Brushes and Shelters as Low-Cost Environmental Enrichment Devices for Loggerhead Turtles (Caretta caretta) During Rehabilitation
ABSTRACT
Environmental enrichment (EE) can significantly improve the welfare and health of captive animals. Despite the known benefits of EE, there has been a lack of research on the effects of EE devices (EED) for wild animals. Here, we quantified the time loggerhead turtles (Caretta caretta) in a rehabilitation center spent interacting with EEDs, specifically brushes and sheltering objects. We found turtles spent up to 50% of their time interacting with EEDs and demonstrated a preference to rest under sheltering objects (assisted resting), when compared with the open. Based on these findings, we propose the inclusion of brushes and sheltering objects as low-cost EEDs for improving the welfare of sea turtles in captivity.
A widespread conservation intervention for sea turtles is to treat injured or diseased animals at rehabilitation centers (Ullmann and Stachowitsch 2015; Monreal-Pawlowsky et al. 2017). Recovery may take several years, so efforts to improve the welfare of animals during rehabilitation is of prime importance. Environmental enrichment (EE)—defined as the enhancement of captive environments based on behavioral biology—can increase welfare, reduce stress, and minimize further risk of illness for animals in captivity (e.g., Newberry 1995; Alaniz et al. 2022; Case et al. 2005). Environmental enrichment devices (EEDs) are recommended for improving the welfare and reducing stereotyping in captive sea turtles (USFWS 2019), but little information exists on the usefulness of EE to sea turtles undergoing rehabilitation (Diggins et al. 2022; Escobedo-Bonilla et al. 2022).
According to the five freedoms paradigm of animal welfare, expression of natural behavior is critical for positive welfare, and can also be considered a basic right of captive animals (McCausland 2014; Webster 2016). Both wild and captive turtles have been documented scratching flippers, heads, and carapaces on coral, rocks, and marine debris (Keuper-Bennett and Bennett 2000; Frick and McFall 2007). It has been proposed that this scratching allows the turtle to remove epibionts (Frick and McFall 2007). Scratching may also be desirable to remove dead skin, as is the case in many other vertebrates (Meynecke et al. 2023). The sensation of scratching may activate the neural reward pathways (Martínez-García and Lanuza 2018).
Assisted resting (whereby the turtle rests ‘anchored’ or ‘wedged’ under corals and rocks) is a common behavior in wild hawksbill (Eretmochelys imbricata) and green turtles (Chelonia mydas). Assisted resting most commonly occurs during the night, and is thought to be an important behavioral strategy with ecological and energetic benefits (Houghton et al. 2003; Stimmelmayr et al. 2010; Reisser et al. 2013).
No studies to date have looked at the use of objects to promote important natural behavior such as scratching and assisted resting in turtles during rehabilitation. Previous studies examining EE in sea turtles are limited and have focused on the use of polyvinyl chloride (PVC) pipes or nutritional enrichment. Specifically, the use of PVC pipes reduced injuries from conspecifics in group-housed green turtles (Kanghae et al. 2021). Food provisioning in PVC pipes or other EEDs also led to a reduction in stereotyped swimming in green and loggerhead turtles (Caretta caretta) reared in captivity (Therrien et al. 2007). The use of nutritional enrichment (including crabs and jellyfish) has been proposed to encourage natural feeding behavior in loggerhead turtles (Monreal-Pawlowsky et al. 2017).
The most detailed enrichment study to date assessed stereotypical swimming in captive-reared sea turtles before and after addition of EEDs (Therrien et al. 2007). The sole use of stereotypical swimming as a welfare indicator is problematic because it will only enable cases of poor welfare to be detected. There is a need for more research that focuses on the quantification of varying degrees of moderate to good welfare and the prevention of initial poor welfare.
To encourage expression of natural behavior, we tested 3 simple, low-cost EEDs for use with loggerhead turtles in rehabilitation centers: 1) a shelter to imitate corals and rocks that wild turtles have been observed resting under (Kirkden and Pajor 2006; Sobin and Tucker 2008), 2) brushes to allow scratching, 3) a pipe to allow assisted resting. The main objectives of this study were 1) to determine how much time loggerhead turtles interact with EEDs; and 2) to assess preferences in loggerhead resting behavior.
METHODS
Research was conducted at Arca del Mar, a marine rehabilitation center at the Oceanogràfic aquarium in Valencia, Spain. All procedures complied with the Oceanogràfic Animal Care & Welfare Committee (project reference OCE-32-22).
Animal Information. — This study assessed 6 wild loggerhead turtles (Table 1) caught as bycatch by fisheries. Two female (mean curved carapace length [CCL] = 34.5 cm, standard deviation [SD] = 7.78) and 4 male (mean CCL = 37.75 cm, SD = 1.44) loggerhead turtles were used. These turtles were medically treated at the Arca del Mar and were declared clinically healthy by veterinary staff before inclusion in this study. Turtles had been housed in the center for a mean of 19 d (SD = 22.58).
During experiments, each turtle was housed in a single circular tank (3 m in diameter and 1 m in depth; Fig. 1). The water temperature was 24°C; salinity ranged between 39.33 g/L and 40.70 g/L, and turtles were kept on a natural lighting cycle. Turtles were provided with natural feed (a variation of herring, squid, hake, cod, and jellyfish with carrots and spinach) 4 d/wk, at 3% to 5% of total body mass. Turtles were individually housed in the experimental tank and rotated every third day. Fourteen-gram Accelerometers (Axy-trek Marine, TechnoSmart Europe) were attached to turtles for a different study using Velcro and T-Rex waterproof tape. Accelerometers did not exceed 1% of the individual turtle’s body weight.
Citation: Chelonian Conservation and Biology: Celebrating 25 Years as the World’s Turtle and Tortoise Journal 22, 2; 10.2744/CCB-1596.1
Enrichment Design. — Three EEDs were tested: 1) A brush, to allow for scratching: 3 large brush heads were attached to a metal pole. Two brush heads were medium-sized bristle brushes (DocaPole), while the final brush head consisted of longer, plastic bristles (Bosque Verde Outdoor broom; Fig. 1). The metal pole was fitted with a 5-kg disk weight and hung from the side of the tank. The brush heads were suspended approximately 30 cm from the tank floor; 2) A shelter, to allow assisted resting and hiding: a solid plastic pet bed (95 × 70 × 29 cm) was inverted and fitted with 2 × 3-kg disk weights; 3) A PVC pipe, to allow for assisted resting: the experimental tank had a PVC pipe fitted with a gap underneath to allow for water filtration.
The EEDs used in this study were designed to aid the specific behavioral needs of resting and scratching. Furthermore, the EEDs were designed to require little cognitive demand to ensure ease of use: the positioning of the brush close to the floor allowed for the option of initial learning through incidental interaction. To prevent accidental ingestion of bristles, the brush was removed every night and returned to the tank each morning because individual turtles could only be monitored for ∼ 8 hrs on the first day of exposure to the brush. EEDs were returned to the same location each day (Fig. 1).
Behavioral Filming. — Direct overhead recordings were taken on a GoPro hero 11. Filming of turtles began the day after moving to the enclosure. Six turtles were recorded for a minimum of 7 hrs. To document scratching behavior without accelerometer attachments, accelerometers were removed from 5 turtles after a mean of 5 hrs 30 min (SD = 0.05) and carapaces were cleaned. These five turtles were recorded without accelerometers for a mean of 2 hrs 39 min (SD = 0.06).
Data Analysis. — Videos were analyzed to determine the time spent performing each of the behaviors presented in Table 2. The time spent performing behaviors was documented for turtles with both accelerometers attached, and those with a clean carapace without attachments. The total time spent performing behaviors, regardless of attachments, was also calculated. The percentage time spent performing behaviors in Table 2 was then calculated for 1) overall behavior, regardless of attachments; 2) behavior with accelerometers attached; and 3) behavior with a clean carapace. Percentages were calculated as time spent performing behavior/time observed × 100.
Statistical analyses were conducted in Program R (v. 4.2.2). Code and raw data are available at https://github.com/jharv3y/Loggerhead-enrichment-analaysis. To assess the potential impact of accelerometers that were attached for a different study on behavior, a linear mixed-effect model was conducted. The model assessed the effect of accelerometer attachment on percentage time spent interacting with EEDs (behaviors listed in Table 2). Pipe interactions were not included in the analyses because of limited sample size (n = 4). Turtle ID was treated as a random effect (R packages “lme4” and “lmerTest”). The presence of accelerometers was not found to alter time spent performing behaviors (Table 3). Individual turtles exhibited small variation in behavioral allocations (Variance = 0.45, SD = 0.67, n = 5). The accelerometers did not significantly alter EED use, so the attachments were disregarded.
A fundamental concept in animal welfare is that animals will actively choose something that is in their best interest. Preference testing is used to measure an animal’s choice for a certain behavior (Kirkden and Pajor 2006; Hemsworth et al. 2014). To investigate resting preferences, the percentage of time spent in assisted resting (resting ‘anchored’ to an EED) was combined and compared with time spent under the “Resting exposed” behavior. Data did not fulfil the assumptions of a parametric t-test; therefore, a Wilcoxon signed rank test was conducted.
RESULTS
EED Use. — All turtles used both the brush and shelter. Four out of 6 also interacted with the pipe. Turtles spent a mean of 29.67% (SD = 11.75) of time engaging with enrichment devices (Range = 19.66%–50.00%). Turtles interacted with the three EEDs unevenly (Brush: mean = 12.87%, SD = 17.0. Shelter: mean = 14.22%, SD = 11.08. Pipe: mean = 2.58%, SD = 3.11 [Fig. 2]).
Citation: Chelonian Conservation and Biology: Celebrating 25 Years as the World’s Turtle and Tortoise Journal 22, 2; 10.2744/CCB-1596.1
Turtles spent a mean 6.5% (SD = 8.48) of time scratching (Fig. 2), and 6.32% (SD = 9.87) of time using the brush for assisted resting. Strong variation in percentage time scratching was observed between individuals (Range = 0.51%–28.28%).
Resting Preferences. — All turtles spent significantly more time in assisted resting than exposed resting (Variance = 21, p = 0.031. Assisted resting: mean = 16.65%, SD = 8.3; Exposed resting: mean = 6.15%, SD = 5.39 [Fig. 3)]. In one incident, a turtle was documented sleeping under the shelter after the unit lights were turned off.
Citation: Chelonian Conservation and Biology: Celebrating 25 Years as the World’s Turtle and Tortoise Journal 22, 2; 10.2744/CCB-1596.1
DISCUSSION
We present the first quantification of EED use by loggerhead turtles in rehabilitation settings. Previous EE research has focused on captive-reared loggerhead turtles (Therrien et al. 2007). On average turtles spent nearly a third (mean = 29.67%) of recorded time positively interacting with EEDs, demonstrating successful implementation of EEDs. One turtle was recorded interacting with EEDs half of the observed time (50.01%). All turtles engaged with the brush (scratching the carapace and head or resting) and the shelter. On average, turtles spent more time interacting with the shelter (14.22%, SD = 11.08) than the brush (12.87%, SD = 17.00) and the pipe (2.58%, SD = 3.11), highlighting the importance of providing a shelter for turtles in a rehabilitation setting. The time turtles spent interacting with the brush consisted of scratching 6.5% (SD = 8.48) and assisted resting 6.32% (SD = 9.87). The roughly equal time spent by turtles scratching and in assisted resting with the brush again highlights the importance of providing an object for turtles to undergo assisted resting during rehabilitation. The multipurpose function of the brush as both scratching and an assisted resting EED suggests that providing a single EED could improve welfare substantially, when compared with housing in empty tanks.
Turtles spent significantly more time (10.5%) in assisted resting (‘anchored’ to an EED) than resting in the open. Assisted resting in the wild likely provides benefits such as passive predator avoidance and buoyancy aid, which could result in a lower energetic cost than the cost of exposed resting (Stimmelmayr et al. 2010). Facilitating this important natural behavior during rehabilitation could reduce energetic expenditure and stress levels due to decreased vigilance. Given the preference for turtles to rest ‘anchored’ to an object, our results suggest that EEDs could contribute to the welfare of loggerhead turtles undergoing rehabilitation, increasing perceived safety and promoting recovery during illness (Warwick 2023).
Although not directly assessed in this study, it is likely that the use of enrichment objects directly promotes positive welfare for turtles in a rehabilitation setting, by encouraging natural behaviors. This may have further benefits, such as lowering stress levels, improving health, and improving recovery, leading to a greater chance of postrelease survival (e.g., Bonnet et al. 2013; Burghardt 2013; Näslund and Johnsson 2016). Furthermore, the use of enrichment devices may improve welfare, enhance quality of life, and prevent development of stereotypical behaviors for both resident and nonreleasable sea turtles (USFWS 2019).
The major limitation of the observations presented here is that we only assessed the short-term use of EEDs and did not assess the behavioral impacts of EED addition (pre- and post-behavior comparisons). Despite the aforementioned limitations, we have implemented low to no maintenance, cheap enrichment devices; loggerhead turtles during rehabilitation interacted with these devices for up to 50% of their time. Furthermore, the EEDs used in this study are easy to clean and disinfect, as well as being safe and easy to use, which are all key requirements for rehabilitation facilities for both short-term and longer-term admissions. Adoption of low-cost enrichment objects like the ones designed for this study can mitigate perceived barriers associated with enrichment implementation; namely cost and practicality (from both a husbandry and logistical view; Riley and Rose 2020; Diggins et al. 2022). Previously studied enrichment may be difficult to implement for rehabilitation centers. For example, additional enrichment in Therrien et al. 2007 for a blind green turtle involved being periodically scratched by a researcher. This reduces the accessibility of the tool to centers with limited time and personnel.
We present information demonstrating that when given ad libitum access to a brush, loggerhead turtles will perform self-regulated scratching behaviors. We also identified a significant preference for turtles to rest ‘anchored’ to an object, which should be explored in future experiments. This suggests a potential application for improving the welfare of turtles held in rehabilitation centers. Regardless of the reason behind assisted resting, we suggest turtles in rehabilitation facilities (for any amount of time) should be given the option to perform assisted resting, alongside access to brushes. EEDs that enable scratching and assisted resting should also be considered to improve welfare of turtles in core aquarium collections. The data presented here will provide relevant information for future welfare experiments for turtles.
(A) Tank setup. i: Brush; ii: Shelter; iii: PVC Pipe. (B) Brush design. Readily available, low-cost brush heads were used. For suspension in tank, a 5-kg disk weight was added, the brush heads were suspended approximately 30 cm from the tank floor. A black plastic barrier was added in the gap between the vertical brush and tank wall to prevent turtles becoming stuck. (C) Shelter. Inverted plastic pet bed with 5-kg weights attached either side, allowing the shelter to be weighted while turtles may enter and exit either side. (D) PVC pipe for exposed resting. Here a turtle has intentionally anchored herself to the gap in the pipe by the flipper and is performing assisted resting. Photographers J. Harvey-Carroll and M. Saubidet.
Ethogram for loggerhead turtles in tanks with access to environmental enrichment devices. Percentage time spent performing discrete behaviors at enrichment devices. “Shelter other” consists of biting the shelter (3 occurrences), remaining stationary on the shelter, apparent scratching of the plastron and/or rubbing on the shelter, presumed foraging behavior and trying to push under the shelter from opposite sides to the entrance. “Brush other” consists of biting the brush. ‘Other’ behavior (grey) consisted of swimming, eating and exposed resting. See Table 1 for further description of the different behaviors.
Percentage time spent in assisted resting versus exposed resting. Time spent in assisted resting was pooled across time spent resting under brush, pipe, and shelter. Boxplots depict median relative expression levels and the 25th and 75th percentiles. Whiskers are 1.5× the interquartile range, data points outside this range are marked as outliers (circles). Sample size n = 6. * indicates significant differences (p < 0.05).
Contributor Notes
Handling Editor: Jeffrey A. Seminoff
