Understanding how and why animals move is a key topic in animal spatial ecology (Morris, 2003). This knowledge assists in predicting where animals will occur and how they will move, and further informs how animals will respond to anthropogenic influences such as climate change or habitat degradation (Sakabe and Lyle, 2010; Speed et al., 2012; Vaudo and Heithaus, 2013; Schlaff et al., 2014). However, several literature reviews have shown that most tracking studies report on the when and where animals occur and only a few have identified the drivers (Sims, 2010; Hammerschlag et al., 2011; Flowers et al., 2016). Furthermore, habitat selection will be influenced by the link between an animal's physiology and its external environment. The physiological state of an animal will be optimized under a subset of environmental conditions and animals may avoid habitats where such environmental conditions are sub-optimal (Brown et al., 2004). These environmental thresholds may be species- or population- specific, allowing each to exploit different ecosystems (Fangue and Bennett, 2003; Dabruzzi et al., 2013).

Shallow marine environments (e.g., intertidal zones) are dynamic in nature and the selection of appropriate habitats within these environments will depend on a variety of physical factors (such as temperature, tidal flow, salinity) and biological processes (such as predation risk and prey availability) (Sims, 2003; Carlisle and Starr, 2010; Speed et al., 2010). The dynamism of these factors is often manifested over a variety of temporal scales and animal movement patterns may change over large-scale seasonal cycles, for example annual migrations (Bauer and Klaassen, 2013; Chapman et al., 2015), to smaller-scale diel and tidal cycles, for example a change in habitat use during nocturnal/crepuscular periods (Hammerschlag et al., 2017) or during high tides (Lea et al., 2020).

Tides specifically, and the associated changes in water levels, will potentially have one of the strongest impacts on animal movement in shallow environments as it influences all other physical aspects of the water column such as temperature and oxygen concentration (as seen in Carlisle and Starr, 2010; Brinton and Curran, 2017; Smith and Curran, 2017). Tides also influence biotic factors including predation risk, as predators may become physically unable to access very shallow waters during low tide periods, and by changes in prey availability within inter-tidal flats. Indeed, tides are a significant driver of the movement of a number of marine taxa including seabirds, where individuals alter habitat use or foraging activities with the tide (Rogers et al., 2006; Rey et al., 2010), teleosts which have been found to move with the ebb and flow of the tide (Næsje et al., 2012), and elasmobranchs who move with the tide to forage over larger areas or to find refuge areas at high tide (Carlisle and Starr, 2010; Filmalter et al., 2013; Davy et al., 2015; Lea et al., 2020).

When considering the biology of elasmobranchs, their ectothermic nature means that temperature, in particular, can have a significant effect on physiological and biological aspects such as growth and metabolic rates, embryonic development, parturition, and the rate of digestion (Fangue and Bennett, 2003; Hight and Lowe, 2007; DiGirolamo et al., 2012; Dale et al., 2013). Consequently, temperature influences elasmobranch movement and certain species exhibit behavioral thermoregulation (whereby individuals will select habitats in a thermally heterogeneous environment to maintain body temperatures within an optimal range) in both field and laboratory environments. Thermoregulatory behaviors often manifest over diel and seasonal scales, given that temperature fluctuates across these temporal scales (Matern et al., 2000; Hopkins and Cech, 2003; Sims et al., 2006; Vaudo and Lowe, 2006; DiGirolamo et al., 2012; Speed et al., 2012).

As there are multiple drivers behind habitat selection, the relative importance of these will be context-dependent and trade-offs are likely to be necessary (Schlaff et al., 2014). For example, batoids have been shown to select shallow nearshore habitats at low tide, which may serve as a refuge from predation or competitive effects, but these habitats are likely metabolically expensive due to high temperatures and lower levels of dissolved oxygen (Vaudo and Heithaus, 2013; Dabruzzi and Bennett, 2014). Understanding which aspects of the environment drive habitat choice is necessary to elucidate the habitat attributes that maintain biodiversity, which is useful for prioritizing management or conservation initiatives for vulnerable faunal groups such as stingrays, where approximately 56% of species considered as threatened (Lea et al., 2016; Dulvy et al., 2021).

Consequently, this study aimed to investigate the influence of selected environmental factors on the daily presence to, and habitat use within, a shallow remote atoll by 3 sympatric stingray species. Firstly, it was investigated if presence to this offshore atoll by stingrays was influenced by seasonal variables such as temperature, precipitation, and wind speeds. Secondly, habitat use within the atoll was investigated to determine if presence in the deeper lagoon or shallow reef flat habitats were influenced by heterogenous environmental variables such as water temperature, tidal height, precipitation and diel stage. The extent to which these environmental variables influence habitat selection may inform why stingrays occur where they do.

These methods were approved by the South African Institute for Aquatic Biodiversity Animal Ethics Committee (reference number 2014/12) and by the Ministry of Environment, Energy, and Climate Change, Seychelles. All analyses were conducted in R (version 3.6.3) (R Core Team, 2020).

Stingrays were monitored for a mean of 345 ± 307 (SD) days from April 2014–October 2017. Most individuals (87%) were considered juveniles at the time of tagging. Residency to the atoll was found to be variable between individuals, but overall quite high, and periods of no detections (which could be related to individuals being present in the atoll but not detected, or having left the atoll) were mostly on the scale of days, but could range up to a few months of no detections (Elston et al., 2019).

This study revealed the nuanced ways in which the environment can effect stingray movement and habitat use patterns over a variety of temporal scales. Specifically, the interplay between tide and temperature, and how these changed across seasonal and diel scales, significantly effected the way that stingrays used the dominant habitats associated with the St. Joseph Atoll. Previous work found that juvenile stingrays in this atoll preferentially utilized the shallow sand flat habitat, but did at times occur in the deeper lagoon (Elston et al., 2019, 2020). The present study suggests that movements between these two habitat types were influenced by temperature and tidal height.

The daily presence of stingrays to the atoll was, in part, correlated with changes in temperature and wind speed. Stingrays were more likely to be detected in the atoll when daily temperatures were higher and when wind speeds were lower, conditions which correspond to the NW-monsoon season. Furthermore, individuals were found to increase the spatial extent of their movements in the atoll in the warmer NW-monsoon season, similarly to what was previously found for U. asperrimus juveniles (Elston et al., 2017). This trend was partially driven by the increased utilization of the lagoon and fringe reef habitats in the NW-monsoon season; habitats that are deeper and cooler compared to the shallow sand flats. This suggests that one of the reasons behind this movement pattern was temperature and that stingrays were potentially behaviorally thermoregulating, seeking out cooler habitats during the warmest time of the year. A similar finding was reported for juvenile stingrays that roamed less widely in a coastal bay in Australia during cooler months (Cerutti-Pereyra et al., 2014) and many other studies have identified batoids moving to actively exploit optimal thermal conditions (Matern et al., 2000; Fangue and Bennett, 2003; Vaudo and Lowe, 2006; Dabruzzi et al., 2013).

Temperature was also found to influence habitat use over fine temporal scales, a response that was found to differ between seasons. Specifically, stingrays appeared to avoid temperature extremes, seeking out the thermal stability of the lagoon when temperatures on the flats became too high or low. The tropical location of the atoll means temperatures do not fluctuate between seasons as much as they do in temperate regions. However, the physical properties of the different habitats means that temperatures on the shallow flats fluctuate quite significantly (approximately 15^oC). For both P. ater and U. asperrimus, when temperatures on the flats were lower during the cooler SE-monsoon season, individuals occurred more frequently at these times in the contextually warmer lagoon. Similarly, when temperatures on the flats were higher during the warmer NW-monsoon season, juveniles occurred more frequently at these times in the contextually cooler lagoon. Furthermore, warm temperatures appeared to influence habitat use more than cold temperatures, as models predicted temperature to have a larger influence on habitat use in the warmer NW-monsoon season compared to the cooler SE-monsoon season. These results suggest that the use of the lagoon habitat was, at least in part, influenced by temperature. This apparent behavior to remain within a thermal niche and to avoid temperature extremes may ensure that physiological processes are continually optimized. High temperatures may increase metabolic rates to sub-optimal levels, while low temperatures may slow down growth rates too much (Dale et al., 2013). The preference to remain in a narrow thermal niche has been witnessed in a laboratory setting where tropical juvenile ribbontail stingrays Taeniura lymna were found to have small thermal niches and individuals actively moved in tanks to remain in this narrow thermal niche (Dabruzzi et al., 2013).

Conversely, when considering U. granulatus, temperature was not found to significantly influence habitat use, raising questions as to whether this species may be more thermally tolerant in comparison to the other two. Laboratory studies show that different species can have different thermal tolerances (Fangue and Bennett, 2003; Dabruzzi et al., 2013) however, species-specific responses to temperature changes in the wild remain an enigma as the majority of studies that investigate the influence of temperature on batoid movement only consider one species (Cartamil et al., 2003; Vaudo and Lowe, 2006; Jirik and Lowe, 2012). It is important to note that these results must be treated with caution as there was only limited data for one season (SE-monsoon) for this species, and that more multi-specific research is needed to determine how different species respond to the same thermal environment.

Stingray habitat use was not only influenced by temperature, but also by the tides. Detections on the flats occurred throughout the tidal cycle (although were somewhat more frequent at high tides), while detections in the lagoon occurred at a significantly higher proportion at low tides. Given that stingrays are dorso-ventrally flattened, they can occur in very shallow waters and indeed, the majority of detections for these stingrays occurred on the shallow flats (Elston et al., 2021). The high use of the flats by stingrays was likely a function of predator avoidance. The lagoon presents a significant predation risk as sharks display fidelity to this habitat (Filmalter et al., 2013; Lea, 2017) and juvenile stingrays were occasionally observed with shark bites on their pectoral fins (personal observations). Conversely, while sharks have also been shown to increase their use of the flats at high tide in the atoll (Lea, 2017; Lea et al., 2020), this habitat likely still provides refuge opportunities for dorsoventrally-flattened stingrays that can make use of shallow island fringes or sand banks that are more inaccessible to sharks. Further support for this predator avoidance hypothesis comes from the evidence that juveniles preferentially occurred on the flats, whilst adults (which are at lesser risk to predation) preferentially occurred in the lagoon (Elston et al., 2021). This behavior has been observed in an inter-tidal bay, where U. granulatus juveniles moved in a directed way during flooding and ebbing tides, remaining in shallow water, which was hypothesized to be a predator avoidance strategy (Davy et al., 2015). However, despite the reduction of predation risk on the flats, some regions of this habitat may become exposed at low tide, thus presenting a high risk of stranding and restricting movement in this habitat. This likely necessitates that stingrays seek out deeper waters at low tide, and thus they were found to occur in the lagoon more frequently during low tides. This is notwithstanding that the increased use of the lagoon at low tides may also relate to the avoidance of temperature extremes. The shallow waters at low tides on the flats become very warm (over 40^oC), which may be approaching the upper thermal maximum for these stingrays (Dabruzzi et al., 2013), driving them to seek out the cooler waters of the lagoon at low tides. Finally, stingrays in St. Joseph Atoll are known to feed on the flats (Elston et al., 2020) and the preferential use of this habitat may also be linked to foraging opportunities.

In this study, most individuals had diel peaks in detections and the diel stage significantly influenced habitat use. Previously identified elasmobranch diel movement patterns have been attributed to foraging or predator avoidance (Cartamil et al., 2003; Carlisle and Starr, 2010; Hammerschlag et al., 2017). If these were indeed the drivers behind diel movement patterns, one would expect a consistent pattern (Holland et al., 1993; Cartamil et al., 2003). However, in this study, there was high individual variability in the periodicity of diel peaks. This results reiterates that this atoll is a dynamic environment, where individuals likely have to balance a variety of environmental factors when using different habitat types.

The caveats of acoustic telemetry need to be addressed, specifically the potential effect that environmental factors may play on the detection efficiency of receivers and its subsequent role in the observed daily presence patterns. In the present study, the probability of daily detection was significantly negatively correlated with wind speeds and this trend could be influenced by receiver detection efficiency. Wind speed is known to reduce detection efficiency of receivers, both in this and other receiver arrays (Heupel et al., 2006; Elston et al., 2019). The SE-monsoon season is characterized by persistent strong winds and detections in this season were restricted when compared to the detections in the NW-monsoon season. However, the increase in detections recorded in the NW-monsoon season largely occurred in the deeper lagoon habitat, where wind is unlikely to have as large an effect on detection efficiency as on the shallow flats. This discrepancy suggests that these seasonal patterns in detections may not only be due to the changes in receiver detection efficiency but are likely driven by actual changes in movement patterns. A further caveat is that the different transmitters types used in the study will have different rates of detectability due to factors such as power output and ping rate. The larger transmitters were implanted into the larger individuals, and indeed, larger individuals were found to have significantly higher edge and node densities in the networks constructed. This trend may, at least in part, be influenced by the different transmitters used.

Elasmobranchs have been hypothesized to be highly susceptible to the effects of climate change, and this is specifically true for U. asperrimus (Chin et al., 2010). The significant influence of temperature on habitat use found in the present study supports this notion, and may suggest that these species will be significantly effected by climate change induced temperatures. Specifically, with increasing temperatures, individuals may need to increase their utilization of the lagoon habitat to keep metabolic rates at optimal levels. This is important as increased metabolic rates have been linked to losing body mass in elasmobranchs, potentially reaching lethal levels under future climate change scenarios (Lear et al., 2020). Thus, species using shallow water environments, such as atolls, with may be at greater risk to the effects of warming water temperatures and may need to significantly alter habitat use patterns to survive. Future work could focus on exactly how temperature may influence stingray physiology, for example determining metabolic rates under various temperature regimes, to fill in the gaps of understanding how changing temperatures may influence the survival potential of these animals. The interplay between how tides and temperature effected habitat use in stingrays was dynamic, but relatively consistent between stingray species, highlighting how important the physical environment is on determining habitat use in stingrays.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The animal study was reviewed and approved by South African Institute for Aquatic Biodiversity Ethics Committee.

CE, PC, and RB conceived and developed the project. CE performed data analysis and wrote the first draft of the manuscript. PC, RB, and JL edited the manuscript. PC and RB were supervisors of the project. All authors contributed toward data collection.

This research was supported by a grant obtained from the Save Our Seas Foundation Keystone (Grant number 288). CE received a scholarship from the South African National Research Foundation.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.