Turtles are unique and recognizable vertebrates. With over 300 extant species and residence on every continent except Antarctica (Turtle Taxonomy Working Group 2014), they have been incorporated into many human cultures as a source of food, tools, and inspiration for folklore and creation myths. Turtles are also an ancient taxon with representatives that evolved shelled forms appearing in the fossil record over 200 million years ago (Ernst and Lovich 2009). Despite the apparent success of turtles as an enduring taxon, and the high regard in which most cultures hold these animals, turtles today are in serious decline with over half the species threatened with extinction (Turtle Taxonomy Working Group 2014). Many factors precipitating these declines are similar to those faced by all animals and include habitat loss and degradation, unsustainable exploitation for food markets and the pet trade, environmental pollution, and global climate change. However, the argument can be made that these stressors are amplified in turtles because they generally rely on individual longevity to maintain healthy populations (e.g., Congdon et al. 1994).

We examined how adult painted turtles (Chrysemys picta) responded to the presence of humans during a critical life stage, the nesting season. Can wild C. picta accommodate the presence of humans or do turtles perceive humans as a continual predation threat? Many wild vertebrates respond to human-disturbance stimuli as they would to predation attempts (Frid and Dill 2002), but some species can habituate to human presence (Ellenberg et al. 2009; Selman et al. 2013). In the wild, basking is an important and widely displayed behavior in freshwater turtles and one which could be affected by human disturbances. Turtles that cannot accommodate human presence may decline in health (Moore and Seigel 2006). With human populations growing, more turtle populations experience contact with humans throughout their life cycles. Therefore, knowing whether or not freshwater turtle species can tolerate human presence can provide useful insight into whether these species will be able to persist in their native habitats despite human encroachment.

Chelonians have numerous predators. This is especially true of hatchling and juvenile turtles (e.g., Janzen et al. 2000), but may occur at any life stage (Greene 1988). In response, chelonians have evolved a multitude of physical and behavioral defenses, including retreat into their iconic shells. Other antipredator behaviors emphasize crypsis, cloacal discharge, postural threats, and biting (Greene 1988). Of particular note is the fact that basking freshwater turtles often flee into nearby water to avoid predation. Similar escape behaviors have been studied in other taxa, particularly birds (Tarlow and Blumstein 2007; Blumstein and Fernández-Juricic 2010). There is the potential for wild animals to perceive human recreational activities as predation attempts. They may not have evolved the means to appropriately measure the threat levels of various human recreational activities. Thus, they may overreact to perceived anthropogenic threats, diverting resources to immediate survival and away from important life-history components such as reproduction. It is therefore critical to measure and understand how animals react to potential anthropogenic stressors, including human recreational activities.

The response of animals to human approach is often measured by flight initiation distance (FID), which is the distance to which a putative predator can approach a wild animal before the animal flees (Blumstein 2003; Cooper 2008, 2009; Schwanz et al. 2011). FID is highly germane to freshwater turtles because of its relationship to the basking habit, which is linked to regulation of metabolism, decrease in parasite load, and egg deposition (Ernst and Lovich 2009). Although basking is beneficial for freshwater turtles, it is potentially dangerous as it leaves the animals more vulnerable to predation. Indeed, many freshwater turtle species rarely emerge from the water except to nest or bask on supra-aquatic surfaces (Steen and Gibbs 2004; Ernst and Lovich 2009). Thus, an estimate of how sensitive basking animals are to disturbance by humans can give insight into adverse impacts on important life-history processes.

We compared the average FID in two groups of painted turtles, one exposed to human recreation and one not experiencing recreation, for which raccoons (Procyon lotor) and otters (Lontra canadensis) were the only known nonhuman predators on the adults (F. Janzen, pers. obs.). We hypothesized that painted turtles at the site with frequent recreation activity would perceive humans as less of a threat and have a shorter average FID compared with painted turtles at a field site with minimal human activity (sensu Bowen and Janzen 2008; Spencer and Janzen 2010; Strickland et al. 2010 for nesting behavior).

METHODS

The 2 field sites were approximately 32 km apart along the Mississippi River in northwest Illinois (Fig. 1). The first population was located at Lost Mound Unit (hereafter rural), an ex-military base that has long been off-limits to human recreational activities. Although this site was used to fire artillery, by 1920 it served primarily as a munitions storage and recycling facility. As the munitions storage facilities were located upland and the base had few personnel present for many years, the turtle population in the river had not been exposed to many humans for over a century. The other population is found at Thomson Causeway (hereafter urban), a popular camping and fishing area that experiences substantial human activity. For example, 7301 campers visited the area from 31 March to 31 October 2013, a figure that does not include additional day visitors (K. Zidarich, pers. comm., March 2015). The type of human activity to which a given turtle was likely exposed was based on whether the turtle was engaged in aquatic or terrestrial activity. Most of the turtles that emerged onto the land during the time period in which this study was conducted were gravid females searching for a suitable nest site location or actively nesting. These animals were likely to be exposed to human approaches on foot or approaches from vehicles. Alternatively, turtles engaged in aquatic activities such as aquatic basking were likely to be approached by on-shore fishermen or boats. Despite these various human activities, turtles at this site frequently nest in close proximity to humans (Bowen and Janzen 2008; Strickland et al. 2010).

View Figure

• Download as Powerpoint
• Download Figure

For each field site, multiple locations to assess the FID of adult painted turtles basking on logs were chosen based on ease of access and how well they represented basking locations within each site. At the urban site, 12 locations were observed that were close to popular campsites or near areas where people would regularly fish. At the rural site we observed 2 locations. Not only was human recreational activity precluded at this site, but most locations were strictly off-limits because of the potential for unexploded ordinance in the water. The locations at both sites were characterized by the presence of large logs in the water and therefore were frequented by basking painted turtles.

Every day from 26 May to 20 June 2013, one person approached the turtle(s) while a second person recorded FID. To control for climatic variables, such as daily fluctuations in temperature, FID trials were performed at both sites at the same time each day by 2 groups of 2 researchers. For consistency's sake, R.L.P. was always the approaching partner at the rural site and M.B. was always the approaching partner at the urban site, targeting turtles from the same chosen spot at each location every time FID was measured. Starting distance was standardized for each location within each site such that a researcher always approached basking turtles at a given site from the same standardized distance, but the distances varied based on location and site. At our rural site, starting distances were 47 and 109 m for 2 sites. At our urban site, starting distances ranged between 36 and 125 m for 12 sites. The recorder used a laser-range finder (model: Nikon Prostaff 550, 0.0144-m accuracy from 91 to 503 m) to obtain the initial distance between the researcher and a turtle as well as the distance between the researcher and the turtle when it fled. For the second measurement, in instances where multiple turtles fled from the log at once, we conservatively recorded only a single FID measurement taken from the center of the group of basking turtles. In instances when some turtles fled but others did not, the researcher continued to approach the turtles and, each time a new group of turtles fled, a new FID was recorded. Evidence from a basking study that occurred simultaneously with this one (and that involved trapping and marking individuals) suggests that individual turtles infrequently returned to the same spots to bask day after day (R.L.P., unpubl. data, June 2013). Thus, we expect that we rarely sampled the same turtles more than once.

We analyzed the FID data with SAS software v. 9.4 (SAS 2015). We performed a nested analysis of covariance (ANCOVA), using the model:

where the covariates tested included “date,” the day of the year on which an FID trial was performed; “time,” the exact time (military time) of day that an FID trial occurred; “site,” whether rural or urban; and “site(location),” the FID location within each field site. In this model we also tested the interactions “date × site” and “time × site” to see if the date on which FID was assessed, or if time of day at which FID was recorded, had context-dependent impacts on FID.

RESULTS

Over 26 days, we collected 97 observations from 2 locations at the rural site and 238 observations from 12 locations at the urban site. The covariate time was statistically significant, as were the interactions date × site and time × site (Table 1). For the time × site interaction, turtles at the rural site exhibited decreased FIDs later in a given day. Overall, we detected a significant difference in the approach distance tolerated by painted turtles in the 2 populations. Turtles at the rural site fled at greater distances than turtles at the urban site (Table 1). On average, painted turtles at the urban site fled at 40.7 ± 1.451 m SE, while painted turtles at the rural site fled at 60.0 ± 8.218 m SE. The factors date and site(location) were not significant main effects (Table 1).

Table 1. ANCOVA table for full flight initiation distance model with sums of squares (SS), degrees of freedom (df), mean squares (MS), F-values, and p-values. Significant terms are printed in bold.

View Fullscreen

DISCUSSION

We hypothesized that animals in populations where human presence is common would flee at shorter distances, showing greater tolerance of human activity, than those not regularly experiencing human approach. Our results are consistent with this hypothesis. We found that adult painted turtles basking on logs where human disturbance was not common did flee at much greater distances than those at the more regularly human-visited site. By presumably adjusting FID based on how often they were exposed to humans and, therefore, on the threat-level that they perceived from humans, the more urban population of painted turtles minimized the loss of critical basking opportunities that fleeing more frequently would have incurred.

This FID information is especially relevant because turtles disturbed from their aquatic basking sites can take up to an hour to return (Moore and Seigel 2006). Therefore, if turtles are consistently frightened off logs, they could potentially lose significant amounts of basking time. A reduction in basking could decrease fitness due to inability to thermoregulate and perform critical reproductive and self-maintenance tasks, including regulation of vitellogenesis (which may be especially crucial during the nesting season, when we conducted this study) as well as proper regulation of metabolism and control of parasite loads (Ernst and Lovich 2009). Indeed, turtles in poor health condition, as measured by higher parasite loads and lower white blood cell counts, may spend more time basking than their peers in better health, suggesting that turtles may actively bask as a mechanism to improve their immune system (Ibáñez et al. 2014).

Why did C. picta at the urban site exhibit a shorter FID than painted turtles at the rural site? This study did not examine all possible explanations for this difference, but one potential explanation is that the turtles at the urban site learned via habituation (Rodriguez-Prieto et al. 2007) that most humans are not a threat to them while the turtles are basking. Habituation is often considered a learning process; as animals are repeatedly exposed to a putative stressor, they eventually cease to regard it as dangerous. Such flexible behavioral capacity may help turtle populations survive in the presence of humans whose behavior is not directly harmful or lethal to turtles. Another possible explanation relates to turtle size. Ibáñez et al. (2014) found that the amount of time male Spanish terrapins (Mauremys leprosa) hid in their shells after predator attacks depended on size, with larger animals spending more time hidden before emerging. If the turtles at our 2 study sites differed in size, this could be a factor in the difference in FID between the two C. picta populations. Turtle health is another possibility. Polo-Cavia et al. (2009) studied western pond turtles (Emys marmorata) residing in 2 populations in northern California. One population resided in a pristine, unaltered habitat. The other was found at a site with altered habitat where turtles were potentially exposed to both rural and urban anthropogenic water contaminants. This study found that turtles residing in the altered habitat (higher concentrations of aquatic contaminants) had lower immunocompetence compared with their conspecifics residing in an unaltered habitat (Polo-Cavia et al. 2009). Because turtles at both sites for this study shared the same watercourse, we assumed that both populations were exposed to the same contaminants and therefore experienced the same impacts.

It is also possible that the C. picta at our urban site exhibited a shorter FID than did the C. picta at our rural site because of differences in starting distance for our FID trials at the different locations within each site. This effect has been studied within the flight initiation distance literature, particularly in birds (Blumstein 2003) and lizards (Cooper 2005, 2008). However, we think it is unlikely that the results we observed in this study were owing to an artifact of starting distance. This is due to the fact that the starting distances for the 2 locations within our rural site fell within the range of those that we used for our 12 urban sites. Thus, if starting distance had a statistically significant effect, we should have seen that effect in our ANCOVA when we tested for an effect of location.

Our results showed that, although rural turtles had larger FIDs than the urban animals overall, flight distances were especially large in the morning when turtles were cold. Turtles allowed us to approach more closely before fleeing during the warmer part of the day, perhaps because, when temperatures rise, turtles may gain confidence in their ability to escape and thus allow potential threats to approach closer before fleeing. This effect has been noted in other ectothermic vertebrates including Anolis lizards (Anolis lineatopus), tropical whiptail lizards (Ameiva festiva), and brown trout (Salmo trutta) (Cooper 2006, Lattanzio 2014, Öhlund et al. 2014). It is important to understand how other important factors influence the way turtles respond to predation threats such as approaching humans. We also found a date × site effect, but this finding is likely attributable to the scarcity of painted turtles at the rural site at the beginning of this study.

If painted turtles, in general, can tolerate close human approach when regularly exposed to nonthreatening human approach, this response may allow painted turtle populations to persist in the presence of certain recreational uses. Managers should collect more data on the response of basking turtles to different types of human approaches while including important environmental factors in order to understand the types of recreation that pose the lowest threat to species that depend on basking.