Coral growth persistence amidst bleaching events

As mass bleaching events decimate stony coral populations, production of calcium carbonate is diminished on reefs, dampening their capacity to keep pace with rising sea levels. However, perturbations to the calcification process of surviving wild corals during bleaching are poorly constrained, owing to the lack of suitable techniques to retroactively extract this information from coral skeletons at sufficient resolution. Here, we use novel Raman spectrometry techniques to test the biogeochemical response of long‐lived corals before, during, and after bleaching. Maintenance of high aragonite saturation state (ΩAr) in the coral calcifying fluid is key to driving rapid skeletal growth but would be expected to decrease when corals become energetically depleted without their symbionts. Contrary to this expectation, our results demonstrate that corals upregulate calcifying fluid ΩAr during bleaching and for at least 2 yr after recovery. This indicates that the calcification process of coral‐bleaching survivors is unexpectedly resilient.

. This thermal stress threatens coral calcification processes necessary for constructing the calcium carbonate (CaCO 3 ) building blocks that are eventually cemented together into reef structures (Harris et al. 2018).
Stony corals form their skeletons by sequentially depositing granular and fibrous crystals of the mineral aragonite, a polymorph of CaCO 3 , at a rate dependent on the geochemistry of the semi-isolated space between the tissue and skeleton.This calcification process is both biologically and geochemically controlled in the small fluid-filled space between the coral tissue and skeleton (Cohen and McConnaughey 2003;McCulloch et al. 2017;Tambutté et al. 2011;Thompson et al. 2022).Corals passively and actively transport external seawater into the calcifying fluid to supply the ions and dissolved inorganic carbon (DIC) needed to calcify, while simultaneously increasing the calcifying fluid pH to favor rapid aragonite crystal growth (Furla et al. 2000).Specifically, the thermodynamic favorability of aragonite deposition is dependent on the Ω Ar of the calcifying fluid, expressed by the equation , where K sp is the solubility product of aragonite that is affected by temperature, salinity, and pressure (Mucci 1983).As corals increase calcifying fluid pH, the seawater carbonate system shifts toward CO 3 2À , increasing Ω Ar .Generally, the greater the Ω Ar , the more calcification is favored over dissolution, and the more rapidly aragonite crystals will precipitate (Burton and Walter 1987;Gattuso 1998).Additionally, aragonite crystals precipitated at higher Ω Ar will be more disordered than those formed in lower Ω Ar conditions because of their more rapid formation and incorporation of impurities (DeCarlo et al. 2017;Farfan et al. 2018Farfan et al. , 2022)).
The energy required for the upregulation of the semi-isolated calcifying fluid and ultimately precipitation of the coral skeleton is obtained via coral heterotrophy and symbiont photosynthesis, with the bulk of the energetic requirements being met by the symbiont photosynthate (Cohen and Holcomb 2009).Thus, retention of symbionts is thought to be crucial to not only coral survival but also their rapid calcification (Davies 1984), as exemplified by faster growth rates in symbiont-bearing corals (Dimond and Carrington 2007).However, there is debate as to whether the active transport of DIC into the calcifying fluid is predominantly controlled by energetic limitations or surrounding DIC concentrations (McCulloch et al. 2017;Thompson et al. 2022).Yet, reductions in lipid reserves available to use for calcification have been observed with the loss of symbionts and/or impairment of photosynthetic processes (Iglesias-Prieto et al. 1992;Jones et al. 1998;Grottoli et al. 2004;Cheung et al. 2021;Thompson et al. 2022).Bleached corals are effectively starving and can only survive as long as their lipid reserves last or they can compensate with heterotrophic feeding (Mendes and Woodley 2002).This disruption of calcification due to bleaching is also supported by the perturbation of geochemical proxies in coral skeletons during bleaching (DeCarlo and Cohen 2017; D 'Olivo and McCulloch 2017;D'Olivo et al. 2019).
However, the link between bleaching and the regulation of calcifying fluid carbonate chemistry to facilitate calcification despite decreased energy reserves remains unclear.
The coral calcifying fluid Ω Ar is difficult to directly study due to its small volume and isolation beneath living tissue layers.Indeed, existing techniques to quantify the chemistry of this fluid rely on either invasive microprobes or bulk isotopic analyses that cannot yet be applied on the timescales of weeks over which bleaching events commonly occur (Leder et al. 1991;Schoepf et al. 2014;D'Olivo and McCulloch 2017;Thompson et al. 2022).A new approach that is currently the only method with the ability to measure Ω Ar of the calcifying fluid and resolve variations on weekly timescales is using the full width at half maximum (FWHM) of Raman v 1 peaks (DeCarlo et al. 2017(DeCarlo et al. , 2019)).This nongeochemical approach not only allows increased resolution of calcifying fluid Ω Ar measurements but also removes some of the uncertainty associated with the assumptions necessary to perform boron-based Ω Ar reconstructions (DeCarlo et al. 2018a,b;Thompson 2022).Thus, Raman spectroscopy allows the reconstruction of calcifying fluid Ω Ar , which controls crystal precipitation rates, at higher resolutions and certainty than any other current methodology.
Raman spectrometry measures the shifts in wavelength of scattered light returned when a laser is directed at a material.The shift differs depending on the characteristics of the material, providing insight on the differences in coral skeletal composition that accompany stress events.One such difference between Raman spectra is the FWHM of peaks.The FWHM of v 1 peaks is used as a measure of aragonite crystal disorder, which is directly impacted by Ω Ar (DeCarlo et al. 2017;Farfan et al. 2022).Thus, with Raman Spectroscopy, v 1 FWHM measurements can be used to retroactively calculate coral calcifying fluid Ω Ar at the time of skeletal formation.Using this technique, for the first time, we were able to observe changes in calcifying fluid chemistry at high resolution for over 2 yr before and after a severe bleaching event in the Red Sea.

Site description
Southern Red Sea coral reefs are unique in their exposure to very warm (up to 33 C) waters, but with high variability due to upwelling (DeCarlo et al. 2021).The oceanographic isolation, combined with strong environmental gradients, has led to relatively high levels of endemism in Red Sea reef fauna (DiBattista et al. 2016).While corals living in this region are adapted to these unique conditions (Ziegler et al. 2019), enabled in part by their associations with heat-tolerant endosymbionts (Terraneo et al. 2019), they are still sensitive to anomalously warm summers and have bleached during heat extremes (DeCarlo et al. 2021;Gajdzik et al. 2021).

Samples
In summer 2015, the southern Red Sea experienced one of the most severe coral bleaching events ever observed.At some reefs, 99% of corals experienced bleaching (Ormond et al. 2016;Osman et al. 2018), and the region as a whole lost 44% of live coral cover (Anton et al. 2020).Based on coral skeletal cores-the same ones used in this study-the severity of the bleaching event was identified as being unprecedented over at least the past century (DeCarlo 2020).Within the skeletons of surviving massive corals (in the genus Porites), anomalous high-density skeletal bands, called "stress bands," archive the event (DeCarlo et al. 2020).As a result of this particularly intense bleaching event, the stress bands formed within these coral skeletons are among the most intense ever observed (DeCarlo et al. 2020).

Sample preparation
Thirteen Porites spp.coral cores from the Farasan Banks region of the Red Sea previously collected and described by DeCarlo et al. (2020) were analyzed here for changes in calcifying fluid chemistry across the 2015 bleaching event (Fig. 1).All 13 cores contained clear high-density skeletal bands indicative of bleaching events.These stress bands were visually identified and measured from computed tomography (CT) scans using Horos software (Fig. 2).Cores were cut in 3-5 mm thick slabs along the primary growth axes for Raman analyses using a wet tile saw.Cut slabs were sonicated for 10 min and then dried to remove loose CaCO 3 particles.

Raman data collection
A Renishaw inVia confocal Raman microscope with a 785 nm laser source was used to obtain all Raman measurements.Before any spectra collection, the spectrometer was calibrated using an internal silicon chip embedded within the instrument.Such calibrations included alignments of the laser, slits, charged coupled device area, and silicon reference sample.All spectra were then collected using a 50Â objective, 1% laser power (from a 45 mW source), 0.5 s acquisition time, and a 1200 mm À1 grating.We followed established methods (DeCarlo et al. 2017) to create a calibration of v1 FWHM ¼ 0:321 Â ln ΩAr ð Þþ3:21; R 2 = 0.95 (see Supporting Information Fig. S1).

Raman mapping
Starting on the visually identified stress bands, coral core spectra were collected along the core every 10 μm $2 cm above and below the stress band.Porites corals typically grow $1 cm yr À1 , therefore, this sampling resolution gave us approximately two Raman measurements per day of coral growth.Our age model, however, is limited by annual-band tie points and because growth rates may vary seasonally, we cannot assign spectra to specific calendar days (i.e., our time series have replication equating to sub-daily, but our age models are lower resolution).Three replicate lines were measured along this path spaced over 1 cm perpendicular to the growth axis to capture spatial variability within the core.This analysis produced $12,000 spectra per coral core (Mantanona and DeCarlo 2023).Poor quality and out-of-focus spectrum were filtered out before statistical analyses were performed (see Supporting Information Data S1 for additional details).

Extension analyses
Corals' linear upward growth, or linear extension, was measured from CT scans by counting the annual bands along the Raman analysis line before and after the stress band.Distance measured along the analysis line was then divided by years derived from CT images to obtain linear extension rates corresponding to our Raman data.

Statistical analyses
To account for differences in mean Ω Ar among cores, we calculated Ω Ar anomalies by subtracting the core-mean Ω Ar from each individual Ω Ar measurement.These saturation state anomalies for each coral were then binned in R (2021) as "before," "during," or "after" the bleaching event by distance from the stress band using the width of the stress band estimated from CT scans of each core using Horos software."Before" the bleaching event was defined as all measurements down the core from the identified stress band and the subsequent recovery period after bleaching was defined all measurements above the stress band.For each core, values in each bin were averaged.Each core's average Ω Ar values were then compiled into a new list containing the "before," "during," and "after" bleaching Ω Ar anomaly values for all 13 replicate cores.
All statistical analyses were performed in R (2021).To test if there was significant change in Ω Ar anomaly before, during, and after bleaching, we performed a repeated measures ANOVA.All assumptions of ANOVA were met, including homogeneity of variances with Bartlett's test.A significant difference in Ω Ar anomaly was found.Therefore, paired pairwise t-tests were used to determine which binned groups differed significantly.Additionally, temporal changes in Ω Ar anomaly for each core were examined by creating a plot of the averages of all replicate cores' Ω Ar anomaly with subsequent standard error values (calculated with the 13 cores as the units of replication) over the distance from the stress band.To do this, Ω Ar anomalies of each core's three replicate transect lines were averaged so each down-core distance had only one Ω Ar anomaly value.These values for each core were then compiled and replicate Ω Ar anomalies for each down-core distance was averaged among cores.
To estimate time elapsed over the sampled section of each core, light/dark bands representative of 1 yr of growth were counted from CT scans before and after the stress event.
Under the assumption that growth was linear, these years were set as tie-points of the sampling period, which allowed time to be linearly interpolated from distance values along the core taken during Raman map acquisitions.This enabled us to interpolate to sub-annual timescales based on each core's annual banding pattern.However, the resolution of specific days or months cannot be made since the tie-points of our interpolation are annual.For analyses of extension rate, a paired t-test was performed to test for differences in extension rate before and after bleaching.

Results and discussion
Calcifying fluid Ω Ar increases during bleaching and recovery Surprisingly, despite undergoing one of the most intense bleaching events recorded, we found that these corals showed the ability to elevate Ω Ar during bleaching and recovery (Fig. 3; Supporting Information Figs.S3 and S4).Ramanderived Ω Ar anomalies from the 13 replicate cores showed differences in Ω Ar anomaly in response to bleaching (Fig. 4a; F [2,24] = 8.55, p = 0.002), revealing that bleaching significantly affects the calcifying fluid chemistry.Mean Ω Ar anomalies (standardized to each core's average) before bleaching (À0.602AE 0.587) were significantly lower than Ω Ar values both during (0.415 AE 0.796, p = 0.03) and after bleaching (0.441 AE 0.459, p = 0.01).No significant difference in Ω Ar anomaly was observed between the during-and afterbleaching periods ( p = 1.0), indicating that Ω Ar remained upregulated during recovery even when symbionts returned.
There are several potential mechanisms to explain our observed enhancement of coral calcification processes during and after bleaching.First, this response may be the result of reduced negative competition with symbionts for the inorganic carbon needed for both calcification and photosynthesis (Davy et al. 2012).Thus, increased pools of available DIC, and subsequently CO 3 2À , would increase Ω Ar .Similarly, the increased Ω Ar that we observed may be explained by the proton flux hypothesis, which states that the ability for corals to calcify is primarily dependent on their ability to pump H + against the gradient between the calcifying fluid (lower H + ) and external seawater (higher H + ) to facilitate the formation of CaCO 3 from HCO 3 À (Jokiel 2011a,b).Symbiont photosynthetic activity within the coral tissues increases internal pH (i.e., lower H + concentration; Inoue et al. 2018), therefore, the loss of symbionts would decrease the proton gradient between the coral and external seawater, making the conversion of HCO 3 À to H + and CO 3 2À more favorable, and in turn elevate Ω Ar .However, ongoing ocean acidification increases this proton gradient, opposing the mechanisms driving skeletal formation, potentially hindering any positive effect of decreased competition with symbionts for DIC (Jokiel 2011b).While these mechanisms offer possible explanations for the increased Ω Ar in the absence of symbionts during bleaching, the maintenance of elevated Ω Ar for multiple years after bleaching is more enigmatic.Our post-bleaching analyses extend for 2 yr following the event, meaning that the increase in Ω Ar may not be a short-lived perturbation but rather represents a permanent shift in the calcification mechanism.First, it is possible that symbiont populations did not recover to their pre-disturbance levels.Pigmentation is a poor indicator of symbiont densities (Apprill et al. 2007), thus, even apparently recovered corals may still contain depleted symbiont populations, prolonging the elevated Ω Ar that arose during bleaching.Additionally, it has been observed that some corals may shuffle their symbiont communities to favor higher relative densities of thermally tolerant genera after bleaching recovery, especially in circumstances of severe thermal stress such as corals experienced in this study (Stat et al. 2009;Cunning et al. 2015;Webster and Reusch 2017;Quigley et al. 2022).This change in symbiont community composition inherently changes the carbon-energy dynamics between symbionts and their coral hosts, as there are various trade-offs associated with increased thermal tolerance.One such trade-off is decreased photochemical efficiency (Jones and Berkelmans 2011;Cunning et al. 2015), which increases the carbon available for the coral, as its symbionts would no longer fix as much carbon as less-  thermally tolerant symbionts.This increased DIC pool available for corals to use in calcification may explain why Ω Ar and linear extension were elevated years after corals recovered symbionts.
Finally, the increases in Ω Ar during and after bleaching may be driven by changes in the coral animal, rather than host-symbiont dynamics.In particular, our results could be indicative of a shift toward more nighttime relative to daytime calcification.It is thought that granular centers of calcification-the nuclei of incipient fibrous crystalsprecipitate at nighttime (Cohen and McConnaughey 2003) and are characteristically formed at elevated Ω Ar (DeCarlo et al. 2018b), likely a result of the higher saturation needed to overcome the thermodynamic barrier to crystal nucleation.Thus, during bleaching when light-enhanced calcification and symbiont-derived energy for daytime fibrous crystal growth decreases, the resulting skeleton would have a relatively high ratio of granular to fibrous crystals, resulting in an apparent increase in bulk Ω Ar .This explanation, however, is not consistent with the observed increase in extension rate because fibrous crystals compose the majority of the skeleton and are the main contributor to linear extension, not COCs.Thus, it is possible that a change in nighttime vs. daytime calcification patterns could contribute to our findings, especially the elevated Ω Ar during bleaching, but it is unlikely to explain the maintenance of elevated Ω Ar for years afterwards.
Our results provide insight into the mechanisms involved in coral responses to bleaching on timescales not yet observed.This increased understanding of mechanisms driving skeletal formation informs the capacity for corals to regulate their calcifying fluids and therefore survive and grow despite bleaching.Regardless of the underlying mechanism, the capacity for bleaching resilience shown here helps determine how reefs will change in the future, impacting not just corals but the ecosystems they support.It is important to note, however, that corals in the Red Sea (like those used in this study) are unique in their frequent exposure to thermal extremes.This repeated exposure thermal stress may lead to acclimatization, contributing to the observed resilient response to bleaching, a response that may differ compared corals in areas, which have historically experienced more stable temperature regimes (Thompson and van Woesik 2009).Additionally, the increasing acidity of the ocean and its compounding effects with bleaching may lower the coral calcifying fluid Ω Ar beyond what corals can upregulate to maintain calcification, as decreases in just one Ω Ar unit of seawater have been shown to decrease coral calcification by 15% (Chan and Connolly 2013;Thompson et al. 2022).Similar studies need to be conducted using corals from additional sites to explore the extent that site and species specificity influence bleaching responses in the calcifying fluid.Although mass bleaching events are still expected to be the leading cause of future reef decline (Cornwall et al. 2021), those corals that survive bleaching may exhibit more calcification resilience to heatinduced stress than previously recognized.

Fig. 1 .
Fig. 1.Coral core sample locations.(a) Elevation map of the Red Sea region, with inset map showing location of the Red Sea in the northwestern Indian Ocean and the red box indicating the Farasan Banks region.(b) Farasan Banks region of the Saudi Arabian Red Sea with coral reefs shown in black outlines and core sampling locations indicated by red points.

Fig. 2 .
Fig. 2. Coral growth and stress banding.Examples of CT scans of coral cores showing the annual banding pattern (orange bracket) and high-density stress band from a bleaching event (blue arrows).Light shading indicates relative high-density areas as opposed to dark shaded areas.Scale bars are 1 cm.

Fig. 3 .
Fig. 3. Temporal variability of Ω Ar anomalies.Time series of mean (black line) Ω Ar anomaly from the 13 cores, with blue shading showing standard error among cores.The x-axis is years relative to bleaching, with the gray bar showing the approximate width of high-density stress bands.

Fig. 4 .
Fig. 4. Coral calcifying fluid saturation state across coral a bleaching event.(a) Saturation state (Ω Ar ) anomaly distributions before, during, and after bleaching and (b) Linear extension rates before and after bleaching.Individual core mean values are indicated with gray circles, which are randomly offset horizontal to aid visualization.Shading of bars is indicative of the density distribution of datapoints and median values for each time bin are indicated with a white dot.Thickness of black vertical bar indicates 95% and 50% of the data distribution.