PHENOTYPIC PLASTICITY IN ACROPORA PULCHRA UNDER VARIED ENVIRONMENTAL CONDITIONS

The extent to which coral reefs have declined globally has triggered major scientific investment in coral restoration research. However, much of the predictions for reef futures do not include the capacity for coral acclimatization, or phenotypic plasticity, and how this plasticity varies across seasons. In light of this, we outlplanted clonal replicates of distinct genotypes of the reef building coral Acropora pulchra from an existing coral nursery common garden site to three sites on the North Shore of Mo’orea, French Polynesia. After transplantation (October 2019), the outplanted colonies were sampled at all three sites in January and November of 2020, for the following physiological metrics; maximum photosynthetic rate (Am), photosynthetic efficiency (AQY), dark respiration (Rd), chlorophyll concentration, symbiont density, total protein, and ash free dry weight. Nursery genotypes and wild corals from two of the outplant sites were sampled at the outset of the experiment (October 2019) to provide a physiological baseline, which identified differences in coral physiology between the common garden and the wild corals from the two sites. Our results show that outplanted corals displayed significantly different phenotypes both through time and between sites. Our data show that plasticity score (calculated as the differences in multivariate space between October 2019 and each of the other timepoints for each genotype) was highest across all sites 3 months after transplantation (January 2020) and decreased across all sites by 13 months (November 2020) This identifies a capacity for site driven phenotypic plasticity that became more similar to the baseline common garden phenotype by 13 months, due to location acclimation, or seasonal environmental similarities. This study demonstrates that plasticity varies by site and through seasons, highlighting the need for physiological time series research to interpret performance in following human interventions

colonies were sampled at all three sites in January and November of 2020, for the following physiological metrics; maximum photosynthetic rate (Am), photosynthetic efficiency (AQY), dark respiration (Rd), chlorophyll concentration, symbiont density, total protein, and ash free dry weight. Nursery genotypes and wild corals from two of the outplant sites were sampled at the outset of the experiment (October 2019) to provide a physiological baseline, which identified differences in coral physiology between the common garden and the wild corals from the two sites. Our results show that outplanted corals displayed significantly different phenotypes both through time and between sites.
Our data show that plasticity score (calculated as the differences in multivariate space between October 2019 and each of the other timepoints for each genotype) was highest across all sites 3 months after transplantation (January 2020) and decreased across all sites by 13 months (November 2020) This identifies a capacity for site driven phenotypic plasticity that became more similar to the baseline common garden phenotype by 13 months, due to location acclimation, or seasonal environmental similarities. This study demonstrates that plasticity varies by site and through seasons,  The extent to which coral reefs have declined globally has triggered major scientific investment in coral restoration research. However, much of the predictions for reef futures do not include the capacity for coral acclimatization, or phenotypic plasticity, and how this plasticity varies across seasons. In light of this, we outlplanted clonal replicates of distinct genotypes of the reef building coral Acropora pulchra from an existing coral nursery common garden site to three sites on the North Shore of Mo'orea, French Polynesia. After transplantation (October 2019), the outplanted colonies were sampled at all three sites in January and November of 2020, for the following physiological metrics; maximum photosynthetic rate (Am), photosynthetic efficiency (AQY), dark respiration (Rd), chlorophyll concentration, symbiont density, total protein, and ash free dry weight. Nursery genotypes and wild corals from two of the outplant sites were sampled at the outset of the experiment (October 2019) to provide a physiological baseline, which identified differences in coral physiology between the common garden and the wild corals from the two sites. Our results show that outplanted corals displayed significantly different phenotypes both through time and between sites.
Our data show that plasticity score (calculated as the differences in multivariate space between October 2019 and each of the other timepoints for each genotype) was highest across all sites 3 months after transplantation (January 2020) and decreased across all sites by 13 months (November 2020) This identifies a capacity for site driven phenotypic plasticity that became more similar to the baseline common garden phenotype by 13 months, due to location acclimation, or seasonal environmental

Introduction
Coral reefs account for less than 0.1% of the global surface area, yet they account for more than 30% of all marine species (Reaka-Kudla 2005;Fisher et al. 2015) . Coral reefs are created by their ecosystem engineers, scleractinian corals. These colonial cnidarians house photosymbiotic dinoflagellates in the family Symbiodiniaceae, which supply their coral hosts with the majority of their daily metabolic requirements (Muscatine, McCloskey, and Marian 1981) through the allocation of photosynthates.
The coral in turn uses these substrates to generate energy needed to secrete a calcium carbonate skeleton (Tambutt et al. 2008) , which ultimately creates the threedimensional structure that is essential for fish and invertebrate habitat (Cesar 2002 The global degradation of coral reefs has reached a point where the natural processes alone that lead to recovery are unlikely to assure that the biodiversity and ecological functions of coral reefs will be maintained in the future (Veron et al. 2009 (Fabian, Beck, and Potts 2013;Young, Schopmeyer, and Lirman 2012) . In corals if a branch or fragment of coral is broken off from a parent colony it will give rise to a genetic clone, thus allowing parent colonies to give rise to multiple genetically identical offspring.
About half of current conservation plans to date utilize fragmentation to grow coral fragments in nurseries (Boström-Einarsson et al. 2020). Of the many species targeted for restoration purposes, approximately 30% are of the genus Acropora (Boström-Einarsson et al. 2020). Because of their high growth rates, natural use fragmentation for asexual reproduction, ability to rapidly heal wounds, comparatively high survivorship as fragments, and their endangered status in the Caribbean (Tunnicliffe 1981;Bak and Criens 1981;Gladfelter, Monahan, and Gladfelter 1978;Highsmith 1982;D. Lirman et al. 2010) this genus is a prime candidate for restoration projects.
However, despite the growing focus on reef conservation today, repeated temporal assessment of physiological plasticity following transplantation remains limited.
Most coral conservation programs and research are aimed at increasing survival and the biomass of outplants at restored sites because these are tangible results that are comparable and easy to mark success or failure. For example, (Schopmeyer et al. 2017), created a set of major benchmarks to measure the success of Acropora cervicornis restoration for six programs in Florida and Puerto Rico, which were centered on nursery and ouplant growth and survival rates. While growth and survival provide tangible benchmarks for conservation programs, these metrics do not fully assess the long-term physiological impacts of outplanted corals, or explain how they acclimate to these new environments. In fact, (Ware et al. 2020) found that while initial outplant survival remains high, after 2 years survival tends to decline. Thus, it is vital to understand the physiological capabilities of coral used in restoration and how they vary through time, to untangle the factors that can contribute to a successful or failed restoration effort.
Measuring a suite of targeted physiological metrics would provide insight to the plasticity and acclimation potential of key factors of metabolism, symbiont growth, symbiont productivity, stress tolerance and energetic content (Gardner et al. 2017;Wall et al. 2021). By integrating across biological scales in a set of genetic lines of a given species exposed to multiple environmental conditions, it is possible to gain a better understanding of the mechanisms driving phenotypic plasticity. Phenotypic plasticity is the ability for a single individual or genotype to produce a wide array of phenotypes in response to their environment. The ability to acclimate to new environmental challenges varies by genotype (Jury, Delano, and Toonen 2019;Shaw et al. 2016) and the interaction with environmental history (Wong et al. 2021;Wall et al. 2018). Safaie et al. (Safaie et al. 2018), demonstrates that corals that experience high frequency temperature variability were less likely to experience bleaching even when temperatures would rise 1 °C above daily means. Similarly, (Morikawa and Palumbi 2019) demonstrates that growing nursery fragments under more variable temperature conditions displayed higher rates of bleaching resistance than the same genotypes grown in more stable temperature conditions. Therefore, characterizing the interplay of environment and genotype and its role in shaping coral phenotype is crucial to understanding of acclimatory mechanisms and the optimization of restoration success.
Here we examined phenotypic plasticity in corals from a common garden that were outplanted to sites ranging in environmental conditions. Using an existing coral nursery in Mo'orea as a common garden setting, clonal replicates of 10 Acropora pulchra genotypes were outplanted to three different reef sites located in the back reefs on the north shore of Mo'orea, French Polynesia (Fig. 1). These three sites represent an established nutrient gradient (Adam et al. 2021). Environmental conditions were also monitored in real time by deploying sensors to monitor temperature, light, and pH. Additionally, the locally occurring macroalgal species Turbinaria ornata was collected throughout the course of this study to provide integrated environmental nutrient assessment (Adam et al. 2021). This research will provide meaningful insight to the physiological differences of nursery and wild type corals, metabolic impacts of nutrients to outplanted corals, and effectively quantify plasticity through time.

Experimental Design
Ten distinct genotypes of the reef building coral Acropora pulchra were identified within an existing coral nursery (17°29'01.1"S, 149°50'03.9''W) on the island of Mo'orea, French Polynesia at the outset of the experiment in October of 2019.
Acropora pulchra was chosen for this experiment because it is a fast growing species reef building coral (Alcala, Alcala, and Gomez 1981;Soong and Chen 2003)  In-situ sampling of each coral colony for every time point, which occurred in October 2019, January 2020, and November 2020, included taking a 2-3cm apical clipping of each Acropora pulchra colony which were transferred in quart (~946mL) ziplock bags filled with seawater from each site to the University of California Berkeley Gump research station on Mo'orea. These live fragments were immediately placed in flowthrough seawater tables where they were exposed to ambient water from Cook's Bay, Mo'orea for 2 days prior to physiological sampling. All colonies (nursery and outplant) were sampled for the following 8 physiological metrics; respiration (Rd), photosynthetic efficiency or apparent quantum yield (AQY), maximum rate of photosynthesis (Am), chlorophyll concentration, symbiont density, total protein, and ash free dry weight.

Photosynthesis:Irradiance curves
Each labeled coral was taken from the outdoor water table and placed into individual 620 ml acrylic respirometry chambers with a magnetic stir bar filled with seawater from the water tables. Chambers were sealed and placed in magnetic stir plate stands.
The chambers were fully submerged in ambient temperature seawater, which was controlled by an aquarium heater (Finnex 300W Titanium Heater, Burnaby, British Columbia, Canada) and pump (Pulaco 400GPH Submersible Pump, Baiyun District, Guangzhou, China). A temperature sensor (PreSens Pt1000) and a fibre-optic oxygen probe [Presens dipping probe (DP-PSt7-10-L2.5-ST10-YOP)] was placed in each chamber to measure temperature corrected oxygen flux in µmol l -1 . The corals were exposed to ten different irradiance levels for 10 minutes each (~0, 18, 68, 113, 169, 243, 499, 709, 844 photosynthetic efficiency or apparent quantum yield (AQY), and maximum rate of photosynthesis (Am). Once all light levels were completed, the corals were removed from the chambers, transferred to 4oz sterile plastic bags (Grainger Industrial, Lake Forest, Illinois, USA), the seawater was removed, fragments were snap-frozen in liquid nitrogen, and stored at -40°C until processing. The volume of water in each chamber was measured and used to correct for differences in coral displacement for each fragment. Rates of photosynthesis and respiration were calculated in final units of µmols O2 cm -2 h -1 .

Tissue Removal for Assays
Coral tissue from the snap frozen fragments was stripped from their skeletons using a

Surface Area Quantification
The coral skeletons were placed in a drying oven (Fisher Scientific were calculated using equations for dinoflagellates in 100% acetone described in (Jeffrey and Humphrey 1975) and corrected for path length of the 96-well plate (0.66cm; Hellma Analytics Quartz Microplate). Chlorophyll concentrations represent total chlorophyll (a and c2 values combined) standardized to sampled fragment surface area (µg cm -2 ), and endosymbiont cell density (µg cell -1 ).

Symbiont Density
To quantify the Symbiodiniaceae densities, repeated cell counts (n = 8) were conducted on the homogenate aliquots using an Improved Neubauer Haemocytometer (Marienfeld Superior, Lauda-Königshofen, Germany). The fixed volume within the defined area of the haemocytometer was used to calculate symbiont density of the sample and was normalized to the surface area of the coral skeleton (cells cm -2 ).

Ash Free Dry Weight (Biomass)
Tissue biomass was quantified as Ash Free Dry Weight (AFDW) from both the host and symbiont fractions. An aliquot of 5 mL of coral tissue homogenate was centrifuged at 3500 rpm for 3 minutes. 4 mL of the supernatant (host) was removed and transferred into a pre-burned aluminum pan. The algal pellet (symbiont) was resuspended in 1 mL of cold 1X PBS before being placed in a separate pre-burned aluminum pan, and both pans for each sample were placed in a drying oven at 80 °C for 24 h and dry weight was measured. Pans were subsequently burned in a muffle furnace at 450 °C for 4-6 h. Dry mass was calculated as dry tissue and pan massinitial pan mass. AFDW was calculated as dry tissue and pan mass -burned tissue and pan mass, normalized by homogenate volume and standardized surface area (mg cm -2 ).

Statistical analysis
To test that hypothesis that there was no difference in the physiology of the common garden (n=10) and wild coral colonies (n=3) from sites 1 and 2 (October 2019), we analysed each of the 8 physiological metrics using one-way ANOVA. Further, to test the hypothesis that each variable differed between sites (3 sites, fixed factor) and time point (January and November 2020, fixed factor), we analysed each of the aforementioned 8 physiological metrics using a two-way ANOVA. Site was used as a proxy for environmental conditions as the environmental data set did not capture data for all metrics for the entirety of the experiment. The environmental data was cleaned which resulted in n= 68 days of continuous data at all three sites for all metrics which was used in the ANOVA analyses of all metrics used in Figure 3. Repeated measures were not considered for this analysis because additional replicates of the genotypes displaying reproductive viability were outplanted for separate reproduction work but were sampled at times due to loss of colonies (unidentifiable, lost, or died) so despite sampling the same genotypes throughout the study they do not always represent the same colony. Posthoc tests (Tukey's HSD) were used to parse out significance of pairwise comparisons for both datasets.  (Fig. 4). These plasticity scores were then analyzed using a two-way ANOVA to test the significance of site and time point on multivariate plasticity, followed by posthoc tests (Tukey's HSD). The assumptions of ANOVA were tested with graphical analysis of the residuals and all variables were log transformed to meet the assumptions of normality and distribution.

Coral Energy Production and Demand (Am, AQY, Rd)
Analysis of P:I curves identified that Am was significantly different between nursery and wild sites (F2,13 = 6.779, p = 0.01 ), with the higher rates of photosynthesis at sites 1 and 2 in comparison to the nursery site (Table S1, Fig. 2). Apparent quantum yield (AQY) displayed high variation at each site and no significant differences were detected (F2,13 = 2.2249, p-value = 0.148). Dark respiration (Rd) were significantly different at each site (F2,13 = 15.402, p < 0.0004), with higher respiration rates at site 1 and site 2 in comparison to the common garden site (Table S1, Fig 2). These results show that energy production and demand are both higher at both wild sites than the nursery. Thus the environments at the wild sites and nursery are imposing different challenges that trigger differing energetic production and demand.

Symbiont Variables [Total Chl (ug cm -2 , ug cell -1 ), Symbiont Density (cells cm -2 )]
Total chlorophyll concentrations of the sampled fragments (ug cm -2 ) were significantly different between sites (F2,13 = 36.524, p = 4.622e-06) with all three sites being significantly different from one another (Table S2) Fig 2). Total chlorophyll concentrations per symbiont (ug cell -1 ) were significantly different between sites as well (F2,13 = 15.642, p < 0.0004), with site 1 having the highest concentration of chlorophyll per symbiont than both site 2 and the common garden sites (Table S1, Fig.   2). Symbiont density was marginally different between sites (F2,13 = 3.774, p-value = 0.051). These results show that each site imposes varying environmental challenges that cause chlorophyll concentrations to fluctuate by site. The differences in chlorophyll concentration, therefore, are mostly attributed to the symbiont physiology and the degree to which they are concentrated with chlorophyll by site.

Host Variables [Host Protein (ug cm -2 ), AFDW (mg cm -2 )]
Host protein concentrations were significantly different between sites (F2,13 = 17.684 , p < 0.0002), with higher concentrations of host protein at site 1 and site 2 in comparison to the common garden site (Table S1, Fig. 2). Ash free dry weight values were significantly different between sites (F2,13 = 11.51 , p = 0.001), with site 1 having significantly higher concentrations than the nursery (( Table S1, Fig. 2)). Site 1 protein concentrations are not significantly different than site 2 and site 2 protein concentrations are not significantly different than the nursery (Table S1, Fig. 2) while site 2 was not significantly different than the nursery (TukeyHSD, p-value = 0.052).
These results suggest that while the wild sites produce higher values of biomass than the nursery, the protein contribution to this biomass is not the only thing driving these differences between sites because site 1 and 2 were not both significantly higher in biomass and protein.

Transplant Timeseries Univariate Phenotype Analysis:
While 10 genotypes were outplanted, the full data set of values for all eight physiological metrics were limited to only four genotypes, due to either partial/full mortality of fragments, or not being able to locate a fragment in the field at a particular time point.
Environmental Data: From the data set we see that site 1 had the highest light intensity, was the warmest, the least variable in daily temperature and had relatively high pH but not the highest pH. Site 2 displayed moderate levels of light and was the most variable in temperature and pH of all three sites. Lastly, site 3 was characterized by having the lowest temperatures, pH values, light intensity, and was the least variable for pH and temperature. This data showed that out of the 12 metrics tested in However, respiration rates were significantly different between timepoints (F1, 18= 17.868, p = 0.000507), where the average respiration rates were higher at all three sites in January of 2020 compared to those in November of 2020 (Table S2, Fig. 3).
These results suggest that photosynthetic capacity is unchanging spatiotemporally but coral energetic demands are higher in January compared to October/November. These higher metabolic demands could be the product of warmer austral summer temperatures or potentially a product of acclimatization, which requires more energetic input to help the corals overcome the stress of transplantation.

Symbiont Variables [Total Chl (ug cm -2 , ug cell -1 ), Symbiont Density (cells cm -2 )]
Total chlorophyll concentrations per coral fragment (ug cm -2 ) were significantly different between sites (F2, 18 = 4.307, p = 0.0296), with site 3 having overall lower total chlorophyll concentrations than site 1 and significantly lower concentrations than site 2 (Table S2, Fig. 3). Total chlorophyll concentrations per fragment were also significantly different between time point (F1, 18= 8.282, p = 0.0100), with January chlorophyll concentrations at site 1 and site 2 being significantly higher than the same sites sampled in November 2020 ( values at all three sites in January 2020 compared to those in November 2020 (Table   S2, Fig. 3). These results suggest that while symbiont cells may be more abundant in corals in January, the few cells in November 2020 are much more concentrated with chlorophyll than those in January. Therefore, there is this physiological tradeoff where in the austral summer and earlier in the acclimatization period, corals increase the density of symbionts they house but overtime these cell densities revert back to common garden levels but increase the chlorophyll concentrations per symbiont.

Host Variables [Host Protein (ug cm -2 ), AFDW (mg cm -2 )]
Protein concentrations (ug cm -2 ) were significantly different at each site (F2, 18 = 12.60, p = 0.000378), with all three sites being significantly different from one another. Site 2 had the highest, site 1 had intermediate, and site 3 had the lowest protein concentrations (Table S2, Fig. 3). Protein concentrations were also significantly different between time points (F1, 18= 51.66, p = 1.09e-06), with January 2020 having higher protein concentrations at all three sites (Table S2,  Chlorophyll (ug cm -2 and ug cell -1 ) and symbiont cell densities provided significant contributions to the variation seen in the phenotypes across all sites and time points ( Fig. S1 and S2, Appendix 1).

Plasticity Scores
Here we used plasticity scores to indicate the change in the phenotypes between the baseline common garden and each site and time point. Plasticity score was not significantly different between sites (ANOVA; F2, 18, = 0.469 p = 0.633), nor was the interaction between site and time point significant (ANOVA; F2, 18, = 1.406, p = 0.271). However, the plasticity score was significant between time points (ANOVA; F1, 18, = 21.035, p < 0.0003), where January 2020 (timepoint1) plasticity scores were significantly higher than November 2020 (timepoint4) plasticity scores. These results suggest that phenotypic plasticity varied across time and were greater at the January 2020 time point as compared to the November 2020 time point (Fig. 5). These results suggest that corals exhibit a more plastic or diverse phenotypic response to outplanting to various sites 3 months after transplantation (average plasticity value = 4.25) compared to 13 months after transplantation when the plasticity scores are about half (average plasticity value = 2.0) the January 2020 values.

Discussion
As human interventions become more widely adopted to address global declines of coral reefs, it is essential to investigate all possible mechanisms that drive restoration success and failures. While increasing biomass is important to restore ecological services for the reef communities, it is important to understand how restoration techniques like coral gardening or using coral nurseries will impact outplants. For example, understanding the impact of the environment of the nursery will be instrumental to predicting the outplants' responses to future stressors and therefore dictate the restored coral community. [Morikawa and Palumbi 2019], demonstrated that temperature variability of a nursery site is a great predictor of bleaching resistance regardless of nursery genotype. In this case we saw the outplants thrived in conditions of low ambient nutrients, moderate light, variable pH and temperature whereas wild corals seemed to thrive in high nutrient conditions, indicating that corals originating from this nursery were able to thrive in the most optimal growth settings while those found on the reef who acclimated to their environments.

Initial differences between common garden and wild corals
This study shows that nursery reared corals displayed differing phenotypic profiles compared to wild colonies naturally occurring on the backreef. Specifically across all 8 metrics, wild corals showed equal or higher values. This indicates that there was an environmental effect or inherent differences of nursery reared corals which should be further explored.

Physiological plasticity occurs in space and time
Outplanted corals acclimate to a new environment by inducing a more plastic phenotypic response and then revert back to their original nursery phenotype overtime.
Phenotypic plasticity in this case serves as a short-term acclimation strategy but the long-term metabolic phenotype is dependent on the nursery environment.
These data provide insight to the biological complexities surrounding coral restoration techniques commonly used today on a commonly used restoration species (Boström-Einarsson et al. 2020). This work demonstrates the importance of nursery environments and elucidates how plasticity could be a short-term solution that drives high success in restoration programs, while the environment of the nursery these outplants are reared in can dictate the eventual metabolic phenotype. Coral gardening and other restoration programs and methods have been hugely successful at increasing overall biomass of transplant to degraded reefs, which could be the result of programs primarily focusing on fast growing species that are poor competitors ecologically.
Therefore, restoration outcomes may be successful in the short-term because of corals' ability to acclimate to varied outplant environments, but lack the longevity. (Ware et al. 2020) indicates that coral outplant survivorship declines after 2 years, which is beyond the 18-month average outplant monitoring timeline of most conservation programs currently (Boström-Einarsson et al. 2020). I propose that this decline in survival of outplant to be the result of metabolic and environmental match or mismatch, in which after the acclimation period and corals revert back to their baseline phenotypes, these phenotypes are either compatible or at odds with the environmental challenges imposed by their outplant site. Thus leading to overall mortality over time. In fact, (Barott et al. 2021) corroborated this finding where the phenotype of outplants of Montipora capitata and Pocillopora acuta were more reflective of their origin site rather than the outplant site but those that came from a less challenging environment were more successful and more fecund because of their phenotype. Another important finding from (Barott et al. 2021), is that this response was consistent across widely disparate genera on the phylogenetic tree, suggesting that this could be a highly conserved phenomena across corals but future studies would need to investigate if this response is true outside of branching corals as well.

Conclusions
Nursery Acropora pulchra colonies displayed overall depressed metabolic phenotypes compared to wild Acropora pulchra colonies found in the backreef lagoon on the north shore of Mo'orea, French Polynesia. When outplanted to three sites spread across the north shore backreef, colonies within 4 fixed genetic lines displayed a highly plastic metabolic response 3 months post-transplantation whereas 13 months post-transplantation plasticity was significantly reduced, and their phenotype was more similar to their original nursery phenotypes. Plasticity in this case is the product of short-term acclimatization to a new environment but long-term phenotype is influenced more by environmental origin which was the nursery in this case. Figures   Figure 1. A) a map of the North Shore of the island Mo'orea, French Polynesia and the three outplant sites in the backreef lagoon chosen for this study. B) Colonies of known genotypes (n = 4) were tagged and sampled prior to and after transplantation to the three sites such that 1 colony of each genotype was present at each site. C) Environmental loggers deployed at each of the sites collecting light, pH, and temperature. These loggers were set at each of three transplant sites in October of 2019 and were offloaded and monitored throughout the experiment (Oct. 2019 -Nov. 2020).       Figure 1 by multiple 8" zipties. Each temperature logger was wrapped in white electrical tape to reduce marine biofouling on the actual sensor.
pH Logger HOBO MX pH Logger were manually calibrated using Bluetooth capabilities of the sensor and the HOBOmobile App on a smartphone. Sensor was calibrated by, first activating the bluetooth pairing between the smart device and the HOBO MX pH logger. The sync was completed by opening the HOBOmobile app. The correct device was selected by hitting the Devices icon and tapping the individual logger that appeared for the sync. Immediately the user was asked if they wanted to calibrate the sensor or the calibrate button is located at the bottom right of the screen. "Yes" was selected and then the two-point pH calibration option (using pH 7.00 and 4.01 standards) was selected. Using a squirt bottle the pH sensor was rinsed with deionized or distilled water prior to calibration and dried with a paper towel. The clear storage cap was then filled with Metrohm NBS buffer 7.00 before submerging the sensor in the solution, and secured onto the sensor by screwing on the cap. Once the top was secured, the 7.00 calibration was selected for the logger to take a reading. The reading then stabilized, and the 'Confirm Buffer' option was selected. This saved the calibration before calibrating for the 4.01 standard. Sensor was again rinsed and dried and the calibration process was repeated before finally selecting 'Save Calibration'.
Sensor was rinsed one last time with deionized or distilled water and then deployed. If deployment was not done immediately, logger was placed in pH electrode storage solution. During the March to September deployment, the sensors at sites 1 and 2 were damaged and were rendered unusable after (glass electrode was shattered).

Light Intensity
Odyssey Xtreem PAR Loggers were calibrated and deployed to capture light intensity at each site. Calibration required a two-step process requiring an Android device to connect to the logger via the Bluetooth capabilities of the logger via the Xtract App and then accessing the data through www.xpert.nz. Each logger was battery operated, so the loggers were filled with new AA batteries to activate the logger and connected to it using the Xtract App. The calibration was conducted on the Xpert website through the "Configure" tab. Here each logger, identified by its serial number, was labeled with a new name. Then the "Reset" option was selected on the Xpert website under the "Configure" tab, followed by selecting the "Submit" button. These loggers were then set in a bucket of seawater under a Prime™ 16HD Reef Aquarium Light set to a known light intensity alongside a LiCor to get discrete values alongside the loggers for 5-6 hours. The data was then off-loaded from the logger by re-opening the Xtract App and connecting to the loggers one at a time (connectivity based on proximity). The data was exported via .csv files sent to the Xpert account holder's email by selecting the "Send Report'' button under the "Data" tab. Using the "Calibration" tab in Xpert, select the "Test Duration'' box and input the time of the calibration test for the loggers. The known light intensity value was then input to the "Reference Reading Average" box on the "Calibration" page and the submit option was selected. The calibrations were applied and the chart in Xpert was adjusted to the calibration. The loggers were recalibrated after each deployment period for the next.
The logger at Site 1 unfortunately stopped working during the March -September deployment (probably due to water seepage frying the circuit) and light was not obtained for this site past March.
Table S1. One way ANOVA (site) and post-hoc results for the 8 univariate metrics sampled on the nursery genotypes (n=10) and the wild type corals (n=3) at site 1 and 2 in October of 2019. Table S2. Two-way ANOVA (site*timepoint) and post-hoc results for the 8 univariate metrics sampled on the outplanted genotypes at all three sites in January and November 2020 time points.