Heating rate explains species-specific coral bleaching severity during a simulated marine heatwave

: Marine heatwaves (MHWs) are becoming more frequent as a consequence of climate change. These discrete events are causing widespread stress and mortality in marine ecosystems, including coral reefs. The heat tolerance of different coral species is often complex and depends on a combination of environmental and biological factors, making accurate predictions of the impact of MHWs on individual species challenging. Heating rate has been shown to influence coral bleaching in Acropora species, but it remains unknown how heating rate influences bleaching in other corals with contrasting morphology and bleaching sensitivities. In this study, we explored the sensitivity of Pocillopora damicornis and Plesiastrea versipora , representing branching and encrusting growth forms, respectively, to heating rate. We experimentally simulated MHWs with slow (0.5°C d −1 ) and fast (1°C d −1 ) heating rates and measured physiological responses to quantify changes in coral health, including photochemical efficiency, holobiont metabolism, tissue biomass, chl a , and symbiont density. Our results confirm that heating rate is a good predictor of coral bleaching sensitivity for these species, with faster heating rates causing more severe bleaching and declines in coral health. However, bleaching sensitivity differed between P. damicornis and P. versipora , with P. damicornis more affected by the faster heating rate. The use of heating rate, in addition to other metrics such as duration and intensity of heat, will enhance our capacity to predict the local impact of MHW events and their overarching ecological consequences for coral ecosystems.


INTRODUCTION
Rising ocean temperatures are severely impacting marine ecosystems (Hoegh-Guldberg & Bruno 2010, Pecl et al. 2017, Pinsky et al. 2020), a trend expected to continue with increasing anthropogenic activity (Intergovernmental Panel on Climate Change [IPCC] 2021). In addition to gradual sustained temperature increases, marine ecosystems are further threatened by acute warming events, known as marine heatwaves (MHWs) (Smale et al. 2019). The frequency of MHWs has increased by over 50% during the last decade (Oliver et al. 2018, IPCC 2021, and more than 50% of the world's oceans are projected to be in a permanent MHW state (compared to today) by the end of the 21st century . The predicted increases in MHWs pose a huge challenge to marine organisms, with upper thermal thresholds expected to be exceeded more regularly. Therefore, understanding organismal re sponses to extreme temperature events is imperative to understanding the impacts of MHWs on marine biodiversity.
Coral reefs are increasingly experiencing mass bleaching events and one-third of all scleractinian corals are at elevated risk of extinction (Frieler et al. 2013, Hughes et al. 2018. Most coral reefs worldwide have already suffered from widespread coral bleaching (Sully et al. 2019), a phenomenon which results in the loss of photosynthetic symbiotic dinoflagellates (Symbiodiniaceae) (Glynn 1996, La -Jeunesse et al. 2018. As corals typically rely on their endosymbionts for the majority of their daily energy requirements (Muscatine et al. 1981, Dubinsky et al. 1990, the breakdown of this symbiotic relationship can lead to impairment of metabolic processes and, in severe cases, starvation and partial or whole colony mortality (Fitt et al. 2000, Grottoli et al. 2004, Rodrigues & Grottoli 2007, Rädecker et al. 2021. Heat stress is one of the most common causes of coral bleaching and the severity of the bleaching event depends on both the intensity and duration of heat stress (Middlebrook et al. 2008); thus, the degree heating week (DHW) metric was developed (Liu et al. 2003) to better characterize the cumulative stress associated with warming events. DHW integrates sea surface temperature (SST) anomalies that are 1°C above the regional maximum climatological mean, with 1 DHW representing 1°C above the maximum monthly climatological mean for a period of 1 wk (Skirving et al. 2020). When DHW reaches 4°C, significant coral bleaching is likely, while > 8°C DHW has been associated with widespread coral mortality (Liu et al. 2005). The DHW metric allows a general comparison of cumulative heat stress between events, but it can fail to accurately predict the impact of MHWs (i.e. events may be underestimated, or fine-scale events may be missed). A similar DHW may cause different bleaching responses between species and locations (Weeks et al. 2008, Bainbridge 2017, DeCarlo et al. 2017. This is, in part, because the capacity of corals to withstand heat stress is complex and influenced by a combination of environmental (e.g. Safaie et al. 2018) and biological factors including, but not limited to, coral taxa (Marshall & Baird 2000), colony morphology (Loya et al. 2001), symbiont genotype (Barshis et al. 2010), tissue thickness, and feeding behaviour (Rodrigues & Grottoli 2007, Grottoli et al. 2014. For instance, growth forms (i.e. colony morphology) are generally, but not always (Guest et al. 2012), informative in explaining bleaching tolerance: branching corals are generally more susceptible than massive and encrusting corals (reviewed by McCowan et al. 2012). Although the underlying mechanisms are not fully understood, this may be, in part, because massive and encrusting corals generally have thicker tissues (Marshall & Baird 2000, Loya et al. 2001. Additionally, corals that can increase their heterotrophic feeding whilst bleached appear to have an ecological advantage and higher chance of survival (Grottoli et al. 2004, Rodrigues & Grottoli 2007, Anthony et al. 2009, Conti-Jerpe et al. 2020. Bleaching response can be further influenced by regular exposure to anomalously high temperatures, particularly on tidal/daily time scales (Ainsworth et al. 2016, Camp et al. 2018, Safaie et al. 2018, with increased heat tolerance in corals located in thermally variable reef environments, such as back-reef pools (e.g. Palumbi et al. 2014) or intertidal reef environments (e.g. Schoepf et al. 2015).
Recently, the rate of heating from the onset of the MHW to the maximum intensity has been shown to be a better predictor of bleaching severity compared to DHWs for some species (Maynard et al. 2008, Hobday et al. 2016, Li & Donner 2022. Acropora spp. showed higher physiological stress when exposed to fast versus slow heating rates (Middlebrook et al. 2010, Martell & Zimmerman 2021. Therefore, understanding susceptibility to different heating rates may improve heat stress and bleaching predictions in combination with existing metrics (i.e. event intensity, duration, DHW) by enabling a unique description and comparison between different MHW events (Hobday et al. 2016). However, to date, the effects of heating rate have only been studied in Acropora spp. and it therefore remains unknown if other species and growth forms show similar responses, parti cularly massive and encrusting corals, which have often been shown to be less susceptible to heat stress.
To test the effect of heating rate on bleaching sensitivity, we experimentally simulated 2 different MHWs with the same cumulative intensity, but different rates of heating (0.5 vs. 1°C d −1 for slow and fast heating rates, respectively), replicating observed MHW temperature profiles described by Hobday et al. (2016). We investigated the effect of heating rate on photochemical efficiency, holobiont metabolism, tissue biomass, chl a, and symbiont density in 2 widespread coral species, Pocillopora damicornis and Plesiastrea versipora, representing branching and encrusting growth forms, respectively. We focused on subtropical Western Australia, as it is a global warming hotspot, warming faster than 90% of the global ocean and therefore a sentinel for climate change (Hobday & Pecl 2014). This region also experienced an extreme MHW in 2011 with DHW > 16, SST anomalies reaching up to 5°C above the longterm averages, and a heating rate of 0.3°C d −1 (Wernberg et al. 2013, Schlegel & Smit 2018. Although this event resulted in widespread bleaching across the majority of the Western Australian coast (Moore et al. 2012), some corals in the subtropical region, including P. versipora, increased 4-fold in abundance after this event (Tuckett et al. 2017). We hypothesized that fast onset MHWs would have more damaging effects on the physiology of the 2 species compared to slow onset MHWs, and that these effects would be greater in the branching species (P. damicornis) compared to the encrusting species (P. versipora).

Coral collection
The encrusting coral Plesiastrea versipora (Lamarck, 1816) and the branching coral Pocillopora damicornis (Linnaeus, 1758) are both widely distributed across the Indian and Pacific Oceans, and locally abundant along the mid-West Australian coast (Veron & Marsh 1988). We chose these 2 species as representatives of winners (P. versipora) and losers (P. damicornis) of the extreme MHW in 2011 (Tuckett et al. 2017). Eight visibly healthy parent colonies (approx. 15 cm in diameter) of each species were collected on 12 December 2019 (the austral summer) from Port Gregory (28° 20' 04.6" S, 114° 24' 62.8" E), Western Australia. Due to differences in habitat preferences (Veron & Marsh 1988, Veron 2000, P. damicornis colonies were collected from open reef areas at 8.5 m depth, where light intensity reached a maximum of approximately 290 μmol photons m −2 s −1 , whereas P. versipora colonies were collected from shaded crevices at 8 m depth, where the light intensity reached a maximum of approximately 200 μmol photons m −2 s −1 during the sampling period. Both light measurements were collected using a Li-Cor underwater quantum sensor (LI-192; sample rate 1 Hz) deployed around midday (11:00 to 13:00 h) over a few cloudless days in summer (10 to 12 December 2019) to determine maximum light intensities at our study sites. Colonies of both species were collected using hammer and chisel at least 10 m apart from each other to maximize the likelihood of obtaining different genotypes (Baums et al. 2006). Following collection, colonies were submerged in seawater in portable coolers and transported to laboratory facilities (5 h) where they were maintained in indoor, flow-through aquaria at 23.5 ± 0.2°C (ambient temperature at collection site) at approximately 290 μmol photons m −2 s −1 for 2 wk to allow recovery and acclimation to tank conditions.

Experimental design
Following a 14 d recovery and acclimation period, each colony was fragmented into 3 equal-sized pieces (i.e. P. versipora: 4 cm diameter; P. damicornis: 3 × 1 cm branches) and glued onto pre-labelled plastic tiles using super glue. Corals were kept for another 14 d at 23.5 ± 0.2°C prior to the start of the experiment under a reduced light intensity of approximately 120 μmol photons m −2 s −1 to aid recovery from fragmentation. Water motion was provided using submersible wave makers (JP-067, Sunsun, 2300 l h −1 ). Seawater renewal rate was 1.5 l min −1 , resulting in a turnover time of approximately 30 min in 50 l tanks. Light was provided using customdesigned LED arrangements and colours (Ledzeal S150 Plus, 150 W). Lights were programmed to a 12 h light:12 h dark cycle, following a natural diurnal light cycle with gradual increase up to 160 μmol photons m −2 s −1 at noon. The experimental light levels were lower than those measured at the collection sites for P. damicornis to avoid light stress in P. versipora, which naturally grows in shaded areas and crevices. Corals were fed 3 times a week with freshly hatched Artemia nauplii at night. Approximately 2.5 g of brine shrimp eggs were hatched in seawater for 36 h, and the stock solution was equally distributed within all tanks. Pumps were turned off for 30 min to allow for feeding, and the remaining brine shrimp was presumed flushed out within the next 30 min after pumps were turned back on.
The MHW simulations started following the second 2 wk acclimation period on 6 January 2020 (i.e. 30 d after collection). One coral fragment from each parent colony was randomly assigned to each of 3 temperature treatments: (1) ambient control, (2) slow onset MHW, and (3) fast onset MHW (see details below), resulting in n = 8 coral parent colonies per species per treatment. The temperature profiles replicated observed MHW profiles described by Hobday et al. (2016). Each temperature treatment consisted of 3 replicate 50 l transparent, plastic flowthrough tanks. Temperature increases were done using glass heaters (AquaManta 300W) corresponding to the simulated MHW onset; slow onset 0.5 ± 0.2°C d −1 and fast onset 1 ± 0.2°C d −1 (Fig. 1). Temperature in each tank (n = 9) was controlled with an Aquatronic temperature sensor and logged every 10 min. Temperature sensors were calibrated every week using a high-precision (0.01°C) thermometer (Fisher Scientific).
Slow onset treatments were kept an additional 1.5 d at the desired heat stress temperature to ensure that both treatments had the same DHWs at the end of the experiment (13 DHW for P. damicornis and 20 DHW for P. versipora). For the first phase of the simulated MHW, temperature was increased from 23.5 ± 0.2 to 26.5 ± 0.2°C in both heat stress treatments. To ensure comparability with DHWs calculated for real MHWs, experimental DHWs were calculated using 1°C increases from the maximum monthly mean (MMM) of 23.5°C (MMM provided for the nearby Houtman Abrolhos Islands, approximately 60 km to the west, by NOAA Coral Reef Watch, Skirving et al. 2020). However, 10 d at 26.5°C (4 DHW) proved insufficient to elicit changes in visual coral health (see Fig. 2), which may suggest that corals did not experience enough heat stress. Therefore, temperature was gradually increased to 28.5 ± 0.2°C (corresponding to the simulated MHW onset) and maintained for 6 d for P. damicornis and 26 d for P. versipora. The longer duration for P. versipora was necessary to induce bleaching (see Fig. 2). In total, P. damicornis fragments were subjected to heat treatment for 40 and 38 d and P. versipora fragments were subjected for 60 and 58 d for slow and fast onset treatments, respectively.

Visual coral health
Coral health was scored once a week to visually assess and quantify colour loss (i.e. paling or bleaching). Corals were scored on the upper surface of the fragments using the CoralWatch ® Coral Health Chart, where a change of 2 units corresponds to significant changes in pigment concentration (i.e. bleaching, Siebeck et al. 2006).

Photochemical efficiency of Photosystem II
Maximum dark-adapted quantum yield (F V /F M ) of chl a fluorescence (Schreiber et al. 1995) was measured 45 min after sunset (i.e. when the lights were turned off) to evaluate changes in photo-physiology throughout the experiment (War ner et al. 1996, Hill et al. 2004). All coral fragments were measured every 2 to 3 d. Measurements were done using a mini-PAM 36 Fig. 1. Temperature regimes during the heat stress test. Shaded areas indicate the first and second ramping periods; horizontal dashed line: long-term maximum monthly mean (MMM) temperature at the collection site; horizontal solid line: presumed bleaching threshold (MMM + 1°C) and the short-term in situ collected MMM values; vertical dashed line: day the experiment was ended for Pocillopora damicornis. Degree heating weeks (DHW) are given at times when the experiment was ended for each species. Data collection times are indicated as end of the experiment for *P. damicornis, and **Plesiastrea versipora fluorometer (Walz) using the following settings: measuring light intensity 3, saturation pulse intensity 12, saturation pulse width 0.8 s, gain 3, and damping 2.
Measurements were made at a constant distance of 3 mm from the coral tissue using a piece of clear plastic tubing. The same part of each fragment was measured throughout the experiment, given the fixed position of each coral.

Holobiont metabolism (P :R)
Holobiont metabolism was measured as the ratio of oxygen production through photosynthesis (gross primary productivity, P gross ) (μmol O 2 h −1 cm 2 ) to oxygen consumption through respiration (R) (μmol O 2 h −1 cm 2 ). P gross :R (hereafter P:R) was measured immediately following the end of heat stress (i.e. after 40 and 60 d for P. damicornis and P. versipora, respectively). Six fragments per treatment (representing control, slow onset, and fast onset) were randomly selected from 6 parent colonies. Fragments from the remaining 2 parent colonies were excluded due to time limitations. All incubations were done at control temperatures (~23°C) to avoid errors caused by the differences in oxygen solubility between different temperatures. This drop in temperature may have stressed the corals, although we consider this unlikely given that the cooler temperatures were well within the seasonal temperature range and may, in fact, have provided relief from heat stress. Corals were kept under constant light intensity of 160 photons m −2 s −1 during light incubations to match the maximum intensity in the experimental tanks. Incubations were performed between 10:00 and 14:00 h to avoid circadian effects of photosynthesis. All fragments were dark acclimated for at least 1 h before dark incubations and all incubations from the same fragments were performed on the same day. Oxygen, temperature, salinity, and pressure were measured continuously during incubation using a Presens SMA-OXY4 unit (Presens) calibrated once before the experiments started according to the manufacturer's instructions in an oxygen-free environment (i.e. using sodium sulphate). Corals were incubated for 60 to 75 min, depending on the size of the coral fragment, to achieve an approximately 15% change in O 2 saturation. Incubations were done in 1 l sealed, transparent plastic chambers placed in a water bath to keep the temperature constant during incubation. Each incubation round included 1 control chamber which was used to correct for the seawater microbial P gross and R. The control chamber contained a plastic tile and a bleached, clean coral skeleton of similar size to displace a roughly equal amount of water. Mixing in the chamber was provided with a submersible magnetic stirrer. Slow onset treatments were incubated for 1.5 d after the fast onset treatments to ensure both treatments had reached the same DHW. Both rounds of incubations (i.e. light and dark) were completed over 3 d. Oxygen data were normalized to incubation volume, incubation duration, and surface area of the fragment (see Section 2.6.3). P:R ratios were calculated as 12 h of gross P (= net P + R) (μmol h −1 cm −2 ) divided by 24 h of R (μmol h −1 cm −2 ) given the 12 h light:12 h dark cycle.

Sample processing
Coral fragments were sacrificed immediately at the end of their respective heat stress period and preserved at −80°C for 4 mo. Coral tissue was removed from the skeletons with deionized (DI) water using an airbrush (P. damicornis) or Waterpik (P. versipora), as these were determined to be the most appropriate methods to suit each morphology (Schoepf et al. 2015).
2.6.1. Tissue biomass. A 3 to 4 ml aliquot per sample of the tissue slurry was dried at 60°C in pre-combusted aluminium pans to constant weight and ashed in a muffle furnace at 500°C for 4 h. Ash-free dry weight (tissue biomass) was determined as the difference between dry and ash weight (mg) and normalized to surface area (mg cm −2 , see Section 2.6.3).
2.6.2. Chl a and symbiont density. The remaining tissue slurry was centrifuged for 10 min (3220 × g) to separate the animal host and symbiont fraction. The symbiont fraction was resuspended in DI water and used for measuring chl a and symbiont density. Chl a was extracted in 100% acetone in the dark at 4°C for 24 h. Concentrations of chl a from each sample were determined spectrophotometrically using individual glass cuvettes following the equations of Jeffrey & Humphrey (1975) for dinoflagellates. Chl a concentrations were normalized to both surface area (μg cm −2 ) and cell density (pg cell −1 ). Symbiont cell density was calculated using 8 replicate counts on an improved Neubauer hemacytometer and standardized to surface area (10 6 cells cm −2 ). 2.6.3. Surface area. Surface area was estimated using the most appropriate methods suiting each morphology (e.g. Schoepf et al. 2015). To this end, the wax-dipping technique (Veal et al. 2010) was used for P. damicornis and the aluminium foil technique was used for P. versipora (Marsh 1970).

Statistical analyses
All statistical analyses were done using R version 3.5.3 (R Development Core Team 2019). Generalized linear mixed models (GLMM) were computed using the 'glmer' function from the 'lme4' package (Bates et al. 2015) to analyse the effect of heating rate (fixed effect with 3 levels: control, slow onset, fast onset) on tissue biomass, endosymbiont density, chl a, P:R, respiration, gross photosynthesis, and F V /F M measurements for each species. Models were built separately because each species experienced different durations of heat stress and DHWs. Tukey's tests using Kenward-Roger degrees-of-freedom adjusted p-values were used to test for the effect of heating rate when main effects were significant, using the 'multcomp' package (Hothorn et al. 2008). Parent colonies were used as random group intercepts but had no influence on any of the response variables. We selected the model with the error distribution that best fit the data. Different models were compared using Akaike's Information Criterion (AIC), and a Gamma error distribution with a log link function was selected as it had the best balanced fit and parsimony. Residuals were visually inspected using the 'performance' package for violations of statistical assumptions (Lüdecke et al. 2021). Significance of fixed effects was determined using the ANOVA function of the 'rstatix' package (Kassambara 2021), which provided Analysis of Deviance tables (Type III Wald chi-square tests). Multiple pairwise comparisons were done for F V /F M for the days of the experiment that had the same cumulative heat stress exposure (13 DHW for P. damicornis and 20 DHW for P. versipora). This corresponded to the Day 40 value for slow onset with Day 38 value for fast onset for P. damicornis, and the Day 60 value for slow onset with Day 58 value for fast onset for P. versipora.

Visual coral health
All fragments appeared visibly healthy with full tissue cover, no discolouration, and no signs of disease at the start of the experiment, and control fragments remained visibly healthy throughout the experiment. Complete mortality of the tissue surface only occurred in heat-stressed fragments belonging to 1 Pocillopora damicornis parent colony. At the end of the experiment (i.e. on Days 40 and 60 for P. dami -cornis and Plesiastrea versipora, respectively), heatstressed fragments of both species were visibly paler (health score 2 ± 1; mean ± SD, n = 8 per species and treatment) than the controls (health score 4 ± 1; n = 8).

Photochemical efficiency of photosystem II
F V /F M remained high and stable (P. damicornis 0.69 ± 0.05; P. versipora 0.67 ± 0.07; n = 8) within control fragments over the course of the experiment (Fig. 2). Heated P. versipora fragments maintained F V /F M values similar to their control treatments for 20 d longer than P. damicornis fragments, before experiencing a similar decrease (Fig. 2B).
Heating had a significant negative effect on F V /F M in both species. For P. damicornis, heat stress treatments had significantly lower (by 20%) F V /F M values at the end of the experiment compared to their controls (Tables 1 & 2) (Days 40 and 38 for slow and fast onset, respectively). However, there were no differences in reduction of F V /F M between slow and fast onset treatments ( Fig. 2A). For P. versipora, declines in fast onset treatment corals were greater than in slow onset treatment corals, and only fast onset treatments were significantly different to the control treatments (12% lower) (p = 0.007) at the end of the experiment (Days 60 and 58 for slow and fast onset, respectively) (Fig. 2B, Table 2).

Holobiont metabolism (P :R)
Heating rate had a significant effect on P:R at the end of the experiment in P. damicornis (p < 0.001) ( Table 1), but not P. versipora (Fig. 3C,F, Table 2). Fast heating rate resulted in greater declines in P:R than slow heating rate in P. damicornis, with fragments experiencing a 65% decrease compared to the controls and the loss of net autotrophy (P:R < 1) at the end of their heating treatment and slow onset fragments experiencing 26% decrease. No significant change was observed in the rate of gross photosynthesis (P gross ) for either species at the time of incubation (Tables 1 & 2). Similarly, heating rate had no significant impact on the respiration (R) of P. versipora; however, it had a significant effect on P. damicornis, with fast onset corals having significantly higher respiration rates (0.5 ± 0.2 μmol h −1 cm −2 , mean ± SE) than both control (0.2 μmol ± 0.1 h −1 cm −2 ) and slow onset corals (0.2 μmol ± 0.1 h −1 cm −2 ) (Fig. 3B

Chl a, symbiont density, and tissue biomass
Chl a per area decreased significantly (85 to 87%) as a consequence of heat stress in P. damicornis, but there was no significant difference between the slow and fast onset treatments (Fig. 4A, Table 1). However, heating rate had a significant effect when chl a was normalized per cell (Fig. 4B) rather than per area. In fact, we found significantly higher chl a per cell in corals in fast onset treatments than corals in either slow onset (p = 0.018) or control treatments (p < 0.001). This was associated with significant declines in symbiont densities (87 to 90%) (Fig. 4C) in both slow and fast onset treatments compared to the control treatments, which were independent of heating rate.
Heating also had a significant negative effect on symbiont-related parameters in P. versipora corals, where chl a per area (Fig. 4E) and symbiont density (Fig. 4G) were both significantly lower than in the control (63 and 46%, respectively) ( Table 2), but heating rate had no significant effect. However, unlike in P. damicornis, we found no difference between slow and fast onset treatments when chl a was normalized per cell rather than area, although heated corals in both treatments had significantly lower values (−30%) than the control. At the end of the heating period, heated P. damicornis had 36% less biomass (Fig. 4D) than the controls; however, heating rate had no significant effect. In contrast, heating rate had no significant impact on P. versipora tissue biomass (Fig. 4H, Table 2)

DISCUSSION
The present study compared the responses of 2 coral species, representing different growth forms, to 2 ecologically relevant MHW scenarios with the same cumulative heat exposure but different heating rates for each species, testing the hypothesis that heating rate can play an important role in bleaching responses. Consistent with other studies that have shown detrimental effects of rapid heating on branching Acropora corals (Middlebrook et al. 2010, Martell & Zimmerman 2021, our results showed sensitivity to heating rates but these responses were species-specific. Species-specific responses to heating rate were most evident in P:R ratios, with encrusting Plesiastrea versipora not impacted by either MHW simulation, while rapidly progressing MHW conditions had detrimental effects on branching Pocillopora damicornis, which ultimately resulted in the loss of net autotrophy (i.e. P:R < 1). In contrast, under slowly progressing MHW conditions, P. damicornis fragments were able to maintain net autotrophy, although the lowered P:R ratios observed here may have led to partial depletion of energy reserves. Photosynthesis (P gross ) rates under rapidly evolving MHW conditions remained relatively comparable to control treatments at the end of the experiment, which might be explained by the fact that photosynthesis rates might be independent of total available chl a and inversely correlated with symbiont density (Dubin-   sky et al. 1990, Rodrigues & Grottoli 2007). Therefore, the significant change in P:R ratios of rapidly heated P. damicornis corals appears to be driven by the significant increase in respiration, which has also been observed in several other studies (Coles & Jokiel 1977, Hoogenboom et al. 2010, Baker et al. 2018 and might be one of the key mechanisms to explain susceptibility (e.g. Loya et al. 2001) of this species to rapid onset MHWs.
Additionally, our study showed that faster heating rate had a negative effect on photochemical efficiency, as highlighted by the greater declines in rapidly heated P. versipora compared to the slowly heated treatments. This finding is consistent with the previous studies on the negative effects of rapid heating rates on branching Acropora corals (Middle-brook et al. 2010, Martell & Zimmerman 2021. A similar response might have been expected from rapidly heated P. damicornis fragments, especially considering the significant reductions we observed in holobiont P:R ratios. However, as this species is generally sensitive to heating (i.e. Marshall & Baird 2000, Loya et al. 2001, it is likely that 0.5 and 1°C d −1 may have both been too rapid and/or similar to each other to allow a significant change for this species to be observed. In general, the fast onset treatment corals showed earlier declines in F V /F M than the slow onset treatments in both species, even though at the Lowercase letters indicate results from Tukey-adjusted multiple pairwise comparisons when the effect of treatment was significant. Note the different scaling for the 2 species for all parameters end of the experiment they experienced the same overall declines in magnitude. Lack of a strong change in the F V /F M may also be due to the nature of the experimental design used in this study. We used more ecologically relevant heating rates (0.5 and 1°C d −1 ) compared to the acute heat stress assays that are most often used (reviewed by McLachlan et al. 2020). As such, heating rate effects may have revealed some differences which could be difficult to detect in a short-term stress test. This further highlights the importance of employing more longer-term, ecologically relevant experiments. It is further important to highlight that while the maximum light levels in the experiment were similar to natural levels at the collection site for P. versipora, they were somewhat lower than the maximum light levels that P. damicornis is naturally exposed to (i.e. 160 vs. maximum 290 photons m −2 s −1 , respectively). This could have contributed to the observed differences in bleaching sensitivity between species. However, the use of low light levels can be informative to understand how much heat stress is required to trigger thermally induced bleaching (Anthony et al. 2007, Middlebrook et al. 2010. P. damicornis was more sensitive to heat stress compared to P. versipora, indicated by the ability to maintain high and stable F V /F M 20 d longer during the experiment. The sustained photosynthetic response of P. versipora relative to P. damicornis suggests species-specific differences in tolerance of sustained heat stress, which is consistent with previous studies on heat tolerance within this region (Tuckett et al. 2017). It is important to note, however, that these different responses may also have been caused by the different types of Symbiodiniaceae hosted by these coral species in Western Australia (Breviolum and Cladocopium in P. versipora and P. damicornis, respectively) (Silverstein et al. 2011, LaJeunesse et al. 2018. For instance, while some symbiont genera and/or species are known to be susceptible to higher temperatures, others are exceptionally resilient (Hoadley et al. 2015, Hoogenboom et al. 2017). Nevertheless, higher heat tolerance could be a critical feature for P. versipora to adapt to the gradual increases in ocean temperature, which could promote resistance and recovery to future MHW events.
Our results did not show an effect of heating rate on colony biomass, which may be an important indicator for coral health and bleaching sensitivity (Thornhill et al. 2011, Grottoli et al. 2014, but see Wall et al. 2019). During bleaching where severe losses of symbionts take place, corals depend on alternative strategies to meet their daily metabolic energy requirements (Muscatine et al. 1981). Catabolism of energy reserves (i.e. lipid, protein, and carbohydrate levels) and heterotrophy are two such strategies (Fitt et al. 2000, Grottoli et al. 2006, Rodrigues & Grottoli 2007. Species that are able to increase their heterotrophic input of fixed carbon under heat stress are thought to have an ecological advantage in survival compared to species that deplete their re stricted source of energy (i.e. Grottoli et al. 2006, Conti-Jerpe et al. 2020). P. versipora was able to maintain its tissue biomass under both heating rates. This could potentially indicate greater capacity for heterotrophic plasticity and may have provided bleaching resilience (Anthony et al. 2009). Alternatively, this could also be the result of this species' ability to maintain P:R ratios and net autotrophy under heat stress. Conversely, P. damicornis experienced a significant decline in tissue biomass which may be explained by the catabolism of stored energy reserves (Porter et al. 1989, Rodrigues & Grottoli 2007. A reduction in tissue biomass under ocean acidification, another climate change stressor, has also been shown for Pocillopora acuta, a species closely related to P. damicornis (Wall et al. 2017, but see Schoepf et al. 2013). Colonies with high tissue biomass are known to be less susceptible when exposed to thermal stress compared to those with low tissue biomass (Fitt et al. 2000, Loya et al. 2001, Thornhill et al. 2011. As an encrusting coral species, the tissue biomass in the control treatment for P. versipora was 3 times higher than in the control treatment for P. damicornis, which might have provided protection from light and more effective self-shading of the symbiont cells (Hoegh-Guldberg 1999, Cunning & Baker 2014). This finding is consistent with an ecological advantage provided by the differences in surface area to volume ratios and thicker tissues in massive and encrusting species compared to branching colonies (i.e. Loya et al. 2001).
The differences in heat tolerance and sensitivity to heating rate between 2 common coral species may have broader implications for the reef community. This might especially be the case in tropical−temperate transition zones where species are undergoing major redistributions (Vergés et al. 2014, Pecl et al. 2017. While many tropical coral reefs and temperate kelp forests continue to degrade and homogenize with warming oceans (Alvarez-Filip et al. 2009, Krumhansl et al. 2016, Wernberg et al. 2016, Perry & Alvarez-Filip 2019, the tropical−temperate transition zones are becoming 'tropicalized' with poleward-moving tropically affiliated corals (Precht & Aronson 2004: Acropora spp.; Thomson 2010: Gonio-pora spp.; Yamano et al. 2011: Acropora spp., Pavona spp.; Baird et al. 2012: Acropora spp.). More opportunistic corals are increasing in abundance, enabled by competitive release from historic habitats. This has been observed in Western Australia as a result of the 2011 MHW, which resulted in major loss of canopy-forming kelps and the indirect increase in coral abundance (Wernberg et al. 2016, Tuckett et al. 2017. However, the structural complexity and biodiversity which might be supported by novel coral assemblages will ultimately depend on their resilience to future warming events (Vergés et al. 2019). For instance, in the tropical−temperate transition zone in Western Australia, P. versipora has shown high heat tolerance and survival under an extreme MHW (i.e. 16 DHW), particularly when compared to other locally abundant species (i.e. P. damicornis) (Tuckett et al. 2017). This is consistent with our findings of a sharp decline in F V /F M when DHW reached 18 for P. damicornis, yet P. versipora re quired an additional 20 d to show similar declines. While our study identifies some of the mechanisms behind this outcome, it also suggests that, under rapidly evolving MHWs, these novel coral communities might shift towards less complex morphologies (i.e. dominated by the encrusting P. versipora). Therefore, with the projected increases in frequency and intensity of MHWs , the functional attributes that can be supported by these novel coral communities (i.e. tropicalized coral assemblages) (Vergés et al. 2019) might be lower than estimated.
Overall, our results confirm that in addition to Acropora species, corals from other genera are also susceptible to heating profiles of MHWs, such as the heating rate. Our experiment as a mid-term, gradual heat-stress onset design also allows comparison with many natural bleaching events (McLachlan et al. 2020, Grottoli et al. 2021. Although the responses were species-specific, the differences in heating rate appear to be consistent with the idea that massive and encrusting corals may be less sensitive to faster onset in heating than branching corals. Additionally, highlighted by the loss of net autotrophy, these corals may rely more upon heterotrophic feeding to overcome this disadvantage. Understanding speciesspecific sensitivities to heating rates in combination with existing metrics (i.e. DHW) is likely to enable us to make better predictions on impacts of MHW generally, but also particularly for species-specific responses. These subtle changes (i.e. heating rates) can influence responses of some coral species more than others, which can have consequences for future community assemblages and reef functions. Data availability. The data and R code supporting the findings of this study are openly available in https://github. com/defneasahin/Heating_rates.