Juvenile waiting stage crown‐of‐thorns sea stars are resilient in heatwave conditions that bleach and kill corals

The juveniles of predatory sea stars can remain in their recruitment–nursery habitat for some time before their ontogenetic shift to the adult habitat and diet. These small juveniles are vulnerable to a range of factors with their sensitivity amplified by climate change‐driven ocean warming. We investigate the thermal tolerance of the waiting stage herbivorous juveniles of the keystone coral predator, the crown‐of‐thorns sea star (COTS, Acanthaster sp.), in context with the degree heating weeks (DHW) model that predicts coral bleaching and mass mortality. In temperature treatments ranging from +1 to 3°C in prolonged heatwave acclimation conditions, the juveniles exhibited ~100% survival in DHW scenarios that trigger coral bleaching (4 DHW), resulting in mass mortality of corals (8 DHW) and extreme conditions well beyond those that kill corals (12 DHW). This indicates that herbivorous juvenile COTS are far more resistant to heatwave conditions than the coral prey of the adults. The juveniles exhibited higher activity (righting) and metabolic rate after weeks in increased temperature. In separate acute temperature experiments, the upper thermal limit of the juveniles was 34–36°C. In a warming world, juvenile COTS residing in their coral rubble nursery habitat will benefit from an increase in the extent of this habitat due to coral mortality. The juveniles have potential for long‐term persistence as herbivores as they wait for live coral to recover before becoming coral predators, thereby serving as a proximate source of COTS outbreaks on reefs already in a tenuous state due to climate change.

frequency causing collapse of coral reef structure and associated communities (Baum et al., 2023;Hughes et al., 2017Hughes et al., , 2018;;Stuart-Smith et al., 2018).This ecological disruption can lead to a shift from coral-dominated to algal-dominated coral rubble ecosystems (Hughes, 1994;Kenyon et al., 2023).Impacts on associated species have been documented, especially for species that depend on coral for food and/or shelter with flow-on effects for the ecosystem (Jones et al., 2004;Przeslawski et al., 2008;Rice et al., 2019;Stella et al., 2022).Following bleaching mortality, coral skeletons are colonised by algae and other organisms and erosive agents (e.g.cyclones and bioeroders) accelerate breakdown of the skeleton into rubble.While thermal stress and changing habitat are deleterious for live coral-associated species, the breakdown of dead coral into rubble gives rise to a diversity of crypto-benthic invertebrates and fishes that rapidly avail of this newly created habitat (Kenyon et al., 2023;Stella et al., 2022;Timmers et al., 2021;Wolfe et al., 2020Wolfe et al., , 2021)).
To predict future community composition of coral reef habitat, it is important to determine how species are directly (thermal stress) or indirectly (loss/gain of habitat) impacted by heatwave anomalies.As a resident of coral rubble habitat, the herbivorous juvenile stage of the coral predator, the crown-of-thorns sea star (COTS, Acanthaster) may benefit from increased habitat following coral mortality.This sea star is one of the most influential invertebrates in Indo-Pacific coral reefs causing widespread coral mortality when their populations outbreak (Deaker & Byrne, 2022a;Kayal et al., 2012;Pratchett et al., 2021;Roche et al., 2015).Due to their ecological influence, it is important to understand the impacts of habitat warming on COTS, especially in context with the thermal sensitivity of their coral prey.
A study that tested temperatures across present-day and heatwave conditions showed that COTS development has a broad thermal optimum with high survival and normal development across the 25-32°C range, indicating resilience to heatwave conditions (Lamare et al., 2014;Uthicke et al., 2015).This thermal resilience may increase in progeny of warm-adapted parents (Hue et al., 2022).Competent larvae reared in control temperatures of the northern Great Barrier Reef (GBR, 28°C) followed by exposure to warm treatments (30-34°C) were able to settle at high rates across all temperatures, but postsettlement survival decreased markedly at ≥34°C (Lang et al., 2023).For juvenile COTS, increased temperature (30°C) had a positive effect on growth given ad libitum access to food (Kamya et al., 2018).Adult COTS exhibit increased movement, feeding and respiration with increased temperature and had decreased survival at 32°C (Lang et al., 2021(Lang et al., , 2022)).
While mortality of corals in response to warming indicates that the future of COTS is bleak due to the demise of its prey, this sea star has flexibility in its diet at the initial herbivorous juvenile stage availing of a range of food sources, including biofilms (Deaker, Agüera, et al., 2020;Deaker, Mos, et al., 2020).In addition, adult COTS prey on soft corals, bivalves and algae when hard coral availability is low (Chesher, 1969;Kuo et al., 2022;Yamaguchi, 1975).As seen for several ecologically important predatory sea stars (Byrne et al., 2021;Nauen, 1978), juvenile COTS can remain in their recruitment-nursery habitat for a variable amount of time where they exist on an algal diet before they transition to the corallivorous adult stage (Wilmes, Schultz, et al., 2020).They start their benthic life as herbivores with a preference for coralline algae and have the potential to persist for years (at least 6 years) in the herbivorous phase in the absence of coral prey (Deaker, Agüera, et al., 2020;Deaker, Mos, et al., 2020;Wilmes, Hoey, et al., 2020;Wilmes, Schultz, et al., 2020).Thus, in a situation of low coral prey following mass bleaching or a COTS outbreak, the juveniles have the potential to persist and build up in number over multiple years before transitioning to corallivory and the mature adult stage (Deaker, Agüera, et al., 2020;Deaker, Mos, et al., 2020;Wilmes, Hoey, et al., 2020;Wilmes, Schultz, et al., 2020).
These findings prompted the waiting stage hypothesis, positing that accumulation of juvenile COTS in the reef infrastructure over many years may be a proximate driver of outbreaks (Deaker, Agüera, et al., 2020).
We investigated the thermal tolerance of the early juvenile 'waiting stage' COTS in conditions that cause bleaching and mortality in corals.The degree heating weeks (DHW) model is used to predict the onset and severity of coral bleaching, incorporating the magnitude of temperature increase and heatwave duration (Skirving et al., 2020).Accumulated heat stress as measured by the DHW model is calculated based on the scenario of 1°C above the maximum historic mean temperature with 1 week of +1°C being 1 DHW and a +2°C rise for 1 week being 2 DHWs and so on (i.e.pro rata) with respect to degree of warming and how long warm temperatures persist (Skirving et al., 2020).The DHW model is well correlated with bleaching onset, severity and coral mortality (Hughes et al., 2017;Skirving et al., 2020) and forms the basis of the bleaching alert system (www.coralreefw atch.noaa.gov).We used the DHW framework in experiments with juvenile COTS exposing them to heatwave scenarios.
As coral rubble provides both habitat and algal food for juvenile COTS, they have the potential to persist and benefit from the increased availability of this habitat on bleached reefs.Among many factors such as predation (Desbiens et al., 2023) and reduced cover of crustose coralline algae (CCA) due to bleaching (Lang et al., 2023), juvenile survival would depend on their resistance to heatwave anomalies, which are projected to increase in incidence and duration.Many studies of adult and juvenile benthic marine invertebrates show tolerance to warm conditions for short pulses of heat stress (days/weeks), but deleterious effects and mortality occur the longer (weeks/months) warm conditions remain (Balogh & Byrne, 2021;Byrne, 2011;Hughes et al., 2017Hughes et al., , 2018;;Wolfe et al., 2013).We tested the response of herbivorous juvenile stage of COTS to habitat warming in two experiments.The first involved gradual warming to a set temperature where they were held for weeks (4-12 DHW) to emulate heatwave conditions in the framework of the DHW model.This allowed us to consider the thermal tolerance of juveniles in context with that determined for corals (Hughes et al., 2017(Hughes et al., , 2018;;Skirving et al., 2020).In the second experiment, we assessed the response to acute temperature exposure as an indication of maximum thermal tolerance and to determine the upper lethal temperature (ULT).The ecological success of COTS is associated with its extreme morphological and physiological plasticity (Byrne, 2022;Deaker & Byrne, 2022a), and so we expected that the juveniles would tolerate heatwave conditions.Our aim was to determine the magnitude of this tolerance and place our findings in context with the DHW model of thermal stress determined for their coral prey.

| Juvenile rearing
As the taxonomy of the Acanthaster species in the western Pacific is uncertain (Haszprunar & Spies, 2014), we use the genus name or COTS when referring to the species investigated here.Adult Acanthaster sp. were collected from the northern GBR near Cairns by the COTS control team in November 2019.They were shipped to Sydney and maintained in flow-through aquaria supplied with filtered sea water (FSW 5.0 μm) at temperatures around the time of collection 26-27°C (http://data.aims.gov.au/aimsrtds/yearl ytren ds.xhtml) at the Sydney Institute of Marine Science.The juveniles were generated from a spontaneous mass spawn of COTS (four males and four females) in the aquarium system in January 2020.The larvae were reared in aerated 1 L beakers of 1 μm FSW at 26°C and fed 20,000 cells mL −1 of Proteomonas sulcata every 2 days until the competent brachiolaria stage.Competent larvae (16-day postfertilisation) were transferred into Petri dishes (10 mm diameter) with 1.0 μm FSW (26-27°C) and a settlement inducer, a small piece of the coralline alga, Amphiroa sp.This alga is also a good food source for the juveniles (Deaker & Byrne, 2022b).The juveniles were fed fresh Amphiroa sp.every 3 days in parallel with water changes for 4 months and then used for thermal experiments.The temperature at which the adults were maintained for 2 months prior to spawning was used as the control temperature for the experiments because the temperature at which COTS parents are held can influence the thermal response of their offspring (Hue et al., 2022).

| Marine heatwave: Gradual warming
The control temperature (27°C) is the maximum mean summer sea surface temperature experienced by COTS populations in the southern GBR, and we used the coral bleaching data for this region (see Brown et al., 2023) to base our experimental DHW scenarios.In the southern GBR, the coral bleaching threshold is 28.3°C (Brown et al., 2023).Three heatwave scenarios, +1°C (28°C), +2°C (29°C) and +3°C (30°C), above the maximum mean sea surface temperature were used.
Temperature treatments for the heatwave experiment were created using two aluminium blocks set up in parallel that allowed for a stable, static thermal gradient (27.3-32°C) established by using warm-and cold-water inputs at either end.Each heat block had four columns of holes (20 mm diameter) to fit the vials (40 mL), with each row representing a temperature treatment and set up to provide a gradual (~0.3°C increments between columns) temperature increase.
At the start of the experiment, the juveniles were 4-month postsettlement (mean diameter ± SE = 2.13 ± 0.45 mm, range = 1.39-2.90mm).They were randomly placed in individual vials and fed with a frond of Amphiroa sp.approximately 1.5 cm long.The vials, FSW and algae were replaced every 2 days to ensure that the juveniles were fed ad libitum and that the Amphiroa sp. was fresh and to prevent the development of a biofilm.During the months-long experiment, we noted that the juveniles continued to increase in size but did not measure them to avoid handling stress.
Juveniles (n = 14) that remained at control conditions were maintained in a temperature-controlled water bath at a stable temperature of 27.1°C.The vials with juveniles that were subjected to a heatwave scenario were placed in the cool end of the gradient starting at 27.3°C.In a staggered start, groups of juveniles (n = 8 per scenario) were moved by stepping them one row (~0.2-0.4°C)approximately every 3 days until they reached the final target temperature of 28, 29 or 30°C.The juveniles in the 28°C heatwave scenario remained at temperature for 29 days (mean = 28.14 ± 0.06°C, n = 27) representing 4.1 DHW, the threshold for coral bleaching (Liu et al., 2006), and suggested to be the threshold for onset of severe bleaching (Hughes et al., 2018).Juveniles stepped up to the 29°C heatwave scenario were maintained at 28-28.5°C for 1 week and at 29°C for 23 days (mean ± SE = 28.96± 0.03°C, n = 22), representing 7.6 DHW.Those being stepped up to the 30°C heatwave scenario were maintained at 28.5-29.5°Cfor 2 weeks and then moved to 30°C.As we did not observe juvenile mortality at 30°C, we maintained that group at 30°C for 20 days (mean ± SE = 29.88 ± 0.05°C, n = 18), representing 12 DHW, an extreme warming scenario well beyond levels that causes mass mortality of corals (Hughes et al., 2017).The DHW levels were calculated with respect to the DHW metric used to predict coral bleaching (Kayanne, 2017;Skirving et al., 2020).The temperature in the heat blocks and the water bath was monitored daily (Vernier Temperature Probe; Vernier Software and Technology).Dissolved oxygen (DO) was measured before water change with a Pyro Firestring oxygen probe (PyroScience) and was >90% across all treatments.
To assess muscular coordination activity, the righting response, a behaviour used to assess the health of echinoderms (Lawrence & Cowell, 1996;Minuti et al., 2021;Peck et al., 2008), was determined for the juveniles after they had been at the target temperatures for 12 days.The juveniles were placed in a small Petri dish (22 mm diameter × 20 mm height) with 3 mL of FSW at experimental temperature.To maintain the temperature in the dishes, the room temperature was also set to match the juvenile treatment temperature.Juveniles were flipped using an eyelash brush under a dissecting microscope (Olympus SZH10).The righting response was timed from when they were placed on their aboral side to when they had completely returned to a flat profile on their oral side.To account for individual variability (Lawrence & Cowell, 1996), the experiments were performed twice for each juvenile with the second performed 2 h later and the righting time was averaged per individual.The juveniles were photographed using an Olympus DP73 digital camera mounted on an Olympus BX60, and their diameter was measured from the arm tip to the opposite arm tip using ImageJ (ver.1.52a; NIH).
The mass-specific metabolic rate of the juveniles was determined using constant volume respirometry after juveniles had been at temperature for 3 weeks with the control juveniles having been at 27°C for their entire development.A single juvenile was placed into a 2 mL vial (PreSens SV-PSt5) fitted with an oxygen sensor spot (PreSens, 5 mm sensor spots; AS1 Ltd.).Each vial was filled with FSW and sealed ensuring that no air bubbles were trapped within the vial.
Vials were placed in the wells of the Sensor Dish Reader (PreSens SDR Oxodish; AS1 Ltd.) in the dark within a temperature incubator that maintained a constant temperature of 27, 28, 29 and 30°C for each treatment.DO within each vial was measured at 15-s intervals until the oxygen concentration decreased by ~20% or once there was a measurable decrease over 4 h.Controls (n = 4) without a juvenile were measured to correct for background changes in oxygen levels.The juveniles were photographed after the respiration and measured (Image J).Across the four temperatures, 27, 28, 29 and 30°C, the diameters of the juveniles used were 3.

| Acute thermal tolerance: Upper thermal limits
As the juveniles exhibited resistance to warming in the DHW experiment, a separate set of juveniles (~5 months old, 4.5-7.5 mm diameter) were exposed to acute warming to determine the ULT.
The juveniles were placed in 40 mL vials (as above) and transferred directly from the control temperature of 27°C into seven experimental temperatures established using the heat block (see above); 28°C (+1°C, n = 7), 29°C (+2°C, n = 8), 30°C (+3°C, n = 7), 31°C (+4°C, n = 7), 32°C (+5°C, n = 7), 34°C (+7°C, n = 6) and 36°C (+9°C, n = 4).A group of juveniles (n = 7) was maintained in parallel at the control temperature (27.1°C).To monitor survival, the juveniles were checked daily for a maximum experimental duration of 30 days (depending on survival).The juveniles at the highest temperature were also examined after 12 h.The juveniles were fed Amphiroa sp., and the alga, FSW and vials were replaced every 2 days (as above, depending on survival).DO was measured before water change (as above) and remained >90% in the 27-32°C treatments, but dropped to 87% at 34°C.We did not measure the DO in the 36°C treatment because most of the juveniles were dead within 12 h.

| Statistical analysis
All data were analysed using R (ver 4.0.3,R Core Team, 2020), and figures were made using ggplot2 (Wickham et al., 2018).The metabolic rate of juveniles in the different temperature treatments (27, 28, 29 and 30°C) was analysed using a one-way analysis of variance (ANOVA) (lm function, stats package).As the righting response of juvenile COTS is influenced by size (Deaker et al., 2021), the righting time was divided by juvenile diameter and the corrected data were analysed using a one-way ANOVA.Homogeneity of variance was confirmed for the righting, and respiration data using Levene's test and normality were assessed by visually inspecting plots of the residuals.All data were heteroscedastic, and the data for the righting response were square root transformed to be normally distributed.Post hoc analysis was computed using Tukey-adjusted pairwise comparisons to assess significant differences in the righting and respiration data between temperatures (emmeans package, Lenth, 2020).
For the acute temperature tolerance data, Kaplan-Meier survival probability curves were constructed with 95% confidence intervals (Pocock et al., 2002) using the survival package (Therneau, 2020) and survminer package in R (Kassambara et al., 2021).Post hoc pairwise comparisons were made using Benjamini-Hochberg adjusted p-values to determine the temperature levels that had a significant effect on survival.

| Heatwave experiment
Apart from one juvenile in the 28°C treatment that died on Day 3, there was no mortality (100% survival across DHW scenarios) in the heatwave experiment.The single mortality at 28°C was unlikely to have been caused by the 1°C increase in temperature.
There was a significant difference in the righting response (ANOVA: F 3,23 = 3.144, p = .045).Post hoc comparison indicated that the juveniles from the 30°C heatwave treatment righted themselves significantly faster than those from the 28°C treatment (Figure 1A).

Δ O 2 × V W
There was a significant effect of temperature on the metabolic rate (ANOVA: F 3,32 = 4.78, p = .007).Post hoc comparison indicated that the juveniles in the 30°C treatment had a higher metabolic rate compared with the juveniles at 27 and 28°C (Figure 1B).Across the four temperatures, the mean (±SE) mass-specific metabolic rates of the juveniles were 0.058 ± 0.006, 0.043 ± 0.004, 0.064 ± 0.006 and 0.084 ± 0.012 mg O 2 g −1 h −1 , respectively (Figure 1B).

| Acute thermal tolerance experiment-Upper thermal limits
Juveniles were tolerant of elevated temperatures below 31°C (+4°C) with no mortality in the 27-30°C treatments over 30 days (Figure 2).
There was a significant effect of temperature on survival (χ 2 = 58.> 36°C, Figure 2) with 100% mortality within 12-24 h for juveniles at 36°C (+9°C), after 4-11 days at 34°C (+7°C) and 20 days at 32°C (+5°C).The ULT for rapid death was 36°C with 50% death at 34°C by 5 days.The ULTs of the juveniles approximate 34-36°C.There was high confidence of probability for survival in the extreme temperatures and variable survival in the sublethal temperature (31°C; Figure 2).

| DISCUSS ION
As ocean temperatures increase with extreme heatwave anomalies becoming more frequent, marine communities are shifting to warm, stress-tolerant species (Poloczanska et al., 2016).For coral reefs hit hard by recurrent bleaching events, community composition is being influenced by the types of coral that recover quickly following bleaching events such as Acropora, the species favoured by COTS (Keesing et al., 2019) and those that do not recover quickly (e.g.Porites; Hughes et al., 2018) as well as by the community dynamics of the vast rubble habitat created by dead coral.This habitat supports a diversity of species, including juvenile COTS (Stella et al., 2022;Wilmes, Hoey, et al., 2020;Wilmes, Schultz, et al., 2020;Wolfe et al., 2021) and their predators (Desbiens et al., 2023).Given the importance of the rubble recruitment-nursery habitat for juvenile COTS, it is essential to understand the thermal tolerance of this life stage, which may accumulate in the rubble as they wait for live coral to increase.
We exposed juvenile COTS to thermal stress scenarios at a tempo and temperature duration designed to reflect the DHW conditions that cause coral bleaching and mortality.The juveniles exhibited remarkable stress tolerance with ~100% survival in conditions ranging from 4 DHW, the threshold for coral bleaching, to 8 DHW, which causes mass bleaching mortality and even to 12 DHW when the experiment was stopped.Thus, the juveniles exhibited tolerance to heatwave conditions well above levels that kill corals (Hughes et al., 2017).The juveniles were fed ad libitum, and their continued growth over the experiment indicated that they were in good condition through the heatwave durations.Ready access to food may have enhanced their resilience and thereby their temperature tolerance in our experiment.Juvenile herbivorous COTS have a flexible diet being able to survive on a range of algal food sources (Deaker, Mos, et al., 2020).With the high cover of algae that characterises the coral infrastructure and rubble habitat (Kenyon et al., 2023;Klumpp & McKinnon, 1989, 1992), they are unlikely to be food limited in nature.This may help make the juveniles robust to stress, including thermal perturbations.That said, their preferred algal food, CCA, is sensitive to increased temperature (Diaz-Pulido et al., 2012;Short et al., 2015;Vasquez-Elizondo & Enriquez, 2016) with some bleaching observed at 32°C and significant bleaching at 34°C (Lang et al., 2023).Growth of CCA can be negatively impacted by heatwaves, but some species show resilience (Cornwall et al., 2019).
The faster righting time and higher metabolic rate of the juveniles in the warmest treatment (30°C) are similar to that seen in other studies where a gradual increase in temperature was undertaken in an acclimation approach with echinoderms (Christensen et al., 2023;Harianto et al., 2018).The metabolic rate determined here for the juveniles maintained in the 27°C controls is similar to that obtained previously (Deaker & Byrne, 2022b).In adult COTS, the righting time of acclimated individuals was not affected by increased temperature (to 32°C), but metabolic rate was higher at 30°C and decreased at 32°C (Lang et al., 2022).We might have seen a decrease in metabolic rate in the juveniles if we used a 32°C heatwave.
We had 100% survival in the DHW experiment.This contrasts with a recent acclimation study of adult COTS where mortality was observed at 32°C (Lang et al., 2022).In that study, there was 50% mortality of adults after 40 days at 32°C and 75% mortality after 60 days.This response to month(s)-long exposure suggests that 32°C is above the thermal tolerance of COTS adults (Lang et al., 2022).While we did not test 32°C as a maximum heatwave temperature, the juveniles tolerated this temperature for 20 days in the acute trials.
As exposure to the most extreme heatwave scenario (12 DHW) indicated that the juveniles had high temperature tolerance, the acute thermal survival study (+1-9°C) was conducted to determine their ULT.The 9°C increase (36°C) caused rapid mortality and the 7°C increase (34°C) killed 50% of the juveniles by 5 days.This indicates an ULT for the juveniles in the range of 34-36°C, similar to that determined for adult COTS in an acute study (33-34°C, n = 3 animals; Yamaguchi, 1974).In an acute study, small subadult COTS tolerated 36°C (Lang et al., 2021).
Our results with herbivorous juvenile COTS indicate high thermal tolerance and resistance to the ongoing trajectory of ocean warming as well as resistance to the heatwave anomalies that kill the coral prey of adult COTS.In nature, however, the juveniles face a plethora of threats in their rubble habitat.Predation and sublethal injury can be high (Deaker et al., 2021;Keesing et al., 2018;Keesing & Halford, 1992a, 1992b;Wilmes et al., 2019), they may succumb to disease (Zann et al., 1987), and their CCA food may decline in warm conditions (Lang et al., 2023;Short et al., 2015).These threats may be exacerbated by heatwaves in conjunction with other climate change stressors as seen for the early juvenile stages of other species (Balogh & Byrne, 2021;Wolfe et al., 2013).A recent study shows that COTS-eating predatory crabs occur in coral rubble and in the laboratory consumed five juveniles a day (Desbiens et al., 2023).Understanding community dynamics in the rubble habitat in changing climate is a priority (Wolfe et al., 2021).It would be of interest to determine the thermal tolerance of the juvenile-eating crabs and their future performance as COTS predators.
The thermal sensitivity of echinoderms varies across life stages (Byrne et al., 2018;Gall et al., 2021;Peck et al., 2008) as appears in the case for COTS.Adult COTS may be living close to their thermal limits (Lang et al., 2022(Lang et al., , 2023) ) as is the case for many coral reef species (Stuart-Smith et al., 2017).In contrast, the herbivorous juveniles tolerated long-duration heatwave conditions.Our results are counter to expectations that the early benthic juvenile stages of marine invertebrates are more sensitive than the adults and so are a population mortality bottleneck (Gosselin & Qian, 1997;Hunt & Scheibling, 1997;Przeslawski et al., 2015).It appears that juvenile COTS that survive the early postlarval period may persist for some time in their coral rubble recruitment-nursery habitat, as seen for the juveniles of other predatory sea stars (Byrne et al., 2021;Nauen, 1978).An understanding of the thermal biology across individual life stages of COTS with respect to ocean warming is emerging, with most studies being on the larval stage (Hue et al., 2022;Kamya et al., 2014;Lamare et al., 2014;Lang et al., 2023;Uthicke et al., 2015).As found here, studies of juvenile COTS indicate that they are heat stress tolerant (Kamya et al., 2018).If the supply of juveniles resulting from spawning, fertilisation and larval development is reduced at any stage in propagule generation due to impaired thermal performance then the juveniles may not feature in the rubble community.This points to the need for a cross-generational understanding of the prospects for life cycle success in COTS in a warming climate (Byrne et al., 2020;Hue et al., 2022).
Population outbreaks of Acanthaster are likely to be driven by a suite of opportunistic-plastic traits that are characteristic of this sea star (Birkeland, 1989;Byrne, 2022;Deaker & Byrne, 2022a;Wolfe et al., 2015).As found for Asterias and Marthasterias species, the early benthic stage of predatory sea stars appears to be highly resilient and plastic in their metabolic needs and life transitions allowing them to remain as immature 'Peter Pan' juveniles for many years (Byrne et al., 2021;Deaker, Agüera, et al., 2020;Guillou et al., 2012;Nauen, 1978).Nauen (1978) suggests that the high adaptability of sea stars, with respect to environmental conditions, and the timing of the juvenile waiting stage, contributes to their ecological success.These cryptic juveniles are difficult to study but likely hold the secrets to the success of predatory sea stars.
To understand the potential of the juveniles in waiting hypothesis to contribute to the explanation of COTS outbreaks, future studies where cohorts of juveniles of a known age are tracked in situ are needed, as done in other studies (Byrne et al., 2021;Nauen, 1978).This is a challenge and would likely require serendipitous discovery of newly settled sea star cohort (e.g.Byrne et al., 2021).We also need a means to age juveniles found in nature to understand the extent of the 'Peter Pan' phenomenon in COTS.Finally, we do not know the threshold of live coral abundance required and the coral species available (see Neil et al., 2022) for juveniles to make the diet switch.
As the ocean continues to warm across the tropical Indo-Pacific, waiting stage juvenile COTS may benefit from the increase in the spatial extent of rubble habitat generated by coral bleaching and mortality and increase in number over time .These juveniles may be key to the outbreak phenomenon.Their presence in the reef infrastructure may explain the appearance of new coraleating juveniles (as small as 3 cm diameter) after a major outbreak of large adults has passed (Moran et al., 1985) and why reefs require repeated visits by culling teams to curtail population outbreaks (Westcott et al., 2020).In a warming world, juvenile COTS residing in

CO N FLI C T O F I NTE R E S T S TATE M E NT
The authors declare no conflicts of interest.
69 ± 0.26 mm (n = 14), 4.38 ± 0.29 mm (n = 8), 3.50 ± 0.21 mm (n = 6) and 4.36 ± 0.31 mm (n = 8), respectively.Immediately after each respiration measurement, the vials containing FSW and a juvenile were weighed on an electronic balance (Mettler Toledo PL403, three decimal places).Juveniles were then removed from the vial, patted dry and weighed.The weight of FSW in the respirometry vial was determined by subtracting the weight of the empty vial and juvenile.To calculate the volume of FSW in each vial, the weight of the FSW was multiplied by its density to account for the salinity and temperature of each treatment.The metabolic rate (MO 2 , mg O 2 h −1 g −1 ) was calculated using the equation: where W is juvenile wet weight, V is the volume of FSW in the vial and Δ O 2 is the linear regression (i.e.slope) of oxygen concentration over time (mg O 2 L −1 h −1 , regression method, see Harianto et al., 2018).
the rubble habitat are well placed to avail of new coral growth such as that seen in the recent recovery of fast-growing Acropora, the preferred food of COTS, in areas of the GBR devastated by bleaching(AIMS, 2022).Given such opportunity, these juveniles are likely to emerge as coral predators in reefs already vulnerable due to climate change (Figure3a-e).ACK N O WLE D G E M ENTSWe thank Monique Webb, Natasha Rae and Natasha Garner for assistance in rearing the juveniles and Dr Laura Parker for providing the respirometry equipment.Dr. Ana Christensen provided advice on respirometry.Staff of the Sydney Institute of Marine Science (SIMS), particularly Sergio Torres Gabarda and Andrew Niccum, are thanked for assistance.Research was supported by a PhD scholarship from the University of Sydney (DJD) and partially supported by the Lizard Island Reef Research Foundation, Ian Potter Grants for COTS research.We thank two reviewers for comments that helped improve the manuscript.This is SIMS contribution number 310.Open access publishing facilitated by The University of Sydney, as part of the Wiley -The University of Sydney agreement via the Council of Australian University Librarians.

F
Coral habitat prebleaching (a) and changes that occur to coral reefs during bleaching (b) and subsequent mortality of coral leading to a collapsed reef and algae-covered coral rubble (c, d).Juvenile herbivorous COTS reside in the rubble (a-d) and after bleaching mortality wait until the return of live coral to allow them to transition to the predatory stage (e).Symbols sourced from Integration and Applied Network, University of Maryland Center for Environmental Science (https://ian.umces.edu/media-libra ry/) and Byrne Laboratory images.COTS, crown-of-thorns sea star.