Beach Condition and Marine Debris: New Hurdles for Sea Turtle Hatchling Survival
Abstract
The dash of freshly emerged hatchlings to the sea is a short but critical phase in the life cycle of sea turtles. We examine the time spent on a Mediterranean beach from the perspective of substrate composition and temperature, marine debris, and hatchling crawling speeds and times. Crawling speeds depended on beach quality. The hatchlings crawled significantly faster on sand than on cobbles, and cooler temperatures considerably prolonged the crawls. Marine debris was a major impediment: hatchlings were severely entangled in fishing nets and entrapped in simple containers such as plastic cups and cut-open canisters. They never avoided contact with such experimentally deployed debris or reversed their direction to escape. The overall debris density at the study site averaged 1.03 items m−2, mostly plastic, and 2 out of 3 hatchlings had contact with such debris on the way to the sea. Marine debris is a new aspect of habitat quality for sea turtle nesting site monitoring and conservation efforts and may help explain the long-term decline in nest numbers on this beach. This serves as a case study for the role of habitat quality in the survival of endangered species.
All 7 species of sea turtles are listed on the International Union for Conservation of Nature (IUCN) Red List of Threatened Animals (IUCN 2011). The threats facing certain stages in the life cycle are well documented for adults at sea or for eggs inside the nest. Few sea turtles survive from egg to adulthood, with various models yielding an often-cited 1∶1000 survival rate (Frazer 1986). It is assumed that the most mortalities occur in the hatchling and juvenile stages, although few quantitative data are available on hatchling mortality. Values range from 2.5% in the first 15 min of leaving shore for loggerhead hatchlings (Glenn 1998) to 6.9% in the first 20 min for hawksbill hatchlings (Harewood and Horrocks 2008) to 46.7% for green turtle hatchlings followed for 1 hr over or near reefs with predaceous fishes (Pilcher et al. 2000).
Sea turtles—and hatchlings in particular—are among the few organisms facing threats in both the marine and terrestrial ecosystems. In the sea, hatchling numbers are mainly reduced by predators (Stewart and Wyneken 2004). The effect of well-known anthropogenic threats to adults and juveniles such as entanglement, toxins, and ingestion of plastic or tar balls (Carr 1987; Gramentz 1988; Bjorndal et al. 1994; Bugoni et al. 2001; Tomás et al. 2002; Mascarenhas et al. 2004) are either poorly documented or unknown for hatchlings (National Research Council 1990). On land, hatchlings are susceptible to natural predators while still in the nest (Türkozan and Baran 1996; Engeman 2003). Dogs, jackals, foxes, raccoons, pigs, and monitor lizards (Brown and Macdonald 1995; Blamires et al. 2003; Barton and Roth 2008; Engeman et al. 2010) can dig up nests, consume the eggs and hatchlings directly, or indirectly damage eggs due to exposure and secondary bacterial infestation (Bustard 1972). After successful emergence, while still on the beach, they are prey for opportunistic predators such as ghost crabs, rats, birds, and crocodiles (Limpus et al. 1983; Fowler 1987; Stancyk 1995; Sutherland and Sutherland 2003; Stewart and Wyneken 2004; Caut et al. 2008; Santidrián Tomillo et al. 2010). Emerged hatchlings also face anthropogenic threats, above all light pollution, which misorients them (Witherington and Bjorndal 1991; Tuxbury and Salmon 2005), resulting in desiccation (Bustard 1972), increased predation (Stancyk 1995), and compromised ability to respond to later orientation cues at sea (Lorne and Salmon 2007). These factors, coupled with habitat deterioration (Bjorndal and Jackson 2003), have accelerated the decline of many populations (Spotila et al. 2000; Jackson et al. 2001; Ilgaz et al. 2007).
The dash across the beach is a run for life and a run against time. It is very short compared to the time spent in the nest and at sea, but losses here are thought to be substantial (Davenport 1997): generally, the shorter the duration, the better. We address the gap in our knowledge about this critical phase using a Mediterranean population of the endangered loggerhead turtle Caretta caretta (L.) to determine factors prolonging crawl times and reducing hatchling survival.
We begin by examining crawling speeds and the range of factors potentially impacting hatchling performance, including beach grain size, temperature, hatchling size, and what we define as orientational motivation. We then focus on one anthropogenic threat—marine debris, which is omnipresent in the marine environment and a major threat to seabirds (Huin and Croxall 1996), marine mammals (Page et al. 2004), adult sea turtles (Carr 1987), and even fish (Sazima et al. 2002). Plastic makes up most of the marine litter worldwide (Derraik 2002) and in the Mediterranean (Gabrielides et al. 1991). Remote or inaccessible beaches, which are not cleaned, often have the greatest densities of plastic pollution (Ryan and Moloney 1993; Gregory 1999).
Only anecdotal evidence and brief mentions of debris-induced hatchling mortality are available (Dial 1987; Al-Gheilani 1996; Claereboudt 2004; Özdilek et al. 2006; Santidrián Tomillo et al. 2010). Our in situ experiments used some of the most common items found on beaches to show that almost all pose an acute entanglement or entrapment threat. This has significant implications for conservation efforts on nesting beaches, whose goal is to ensure that as many hatchlings as possible reach the sea. It is also relevant for beach clean-ups—the largest volunteer activity in the world for the marine environment.
MATERIAL AND METHODS
This study was conducted from mid-July to mid-September 2004 in the Gulf of Fethiye (lat 36°13′12″N, long 29°4′24″E), southwest Turkey, Mediterranean Sea. Fethiye Beach is one of 17 key Turkish sea turtle nesting sites (Margaritoulis et al. 2003), with a relatively flat beach slope without conspicuous topography. Of a total of 58 nests in 2004 (Ilgaz et al. 2007), 19 were monitored at hatching between 2200 and 0500 hrs by teams of 3 observers. The number of hatchlings observed per nest varied considerably depending on the number emerging on a particular night. Special dim diode flashlights (Kingbright Hyper Orange L-7104SEC-H) with maximum transmission at 630 nm were used for nest surveillance and to follow crawling hatchlings if necessary; these elicited no visible reaction from hatchlings. Two sets of experiments were conducted: determination of natural crawling speed and hatchling reaction to marine debris barriers.
Natural Crawling Speed
The total time, total distance, and pattern of every crawl was recorded from the top of the nest until shortly before the water's edge. Timing began after a hatchling started to crawl in one direction. At the end of the crawl, hatchlings were caught, weighed (± 0.05 g), sized (± 0.05 mm; straight carapace length, width, carapace height, and flipper length), and released.
Hatchling crawls were broken down into 4 categories: 1) crawl (turtle moving, separated according to surface type), 2) number and length of pauses (immobile except for head movement), 3) upside-down position (limbs moving or not), and 4) obstacle encounter (forward movement hindered). The sequence of these categories and their durations were timed with stopwatches. The overall distances were divided into 2 in order to compare the crawling speeds of the first vs. the second half of the crawl.
Two subsets of hatchlings were distinguished. One subset (n = 24) involved unrestrained hatchlings. The second subset involved hatchlings retained in a plastic cage (1 cm mesh size) until earlier hatchlings completed their crawls, so that the animals could be individually timed.
We also determined the hatchlings' “orientational motivation”—the degree to which a hatchling's perception of correct orientation is established, reflected here in the number of pauses (see Discussion).
The grain size composition at the surface (sand 0.125–2 mm, pebbles 2.1–30 mm, coarse pebbles 30.1–80 mm, and cobbles 80.1–150 mm) between the nests and the sea was determined visually (sand) and with calipers (larger grain sizes) at visible boundaries. Dry sand, wet sand, and sea grass (Cymodocea nodosa) were also distinguished. Sand temperatures (surface, 15-cm depth) and air temperature (2 cm above surface) next to the nest were recorded with a digital thermometer (Moneray International LTD DT250; ± 0.1°C) at the beginning of each crawl, along with inclination and declination of the moon (protractor and compass bearing in degrees, respectively).
Marine Debris Barriers
Four types of items, all plastic, were selected to cover the range of sizes, weights, and shapes of marine debris known to accumulate on beaches worldwide and that hatchlings would typically encounter. The items were arranged in rows (between 3.6 m and 5 m long) on a smoothed sand surface at least 3 m from the nest. Experiments were conducted on each item category separately. The configurations (spacing) and number of elements were chosen to maximize the encounter rate by hatchlings but to enable evasion.
-
Empty plastic bottles (1.5 l): lightweight and large relative to hatchling, yet with opening too small to enter; beverage bottles are listed on the “Top Ten List of Debris Items” collected worldwide (Ocean Conservancy 2009). Sixteen bottles, positioned lengthwise, parallel to waterline in 2 staggered rows (each ca. 5 m long), 50 cm between rows, 5 cm between bottles.
-
Styrofoam cups (175 ml): lightweight, small internal volume relative to hatchling, entry possible; small plastic cups or other drinking containers are also on the Top Ten List (see above). Positioned in 1 continuous row (ca. 3.6 m long) with openings facing nest.
-
Plastic canisters (5 l): floor cut away to create large opening; heavy and large compared to hatchling, with large internal volume, entry possible; representative of a wide range of household and industrial containers. Positioned side-down with large openings facing nest in 2 staggered rows (ca. 2 m long), 50 cm between rows, 25 cm between canisters.
-
Fishing nets: combined textile and plastic fibers; mesh size 38 mm, twine diameter ca. 0.75 mm. Two configurations: heaps (100 × 25 × 8 cm) and single layers (100 × 25 cm). Both were presented in a ca. 3-m-long row with 10 cm space between individual nets or heaps.
Delays caused by debris were recorded as the time between first contact and successful passage. Entrapment for 10 min was considered “permanent”; such crawls were recorded as “unsuccessful” and the experiment terminated and the hatchling released. “Successful” hatchlings overcame the obstacles before the 10-min limit. Most experiments were recorded using a digital video camcorder (Sony DHC-HC30E) with infrared light.
To determine overall debris density, a survey was conducted at the study site during the nesting season from 14 July to 30 August 2004. Three randomly chosen stretches were examined. Within the first and second stretches, 3 samples, each with 4 transects (100 m long and 1 m wide), were taken parallel to the waterline. The first transects were positioned at the high tide line, the fourth near the vegetation line. The distance between each sample was at least 100 m. The distance between transects ranged from 4 to 14 m, depending on beach width. Within the third stretch, only 4 transects were taken (total transect area = 2800 m2). All visible objects in the sand surface were counted, classified, and measured (for more details and figures on both the debris survey and the experimental debris setup, see Triessnig 2006).
To calculate the average contact frequency of hatchlings with marine debris items other than the experimentally positioned ones, we used the formula of Laplace in Schärf (1994):
whereby P(E) is the probability of encounter, g the number of encounters and m the total number of hatchlings. Statistical analyses were performed with Microsoft Excel XP and SPSS software version 13.0. Multiple regression analysis with stepwise elimination of nonsignificant terms (“overall model”) was employed to estimate important independent variables contributing to crawling speed (F-value for accepting ≤ 0.050; F-value for excluding ≥ 0.100). Statistical tests are 2-tailed with α set at the 0.01 (*) or 0.05 level (**).
RESULTS
Natural Crawling Speeds and Times
We evaluated 173 hatchlings (not all variables apply to each hatchling, therefore different hatchling numbers are stated below). The average overall speed was 1.64 cm sec−1 (range: 0.35–7.44 cm sec−1). The overall crawl durations varied between 55 sec and 3 hrs, 35 min, 26 sec (mean = 16 min, 19 sec), equivalent to a factor of 235. Distance, crawling speeds, and obstacles determined the time spent on the beach. Crawling speeds differed considerably between and within nests: within a nest, this was true even for simultaneously emerging animals (e.g., 2.14 vs. 7.41 cm sec−1), whereby hatchlings emerging on later nights (up to 5 days after first hatch) tended to be slower than those on the first night (R2 = 0.031; p = 0.023; n = 170)** (r = −0.175).** According to the overall regression model presented here, speed was mainly determined by sediment grain size (standardized β coefficient −0.558), surface temperature (0.525), and orientational motivation (−0.186) (Fig. 1). Height, phase, and position of the moon also played a role (Fig. 2e, f), as did hatching date and time spent retained. Hatchling mass and size (i.e., carapace width, height, flipper length as well as various indices of these measurements) had no significant effect on crawling speed.
Citation: Chelonian Conservation and Biology 11, 1; 10.2744/CCB-0899.1
Citation: Chelonian Conservation and Biology 11, 1; 10.2744/CCB-0899.1
Crawl times were primarily affected by substrate type. Speeds decreased in the following order: wet sand > dry sand > pebbles > sea grass (Cymodocea nodosa) > coarse pebbles > cobbles (Fig. 2b) and rapidly decreased with grain size (R2 = 0.29; p < 0.0001; n = 172)* (r = −0.538) (Fig. 2c, d). The somewhat steeper final sections of the beach before the wet sand had the biggest grain diameters and yielded the slowest values. According to Fig. 2d, a cobble diameter of ca. 12–15 cm marks the size limit for successful crawls: even at somewhat smaller diameters, hatchlings tended to become entrapped headfirst in larger interspaces (Fig. 3). Fastest speeds were recorded at the end of the crawl on wet sand. Most hatchlings traversed all 6 substrate categories (average 34.7% dry sand, 9.7% pebbles, 34.1% coarse pebbles, 2.2% cobbles, 3.4% sea grass, and 16% wet sand). The grain size distributions were similar along the beach, enabling a comparison of overall crawls based on temperatures.
Citation: Chelonian Conservation and Biology 11, 1; 10.2744/CCB-0899.1
Surface temperature (23.5° ± 2.3°C SE, n = 171, all measured next to the nests) was linearly correlated with crawl speed (Fig. 2a; R2 = 0.381; p < 0.0001; n = 169)* (r = 0.617). At coolest temperatures (16°–18°C), speeds were 5 times slower than at 26°–28°C. Surface temperature reflected air temperature, which decreased during the hatching season and during the night. Thus, hatchlings retained in mesh cages were cooler and ran slower (although an energy loss in trying to escape from the cage may play a subordinate role). Sand temperature at 15 cm depth (where hatchlings may aggregate before emerging and which presumably reflects their immediate postemergence body temperature) had no significant effect on crawling speed.
Crawl times were also affected by pauses and overturning. Only 11 animals (6.4%) traversed the beach without interruptions. The average overall pause duration of the 54 hatchlings whose pauses were timed (rather than merely counted) was 39 s crawl−1 (range: 2 sec–4 min, 26 sec). The average number of pauses per hatchling was 12.3 (range: 1–151). Slower-crawling hatchlings made more pauses (r = −0.433; p < 0.0001; n = 169).* Number of pauses decreased over a run (first half: x¯ = 6.9 ± 9.5 SD; second half: x¯ = 5.6 ± 8.7 SD) (p < 0.0001; n = 161; nonparametric 2-tailed Wilcoxon test).** Unrestrained hatchlings made more pauses than those retained (p = 0.009; n = 132; Mann-Whitney U-test, 2-tailed).* Moreover, the longer they were restrained in the mesh cages, the fewer pauses they made when their crawl began (p < 0.0001; n = 109; nonparametric Kruskal-Wallis-Test).*
Overturning increased crawl time. On average, a hatchling spent 40.5 s crawl−1 (range: 1.2 sec–4 min, 39.2 sec; n = 95) overturned (average of 2.1 overturning events crawl−1; n = 172). Overturning was typically associated with steeper coarse pebble/cobble stretches or stronger irregularities in the sand (human footprints and horse hoofprints). Subsequent reorientation consumed additional time.
The average distance crawled was 14.7 m (range: 2.8–129.8 m). Distance did not significantly affect speed, although the average speed during the first half of the nest–sea distance (6.9 cm sec−1) was somewhat higher than in the second half (5.6 cm sec−1); however, 90 out of 171 individuals crawled faster in the second half. Note that although the fastest crawl speeds were on the wet sand at the end of the crawl; this stretch was relatively short and did not affect overall crawl speed.
Orientational stimuli directly affected speed. Fastest values and quickest orientation were recorded when the moon was full and at a flatter (< 30°) angle (Fig. 2e), either behind (landward of) or in front of the nest (Fig. 2f); weaker/conflicting cues caused more pauses and slower speeds. Some rapid runs also occurred on moonless nights. Distant city lights caused almost all runs to deviate to one side (to the left, or southeast). Artificial beach lighting from a hotel was responsible for the longest-distance run (129.8 m; 81 min). After these hotel lights were turned off, 6 individuals from that same nest crawled only 8–12 mins: the light therefore increased time (and distance) by a factor of 8. The longest time (3 hrs, 35 min, 26 sec for 77.7 m) reflected a worst-case scenario: cold temperature, poor orientation, and cobbles.
Marine Debris
None of the 199 hatchlings tested actively evaded experimentally deployed marine debris obstacles, i.e., all made initial contact. After contact, hatchlings were able to circumvent or overcome only the plastic bottles; most were “permanently” entrapped (in cups and canisters) or entangled (in nets) (Fig. 4a).
Citation: Chelonian Conservation and Biology 11, 1; 10.2744/CCB-0899.1
All hatchlings in the plastic bottle experiment (n = 33) initially touched a bottle. Eighty-eight percent then circumvented the bottle, whereby about half (58%) displaced it up to 10 cm. Twelve percent went beneath the bottle (displacement by up to 24 cm). The average time spent at a bottle was 21 sec (range: 2 sec–1.4 min; highest values reflect long pauses in front of the bottle, indicating orientation problems).
The lowest success rate was at the cut-open canisters, where 84% (n = 44) were permanently (i.e., 10 min) trapped. These turtles crawled inside up to the seaward-facing end and attempted to continue going straight forward by butting against the wall or against the somewhat funnel-shaped, capped opening (Fig. 5). Canisters also delayed the 7 successful hatchlings (16%) the longest (median: 1.4 min, range: 19 sec–3.3 min; Fig. 4b). Some of these hatchlings escaped by crawling along the inner wall until they reached the open end, others crawled haphazardly.
Citation: Chelonian Conservation and Biology 11, 1; 10.2744/CCB-0899.1
Heaped fishing nets trapped 82% of the animals (n = 44). Unsuccessful hatchlings crawled an average of 15 cm on the net before entanglement and continued to crawl forward against the net's resistance; none moved backward to escape. Eight percent crawled beneath the net but also became trapped. The animals were entangled by an average of 7 loops (range: 1–22), mainly in the front part of the body (68%: around the head, neck, and/or front flippers) (Fig. 6). Even a single loop permanently stopped the animals. Three of the hatchlings released after 10 min were upside down. Only 8 hatchlings (18%) were successful: 2 became temporarily entangled by 1 loop (head entanglement 44 sec; front flipper 9 min), but both escaped by slightly changing direction. The remaining 6 individuals circumvented the net after initial contact.
Citation: Chelonian Conservation and Biology 11, 1; 10.2744/CCB-0899.1
At the single-layered nets, 55% (n = 29) were unsuccessful. One difference to the heaped configuration was that these unsuccessful hatchlings were able to crawl farther (average 35 cm; range: 4–63 cm) before being stopped. The hatchlings were entangled by an average of 7 loops, but all remained upright. One turtle crawled beneath the net but was stopped after 10 cm by a loop around its neck. Thirteen individuals (45%) crossed this net. Six became temporarily entangled by 1 loop but escaped (2.5–3 min); the remaining 7 required 16 to 53 sec, depending on the number of pauses on the net.
The capture rate at the Styrofoam cups was 41% (n = 49). As in the canister, the turtles continued to butt against the inside end facing the sea. One hatchling reentered the same cup after tumbling out. Of the successful animals, 35% escaped by standing upright against the side wall, sliding back down sideways with their head facing the opening and therefore effectively turning around. The second group (12%) escaped after overturning; these animals initially stood upright in the cup and then tumbled out on their backs. After up to 1 min, they flipped back over and continued crawling towards the sea. The remaining 12% pushed the cup away after bumping against its rim. Again, none escaped by going backward. The average time spent in the cup was 1.4 min (range: 7 sec–8.5 min). In contrast, those that pushed the cup away required only 22 sec (range: 7–42 sec) to pass the array. In general, body weight and body size did not significantly affect the success rate at any obstacle.
The beach debris survey yielded a total of 2871 objects, equivalent to an overall density of 1.03 items m−2. The most common category of litter was plastic (47.6%). Overall (2 experimental setups: crawling speed and debris), 372 hatchlings from 19 different nests encountered 247 (not experimentally deployed) marine debris items on their way to the sea. The average frequency of contact was 0.66, i.e., 2 out of 3 hatchlings touched an object lying on the beach. Some items that hatchlings contacted were similar to those in the experimental arrays, i.e., plastic bottles or plastic cups. The time they needed to overcome these different litter items ranged from 3 sec (plastic bottle) to 35 sec (plastic cup).
DISCUSSION
The dash of hatchlings to the sea is the briefest period in the life cycle of sea turtles. Its duration and threat potential depend on numerous factors such as the nest–sea distance and the presence or absence of predators. Hatchlings show a frenzy motile behaviour during their crawl to the sea, apparently to reduce predation (Dial 1987), whereby difficulties in “sea-finding” may increase the probability of being predated (Santidrián Tomillo et al. 2010). This study identifies and quantifies a series of previously neglected factors influencing crawl speeds and times: substrate grain size, substrate temperature, orientational motivation, and marine debris. We believe that sand mining (which contributes to suboptimal grain sizes), artificial lights (which contribute to poor orientational motivation), and expanding beach bars and restaurants, combined with marine debris, help explain the long-term decline in nest numbers recorded here (Ilgaz et al. 2007).
An unfavorable combination of the above factors led to crawling speeds only 15% of those reported for loggerhead hatchlings elsewhere (Dial 1987); genetic factors alone would probably not fully explain this difference. Speeds decreased linearly with temperature; larger grain sizes slowed down hatchlings considerably, and conflicting orientational cues (artificial lights) led to longer-distance crawls, slower crawling speeds, and more pauses. Most debris types we tested stopped crawls entirely.
Substrate characteristics influence adult nest site selection (Kikukawa et al. 1999) and therefore directly determine hatchling crawl distances and durations. The larger, smooth-surfaced pebbles at Fethiye Beach prolonged crawl times because they are difficult to negotiate by hatchling gaits (alternate-limb crawling, crutching; Wyneken 1997), leading to ineffective limb movements in the air. Hatchlings also became trapped in the interspaces between large pebbles or cobbles. Finally, overturned individuals were recorded almost exclusively on such large grain sizes. Even optimal grain sizes had differentiated effects: speeds were slower on dry vs. wet sand, which we attribute to more energy being expended on (dry) sand displacement than on forward movement.
Temperature plays a major role in hatchling crawls. Multiple regression analysis ranks the importance of temperature only slightly below grain size. Crawling speeds were clearly slower at colder surface temperatures. At night, hatchlings emerge from warmer subsurface temperatures and then cool off. At our Mediterranean site, the difference at 15 cm and the surface exceeded 10°C, with air temperature dropping to 17°C. Hatchlings emerging later at night, as the temperature drops, will be slower and should become slower the longer the run. Colder animals may also have greater difficulty surmounting obstacles. At 15°C, locomotory integration becomes degraded (Davenport et al. 1997). This is supported by our observations: certain individuals whose runs had come to a virtual standstill during the night resumed their runs when warmed up by the morning sun.
Our results reveal the importance of a third factor, which we define as orientational motivation: the more highly “motivated” or certain a hatchling is about its direction, the faster its crawling speed. Numerous studies document the importance of light in the sea-finding orientation of young sea turtles (Mrosovsky 1970; Witherington 1995; Tuxbury and Salmon 2005). Our study confirms that light influences the direction of a run, but also reveals that it affects crawling speed itself. The moon behind the hatchlings favored faster speeds (probably due to stronger reflection on the sea surface), as did the moon ahead over the sea at a shallow angle. Those hatchlings also made fewer pauses. We conclude that pauses involve orientation rather than rest because their number decreased over a run and because they were negatively correlated with speed. Moreover, unrestrained hatchlings made more pauses, and retained animals (having had more time to orient themselves) made fewer pauses the longer they had to wait before being released from enclosures. We therefore used the number of pauses as an index of orientational motivation. Note, however, that running fast is advantageous only if the direction is correct.
Brief periods of nonmovement have been reported in hawksbill turtle hatchlings during their offshore migration. Chung et al. (2009) assumed that this is an adaptive strategy against predators that mimics inanimate floating objects. Whether such pauses make hatchlings less prone to predators on the beach as well would require a separate study but is unlikely considering the frenzied motile behavior of hatchlings during their way to the sea (Dial 1987).
From an evolutionary perspective, barriers such as beach wrack have apparently never posed a threat to sea turtle populations, perhaps explaining why hatchlings never actively avoided obstacles. Modern marine debris represents a new, lethal dimension, simply because of the large additional increase in the overall density of material, coupled with the entrapment potential inherent in the cavities of plastic debris and net mesh-works. Our most surprising result was that hatchlings, beyond being captured by nets, were typically unable to escape even from “simple” items such as canisters or drinking cups. The slightly better success rate in escaping from cups is probably size-related: hatchlings that stood upright inside cups more easily slid or tumbled out directly than when in the larger canisters. Entrapped hatchlings always showed the same behavior pattern. Upon entering a container, they crawled to the seaward-facing end and attempted to go straight forward by butting against it (Fig. 5). Container translucence, i.e., the outside illumination shimmering through from the sea, may have promoted this directionality. Even if uncapped, the spouts of such canisters would be too small (and the inner walls too smooth) to permit hatchlings to escape in the seaward direction.
Fishing nets, regardless of material or age, cannot be torn by a hatchling, making entanglement in even a single loop fatal. Mesh size apparently plays little role: green turtle hatchlings, for example, were entangled and killed in considerably larger meshes than deployed here (Al-Gheilani 1996). According to G. Minton (pers. comm., cited in Claereboudt 2004), hatchlings also failed to reach the ocean after entanglement in monofilament fishing lines abandoned on the beach. Such lines often form tangled clumps of loops, creating a net-like structure (Ocean Conservancy 2009). In both fishing net configurations, hatchlings continued crawling forward against the resistance. Crawling backwards is not a motor pattern found in the crawling frenzy of sea turtle hatchlings (J. Wyneken, pers. comm.). Although presumably not disadvantageous on undisturbed beaches, this is clearly unfavorable on modern, debris-strewn beaches.
At the study site, which had a litter density of 1.03 items m−2 (Triessnig 2006), 247 of 372 hatchlings, i.e., 2 out of 3, encountered litter. Even higher densities elsewhere (e.g., Japan: 3.41 items m−2 [Kusui and Noda 2003], coast of Panama: 3.6 items m−2 [Garrity and Levings 1993], Baja California, Mexico: 1.5 items m−2 [Silva-Iñiguez and Fischer 2003]) would mean more contacts and captures. Dense debris may also increase the recapture rate of hatchlings that escape a previous entanglement or entrapment. Debris orientation is important. At our study site, the openings of nearly 45% of all cups counted faced landwards—the harmful orientation (Triessnig 2006). In our experience, the closed ends of cups face the wind, so that the common onshore winds would promote the disadvantageous orientation.
The factors of grain size, temperature, and orientational motivation clearly act in concert to affect crawling speed. In a positive example, the highest speeds on wet sand at the end of the run probably reflect the harder, even surface structure; the strong cumulative orientational motivation at the end of a crawl; and perhaps even warmer sand due to incoming waves (in August, ca. 10°C difference between water and sand/air temperature at night). In a negative example, characteristic for Fethiye Beach, these factors reinforce each other when their individual states are poor (large grain size, cold air and substrate, poor orientational motivation), and are further amplified by other disturbances such as light pollution. Crawls that deviate to one direction increase the crawl distance and lower speeds due to cooling. They also aggravate overturning because oblique crawls mean longer distances across pebble/cobble zones. The final factor we identify is marine debris. Lengthier crawls also increase the chances of encountering such items. Debris not only affects crawl times, but actually traps and kills hatchlings. Finally, slower, longer crawls and entrapment expose hatchlings to predation (Stancyk 1995), hyperthermia, desiccation (Bustard 1972), and human-related mortalities (vehicles) (National Research Council 1990) and may impact imprinting mechanisms (Barrett 2004).
The combination of critical factors we identify and quantify impact the sea turtle population in Fethiye. Thus, this beach, which has wide sections of larger grain sizes and in which beach debris is not removed (and which suffers from light pollution), show a long-term decline in nest numbers, whereas nearby beaches that consist entirely of sand, that are cleaned, and that have no artificial lights show no such trend (e.g., on Dalyan Beach; Ilgaz et al. 2007; Türkozan and Yilmaz 2008).
These results have direct implications for efforts to protect endangered sea turtle species here and elsewhere. Conservation efforts must focus on eliminating the factors that prolong the time spent on the beach during this crucial life cycle phase. One focus should therefore be on changes in beach grain size compositions and distributions, particularly an increase in larger pebble and cobble components. The slower crawls under cooler conditions also mean that sea turtle programs should avoid delayed hatchling releases. Greater efforts must be made to address light pollution; artificial lights increase crawl distances and reduce orientational motivation, slowing down crawling speeds. Coastal cleanups—one of the largest and most effective volunteer-based conservation projects in the world (Ocean Conservancy 2009)—should be scheduled before the hatching season (today, annual cleanups typically take place later, in mid-September) and focus on critical items such as canisters, drinking cups, and fishing nets. This calls for environmental awareness of all stakeholders to promote an obstacle-free start in life for all sea turtle species. Beach quality in the broadest sense (Alkalay et al. 2007) therefore represents a rallying point for conservation efforts, underlining the direct link between marine pollution, habitat integrity, and species protection.
Conceptual model of key factors influencing hatchling crawling speed (+, positive effects; −, negative effects). More pauses reflect low orientational motivation (and speed). Overturning increases with grain size. Worst-case scenarios involve combinations of cold temperature, poor orientation, and large grain sizes: cobbles and marine debris can stop crawls altogether.
Hatchling crawl speeds, excluding pauses and obstacles, related to a) surface temperature, individual crawls; b) substrate categories; c) average grain diameter of each crawl (pure sand crawls on left); d) crawls additionally broken down according to individual substrate categories (or mixtures thereof in some timing units); e) height of moon (0°, below horizon) and f) its compass direction in degrees; 2 peaks mark landward and seaward positions.
The strip of largest grain sizes on the beach, cobbles > ca. 10 cm in diameter, can stop hatchling crawls entirely and kill them. (Photo: M. Stachowitsch.)
a) Hatching success in overcoming marine debris items (shaded bars = unsuccessful, open bars = successful hatchlings). b) Time successful hatchlings spent at the obstacles. Thick horizontal lines in the boxes represent median values, dots are the outlying values, and asterisks extreme values.
Five-liter canisters with cut-off bottoms—simulating a common marine debris item—entrapped the most hatchlings and retained those that escaped the longest of any debris. Left panel depicts overview; right panel shows detail. (Photos: P. Triessnig.)
Both a single layer of net and heaped nets typically entangled hatchlings with several loops around neck and limbs. (Photo: P. Triessnig.)
