Differential survival of nursery‐reared Acropora cervicornis outplants along the Florida reef tract

In recent decades, the Florida reef tract has lost over 95% of its coral cover. Although isolated coral assemblages persist, coral restoration programs are attempting to recover local coral populations. Listed as threatened under the Endangered Species Act, Acropora cervicornis is the most widely targeted coral species for restoration in Florida. Yet strategies are still maturing to enhance the survival of nursery‐reared outplants of A. cervicornis colonies on natural reefs. This study examined the survival of 22,634 A. cervicornis colonies raised in nurseries along the Florida reef tract and outplanted to six reef habitats in seven geographical subregions between 2012 and 2018. A Cox proportional hazards regression was used within a Bayesian framework to examine the effects of seven variables: (1) coral‐colony size at outplanting, (2) coral‐colony attachment method, (3) genotypic diversity of outplanted A. cervicornis clusters, (4) reef habitat, (5) geographical subregion, (6) latitude, and (7) the year of monitoring. The best models included coral‐colony size at outplanting, reef habitat, geographical subregion, and the year of monitoring. Survival was highest when colonies were larger than 15 cm (total linear extension), when outplanted to back‐reef and fore‐reef habitats, and when outplanted in Biscayne Bay and Broward–Miami subregions, in the higher latitudes of the Florida reef tract. This study points to several variables that influence the survival of outplanted A. cervicornis colonies and highlights a need to refine restoration strategies to help restore their population along the Florida reef tract.


Introduction
Over the last four decades, thermal-stress events and disease have caused rapid declines in coral populations worldwide (Edwards & Gomez 2007;Hoegh-Guldberg et al. 2007;Hughes et al. 2018).
Some of the most heavily impacted regions have been the Caribbean (Aronson & Precht 2001) and Florida (Porter & Meier 1992;Toth et al. 2014;Precht et al. 2016;Walton Author contributions: RvW conceived the idea of amalgamating the datasets and wrote the first draft of the manuscript, and all authors contributed to editing subsequent drafts; EB, DSG, EAG, CL, KM, AM, EMM, SS, DL conducted fieldwork; RvW, RBB analyzed the data; RvW wrote the R code. et al. 2018). This decline included unprecedented mortality of Acropora cervicornis and Acropora palmata, two historically important reef-building coral species in the Caribbean. Historically, in Florida, A. palmata was dominant on reef crests on fore reefs, and A. cervicornis was dominant between 5 and 25 m on fore reefs and in shallower habitats on sheltered patch and back reefs (Agassiz 1885;Vaughan 1919;Goldberg 1973;Marszalek et al. 1977;Precht & Miller 2007). However, both acroporids suffered major declines because of white-band disease in the late 1970s and early 1980s (Aronson & Precht 2001;Gardner et al. 2003). More recently the coral assemblages along the Florida reef tract have become homogeneous (Burman et al. 2012) with the loss of reef-building species such as acroporids (Precht & Miller 2007). In 2006, their population declines prompted the listing of both acroporid species as threatened under the U.S. Endangered Species Act (NMFS 2006), and in 2008 as critically endangered and placed on the International Union for Conservation of Nature red list (IUCN 2020). Four decades after the initial mortality events, acroporid populations along the Florida reef tract continue to decline (Ruzicka et al. 2013;Toth et al. 2014).
To facilitate the recovery of acroporid populations, coral restoration programs have expanded along the Florida reef tract over the last 15 years (Johnson et al. 2011;Schopmeyer et al. 2012;Young et al. 2012). These restoration programs have focused particularly on restoring populations of A. cervicornis (Goergen et al. 2019;Ware et al. 2020). Although coral restoration may be a useful option to increase coral populations, strategies to optimize survival of nursery-reared outplants are still in their infancy. Previous studies have shown a positive relationship between colony size and survival of natural coral populations (Hughes et al. 1992), and several restoration studies have also shown that colonies with greater than 15 cm total linear extension (TLE) survive better than smaller colonies (Bowden-Kerby 2001; Herlan & Lirman 2008;Lirman et al. 2014;Goergen & Gilliam 2018). Yet, large colonies take longer to grow in nurseries than small colonies, and outplanting small colonies is often most practical. Attachment method could also play a role in survival. Some studies suggest that the "nail" method is the most efficient (Bruckner & Hourigan 2000;Young et al. 2012;Goergen & Gilliam 2018), the most inexpensive (Goergen & Gilliam 2018), and the most stable method for coral restoration in high-energy environments (Bruckner & Hourigan 2000;Young et al. 2012). Still, the epoxy method is convenient in moderate-to low-energy environments especially when coral colonies are large.
Coral survival is also a consequence of genotypic tolerance (Drury et al. 2017), and bet-hedging theory suggests that diverse clusters of A. cervicornis outplants should have the highest likelihood of survival (Hughes et al. 2008). Environmental differences across habitats and regions, such as differences in flow regimes and irradiance, are also likely to influence survival of coral outplants, especially since spatial differences in environmental conditions influence natural distributions of A. cervicornis (Marszalek et al. 1977;Ginsburg & Shinn 1995;Toth et al. 2018;van Woesik et al. 2020). Historically, A. cervicornis was ubiquitous in clear oligotrophic waters (Precht & Miller 2007), although recent studies suggest that turbid conditions are favorable during high-heat stress events (van Woesik & McCaffrey 2017). Therefore, nearshore reefs may become important habitats for the restoration of A. cervicornis along the Florida reef tract as the ocean temperatures continue to increase.
In order to facilitate the recovery of coral populations, restoration programs need answers to a suite of questions, which include: (1) What is the optimal size of an outplanted coral colony? (2) Which attachment method is best for outplanting nursery-reared colonies to natural reefs? and (3) Which habitats and geographical locations will show the highest survival? Here, we compile data from six different coral restoration programs throughout Florida to determine the conditions that may influence the survival of 22,634 nursery-reared outplanted A. cervicornis colonies. The colonies were outplanted to six natural reef habitats in seven geographical subregions along the Florida reef tract between 2012 and 2018. The objectives of the study were to examine the influence of seven variables on survival, which included: (1) coral-colony size at outplanting, (2) coral-colony attachment method, (3) genetic diversity of outplanted A. cervicornis clusters, (4) reef habitat, (5) geographical subregion, (6) latitude, and (7) the year of monitoring.

Methods
We collated data on the survival of 22,634 Acropora cervicornis colonies raised in nurseries along the Florida reef tract and outplanted to six natural reef habitats in seven geographical subregions between 2012 and 2018 from the following six coral restoration programs (i.e. organizations, agencies, or universities): (1) The Nature Conservancy, (2) the Mote Marine Laboratory, (3) the Florida Fish and Wildlife Conservation Commission, (4) the Coral Restoration Foundation, (5) the University of Miami, and (6) Nova Southeastern University ( Fig. 1; Table 1; extended details of methods are provided in Supplement S1). This current study examined the relationships between the survival of the A. cervicornis colony outplants and the effects of seven variables: (1) coral-colony size at outplanting, (2) coralcolony attachment method, (3) genetic diversity of outplanted A. cervicornis clusters, (4) reef habitat, (5) geographical subregion, (6) latitude, and (7) the year of monitoring.
The six different coral restoration programs identified in this study used either one of two methods for documenting the size of A. cervicornis colonies. While most of the groups reported the size of A. cervicornis colonies in terms of TLE (Johnson et al. 2011), which is the sum of the lengths of all the branches, the Coral Restoration Foundation reported maximum colony diameter. Therefore, the Coral Restoration Foundation maximum diameter measurements were aligned (using TLE = 2.95*diameter − 9.17) to the TLE size-class categories used in this study: (1) 1-15 cm TLE were classified as small colonies, (2) 16-50 cm TLE were classified as medium colonies, and (3) 51-160 cm TLE were classified as large colonies (Table 1).
All six coral restoration programs prepared the point of coral attachment by clearing the algae and sediment-bound turf with scrubbers from the specific area where corals were to be outplanted, then used either one, or both, of two attachment Restoration Ecology January 2021 methods: (1) nail and cable tie and (2) epoxy. The nail and cabletie method involved hammering a masonry nail vertically into the reef substrate and securing a fragment of A. cervicornis to the nail with cable ties (Fig. 2). The epoxy method involved adhering a colony directly to the reef substrate with a small amount of epoxy. The Nature Conservancy, the Mote Marine Laboratory, the University of Miami, and Florida Fish and Wildlife used nail and cable ties. The Coral Restoration Foundation used epoxy, and Nova Southeastern University included data on both methods (Table 1).
Coral host genotypes were characterized using four host (diploid) microsatellite markers following Baums et al. (2005). The current study tested whether the aggregation of genotypic diversity in clusters of A. cervicornis outplants influenced their survival (Drury et al. 2019; see Supplement S1 for details of clustering). The number of A. cervicornis genotypes in each cluster of outplanted coral colonies at each reef habitat were examined using three classes of genotypic diversity: (1) high (≥21 genotypes), (2) moderate (7-20 genotypes), and (3) low (<7 genotypes).
A. cervicornis colonies were outplanted by the six different coral restoration programs at six standardized reef habitats along the Florida reef tract between 2012 and 2018 (Florida Fish and Wildlife habitat shapefiles, https://myfwc.com/research/gis/ regional-projects/unified-reef-map) in <8 m water: (1) back reefs, (2) bank/shelf, (3) fore reefs, (4) lagoons (i.e. patch reefs), (5) reef crest, and (6) unknown or unidentified habitats.  In addition to being outplanted at six different reef habitats, A. cervicornis colonies were distributed across seven different geographical locations of the Florida reef tract, herein called subregions: (1) Dry Tortugas, (2) Marquesas, (3) lower Florida Keys, (4) middle Florida Keys, (5) upper Florida Keys, (6) Biscayne Bay, and (7) Broward-Miami ( Fig. 1). All outplanted A. cervicornis colonies analyzed in this study were considered shallow (<8 m) outplants. To examine a potential trend with latitude, the outplant sites were designated a latitude and longitude coordinate, and each outplant site was categorized into one of five latitudinal ranges: 24 -24.5 All six coral restoration programs monitored the survival of A. cervicornis outplants along the Florida reef tract between 2012 and 2018. All programs monitored outplants after 1 month and 1 year, whereas only some programs monitored annually for 4 years. At every monitoring interval, each outplanted colony was visually assessed to determine if it was alive or dead. An A. cervicornis colony was considered censored, which is the terminology used in the medical literature, if the colony was still alive at the last monitoring interval.

Data Analysis
A semi-parametric Cox proportional hazards regression was used within a Bayesian framework to examine the survival of A. cervicornis outplants. The technique is a rigorous model that determines the effects of different covariates on the outcome of survival; it is semi-parametric because it has the advantage of the baseline hazard taking any form whereas the covariates are linear. We were interested in determining the relative risk of mortality that may have been attributed to the following seven covariates: (1) coral-colony size at outplanting (three levels), (2) coral-colony attachment method (two levels), (3) genotypic diversity of outplanted A. cervicornis clusters (three levels), (4) reef habitat (six levels), (5) geographical subregion (seven levels), (6) latitude (five levels), and (7) the year of outplanting (six levels). The Cox proportional hazards model is represented as: where h i is the hazard at observation i at time t, h 0 is the baseline hazard (when all covariates are equal to zero), B i are the intercepts, and x i are the environmental covariates of interest. The model was used to quantify the relative risk attributed to each covariate on the likelihood of A. cervicornis survival. The analyses were run in "spBayesSurv" (Zhou et al. 2018) in R (R Core Team 2019) using noninformative priors. Multiple models were run to find the most informative model, with the highest log-pseudo marginal likelihood. Latitude was examined using a Cox proportional hazards model that was independent of subregional effects (but included colony size, habitat, and year), because of the confounding effects between latitude and subregions. The models captured the effects of multiple covariates on coral survival; however, when a covariate showed an effect we did not pool that data with other covariates. Therefore, we graphically display the response of A. cervicornis outplant survival for similar habitats and subregions for the year 2016. Although the year 2016 was one of the best years for survival, it was also the year when all six agencies were simultaneously monitoring survival. The results for other habitats, subregions, and years are presented in supplementary document. The R script files and data files can be located at https://github.com/rvanwoesik/Acropora_survival.

Results
The most optimal Cox proportional hazards model that assessed the survival of Acropora cervicornis outplants along the Florida reef tract between 2012 and 2018, with the highest log-pseudomarginal likelihood, included the variables coral-colony size at outplanting, reef habitat, geographical subregion, and the year of monitoring. The highest survival of A. cervicornis outplants was apparent for medium-sized colonies, 16-50 cm TLE, and large-sized colonies, 51-160 cm TLE ( Fig. 3; Table 2, Supplement S1, Figs. S6-S11). Small-sized colonies, between 1 and 15 cm TLE, showed lower survival than medium-and largesized colonies ( Fig. 3; Table 2). Although survival was not vastly different across habitats, the Cox proportional hazards model did indicate that back-reef and fore-reef habitats showed the highest level of survival of A. cervicornis outplants (Fig. 4), with lowest survival on the reef crest (Table 2, Supplement S1).
The results of the Cox proportional hazards model also considered survival of A. cervicornis outplants in geographical subregions independently of coral colony size, reef habitat, and year of monitoring (Table 2). These results indicated that the overall survival of A. cervicornis outplants in the different subregions (listed from the highest to the lowest likelihood of survival) occurred in the Biscayne Bay and Broward-Miami subregions; followed by the Dry Tortugas, the lower Florida Keys, and the Marquesas subregions; and then the middle and upper Florida Keys. Even when considering the variables coral-colony size, reef habitat, and year of monitoring, together with geographical subregions, Biscayne Bay, Broward-Miami, and the Dry Tortugas subregions still consistently exhibited the highest survival of A.cervicornis outplants (Fig. 5, Supplement S1, Figs. S6-S17).
The years with the highest survival for A. cervicornis outplant survival were 2012 and 2016, and the years with the lowest survival were 2018, 2017, and 2014 (Table 2). Examining survival across latitudes using a pooled Cox proportional hazards model (without including the covariate subregions, but including colony size, reef habitat, and year of monitoring) suggested that although there was considerable overlap in credible intervals, survival of A. cervicornis outplants increased with increasing latitude along the Florida reef tract (Fig. 6).
Including coral-colony attachment method and genotypic diversity of outplanted A. cervicornis clusters within the model decreased the log-pseudo marginal likelihood value, suggesting that these two predictive variables did not have a major measured effect on A. cervicornis outplant survival along the Florida reef tract between 2012 and 2018.

Discussion
This study found that five variables-namely, coral-colony size at outplanting, reef habitat, geographical subregion, latitude, and the year of monitoring-influenced the survival of nursery-reared Acropora cervicornis colonies that were outplanted to reefs along the Florida reef tract between 2012 and 2018. By contrast, coralcolony attachment method and genotypic diversity of outplanted A. cervicornis clusters did not have significant effects on outplant survival. Therefore, when identifying outplanting sites, coral restoration programs should not rely solely upon historical distributions of A. cervicornis but rather take into consideration contemporary niche space (van Woesik et al. 2020) and the suite of variables identified by this study.
Considering coral-colony size at outplanting, medium-(16-50 cm TLE) and large-sized (51-160 cm TLE) A. cervicornis outplants had higher survival rates than small-sized (<15 cm TLE) outplants. It is uncertain whether clipping the small colonies from large nursery-reared colonies created any further disadvantage to the outplanted fragments by limiting resources. Small Acropora colonies, however, are known to have generally lower survival than large coral colonies of the same species (Hughes et al. 1992), in part because they are more vulnerable than larger colonies to disturbances, such as predation by fireworms (Hermodice carunculata; Goergen et al. 2019) or by gastropods (Coralliophila abbreviata; Goergen & Gilliam 2018), abrasion by gorgonians (Sebens & Miles 1988), the presence of macroalgae , and by sedimentation (De Marchis 2017). These disturbances can cause partial coral colony mortality, which have disproportionate consequences on small colonies, that can lead to total colony mortality (van Woesik & Jordán-Garza 2011). Indeed, small colonies are generally disadvantaged on coral reefs (Hughes et al. 1992) except during thermal-stress events when small colonies have an advantage because of comparatively high rates of mass transfer (Patterson 1992;Loya et al. 2001;Nakamura & van Woesik 2001). The current study also showed that survival of A. cervicornis colonies outplanted to back-reef and fore-reef habitats was higher than survival of colonies outplanted elsewhere. A. cervicornis outplants had the lowest survival on highly exposed reef crests, which is not surprising because historically A. cervicornis has not been a reef-crest species (Precht & Miller 2007). Coral reef habitats vary in a variety of features, most characteristically differing in flow rates and irradiance (Done 1983). Average flow rates affect rates of mass transfer of metabolites and gases that directly influence coral physiology and survival (Patterson 1992;van Woesik et al. 2012). D'Antonio et al. (2016) showed that contemporary colonies of A. cervicornis along the Florida reef tract were most commonly found close to reef edges, where water-flow rates were high. In addition, van Woesik et al. (2020) showed that moderate wave energy, between 0.5 and 1.5 kJ/m 2 , and moderate turbidity, between 0.15 and 0.25 kd490 (per m), were the best predictors of site occupancy of A. cervicornis along the Florida reef tract. Physiological experiments have also shown that Acropora colonies are particularly intolerant to stagnant waters (Nakamura & van Woesik 2001). Therefore, outplanting A. cervicornis colonies into low-flow habitats, with low rates of mass transfer, is likely to reduce survival, whereas outplanting them into moderate-flow environments is likely to increase their survival. Table 2. Posterior inference of regression coefficients of the Cox proportional hazards model, using large-sized Acropora cervicornis colony outplants (51-160 cm TLE) inputting the data from six coral restoration programs from 2012-2018 along the Florida reef tract. TLE refers to total linear extension (Johnson et al. 2011), which is the sum of the lengths (cm) of all the branches. The six coral restoration programs included: (1) The Nature Conservancy, (2) the Mote Marine Laboratory, (3) the Florida Fish and Wildlife Conservation Commission, (4) the Coral Restoration Foundation, (5) the University of Miami, and (6) Nova Southeastern University. The 23 records from 2013 were removed for this analysis. The bases for the models were: large colonies, back reef, Biscayne Bay, and the year 2012.   The present study identified Biscayne Bay, Broward-Miami, the Dry Tortugas, then the lower Florida Keys subregions as geographical subregions where the likelihood of survival of A. cervicornis outplants is highest. These results agree with recent niche models that show the highest probability of occurrence of colonies of A. cervicornis are in the Dry Tortugas, the lower Florida Keys, Biscayne Bay, and Broward-Miami, although the niche model also showed a high probability of occurrence in the upper Florida Keys; and that the middle Florida Keys is less likely to support A. cervicornis (van Woesik et al. 2020). The present study also found a low likelihood of survival in the middle Florida Keys. Previously, Ginsburg and Shinn (1995) reported on the negative influence of Florida Bay on the middle Florida Keys, and Toth et al. (2018) showed geological evidence that the negative influences of Florida Bay terminated reef accretion in the middle Florida Keys considerably earlier than elsewhere. We suspect, therefore, that Florida Bay may continue to have negative influences on A. cervicornis restoration efforts in the middle Florida Keys.
We also found an increase in survival with increasing latitude. The conditions that change with increasing latitude, such as lower maximum sea-surface temperatures or higher nearshore turbidity along the northern Florida reef tract, could moderate the effects of thermal-stress events (van Woesik & McCaffrey 2017). Yet, there are some potentially confounding effects associated with latitude in this study. For example, the six different coral restoration programs work in different subregions, except for the Coral Restoration Foundation, which works in both the upper and lower Florida Keys. Therefore, the latitudinal and subregional effects could be a consequence of some other latent effects that were not quantified here, such as the conditions in the nurseries from which the corals originated, impacts related to the transportation of corals for outplanting, or different suites of genotypes. Indeed, coral physiology may be influenced by nursery conditions, such as water quality, temperature, light, or currents, and survival may be partially dependent on the coupling between the nursery conditions and the reef conditions.
Predictions of coral survival therefore may benefit from more information from nursery sites, such as light dynamics, waterflow rates, and diseases.
This study identified the year of outplanting as a significant influence on the survival of A. cervicornis colony outplants along the Florida reef tract. It is not necessarily the years themselves that offer any predictive significance, but rather the conditions during each year that influence survival. For example, the worst years for survival of A. cervicornis outplants were 2018, 2017, and 2014. In September 2017, Hurricane Irma severely affected the study area (personal observations), particularly the Florida Keys, altering the physical structure of many reefs and increasing sedimentation and turbidity. The survival of A. cervicornis coral outplants through Hurricane Irma was higher on patch reefs than on fore reefs (Lohr et al. 2020), which agrees with the impacts of hurricanes to Florida reefs from past studies (Lirman & Fong 1996). In addition, high thermal-stress conditions were associated with the 2014-2017 El-Niño conditions with back-to-back bleaching events occurring in 2014 and 2015 (Manzello 2015;Drury et al. 2017). The summer and winter sea-surface temperatures of 2014 were the highest on record, resulting in a coral-bleaching event throughout the Florida Keys (Manzello 2015). Although Drury et al. (2017) showed high thermal stress and coral bleaching in 2015 in the Florida Keys, our results show higher coral survival in 2015 than in 2014.
The inclusion of the two variables, coral-colony attachment method (either nails and cable-ties or epoxy) and genotypic diversity of outplanted A. cervicornis clusters, reduced the model's overall strength. In other words, these two variables did not significantly influence the survivorship of outplanted A. cervicornis colonies. However, the effects of genotypic diversity of outplanted A. cervicornis clusters on their survival are complex. In theory, and over the long term, genotypically diverse clusters of A. cervicornis outplants would be more likely to survive stress events or disease outbreaks than less diverse clusters of outplants (Hughes et al. 2008;Vollmer & Kline 2008). However, survival may be less dependent on the number of genotypes present than on the types of genotypes present and their tolerance to environmental stress (Baums et al. 2010;Drury et al. 2017Drury et al. , 2019. Moreover, our study lacked consistent methodological data on the nature of coral outplant clusters at restoration sites. Ensuring consistency in future studies could contribute toward a better understanding of coral outplant survival. Coral restoration programs therefore need to develop a coordinated effort to record clusters and investigate the role that different genotypes of outplanted A. cervicornis colonies have on population restoration efforts along the Florida reef tract (Drury et al. 2017). Key to those studies is a need to relate epigenetic and genetic profiles with phenotypic responses through environmental-stress events (Johnson et al. 2011;van Oppen et al. 2015;Muller et al. 2018).
Although the present study found both significant trends and differences between the survival of nursery-reared A. cervicornis colonies outplanted by six coral restoration programs, it also clearly indicated a need for standardized monitoring. This study used data on the survival of individually tagged A. cervicornis colony outplants to ensure that each colony was reidentified in the field. This approach, however, may have resulted in an underestimation of outplant survival and overall outplant biomass, either because some outplants may have been dislodged from their holdfasts (Goergen & Gilliam 2018) or because they may have become fragmented over time and survived at some distance from the original outplant locality. Most of the coral restoration programs did not record the dislodged outplants because they were difficult to distinguish from natural fragments. Standardized monitoring that goes beyond tracking individual fragments would greatly benefit coral restoration efforts in the future.
We acknowledge the complexity of understanding coral-outplant survival, especially as the discipline of coral restoration is still in its infancy. Just as importantly, we acknowledge the complexity of choices and decisions taken when analyzing such a complex dataset such as ours through the "garden of forking paths" (Gelman & Loken 2014). For example, a more geographically focused approach may have led to stronger inference, but we would have lost valuable insight on differential survival across the region. In addition, the seven predictive variables examined in the present study are not exhaustive, and future studies may ignore some of the variables, such as attachment method, and target others, such as genotype. We used noninformative priors throughout the analysis because of the sparseness of prior data in this newly emerging field of coral restoration, whereas weakly informative regularizing priors may have provided stronger inference (Banner et al. 2020). Indeed, eliciting informative priors is a highly recommended analytical approach for future coral restoration studies. While our results provide broad insight on the survival of coral outplants (and our analytical approach has been annotated for reproducibility), it is recommended that future studies build on this work by using more extensive and standardized field data and by applying other analytical approaches (Gabry et al. 2019) that may further optimize restoration efforts and enhance populations of endangered coral species.
In conclusion, the present study identified that five variables, namely coral-colony size at outplanting, reef habitat, geographical subregion, latitude, and the year of monitoring, all influenced the survival of nursery-reared A. cervicornis outplants. However, coral-colony attachment method and genotypic diversity of outplanted A. cervicornis clusters did not significantly influence the survivorship of A. cervicornis outplants in natural reef habitats. We recommend continued communication and coordination across all six coral restoration programs to allow for: (1) standardized monitoring, (2) the examination of effects of genotype and phenotypic expression on coral outplant survival, and (3) the determination of optimal macro-and microenvironments for restoring populations of A. cervicornis and other corals along the Florida reef tract.

Supporting Information
The following information may be found in the online version of this article: Supplementary document. Supplement S1. Extended methods. Table S1. The number of coral nursery outplants per year, of the 22,634 outplanted Acropora cervicornis colonies used in this study, from each of six coral restoration agencies along the Florida reef tract from 2012 to 2018. Figure S1. The number of outplanted Acropora cervicornis colonies along the Florida reef tract between 2012 and 2018 for each agency. Figure S2. The number of outplanted Acropora cervicornis colonies along the Florida reef tract between 2012 and 2018 for each month. Figure S3. The number of outplanted Acropora cervicornis colonies along the Florida reef tract between 2012 and 2018 for each of six reef habitats. Figure S4. The number of outplanted Acropora cervicornis colonies along the Florida reef tract between 2012 and 2018 for each of seven geographical subregions. Figure S5. The number and size of outplanted Acropora cervicornis colonies along the Florida reef tract for each agency. Figure S6. Survival of three size classes of Acropora cervicornis colony outplants for fore-reef habitats in the Broward-Miami subregion, in 2016. Figure S7. Survival of three size classes of Acropora cervicornis colony outplants for fore-reef habitats in the Biscayne subregion, in 2016. Figure S8. Survival of sizes of Acropora cervicornis outplants for fore-reef habitats in the upper Florida Keys subregion, in 2016. Figure S9. Survival of sizes of Acropora cervicornis outplants for fore-reef habitats in the middle Florida Keys subregion, in 2016. Figure S10. Survival of sizes of Acropora cervicornis outplants for fore-reef habitats in the Marquesas subregion, in 2016. Figure S11. Survival of sizes of Acropora cervicornis outplants for fore-reef habitats in the Dry Tortugas subregion, in 2016. Figure S12. Survival of Acropora cervicornis outplants in six different habitats for colonies between 16 and 50 cm in the Broward-Miami subregion, in 2016. Figure S13. Survival of Acropora cervicornis outplants in six different habitats for colonies between 16 and 50 cm in the Biscayne subregion, in 2016. Figure S14. Survival of Acropora cervicornis outplants in six different habitats for colonies between 16 and 50 cm in the upper Florida Keys subregion, in 2016. Figure S15. Survival of Acropora cervicornis outplants in six different habitats for colonies between 16 and 50 cm in the middle Florida Keys subregion, in 2016. Figure S16. Survival of Acropora cervicornis outplants in six different habitats for colonies between 16 and 50 cm in the Marquesas subregion, in 2016. Figure S17. Survival of Acropora cervicornis outplants in six different habitats for colonies between 16 and 50 cm in the Dry Tortugas subregion, in 2016.