resting and/or foraging locations.
The abundance, growth rates, and home-range sizes among this Hawksbill aggregation are comparable to those reported from other Caribbean sites (Wood et al., 2013), further suggesting that reefs of Palm Beach, although not formerly considered important Hawksbill habitat, strongly support their development and are worthy of continued study and targeted conservation measures. Understanding where Hawksbills are most likely to be found can greatly streamline management strategies. Although direct harvest is strictly prohibited in U.S. waters, these turtles are affected by a number of anthropogenic factors such as discarded fishing line, boat propeller strikes, and pervasive degradation of reef habitat. Their consistent presence on the 15–25 m deep reef complex underscores their reliance on it for both food and shelter. Because covered refuges have emerged as important environmental features for Hawksbills, a carefully planned artificial reef program, similar to that of Palm Beach County, may augment Hawksbill recruitment in areas where similar “turtle-friendly” features are lacking. Additionally, the efficiency of future in-water surveys and/or mark-recapture studies could be considerably increased by focusing on areas where plenty of cover is available, particularly after dark.
This study underscores the value of employing Fastloc® GPS-linked remote sensing for marine turtle research. Although comparatively costly, transmitters equipped with Fastloc® GPS capabilities can provide the levels of detail that wildlife managers desperately need to assign critical habitat and wisely target conservation resources (Dujon et al., 2014). In the case of subadult Hawksbills, the small home ranges they appear to occupy facilitate transmitter retrieval, refurbishment, and reuse through the increased likelihood of subsequent turtle recapture, potentially providing a less costly approach to increase sample sizes. Additionally, the remote sensing of marine turtles of any kind generates graphics that help translate to a wide audience the often complex biology of these endangered species and more easily draw the public's attention to areas of conservation priority.
Reefs of central and northern Palm Beach County, Florida.
(A–E) Home-range estimates for five Hawksbill Turtles. The shaded polygons represent the HRE as calculated by the minimum convex polygon (MCP) method, and the light blue contours as calculated by the 95% kernel density estimate (KDE95) method. Core use areas (KDE50 and KDE25) are represented by bright green and white contours, respectively. Turtle 22 (C) moved to the south during the transmitter deployment, creating two distinct home ranges.
(A) Home-range overlap (KDE95, 50, 25) for Hawksbill Turtles 19 and 24. Both were captured at an artificial reef site consisting of scuttled vessels, limerock, and concrete debris. The KDE50 areas (bright green) overlapped nearly perfectly at this site, and the KDE25 areas (white) for both animals were located at a shipwreck known as the Amaryllis. (B) Home-range overlap (KDE95, 50, 25) for turtles 21–23 (continental reef tract). For all three turtles, the KDE25 areas (white) were located at or near the center of each KDE50 (bright green).
Adjacent home ranges of Hawksbill Turtle 22 before and after an abrupt southward move that occurred 346 days into a 429-day deployment.
Mean proportion (%) of diurnal/nocturnal coordinates located in the area between the KDE50 and KDE95 isopleths, within the KDE50 isopleth, and within the KDE25 isopleth for six Hawksbill Turtles.
Home-range size (km2) versus number of coordinates (dashed line) and straight carapace length (SCL notch–tip in cm; solid line) for six turtles. Squares (□) = SCL; diamonds (♦) = total KDE95; crosshairs (×) = nocturnal KDE95; triangles (▴) = nocturnal KDE50.
Contributor Notes
