Along Florida’s Coral Reef, advances in the efficacy and scalability of nursery coral propagation and outplanting coupled with stress-testing have great potential to improve ecological restoration success and buffer degraded ecosystems. Yet, the most ecologically relevant method to assess species responses to climate change-associated stressors is long-term exposures (LT; 1–2 months) that are limited in scalability (number and replication of genotypes), efficiency (time and cost), and precision (potential for tank acclimation). Here, we compared the physiological responses to chronic (2 mo) and acute (18 hr) thermal stress in the prioritized Caribbean restoration coral species A. cervicornis, with the aim to increase scalability and standardization of thermal resistance screening. An advantage of using CBASS over LT experiments was the ability to determine nursery population thermal thresholds15, which was 34.37°C, and individual thresholds, which ranged from 34.00-34.72°C. In contrast, LT thermal exposures either need multiple temperature treatments36,37 or preliminary assessments to identify temperatures at which the strongest differential responses can be detected at the individual level. Moreover, it took over 197 hrs of chronic thermal stress to achieve the same bleaching response (ED25) detected within 5 hrs under acute thermal stress with the CBASS. In a restoration setting, there may be value in replacing chronic exposures with CBASS to improve scalability, especially if fine-scale physiological differences are conferred between the two approaches.
Across all phenotypic traits measured – photochemical efficiency of PSII (Fv/Fm), photosynthetic pigment (total chlorophyll [a + c2]), and host soluble protein – responses were most similar between the CBASS 34.3°C treatment and the LT ocean warming (LT-OW; 30.5°C) treatment. Importantly, the CBASS 34.3°C treatment was closest to the nursery population’s upper bleaching threshold (ED50) and resolved most bleaching responses found under LT-OW. Yet, when each trait was independently compared among the two experimental treatments, bleaching responses did indeed differ, likely as a result of different heating rates and heat accumulation under chronic and acute exposures. These findings are in contrast to other comparative thermal stress studies36,37, which observed comparable bleaching responses between treatments under ‘classic’ and acute thermal stress. It is notable that these ‘classic’ bleaching experiments refer to 10-14d of treatment whereas the chronic heat treatment herein occurred over two months. Although two months of thermal stress would be considered more ecologically relevant than a 10-14d exposure to achieve realistic measures of bleaching responses in longer-to-respond physiological traits, such as host protein concentrations33,46, it could be too long a time frame to achieve comparable results to the 18hr exposure of a CBASS experiment. Moreover, longer exposure periods can introduce photoacclimation to tank effects35 and confounds interpretations for heated corals, evident in the significant decline of the LT-Control Fv/Fm values observed over the course of the experiment, despite ~ three weeks of post-collection acclimation. Since chlorophyll measures were not taken prior to the LT experiment, we cannot know whether chlorophyll values also acclimate to the experimental light regime, but it appears chlorophyll photophysiology adjusted differently than photochemistry given the significant effect of treatment for chlorophyll under the LT experiment. In contrast, there was not enough time for photoacclimation under acute exposures, as demonstrated by CBASS, where all treatments equal to or greater than the bleaching temperature of 34.3°C were different from control values. Consequently, the varying heating durations between the two experiments, which represent the extreme ends of typical thermal stress exposures, likely exert different influences on the molecular, cellular, and physiological responses to thermal stress.
Most research utilizing CBASS to rapidly determine bleaching thresholds has focused on detecting population or regional-level distinctions36,37, with only one study incorporating colony-level assessment15. Hence, this study marks the first attempt to investigate thermal tolerance rankings among genotypes. Restoration programs with objectives to increase climate resilience would benefit from rapidly generated information regarding which nursery genotypes are likely to survive thermal stress. Instead of having clearly defined ‘winner’ genotype(s), genotype rankings were not straightforward and depended on the trait investigated and experiment type. Similar to the findings from Cunning et al.15, in each nursery population along FCR, there appears to be a mixture of tolerant and susceptible genotypes. The relative similarity in ED50 thresholds and bleaching response values across the genotypes examined herein support broad genetic and phenotypic variation sufficient for a diversity of responses against natural threats, especially if there is a lack of tradeoffs in disease resistance, fecundity, and/or growth17,33. Individuals of A. cervicornis in restoration nurseries are the resilient survivors from multiple decades of environmental stress in FCR so it is feasible that these common-gardened genotypes have similar bleaching outcomes. However, we did detect a consistent ‘loser’ genotype, UK12, between both studies. If 90% of restoration genotypes have similar thermal tolerances, as observed in this study, identifying those most susceptible may be a goal for restoration groups to potentially outplant these thermally sensitive genets to sites buffered from increasing water temperatures.
Although the ten A. cervicornis genotypes examined here have been common-gardened within an in situ nursery for a minimum of five years, fixed regional differences in coral source locations had an impact on bleaching responses, where ED50 values were higher for Upper Keys corals in comparison to Lower Keys genotypes. Most Upper Keys corals were collected from shallow patch reefs prior to entering MML’s restoration propagation pipeline in the Lower Keys (Table S1). Alternatively, Lower Keys genotypes were sourced from deeper midchannel reef sites and experience more stable temperatures annually, contributing to lower thermal thresholds in comparison to corals from shallow patch reef sites. Most thermal tolerance research attributes elevated thermotolerance in populations from shallower reef habitats that experience high thermal variability3,18,47, and the results from this study corroborate long-standing genetic effects on thermal tolerance. In contrast, Cunning et al.15 did not find any correlations between original source colony environmental parameters with nursery ED50 values, despite the six nurseries covering 300km, but this could be attributed to satellite-derived SST information instead of capturing microhabitat variation at the in situ level18.
Patterns of fixed genetic effects of the source population also varied by experiment type and trait measured36. The LT-OW resolved coral source regional differences across all traits whereas CBASS indicated population-level differences for the relative change in Fv/Fm only and in the opposite pattern as LT-OW. For example, Lower Keys sourced corals had lower Fv/Fm values than Upper Keys corals under CBASS, but greater amounts of chlorophyll than Upper Keys corals under LT. Lower Keys corals are likely adapted to gradual changes in environmental stress and their ED50 suggests this population was more affected during acute heat stress but withstood the gradual chronic heat stress during LT exposures, resulting in chlorophyll values that were greater than Upper Keys corals. Within the Lower Keys population, genotype LK50 had one of the lowest Fv/Fm under CBASS, but one of the highest amounts of chlorophyll under LT heat. Alternatively, Upper Keys-sourced corals could be considered adapted to the daily variability in their shallow environment15 and therefore would be less affected by the rapid heating rate under CBASS treatments, resulting in higher ED50 values than Lower Keys ED50s. For example, genotype UK12, which had the highest ED50 thermal threshold of 34.72°C, had lower chlorophyll and host protein values in comparison to Lower Keys corals. Although sourced from shallow, variable reef habitats, it is possible that latitudinal differences in summer mean maxima could influence performance under chronic thermal stress and impair long-term survival after outplanting at restoration sites (but see48). Karp et al. 2017 also found latitudinal source effects on thermal tolerance in Miami-Dade genotypes of A. cervicornis under a moderate thermal stress experiment (three weeks at 32.5°C), although there was no relationship with source depth15.
Another possibility influencing the switch in genotype/population performance would be the observed discrepancy among the two fluorometric traits measured in the CBASS experiment, which could be attributed to the biological timing of measurable bleaching responses. Genotype UK12, which had the highest ED50 thermal threshold of 34.72°C, was the lowest-ranked performer with regard to delayed bleaching traits (chlorophyll and protein). Some genotypes with higher Fv/Fm retention (measured during CBASS) had lower chlorophyll values (measured at the end of CBASS) and vice versa (Fig. S5B). Recent work using Fv/Fm values measured right after the CBASS thermal ramp profile before recovery has found thermal threshold differences in Fv/Fm that corroborate experimental and natural bleaching responses in coral populations18,36,37. In contrast, we found a negative or no relationship between individual genotype ED50 values and the relative change in chlorophyll and protein among CBASS or LT exposures (Fig. 5). Voolstra et al.36 also found no correlations between experiment types for destructive bleaching responses (chlorophyll, protein) but did detect a correlation between Fv/Fm, thus arguing for that metric as a reliable indicator of bleaching responses. A line of evidence for the disparity in Fv/Fm between this study and Voostra et al.36 could be attributed to population-level effects detected in Voolstra et al.36, where individuals from the protected reef site consistently performed better across all traits. Here, top-performing genotypes detected with Fv/Fm during CBASS were not the top performers regarding bleaching outcomes, i.e., chlorophyll. This observed ‘switch’ in performance cautions conclusions drawn from utilizing only one photobiological trait to identify thermal tolerance (Cunning et al. 2021) and suggests that measurements should be conducted in parallel. Additionally, these results also suggest that CBASS may be useful for differentiating sub-populations among regions with different environmental conditions, but less so for differentiating genotype performance from corals within a common garden nursery.
The disparity among trait responses across experiment type and coral source region raises two important points, 1) is there a tradeoff between thermal tolerance and other physiological traits, and, 2) what is the best metric to assess heat stress tolerance? Initial comparative studies directly37 and indirectly25 demonstrated that CBASS can serve as a standardized approach to assess natural variation in thermal tolerance38. In this study, initial genotype or population resistance and physiological responses post-heat stress was mostly incongruent between acute and chronic thermal stress, despite having similar trait mean values when comparing bleaching treatments across experiment types. A complementary study by Nielsen et al.45 measured multiple physiological traits – Fv/Fm, color, chlorophyll, catalase, and protein – in Acropora tenuis from five reef sites and found that Fv/Fm, color, chlorophyll, and protein were all correlated and accounted for 43% of trait variation, similar to our findings in Fig. 2. Nielsen et al. also found a strong correlation between non-invasive metric(s) of Fv/Fm and tissue color with laboratory-derived metrics, whereas we did not find a strong correlation between metrics within CBASS (Fig. S5) and between the experiment types (Fig. 5). This difference observed in this study is likely a result of the timing of measurements made, where Nielsen et al. sampled Fv/Fm at the same time as all laboratory-derived metrics whereas we sampled Fv/Fm at 0 h of recovery and laboratory-derived metrics at 12 h recovery similar to Voolstra et al. 2021?. Thus, sampling traits at different time points provides contrasting snapshots of the underlying bleaching response in an individual coral, which could erroneously conclude that tradeoffs exist.
In Nielsen et al., each bleaching trait varied in its rate of decline under CBASS, signifying that integrative responses are occurring within or between different holobiont partners at different rates and thus challenge final bleaching response interpretations. Therefore, the particular trait measured, in addition to the timing of sampling, can confound CBASS45 and LT interpretations25,37. This study corroborates that the amount of decline in phenotypic traits can vary within45 and between25,37 heat stress exposure types. For example, Fv/Fm was not related to total chlorophyll under CBASS in this study and Fv/Fm was relatively stable throughout the Nielsen et al. study, up until 24hr post heat ramp, suggesting that Fv/Fm as a bleaching response may not be as informative immediately after acute heat application in comparison to other higher-order traits such as tissue colour45. In contrast, other studies using stress tolerant corals such as Stylophora pistillata from protected sites in the Red Sea36 and Siderastrea sidera from nearshore reefs in Belize49 found that changes in Fv/Fm values correspond strongly to other bleaching traits. Our study utilized a relatively thermally sensitive coral species maintained in a sandy bottom nursery. Despite different ocean basins and inherent thermal tolerance differences, this suggests that some bleaching trait metrics may not be as informative under a CBASS setting and may require chronic levels of stress for final bleaching outcomes. For example, host soluble protein concentrations are typically slower to respond, attributed to precursory molecular and physiological processes prior to measurable changes46 and does not display similar responses to acute heat stress like Fv/Fm and chlorophyll (Fig. S2). Combining CBASS and LT studies with fine-scale temporal sampling45 and molecular analyses could help explain differences in physiological responses under different thermal exposures. Together, there are definite nuances in measuring and interpreting bleaching responses under CBASS and continued exploration of the methodology is needed to validate and resolve acute heat stress responses.
The increased frequency and severity of global thermal stress is threatening coral survival and ecosystem functioning, spurring the urgency to rapidly and accurately quantify coral thermal tolerance and identify tolerant individuals, populations, and species for conservation and restoration efforts. With an increased interest in high-throughput heat stress assays like the CBASS, it is prudent to conduct comparative studies between different length experimental thermal exposures to effectively corroborate natural bleaching responses and quantitatively compare the upper thermal limits of coral species, populations, and individuals25,37. Acute heat stress assays quickly screen multiple individuals to determine relative thermal tolerances and have resolved tolerance differences in many coral populations4,15,18,23,26,30,36,37,40,50. We assessed the applicability of using the acute CBASS to determine whether bleaching responses are similar in restoration nursery genotypes when compared to long-term thermal exposures. Our findings conclude that although CBASS can achieve bleaching responses more rapidly (time scalability), there were differences observed in bleaching traits between CBASS and LT experiments, some of which can be attributed to common-garden and/or fixed effects on nursery A. cervicornis and the timing of measurements. Among similarly ranked genotypes we did detect one poor performer (CU12), yet initial rankings from Fv/Fm and ED50 threshold values measured right after acute heat were not comparable to endpoint bleaching response rankings (Fig S6). Despite being common-gardened for > 5 yrs, our results suggest nursery corals sourced from the Lower Keys may be more adapted to chronic heat than Upper Keys corals while Upper Keys corals appeared to be initially less affected by acute heat. To better assess variation in heat tolerance and heat tolerant traits among restoration corals, we recommend comparing nursery with wild/outplanted genotypes as well as incorporating additional Fv/Fm measurement timepoints that match other sampling measures for cross-comparison of bleaching phenotypes and detection of meaningful bleaching traits. Moreover, coupling physiological and molecular sampling would further our understanding of the timing of biological responses to make the most meaningful conclusions51, while correlating CBASS results to natural bleaching responses in restored populations can corroborate the utility of acute stress assays in a restoration context. Importantly, incorporating rapid, scalable non-invasive bleaching metrics38,45,52 to predict bleaching susceptibility alongside restoration interventions could improve restoration outcomes in lieu of declining coral cover in a warming world.