Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Phenotypic plasticity and its ecological and evolutionary significance for reef building coral
(USC Thesis Other)
Phenotypic plasticity and its ecological and evolutionary significance for reef building coral
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
i
Phenotypic plasticity and its ecological and evolutionary significance for reef building coral
by
Wyatt C Million
A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(MARINE BIOLOGY AND BIOLOGICAL OCEANOGRAPHY)
August 2022
ii
Acknowledgements
I thank my advisor Carly who has inspired me to pursue my curiosities and has provided an
example of how to endure hardship with respect and compassion for myself and others.
Thank you to my committee for your direction, support, and example of what
dedicated and impassioned scientists can accomplish.
Thanks to my fellow lab mates for fruitful discussions, exciting efforts to troubleshoot molecular
and bioinformatic workflows, and memorable experiences in the field. Special thanks to Maria
who proved to be the best friend, lab mate, and dive partner I could have ever had. Thanks to my
MEB peers for creating a supportive and fun environment in which it has been a pleasure to
participate.
I’d like to acknowledge the group of scientists whom I worked with at Mote Marine Lab who
were vital to my success by leading and aiding in strenuous data collection in the field.
I must also acknowledge the wealth of support from my parents who fostered my passion for
science and the environment. Their unconditional encouragement during my career allowed me
to take the chances that have allowed me to arrive at this point. I am privileged to have such a
network of support in my life.
I found my escape from the stress of school and the city in the mountain trails surrounding Los
Angeles. In particular, the Backbone Trail, Kenyon Devore Trail, and Pacific Crest Trail
reminded me that life was beautiful and bigger than the problems I faced at a particular time.
Trail running and the community that came with it provided me with the crucial balance and
perspective that I needed to get through the lows and to celebrate the highs.
iii
Table of Contents
Acknowledgements ii
List of Tables vii
List of Figures viii
Abstract xi
Chapter 1 Introduction 1
1.1 Phenotypic plasticity in the context of reaction norms 1
1.2 Phenotypic plasticity’s influence on trait evolution 2
1.3 Evolution of plasticity 4
1.4 Conditions influencing the evolution of plasticity 7
1.5 Plasticity in coral 8
1.6 Acropora cervicornis as a system to study plasticity 10
1.7 Goals of this dissertation 11
Chapter 2 Development of 3D photogrammetry method to measure individual coral colony
morphology 13
Summary of Contribution 13
2.1 Abstract 13
2.2 Introduction 15
2.3 Methods 19
2.3.1 Study Design 19
2.3.2 Image Capture 20
2.3.3 Image Pre-processing and Model Building 20
2.3.4 3D Phenotyping 21
2.3.5 Statistical Analysis 22
2.4 Results 23
2.4.1 Modeling Performance and Precision 23
2.4.2 Photogrammetry vs. Traditional Methods 24
2.4.3 TLE as a Proxy for Higher Order Traits 25
2.4.4 Relationships Among Growth Traits Over Time 26
2.5 Discussion 29
2.5.1 Relationships Between TLE and Higher Order Traits 30
iv
2.5.2 Predictive Power of Initial Growth 32
2.5.3 Advantages of 3D Photogrammetry 34
2.5.4 Broader Applications 36
2.6 Supplementary Material 39
Chapter 3 Evidence of adaptive morphological plasticity in the endangered coral, Acropora
cervicornis. 46
Summary of contribution 46
3.1 Abstract 46
3.2 Significance Statement 47
3.3 Introduction 48
3.4 Results 53
3.4.1 Ramet survival is a function of genotype, outplant site, and the interaction 53
3.4.2 Non-random fragmentation in A. cervicornis 54
3.4.3 Morphology exhibits plasticity that varies by genotype (GxE) 54
3.4.4 Growth rate is dependent on genotypic and environmental characteristics 55
3.4.5 The capacity for morphological plasticity is correlated with improved survival and
growth 57
3.4.6 Variation among offshore reefs may contribute to site-specific growth and survival 58
3.5 Discussion 60
3.5.1 Adaptive phenotypic plasticity in coral 61
3.5.2 Genotype-environment interactions limit survival and growth predictions 62
3.5.4 Multivariate environmental conditions distinguish reefs 66
3.5.5 Conclusions 68
3.6 Methods 68
3.6.1 Experimental design 68
3.6.2 Phenotyping 69
3.6.3 Environmental data 70
3.6.4 Statistical analysis 70
3.7 Supplementary Material 72
3.7.1 Supplemental Methods 72
Phenotyping 73
v
Environmental data 74
Statistical analysis 74
3.7.2 Supplemental Figures and Tables 78
Chapter 4 Plasticity and GxE in transcriptomic response in Acropora cervicornis across
environmentally variable reefs. 101
Summary of Contribution 101
4.1 Abstract 101
4.2 Introduction 102
4.3 Methods 106
4.3.1 Experimental Design 106
4.3.3 Genet and reef site characterization 108
4.3.2 Bioinformatic analyses 109
4.3.3 Gene and gene network expression analyses 111
4.4 Results 113
4.4.1 Transcriptome-wide profiles 113
4.4.2 Genet, site, and GxE contributions to gene expression 114
4.4.3 Gene network expression 118
4.5 Discussion 121
4.5.1 Genet-level differences in expression persist after outplanting 121
4.5.2 Abundant gene expression plasticity may contribute to coral acclimatization 124
4.5.3 Contributions of intraspecific variation in expression plasticity in higher order traits
129
4.5.4 Conclusions 131
4.6 Supplementary material 132
Chapter 5 Intraspecific variation in physiological plasticity in Acropora cervicornis under
thermal and acidification stress 149
Summary of Contribution 149
5.2 Introduction 150
5.3 Methods 155
5.3.1 Experimental Design 155
5.3.2 Physiological phenotyping 156
5.3.3 Statistical analysis 157
vi
5.4 Results 158
5.4.1 Holobiont productivity 158
5.4.2 Host protein modulation 159
5.4.3 Symbiont physiology 160
5.4.4 Relationship of physiology plasticity across traits and treatments 161
5.5 Discussion 163
5.5.1 Population-level plasticity to acidification, temperature and combined stress 163
5.5.2 Intraspecific variation in physiology 166
5.5.3 Conclusions 169
5.6 Supplementary Material 170
References 171
vii
List of Tables
Table S3.1:Outplant sites and SERC water quality monitoring station coordinates
Table S3.2: Results of Cox proportional hazard ratio models for genets
Table S3.3: Results of Cox proportional hazard ratio models for sites
Table S3.4: Results of cumulative linked model for effects of genet and site on fragmentation
Table S3.5: Summary of results for Fisher Exact Test of size effects on fragmentation
Table S3.6: Summary of fixed effects for absolute size in TLE, SA, V and Vinter
Table S3.7: Summary of random effects on absolute size in TLE, SA, V, and Vinter
Table S3.8: Summary of linear fixed effects for growth rate in TLE, SA, V and Vinter
Table S3.9: Summary of random effects on growth rate in TLE, SA, V, and Vinter
Table S3.10: Summary of thermal characteristics for experimental reef sites
Table S3.11: ANOVA results for SERC water quality differences between sites
viii
List of Figures
Figure 1.1: Hypothetical reaction norm displaying variation in phenotypic plasticity
Figure 2.1: Relationship between by-hand and 3D photogrammetry measurements of TLE
Figure 2.2: Correlations between growth morphological traits over time
Figure 2.3: Non-linear relationship between TLE and volume of interstitial space
Figure 2.4: Relationship between early and late growth of outplanted A. cervicornis colonies
Figure S2.1: Image of T0 3D photogrammetry image capture set up at the in situ nursery
Figure S2.2: Convex hull overlaid 3D coral model in MeshLab
Figure S2.3: Example of nonlethal fragmentation in outplanted colony
Figure S2.4: Relationship between number of photos taken and size of coral colony
Figure S2.5: Capability of TLE to predict growth in surface area, volume, and volume of
interstitial space
Figure S2.6: Capability of early growth to predict later growth in each morphological trait
Figure S2.7: Impact of breakage on predictive capability of early volumetric growth for later
growth
Figure S2.8: Images of a single colony over time and an example its high quality 3D model
Figure S2.9: Example of ability of protocol to build 3D models for highly complex structures
Figure 3.1: Experimental design and environmental conditions for in situ experiment
Figure 3.2: Genotype, environment, and GxE patterns of survival
Figure 3.3: Average size in TLE for each genotype over experimental period
Figure 3.4: Morphological plasticity and its relationship to growth and survival
Figure S3.1: Genet rank correlations across reef sites ordered by geographical location
Figure S3.2: Total number of observed fragmentation events across genets and sites
Figure S3.3: Genet surface area growth curves at each site over the experimental period
Figure S3.4: Genet volume growth curves at each site over the experimental period
Figure S3.5: Genet volume of interstitial space growth curves at each site over the experimental
period
Figure S3.6: Relationship between growth rate in TLE and colony size
Figure S3.7: Relationship between growth rate in TLE standardized by biomass and colony size
Figure S3.8: Correlations between genet plasticity in each trait and genet risk score
Figure S3.9: Correlation between genet plasticity and mean growth rate in each trait
Figure S3.10: Correlation between genet mean growth rate in each trait and genet risk score
Figure S3.11: PCA of invariant morphology colored by outplant location
Figure S3.12: Hourly temperature profiles for the nine experimental reef sites
Figure S3.13: PCA of contemporary SERC water quality metrics and experimental thermal
characteristics
Figure S3.14: PCA of historical SERC water quality metrics and experimental thermal
characteristics
Figure S3.15: Bayesian model results for the effect of water quality on coral TLE
ix
Figure S3.16: Bayesian model results for the effect of water quality on coral surface area
Figure S3.17: Bayesian model results for the effect of water quality on coral volume
Figure S3.18: Bayesian model results for the effect of water quality on coral volume of
interstitial space
Figure S3.19: Bayesian model results for the effect of water quality on coral risk score
Figure 4.1: Principal component analysis of high expression genes at T0 and T12
Figure 4.2: Site-specific DEGs clustered reaction norms across outplant reef sites
Figure 4.3: Genet-specific DEGs cluster patterns across A. cervicornis genets
Figure 4.4: WGCNA module-trait heatmaps for genet- and site-specific parameters
Figure S4.1: Principal component analysis of combined T0 and T12 datasets
Figure S4.2: Distribution of genet-, site-, and GxE-specific DEGs and enrich GO terms
Figure S4.3: All site-specific DEG cluster reaction norms
Figure S4.4: Patterns of all genet-specific DEG clusters across experimental genets
Figure S4.5: Enriched GO terms in GxE-specific DEGs
Figure S4.6: Full genet and site associated WGCNA module-trait heatmap
Figure S4.7: GxE WGCNA module-trait heatmap
Figure S4.8: Module membership vs gene significance relationship for modules associated with
total number of breaks
Figure S4.9: Module membership vs gene significance relationship for modules associated with
TLE growth
Figure S4.10: Module membership vs gene significance relationship for modules associated with
SA growth
Figure S4.11:Module membership vs gene significance relationship for modules associated with
V growth
Figure S4.12: Module membership vs gene significance relationship for modules associated with
Vinter growth
Figure S4.13: Module membership vs gene significance relationship for modules associated with
SA-to-V ratio
Figure S4.14:Module membership vs gene significance relationship for modules associated with
sphericity
Figure S4.15: Module membership vs gene significance relationship for modules associated with
convexity
Figure S4.16: Module membership vs gene significance relationship for modules associated with
SA-to-TLE ratio
Figure S4.17: Module membership vs gene significance relationship for modules associated with
genet-specific risk of mortality
Figure S4.18: Module membership vs gene significance relationship for modules associated with
site-specific risk of mortality
Figure S4.19: Module membership vs gene significance relationship for modules associated with
longitude
x
Figure S4.20: Module membership vs gene significance relationship for modules associated with
summer thermal predictability
Figure S4.21: Module membership vs gene significance relationship for modules associated with
average daily temperature range
Figure S4.22: Module membership vs gene significance relationship for modules associated with
MMM
Figure S4.23: Salmon gene heatmap across samples order by genet-specific risk of mortality
Figure S4.24: Light Steel Blue1 gene heatmap across samples order by genet-specific risk of
mortality
Figure 5.1: Genotypic reaction norms for holobiont physiological traits
Figure 5.2: Genotypic reaction norms for host physiological traits
Figure 5.3: Genotypic reaction norms for symbiont physiological traits
Figure 5.4: Correlation between physiological plasticity in traits across experimental treatment
Figure S5.1: Correlations of physiological plasticity across traits and experimental treatment
xi
Abstract
The potential of phenotypic plasticity to enhance or limit survival, accelerate or hamper
evolution, and to even generate novel phenotypic variation make this trait a crucial part of the
ecology and evolution of species. Despite these eco-evolutionary implications, our understanding
of plasticity’s role in evolution and its potential to evolve has been limited to theory and sparse
empirical evidence from choice model systems. While not traditionally considered a model for
studying phenotypic plasticity, the sessile and long-lived nature of reef building coral suggest
that plasticity, rather than avoidance or migration, may be the predominant strategy to deal with
environmental change. Given the threatened or endangered status of coral around the world,
understanding the adaptive role of plasticity will be critical to predicting if and how these species
will respond to climate change. This dissertation investigates the abundance of intraspecific
variation in morphological, physiological, and molecular plasticity in the critically endangered
Caribbean coral, Acropora cervicornis. To quantify plasticity’s role on fitness and provide
insight into its adaptive potential, I utilized a combination of in situ and lab-based
experimentation to induce plasticity in A. cervicornis genotypes across a range of conditions and
compared an individual’s potential to be plastic with its overall fitness. I uncover novel
intraspecific variation in morphological plasticity and add to previous evidence of variation in
physiological and molecular plasticity. While morphological plasticity appears to be adaptive,
physiological plasticity showed no impact on coral fitness. Similarly, both fixed and plastic
variation in gene expression seem to contribute to survival. These results suggest phenotypic
plasticity is an important process in A. cervicornis that contributes to fitness in a trait-specific
fashion. Moreover, the work presented here helps establish A. cervicornis as a system for
studying the eco-evolutionary dynamics of phenotypic plasticity that also can inform genetic-
and environment-based strategies for coral restoration.
1
Chapter 1 Introduction
1.1 Phenotypic plasticity in the context of reaction norms
Phenotypic variation arises as a product of fixed genetic and environmental effects.
Mechanisms of evolution are dependent on genetic variation because only heritable variation will
be passed on to subsequent generations. However, because selection ultimately acts on traits, the
contributions of the environment to phenotypic variation, such as phenotypic plasticity, have the
potential to shape the evolution of heritable traits by altering fitness (Price, Qvarnström, and
Irwin 2003; Stearns 1989). Phenotypic plasticity, or the ability of an individual to change its
phenotype in response to environmental change, has often been regarded as a property of a trait
(West-Eberhard 1989; Scheiner 1993). Plasticity offers a rapid mechanism to respond to
environmental change, allowing individuals to acclimate and even persist long enough to
facilitate genetically-based adaptation (Cheviron, Bachman, and Storz 2013; Zenni et al. 2014;
Fox et al. 2019). Fixed and plastic variation can be visualized through reaction norms that show
how the trait values of individuals differ within and among environments (Fig. 1.1). Here, fixed
differences between individuals is attributed to genetic variation while the differences in trait
values between environments is indicative of phenotypic plasticity. Plasticity can be measured at
the population level, as the change in mean trait value among environments, or at the individual
level as a genotype’s reaction norm (Fig. 1.1). Flat reaction norms indicate a canalized
phenotype (Individual D) while non-zero slopes (A, B, C) indicate the phenotype is dependent to
some degree on the environment an individual inhabits (Fig. 1.1). Plasticity (the slope of the
reaction norm) may be constant across individuals (C and B) or can vary within a population (A
vs. C), indicating that environmental responsiveness depends on the individual (Fig. 1.1).
2
While population-wide plastic changes allow entire groups to respond to environmental
change, variation in reaction norms among individuals (Fig. 1.1) indicate genetic variation in
environmental responsiveness. When plasticity is viewed as a reaction norm that can differ
among individuals, it also becomes a trait with the opportunity to shape mean trait evolution and
to evolve itself (Gavrilets and Scheiner 1993; Tufto
2000; Via and Lande 1985). However, evidence of
plasticity’s influence on trait evolution remains limited
(Ghalambor et al. 2015; Levis and Pfennig 2021) and
data on its potential to evolve (Leung et al. 2020;
Velotta et al. 2018) is often contrasting, suggesting that
ecological and evolutionary outcomes will depend on
the direction of the phenotypic change and the species,
traits, and environments in which it occurs (Van
Buskirk and Steiner 2009; Hendry 2016; Velotta and
Cheviron 2018). Therefore, additional experiments in a
diversity of systems and across traits will be required
before we are able to draw conclusions on the
ecological and evolutionary role of phenotypic
plasticity in nature.
1.2 Phenotypic plasticity’s influence on trait evolution
The influence of plasticity becomes clearer when framed in the context of the changing
selection pressure that may be associated with environmental change. Using Figure 1.1 as an
example, while the same amount of genetic variation is present in both environments, E1 will
3
favor a low trait value while E2 favors a high trait value. In the absence of plasticity, predicting
the response to selection in E2 based on observed trait values in E1 is straightforward. However,
plasticity shifts trait values among environments, allowing individuals to track environmental
change and thus diverge from additive predictions. When plastic changes align phenotypes with
the changing fitness landscape (e.g. individual A in Fig. 1.1), plasticity is considered adaptive. A
classic example of adaptive plasticity can be seen in Daphnia that develop defensive spines when
exposed to predator cues, ultimately lowering their susceptibility to predation (Boersma, Spaak,
and De Meester 1998). Alternatively, plasticity can be maladaptive if plastic responses shift trait
values away from fitness optima (Dewitt, Sih, and Wilson 1998). For example, exposure to high
elevation in native, low-elevation mice leads to a maladaptive increase in right heart ventricle
sizes that can precipitate heart failure and death (Velotta et al. 2018), suggesting that plastic
responses in this trait lower fitness compared to individuals that maintain or decrease heart size.
Maladaptive plasticity may also arise if there are significant fitness costs associated with
producing new phenotypes (Dewitt, Sih, and Wilson 1998; Murren et al. 2015; Hendry 2016)
even when traits change in an adaptive direction. Costs of plasticity become especially important
for determining selection on plasticity itself which considers the price of being a plastic
individual in multiple environments, even those where adaptive changes are not necessary (see
Section 1.3).
Understanding the direction of a plastic change in relation to changes in fitness optima
will be an important step in determining if and how plasticity will shape the evolution of a trait.
Adaptive plasticity may indirectly facilitate trait evolution when plastic changes improve fitness
and allow individuals to persist (“phenotypic rescue”) long enough for populations to diverge
(Simpson 1953; Price, Qvarnström, and Irwin 2003; Hendry 2016; Schmid and Guillaume 2017).
4
Adaptive plasticity is also hypothesized to contribute to the generation of novel adaptive
variation when a plastic response is favored due to its adaptive benefits (genetic accommodation)
and is produced so consistently in a group that genetic changes canalizing the trait become fixed
(genetic assimilation) (Kelly 2019; Levis, Isdaner, and Pfennig 2018; Levis and Pfennig 2021) .
However, adaptive plasticity may also constrain the evolution of mean trait values when all or
most individuals are able to survive environmentally-induced selection events, when plastic
individuals maintain population connectivity across environmental gradients, or when selection
acts on novel phenotypes without altering allele frequencies (Hendry 2016; Kelly 2019).
Maladaptive plasticity may also serve as a mechanism for adaptive evolution by exposing cryptic
genetic variation to selection thus accelerating changes in mean trait values within populations
(Ghalambor et al. 2007; Suzuki and Nijhout 2006; Ghalambor et al. 2015; Velotta et al. 2018).
Taken together, plasticity can be adaptive, maladaptive, or even neutral (Hendry 2016;
Ghalambor et al. 2007), and the system-specific effects will determine both the evolution of
mean trait values as well as the evolution of plasticity itself.
1.3 Evolution of plasticity
Beyond its role in the evolution of mean trait values, phenotypic plasticity itself can be
viewed as a trait, subject to evolutionary mechanisms dependent on heritable genetic variation
and an association with fitness. Intraspecific variation in plasticity is evidenced by genotype-by-
environment interactions (GxE) on trait values which imply that a phenotype is environmentally
responsive but the magnitude and/or direction of the response is dependent on an individual.
While abundant evidence of GxE in nature supports the classification of plasticity as a trait, there
has been much less work testing the relationship between the ability to be plastic and
overall/global fitness. Global fitness integrates selection outcomes across multiple environments
5
and is therefore able to account for costs of plasticity that may not be obvious under certain
conditions but are ever present, even when plastic changes are not needed. Potential costs of
plasticity include the energetic requirements of sensing and interpreting environmental cues, the
maintenance of pathways to produce new traits, reductions in fitness associated with producing
an imperfect phenotype relative to locally adapted individuals (limits to plasticity), or tradeoffs
between the ability to be plastic and mean trait values (Dewitt, Sih, and Wilson 1998; Murren et
al. 2015; Auld, Agrawal, and Relyea 2010).
While intraspecific variation in reaction norms is required for the evolution of plasticity,
selection does not act on plasticity directly. For this reason, plasticity is expected to be beneficial
in highly variable environments (see Section 1.4) where selection alternates, favoring different
phenotypes over time. As the environment fluctuates more frequently, plastic individuals capable
of producing appropriate phenotypes will have higher overall fitness compared to canalized
individuals that only experience high fitness for short periods during fluctuations (Hendry 2016).
In this scenario, limits to plasticity (the difference between a plastic phenotype and a canalized
locally adapted phenotype) can be overcome. However, if environments become more constant,
if plasticity is too costly, or if plasticity cannot accurately track the fitness landscape
(maladaptive plasticity), plastic individuals will be less fit overall relative to canalized, locally
adapted individuals and selection will favor decreased plasticity. Applying these theoretical rules
to Figure 1.1, if individuals experience E1 and E2 equally, overall fitness will be highest in
Individual A followed by B & C. However, if individuals are isolated to E2, selection will favor
Individual D and depending on the strength of selection, limits of plasticity arising in A, B, & C
will cause a reduction in plasticity of the population over time.
6
Evidence for the evolution of plasticity is sparse. The strongest support comes from
closely related populations and species where increased adaptive and decreased maladaptive
plasticity follows selection gradients in natural habitats (Levis, Isdaner, and Pfennig 2018;
Velotta et al. 2018). Experimental evolution studies have found plasticity can evolve over time
but often in ways contradictory to theoretical expectations (Schaum, Rost, and Collins 2016;
Leung et al. 2020; Fragata et al. 2016). While stable environmental conditions are expected to
reduce plasticity, in some cases, the reserve was true (Fragata et al. 2016). Similarly, while
fluctuating environments appear to select for increased plasticity in some systems, as expected
(Schaum, Rost, and Collins 2016), these conditions had no effect on the evolution of plasticity
(Leung et al. 2020). Understanding how plasticity itself will respond to selection becomes
increasingly important in light of its putative role in the adaptive divergence of life.
Increasing knowledge of plasticity and its potential to evolve has led to the formation of
new hypotheses for the generation of novel traits that is distinct from the traditional mutation-led
evolution. Plasticity-led evolution (PLE) proposes that fixed phenotypic variation can arise from
once environmentally sensitive traits (Pigliucci, Murren, and Schlichting 2006; Levis and
Pfennig 2019, 2021). PLE relies on the ability for adaptive plasticity to be selected for and
refined over time (genetic accommodation) until the new phenotype becomes fixed (genetic
assimilation) through genomic changes that canalize the trait. Evidence for PLE is sparse but this
hypothesis has been supported by examples such as the Spadefoot toad species complex where
divergent developmental pathways of two species appear to have originated from ancestral
developmental plasticity of tadpoles in desiccation susceptible vernal pools (Levis, Isdaner, and
Pfennig 2018). Considering that this system has only recently received attention, it illustrates the
utility of non-model organisms for addressing basic biological questions regarding the evolution
7
of plasticity. Moreover, PLE demonstrates that understanding how and under what scenarios
plasticity will evolve can help reshape our view of the origin of biological diversity.
1.4 Conditions influencing the evolution of plasticity
Trait plasticity is only relevant in the presence of environmental change, so the adaptive
nature of plasticity, its influences on trait evolution, and its own evolution will be highly
dependent on patterns of environmental fluctuation. Constant conditions render plasticity useless
and instead favor canalization of a single phenotype. However, fluctuating conditions make
canalized individuals fall outside of fitness optima more often and instead favor individuals
capable of tracking environmental conditions. Therefore, habitats with greater temporal or spatial
variation will be expected to favor plasticity over canalization (Hendry 2016; Bitter et al. 2021).
While natural populations follow this trend (De Meester 1996; Lind et al. 2011) experimental
evolution studies provide conflicting results where fluctuating environments can both increase
and decrease plasticity in different species (Schaum, Rost, and Collins 2016; Leung et al. 2020;
Fragata et al. 2016; Barley et al. 2021). Rather than variability alone, the predictability of
environmental fluctuations likely plays a key role in the evolution of plasticity because reliable
environmental cues will be necessary for organisms to produce the appropriate response for
future conditions (Hendry 2016; Bitter et al. 2021). Highly predictable environments allow
individuals to better track fitness optima, reducing the risk of inappropriate phenotypes, and
supporting the maintenance or enhancement of plasticity.
While particular environments can alter the ecological and evolutionary dynamics of
plasticity between populations of species, characteristics of species themselves can also influence
the evolution of plasticity. For instance, species with high dispersal will likely be more reliant on
plasticity as new generations have a heightened potential of experiencing habitats with differing
8
environmental conditions (Hendry 2016; Eierman and Hare 2016). Species with long life spans
are expected to encounter environmental variation over time, and if incapable of behavioral
evasion, acclimation, tolerance, or resistance via plasticity may be the only strategies to cope
with environmental stress. Therefore, organisms with these characteristics are promising
candidates for exploring plasticity’s role in ecology and evolution.
1.5 Plasticity in coral
Scleractinian corals are clonal cnidarians that deposit a hard calcium carbonate as they
grow, making them responsible for much of the structural complexity of tropical reef ecosystems
(Alvarez-Filip et al. 2009; Zawada, Piniak, and Hearn 2010). Although not a traditional model
organism in ecology and evolution, reef building corals have numerous characteristics that make
them interesting subjects in which to explore the roles of phenotypic plasticity. For example,
corals are sessile, long-lived, slow to reach reproductive maturity (Babcock 1991b), and have
high dispersal potential (Drury et al. 2018; Hemond and Vollmer 2010) suggesting that
individuals will likely experience environmental change that will require phenotypic plasticity to
survive. In fact, plasticity has been commonly documented in coral morphology (Todd 2008;
Tambutté et al. 2015), physiology (Muller et al. 2021; Ziegler et al. 2014), and gene expression
(Bay and Palumbi 2015; Rocker et al. 2019) in response to a variety of abiotic factors, indicating
the environmental responsiveness of phenotypes. Skeletal shape and structural integrity can vary
in response to light, nutrients, and depth (Hoogenboom, Connolly, and Anthony 2008; Muko et
al. 2000) (Holcomb et al., 2015) and chlorophyll, protein, and fatty acid concentrations have
been shown to be plastic as well (Ziegler et al. 2014; Rocker et al. 2019). Plastic changes at the
molecular level also stand out as important mechanisms underlying responses of populations to
environmental variation (Bay and Palumbi 2015; Kenkel and Matz 2016; Bay et al. 2009).
9
Although documented plastic changes in coral phenotypes appear to be adaptive for populations
(Kenkel and Matz 2016; Muko et al. 2000; Hoogenboom, Connolly, and Anthony 2008; Grottoli,
Rodrigues, and Palardy 2006), it is unclear if plastic individuals exhibit a higher overall fitness.
This gap in knowledge arises because 1) plasticity within-populations has rarely been
investigated (Drury and Lirman 2021; Bruno and Edmunds 1997; Barott et al. 2021) limiting our
knowledge of variation occurring between individuals and 2) cost of plasticity across multiple
environments are often ignored when considering whether plasticity improves fitness. Further
exploration of intraspecific variation in coral plasticity is especially warranted because they have
many of the characteristics that are expected to promote the evolution of adaptive plasticity
(Hendry 2016).
Tropical ecosystems are experiencing dramatic changes in environmental conditions that
will continue to challenge coral species around the world (Hoegh-Guldberg et al. 2007; Manzello
2015). While adaptation over multiple generations will be required for long term persistence,
acclimation via phenotypic plasticity can be an immediate strategy that will “buy time” under
climate change. For coral targeted for restoration, the importance of quantifying intraspecific
variation in plasticity is two-fold. First, genotype-dependent responses to environmental
conditions suggest that trait variation between individuals in one environment will not be
consistent across environments (Fig. 1.1). Patterns of GxE may then be useful when determining
which individuals can be used to repopulate certain reef environments. Alternatively, if GxE
does not exist, fixed differences between individuals could be used to identify global winners
and losers for restoration (Barott et al. 2021; Morikawa and Palumbi 2019). Second, the costs
and benefits of plasticity should be considered when predicting the trajectory of ecologically
important coral traits under natural or artificial selection. For example, increased thermal
10
tolerance will be preferred as oceans warm during climate change (Manzello 2015) but if
tolerance is found to be inversely related to plasticity, selection for heat tolerance will come at
the cost of environmental responsiveness. Positive and negative relationships between plasticity
and other traits will therefore influence the adaptive capacity of coral and justify further
investigation in species threatened by climate change, especially those that are targets for human-
assisted evolution strategies.
1.6 Acropora cervicornis as a system to study plasticity
Acropora cervicornis, Staghorn coral, is a branching coral that significantly contributes
to the structural complexity of Caribbean reefs. Once dominant, A. cervicornis populations have
declined precipitously since the 1970’s (Cramer et al. 2020) with losses of coral cover up to 80%
in regions like the Florida Keys Reef Tract (Gardner et al. 2003; Miller, Bourque, and Bohnsack
2002). Now an IUCN critically endangered species (Aronson et al. 2008), A. cervicornis has
become the focus of massive efforts that aim to ensure its persistence through a variety of
ecological and evolutionary strategies (Young, Schopmeyer, and Lirman 2012; Lohr et al. 2015;
Vollmer and Palumbi 2007; Calle-Triviño et al. 2018). Reef restoration programs have begun
asexually propagating coral genotypes and using clones to repopulate degraded reefs across
Florida and the Caribbean (Lohr et al. 2015; Young, Schopmeyer, and Lirman 2012).
Meanwhile, research efforts are improving our understanding of the genetic and environmental
factors contributing to coral performance (O’Donnell, Lohr, and Bartels 2017; Woesik et al.
2021; Drury, Manzello, and Lirman 2017) and are contributing to the development of
intervention strategies such as assisted migration and artificial selection. This increasing
emphasis on A. cervicornis restoration and research, paired with its life history characteristics,
11
provide the framework and the justification for exploring the role of phenotypic plasticity in
coral.
1.7 Goals of this dissertation
While intraspecific variation in plasticity has been extensively documented in multiple
systems, there are disproportionately few empirical studies testing plasticity’s adaptive
significance. Traditional model organisms may often lack characteristics found in nonmodel
systems that are particularly beneficial for answering questions regarding the ecological and
evolutionary implications of phenotypic plasticity. This dissertation seeks to address these gaps
in knowledge using the endangered coral Acropora cervicornis to 1) quantify intraspecific
variation in plasticity at the morphological, physiological, and molecular levels and 2) determine
the relationship between plasticity and overall fitness.
An in situ transplant experiment conducted in the Florida Keys was used to determine the
genetic and environmental factors influencing morphology and gene expression in a set of
restoration-stock A. cervicornis genotypes. To quantify changes in coral morphology, I first
developed a workflow using 3D photogrammetry technology to generate and analyze 3D models
of coral colonies created from photographs taken at experimental locations (Ch. 2). Performance
metrics from experimental individuals were then used to identify fixed and plastic effects on
coral morphology and survival and to determine the relationship between morphological
plasticity and coral fitness (Ch. 3). Changes in gene expression profiles of individuals after
transplantation were then used to quantify molecular plasticity and its relationship to coral
survival, growth, and morphological plasticity (Ch. 4). Finally, an aquarium-based experiment
was used to determine individual responses to temperature and acidification stress (Ch. 5).
Plasticity in response to temperature and pH was measured for symbiont, host, and holobiont
12
traits and then compared to individual performance to assess the relationship between
physiological plasticity and fitness under ecologically relevant stressors (Hoegh-Guldberg et al.
2007; Manzello 2015).
13
Chapter 2 Development of 3D photogrammetry method to measure individual
coral colony morphology
This chapter appears as published in Frontiers in Marine Science 8 (2021): 384.
Summary of Contribution
This project was completed as part of a larger experiment conceptualized by Carly Kenkel, my
PhD advisor, and Cory Krediet of Eckerd College. I developed methods for 3D photogrammetry
to be applied to this experiment and this project represents a proof of concept for these methods.
Carly, Cory, and Erich Bartels (Mote Marine Laboratory) contributed to field work and image
collection. Undergraduate Sibelle O’Donnell significantly contributed to data collection. I led
manuscript writing while Carly, Cory, Erich, and Sibelle contributed to the published version.
2.1 Abstract
The ability to quantify changes in the structural complexity of reefs and individual coral
colonies that build them is vital to understanding, managing, and restoring the function of these
ecosystems. However, traditional methods for quantifying coral growth in situ fail to accurately
quantify the diversity of morphologies observed both among and within species that contribute to
topographical complexity. Three-dimensional (3D) photogrammetry has emerged as a powerful
tool for the quantification of reefscape complexity but has yet to be broadly adopted for
quantifying the growth and morphology of individual coral colonies. Here we debut a high-
throughput method for colony-level 3D photogrammetry and apply this technique to explore the
relationship between linear extension and other growth metrics in Acropora cervicornis. We
fate-tracked 156 individual coral transplants to test whether initial growth can be used to predict
subsequent patterns of growth. We generated photographic series of fragments in a restoration
nursery immediately before transplanting to natural reef sites and re-photographed coral at 6
14
months and 1 year post-transplantation. Photosets were used to build 3D models with Agisoft
Metashape, which was automated to run on a high-performance computing system using a
custom script to serially process models without the need for additional user input. Coral models
were phenotyped in MeshLab to obtain measures of total linear extension (TLE), surface area,
volume, and volume of interstitial space (i.e., the space between branches). 3D-model based
measures of TLE were highly similar to by-hand measurements made in the field (r = 0.98),
demonstrating that this method is compatible with established techniques without additional in
water effort. However, we identified an allometric relationship between the change in TLE and
the volume of interstitial space, indicating that growth in higher order traits is not necessarily a
linear function of growth in branch length. Additionally, relationships among growth measures
weakened when comparisons were made across time points, implying that the use of early
growth to predict future performance is limited. Taken together, results show that 3D
photogrammetry is an information rich method for quantifying colony-level growth and its
application can help address contemporary questions in coral biology.
15
2.2 Introduction
The three-dimensional structural complexity of coral is central to the ecological function
of reefs (Alvarez-Filip et al. 2009; Zawada, Piniak, and Hearn 2010). Structurally intricate reefs
sustain high biological diversity (Risk 1972; Graham and Nash 2013), increase productivity
(Szmant 1997), and reduce hydrodynamic energy (Lugo-Fernández et al. 1998; Monismith
2007). Colony morphology also influences small scale water flow that controls the size of the
diffusion boundary layer, influencing heat and mass transfer (Stocking et al. 2018) and pH of the
tissue surface (Chan et al. 2016). These physicochemical processes support key aspects of a
coral’s biology, including nutrient uptake, and mitigation of acidification and thermal stress
(Dennison and Barnes 1988; Lesser et al. 1994; Jimenez et al. 2011). Consequently, the ability to
quantify changes in the structural complexity of reefs and the individual coral colonies that build
them is vital to understanding, managing, and restoring the function of these ecosystems.
Traditional methods for measuring coral growth have provided foundational knowledge
of extremely fast growth rates in Acropora cervicornis compared to other species (Lirman et al.
2014), trade-offs between growth and thermal susceptibility (Jones and Berkelmans 2010; R.
Cunning et al. 2015) and the capacity for morphological plasticity (Bruno and Edmunds 1997;
Todd 2008; Drury, Manzello, and Lirman 2017). However, the type and quality of information
that can be obtained from each individual metric is limited. Moreover, many require handling, or
even destructive sampling, hindering the potential to investigate temporal changes in growth
using repeated measures. Non-invasive methods for measuring growth in situ include total linear
extension (Johnson et al. 2011) and estimated ellipsoidal volumes (Kiel, Huntington, and Miller
2012). TLE relies on linear measures of branch lengths, typically taken by hand with calipers or
a ruler, to quantify skeletal size (Johnson et al. 2011). However, this measure is only applicable
16
to branching species, and can be arduous to complete in situ as colonies increase in size (Lirman
et al. 2014). Alternatively, colony size can be calculated from the volume of an ellipsoid with the
same length, width, and height as a coral colony (Kiel, Huntington, and Miller 2012). However,
this method is likely unable to resolve changes in the pattern of growth that do not increase the
maximum dimensions of a colony. Invasive techniques to measure coral growth include direct
measures of surface area (SA) via wax dipping (Stimson and Kinzie 1991; Veal et al. 2010) ,
volume (V) via water displacement (Jokiel, Maragos, and Franzisket 1978; Herler and
Dirnwöber 2011), and mass via buoyant weighing (Davies 1989). However, each of these
methods requires handling and, for wax dipping, sacrificing the coral completely. Measuring
growth in one dimension also ignores structural differences. For example, two colonies may have
the same linear extension or ellipsoidal volume but encompass vastly different morphological
forms (Pratchett et al. 2015). Therefore, multiple independent methods would be required to gain
a comprehensive understanding of colony growth and morphology.
Three-dimensional (3D) photogrammetry, i.e., obtaining measurements from digital,
scaled 3D representations of objects, has become a powerful tool for quantifying structural
components of coral reefs (Burns et al. 2015; D’Urban Jackson et al. 2020; Hernández-Landa,
Barrera-Falcon, and Rioja-Nieto 2020). There has been growing application of this technology to
quantify reefscape complexity (McKinnon et al. 2011; Burns et al. 2015; Leon et al. 2015) , but
3D photogrammetry has yet to be broadly adopted for quantifying growth and morphology of
individual coral colonies, despite several studies validating its application (Figueira et al. 2015;
Lavy et al. 2015; Agudo-Adriani et al. 2016; Ferrari et al. 2017; House et al. 2018; Doszpot et al.
2019; Lange and Perry 2020). One reason for this may be that although model building
algorithms and the user-friendliness of applications continue to improve, methodological
17
nuances can result in differences in the required processing time, quality of models produced,
and traits measured (Lange and Perry 2020; Lavy et al. 2015). Similar to open-source software
that allows for community contribution that drives innovation (Hippel and Krogh 2003),
increasing accessibility and reproducibility of 3D photogrammetry methods can help facilitate
broader adoption and advancement of this technique. The information richness of a single 3D
model is a major methodological advantage, expanding the range of biological questions that can
be investigated. For example, in addition to quantifying coral growth via TLE, SA, or V,
quantifying the volume of interstitial space (Vinter) can give insights into the ecological function
of coral growth by providing an estimate of the amount of resulting habitat space that is created
(Coker, Wilson, and Pratchett 2014; Agudo-Adriani et al. 2016). 3D models can also provide
digital structures to enable more accurate mapping of the hydrodynamics of flow around colonies
(Stocking et al. 2018). Finally, the permanent record of 3D models allows for researchers to
revisit historical datasets to verify previous measures or test new hypotheses. This unprecedented
access to multiple traits once hard to measure comes without the need to touch or manipulate the
coral itself, making it especially useful for conducting repeated measures over time.
The ability to non-invasively quantify changes in key performance traits also has
important practical applications, particularly in the context of reef restoration. The structural
complexity of Caribbean coral reefs has been altered by dramatic losses of branching acroporids
(Miller, Bourque, and Bohnsack 2002; Alvarez-Filip et al. 2009). As a result, A. cervicornis and
A. palmata have been the focus of broad-scale restoration efforts in Florida and the broader
Caribbean. Significant effort has been invested in the development of in-water and ex situ
nurseries in order to generate biological material to outplant to degraded reef sites (Rinkevich
1995; Young, Schopmeyer, and Lirman 2012; Lohr et al. 2015). However, the ultimate success
18
of these efforts has been hampered by the inability to predict the performance of coral outplants,
given that both genetic and environmental variables can influence survival and growth (Drury,
Manzello, and Lirman 2017; O’Donnell, Lohr, and Bartels 2017). The ability to use early, non-
invasive indicators to quantify outplant success would be beneficial in restoration settings that
aim to optimize efforts to repopulate Caribbean reefs (Edmunds and Putnam 2020; Parkinson et
al. 2020).
Reliable predictors of coral performance that can be evaluated before or soon after
outplanting can guide restoration programs to strategically enhance efforts. However, prior work
suggests that the predictive power of traditional growth metrics may be limited. Early growth
rate in TLE for Acropora cervicornis was found to be a poor predictor of future growth rate for
nursery-grown (O’Donnell et al. 2018) and wild corals (Edmunds 2017). Similarly, early growth
rates for two massive coral species as measured in nurseries explained only a portion of future
performance, and the predictive power of these metrics varied over sampling points and traits
(Edmunds and Putnam 2020). The utility of higher order traits, such as SA, V, and Vinter, for
predicting subsequent growth remains unknown because handling or destructive sampling of
outplants is not possible in a restoration context.
Here we apply high-throughput colony-level 3D photogrammetry to fill this knowledge
gap. First, we develop an open-access protocol for conducting colony-level 3D photogrammetry
that provides guidance for in situ image collection, makes high-throughput model building more
readily accessible, and creates a repeatable, standardized method for coral phenotyping from 3D
models. We then employ this method to address questions regarding the predictive power of
higher order morphological traits in A. cervicornis by quantifying the growth of outplants at
different intervals throughout 1 year. We test to what degree total linear extension (TLE), a
19
commonly used non-invasive metric of growth in A. cervicornis (Johnson et al. 2011), can be
used to estimate growth in higher order traits, such as surface area (SA), volume (V), and the
volume of the interstitial space (Vinter), in addition to testing for predictive correlations among
these traits.
2.3 Methods
2.3.1 Study Design
In April 2018, three ramets of each of ten genotypes of A. cervicornis from Mote Marine
Laboratory’s in situ nursery were transplanted in triplicate to nine offshore reef sites throughout
the Lower Florida Keys under FKNMS permits 2015-163-A1 and 2018-035 (n = 27 initial
ramets/genotype). Triplicate ramets were attached to the reef substrate, with one ramet per
genotype assigned to each of 3 randomized arrays within each site. Coral morphology was
quantified with 3D photogrammetry at 3 time points: in the in situ nursery prior to
transplantation (April 2018), and 6 months (October 2018) and 12 months (April 2019) post-
transplantation. Of the 270 outplanted coral fragments, 72 died or were lost from the transplant
after 12 months. An additional 43 coral fragments, including 40 nursery fragments, 2 fragments
at 6 months, and 1 fragment at 12 months were removed from the dataset due to insufficient
photographic coverage causing incomplete models, resulting in a final dataset of 156 fragments
measured for all traits at all-time points. Additionally in April 2018 and October 2018, TLE was
measured by-hand with a flexible ruler following the protocol of Lirman et al. (2014) and these
measurements were used to compare with TLE obtained through 3D photogrammetry in order to
assess the potential for backward compatibility.
20
2.3.2 Image Capture
Photographs of outplanted corals were taken by SCUBA divers following the protocol
available at dx.doi.org/10.17504/protocols.io.bgdcjs2w. Briefly, photographs were taken from
multiple angles using the Olympus Tough TG-4/TG-5 (Olympus America Inc.), maintaining
roughly 80% overlap until the entire coral colony was captured with 2D images. The Olympus
Tough TG-4/TG-5 was chosen for its low price point, its built-in underwater capabilities which
avoids the need for auxiliary underwater housings, and its small size which allowed divers to
operate the camera with one hand in high wave-action environments. Cameras were set to
“underwater” or “auto” mode with flash turned off, and photos were taken by either manually
pressing the shutter or set to capture images automatically at 1 s intervals, depending on
photographer preference. In April 2018, coral ramets were photographed in the nursery prior to
transplantation in groups of 10 (Supplementary Figure 1) requiring 120–383 photos per group.
Post-transplantation, in October 2018 and April 2019, a custom-built scaling object incorporating
Agisoft Metashape markers was included around each coral to provide a fixed reference for
model generation. The number of photos taken for each colony ranged from 7 to 293, with the
number of photographs depending on size of the colony and water quality conditions at a given
location.
2.3.3 Image Pre-processing and Model Building
Photos were downloaded and sorted into separate photo sets for each group of 10 nursery
corals (April 2018) or individual outplanted colony (October 2018, April 2019). Photographs
where the coral was out of focus or out of frame were manually removed. Photographs shot in
“auto” mode were color corrected in Lightroom (Adobe® Lightroom® software) to remove
21
green tint and enhance contrast of the coral when needed, for example, when model building was
hindered by bright white branch tips.
Three-dimensional models were generated from photo sets using Agisoft Metashape
(Agisoft LLC, St. Petersburg, Russia). Metashape is processing intensive so in order to rapidly
build 3D models with limited user input, we completed all model building on a Dell PowerEdge
R910 with an Intel® Xeon® Processor E7-4850 with Metashape manually limited to 20–40
CPUs and 250 GB of RAM. Due to limited licensing capabilities of Metashape, a custom Python
script was used to serially process models and Metashape settings were adjusted when needed for
individual models. Models were exported as Wavefront (.OBJ) files which are capable of
maintaining color pulled from 2D photos. Models with portions of colonies missing were rebuilt
with a modified script that expands the size of the bounding box during the “Align Photos” step
to include all calculated tie points. All bioinformatic scripts used to run Metashape on the
command line can be found at https://github.com/wyattmillion/Coral3DPhotogram. We assessed
the performance of 200 automated Agisoft Metashape runs through the number of tie points
found from each photoset (overlapping points found in two or more photos) and the number of
faces making up each mesh.
2.3.4 3D Phenotyping
Models generated in Metashape were imported into the free 3D model editing software,
Meshlab v2016.12 (Cignoni et al., 2008) to obtain measurements of TLE, SA, V, and Vinter
following protocols outlined at dx.doi.org/10.17504/protocols.io.bgbpjsmn. Briefly, models were
scaled using Agisoft Metashape markers and scaling accuracy was verified by remeasuring
markers. Models were manually trimmed using the Select vertices and Remove current set of
selected vertices tools leaving only living coral in the models. Small holes in the areas of coral
22
tissue were filled prior to phenotyping models using the Close Holes filter. To calculate TLE, the
lengths of all branches were manually measured and summed using the Measuring Tool. Surface
area was automatically quantified using MeshLab’s Compute Geometric Measure filter. The
bottom of coral models were then closed using the Close Holes filter in order to make the mesh
“watertight” for the Compute Geometric Measure filter to measure the volume of the coral. The
volume of a colony was subtracted from the volume of a convex hull, i.e., a mesh overlaid over
the most extreme branch tips (Supplementary Figure 2), in order to quantify Vinter. Traits such as
“proportion occupied” (Doszpot et al. 2019) and “convexity” (Zawada, Dornelas, and Madin
2019) utilize a convex hull but instead derive morphometric features standardized to coral
volume or surface area. Here we used Vinter to measure the absolute volume of empty space
between coral branches.
Five colonies suffered severe breakage and only a small portion of living tissue remained
after 12 months (Supplementary Figure 3). 3D models were not built for these colonies and
instead, SA was measured in ImageJ (Rasband) following (Kenkel, Setta, and Matz 2015) and V
and Vinter were recorded as 0 because the colony tissue represented a 2D area with no living 3D
structure beyond the height of a single polyp (Supplementary Figure 3). TLE was recorded as 1
mm as this was the average height of the coral polyps on the colonies at the 6-month time point.
2.3.5 Statistical Analysis
All statistical analysis was completed in R 3.6.3 (R Core Team, 2020). One coral was
randomly chosen to assess precision of the 3D photogrammetry method. To do this, we
determined the coefficient of variation (CV) for each focal trait measure (TLE, SA, V, and Vinter)
across 6 replicate models built from a reduced photoset where a random 10% of the photographs
were removed, similar to Ferrari et al. (Ferrari et al. 2017) and Lange and Perry (2020). The
23
corals used in this study are part of restoration efforts and as such could not be removed from the
substrate to measure volume or surface through traditional methods. Instead, TLE measured by-
hand in situ was used to ground-truth measurements of TLE obtained from 3D models. Pearson
correlations were used to assess the relationship between by-hand and 3D model based
measurements of TLE collected at two time points (initial, and 6 months, n = 311) and to assess
whether the difference in methods is a function of coral size. To assess the strength of predictive
relationships among trait measurements within and across time-points, Kendall’s tau correlations
were used to relate the change in trait values between 0–6 months and 6–12 months (n = 156).
Kendall’s tau was used for its ability to handle non-normally distributed data and heteroscedastic
residuals, as well as outliers, which here represent true biological variation, and non-linearity
(Newson, 2002). Additive polynomial regressions were used to maximize the fit of non-linear
relationships observed for TLE and Vinter, while adding the fewest polynomial terms.
A. cervicornis naturally reproduces through fragmentation (Tunnicliffe 1981; Drury et al.
2019). To assess whether natural fragmentation events influenced the predictive relationship
among traits, we recorded the number of breakage events experienced by each coral by
comparing photographic records across time points. Breakage was coded as a binary with coral
experiencing one or more breakage events considered “Broken” and corals experiencing no
breakage considered “Unbroken.”
2.4 Results
2.4.1 Modeling Performance and Precision
3D models of individual colonies (6- and 12-month time points) were built from an
average of 91 photos, while as few as 7 photos (for the smallest colonies) and as many as 293
photos (for the largest colonies) were used in successful runs. The number of photos used as
24
input into the model building step was loosely related to the size of the colony following a linear
pattern (Fig. S2.4). We used a subset of 200 models to evaluate our model building pipeline
which took about 1.5 h on average to build models of colonies ranging from 0.1 to 74 cm TLE
and with 1 to 21 branches. The number of tie points for the 200 model runs ranged from 920 to
151,132, and the number of faces ranged from 137,873 to 5,337,592.
Replicate models of a single coral colony at the 12-month time point were measured to
assess the precision of model building and phenotyping steps. This coral represented a relatively
large and complex colony, with 21 branches total (TLE = 74 cm, SA = 261 cm2, V = 85 cm3,
Vinter = 434 cm3). Replicate 3D models had similar resolution with an average of 2,125,932 faces
per model (CV = 1.5%). The replicate models produced very similar trait measurements, once
individually scaled, with a CV of 1.3% for TLE, 3.7% for SA, 5.7% for V, and 4.9% for Vinter.
2.4.2 Photogrammetry vs. Traditional Methods
A strong correlation was observed between TLE measures made by-hand in situ and
those obtained from the 3D models (r = 0.98, n = 311, Fig. 2.1A). The average absolute
difference between the two methods for a single colony was 12.8 mm (Fig. 2.1A). Additionally,
there was a significant but weak positive correlation between the absolute size of a focal colony
and accuracy, as quantified by the difference in the value of TLE, based on how the
measurement was made (r = 0.29, p < 0.001, Fig. 2.1B). A greater agreement between by-hand
and 3D model based measurements (i.e., a lower difference value) indicates higher accuracy,
suggesting that discrepancies between the two methods increase as colonies increase in size.
25
Figure 2.1: The relationship between (A) TLE measured by-hand in the field and from 3D
models and (B) the absolute difference between by-hand and 3D based measures and the size of
the focal coral colony (n = 311).
2.4.3 TLE as a Proxy for Higher Order Traits
TLE was highly correlated with growth in higher order traits when compared within the
same time point (Fig. 2.2 and Fig. S2.5). The change in TLE over the first 6 months of
outplanting was very strongly correlated with growth of SA, V, and Vinter over the same time
window (τ = 0.76–0.85). The strength of correlations was slightly reduced in the second 6-month
window, but TLE was still strongly correlated with other three-dimensional traits (τ = 0.64–0.8).
Regression plots of individual trait correlations suggest that nonlinear dynamics likely
influenced trait relationships (Fig. S2.5). For example, a second order polynomial better
explained the relationship between growth in TLE and growth in Vinter during the first 6 months
(r = 0.922, p = 2.2e-16) and during the second 6 months (r = 0.864, p = 2.2e-16; Fig. S2.3), with
Vinter increasing at a faster rate than TLE.
26
Figure 2.2: Correlation matrix showing the relationships among growth in traits over time.
Values indicate Kendall’s tau correlation coefficient. All relationships were significant (p <
0.05). Color indicates the magnitude and direction of the correlation.
2.4.4 Relationships Among Growth Traits Over Time
Relationships among growth measurements weakened markedly when comparisons were
made across time points (Fig. 2.2). Trait growth in the first 6 months post-outplant was weakly
correlated with growth of the focal trait over the second 6 months (τ = 0.35–0.45, Fig. 2.2). The
use of non-linear regressions explained an additional 2.5% of the variation in subsequent TLE
growth and an additional 1.7% of the variation in subsequent Vinter growth (Fig. 2.3). The
relationships between future and subsequent growth in SA and V were not significantly
improved by non-linear models (Fig. S2.6).
27
Figure 2.3: Growth in TLE compared to growth in Vinter over the first 6 months (A) and over the
second 6 months (B). The solid blue line shows the second order polynomial fit to each
relationship. Shaded areas around each line represent the 95% confidence intervals (n = 156).
Random coral breakage influences the predictive power of initial growth measurements,
but relationships remain weaker than for the within-time-point correlations even when broken
corals are excluded from models (Fig. 2.4). For example, the correlation between initial and
subsequent Vinter growth improved when broken coral were excluded (Broken included: τ =
0.45, p = 5.05e-17; Fig. 2.2; Broken excluded: τ = 0.63, p = 3e-9, Fig. 2.4D). Similar effects of
breakage diminish the power of initial growth to predict future growth in other traits as Kendall’s
tau coefficients also increased for correlations between initial and subsequent growth in TLE and
SA (τ = 0.57 and τ = 0.49, respectively) when broken corals were excluded (Fig. 2.4A,B). No
significant relationships between initial and subsequent growth were observed for corals that
experienced breakage (Fig. 2.4).
28
Figure 2.4: Growth over the first 6 months compared to growth over the second 6 months for
TLE (A), SA (B), V (C), and Vinter (D) colored by whether coral experienced breakage (open
circles, dashed red line) or not (open triangles, solid blue line). Kendall’s tau coefficients and p-
values are included for correlations. Shaded areas around each line represent the 95% confidence
intervals (n = 156).
The change in TLE over the first 6 months of outplanting was generally a poor predictor
of future growth. The strongest predictive relationships for an initial measure of growth were
observed for V, which had the strongest correlations with subsequent growth in all traits
regardless of breakage events (τ = 0.37–0.48, Fig. 2.2), and when broken corals were excluded (τ
= 0.37–0.49, Fig. S2.7). Interestingly, initial growth in all traits displayed stronger relationships
with subsequent growth in Vinter, as evidenced by stronger correlation coefficients in the 6–12
month Vinter column, than for other initial to subsequent growth comparisons (TLE, SA, V, Fig.
2.2).
29
2.5 Discussion
The complex structure of a coral reef is vital to its function (Alvarez-Filip et al. 2009;
Zawada, Piniak, and Hearn 2010). Consequently, declines in populations of reef building corals
dramatically change the topography of these biologically diverse ecosystems (Denis et al. 2017) .
Acropora cervicornis was once a prolific reef builder in the Caribbean and its decline has
sparked massive restoration efforts in areas like the Florida Keys (Young, Schopmeyer, and
Lirman 2012). To quantify restoration outcomes, practitioners often desire to track colony
growth as a performance metric (Young, Schopmeyer, and Lirman 2012; Ware et al. 2020)
because coral size correlates with fitness-related traits like survival and fecundity (Babcock
1991b; Hughes, Ayre, and Connell 1992; Álvarez-Noriega et al. 2016). However, given the scale
of outplanting efforts, invasive methods to track coral growth are inefficient and counter to the
goals of restoration when destructive sampling is necessary (Johnson et al. 2011). Similarly,
while non-invasive measurements, such as TLE, are informative (Johnson et al. 2011), they can
be painstaking and slow if done by-hand (Lirman et al. 2014). Here we show that 3D
photogrammetry can generate accurate measurements of linear growth while simultaneously
providing information on higher order traits reflective of an individual coral’s condition, like
surface area, and volume, as well as its function, such as the volume of interstitial space
produced by its branching morphology. We find that linear extension is a strong correlate of
growth in higher order traits within a monitoring period. However, predictive relationships
across time-points are poor. Moreover, breakage and non-linear relationships among traits
present obstacles to predicting future growth in coral colonies. Variation in morphological
growth brought on by environmental variation may further complicate projections made from
initial growth rates.
30
2.5.1 Relationships Between TLE and Higher Order Traits
Growth in TLE was a reliable estimator for growth in SA, V, and Vinter within a particular
monitoring window. Relationships were the strongest within the first 6 months post-outplant
suggesting that TLE could be used to approximate higher order traits in smaller, less complex
colonies. However, this relationship weakened on average during the second 6-month monitoring
window. This decrease in the strength of relationships among traits over time could be due to
allometric variation in the growth rates of traits relative to overall colony size.
Allometry broadly describes the scaling of two biological traits (West, Brown, and
Enquist 1997) but often implies differences in growth rate between these traits. For example,
horn length in male rhinoceros beetles showed a positive allometric relationship to prothorax
size, meaning that increases in body size resulted in disproportionately large increases in horn
size (McCullough et al. 2015). Allometric scaling has also been used to describe size-dependent
growth rates in plants that show negative allometry—growth rate decreases as overall biomass
increases (Wesselingh et al. 1997; Félix-Burruel et al. 2019). Additionally, allometric
relationships can change due to environmental effects, resulting in inconsistent scaling between
traits within a species. This phenomenon has been reported in height-stem diameter relationships
across environmental gradients for some tree species (Lines et al. 2012). Finally, intraspecific
variation in allometric scaling between traits can result in further phenotypic variation, for
example, scaling between horn and body size can vary among competing males of a single
rhinoceros species (McCullough et al. 2015).
We uncover an exponential relationship between TLE and Vinter where marginal increases
in TLE result in larger increases in Vinter as corals grow, indicating positive allometry (Fig. 2.3).
Similar positive allometry was observed in multiple coral morphotypes for shelter volume-
31
colony diameter and planar area relationships (Urbina-Barreto et al. 2021), and between colony
surface complexity and colony volume (Zawada, Dornelas, and Madin 2019). The allometric
relationship observed here between TLE and Vinter means that small colonies create interstitial
space at a slower rate than larger colonies. This may slow restoration of the ecological function
of A. cervicornis reefs if colonies are not able to reach large sizes due to natural breakage
(Tunnicliffe 1981; Madin et al. 2014) in addition to intrinsic negative scaling that can limit the
overall size of an otherwise healthy coral (Dornelas et al. 2017). Negative allometric
relationships can also impact morphological growth. Allometry in colony geometry and volume
(Zawada, Dornelas, and Madin 2019) indicate that growth in higher order traits has the potential
to plateau or regress as coral size increases. Changes in colony shape and geometry would result
in shifts of ecological functionality over colony size ranges. The complex growth dynamics
influenced by positive and negative allometry underscore the need to track various forms of
morphological growth to understand the ecological consequences (Lesser et al. 1994; Agudo-
Adriani et al. 2016; Chan et al. 2016), a need which can be met by 3D photogrammetry.
The remaining correlations between TLE and SA or V in A. cervicornis were explained
by linear functions; however, the strength of these relationships decreased over the second 6
months. Explanations for this pattern likely include both environmental or genetic variation in
allometric relationships. For example, location-specific variation in the relationship between
ellipsoid volume and TLE has been reported in A. cervicornis (Huntington and Miller 2014).
Moreover, while intraspecific allometric variation was not found for negative size-dependent
growth (Dornelas et al. 2017), an extensive body of work has identified genotypic variation in A.
cervicornis morphology (Bowden-Kerby 2008; Kuffner et al. 2017; Lohr and Patterson 2017;
Drury et al. 2019) supporting the potential for intraspecific variation in allometric scaling
32
relationships among traits. Here, colonies representing multiple genotypes were outplanted to
multiple reef sites likely to vary in environmental conditions; however, our current dataset lacks
the temporal resolution with which to identify patterns of site-specific and intraspecific variation
in allometry between TLE and higher order traits. Tracking individuals with high temporal
resolution would help capture site-specific and genotypic growth curves spanning multiple size
classes. Allometric variation among colonies can also result in varying trait growth relationships
within or between reef sites, which could play a role in the weakened among-trait correlations
seen in this study. Ultimately, quantifying the potential for allometric variation among
individuals is important for understanding the ecology and evolution of populations and species,
especially when that variation arises in fitness-related traits, such as growth in coral (Babcock
1991; Hughes, Ayre, and Connell 1992; Álvarez-Noriega et al. 2016).
2.5.2 Predictive Power of Initial Growth
Coral restoration can benefit from reliable predictors of key performance traits that can
be used to guide propagation and outplanting of corals with desired traits (van Oppen et al.
2015). We tested the predictive power of initial growth by relating it to growth in subsequent
time intervals. Initial growth was weakly correlated with subsequent growth in each
morphological trait (τ = 0.35–0.45). This is significantly higher than the predictive power of
growth rates measured in an in situ field nursery for growth post-outplant in A. cervicornis
(O’Donnell et al., 2018). This suggests that post-outplant growth, rather than nursery growth, can
more reliably project the performance of A. cervicornis transplants. However, the moderate
correlations seen here, which are similar to the correlations between initial and subsequent
growth previously reported for massive species (Edmunds and Putnam 2020), imply that colony
growth is determined by more than just intrinsic growth rate.
33
One random factor exerting a negative effect on growth in A. cervicornis is fragmentation, or
breakage. Breakage is a natural ecological process that can lead to asexual propagation
(Tunnicliffe 1981; Drury et al. 2019), partial mortality, or death (Madin et al. 2014). Breakage
was frequent in this study, with 54% of colonies experiencing at least one breakage event over
the 1-year monitoring window. The effects of breakage on the predictive power of initial growth
in TLE, SA, and Vinter were evident as correlations with subsequent growth improved once
broken colonies were removed from models. Interestingly, the exclusion of broken coral resulted
in no increase in the predictive power of initial growth in V, suggesting that measures of growth
in V may be robust to breakage. Initial growth in V was also the best predictor of subsequent
growth in TLE, SA, and Vinter when broken corals were included which may indicate a consistent
relationship between the growth in V and growth in other traits. This is similar to the findings of
Zawada et al. (2019) who showed that for a variety of coral morphotypes including branching
species, colony volume scaled consistently with many higher order traits as colony size increased
(i.e., limited allometric scaling).
Seasonality may also play a role in modulating growth rates over time as unexplained
variation in subsequent trait growth remained even after accounting for breakage. Morphological
traits that respond to environmental changes can result in differences in growth patterns over
time that further reduce the predictive power of initial growth. Our study spanned four seasons in
the Florida Keys with the first 6 months occurring during roughly Summer and Fall while the
second 6 months occurred during Winter and Spring. Seasonal variation in environmental
conditions like wave energy, which is dictated by wind speed in the Florida Keys (Ahn, Haas,
and Neary 2020), can result in fluctuating selection on different aspects of colony growth. For
example, high wave energy increased the number of smaller branches produced in A. cervicornis
34
colonies (Bottjer 1980), hypothetically increasing TLE, SA, and V but decreasing Vinter if new
branches did not extend beyond the initial branch length. Similarly, trade-offs between linear
growth and skeletal density were evident in A. cervicornis grown using two grow-out methods
(blocks vs. suspended trees) thought to modulate the amount of force exerted on colonies by
water movement (Kuffner et al. 2017). Environmental conditions can therefore cause differences
in how colonies grow and if environmental conditions are not constant over time, growth in one
time period will not be indicative of growth in another. Moreover, Edmunds and Putnam (2020)
also noted that the predictive power of initial growth varied by season, further confirming that
projections of future growth are complicated by temporal changes in environmental conditions.
2.5.3 Advantages of 3D Photogrammetry
It has been nearly 20 years since 3D models built from photographs were first used in the
study of coral (Bythell, Pan, and Lee 2001) and in that time, technological improvements have
made 3D photogrammetry a reliable method for exploring complex features of coral and coral
reefs. We show that small-scale colony-level 3D photogrammetry yields measures of TLE that
have a near-perfect correlation with measures of TLE made by-hand and this backward
compatibility allows for seamless transitions between methods. 3D photogrammetry could be
adopted at any time during the course of a study without the need to apply a correction factor.
Alternatively, high correlations between by-hand and 3D measures ensure that those unable to
adopt 3D techniques, such as small non-profit organizations, can continue to use traditional
methods while still being able to compare results with novel techniques. Measures of SA and V
from 3D models could not be directly compared to traditional measures due to the nature of this
experiment; however, previous efforts have confirmed the accuracy of 3D photogrammetry
(Figueira et al. 2015; House et al. 2018). In addition to facilitating meta-analysis through
35
backward compatibility, 3D photogrammetry also offers several advantages over traditional
measurements in terms of its accuracy and information rich digital record, and the ability to
obtain multiple trait measurements using a single method.
While the difference between 3D photogrammetry and traditional methods does increase
as the coral increases in size, this is likely due to error in by-hand measurements rather than error
in the models. Repeated, independent measurements of 3D models showed high precision, even
for the larger, more complex colonies. Moreover, we were able to identify errors that would have
otherwise gone undetected in the by-hand dataset through re-reviewing the extensive
photographic record. For example, two nearby colonies were wrongly identified due to missing
identification tags, but we were able to correctly re-identify these individuals using photographs
from previous time points. Additionally, although one colony was identified as an outlier for
growth in Vinter over the second 6 months, the photographic record allowed us to verify that this
was the result of true biological variation rather than sampling error (Fig. S2.8). 3D
photogrammetry also allowed us to measure individual colonies even after a dislodged colony
became intertwined with the focal colony (Fig. S2.9), which would have been impossible to
accurately measure by-hand. Despite the advantages of 3D photogrammetry, its application in
large scale projects remains hampered by hurdles in downstream image and model processing
procedures.
Beyond in situ image acquisition, 3D photogrammetry requires additional computational
effort which can limit the overall adoption of this technology. Model building software, such as
Agisoft Metashape, require large amounts of processing power and time. While the graphical
user interface of these programs can be straightforward, in the absence of a dedicated computer
3D model building can be difficult to upscale. Agisoft Metashape also has a command line-based
36
interface that can be run remotely on high-power computing (HPC) systems making model
building high-throughput for those with coding knowledge and access to such machines. We
sought to lower the accessibility barrier to high-throughput 3D photogrammetry by providing
directions and scripts for using Agisoft Metashape on remote HPC systems. This experiment
relied on a personally managed remote system but increasing availability of pay-as-you-go cloud
computing services are viable alternatives for researchers wanting to increase processing power
for 3D model building without the need to purchase and maintain a dedicated computing system.
2.5.4 Broader Applications
The advent of 3D photogrammetry has facilitated investigation of new traits without the
need for additional effort in the field or destructive sampling. Traits such as “compactness,”
“top-heaviness” (Zawada, Dornelas, and Madin 2019), and “proportion occupied” (Doszpot et al.
2019) quantify complex, ecologically relevant aspects of colony morphology that can easily be
accessed using digital 3D models. For example, Vinter quantifies the absolute amount of space
created between the branches of a coral, similar to shelter volume (Urbina-Barreto et al. 2021).
Interstitial space provides important habitat for reef fishes (Wilson et al. 2008; Noonan, Jones,
and Pratchett 2012; Urbina-Barreto et al. 2021) helping to maintain populations (Wilson et al.
2008; Graham 2014). The volume of interstitial space also influences the physical environment
of a colony, as it can result in pockets of reduced flow (Reidenbach et al. 2006). This affects
tissue boundary-layer conditions for heat transfer and pH buffering, which influence bleaching
and calcification rate (Chan et al. 2016; Stocking et al. 2018). Because Vinter exponentially
increases with TLE, the ecological functions associated with Vinter are also amplified by marginal
increases in linear extension. Additionally, large scale experiments addressing local adaption,
phenotypic plasticity, and trade-offs in these newly accessible traits are now feasible with high-
37
throughput 3D photogrammetry. Studies requiring large sample sizes, such as those exploring
genotype and environment relationships in A. cervicornis and A. palmata, could be implemented
immediately and would produce valuable information for restoration programs.
Here, we show that 3D photogrammetry can be applied efficiently to large sample sizes,
but we acknowledge that the colonies used in this study were small relative to the maximum size
A. cervicornis colonies can reach. However, the success in modeling complex structures like
those seen in Supplementary Figure 9 highlight the potential of 3D photogrammetry to capture
increasingly complex morphologies by simply taking more photographs in situ (Fig. S2.4).
Furthermore, the use of 3D photogrammetry in studies of densely branched species (Doszpot et
al. 2019) provides further confidence this technology can maintain accuracy while improving
trait accessibility for research and restoration.
Colony-level 3D photogrammetry could also serve restoration practitioners who seek
more detailed information on outplant performance. For example, the tracking of individuals can
identify reef sites or genotypes that cultivate large interstitial space crucial for fish habitat
(Wilson et al. 2008; Noonan, Jones, and Pratchett 2012; Urbina-Barreto et al. 2021) ultimately
helping to strategize restoration efforts. While reef-scape 3D photogrammetry (McKinnon et al.
2011; Burns et al. 2015; Leon et al. 2015) may prove more effective for tracking restoration
outcomes in terms of percent cover and ecosystem complexity, colony-level methods could be
used to ground-truth the lower resolution photogrammetry that practitioners employ over
hundreds of square meters to determine if accurate measures of colony-level phenotype can be
obtained from this coarser method. While 3D photogrammetry (at both large and fine-scales) has
yet to become a common tool in restoration, its potential for data collection, its methodological
simplicity in the field, and recently improved accessibility of post-processing methods have set
38
the stage for its utility. Taken together, colony-level 3D photogrammetry is a powerful tool for
investigating organismal processes such as phenotypic plasticity and ecosystem-level patterns
that could prove vital for restoration efforts.
39
2.6 Supplementary Material
Supplementary Data
The datasets generated and analyzed for this study can be found at
https://github.com/wyattmillion/Frontiers3Dmorphology.
Supplementary Figures
Figure S2.1: A group of ten coral fragments prepared for outplanting in the Mote Marine
Laboratory in situ nursery. The entire array was photographed and 3D models were built to
include all ten fragments in one model. A PVC frame was used to bring Agisoft Metashape
markers in line with coral fragments.
40
Figure S2.2: A) Convex hull (wire-like projection) overlaid over a coral model and B) convex
hull alone with coral removed viewed in MeshLab.
Figure S2.3: Example coral colony where breakage and partial mortality reduced the area of
living tissue to less than 5 cm
2
. V and Vinter were assumed to be 0 because a lack of photographic
coverage prevented model building and because of the limited living 3D structure present.
41
Figure S2.4: Relationship between the number of photos taken for each 3D model and the coral
size measured in TLE (A), SA (B), V (C), Vinter (D). (n=427). Blue lines represent a moving local
regression applied using method = loess in ggplot2. Shaded areas represent the 95% confidence
intervals.
Figure S2.5: Scatter plots showing growth in TLE measured in centimeters compared to growth
in SA in cm
2
(A, D), V in cm
3
(B, E), and Vinter in cm
3
(C, F) within as single time point (0.6 =
zero to 6 months; 6.12 = 6 to 12 months). Blue lines represent linear regressions over the
42
relationship between trait growth to show nonlinear relationships. Shaded areas represent 95%
confidence intervals.
Figure S2.6: Second order polynomials (solid blue line) fit over the predictive relationship
between initial and subsequent growth in TLE (A), SA (B), V (C), and Vinter (D) compared to the
linear model (dashed red line).
43
Figure S2.7: The relationship between initial growth in V compared to subsequent growth in
TLE (A), SA (B), V (C), and Vinter (D). Regressions were fit by whether a coral was broken
(pink) or not (turquoise). R values represent Kendall’s tau correlation coefficients and p values
represents significance of each relationship.
44
Figure S2.8: Photographs of colony EDR9 in immediately before outplanting (A) and 6 months
(B) and 12 months (C) post-outplant and the 3D model of EDR9 after 12 months post-outplant
viewed in MeshLab. This colony was identified as an outlier but was verified to be a result of
biological variation instead of methodological error.
45
Figure S2.9: View of the 3D model in MeshLab made with 151,132 tie points and 5,337,592
faces from 293 photographs. This represents two colonies intertwined after an unknown colony
(right) dislodged and landed next to a tagged individual used in the current study (left). The
unknown colony was removed manually to leave the focal colony able to be measured.
46
Chapter 3 Evidence of adaptive morphological plasticity in the endangered coral,
Acropora cervicornis.
Summary of contribution
This chapter represents an extension of an experiment conceptualized by Carly Kenkel, my PhD
advisor, and Cory Krediet of Eckerd College, which sought to identify molecular biomarkers for
reef restoration. Carly, Cory, Maria Ruggeri and Erich Bartels (Mote Marine Laboratory)
contributed to field work and image collection. Undergraduate Sibelle O’Donnell significantly
contributed to data collection. I led field work excursions, 3D model building, data collection,
data analysis and manuscript writing while all above collaborators contributed to the final
version. This chapter appears as a preprint at
https://www.biorxiv.org/content/10.1101/2022.03.04.483038v1 while under review at PNAS.
3.1 Abstract
Genotype-by-environment interactions (GxE) indicate that variation in organismal traits
cannot be explained by fixed effects of genetics or site-specific plastic responses alone. For
tropical coral reefs experiencing dramatic environmental change, identifying the contributions of
genotype, environment, and GxE on coral performance will be vital for both predicting
persistence and developing restoration strategies. We quantified the impacts of G, E, and GxE on
the morphology and survival of the endangered coral, A. cervicornis, through an in situ
transplant experiment exposing common garden (nursery) raised clones of ten genotypes to nine
reef sites in the Florida Keys. By fate-tracking outplants over one year with colony-level 3D
photogrammetry, we uncovered significant GxE on coral size and survivorship indicating that no
universal winner exists in terms of colony performance. Moreover, the presence of GxE also
implies the existence of intraspecific variation in phenotypic plasticity. Rather than differences in
mean trait values, we find that individual-level morphological plasticity is adaptive in that the
47
most plastic individuals also exhibited the fastest growth and highest survival. This indicates that
adaptive morphological plasticity may continue to evolve, influencing the success of A.
cervicornis and resulting reef communities in a changing climate. As focal reefs are active
restoration sites, the knowledge that variation in phenotype is an important predictor of
performance can be directly applied to restoration planning. Taken together, these results
establish A. cervicornis as a system for studying the eco-evolutionary dynamics of phenotypic
plasticity that also can inform genetic- and environment-based strategies for coral restoration.
3.2 Significance Statement
Phenotypic plasticity, or the ability of organisms to alter traits in response to
environmental change, can facilitate acclimation, expose or buffer genetic variation, and may
even evolve. Understanding how variation in plasticity among individuals impacts fitness will be
critical for determining its adaptive potential during climate change, particularly for organisms
incapable of behavioral escape from environmental extremes. Here, we show that
morphologically plastic Acropora cervicornis individuals had higher survival and growth rates
than less plastic individuals. Therefore, not only does plasticity vary in this species but selection
will be expected to favor its evolution. Considering the absence of a universal top grower or
survivor, our results suggest environmental responsiveness will be more useful for predicting
coral performance across reefs.
48
3.3 Introduction
Intraspecific variation in phenotype provides raw material for selection to act on resulting
in the evolution of trait means (Des Roches et al. 2021). However, trait values may also change
as individuals are exposed to different environments via phenotypic plasticity (West-Eberhard
2003). While plastic trait changes typically occur within a generation, they have the ability to
alter fitness-related traits and promote acclimation, and are therefore relevant for populations
experiencing new or stressful environmental conditions (Simpson 1953; West-Eberhard 1989;
Losos et al. 2000; Kelly 2019). Moreover, variation in the degree of plasticity can magnify
differences among individuals. In the light of intraspecific variation in plasticity, the evolution of
trait means becomes dependent not only on individual trait values but also the environments
those individuals face. Long-standing theory supports a role for plasticity in trait evolution
(Ghalambor et al. 2007; Moczek et al. 2011; Suzuki and Nijhout 2006) and the presence of
significant intraspecific variation in plasticity, i.e. genotype-by-environment interactions (GxE),
suggest that plasticity itself can also evolve (Gavrilets and Scheiner 1993; Scheiner 1993;
Pigliucci 2005).
The evolution of phenotypic plasticity, and consequently its ecological impacts, can
occur if variation in plasticity among individuals results in variation in fitness. Selection is
expected to increase plasticity when the benefits of producing an environment-specific
phenotype outweigh the fitness consequences arising from either the cost of altering that
phenotype, or from the production of an underdeveloped phenotype relative to locally adapted
individuals (Dewitt, Sih, and Wilson 1998; Hendry 2016). Evolutionary models suggest
plasticity will be favored in species with high dispersal that will experience predictably high
spatial or temporal environmental variation and when the costs of plasticity are low (Hendry
49
2016). However, limited empirical tests for an adaptive role of plasticity (Pigliucci 2005; Hendry
2016) provide inconsistent support for model predictions with variation evident among traits,
species, and environments (Van Buskirk and Steiner 2009; Hendry 2016; Velotta and Cheviron
2018). More experiments that quantify the fitness costs or benefits of plasticity, especially in
nonmodel systems, will improve our broad understanding of its ecological and evolutionary role
(Gavrilets and Scheiner 1993; West-Eberhard 2003) while also uncovering system-specific
functions contributing to acclimation to environmental change. This will be particularly
important for species of conservation concern, where persistence may be reliant on both adaptive
plasticity and the ability of human interventions to leverage it.
Reef-building corals form the base of the most biodiverse marine ecosystems, tropical
coral reefs (Reaka-Kudla, Wilson, and Wilson 1996; Knowlton et al. 2010). The ecological and
economic services they provide are determined in part by the complex three-dimensional
structures created by the corals themselves (Graham and Nash 2013; Monismith 2007; Lugo-
Fernández et al. 1998). This structure provides habitat space (Coker, Wilson, and Pratchett
2014), reduces wave energy (Lugo-Fernández et al. 1998), and sustains biological diversity and
productivity (Graham and Nash 2013; Szmant 1997) which support a multi-billion dollar tourism
industry (Cesar, Burke, and Pet-Soede 2003). However, these ecosystem services are being lost
as wild populations decline due to natural and anthropogenic factors (Pratchett, Hoey, and
Wilson 2014; Porter and Meier 1992). For example, populations of Acropora cervicornis, one of
two branching coral species once dominating Caribbean reefs, have declined precipitously since
the 1970s (Cramer et al. 2020) contributing to a loss of structural complexity (Roff, Joseph, and
Mumby 2020; Alvarez-Filip et al. 2011). This decline has prompted a global effort to understand
factors that promote coral survival and persistence under changing ocean conditions. Both
50
differences in fitness-related traits among common-gardened genotypes and of clones under
different conditions (Edmunds 1994; Cunning et al. 2021; Woesik et al. 2021; Drury, Manzello,
and Lirman 2017) suggests that some individuals or environments could be used to reestablish
the structure and function of reefs (Camp, Schoepf, and Suggett 2018; Palumbi et al. 2014;
Hoogenboom et al. 2017). However, as corals experience new conditions, via translocation
during reef restoration (Young, Schopmeyer, and Lirman 2012; Hoegh-Guldberg et al. 2008) or
through climate change (Hoegh-Guldberg et al. 2007; Manzello 2015), it is unclear whether top
performing genotypes will maintain their status (Barott et al. 2021; Morikawa and Palumbi 2019)
or if variation in plasticity will result in new ‘winners’ and ‘losers’ (Drury and Lirman 2021).
Therefore, clarifying the role of both the environment and genetic background on coral
performance as well as GxE will be critical for leveraging naturally occurring biological
variation for the restoration of degraded reefs.
Phenotypic plasticity has been commonly documented in coral morphology (Todd 2008;
Tambutté et al. 2015), physiology (Muller et al. 2021; Ziegler et al. 2014), and gene expression
(Bay and Palumbi 2015; Rocker et al. 2019) in response to a variety of abiotic factors, indicating
the environmental responsiveness of phenotypes, some of which is correlated with fitness-related
traits at the population level (Hoogenboom, Connolly, and Anthony 2008; Grottoli, Rodrigues,
and Palardy 2006; Kenkel and Matz 2016). Within-populations, genotype-by-environment
interactions have been reported less frequently (Drury and Lirman 2021; Bruno and Edmunds
1997; Barott et al. 2021) but the relationship between phenotypic plasticity and overall fitness
remains unresolved. This gap in knowledge limits both our understanding of the evolutionary
potential of plasticity and its role in the natural and human-assisted recovery of coral reefs in a
changing climate.
51
While variation in plasticity exists in wild coral populations (Bruno and Edmunds 1997;
Todd et al. 2004; Barott et al. 2021), it is unclear if such variation is maintained in restoration
corals, like A. cervicornis, that have been propagated in common-garden nurseries for decades.
Given that restoration involves outplanting clonal replicates (ramets) across environmentally-
diverse, natural reefs (Young, Schopmeyer, and Lirman 2012; Lohr et al. 2015), understanding
how individual phenotypes vary can help direct restoration efforts. Previous efforts found no
evidence of GxE in A. cervicornis morphology despite strong site and genotype effects on linear
growth and survival (Young, Schopmeyer, and Lirman 2012; Lohr et al. 2015; Drury, Manzello,
and Lirman 2017) but significant variation in bleaching responses among genotypes was
observed across natural reefs (Drury and Lirman 2021). GxE in fitness-related traits, like
bleaching tolerance, suggest that a multisite phenotype may more accurately describe A.
cervicornis performance in nature.
We rigorously fate-tracked 270 restored coral colonies on natural reefs in a multi-site
transplant experiment to test the effect of genotype, environment and their interaction on growth
rate, size, risk of fragmentation, and survival (Fig. 3.1A-C). We uncovered significant GxE in
survival and absolute size, measured with 3D photogrammetry, and identified relationships
between morphological plasticity, growth, and survival that support the presence of adaptive
plasticity in A. cervicornis. Taken together, our results establish a new system in which to
investigate the ecological and evolutionary impacts of adaptive plasticity while also informing
reef restoration.
52
Figure 3.1: Experimental design and environmental conditions A) Location of outplant sites
(Table S1) and the source restoration nursery in the lower Florida Keys, USA. Ten genets
(genotypes) of A. cervicornis were outplanted in triplicate to nine reef sites in three sub-arrays.
Sites are colored by average survival from dark blue (highest) to dark red (lowest) and the
location of site-specific SERC water quality monitoring stations (Briceño and Boyer 2019) are
indicated by the white diamonds. B) Ramets were reared in the nursery to a mean size of 8 cm
and were measured just prior to outplanting in April 2018 C) To obtain growth and morphology
data, still images were captured using underwater photography, which were used to build 3D
models in Agisoft Metashape, that were subsequently measured for total linear extension (TLE;
example red lines), surface area (SA), volume (V), and volume of the interstitial space (Vinter;
example shaded blue area) D) Average daily temperature of each site (colored by survival rank)
for the one-year experimental period. Inset shows hourly temperatures from July through
September 2018. The dashed line indicates the local bleaching threshold of 30.5°C. E) Principal
components analysis of historical SERC environmental metrics characterizing outplant reef sites
(Looe Key is excluded due to missing data) F) Results of a Bayesian negative binomial
generalized linear mixed effects model testing the association of eleven uncorrelated
environmental parameters on the change in Vinter. Horizontal black lines indicate 95% credible
intervals of the posterior distributions. Values above (red) or below (blue) indicate significant
association between the variable and the change in Vinter across sites.
53
3.4 Results
3.4.1 Ramet survival is a function of genotype, outplant site, and the interaction
Of the 270 outplanted ramets, 48 died and 24 were declared missing leaving 198
surviving ramets at the end of one year. Cox proportional hazard models showed significant
effects of genotype (p=0.002) and site (p=0.005) on survival. The model including the
interaction term, while improving fit, did not converge, so the additive model was used to obtain
risk scores for each genotype and site (Table S2-3). On average, Genotype (G) 36 had the highest
ramet survival (96%) while G41 had the lowest (61.5%) across sites, which incurred a mortality
risk 2.6 times that of G36 on average (Fig. 3.2A, Table S3.2). Three genotypes (G41, G62, G13)
form a group of high-risk genets (>2.3x higher mortality risk in comparison to G36) while
remaining genets display intermediate risk ranging from nearly equivalent to 1.7 times that of
G36 (Fig. 3.2A, Table S3.2). Ramets outplanted to Bahia Honda (60% survival) had a 2.5 times
greater risk of mortality than those outplanted to E. Sambo, the site with the highest survival on
average (96.4%, Fig. 3.2B). Similar to average genotype scores, sites exhibited a continuum of
increasing risk with mortality ranging from 1.1 to 2.5 times higher than the reference, E. Sambo,
with the highest mortality risk occurring at Bahia Honda (Table S3.3).
Pairwise correlations of genet survival ranks across sites show the identity of the best
surviving genet is not maintained across sites, with most correlations being close to 0 (Fig.
3.3C). The highest positive correlations (Pearson’s correlation = 0.54 to 0.70) were observed
among sites with intermediate survival on average (Big Pine, Dave’s Ledge, Looe Key, and
EDR, Fig. 3.2B) and not necessarily among geographically neighboring sites (Fig. 3.1A, S3.1).
54
3.4.2 Non-random fragmentation in A. cervicornis
We documented 177 instances of fragmentation throughout this experiment. Most events
occurred within the first three months post-outplant (84) followed by continually decreasing
occurrences in subsequent time periods. Cumulative linked models show a significant effect of
genotype and site on the likelihood of fragmentation (Table S3.4) with G1and Marker 32
experiencing the least amount of breakage among genotypes and sites, respectively (Fig. S3.2).
G44 and Bahia Honda experienced the most breakage among genotypes and sites, respectively,
during the outplant period. Of the 177 fragmentation events, only 15.8% occurred in the same
time period that the ramet died, indicating that breakage did not result in immediate mortality.
Fragmentation was not overly prevalent in larger size class colonies, and instead was more
common in colonies less than 5 cm and between 5-10 cm in length according to Fisher Exact
tests (Table S3.5).
Figure 3.2: Genotype, environment, and GxE patterns of survival. (A) Survival probability for
genets and (B) sites over the one-year experimental period. Genets and sites are colored by
decreasing overall survival from blue to red. (C) Pairwise correlations of genet survival rank
across outplant sites. Ellipse shape and color is proportional to the strength and direction of the
correlation between two sites. Sites are ordered according to survival as in (B).
3.4.3 Morphology exhibits plasticity that varies by genotype (GxE)
We assessed the fixed effects of genotype, outplant site, time, their interactions, as well
as initial size on the absolute size in four traits: total linear extension (TLE), surface area (SA),
55
volume (V), and volume of interstitial space (Vinter) (Table S3.6). Ramets experiencing
fragmentation were retained, and the site-specific array to which the ramet was outplanted and
the number of fragmentation events it experienced were included as random effects (Table S3.7).
Significant genotype-by-site interactions were detected in all four traits (p<0.001), whereas no
effects were detected for the genotype-by-time and genotype-by-site-by-time interactions in any
trait (p>0.05). Absolute size trajectories varied among genets and inconsistent genet rank order
was evident across sites, confirming the existence of GxE (Fig. 3.3, Fig. S3.3-3.5). Significant
effects of genotype, time, and initial ramet size were also evident for all traits (p<0.0001).
Absolute trait size increased over time and with the initial size of a ramet (Fig. 3.3, S3.3-3.5).
Ramets of G50 were the largest on average, while G13’s ramets were the smallest. The largest
ramet sizes were reached at EDR, a site with the third worst survival on average (Fig. 3.2B).
However, significant fixed effects of outplant site were not evident, while significant site-by-
time interactions were found in SA and V (p<0.01). Random effects of fragmentation
(p<0.0001) and array within site (p<0.0001) were also evident in all trait models (Table S3.7).
3.4.4 Growth rate is dependent on genotypic and environmental characteristics
Growth rate (unit/month) in each trait (TLE, SA, V, Vinter), was modeled as a fixed effect
of genotype, outplant site, time, and their interactions. Size at the beginning of each time interval
was included as a fixed effect in linear mixed models to account for potential size-specific
growth rates (Table S3.8). To accurately quantify growth rate, time periods where a ramet
experienced negative growth due to fragmentation were excluded from the mixed models,
following (Drury, Manzello, and Lirman 2017). Fragmentation and the site-specific array to
which the ramet was outplanted were included as random effects (Table S3.9) and despite being
removed from the analysis, a moderate effect of breakage persisted for V (p = 0.11). Genotype-
56
by-site and the 3-way interaction between genotype, site, and time were not significant for the
growth rate of any traits despite the significant GxE in absolute size. Significant genotype-by-
time and site-by-time interactions were detected in TLE, SA, Vinter (p<0.01) and TLE, V, and
Vinter (p<0.01), respectively. A significant fixed effect of genotype was present in TLE, V, and
Vinter (p<0.05). Growth rate in all traits exhibited a significant fixed effect of time (p<0.0001).
Growth rate increased with increasing initial size and time in all traits (p<0.0001, Fig. S3.6) but
standardized growth, calculated as growth rate per unit of existing tissue following (Lirman et al.
2014), decreased with size (Fig. S3.7). Fixed effects of outplant site were only evident in V
(p=0.03) while variation among arrays within sites was nonsignificant in V and SA despite
significant array effects on TLE and Vinter (p<0.05).
Figure 3.3: Average size in total linear extension (TLE) for each genotype (colored by survival
probability as in Fig. 2A) over time. Reef sites are also ordered by survival probability (left to
right).
57
3.4.5 The capacity for morphological plasticity is correlated with improved survival and growth
Genotype-by-site interactions in the absolute size of all traits indicates significant
variation in the capacity for morphological plasticity among genets. We quantified plasticity
using a joint regression analysis (Stinchcombe, Dorn, and Schmitt 2004) which integrates trait
data across multiple sites to provide a genotype-specific value of plasticity relative to the
population. We found consistent correlations between the degree of trait plasticity, overall
mortality risk, and mean growth rate that support an adaptive role (Fig. 3.4A). Plasticity in
absolute size in TLE, SA, V, and Vinter was only weakly associated with mortality risk at T3 (R=
-0.33 to 0.14, p = 0.14 to 0.87, Fig. S3.8). However, negative relationships, indicating reduced
mortality risk in more plastic genotypes, strengthened over time with significant relationships in
SA and V evident after 9 months (R = -0.65, p = 0.042 and R = -0.82, p = 0.004, respectively),
and in TLE, SA, and V after 12 months (R = -0.69, p = 0.029; R = -0.71, p = 0.02; and R = -0.67,
p = 0.033; Fig. S3.8). Similarly, relationships between average growth rate and trait plasticity
began as neutral/weakly positive and progressed to strong positive correlations over time, with
the strongest correlations found after 12 months (R= 0.81 to 0.85, p= 0.0019 to 0.0043, Fig.
S3.9). Growth rate tended to be negatively correlated with mortality risk, indicating that higher
growth rates were associated with decreased mortality risk, although significant relationships
were only observed at later time points (V, 9 months, R= -0.7, p=0.023; TLE, 12 mos, R=-0.66,
p=0.037; SA, 12 mos, R=-0.66, p=0.039, Fig. S3.10).
As a way to assess changes in colony shape regardless of size, we calculated SA-to-V
ratio, TLE-to-V ratio, packing, convexity, and sphericity (Zawada, Dornelas, and Madin 2019)
for ramets surviving to T12 with no fragmentation. This subset represents ramets occupying a
morpho-space enabling survival and growth uninterrupted by breakage. When plotted in
58
multivariate trait space, ramets did not cluster by genet or site (Fig. 3.4B, S3.11). However,
ramets of genets showing higher average survival occupy a broader area, or trait space, compared
to genets with lower survival (Fig. 3.4B).
Figure 3.4: Morphological plasticity and its relationship to growth and survival. A) Relationship
between genet plasticity in absolute size and average genet growth rate over the final 3 months of
the outplant period. Points are colored by genet mortality risk score. Line and shaded region
show line of best fit and 95% confidence interval for each relationship. B) Principal components
analysis of size-independent morphological traits: sphericity, convexity, packing (Zawada,
Dornelas, and Madin 2019), and SA-to-V ratio and SA-to-TLE ratio (gray vectors labeled in
red). Points represent individual ramets colored by genet identity (n=5-12 ramets per genet) by
decreasing survival from blue to red. Shaded regions (colored by survival) frame the most
extraneous ramets for each genet and outline the morphospace occupied by a genet.
3.4.6 Variation among offshore reefs may contribute to site-specific growth and survival
Benthic temperatures were recorded hourly at all sites for the one-year experimental
period except for Looe Key where data from April 2018 to October 2018 are missing due to a
flooded logger (Fig. S3.12). Therefore, the temperature profiles of 8 sites were used to analyze
thermal differences among reefs. Water temperatures over the experimental period were similar
59
between sites (Fig. 3.1D, Table S3.10). Annual mean temperatures varied by 0.213
o
C between
the warmest (EDR) and coolest (Bahia Honda) sites. Annual temperature ranges varied from
10.12
o
C (E. Sambo) to 12.26
o
C (Bahia Honda) while average daily temperature ranges varied
from 0.59
o
C (Marker 32) to 0.71
o
C (Bahia Honda). Big Pine experienced the most days where
the temperature was at or above 30.5°C (69) while Bahia Honda experienced the fewest (54).
Interestingly, Big Pine and Bahia Honda both had the most days above 32°C (3) while the
majority of sites never reached this temperature throughout the entire experimental period.
Summer thermal predictability, calculated as the sum of positive temporal autocorrelation from
July through September, was highest at W. Sambo and lowest at Bahia Honda. The three sites
with the highest survival probability had the three highest thermal predictability values.
Site-specific biogeochemical parameters, obtained from the long-term SERC water
quality monitoring program (Briceño and Boyer 2019), did not differ when restricting the dataset
to the experimental period. However, analysis of the full historical dataset (1995-2019) revealed
significant differences in nitrate and silica dioxide concentrations among the 9 reef sites (p<0.05,
Fig S3.11). A principal components analysis of all thermal and water quality parameters showed
large aggregate differences between sites despite this limited variation in individual parameters
(Fig. 3.1E, S3.13, S3.14). Sites with the highest survival (E. Sambo, Marker 32, and W. Sambo)
clustered together while the remaining sites were broadly distributed. These high survival sites
were also associated with high thermal predictability and historical high average light attenuation
(Fig. S3.14). Bahia Honda, the site with the lowest survival and growth, consistently stood out as
the most extreme point (Fig. 3.1E, S3.13, S3.14).
After removing highly correlated water quality metrics, eleven parameters were used as
candidate variables in a Bayesian negative binomial generalized linear model to assess their
60
power to predict changes in coral morphology and mortality risk. Average temperature was
significantly negatively associated with the change in TLE, SA, V, and Vinter (Fig. 3.1F, Fig.
S3.15-3.18) while average daily temperature range was negatively associated with only Vinter.
Days above 30°C was significantly positively associated with change in Vinter but no patterns
were evident for other traits (Fig. S3.15-3.18). Risk score was not significantly associated with
any of the environmental parameters (Fig. S3.19).
3.5 Discussion
The presence of a significant GxE interaction indicates that individuals differ in their
sensitivity to environmental variation. This variation in reaction norm slope among individuals
can reduce prediction accuracy and confound selective breeding programs (Ly et al. 2013). Here,
we identify significant GxE in both the size and ultimate survival of restoration lines of A.
cervicornis indicating that no single genet ‘wins’ in all contexts when considering the change in
mean trait values across environments. However, the existence of GxE also means that there is
genetic variation in the capacity for plasticity, or the degree of environmental responsiveness of
individual genotypes. Rather than mean size, we find that this plasticity, or the degree to which a
genet is able to change its size relative to the population mean across sites, is positively
correlated with mean growth rate and survival. This suggests that plasticity may continue to
evolve, although context-dependent trade-offs and the capacity to predict environmental
variation will likely influence the ultimate trajectory. Below we consider potential drivers and
implications of this adaptive plasticity for eco-evolutionary dynamics as well as its applied
relevance for the conservation and restoration of reefs.
61
3.5.1 Adaptive phenotypic plasticity in coral
Phenotypic plasticity can facilitate acclimation over space and time by allowing
organisms to match local phenotypic optima (Via et al. 1995; Kelly 2019). In reef-building
corals, plastic responses in morphology (Todd 2008; Muko et al. 2000), bleaching (Drury and
Lirman 2021), gene expression (Kenkel and Matz 2016), and gene body methylation (Dixon et
al. 2018) alter colony phenotype in ways that are hypothesized to be beneficial, especially in the
context of a stress response. In the absence of other performance data, however, it was unclear
whether such individual plasticity would result in a net fitness benefit, particularly if significant
costs were incurred by plastic individuals in non-stress contexts. Here, by explicitly linking
morphological plasticity with growth and survival we show that an individual’s ability to be
plastic yields a net positive fitness outcome in A. cervicornis. Specifically, increased plasticity
was associated with both a higher average growth rate and an increased probability of survival
(decreased mortality risk, Fig. 3.4A) strongly supporting an adaptive role.
Coral morphology is an important environmental interface as changes in size and shape
can adjust the flow, temperature, and pH in and around a colony (Stocking et al. 2018; Chan et
al. 2016) affecting both normal processes, such as nutrient uptake, as well as responses to
thermal or acidification stress (Dennison and Barnes 1988; Lesser et al. 1994; Jimenez et al.
2011). In the Florida Keys, spatially or temporally variable conditions, such as temperature,
select for local phenotypic optima (Kenkel, Almanza, and Matz 2015; Manzello et al. 2019) that
appear to promote the overall persistence of plastic A. cervicornis genotypes. Although sites did
not appear to select for specific morphologies in terms of their clustering patterns, genets with
the highest average survival occupied a broader morphospace (Fig. 3.4B), again supporting an
adaptive role for morphological plasticity.
62
Coral meet many of the conditions predicted to favor the evolution of increased plasticity
(Hendry 2016). Long generation times and high habitat heterogeneity across dispersal ranges
expand the range of environments individuals experience within their lifetime (Drury et al. 2018;
Hemond and Vollmer 2010). Assuming a genetic basis to plasticity, positive selection in the
form of increased survival and/or reproduction of more plastic individuals can enable genetic
accommodation, making future generations more plastic (Kelly 2019). This evolution of
plasticity may be modulated by strong selection events occurring in A. cervicornis and other reef
building coral in the form of disease or bleaching events (Manzello 2015; Williams and Miller
2005), although the magnitude and direction of the effects will depend on the relationship
between morphological plasticity and the response to stress. We did not detect any tradeoffs
between plasticity, growth rate and survival in ambient conditions, but exploring additional costs
or limits to morphological plasticity will be an important next step in understanding the future
adaptive potential of plasticity, particularly in the face of new environmental extremes. Warming
ocean temperatures are an ever present threat to coral reefs and tradeoffs between plasticity and
bleaching tolerance or recovery would severely limit the benefits of morphological plasticity as
bleaching events become more frequent and severe (Hoegh-Guldberg et al. 2007; Manzello
2015).
3.5.2 Genotype-environment interactions limit survival and growth predictions
Survival and growth are two metrics commonly used to evaluate coral fitness because
improved survival will positively impact population demographic rates (Babcock 1991; Hughes,
Ayre, and Connell 1992; Álvarez-Noriega et al. 2016) while faster growth can shorten the time
to reproductive maturity (Hall and Hughes 1996). Although lifetime reproductive success is
difficult to measure in annual broadcast spawning species such as coral, greater reproductive
63
capacity has been observed in larger colonies (Hall and Hughes 1996). Significant variation in
the Cox Proportional Hazard risk scores indicated that some genotypes (G41, G62, G13) were at
greater risk of mortality than others. Similarly, reef sites also varied in their ability to support
coral survival. These results align with previous findings of strong genotype and site effects on
coral survival (Woesik et al. 2021; Lirman et al. 2014) and suggest that certain genets (G36) and
reef sites (E. Sambo) may be higher quality overall. When looking at the remaining genets and
sites, however, genotype-by-site interactions for survival and lack of preservation in the rank
order of genet survivorship between reef sites (Fig. 3.2C) indicate a limited ability to predict
outcomes based on knowledge of genotypes or sites in isolation. This stands in contrast to a
similar in situ transplant experiment by Drury et al. (2017) that found site effects but no effect of
genotype or the interaction on mortality. However, corals in this prior study endured a thermal
bleaching event and variation in mortality was attributed to variation in bleaching among reef
sites (Drury, Manzello, and Lirman 2017). While no bleaching was observed during the course
of our experiment, our results indicate that under normal conditions, genet survivorship from a
single site alone does not predict survival in other, even geographically neighboring, reefs.
Similarly, coral morphology appears to be influenced by a complex set of interacting
factors that ultimately preclude identification of a globally top performing individual or ‘super
coral’ (Camp, Schoepf, and Suggett 2018). Genotype-by-site interactions were evident in
absolute size which resulted in a different collection of genets representing the largest relative
coral at each site after the 12-month outplant period (Fig. 3.3). This is the first evidence of GxE
in A. cervicornis morphology, although this finding builds on earlier work in other branching
coral species (Bruno and Edmunds 1997; Doszpot et al. 2019; Todd et al. 2004). It is important
to note that statistical models included corals that experienced fragmentation, as this is an
64
ecologically relevant phenomenon contributing to clonal reproduction (Drury et al. 2019). When
we excluded fragmentation, significant genotype-by-site interactions disappeared. Similarly,
(Drury, Manzello, and Lirman 2017) excluded negative growth and showed independent impacts
of genotype and environment on TLE in A. cervicornis, but no GxE was detected. Significant
fixed effects of genotype on both absolute size and growth rate for TLE, SA, V, and Vinter
suggest that the intrinsic growth rate does vary among genets. However, this capacity may be
limited by fragmentation, which also varies as a function of genotype and site. Ultimately
however, population demographic and restoration success are based on size of coral colonies
rather than their growth rate, indicating that GxE must be accounted for when developing
conservation and restoration strategies.
Unlike earlier studies in both A. cervicornis (Drury, Manzello, and Lirman 2017) and its
sister species, A. palmata (Kuffner et al. 2020), we find no effect of site on the majority of
morphological and growth traits despite strong site effects on survival. Once thought of as noise,
GxE allows for the presence of phenotypic variation among individuals in the absence of overall
site effects and may provide a proximate explanation for their absence here. Alternatively,
differences in experimental design may also play a role. Kuffner et al. (2020) compared ramets
outplanted over a large spatial scale (>300 km) compared to the ~60 km span covered in the
present study. Similarly, Drury et al. (2017) selected reef sites spanning inshore and offshore
zones. Although reefs within the offshore zone vary in environmental conditions (Briceño and
Boyer 2019) larger differences are evident between inshore and offshore reefs in temperature,
turbidity, and nutrients (Lirman and Fong 2007; Briceño and Boyer 2019) may have driven the
pronounced site effects on morphology reported earlier (Drury, Manzello, and Lirman 2017).
Finally, temporal variability in environmental stress can create periods of poor growth or high
65
fragmentation followed by periods of recovery that may mask mean site effects and instead
generate a site-by-time signal. We did observe significant site-by-time effects on the majority of
morphological traits indicating that site effects are likely to be time dependent.
3.5.3 The importance of fragmentation
Fragmentation is a vital part of the evolutionary ecology of A. cervicornis as it can significantly
alter the demographic trajectory of populations through asexual propagation (Tunnicliffe 1981;
Drury et al. 2019), partial mortality, or death (Madin et al. 2014). In this experiment numerous
instances of breakage had obvious negative impacts on colony size but were mostly nonfatal.
This suggests that fragmentation can also impact growth beyond the immediate response to
injury, such as increased productivity due to size-dependent growth rates that occur after size
reduction (Lirman et al. 2014). While sometimes considered random, fragmentation in this study
occurred significantly more in some genets and sites. Differences in calcification rate among A.
cervicornis genets (Kuffner et al. 2017; Lohr and Patterson 2017) mean that certain individuals
can produce more dense skeletons faster, potentially making them less prone to breakage.
Moreover, calcification is energetically expensive (Cohen and McConnaughey 2003) and
apparent trade-offs in skeletal density and colony size (Kuffner et al. 2017; Lohr and Patterson
2017) suggest different strategies for skeletal growth in this species that may lead to variation in
the ability to withstand physical stress leading to breakage.
Spatial and temporal variation in hydrodynamic energy (Leichter, Deane, and Stokes
2005; Ahn, Haas, and Neary 2020), likely also imposes variable mechanical stress on coral
colonies. Coral morphology may respond to these conditions (Jokiel, Jury, and Kuffner 2016;
Todd 2008), but sudden or especially strong hydrodynamic forces are common sources of
damage for branching corals in the Caribbean (Woodley et al. 1981; Perkins and Enos 1968).
66
Human activity may also have contributed to fragmentation and anecdotally, higher tourist
activity was observed at Looe Key and EDR. Although fragmentation was usually nonfatal in the
focal ramet, we did not track the fate of newly generated fragments precluding determination of
the ultimate effect on fitness. Regardless, the existence of non-random fragmentation reinforces
the notion that accounting for fragmentation, rather than treating it as experimental error, will be
important for accurately predicting changes in branching coral morphology and performance.
3.5.4 Multivariate environmental conditions distinguish reefs
The offshore reef sites used here are in an area that has historically been treated as a
single environmental unit (Briceño and Boyer 2019). However, site specific variation in coral
performance (Fig. 3.2B, Fig. 3.3) and in environmental condition (Fig. 3.1E) supports the need
for a more nuanced approach. Aggregate differences in abiotic conditions among sites with high
average survival appear to be defined by high nitrogen concentrations, thermal predictability,
light attenuation/turbidity, and low annual and average daily temperature ranges (Fig. S3.13-
3.14). Similarly, the site with the lowest average survival, Bahia Honda, differentiated by
historically high ranges of nitrite and total phosphorus concentrations, turbidity, and light
attenuation (Fig. S3.14). This site also had the highest annual and daily temperature ranges, but
lowest summer thermal predictability during the experimental period. Although no physical or
chemical condition was individually correlated with mortality in the Bayesian models,
fluctuating environmental conditions have been implicated in the conditioning of marine species
to climate change (Kroeker et al. 2020; Ziegler et al. 2021). While temperature variability can
enhance coral tolerance (Safaie et al. 2018; Oliver and Palumbi 2011), the predictability of those
fluctuations should also impact the ability to acclimate and adapt (Bitter et al. 2021; Reed et al.
2010). Thermal predictability was highest among the three sites with the highest survival (E.
67
Sambo, Marker 32, and W. Sambo) and lowest at Bahia Honda yet no correlations were detected,
suggesting the importance of a multivariate approach. The low temporal resolution of water
quality metrics precluded obtaining similar measures of predictability in water chemistry during
the experimental period. Future work quantifying environmental predictability in addition to
fluctuations may yield additional insight into the conditions that support coral performance and
plasticity.
Mean temperature was negatively associated with change in size of all morphological
traits (Fig. 3.1F, S3.15-3.18), suggesting that cooler conditions promoted ramet growth. This is
unsurprising considering the well documented negative impact of high temperatures on coral
performance (Brown 1997) and the fact that mean experimental temperatures were within or
above the apparent optimum thermal range (ca 25-29°C) for A. cervicornis (Paradis, Henry, and
Chadwick 2019; Gladfelter 1984). Interestingly, the number of days above 30.5°C seemed to
encourage growth in Vinter but no other trait. Morphology has been shown to impact flow around
a colony, altering heat flux at the coral surface (Stocking et al. 2018; Jimenez et al. 2011) with
branching morphologies more capable of offloading heat compared to mounding coral (Jimenez
et al. 2008). As A. cervicornis ramets become less compact by increasing the volume of their
interstitial space (Fig. 3.1C), heat dissipation at the coral surface can also be expected to increase
through a reduction in the thermal boundary layer (Jimenez et al. 2008, 2011). While the
offshore reef tract of the Florida Keys is typically considered environmentally contiguous, taken
together these results suggest spatial variation in reef conditions independently and cumulatively
shape coral performance.
68
3.5.5 Conclusions
Current coral restoration strategies rely on transplanting clones across reefs varying in
abiotic and biotic conditions (Young, Schopmeyer, and Lirman 2012; Lohr et al. 2015; Drury,
Manzello, and Lirman 2017) suggesting that plasticity will play a role in the success or failure of
individual colonies. Adaptive morphological plasticity in A. cervicornis may enable genets to
maintain fitness in response to changes in environmental conditions over time or space.
Continued positive selection on intraspecific variation in plasticity, contingent on its freedom
from tradeoffs, should promote the evolution of plasticity and therefore the acclimatory benefits
associated with it. Environmental conditions can also promote or constrain the evolution of
phenotypic plasticity (Bitter et al. 2021; Gavrilets and Scheiner 1993) and while there appears to
be sufficient variation within the A. cervicornis habitat range to induce plasticity at present, its
relative benefits will likely also be dependent on the ability to predict environmental fluctuations,
which may prove challenging in the face of continued climate change.
3.6 Methods
3.6.1 Experimental design
Ten coral genets maintained long-term (5+ years) at Mote Marine Laboratory’s in situ
coral nursery (Table S3.10) were outplanted in a multi-site transplant study under FKMNS
permits 2015-163-A1 and 2018-035. In April 2018, 270 coral (mean TLE of 8.4 cm) ramets
representing 10 genets (27 ramets per genet) affixed to concrete pucks were photographed
following (Million et al. 2021) and manually measured for TLE immediately before
transplantation to nine active restoration sites (Table S3.1, Fig. 3.1). Three ramets per genet
(n=270 fragments) were randomly outplanted at each site with one ramet allocated to each of
69
three ten-coral arrays (Fig. 3.1A). Coral pucks were attached to bare reef substrate using marine
epoxy over 4 days, from April 21 to April 25.
Outplant sites were resurveyed in July 2018, October 2018, January 2019, and April
2019. Ramets were individually re-photographed and measured by-hand at the first four time-
points for TLE. Survivorship was recorded during site surveys and later confirmed with
photographs. Breakage was recorded via the photographic time series and through negative
growth measures in the resulting trait dataset (Supplemental Methods). Ramets where both the
coral tissue and ceramic puck were missing indicated technical failure of the marine epoxy rather
than a true biological loss and were excluded from subsequent analyses.
3.6.2 Phenotyping
Photographs taken in situ were used to generate individual 3D models of each coral ramet
in Metashape 1.5.4 (Agisoft LLC, St. Petersburg, Russia) using a high-throughput pipeline
(Million et al. 2021). Specifications for model building and all scripts can be found at
https://github.com/wyattmillion/Coral3DPhotogram. 3D models were imported into Meshlab
v2020.6 (Cignoni et al. 2008) to measure four growth-related traits following protocols described
in Million et al. (2021) and detailed in the Supplementary Methods: TLE, SA, V, and Vinter. We
assessed the final shape of colonies that survived to T12 with no detectable fragmentation by
calculating SA-to-V and TLE-to-V ratios, in addition to packing, convexity, and sphericity
(Zawada, Dornelas, and Madin 2019). Among this subset, genets were more equally represented
(n=5-12 ramets per genets) than sites (n=2-18 ramets per site). These traits were used in a
principal components analysis to determine how ramets clustered in morphospace as a function
of genet and site.
70
3.6.3 Environmental data
All reef sites are located along the offshore reef tract of the Lower Florida Keys at a
depth of 5.6m to 9.1m. HOBO Pendant Temperature loggers (Onset Computer Corp.; Bourne,
MA), set to record hourly, were attached to the reef substrate directly adjacent to the outplanted
corals, and exchanged with new loggers on subsequent site visits. Hourly temperature records
were used to calculate annual mean, annual range, average daily range, maximum monthly mean,
days and hours above 30.5°C or 32°C, and summer thermal predictability. Thermal predictability
was quantified for July through September only as highly variable temperatures at or above the
bleaching threshold of 30.5°C are expected during this window (Manzello, Berkelmans, and
Hendee 2007). Predictability was calculated as the sum of autocorrelation over a series of lags
until autocorrelation reached zero, i.e. the point at which current temperatures are no longer
informative of future conditions (Walsh, Shapiro, and Shy 2005; Li and Ding 2013). Quarterly
concentrations of benthic nitrite, nitrate, ammonia, dissolved organic and inorganic nitrogen,
soluble reactive phosphorus, total phosphorus, total nitrogen, N:P ratio, silicate, dissolved
oxygen, total organic carbon, as well as turbidity and light attenuation for each outplant site were
obtained from the Southeast Environmental Research Center (SERC) water quality dataset
(Florida International University) associated with each site (Fig. 3.1A,
serc.fiu.edu/wqmnetwork/FKNMS-CD/DataDL.htm).
3.6.4 Statistical analysis
All statistical analyses were performed in R version 3.6.3 (R Core Team 2020) and
scripts can be found at github.com/wyattmillion/Acer_Morphological_Plasticity. Cox
Proportional Hazard models were fitted to outplant survival data using the coxme (Therneau
2020) and survival (Therneau et al. 2022) packages. Consistency in genet rank order across sites
71
was quantified with Pearson’s correlations. Cumulative Linked Mixed Models assessing the
ordinal response of fragmentation were implemented with the package ordinal (Christensen
2019) to test for effects of genotype and outplant site on cumulative breakage events summed
within a ramet over time. Fisher Exact Tests were used to determine enrichment of fragmentation
among A. cervicornis size classes.
Effects of genotype, site, time, all associated interactions, and initial size on colony
morphology (~size) and growth rate in TLE, SA, V, and Vinter were tested with linear mixed
effects models implemented with the package lmer (Bates et al. 2022). Fragmentation and
outplant array nested within-site were included as random effects. When calculating growth rate,
ramets experiencing fragmentation (evidenced by a negative growth rate for a ramet over a 3-
month interval) were removed from the dataset and replaced with NA for the time-point in which
fragmentation occurred.
The extent of plasticity was quantified across each time interval using the joint regression
framework (Stinchcombe, Dorn, and Schmitt 2004; Yates and Cochran 1938; Dorn, Pyle, and
Schmitt 2000) where genet mean sizes at each site were regressed against the site-wide mean.
Regression coefficients representing plasticity in TLE, SA, V, and Vinter were correlated with
Cox mortality risk scores and average genet growth rates using Pearson Correlations.
SERC water quality parameters and thermal characteristics were used to describe
environmental variation among reef sites. We calculated both the overall mean and the range of
benthic nutrient concentrations, turbidity, and light attenuation over the entire length of the
SERC dataset (Spring 1995 to Spring 2019) and over the experimental period (April 2018 to
2019). The historical and contemporary SERC data were used to identify differences between
reef sites along with high resolution temperature data collected during the experimental period.
72
An analysis of variance was used to identify significant differences among sites for independent
parameters. A principal components analysis was used to explain variation among sites using all
parameters simultaneously.
Bayesian negative binomial generalized linear models implemented in R2jags (Su and
Yajima 2021a) were used to test for the impact of environmental parameters on the growth and
survival of ramets at reef sites. Model power was improved by using size, risk score, and
environmental data over each of the four time intervals to increase sample size and by removing
highly correlated environmental variables.
3.7 Supplementary Material
3.7.1 Supplemental Methods
Ten coral genets maintained long-term (5+ years) at the Mote Marine Lab Summerland
Key in situ coral nursery (Table S1 for coordinates) were selected for use in a multi-site
transplant study in partnership with Mote Marine Lab’s Staghorn Restoration Program under
FKMNS permits 2015-163-A1 and 2018-035. On March 19, 2018, 300 coral (mean TLE of 8.4
cm) ramets representing 10 genets (27 ramets per genet) were transferred from hanging trees,
fixed to ceramic pucks with marine epoxy and attached to benthic structures in preparation for
outplanting. In April 2018, each ramet was photographed for 3D photogrammetry following the
protocol described in Million et al. 2021 and manually measured for total linear extension (TLE)
immediately before transplantation to nine established restoration sites (Table S3.1 for
coordinates) in the Lower Florida Keys. Sites were chosen based on their overlap with actively
permitted A. cervicornis restoration activities and proximity to permanent Southeast
Environmental Research Center (SERC) Water Quality Monitoring stations (SERC, Florida
International University). Three ramets per genet (n=270 fragments) were randomly outplanted
73
at each site with one ramet allocated to each of three ten-coral arrays. Arrays were spaced
approximately 1-m apart and the subsite locations were selected based on substrate type,
prioritizing hard bottom. Coral pucks were attached to bare reef substrate using marine epoxy
leaving approximately 10 cm between each puck. Outplanting occurred over 4 days due to
weather and time constraints, from April 21 to April 25.
Outplant sites were resurveyed every three months for a total of four time-points
throughout a one-year period, in July 2018, October 2018, January 2019, and April 2019. Ramets
were individually photographed for 3D photogrammetry-based reconstruction of growth metrics
and total linear extension was measured by-hand for the first four time-points. Survivorship was
recorded during site surveys and confirmed with photographs post-hoc. Ramets were recorded as
alive if any amount of living tissue was visible. Breakage was recorded both via the photographic
time series and through negative growth measures in the resulting trait dataset. Ramets where
both the coral tissue and ceramic puck were missing were discarded from the analyses as we
attribute these losses to technical failure of the marine epoxy upon outplanting rather than true
biological mortality.
Phenotyping
Photographs taken in situ were used to generate individual 3D models of each coral ramet
in Metashape 1.5.4 (Agisoft LLC, St. Petersburg, Russia) using the high-throughput pipeline
described in Million et al. (2021). Models were built on Dell PowerEdge R910 with an Intel®
Xeon® Processor E7-4850 with Metashape manually limited to 20-40 CPUs and 250 GB of
RAM. Specifications for model building and all scripts can be found at
https://github.com/wyattmillion/Coral3DPhotogram. 3D models in Wavefront format (.obj) from
Metashape were imported into Meshlab v2020.6 (Cignoni et al. 2008) to measure four growth-
74
related traits following protocols described in Million et al. (2021): total linear extension (TLE,
(Johnson et al. 2011)), surface area (SA), volume (V), and volume of the convex hull. The
volume of the convex hull represents the smallest sized, convex mesh that encompasses all
points of an object and is used to calculate the volume of interstitial space (Vinter) for each coral
colony by subtracting the volume of the colony from the volume of the convex hull.
Environmental data
HOBO Pendant Temperature loggers (Onset Computer Corp., Bourne, MA) were affixed
to the reef substrate with plastic zip ties and set to record hourly for six-month periods and were
redeployed on subsequent site visits. SERC has monitored water quality quarterly at a network of
270 sites throughout south Florida and the Florida Keys beginning in 1995 (Briceño and Boyer
2019). Concentrations of benthic nitrite, nitrate, ammonia, dissolved organic and inorganic
nitrogen, soluble reactive phosphorus, total phosphorus, total nitrogen, N:P ratio, silicate,
dissolved oxygen, total organic carbon, turbidity, and light attenuation for each monitoring
station associated with an outplant site were obtained from the publicly available SERC database
(SERC, Florida International University). Depth of reef sites range from 5.6m to 9.1m.
Statistical analysis
All statistical analyses were performed in R version 3.6.3 (R Core Team 2020). Raw data
and scripts can be found at github.com/wyattmillion/Acer_Morphological_Plasticity. Cox
Proportional Hazard models were fitted to outplant survival data using coxme (Therneau 2020)
and survival (Therneau et al. 2022) packages in order to test for effects of genotype,
environment, and genotype-by-environment interactions. Effects of fragmentation were included
in mixed effects Cox models by accounting for the cumulative number of breakage events
experienced by each ramet. When models including the interaction of genotype and outplant site
75
did not converge, an additive model was used to reliably estimate coefficients for fixed effects.
Consistency in genet rank, ordered 1 to 10 by increasing mortality risk, across sites was
quantified with Pearson’s correlations. Cumulative Linked Mixed Models were applied with the
package ordinal (Christensen 2019) to test for effects of genotype and outplant site on the
ordinal response variable of cumulative breakage events summed within a ramet across time.
Effects of size preceding the presence/absence of fragmentation was tested for using a binomial
logistic regression.
Effects of genotype, environment, and genotype-by-environment interactions on colony
morphology (~size) and growth rate were tested with linear mixed effects models in the package
lmer (Bates et al. 2014). Morphological change was assessed in two ways for each trait. Both the
absolute size and the monthly growth rate (determined for each 3-month time interval) were
calculated for TLE, SA, V, and Vinter. Trait values were either square-root or log(x+n)
transformed to meet assumptions of normality and account for zero-values. Models for absolute
size included genotype, outplant site, time point, all two and three-way interactions and initial
size as fixed effects. Fragmentation and outplant array nested within site were included as
random effects. To accurately quantify growth rates, ramets experiencing fragmentation
(evidenced by a negative growth rate for a ramet over a 3-month interval) were removed from
the dataset and replaced with NA for only the time-point in which fragmentation occurred. With
this reduced data set, we assessed the fixed effects of genotype, site, time point, all associated
interactions, and ramet size at the start of each 3-month interval on growth rate in each trait,
including a random effect of array nested within site.
Correlations between the extent of plasticity and hazard ratio (mortality risk) for each
genet were used to address the adaptive role of phenotypic plasticity. The extent of plasticity was
76
quantified using the joint regression framework (Stinchcombe, Dorn, and Schmitt 2004; Yates
and Cochran 1938) in which genet mean trait values in each site are regressed against the
environmental mean across all genets at each site. The resulting coefficient gives genet plasticity
relative to the population; a coefficient >1 indicates enhanced plastic response in a genet relative
to the population while a coefficient <1 suggests less responsiveness. Plasticity measured
through joint regression, as compared to traditional two-environment approaches (Dorn, Pyle,
and Schmitt 2000), incorporate data over multiple environments and generate a unitless metric
comparable across traits. Coefficients representing plasticity in TLE, surface area, volume, and
volume of interstitial space in addition to trait means were used as covariates in a linear model to
explain variation in genet hazard ratios estimated from Cox models. This process is commonly
used to estimate selection on plasticity in plant systems (Stinchcombe, Dorn, and Schmitt 2004;
Rausher 1992; Tiffin and Rausher 1999) as it controls for covariance between independent
variables.
SERC water quality parameters and thermal characteristics were used to describe
environmental variation among reef sites. We calculated both the mean and the average annual
range of benthic μM concentrations of nitrite, nitrate, ammonia, total nitrogen, total phosphorus,
total organic nitrogen, total organic carbon, silicate, dissolved oxygen, as well as turbidity and
light attenuation over the entire extent of the SERC dataset (Spring 1995 to Spring 2019) and
over the experimental period (April 2018 to 2019). The historical and contemporary SERC data
were used to identify differences between reef sites along with high resolution temperature data
collected during the experimental period. Annual mean, annual range, average daily range,
maximum monthly mean, days and hours above 30.5°C or 32°C, and thermal predictability for
each site was generated from the hourly temperature data. Thermal predictability was quantified
77
for only the summer months (June through September) because this period represents highly
variable temperatures at or above the bleaching threshold of 30.5°C (Manzello, Berkelmans, and
Hendee 2007). Autocorrelation was calculated with the acf function of the stats package over a
series of lags until autocorrelation reached zero. The sum of autocorrelation coefficients in that
series of lags was then used to summarize the predictability of each site. Sites with larger sums
are expected to be more predictable as current temperatures are more informative of
temperatures longer into the future. An analysis of variance was used to identify significant
differences among sites for independent parameters. A principal components analysis was used
to explain variation among sites using all parameters simultaneously.
Bayesian negative binomial generalized linear models implemented in R2jags (Su and
Yajima 2021b) were used to simultaneously test for the impact of multiple environmental
parameters on the growth and survival of coral colonies at outplant sites. To improve model fit,
change in size and risk score were calculated for each time interval and paired with
environmental parameters over the same interval. For SERC data, a single data point represented
each interval while thermal conditions were calculated within each interval. Highly correlated
environmental parameters, identified with Pearson correlations, were removed from the analysis
leaving 11 metrics to be used in the Bayesian model: total phosphorus, ammonia, total organic
carbon, dissolved oxygen, as well as light attenuation, turbidity, days above 30.5°C, thermal
predictability, average daily range, seasonal range, mean temperature.
78
3.7.2 Supplemental Figures and Tables
Supplementary Table 3.1: Outplant sites and SERC water quality monitoring station coordinates
Outplant sites SERC water quality monitoring stations
Site Lon Lat Station Lon Lat
Bahia Honda -81.24203 24.58908 256 (Bahia Honda) -81.2341 24.585
Big Pine Shoals -81.32655 24.56867 259 (Big Pine) -81.3217 24.5704
Looe Key -81.40214 24.54667 263 (Looe Key) -81.3974 24.5484
Dave's Ledge -81.48734 24.53051 267 (Dave’s Ledge) -81.4873 24.5301
Maryland Shoals -81.56991 24.51036 270 (Maryland Shoals) -81.5643 24.5216
E. Sambo -81.65961 24.49305 273 (E. Sambo) -81.6571 24.4929
W. Sambo -81.70334 24.48268 403 (W. Sambo) -81.7 24.483
Marker 32 -81.7426 24.47408 276 (Marker 32) -81.7374 24.4754
Eastern Dry Rocks -81.84412 24.45948 280 (EDR) -81.8437 24.4536
79
Supplementary Table 3.2: Results of Cox proportional hazard ratio models for genets
Coef exp(coef) se(coef) z p
Genotype1 0.9524501 2.592053 1.228389 0.78 0.44
Genotype50 1.2405797 3.457617 1.155948 1.07 0.28
Genotype3 1.1978105 3.312855 1.158254 1.03 0.3
Genotype44 1.0984207 2.999425 1.155609 0.95 0.34
Genotype7 1.5670542 4.792509 1.120372 1.4 0.16
Genotype31 1.6818513 5.375498 1.099799 1.53 0.13
Genotype13 2.3255781 10.232594 1.063958 2.19 0.029
Genotype62 2.5519734 12.832402 1.059458 2.41 0.016
Genotype41 2.6081417 13.573804 1.052236 2.48 0.01
Supplementary Table 3.3: Results of Cox proportional hazard ratio models for sites
Site Coef exp(coef) se(coef) z p
Marker 32 1.0708734 2.917927 1.232165 0.87 0.38
W. Sambo 1.3784374 3.968695 1.118639 1.23 0.22
Big Pine 1.3089321 3.702218 1.119653 1.17 0.24
Dave's Ledge 1.4134808 4.110237 1.120982 1.26 0.21
Looe Key 2.1242445 8.366574 1.083228 1.96 0.05
EDR 2.0674828 7.904900 1.082746 1.91 0.056
Maryland Shoals 2.1523159 8.604763 1.056446 2.04 0.042
Bahia Honda 2.5649031 12.999399 1.045958 2.45 0.014
80
Supplementary Figure 3.1: Pairwise genet rank correlations
across outplant sites ordered by geographic location
(longitudinally). Shape and color of ellipses are proportional
to the strength and direction of correlation.
81
Supplementary Table 3.4: Results of cumulative linked model testing the effects of genet and site
on ramet fragmentation
Sites Genotypes
Estimate Std. Error z value Pr(>|z|) Estimate Std. Error z value Pr(>|z|)
E. Sambo 0.5077 0.3317 1.531 0.12584 36 0.4645 0.3358 1.383 0.16652
W. Sambo 0.7426 0.3273 2.268 0.02330 50 0.2953 0.3379 0.874 0.38227
Big Pine 0.8121 0.3339 2.432 0.01502 3 0.6699 0.3364 1.991 0.04644
Dave's Ledge 0.9349 0.3287 2.844 0.00445 44 0.6699 0.3383 1.980 0.04766
Looe Key 0.8999 0.3257 2.763 0.00573 7 0.5157 0.3436 1.501 0.13344
EDR 0.6931 0.3358 2.064 0.03902 31 0.6021 0.3347 1.799 0.07206
Maryland Shoals 1.3802 0.3270 4.221 2.44e-05 13 0.7520 0.3365 2.235 0.02544
Bahia Honda 1.5764 0.3291 4.79 1.67e-06 62 0.2679 0.3400 0.788 0.43075
41 0.5835 0.3369 1.732 0.08329
Supplementary Figure 3.2: Total number of observed fragmentation events for genets (A) and
sites (B) across the one-year experimental period.
82
Supplementary Table 3.5: Summary of results for Fisher Exact test of the prevalence of
detectable fragmentation in A. cervicornis size classes.
Size Class n p p.adj p.
>5 cm 897 1.56E-6 9.36E-6
5-10 cm 897 5.52E-5 2.76E-4
10-15 cm 897 1 1
15-20 cm 897 0.485 1
20-30 cm 897 1 1
30+ cm 897 0.358 1
83
Supplementary Table 3.6: Summary of fixed effects for absolute size in TLE, SA, V and Vinter
Trait Factor Sum Sq Mean Sq Num DF DenDF F value Pr(>F)
SA Genotype 228.76 25.42 9 482.92 5.2064 9.06E-07
TLE Genotype 30.545 3.3939 9 496.4 6.4518 1.07E-08
V Genotype 32.874 3.6526 9 481.46 9.5383 1.98E-13
Vinter Genotype 68.554 7.617 9 467.44 6.6733 5.23E-09
SA Genotype:Site 839.91 11.83 71 479.29 2.4231 1.82E-08
TLE Genotype:Site 81.196 1.1436 71 492.65 2.174 8.25E-07
V Genotype:Site 70.096 0.9873 71 478.13 2.5781 1.53E-09
Vinter Genotype:Site 202.854 2.857 71 463.97 2.5031 5.74E-09
SA Genotype:Site:time 594.05 2.87 207 475.75 0.5879 0.999992
TLE Genotype:Site:time 64.679 0.311 208 488.71 0.5911 1
V Genotype:Site:time 47.526 0.2296 207 474.72 0.5996 0.999985
Vinter Genotype:Site:time 116.492 0.571 204 460.7 0.5003 1
SA Genotype:time 164.55 6.09 27 476.13 1.2484 0.183772
TLE Genotype:time 9.829 0.364 27 488.91 0.692 0.8771
V Genotype:time 10.524 0.3898 27 475.08 1.0179 0.441626
Vinter Genotype:time 13.982 0.518 27 461.02 0.4537 0.9925
SA Site 74.99 9.37 8 18.05 1.9198 1.19E-01
TLE Site 3.813 0.4766 8 17.99 0.9059 5.33E-01
V Site 7.544 0.943 8 18.09 2.4624 5.33E-02
Vinter Site 12.019 1.502 8 18.24 1.316 2.96E-01
SA Site:time 219.69 9.15 24 477.38 1.875 0.007664
TLE Site:time 13.699 0.5708 24 490.21 1.0851 3.56E-01
V Site:time 18.814 0.7839 24 476.4 2.0471 2.65E-03
Vinter Site:time 37.087 1.545 24 463.34 1.3538 1.24E-01
SA T0_SA 436.82 436.82 1 483.12 89.4773 < 2.2e-16
TLE T0_TLE 17.872 17.872 1 475.02 33.9751 1.03E-08
V T0_V 17.317 17.3167 1 491.95 45.2208 4.90E-11
Vinter T0_Vinter 55.842 55.842 1 468.85 48.9231 9.23E-12
SA time 1236.44 412.15 3 477.54 84.4239 < 2.2e-16
TLE time 83.662 27.8874 3 490.2 53.0146 < 2.2e-16
V time 64.458 21.4858 3 476.57 56.1081 < 2.2e-16
Vinter time 161.364 53.788 3 464.72 47.1238 < 2.2e-16
84
Supplementary Table 3.7: Summary of random effects on absolute size in TLE, SA, V, and Vinter
trait npar logLik AIC LRT Df Pr(>Chisq)
TLE (1 | Site:Array) 354 -799.21 2306.4 26.338 1 2.87E-07
TLE (1 | CulumativeBreaks) 354 -853.03 2414.1 133.988 1 < 2.2e-16
SA (1 | Site:Array) 353 -1338.9 3383.9 33.072 1 8.88E-09
SA (1 | CulumativeBreaks) 353 -1383.7 3473.3 122.483 1 < 2.2e-16
V (1 | Site:Array) 353 -705.6 2117.2 29.383 1 5.94E-08
V (1 | CulumativeBreaks) 353 -760.71 2227.4 139.605 1 < 2.2e-16
Vinter (1 | Site:Array) 350 -951.64 2603.3 39.209 1 3.81E-10
Vinter (1 | CulumativeBreaks) 350 -983.4 2666.8 102.73 1 < 2.2e-16
85
Supplementary Figure 3.3: Growth curves in SA. Genet averages at
each site are colored in decreasing overall survival from blue to red.
EDR Maryland Shoals Bahia Honda
Big Pine Dave's Ledge Looe Key
E. Sambo Marker 32 W. Sambo
T0 T3 T6 T9 T12 T0 T3 T6 T9 T12 T0 T3 T6 T9 T12
T0 T3 T6 T9 T12 T0 T3 T6 T9 T12 T0 T3 T6 T9 T12
T0 T3 T6 T9 T12 T0 T3 T6 T9 T12 T0 T3 T6 T9 T12
50
100
150
200
0
50
100
150
0
50
100
50
100
150
200
50
100
150
40
80
120
160
100
200
300
40
80
120
160
0
100
200
300
400
SA (cm^2)
Genotype
36
1
50
3
44
7
31
13
62
41
86
Supplementary Figure 3.4: Growth curves in V. Genet averages at
each site are colored in decreasing overall survival from blue to red.
Supplementary Figure 3.5: Growth curves in Vinter. Genet averages at
each site are colored in decreasing overall survival from blue to red.
87
Supplementary Table 3.8: Summary of linear fixed effects for growth rate in TLE, SA, V and
Vinter
Trait Factor Sum Sq Mean Sq Num DF DenDF F value Pr(>F)
SA Genotype 14.294 1.588 9 331.93 1.6839 0.091632
TLE Genotype 6.502 0.7224 9 341.01 4.0994 5.09E-05
V Genotype 6.618 0.7354 9 319.51 2.8875 0.002701
Vinter Genotype 4.655 0.5173 9 314.95 2.0695 0.031896
SA Genotype:Site 70.673 1.024 69 291.44 1.0813 0.324806
TLE Genotype:Site 15.82 0.2228 71 305.87 1.2597 0.0958539
V Genotype:Site 19.679 0.2852 69 307.21 1.1196 0.259268
Vinter Genotype:Site 23.119 0.3303 70 276.28 1.3167 0.063247
SA Genotype:Site:time 144.302 0.756 191 310.69 0.8015 9.53E-01
TLE Genotype:Site:time 34.243 0.1793 191 328.49 1.0182 4.40E-01
V Genotype:Site:time 35.751 0.1902 188 301.91 0.7468 9.85E-01
Vinter Genotype:Site:time 44.74 0.2472 181 297.03 0.9894 5.28E-01
SA Genotype:time 55.574 2.058 27 332.49 2.1831 0.000797
TLE Genotype:time 8.858 0.3281 27 348.14 1.8629 0.0065125
V Genotype:time 9.541 0.3534 27 313.22 1.3875 0.099349
Vinter Genotype:time 13.152 0.4871 27 314.72 1.9495 0.003913
SA Site 13.433 1.679 8 15.21 1.7799 0.158868
TLE Site 2.102 0.2627 8 16.01 1.4912 0.2358322
V Site 6.258 0.7823 8 15.46 3.0699 0.028047
Vinter Site 2.162 0.2702 8 16.81 1.0809 0.421249
SA Site:time 32.699 1.362 24 333.21 1.445 8.35E-02
TLE Site:time 10.609 0.442 24 349.66 2.5098 1.53E-04
V Site:time 10.827 0.4511 24 314.67 1.7714 1.56E-02
Vinter Site:time 11.752 0.4897 24 316.85 1.9594 0.005308
SA size 91.326 91.326 1 26.58 96.8685 2.35E-10
TLE size 15.476 15.4763 1 20.61 87.8828 7.03E-09
V size 11.356 11.356 1 302.72 44.592 1.16E-10
Vinter size 26.01 26.0104 1 72.21 104.099 1.21E-15
SA time 44.401 14.8 3 240.44 15.6703 2.43E-09
TLE time 6.556 2.1855 3 228.84 12.3819 1.56E-07
V time 9.445 3.1485 3 318.51 12.3629 1.14E-07
Vinter time 10.58 3.5268 3 257.45 14.0975 1.55E-08
88
Supplementary Table 3.9: Summary of random effects on growth rate in TLE, SA, V, and Vinter
Trait Factor npar logLik AIC LRT Df Pr(>Chisq)
TLE (1 | CulumativeBreaks) 337 -361.08 1396.2 0 1 1
TLE (1 | Site:Array) 337 -363.41 1400.8 4.6555 1 0.03095
SA (1 | CulumativeBreaks) 335 -641.38 1952.8 0 1 1
SA (1 | Site:Array) 335 -643 1956 3.2458 1 0.07161
V (1 | CulumativeBreaks) 332 -395.19 1454.4 2.5096 1 0.1132
V (1 | Site:Array) 332 -394.92 1453.8 1.9622 1 0.1613
Vinter (1 | CulumativeBreaks) 326 -390.51 1433 0.0041 1 0.9487
Vinter (1 | Site:Array) 326 -398.15 1448.3 15.2826 1 9.26E-05
Supplementary Figure 3.6: Relationship of growth rate in total linear
extension (cm/month) and ramet size at each time interval: Apr -Jul
2018 (0.3), Jul -Oct (3.6), Oct 2018-Jan 2019 (6.9), Jan-Apr 2019
(9.12)
89
Supplementary Figure 3.7: Relationship between growth rate in
total linear extension (cm/month) standardized to existing ramet
biomass and ramet size at each time interval: Apr-Jul 2018 (0.3),
Jul-Oct (3.6), Oct 2018-Jan 2019 (6.9), Jan-Apr 2019 (9.12)
90
Supplementary Figure 3.8: Pearson correlations between genet plasticity calculated with
a joint regression at a given time point and genet risk score calculated across the entire
experimental period. Rows of plots show plasticity-risk score relationships in a trait at
the four time intervals while columns. Black line and shaded region show line of best fit
and 95% confidence interval, respectively.
91
Supplementary Figure 3.9: Pearson correlations between genet plasticity calculated with
a joint regression at a given time point and genet average growth rate (unit/month) at a
given time point. Rows of plots show plasticity-growth rate relationships in a trait at the
four time intervals while columns. Black line and shaded region show line of best fit and
95% confidence interval, respectively.
92
Supplementary Figure 3.10: Pearson correlations between genet average growth rate
(unit/month) at a given time point and genet risk score calculated across the entire
experimental period. Rows of plots show growth rate-risk score relationships in a trait at
the four time intervals while columns. Black line and shaded region show line of best fit
and 95% confidence interval, respectively.
93
Supplemental Figure 3.11: Principal components analysis of ramet invariant morphology
measured with SA-to-V ratio, TLE-to-V ratio, packing, convexity, and sphericity. Points
represent ramets surviving to T12 without experiencing fragmentation (n=81) color by sites
decreasing in survival from blue to red. Polygons define the PCA space occupied by ramets of
each site.
94
Supplementary Figure 3.12: Hourly temperature data from April 2018 to April 2019 for each
outplant reef site.
95
Supplementary Table 3.10: Maximum monthly mean, annual mean, annual range, average daily
range, number of days where temperatures reached at least 30.5C (TDD30.5) or 32C (TDD32),
number of hours where temperatures were at or above 30.5C (TDH30.5) or 32C (TDH32), and
Summer thermal predictability for outplant reef sites calculated over the one-year experimental
period.
Site
Max Monthly
Mean
Annual
Mean
Annual
Range
Avg Daily
Range
TDD
30.5
TDD
32
TDH
30.5
TDH
32 Predictability
E. Sambo 30.833 27.474 10.12 0.628 57 0 824 0 209.141
Marker 32 30.774 27.386 10.489 0.587 54 0 824 0 195.195
W. Sambo 30.810 27.448 10.216 0.658 54 0 806 0 214.138
Big Pine 30.963 27.450 11.772 0.694 69 3 1154 27 134.525
Dave's Ledge 30.803 27.305 11.185 0.636 57 2 870 4 138.193
Looe Key* 28.215 25.809 7.857 0.790 0 0 0 0 NA
EDR 30.882 27.489 11.265 0.650 67 0 1060 0 186.321
Maryland Shoals 30.855 27.408 10.693 0.700 63 0 863 0 182.886
Bahia Honda 30.653 27.276 12.256 0.706 49 3 781 10 105.005
96
Supplementary Table 3.11: Summary of results for analysis of variance for SERC water quality
parameters using the full historical dataset (Spring 1995 -Spring 2019).
Parameter DF Sum Sq Mean Sq F value Pr (>F)
Nitrite 8 0.0117 .00147 1.517 0.147
Nitrate 8 1.12 0.141 2.405 0.0145
Ammonia 8 1.23 0.1543 1.027 0.414
Total Nitrogen 8 450 56.28 0.813 0.591
Total Organic Nitrogen 8 437 54.65 0.77 0.63
Total Phosphorus 8 0.017 0.00215 0.203 0.99
Soluble Reactive
Phosphorus
8 0.0051 0.00064 0.576 0.798
Turbidity 8 21.7 2.718 0.694 0.697
Total Organic Carbon 8 54788 6849 0.934 0.488
Silicon Dioxide 8 16.9 2.112 2.03 0.0405
Salinity 8 0.4 0.04944 0.329 0.955
Dissolved Oxygen 8 0.7 0.0819 0.138 0.997
Light Attenuation 8 0.27 0.0342 0.734 0.661
Oxygen Saturation 8 127 15.88 0.139 0.997
97
Supplementary Figure 3.13: Principal components analysis of contemporary (April 2018-April
2019) SERC water quality data and thermal characteristics for outplant sites used. Grey arrows
show vector loadings with labels at each of the tips in red. Sites are colored by decreasing
survival from blue to red. Looe Key is removed due to a lack of temperature data from April to
October 2018.
Supplementary Figure 3.14: Principal components analysis of historical (April 2018-April
2019) SERC water quality data and thermal characteristics for outplant sites used. Grey arrows
show vector loadings with labels at each of the tips in red. Sites are colored by decreasing
survival from blue to red. Looe Key is removed due to a lack of temperature data from April to
October 2018.
98
Supplementary Figure 3.15: Output from the Bayesian negative binomial generalized
linear model for total linear extension. Change in average TLE over a given time
interval was the response variable for environmental predictors measured over the
respective time interval. Points represent mean effect while error bars represent 95%
credible intervals. Points are colored blue (negative effect on response variable), red
(positive effect) if the credible interval does not overlap zero, and white otherwise.
Supplementary Figure 3.16: Output from the Bayesian negative binomial generalized
linear model for surface area. Change in average SA over a given time interval was the
response variable for environmental predictors measured over the respective time
interval. Points represent mean effect while error bars represent 95% credible intervals.
Points are colored blue (negative effect on response variable), red (positive effect) if
the credible interval does not overlap zero, and white otherwise.
99
Supplementary Figure 3.17: Output from the Bayesian negative binomial
generalized linear model for volume. Change in average V over a given time interval
was the response variable for environmental predictors measured over the respective
time interval. Points represent mean effect while error bars represent 95% credible
intervals. Points are colored blue (negative effect on response variable), red (positive
effect) if the credible interval does not overlap zero, and white otherwise.
100
Supplementary Figure 3.18: Output from the Bayesian negative binomial
generalized linear model for volume of interstitial space. Change in average
Vinter over a given time interval was the response variable for environmental
predictors measured over the respective time interval. Points represent mean
effect while error bars represent 95% credible intervals. Points are colored blue
(negative effect on response variable), red (positive effect) if the credible
interval does not overlap zero, and white otherwise.
Supplementary Figure 3.19: Output from the Bayesian negative binomial generalized
linear model for risk score. Site risk score over a given time interval was the response
variable for environmental predictors measured over the respective time interval.
Points represent mean effect while error bars show 95% credible intervals. Points are
colored blue (negative effect on response variable), red (positive effect) if the
credible interval does not overlap zero, and white otherwise.
101
Chapter 4 Plasticity and GxE in transcriptomic response in Acropora cervicornis
across environmentally variable reefs.
Summary of Contribution
This chapter utilizes the same experimental design in Chapter 3 which was conceptualized by
Carly Kenkel, my PhD advisor, and Cory Krediet of Eckerd College. Carly, Cory, Maria Ruggeri
and Erich Bartels (Mote Marine Laboratory) contributed to field work and RNA sample
collection at both initial (T0) and final (T12) time points. I prepared T0 gene expression libraries
while T12 libraries were prepared by Xuelin Zhao. I completed bioinformatic and differential
gene expression analyses to produce the results described in this chapter.
4.1 Abstract
Genetic and environmental factors can act independently and interactively to shape
phenotypes of individuals, ultimately influencing the ecology and evolution of a population. For
tropical coral reefs experiencing dramatic environmental change, identifying the contributions of
genotype, environment, and GxE on molecular trait production will be useful in determining
strategies of coral response to stress. Here, the impacts of G, E, and GxE on the gene expression
of the endangered coral, A. cervicornis, were quantified through an in situ transplant experiment
exposing common-garden (nursery) raised clones of ten genotypes to nine reef sites in the
Florida Keys. Tag-based RNAseq of coral individuals before and one year-post transplanting
identified a high degree of molecular plasticity with over 80% of the transcriptome showing
differential expression across outplant reef sites. Plastic expression was correlated with thermal
conditions at reef sites, suggesting molecular plasticity may be contributing to acclimation or
resilience during environmental variation. However, the presence of GxE in some genes
indicates that while populations can have plastic molecular phenotypes, individuals differ in their
ability to modulate gene expression. In contrast, significant intraspecific variation in constitutive
102
gene expression observed here may provide support for stress tolerance, an alternative strategy to
deal with environmental change that relies on baseline differences between individuals rather
than plasticity, as a mechanism that can maintain performance in A. cervicornis across reefs. As
environments continue to change, understanding the genetic and environmental contributions in
coral stress responses will be critical for predicting the persistence of this species. The results
presented here provide an important initial assessment of fixed and plastic gene expression in A.
cervicornis and lay the groundwork for future exploration of mechanisms underlying tolerance
and acclimation in coral.
4.2 Introduction
The ecology and evolution of a species is dependent on selection acting on variation in
individual phenotypes that are produced through both genetic and environmental controls.
However, as environments change, phenotypes fully dependent on environmental signals will
fluctuate, creating opportunities for individuals to track changing fitness-optima via acclimation
(West-Eberhard 1989; Wilson and Franklin 2002). Environmentally plastic traits can expose or
buffer genetic variation from selection, potentially accelerating or hindering evolution (Hendry
2016) while plastic traits with no genetic basis will have little contribution to directional
adaptation to stressors (Scheiner 1993). Alternatively, phenotypic variation fully dependent on
genetic contributions will remain constant as environments change allowing selection to interact
directly with genetic variation (Hoffmann and Merilä 1999; Van Buskirk and Steiner 2009).
Baseline differences among individuals in traits like stress tolerance arise from fixed genetic
effects and enable predictable outcomes to directional selection, i.e. as stress increases, higher
tolerance will yield higher fitness. While genetics and the environment both contribute, in some
degree, to the production of most traits, their role in within-generation survival mechanisms
103
(tolerance vs acclimatization) and their influence on the adaptive potential of a species highlight
their importance in understanding if and how organisms will respond to changing environments.
Scleractinian corals are long-lived organisms that form colonies via clonal reproduction
which form the structural basis of tropical coral reef ecosystems (Reaka-Kudla, Wilson, and
Wilson 1996; Knowlton et al. 2010). Although reef-building corals play a fundamental role in
maintaining marine biodiversity, because they already exist near their thermal maxima (Graham
and Nash 2013; Szmant 1997) they are particularly susceptible to natural and anthropogenic
environmental change. Restoration efforts, such as those targeting Acropora cervicornis in the
Caribbean (Young, Schopmeyer, and Lirman 2012), repopulate degraded reefs with nursery
raised coral and, in the process, expose corals to heterogeneous reef environments that may
further challenge individuals. Considering the sessile nature of coral, the ecology of these
systems will be dependent on the use of tolerance or acclimation, rather than avoidance.
Additionally, the genetic and plastic contributions to fitness-related phenotypes, and their
adaptive potential, will be increasingly important when considering that the continued
persistence of coral species into the future will require evolutionary changes to keep up with
environmental change (Donner et al. 2005).
Abundant evidence shows that genetic identity contributes to coral trait production with
broad genotype effects on traits such as growth (Edmunds and Putnam 2020), calcification
(Shaw et al. 2016), heat tolerance (Cunning et al. 2021; Barott et al. 2021; Cornwell et al. 2021),
and disease resistance (Wright et al. 2017). However, plasticity in growth (Todd 2008; Bruno
and Edmunds 1997), calcification (Tambutté et al. 2015; Carricart-Ganivet et al. 2012), heat
tolerance (Bay and Palumbi 2015) and even disease resistance (Muller, Bartels, and Baums
2018) have also all been shown in response to natural and experimental environmental variation,
104
implying a shared role of genetics and the environment in trait production. For fitness-related
traits that show both genotype and environment effects, like heat tolerance, baseline difference
between genotypes suggests differential performance under similar stress conditions (Cunning et
al. 2021; Barott et al. 2021; Cornwell et al. 2021) while a capacity to be plastic suggests
tolerance can be enhanced after environmental exposure (Palumbi et al. 2014; Barott et al. 2021;
Majerova et al. 2021).
Investigations into genetic and environmental effects on coral gene expression have
helped to identify molecular mechanisms supporting production of these higher order traits while
also providing insights into the strategies employed by corals under stress. For instance,
intraspecific variation in heat tolerance has been associated with signatures of fixed differential
gene expression (Barshis et al. 2013; Kenkel, Meyer, and Matz 2013; Bay and Palumbi 2017;
Yetsko et al. 2020) allowing the identification of numerous potential expression biomarkers for
predicting individual responses to heat stress (Louis et al. 2017). Similarly, fixed differential
gene expression underlying baseline resistance to disease also plays a role in the cellular
mechanisms that contribute to immune responses (Libro and Vollmer 2016). Plasticity in gene
expression in response to physical damage (Schlecker et al. 2022; van de Water et al. 2015),
thermal stress (DeSalvo et al. 2008), and coral disease (Daniels et al. 2015) has also provided
into the molecular processes impacted by these stressors. Moreover, gene expression plasticity
appears to support acclimation to environmental change, with numerous examples showing
sustained changes in expression of certain genes as coral persist in varying experimental (Moya
et al. 2015; Brener-Raffalli et al. 2022) or natural environments (Kenkel and Matz 2016; Palumbi
et al. 2014; Bay et al. 2013). In contrast to sustained changes in gene expression, transcriptomic
resilience, defined as expression change followed by a recovery to baseline, has been shown to
105
be a key mechanism supporting coral survival during stress (Savary et al. 2021; Seneca and
Palumbi 2015). Muted gene expression change in response to stressors can be a sign of
constitutively up-regulated genes, which has also been identified as a characteristic of some
thermally tolerant coral (Barshis et al. 2013; Bay and Palumbi 2015; Brener-Raffalli et al. 2022)
suggesting strategies reliant on fixed differences rather than expression plasticity can also
underlie coral responses to stress.
Despite the emphasis on identifying mechanisms underlying the production of higher
order traits and acclimation in other coral systems, transcriptomic investigations in Acropora
cervicornis have been restricted to the development of predictive biomarkers for thermal
tolerance (Parkinson et al. 2018, 2020), or on the ecotoxicology and epidemiology of natural
populations (Morgan, Vogelien, and Snell 2001; Morgan and Snell 2002; Libro and Vollmer
2016). While biomarkers primarily target fixed differences between genets within the population,
A. cervicornis displays abundant plasticity in morphological and physiological traits (Drury and
Lirman 2021; Drury, Manzello, and Lirman 2017; Kuffner et al. 2017) that can make prediction
based on trait values in one environment unreliable as environments change (Million et al. 2021).
Moreover, the molecular mechanisms underlying morphology and survival have received less
attention despite their relationship to fitness of natural and restored populations. Variable
survival and morphology among A. cervicornis genets and across restoration sites indicate both
genetic and environmental influence on coral trait production (Bowden-Kerby 2008; Drury,
Manzello, and Lirman 2017; Lirman et al. 2014). However, the mechanisms that support success
or failure under many environmental stressors are still unclear in this species. For instance,
differential gene expression helps explain variance in heat tolerance (Yetsko et al. 2020) and
disease susceptibility (Libro and Vollmer 2016) in A. cervicornis yet stress, evidenced as poor
106
survival and low growth at particular reefs, even in absence of heat or disease (Chapter 3), has
yet to be explained. Understanding both the mechanisms of, and genetic and environmental
contributions to, fitness-related trait production in A. cervicornis will be vital for identifying if
and how coral acclimate to pending environmental change in current and future generations.
Here, the genetic and environmental effects on A. cervicornis gene expression were
evaluated using 270 restored A. cervicornis colonies, representing 10 unique genotypes (genets)
on natural reefs in a multi-site transplant experiment. Gene expression sampled before
transplantation and one-year post transplant was used to identify genes differentially expressed
between genets, transplant sites, and genet-site interactions (GxE). Gene expression was
correlated with growth, survival, and environmental parameters to identify biological processes
that are associated with colony traits or acclimation to specific types of environments. Large
numbers of differential expressed genes by genet, site, and GxE suggest fixed and plastic effects
contribute to expression phenotypes. Genes displaying GxE were also correlated with
morphological traits and associated with biological functions involved in calcification,
suggesting these candidate genes may contribute to observed morphological plasticity (Chapter
3). Taken together, these results provide unique insights into the molecular processes involved in
A. cervicornis acclimation to natural reefs during restoration.
4.3 Methods
4.3.1 Experimental Design
Ten coral genets maintained long-term (5+ years) at the Mote Marine Lab Summerland
Key in situ coral nursery (Table S1 for coordinates) were selected for use in a multi-site
transplant study in partnership with Mote Marine Lab’s Staghorn Restoration Program under
FKMNS permits 2015-163-A1 and 2018-035. On March 19, 2018, 300 coral (mean TLE of 8.4
107
cm) ramets representing 10 genets (27 ramets per genet) were transferred from hanging trees,
fixed to ceramic pucks with marine epoxy and attached to benthic structures in preparation for
outplanting. In April 2018, each ramet was photographed for 3D photogrammetry (Million et al.
2021) and tissue samples for RNA were taken during a 2-hour window around solar noon. For
RNA, a 1-cm
3
portion of the apical tip of a single branch was removed with wire cutting pliers
and snap frozen in liquid nitrogen within 5 minutes of collection. After photographic and RNA
sampling, coral ramets were transplanted to nine established restoration sites (Table S3.1 for
coordinates) in the Lower Florida Keys. Sites were chosen based on their overlap with actively
permitted A. cervicornis restoration activities and proximity to permanent Southeast
Environmental Research Center (SERC) Water Quality Monitoring stations (SERC, Florida
International University). Three ramets per genet (n=270 fragments) were randomly outplanted
at each site with one ramet allocated to each of three ten-coral arrays. Arrays were spaced
approximately 1-m apart and the subsite locations were selected based on substrate type,
prioritizing hard bottom. Coral pucks were attached to bare reef substrate using marine epoxy
leaving approximately 10 cm between each puck. Outplanting occurred over 4 days due to
weather and time constraints, from April 21 to April 25.
Quarterly resurveys occurring in July 2018, October 2018, January 2019, and April 2019
were used to collect images to assess 3D morphological growth and to document mortality
occurring throughout the first year of outplanting. Survivorship was recorded during site surveys
and confirmed with photographs post-hoc. Ramets were recorded as alive if any amount of living
tissue was visible. Ramets where both the coral tissue and ceramic puck were missing were
discarded from the analyses as we attribute these losses to technical failure of the marine epoxy
upon outplanting rather than true biological mortality. At the one-year time point in April 2019,
108
ramets were also sampled for RNA following the same procedure described above. However,
because transport between outplant sites required RNA sampling to occur over 3 days, all RNA
samples were taken within a 2-hour window of solar noon each day to control for variation in
gene expression occurring throughout the day. Samples were stored at −80 °C until processing.
4.3.3 Genet and reef site characterization
Survival, morphology and environmental parameters were generated as outlined in
Sections 3.6.2 - 3.6.4. Briefly, ramet survival tracked quarterly throughout the first 12 months
post outplant was used to quantify genet- and site-specific risks of mortality via Cox Proportional
Hazard models using the coxme (Therneau 2020) and survival (Therneau et al. 2022) packages
within R (R Core Team 2020). Morphological data was collected from 3D models of each coral
ramet generated in Metashape 1.5.4 (Agisoft LLC, St. Petersburg, Russia) and phenotyped in
Meshlab v2020.6 (Cignoni et al. 2008) following methods outlined in in Million et al. (Million et
al. 2021). Four growth related traits total linear extension (TLE), surface area (SA), volume (V),
and volume of interstitial space (Vinter) were used to describe changes in absolute size of ramets
relative to initial size. The final shape of colonies that survived to T12 was described by four
size-invariant traits: SA-to-V ratio, TLE-to-V ratio, convexity, and sphericity (Zawada,
Dornelas, and Madin 2019). The total number of breaks represent the number of times a ramet
was reported having lost skeletal length. Fragmentation was identified by negative growth in
TLE between quarterly monitoring periods and was verified through the photographic record.
Thermal parameters for each reef site were obtained from hourly temperature data
collected throughout the outplant period by HOBO Pendant Temperature loggers (Onset
Computer Corp.; Bourne, MA). Hourly temperature records were used to calculate average daily
range, maximum monthly mean, summer thermal predictability. Thermal predictability was
109
quantified for July through September only as highly variable temperatures at or above the
bleaching threshold of 30.5°C are expected during this window (Manzello, Berkelmans, and
Hendee 2007). Predictability was calculated as the sum of autocorrelation over a series of lags
until autocorrelation reached zero, i.e. the point at which current temperatures are no longer
informative of future conditions (Walsh, Shapiro, and Shy 2005; Li and Ding 2013). Average
daily temperature range, maximum monthly mean, and summer thermal predictability were
chosen from the available temperature parameters described in Section 3.6.3 because they
showed the most influence on coral phenotypes within the Bayesian framework (Section 3.5.4).
Longitude was also included in this analysis to control for geographical variation in site
environmental conditions.
4.3.3 Library preparation
RNA was extracted using the Aurum Total RNA Mini Kit (Bio-Rad Laboratories, Inc.)
with minor modifications. The final elution step used 25 ul of elution solution instead of the
prescribed 40 ul to optimize RNA concentrations prior to cDNA synthesis. One microgram of
total RNA per sample was used for tag-based RNA-seq, or TagSeq (Meyer, Aglyamova, and
Matz 2011), with modifications for sequencing on the Illumina platform. Briefly, cDNA was
synthesized from RNA template with poly-A selection that incorporated unique oligonucleotides
that are used to remove PCR duplicates during bioinformatic analyses. 30 ng of amplified cDNA
libraries were barcoded then size selected from agarose gel following the “Freeze-and-Squeeze”
method, eluting in purified water instead of sodium acetate/EDTA buffer (Tautz and Renz 1983).
Final libraries were quantified with qPCR and pooled in equal amounts prior to sequencing.
4.3.2 Bioinformatic analyses
110
A total of 462 libraries were sequenced in two runs representing T0 sample (n=275) and T12
samples (n=187). T0 was sequenced on the Illumina HiSeq 4000 at the University of Chicago
Functional Genomics Facility while T12 was sequenced on the Illumina NextSeq2500 at the
University of Southern California Genome Core Facility. Although 270 ramets were outplanted
in April 2018 (3 ramets of each of 10 genets at 9 sites), 48 were lost due to mortality, 24 were
missing after failure of marine epoxy and 11 more were discarded during T12 sample preparation
for poor RNA or cDNA quality. Overall, 272.4 million raw reads were generated for T0 and 2.59
billion raw reads were generated for T12, with individual counts ranging from .59 to 7.84 million
per sample (median = 1.2 million reads) and 2.63 million to 133.7 million reads per sample
(median = 9.04 million reads), for T0 and T12 respectively. Custom perl scripts were used to
discard PCR duplicates and to trim the leader sequence from remaining reads as described by
Kenkel and Matz (2016). The fastx_toolkit (http://hannonlab.cshl.edu/fastx_toolkit) was then
used to remove ‘poly-A’ tails ( ≥ 8 bases) targeted during cDNA synthesis steps, to retain reads
with a minimum sequence length of 20 bases, and to quality filter, only maintaining reads where
70% of bases had PHRED scores of at least 33. A total of 0.27–3.27 million reads per T0 sample
(median = 0.544 million T0 reads) and 1.76 – 57.3 million reads per T12 sample (median =
4.9 million T12 reads) remained after quality filtering. Filtered reads were mapped to the A.
cervicornis reference transcriptome (Libro, Kaluziak, and Vollmer 2013) with gmapper in
SHRiMP (Rumble et al. 2009). Counts of sequences putatively originating from the same gene or
with sufficiently high sequence similarity to justify the assumption that they serve the same
function were grouped into isogroups using a custom perl script (Kenkel and Matz 2016). Reads
mapping to multiple isogroups were discarded. In total, 17,936 – 2,176,282 unique reads per
111
sample (median = 365,414 reads) for T0 and 933,215 – 32,567,082 unique reads per sample
(median = 2,703,052 reads) for T12 mapped to 31,209 isogroups.
4.3.3 Gene and gene network expression analyses
Transcriptomic expression was analyzed using a gene-by-gene approach to identify
individual genes whose expression varies among genets, reef sites, and genet-by-site interactions
(GxE) and with a network-based approach to identify groups of genes correlated with higher
order traits and environmental conditions. Prior to either analysis, low expression genes, or those
with fewer than 10 counts in 90% of samples, were removed from the dataset leaving 10,593
highly expressed genes for the T0 data and 19,527 highly expressed genes for the T12 data. Gene
counts in the T0 and T12 datasets were normalized in DESeq2 (Love, Huber, and Anders 2014)
for gene-by-gene analyses and normalized and regular-log transformed for network analyses.
DESeq2 was used to identify differentially expressed genes (DEGs) across coral genets,
outplant sites, and genet-site combinations at T12 using the Likelihood Ratio Test (LRT). The
LRT method is capable of identifying differentially expressed genes across all pairwise
comparisons of experimental factors and is therefore useful given the large experimental design
used here. Models testing for genet (~genet) or site (~site) effects were compared to reduced null
models (~1) while GxE effects were tested for by comparing a full model (~genet+
site+genet*site) to a reduced model excluding the interaction term (~genet+site). Benjamini-
Hochberg corrected p-values below a threshold of 0.05 were used to identify significantly
differential expressed genes in each category.
In order to visualize complex profiles of DEGs across genets and sites, DEGreport
(Pantano 2022) was used to cluster genes into groups based on similar expression patterns across
genets or sites, therefore providing the dominant patterns of expression within the experimental
112
design. Clustering was applied to only the top 10% most significant DEGs by site and by genet.
Clustering was set to create groups containing a minimum of 15 genes. While this analysis was
helpful to visualize expression profiles, correlations between expression and higher order traits,
including mortality risk score, were completed in the WCGNA framework as described below.
Highly expressed genes at T12 were analyzed using Weighted Gene Correlation Network
Analysis (WGCNA, Langfelder and Horvath 2008, 2012) in order to identify groups of genes
correlated with quantitative traits in experimental ramets. The 19,527 genes were rlog
transformed in DESeq2 using the blind experimental design setting. The quantitative traits
initially used in these analyses included genet mortality risk score, absolute growth in TLE, SA,
V, Vinter, growth rate of TLE, SA, V, and Vinter, T12 values and change relative to T0 in
Sphericity, Convexity, SA-to-V, and SA-to-TLE as well as site characteristics including
mortality risk score, longitude, summer thermal predictability, average daily temperature range,
and maximum monthly mean. Categorical traits identifying ramet genet identity or outplant site
location were also included to identify associated with individual genets or sites. After network
construction, growth rate values and change in Sphericity, Convexity, SA-to-V, and SA-to-TLE
ratios were removed from the analysis because they share similar correlations with related traits
and provide no additional information to the WGCNA analysis. Three ramets were identified
and removed as outliers with a standardized connectivity score of less than -2.5 reducing the
sample size to 184 ramets. A “signed” gene network was built using a soft threshold power value
of 10 and setting the minimum module size to 35 genes. Initially, 52 coexpression modules were
produced however, after merging modules with 80% or more similarity, 40 modules remained.
Candidate modules likely to be associated with quantitative traits were identified by meeting
three criteria. First, candidate modules had significant (p <0.05) and strong (R ≥ 0.15 or R ≤ -
113
0.15) correlations with a quantitative metric. Second, gene module membership vs. gene
significance plots for each candidate module showed that as the membership of a gene to a
module increased, the significance of the gene with the quantitative trait also increased (Fig.
S4.8-4.22). Finally, heat maps of genes in candidate modules showed a consistent up or down
regulation across samples as a quantitative trait changed, suggesting module correlations were
not based solely on a subset of outlier samples with highly differentially expressed genes.
Module eigengene values for each sample were used as a response variable to help visualize how
module expression changed relative to a quantitative trait of interest.
A rank-based gene ontology analysis was performed in the GO_MWU package (Wright
et al. 2015) to identify enriched functional groups within groups of significant site, genet, or GxE
DEGs. Mann-Whitney U tests based on ranking of adjusted p-values were used to identify GO
categories with genes significantly abundant in the most differentially expressed site, genet, or
GxE genes. Functional enrichment analysis of WGCNA modules was completed using Fisher’s
exact tests based on categorical values depending on the presence or absence of a gene in a
module.
4.4 Results
4.4.1 Transcriptome-wide profiles
Transcriptome wide expression patterns evaluated with principal components analysis using all
highly expressed isogroups showed little to no clustering at T0. The first two principal
components explained 33% of variation in T0 expression with higher survival genets tending to
appear lower on PC2 when compared to lower survival genets (Fig. 4.1A). Transcriptome wide
expression at T12 showed strong clustering by genet identity with no clear differentiation by
outplant site (Fig. 4.1B). Genet survival did not appear to be correlated with either principal
114
component although the top five surviving genets tended to have PC1 values greater than -5
while the bottom five surviving genets tended to have PC1 values less than 0. PC1 explained
13% of variation in T12 expression while PC2 explained 10%. When compared directly, T0 and
T12 ramet expression was differentiated primarily along PC1 (19% of variance explained) where
T12 ramets tended to group together (Fig. S4.1).
Figure 4.1: Principal component analysis of variance stabilized transformed expression values
for high expression genes (n = 10,593 T0 isogroups and 19,527 T12 isogroups) for T0 ramets (n
= 275) sampled in the restoration nursery before outplanting (A) and T12 ramets (n = 187)
sampled one year after outplantation to reefs (B).
4.4.2 Genet, site, and GxE contributions to gene expression
Among the 10,593 highly expressed T0 genes, 88.4% of genes (9,367) showed
differential expression between at least one pair of genets. However, no gene ontology terms
were enriched in this set of differentially expressed genes (DEGs). Of the 19,527 highly
expressed T12 genes, 80.3% (15,688) were significantly differentially expressed among any
combination of reef sites (adjusted p < 0.05). Approximately 46.1% (9004) of highly expressed
genes varied significantly among genets (adjusted p < 0.05) while just 2.5% (492) of genes
115
showed GxE effects (adjusted p < 0.05). Numerous genes were shared between categories, 6,242
genes showed independent effects of both genet and site, 278 genes shared genet and GxE
effects, 63 genes shared site and GxE effects, and 147 genes showed genet, site, and GxE effects
(Fig. S4.2A).
Genes with significant site effects were significantly enriched in biological processes
(p<0.05) involving regulation of RNA splicing, mRNA processing and metabolic process,
transcription from RNA polymerase II promoter, activation of protein kinase, activation of NF-
kappaB-inducing kinase, and innate immune response (Fig. S4.2B). In order to identify the
predominant patterns of differential expression across reef sites, the top 10% most significantly
differentially expressed genes among sites (n=1,953) were clustered into groups by similar
expression profile using the DEGreport package. The resulting 19 gene groups contained 16 to
375 genes, each displaying unique patterns across the nine outplant reef sites (Fig. S4.3). When
group reaction norms were organized by decreasing overall site survival, three broad expression
patterns emerged; exemplified in Figure 4.2. First, multiple groups (Group 6, Group 7) showed
similar expression in ramets outplanted across the top 3 highest surviving sites followed by
highly variable expression at the remaining sites (Fig. 4.2). Second, multiple groups including
Groups 23 and 15 showed similar expression among the top 3, middle 3, and bottom 3 sites
based on overall survival (Fig. 4.2). Finally, the majority of clusters, including Groups 13 and
11, show highly variable expression that appears to be independent of site survival (Fig. 4.2).
116
Figure 4.2: Site-specific DEGs clustered reaction norms across outplant reef sites. Points
represent the mean z-score for individual genes averaged within a site. Boxes represent 1st and
3rd quartile of cluster expression at each site, horizontal bars show median while whiskers
represent range of data. Vertical dashed lines help visualize patterns of expression between sites
with high, medium, or low survival when relevant.
Among the most significantly enriched biological processes (p < 0.005) in significant
DEGs by genet include homophilic cell adhesion, regulation of innate immune response,
response to biotic stimulus, and defense response to other organisms (Fig. S4.2C). The top 10%
most significant genet DEGs (1,953) were clustered by similar expression profiles across genets
in the DEGreport, producing 56 distinct groups. Groups ranged from 16 to 50 genes and
displayed complex patterns of expression changes among genets (Fig S4.4). Among these were
groups that showed patterns of up or down regulation that followed the genet rank order in terms
of mortality risk score (Fig. 4.3). Groups 27 and 51 showed increased expression for genes
within the group as genet risk of mortality increased. Genes in these groups are involved in
biological processes such as negative regulation of NF-kappaB transcription factor activity,
117
innate immune response, and response to pH (Group 27) and response to virus, ion
transmembrane transport, and inflammatory response (group 51). Conversely, expression of
genes in Groups 68 and 70 tended to be reduced in genets with increased risk of mortality while
expression of these genes is neutral or slightly increased in genets with low risk of mortality.
Group 68 contains genes involved in processes such as DNA repair, cell fate specification, and
positive regulation of cell proliferation. Group 70 contains genes involved in response to
mechanical stimulus, response to fluid shear stress, and positive regulation of calcium ion
transport into cytosol.
Figure 4.3: Genet-specific DEGs cluster patterns across A. cervicornis genets. Points represent
the mean z-score of individual genes averaged within genets. Boxes represent 1st and 3rd
quartile of cluster expression at each site, horizontal bars show median while whiskers represent
range of data.
There were 73 biological processes that were enriched in the 492 DEGs displaying
genotype-by-site (GxE) interactions (Fig. S4.5). Among the most significantly enriched GO
118
terms (p < 0.001) were those associated with immunity including regulation of immunoglobulin
secretion, regulation of immune effector process, regulation of leukocyte proliferation and
defense response to Gram-negative bacterium. Additional enriched biological processes included
regulation of cellular respiration, response to (laminar) fluid shear stress, mechanosensory
behavior, regulation of osteoblast proliferation, bioluminescence, and cellular response to
chemical stimulus, hormone stimulus, endogenous stimulus, and oxygen containing compounds.
Functional annotations for all genet, site, and GxE DEGs are available in Appendix A.
4.4.3 Gene network expression
Although the expression of numerous genes appears to be impacted by genet, site, GxE,
gene-by-gene approaches are limited in their ability to identify relationships between expression
and quantitative traits such as morphology, survival, and environmental conditions. In order to
further explore these relationships, a weighted gene correlation network analysis (WGCNA,
(Langfelder and Horvath 2008, 2012) was conducted using the T12 dataset and the phenotypic
and environmental measures collected through the outplant period. The 19,527 genes were
assigned to 40 coexpression modules ranging in size from 36 to 4,666 genes. Modules with
significant relationships to quantitative and categorical traits (Figure S4.6 & S4.7) and strong
Module Membership to Gene Significance correlations (Fig. S4.8-4.22) were identified as
candidate modules containing genes putatively associated with phenotype or environmental
conditions. Similarly, module expression was further verified by visually assessing that broad
pattern of expression were consistent among individual genes within a module across samples
(e.g. Fig. S4.23-4.24).
Seven modules showed significant correlations with at least one genet-specific
phenotypic value with correlation coefficients that ranged from -0.27 to 0.57 (Fig. 4.4A). Dark
119
Orange and Light Green modules showed positive relationships with growth in TLE, SA, V, and
Vinter while the Orange and Blue modules showed negative relationships with these traits. While
SA to V ratio, Sphericity, Convexity, and SA to TLE ratio were significantly correlated with
various modules, including those associated with growth, Light Steel Blue 1 was the only
module to be correlated with all traits defining colony shape. The Salmon module showed the
highest correlation with genet mortality risk; however, this relationship was driven by the high
over-expression of module genes in genet (G) 41 (Fig. S4.23). Overexpression of the Light Steel
Blue 1 module was also correlated with increased risk of mortality, however, these relationships
also appeared to be driven by ramets of G41 and G31 (Fig. S4.24). Functional annotation
revealed no significantly enriched GO terms in any of the coexpression models significantly
correlated with genet-specific parameters. Interestingly, 71 genes (14.4%) that showed GxE
effects when analyzed in DESeq2 were also in modules that were correlated with growth and
morphology. Of these modules, Blue contained 26 GxE-specific DEGs, Dark Orange contained
2, Light Green contained 3, Light Steel Blue 1 contained 7, Orange contained 32, and Sky Blue
contained 1. Among the annotated GxE-specific DEGs in the Blue modules were genes involved
in the melanin biosynthesis pathway, regulation of growth, calcium-mediated signaling using
intracellular calcium, and four processes related to the regulation and response to calcium ions.
Among GxE DEGs within the Orange Module were genes involved in response to mechanical
stimulus, regulation of immune activity including regulation of NF-kappaB transcription, and
cellular response to multiple organic and inorganic compounds.
120
Figure 4.4: Module-trait heatmaps showing strength and direction of correlations between
modules (number of genes) and (A) genet-specific traits or (B) site-specific characteristics.
Correlation coefficients are shown when the relationship between module and trait was
significant.
Outplant site risk of mortality, longitude, thermal predictability, average daily range, and
maximum monthly mean were significantly correlated with six coexpression modules with
correlation coefficients ranging from -0.4 to 0.48 (Fig. 4.4B). Only the Turquoise module
showed a significant relationship with the risk of mortality at a given site (R = –0.24) with
expression of this module downregulated at more challenging sites. The Blue module was
negatively associated with site longitude and summer thermal predictability (R =-0.37 and -0.17,
respectively) but was positively associated with average daily temperature range and maximum
monthly mean (R = 0.4 to 0.4, respectively). The remaining site-associated modules (Dark
Magenta, Light Green, Turquoise, Dark Orange, and Brown 4) showed positive relationships
with longitude and summer thermal predictability (R = 0.16 to 0.48) but negative relationships
with average daily temperature range and maximum monthly mean (R = -0.27 to -0.4). While
functional enrichment analysis identified no significant GO terms in any of the six modules,
121
3601 genes within these modules are also site DEGs identified by DESeq2. The Turquoise
module contained site-specific DEGs involved in response to starvation and various aspects of
the immune response. Site-specific DEGs in the Blue module tended to be involved in heat
response processes and membrane-related processes included membrane organization, repair,
and regulation of permeability.
4.5 Discussion
4.5.1 Genet-level differences in expression persist after outplanting
Although initial patterns of gene expression, defined by transcriptomic profiles prior to
outplanting (T0), do not appear to separate ramets by genet identity, trends do appear to be
predictive of survival after outplanting (Fig. 4.1). Specifically, ramets belonging to the top 5
highest surviving genets tended to group lower on PC2 while the bottom 5 lowest survival genets
occurred higher on PC2. While this trend suggests gene expression patterns can potentially
predict outplant survival, similar to other biomarkers for disease and thermal tolerance (Wright et
al. 2017; Louis et al. 2017; Young et al. 2020), power may be limited. Location along PC2 did
not align with survival rank when viewing genets individually outside of this coarse categorical
assignment. For instance, ramet expression of the lowest surviving genet (G62) and highest
surviving genet (G36) often overlapped while the most extreme points along PC2 belonged to the
second worst surviving genet (G41) and the fourth best surviving genet (G50). Mechanisms
determining outplant survival are expected to be complex because rather than tolerance to a
single stressor, survival on natural reefs depends on the ability to cope with a multitude of biotic
and abiotic stressors (Williams and Miller 2005; Williams and Miller 2006; Lustic et al. 2020;
Cramer et al. 2020). Considering responses to these stressors may also rely on environmental
plasticity, fixed molecular biomarkers may be less useful for predicting outplant survival.
122
While genet identity did not appear to have strong influence on T0 expression,
transcriptomic patterns after one year at natural reefs (T12) were predominantly defined by genet
identity (Fig. 4.1B). This suggests that strong fixed effects on expression persist even as
individuals are exposed to different environments. The strong genet-specific expression at T12 is
interesting when compared to the lack of differentiation at T0, but multiple factors could be
contributing to this difference. For instance, in situ coral nurseries are relatively benign habitats
that support high survival and growth in A. cervicornis (Lohr et al. 2015; Herlan and Lirman
2008) when compared to natural reefs that present numerous biotic challenges including
predation (Williams and Miller 2006), competition (Lustic et al. 2020), infection (Cramer et al.
2020; Williams and Miller 2005). Genet-specific responses to biotic stressors may therefore be
contributing to the strong genet differentiation present in overall transcriptome profiles. This is
supported by abundant DEGs that were enriched for processes such as regulation of the innate
immune response, response to virus and bacterium, and positive regulation of defense responses.
Technical differences between sequencing of T0 and T12 expression may also be contributing to
the time-point specific differences. Specifically, T12 samples were sequenced deeper than T0
samples by an order of magnitude, making comparisons between the two datasets difficult.
Regardless of the ability to make direct comparisons, the obvious influence of genet identity that
persists even after a year of outplanting is important for considering the mechanisms that enable
genets to survive environmental challenges in the wild.
Of the highly expressed genes in A. cervicornis ramets at T12, 46% showed significant
effects of genet identity. Many genet-specific DEGs included enriched GO categories related to
immune responses, cell adhesion, and response to lipid, which have been implicated as
mechanisms underlying active disease responses in A. cervicornis (Libro, Kaluziak, and Vollmer
123
2013) and its sister species A. palmata (Young et al. 2020). While no disease was evident
through the experimental monitoring period, differential expression of genes relating to
interactions with microbiota may be representative of the establishment of genet-specific surface
microbiomes seen in these experimental coral upon outplanting (Aguirre et al. 2022). In order to
visualize expression changes across genets, genet-specific DEGs were clustered into groups
exhibiting similar changes across samples. For some groups, expression appeared to be
correlated with mortality risk score, with expression of individual genes increasing or decreasing
as risk of mortality increased (Fig. 4.3). Interestingly, genes that were upregulated in genets with
lower survival were involved in innate immune response, including NK-kappaB regulation and
inflammatory responses. The innate immune system in coral has been shown to be activated in
response to multiple stressors beyond disease (Mydlarz, McGinty, and Harvell 2010; Mydlarz et
al. 2009) so upregulation of these processes in poor surviving corals may be indicative of a broad
stress response in these genets across each reef site. Among the DEG groups whose expression
was upregulated in genets with higher survival were genes involved in cell proliferation (Group
68) and processes putatively related to calcification (Group 70). Expression and methylation
changes in genes relating to cell proliferation in response to pH stress have been linked to
changes in growth in reef building corals (Liew et al. 2018). Additionally, the dependence of
coral growth on chemical homeostasis, especially calcium (Cohen and Holcomb 2009; Allemand
et al. 2011) suggest an important potential role of genes in Group 70 in coral calcification. In
addition to individual genes, two WGCNA modules (Salmon and Light Steel Blue1) were
correlated with genet risk of mortality, however, these trends were driven by expression in one or
two genets (Fig. S4.22 & S4.23). This suggests that expression of these modules is less likely to
124
be associated with the quantitative trait but more indicative of individual genets that may or may
not be poor survivors.
Taken together, expression of numerous genes, including those potentially involved in
immunity and growth, show constitutive differences between genets across all outplant reef sites.
While it is evident that up or down regulation of various genes is associated with increased
survival, the mechanistic link between these fixed differences and coral performance is still
unclear. For instance, upregulation of immune responses is apparent in low survival genets, but
determining whether expression differences are a cause or symptom of compromised health will
be an important next step in explaining the variation in performance seen in A. cervicornis
(Bowden-Kerby 2008; Drury, Manzello, and Lirman 2017; Lirman et al. 2014). Still, the
contribution of fixed differences to expression phenotype will have important implications for
coral ecology and evolution because even in the presence of environmental change, individuals
appear to be able to maintain trait values relative to the population.
4.5.2 Abundant gene expression plasticity may contribute to coral acclimatization
Coral in this experiment were outplanted to reefs that occupy regions that are historically
environmentally divergent (Briceño and Boyer 2019) and were shown to differ in abiotic
conditions throughout the experiment period (Section 3.4.6). Therefore, the abundant gene
expression plasticity in coral individuals may be attributed to environmental variation across
these reefs and adds to the growing evidence of transcriptomic plasticity in coral (Bay and
Palumbi 2015; Kenkel and Matz 2016; Palumbi et al. 2014; Rocker et al. 2019; Savary et al.
2021). Here, nearly 90% of the highly expressed A. cervicornis genes showed significant effects
of reef site suggesting a large majority of the transcriptome is responding to differences between
at least two sites. Considering T12 samples were taken from living coral ramets, the abundant
125
site-specific DEGs represent potentially crucial changes to molecular processes that are
underlying successful acclimation to natural reefs.
The biological processes enriched among site-specific DEGs included the innate immune
response, activation of NK-kappaB-inducing kinase, and general activation of protein kinase.
Protein kinases is a broad family of enzymes that modify other proteins via phosphorylation
(Taylor and Kornev 2011) and have been implicated in numerous processes in coral including
osmolyte production (Mayfield et al. 2010), signaling pathways potentially involved in the
establishment of symbiosis (Rosic et al. 2015), and disease resistance (Young et al. 2020). Here,
the enrichment of protein kinases, especially those activating NF-kappaB, along with innate
immunity processes point to a role for the immune response during acclimation to experimental
reef sites. Biological processes related to transcription and RNA splicing were also enriched in
site-specific DEGs. RNA splicing-related genes are among those downregulated during the first
hours of heat stress (Seneca and Palumbi 2015; Rodriguez-Lanetty, Harii, and Hoegh-Guldberg
2009) suggesting plasticity in these genes is important for immediate responses to environmental
stress. Together, these enriched GO terms suggest differential expression of genes relating to
stress response characterize the transcriptomic patterns of A. cervicornis outplanted to
environmentally diverging reefs (Briceño and Boyer 2019). However, the coral innate immune
response has been shown to be involved in response to heat, acidification, nutrient conditions,
physical damage in addition to direct resistance to infection (Traylor-Knowles and Connelly
2017; Mydlarz, McGinty, and Harvell 2010), raising the possibility that coral are activating
similar pathways to respond to different stressors across sites. While endpoint sampling of RNA
provides only a snapshot of ramet gene expression, making it difficult to precisely identify the
126
conditions that are triggering these responses, broader-scale characterizations of reefs may
provide clues as to how gene expression is responding to environmental conditions.
Groups of site-specific DEGs showed unique expression patterns across experimental
reef sites providing insight into the dynamics of molecular plasticity in A. cervicornis in nature
(Fig. 4.2). Here, cox proportional hazard mortality risk scores were used as a proxy for overall
site quality because this measure assesses the ability of a site to support A. cervicornis survival,
integrating across all environmental parameters that may determine the quality of a site.
Interestingly, expression of some DEGs appeared to follow patterns related to changes in site
quality when sites were grouped into high, medium, and low quality sites. Specifically, certain
site-specific DEG groups were consistently expressed in high quality sites but highly variable in
the remaining sites while other DEGs showed similar expression within, but variable expression
between, high, medium, and low quality sites (Fig. 4.2). Similarly, the Turquoise WGCNA
module (4,666 genes) showed a significant association with site quality, with genes in this
module being downregulated as site quality decreased (Fig. 4.4B). Genes that follow these
quality-related patterns may be useful for identifying the physiological processes that enable
corals to survive at increasingly poor reef sites. Additionally, while the conditions that influence
reef site quality are still unresolved, genes showing plasticity dependent on site quality may
provide insights into types of challenges experienced by coral at these sites. For instance, genes
in Group 23 which are involved in wound healing, response to mechanical stimulus, and
response to starvation are down regulated in high quality sites, up regulated in medium quality
sites, and slightly upregulated in low quality sites. Colony fragmentation results from mechanical
stress and results in wounding that triggers a variety of responses. Considering fragmentation
was variable among experimental reef sites (Section 3.4.2) site-specific expression profiles in
127
Group 23 genes may be indicative of response to fragmentation, a stress that may ultimately
reduce coral performance and characterize high, medium, and low quality reefs.
Regardless of functional annotation, the dependence of gene expression on
environmental contributions highlights the role of phenotypic plasticity during A. cervicornis
acclimation in nature. For example, downregulation of Group 7 genes in corals that survived to
one-year post-outplant at Bahia Honda (BH; Fig. 4.2) may characterize a successful phenotype at
the site with the highest risk of mortality while upregulation of these genes at Maryland Shoals
(MS) appears to be a successful phenotype at the second lowest quality site. While expression
plasticity has been previously shown in coral (Savary et al. 2021; Bay and Palumbi 2015; Kenkel
and Matz 2016) , these results provide crucial insight into the dependence of the environment on
A. cervicornis gene expression. Importantly, expression is highly plastic in this species even in
the absence of extreme thermal, chemical, or biological challenges (Parkinson et al. 2018;
Parkinson et al. 2020; Morgan et al. 2001; Morgan and Snell 2002; Libro and Vollmer 2016)
suggesting gene expression plasticity is likely to be useful for acclimation for individuals
experience environmental change via assisted migration or climate change. However, due to the
single endpoint sampling design, we were unable to document unsuccessful phenotypes so
conclusions based on the necessity or sufficiency of expression plasticity in genes like those in
Group 7 are not fully supported by these results.
Relationships between gene expression and environmental conditions at experimental
sites also highlights the sensitivity of molecular plasticity to specific abiotic conditions on
natural reefs. Multiple WGCNA modules were correlated with reef thermal characteristics
including summer thermal predictability, average daily temperature range, and maximum
monthly mean. Maximum monthly mean is a common characteristic representative of thermal
128
extremes experienced by reefs (Manzello, Berkelmans, and Hendee 2007) while the importance
of temperature range and predictability is being increasingly recognized as a factor contributing
to thermal stress (Oliver and Palumbi 2011; Bitter et al. 2021). The Blue module stood out as the
only module significantly upregulated under high average daily temperature range and high
maximum monthly mean. Numerous site-specific DEGs were present in the Blue module with
many being involved in processes related to response to heat stress suggesting that plasticity in
this module is likely to be important for coral acclimation during ocean warming. Significant
module associations with maximum monthly mean and daily temperature ranges that varied by
less than a degree across sites (Table S3.10) highlight the sensitivity of gene expression in A.
cervicornis, supporting a pattern that has been reported in other coral species (Mayfield et al.
2011; Kenkel, Meyer, and Matz 2013; Poquita-Du et al. 2019).
Recent molecular studies have emphasized the role of gene expression plasticity in the
role of resilience and/or acclimation to stress (Savary et al. 2021; Seneca and Palumbi 2015;
Kenkel and Matz 2016; Palumbi et al. 2014; Bay et al. 2013). While the results presented here
suggest a similar role in acclimation in A. cervicornis, there are important limitations to consider.
Transcriptomic resilience and acclimation describe a temporary or sustained change in
expression from baseline levels and as such require a population of experimental samples
exemplifying baseline expression to compare with plastic responses (Savary et al. 2021; Kenkel
and Matz 2016). While the T0 dataset represents initial gene expression levels, direct
comparisons with the T12 dataset were limited due to differences in sequencing depth.
Additionally, comparisons between T0 and T12 would not account for changes in expression
potentially associated with coral age/size (Schlecker et al. 2022). Similarly, the temporal stability
of gene expression patterns is unresolved, making it difficult to conclude if plastic changes seen
129
here are sustained between reefs (acclimation) or are temporary responses to the most recent
stress event (resilience). Therefore, future studies attempting to identify strategies of coral
responses to environmental change in situ should include additional ramets maintained in their
sourced location (nursery or native reef). Although these limitations preclude our ability to
differentiate between acclimation and resilience, the present study provides strong support for
the influence of environmental conditions on molecular trait production in A. cervicornis.
4.5.3 Contributions of intraspecific variation in expression plasticity in higher order traits
Genotype-by-environment interactions (GxE) provide evidence of intraspecific variation
in phenotypic plasticity, i.e. individuals respond differently to the same environment change.
GxE adds an additional layer of complexity to trait production because relative differences
between individuals in one environment will be exacerbated or minimized in another
environment, potentially creating divergent selection outcomes across environments (Hendry
2016; Kelly 2019). Intraspecific variation in plasticity also implies plasticity, itself, is a trait that
varies within a population which can be the subject of natural selection (Via and Lande 1985;
Scheiner 1993). In the present study, 492 genes showed significant genet-by-site interaction with
these GxE-specific DEGs being enriched in over 70 different biological processes (Fig. S4.5).
Among these enriched biological processes were GO terms involved in immune responses,
apoptosis, and cellular response to chemical and physical stimuli. Fine tuning of immune
responses and apoptosis genes have been implicated in both resistance to (Bellantuono et al.
2012; Kvitt, Rosenfeld, and Tchernov 2016; Mydlarz, McGinty, and Harvell 2010) and
symptomatic of stress (Ainsworth et al. 2008; Fuess et al. 2017). GxE in these suggest A.
cervicornis individuals modulate these processes differently across natural reefs, perhaps due to
intraspecific differences in sensitivity to environmental triggers, differential ability to up or down
130
regulate expression of these genes, or even divergent strategies for dealing with stress (tolerance
vs. resilience/acclimation).
Some of the most compelling GO terms enriched within GxE-specific DEGs were
response to laminar fluid shear stress, response to fluid shear stress, and mechanosensory
behavior which are involved in sensing and integrating cues from the physical environment and
specifically hydrodynamic forces. Similarly, four WGCNA modules were significantly
associated with quantitative growth metrics, while seven were significantly associated with size-
invariant morphological traits. Previous evidence supports the impact of hydrodynamic forces on
cnidarian morphology with water flow often being associated with inducing morphological
plasticity (Todd 2008; Dubé et al. 2017; Bruno and Edmunds 1997). It is tempting to hypothesize
that differential expression of these modules may be contributing to the phenotypic differences
observed in higher order traits. The abundant intraspecific variation in morphological plasticity
displayed by these genets in response to experimental reefs (Chapter 3) creates an opportunity to
make an important connection between GxE in higher order traits and GxE in molecular traits.
To this end, overlap between GxE-specific DEGs and WGCNA modules associated with growth
and morphology provide candidate genes that may be responsible for sensing and interpreting
environmental cues and then producing plastic trait variation. Of the 71 genes showing both GxE
and associations with growth and morphology are those involved in calcium-related processes
and growth and development-related processes. Taken together, the role of calcium and water
flow on coral growth (Allemand et al. 2011; Todd 2008; Cohen and Holcomb 2009) and the
presence of hydrodynamic sensing, calcium regulating, and growth mediating GxE-specific
genes associated with growth suggests these genes are putatively involved in morphological
plasticity. Although this evidence is preliminary at best, few studies have identified gene
131
expression associated with coral growth (Schlecker et al. 2022; Bay and Palumbi 2017) while
none have found associations with morphological plasticity. Considering the role of coral growth
and morphology in establishing many of the ecosystem services provided by coral reefs,
understanding their molecular basis will be important for predicting the ecology and evolution of
these traits.
4.5.4 Conclusions
Transcriptomic investigations are a useful tool to explore the mechanisms underlying changes, or
lack thereof, in higher order traits. The samples collected one-year post-outplanting were taken
from living corals and therefore the patterns of differential gene expression seen at T12 represent
‘successful’ phenotypes. Considering this, the abundant fixed and plastic effects on T12
expression may provide candidate genes contributing to mechanisms for successful stress
tolerance and acclimation/resilience in A. cervicornis. The high overlap between site and genet
specific DEGs (40% of site DEGs were also genet DEGs and 68% of genet DEGs were also site
DEGs) reiterates that the contributions of fixed constitutive differences and environmentally
dependent plasticity, and therefore potential strategies of tolerance and acclimation, may not be
mutually exclusive. Moreover, evidence of GxE imply that the interaction of genetic and
environmental factors is important in trait production in A. cervicornis. In the context of
strategies to deal with stress, intraspecific variation in gene expression plasticity may indicate
that some genets are more capable of utilizing plasticity-derived acclimation than others,
potentially making them more reliant on fixed differences expression to facilitate tolerance. The
GxE in molecular trait production seen here provides an exciting start for exploration into the
mechanism underlying GxE in higher order traits such as survival, morphology, and physiology.
Taken together, the results of this chapter not only provide important insights into the factors
132
influencing trait production in a critically endangered species but add to the potential of A.
cervicornis as a system for studying the eco-evolutionary dynamics of phenotypic plasticity.
4.6 Supplementary material
Figure S4.1: Principal components analysis of variance stabilized transformed expression values
for highly expressed genes in the combined T0 (275 ramets) and T12 (187 ramets) datasets (n=
17,074 genes). Points are colored by genet identity order from highest overall survival after
outplanting (blue) to lowest overall survival (red). Shapes indicate the location where a ramet
was sampled. Nursery ramets were sampled at T0 while all non-nursery ramets were sampled at
T12.
133
Figure S4.2: A) Number of significantly differential expressed genes (adjusted p-value < 0.05)
for each experimental factor; genet, site, or genet-by-site interaction (GxE). B) Gene Ontology
terms (biological processes) significantly enriched within genes displaying significant site effects
or (C) genet effects. The size and color of each term is correlated with the significance of the
enrichment for that term. GO terms are preceded by the fraction of candidate genes (those
passing significance measure) relative to the total number of genes belonging to that category.
134
Figure S4.3: Reaction norms for the 19 site-specific DEG clusters generated in DEGreport.
Points represent the mean z-score for individual genes averaged within a site. Boxes represent
1st and 3rd quartile of cluster expression at each site, horizontal bars show median while
whiskers represent range of data. The number of genes in each cluster are included next to the
numbered name of each cluster (group).
135
Figure S4.4: Patterns of expression by genet for the 56 genet-specific DEG clusters generated in
DEGreport. Points represent the mean z-score for individual genes averaged within a site. Boxes
represent 1st and 3rd quartile of cluster expression at each site, horizontal bars show median
while whiskers represent range of data. The number of genes in each cluster are included next to
the numbered name of each cluster (group).
136
Figure S4.5: Significantly enriched gene ontology terms in genes showing significant GxE
effects. The size and color of each term is correlated with the significance of the enrichment for
that term. GO terms are preceded by the fraction of candidate genes (those passing significance
measure) relative to the total number of genes belonging to that category.
137
Figure S4.6: WGCNA module-trait heatmap. Each row corresponds to a unique module and each
column corresponds to a quantitative or categorical trait: final size at T12 in TLE, SA, V, and
Vinter, growth rate in TLE, SA, V, and Vinter, final T12 value or change in SA-to-V ratio,
sphericity, convexity or SA-to-TLE ratio, genet-specific mortality risk (RiskG), site-specific
mortality risk (RiskS), outplant reef site, genet identity, site longitude, summer thermal
predictability, average daily temperature range, or maximum monthly mean temperature
(MMM). The direction and intensity of correlations between module expression and traits is
represented by color. Correlation coefficients are shown when the relationship between module
and trait was significant.
138
Figure S4.7: WGCNA module-trait heatmap. Each row corresponds to a unique module and each
column corresponds to categorical traits describing the genet-site identity of T12 samples. The
direction and intensity of correlations between module expression and traits is represented by
color. Correlation coefficients are shown when the relationship between module and trait was
significant.
Figure S4.8: Module membership to gene significance relationships with candidate WGCNA
modules associated with the total number of fragmentation events (breaks) experienced during
the experimental monitoring period. Strength and significance of each relationship is included
above each scatter plots. Points represent individual genes within each module.
139
Figure S4.9: Module membership to gene significance relationships with candidate WGCNA
modules associated with growth in TLE over the experimental monitoring period. Strength and
significance of each relationship is included above each scatter plots. Points represent individual
genes within each module.
Figure S4.10: Module membership to gene significance relationships with candidate WGCNA
modules associated with growth in SA over the experimental monitoring period. Strength and
significance of each relationship is included above each scatter plots. Points represent individual
genes within each module.
140
Figure S4.11: Module membership to gene significance relationships with candidate WGCNA
modules associated with growth in V over the experimental monitoring period. Strength and
significance of each relationship is included above each scatter plots. Points represent individual
genes within each module.
141
Figure S4.12: Module membership to gene significance relationships with candidate WGCNA
modules associated with growth in Vinter over the experimental monitoring period. Strength and
significance of each relationship is included above each scatter plots. Points represent individual
genes within each module.
Figure S4.13: Module membership to gene significance relationships with candidate WGCNA
modules associated with T12 SA-to-V ratio. Strength and significance of each relationship is
included above each scatter plots. Points represent individual genes within each module.
142
Figure S4.14: Module membership to gene significance relationships with candidate WGCNA
modules associated with T12 sphericity. Strength and significance of each relationship is
included above each scatter plots. Points represent individual genes within each module.
Figure S4.15: Module membership to gene significance relationships with candidate WGCNA
modules associated with T12 convexity. Strength and significance of each relationship is
included above each scatter plots. Points represent individual genes within each module.
143
Figure S4.16: Module membership to gene significance relationships with candidate WGCNA
modules associated with T12 SA-to-TLE ratio. Strength and significance of each relationship is
included above each scatter plots. Points represent individual genes within each module.
Figure S4.17: Module membership to gene significance relationships with candidate WGCNA
modules associated with genet-specific risk of mortality. Strength and significance of each
relationship is included above each scatter plots. Points represent individual genes within each
module.
144
Figure S4.18: Module membership to gene significance relationships with candidate WGCNA
modules associated with site-specific risk of mortality. Strength and significance of each
relationship is included above each scatter plots. Points represent individual genes within each
module.
Figure S4.19: Module membership to gene significance relationships with candidate WGCNA
modules associated with site longitude. Strength and significance of each relationship is included
above each scatter plots. Points represent individual genes within each module.
145
Figure S4.20: Module membership to gene significance relationships with candidate WGCNA
modules associated with site summer thermal predictability. Strength and significance of each
relationship is included above each scatter plots. Points represent individual genes within each
module.
Figure S4.21: Module membership to gene significance relationships with candidate WGCNA
modules associated with site average daily temperature range. Strength and significance of each
relationship is included above each scatter plots. Points represent individual genes within each
module.
146
Figure S4.22: Module membership to gene significance relationships with candidate WGCNA
modules associated with site maximum monthly temperature mean (MMM). Strength and
significance of each relationship is included above each scatter plots. Points represent individual
genes within each module.
147
Figure S4.23 Heatmap of expression of each gene within the Salmon module (rows) for each
coral ramet (column). Direction and intensity of differential gene regulation is proportional to
color from red (upregulated) to green (downregulated). Sample names show the site, outplant
number, array number, and genet identity. Ramets are ordered by increasing genet-specific risk
score confirmed by the barplot below showing risk of mortality for each sample.
148
Figure S4.24: Heatmap of expression of each gene within the Light Steel Blue1 module (rows)
for each coral ramet (column). Direction and intensity of differential gene regulation is
proportional to color from red (upregulated) to green (downregulated). Sample names show the
site, outplant number, array number, and genet identity. Ramets are ordered by increasing genet-
specific risk score confirmed by the bar plot below showing risk of mortality for each sample.
149
Chapter 5 Intraspecific variation in physiological plasticity in Acropora cervicornis
under thermal and acidification stress
Summary of Contribution
This chapter utilized experimental data that was generated in collaboration with the Mote Marine
Laboratory. The aquarium-based experiment was carried out in 2016 by Mote collaborators and
six of the twelve physiological traits presented in this chapter were measured at Mote before
coral fragments were sent to the University of Southern California. At USC, I carried out sample
processing and phenotyping for the remaining six physiological traits: chlorophyll concentration,
symbiont cell density, total soluble protein concentration, and the activity of four
immunochemical enzymes. The data was originally analyzed for the Proceedings of the Royal
Society publication (Muller et al. 2021) which focused on population-level effects of treatment
variables, tradeoffs between trait means, and heritability of physiological traits. This chapter co-
ops the physiological data collected by myself and Mote collaborators to analyze individual-level
variation in plasticity in response to treatment variables and the relationship between
physiological plasticity and fitness among individuals.
5.1 Abstract
While phenotypic plasticity in physiology can be beneficial by allowing organisms to
maintain or improve fitness as environments change via acclimation, deviations from baseline
trait values may also be symptomatic of reductions in fitness during stress. Organisms whose
physiological processes are tightly linked to environmental conditions (e.g. ectotherms) will
likely experience environment-dependent shifts in traits values but whether the plasticity is
beneficial or detrimental will likely be dependent on many factors including the type of
environmental change. Tropical corals will face numerous stressors as climates continue to
150
change and while physiological plasticity is frequently documented in coral, whether this
plasticity is adaptive, especially under scenarios coral will experience in the coming decades
(temperature and ocean acidification), is still unresolved. Here, clones of Acropora cervicornis
genotypes were exposed to experimental increases in temperature, acidification, and combined
temperature/acidification stress so that plasticity in twelve physiological traits could be measured
at the population and individual level. In addition to constitutive differences between individuals,
population-level changes in the majority of traits during experimental stress were accompanied
by decreases in symbiont cell density (bleaching). Intraspecific variation in growth (buoyant
weight) and dark calcification plasticity confirms the presence of intraspecific variation in
physiological plasticity, however, the ability to be plastic was not positively or negatively
associated with bleaching suggesting a neutral role for plasticity when A. cervicornis were
challenged with thermal and pH stress. Taken together, the constitutive (fixed) differences
between individuals seen here may be useful for predicting winners and losers as reefs become
warmer and more acidic, but differences in the ability to be physiologically plastic may be less
impactful if environment change is expected to mirror the conditions tested here.
5.2 Introduction
Organismal physiology describes the chemical and physical processes operating within
living systems and includes a wide range of functions from biomolecule production to
homeostatic regulation (Schmidt-Nielsen 1997). Physiological processes are often rate-
dependent and operate in a defined, sometimes narrow range of environmental conditions,
making these functions sensitive to changes in the internal and external environment in which
they occur ( Miller and Stillman 2012; Blackman 1905). The dependence of physiology on
environmental conditions has contributed to biological and ecological diversification by, for
151
example, defining species geographic ranges (Barve et al. 2014) or their capacity to deal with
environmental change (Somero 2010). Physiological responses to environmental change are
common across many systems (Storz, Scott, and Cheviron 2010; Valladares et al. 2002; Hoadley
et al. 2015; Breker, Gymrek, and Schuldiner 2013), however whether plasticity is a sign of
acclimation or a symptom of reduced fitness will likely be context dependent. Physiological
processes that are closely related to the fitness of an individual may show no change in response
to environmental variation, making a lack of plasticity in these traits indicative of climate
resistance, and thus beneficial. Metabolic rate in many birds and mammals, for example, is
maintained during high altitude exposure, although this regulation is a result of plasticity in a
suite of underlying traits that manipulate oxygen transport in the lungs, blood, and tissues (Storz,
Scott, and Cheviron 2010). In this case, abundant plasticity in these underlying traits are
signatures of an adaptive response that allows individuals to acclimatize to stress (Storz, Scott,
and Cheviron 2010). Conversely, physiological changes may arise in response to an
environmental change which lower the fitness of an individual in the new environment and
plasticity may instead be symptomatic of loss of regulation or degradation of fitness
representative of passive, maladaptive changes. For example, hypoxia-induced vasoconstriction
can maladaptively limit blood flow and strain the heart even as birds and mammals acclimate to
high altitude (Storz, Scott, and Cheviron 2010). The absence of plasticity in one trait but the
presence of adaptive and maladaptive plasticity in others underscores the importance of
integrating physiological responses across multiple traits in order to contextualize responses to
environmental change.
For species that rely on symbiosis to maintain fitness, the fate of all partners becomes
interdependent on the response of each taxon to the environmental change (Nielsen and Papaj
152
2022). Scleractinian corals engage in an obligate symbiosis with photosynthetic dinoflagellate
microalgae of the family Symbiodiniaceae (Davy, Allemand, and Weis. 2012). In this
relationship, coral hosts obtain up to 100% of their energetic requirements from their
photosymbionts (Muscatine and Marian 1981). Therefore, the fitness of coral individuals under
environmental stress is highly dependent on not only the response of the host but also on the
performance of photosymbionts and the stability of the nutritional symbiosis. The breakdown of
this symbiosis, marked by loss of photosymbiont cells from host tissues, is known as coral
bleaching (Brown 1997; Douglas 2003). While the exact mechanisms underlying bleaching are
still unresolved, environmental conditions such as increased temperature have been identified as
causes of coral bleaching worldwide (Brown 1997; Hoegh-Guldberg et al. 2007). Increased
ocean temperatures expected during climate change will push corals beyond their thermal
maximum (Manzello 2015). Meanwhile, compounding stressors such as ocean acidification,
caused by increased partial pressure of carbon dioxide (pCO2) in sea water, will create additional
physiological challenges for marine species (Tanhua et al. 2015; Gruber et al. 2019). The
detrimental effects of increased temperature on coral are well documented (Paradis, Henry, and
Chadwick 2019; Glynn and D’Croz 1990; McClanahan et al. 2007) and changes to ocean
chemistry are expected to impact chemical dependent processes such as calcification and
photosynthesis (Tambutté et al. 2015; Hoegh-Guldberg et al. 2007; Enochs et al. 2014; Brading
et al. 2011). As climate change alters the temperature and pH of marine habitats, understanding
the capacity of physiological plasticity to facilitate acclimation will be vital to determine how
and if coral will persist into the future.
Physiological plasticity in coral has been widely documented in nature with numerous
traits displaying environmental sensitivity following in situ transplantation (Rocker et al. 2019;
153
Ziegler et al. 2014; Drury and Lirman 2021; Barott et al. 2021).Temperature-induced plasticity
across taxa in photochemical efficiency (Hoadley et al. 2015; Keshavmurthy et al. 2021; Borell
and Bischof 2008; Wall et al. 2018), biochemical composition of hosts and symbionts (Hoadley
et al. 2015; Keshavmurthy et al. 2021; Wall et al. 2018), immune response (Wall et al. 2018),
and energy acquisition method (Borell and Bischof 2008) suggests coral physiology is sensitive
to thermal stress. Similarly, responses in physiological traits to acidification stress implies coral
also responds to changes in ocean pH expected to occur in the near future (Hoadley et al. 2015;
Albright and Langdon 2011; Putnam et al. 2013). Although coral physiology is clearly sensitive
to changes in temperature and pH, the potential for plasticity to be beneficial for individuals is
still unresolved. While changes to some traits, such as symbiont cell density, can be considered
maladaptive plasticity because it indicates a degradation of the symbiosis, plasticity in traits like
photochemical efficiency may indicate an adaptive response to mitigate the loss of symbionts
(Warner, Fitt, and Schmidt 1996; Mydlarz, McGinty, and Harvell 2010). This exemplifies the
interdependence of plasticity across organismal traits and the potential for adaptive and
maladaptive physiological plasticity to exist at the same time. Considering this, identifying
intraspecific variation in plasticity across multiple traits and quantifying trait-to-trait and trait-to-
fitness relationships will be essential steps in understanding the adaptive nature of physiological
plasticity in coral.
While plasticity in physiology appears to be abundant among coral populations, relatively
few studies have uncovered individual genotype-by-environment interactions (GxE) in coral
physiology (Barott et al. 2021; Drury and Lirman 2021; Howells et al. 2013), limiting our
understanding of ecological and evolutionary potential of plasticity. The presence of GxE
implies that physiological plasticity will vary among individuals and will create differences in
154
the amount of variation in trait means present in populations from environment to environment.
For species threatened by climate change, such as Acropora cervicornis, projections of
performance and persistence under increased temperature or pCO2 will be unreliable if we are
unable to account for genotype-dependent responses to environmental pressures. GxE and its
relationship to fitness will be crucial to identifying the adaptive capacity of plasticity itself.
While selection can maintain or enhance acclimatization abilities, costs of plasticity between
traits or environments that result in decreased fitness can lead to the evolution of decreased
plasticity (Dewitt, Sih, and Wilson 1998; Gavrilets and Scheiner 1993; Hendry 2016). GxE in
bleaching susceptibility of A. cervicornis across natural habitats suggest bleaching is not
dependent on genetic or environmental factors alone and the potential for thermal tolerance
exists within populations (Drury and Lirman 2021). Considering that coral bleaching is a higher
order trait likely controlled by numerous host and symbiont processes (Denis et al. 2017; Weis
2019), understanding the extent of GxE in physiology will help clarify the mechanisms that are
promoting acclimation in A. cervicornis.
Anthropogenic climate change has committed oceans to increases in temperature and pH
that will challenge the persistence of coral reefs around the world (Mauritsen and Pincus 2017).
For A. cervicornis, which has already experienced significant declines in recent decades (Cramer
et al. 2020), plasticity-mediated acclimation to temperature and pH will be vital for buying time
for wild and restored populations. Therefore, the relationship between individual plasticity and
fitness under climate change conditions is of particular concern but a dearth of evidence for GxE
in physiology limits our understanding of its adaptive potential. To address these gaps in
knowledge, a chronic experimental exposure of A. cervicornis genotypes to temperature,
acidification, and combined temperature and pH stress was used to quantify genotypic,
155
environmental, and genotype-by-treatment (GxE) effects on coral physiology. Relationships
between genotype values of plasticity across traits and treatments were used to identify if
increased plasticity was associated with decreased bleaching and if plasticity in one trait or to
one stress came at the cost of responsiveness in another trait or environment.
5.3 Methods
5.3.1 Experimental Design
In the summer of 2016, an ex situ experiment was conducted at Mote Marine
Laboratory’s International Center for Coral Reef Research and Restoration to assess the
physiological response of 12 A. cervicornis genotypes under elevated temperature, pCO2, and
combined stress conditions. Twenty fragments (~5 cm) from each genotype were harvested from
nursery raised colonies maintained in Mote's offshore, in situ, coral nursery, located in the lower
Florida Keys (24.56257° N, 81.40009° W). Coral fragments were first acclimated to Mote’s
Climate and Acidification Ocean Simulator system for a week under in situ temperatures (30.35
± 0.2°C) before the treatment temperature and pCO2 conditions were gradually changed over a
period of 2 and 7 days, respectively. Five replicate tanks per treatment were adjusted to ambient
pCO2 and temperature (704 ± 62 µatm pCO2, 27.1 ± 0.05°C), elevated pCO2 (1225 ± 98 µatm
pCO2, 27.0 ± 0.02°C), elevated temperature (798 ± 62 µatm pCO2, 31.0 ± 0.04°C) and
combined elevated pCO2 and temperature (1412 ± 90 µatm pCO2, 31.1 ± 0.05°C). Additional
details of experimental pCO2 and temperature conditions can be found in Muller et al. (2021).
Corals were maintained in experimental conditions for 2 months before subsequent processing
for physiological phenotypes.
156
5.3.2 Physiological phenotyping
Buoyant weight and two photochemical efficiency traits were measured for each
fragment before and after exposure. A diving PAM (Walz, Effeltrich, Germany) was used to
measure maximal quantum yield of PS II (Δ yield) and maximum electron transport rate
(ΔETRm) of algal symbionts. Photosynthesis, respiration, and calcification were measured
immediately after the 2-month exposure (see Muller et al. 2021 for additional details). Fragments
were then snap-frozen and sent to the University of Southern California for subsequent
physiological processing. Coral tissue was removed from the skeleton with a pressurized
airbrush into a known volume of buffer (.5mM DTT, .1M TRIS pH 8.0) kept at 4°C. The
resulting tissue slurry was then homogenized to release algal cells from host tissue and aliquoted
into separate tubes for measurements of chlorophyll concentrations and symbiont cell densities,
respectively. The remaining tissue homogenate was centrifuged to separate host tissue from
symbiont cells. Aliquots of host tissue were stored at -80°C until analyzed for immunochemical
concentrations. Additional symbiont cells were resuspended in the elution buffer and stored at -
80°C. Surface area of coral fragments was obtained by dipping clean and dry coral skeletons into
paraffin wax and quantifying weight of wax added as described in Stimson & Kinzie (Stimson
and Kinzie 1991).
One milliliter of holobiont tissue slurry was centrifuged and the resulting algal pellet was
resuspended in 90% acetone at -4°C in preparation for quantifications of chlorophyll absorbance
(Ritchie 2008). Chlorophyll absorbance at 630, 647, and 664 nm was measured on a SynergyH1
spectrophotometer (BioTek Instruments, Inc., Winooski, VT) and used to calculate Chlorophyll
a, Chlorophyll c, and Total Chlorophyll concentrations (Ritchie 2008). For symbiont cell
densities, tissue homogenate (0.25 ml) was fixed with 0.25ml of 10% formalin to preserve
157
zooxanthellae cells at room temperature. Cell counts using a Neubauer-Improved hemocytometer
(Hausser Scientific, Horsham, PA) were completed to determine the number of algal cells per
milliliter. Host protein concentration was measured using the BioVision BCA Protein Assay Kit
II to quantify sample absorbance at 750nm (e.g. Leuzinger, Anthony, and Willis 2003).
Spectrophotometer- based assays were used to determine peroxidase (POX), prophenoloxidase
(PPO), and phenoloxidase (PO) activity following previously described protocols (Mydlarz and
Palmer 2011). Symbiont cell densities, chlorophyll concentrations and total soluble protein
content were standardized to coral surface area and POX, PPO, and PO were standardized by the
protein content and surface area of each fragment.
5.3.3 Statistical analysis
All data manipulation and statistical analyses were performed in R version 4.1.2 (R Core
Team 2020). Physiological responses were described with twelve response variables: change in
buoyant weight, light and dark calcification, photosynthesis-to-respiration ratio, total soluble
protein concentration, POX, PPO, and PO activity, total chlorophyll concentration, symbiont cell
density, change in maximum quantum yield (Fv/Fm), and change in maximum electron transport
rate (ETRm). Changes in buoyant weight, Fv/Fm, and ETRm were calculated as the differences
between pre- and post-exposure values and all other traits represent endpoint measurements.
Effects of genotype, treatment, and genotype-by-treatment interactions (referred to as GxE here)
were tested using linear mixed effects models in lme4 (Bates et al. 2014). Genotype, treatment,
and their interaction were included as fixed effects while experimental tank was included as a
random effect. Pairwise differences in least squares means between treatments were tested using
lmerTest (Kuznetsova et al. 2020) when significant treatment effects were detected in linear
mixed effect models.
158
When measured across two environments a genotype’s plasticity can be measured as the
slope of its reaction norm or simply the difference between mean trait values across each
environment (Scheiner and Lyman 1989; Via and Lande 1985). Therefore, genotypic responses
to each stress treatment were calculated as the difference between the genotype average trait
value in the control and in the high temperature, high pCO2, or combined high temperature and
pCO2 treatments. Genotypic plasticity values were then correlated across traits and exposure
treatments with Pearson’s Correlations to identify costs or benefits to physiological plasticity.
5.4 Results
5.4.1 Holobiont productivity
The change in buoyant weight and dark calcification rate of coral fragments was
significantly different among experimental treatments (p < 0.01, Fig 5.1) but was not affected by
genotype or the interaction of genotype and treatment (p > 0.05). The change in buoyant weight
was significantly lower in the combined treatment compared to the control, high temperature,
and high pCO2 conditions (p < 0.05). Dark calcification was depressed in all stress conditions
compared to the control (p < 0.001) and in the combined stress condition compared to the high
temperature treatment (p = 0.037). The ratio of photosynthesis to respiration (P:R) varied among
genotypes overall but was not different among treatments or among genotype-treatment
combinations (p > 0.05). Calcification rate in the light showed significant genotype, treatment
and GxE effects (p < 0.05) with rates declining in stress conditions compared to the control (p <
0.001). Light calcification in the combined stress treatment was significantly lower than in the
high pCO2 treatment (p = 0.024) but similar to the high temperature treatment (p > 0.05).
159
Figure 5.1: Holobiont physiological responses to experimental treatments. Box plots show
overall distribution of treatment values for each trait, bolded lines indicate the median value,
boxes show 1st and 3rd quartiles, and whiskers show range of distributions. Points represent
genotype averages in each treatment with lines included to visualize genotypic reaction norms.
Letters indicate significant pairwise differences between treatments (p < 0.05) if overall
treatment effects (T) were observed in linear mixed models.
5.4.2 Host protein modulation
The concentration of total soluble protein in host tissue varied among genotypes and
treatments (p = 0.009 and p = 5.8e-4, respectively). Protein concentration after the two-month
exposure was similar between control and high pCO2 treatments (p > 0.05) but was significantly
different between the high temperature treatment and the high pCO2 treatment (p = 0.013, Fig.
5.2). Protein concentration was significantly lower in the combined treatment than both the high
pCO2 and high temperature treatments (p < 0.001 and p = 0.04, respectively, Fig. 5.2).
Immunological protein activity (PPO, PO, POX) varied among genotypes (p < 0.0001) but
showed no significant effect of treatment or GxE (p > 0.05, Fig. 5.2).
160
Figure 5.2: Host physiological responses to experimental treatments. Box plots show overall
treatment values for each trait. Bolded line indicates the median value, boxes show 1st and 3rd
quartiles, and whiskers show range of distributions. Points represent genotype averages in each
treatment with lines included to visualize genotypic reaction norms. Letters indicate significant
pairwise differences between treatments (p < 0.05) if overall treatment effects (T) were observed
in linear mixed models.
5.4.3 Symbiont physiology
Total chlorophyll concentration, symbiont cell density, change in maximum quantum
yield (Fv/Fm), and change in maximum electron transport rate (ETRm) show variation among
host genotypes (p = 2.31 e-5, p = 0.003, p = 2.45 e-7, and p = 6.65 e-4; respectively). Total
chlorophyll concentration and symbiont cell density varied among treatments (p = 0.003, p=
0.005), however pairwise comparisons showed significant declines in these traits were only
present in combined stressor treatment compared to all other treatments (p < 0.05) while
remaining treatments were similar to each other (Fig. 5.3). Changes in the two photochemical
efficiency traits, Fv/Fm and ETRm, showed similar variation across treatments with significant
decreases from the control to all stressor treatments (p < 0.05) followed by further decreases in
the combined treatment compared to high temperature and high pCO2 treatments (p = 8.20e-5
161
and p = 1.193e-4; respectively). Despite the consistent genotype and treatment effects across
symbiont-specific physiological traits, genotype-by-treatment interactions were only observed in
the change in Fv/Fm over the two-month exposure period (p = 0.005).
Figure 5.3: Symbiont physiological responses to experimental treatments. Box plots show overall
treatment values for each trait, white bolded line indicating the median value, boxes showing 1st
and 3rd quartiles, and whiskers show range of distributions. Points represent genotype averages
in each treatment with lines included to visualize genotypic reaction norms. Letters indicate
significant pairwise differences between treatments (p < 0.05) if overall treatment effects were
observed in linear mixed models.
5.4.4 Relationship of physiology plasticity across traits and treatments
Genotype-by-treatment interactions (GxE) indicate that the responses to temperature,
acidification, and combined stresses depend on the genotype of the fragment. While the majority
of physiological traits showed no variation among individuals, GxE was evident in two
physiological traits: Fv/Fm (p = 0.005) and light calcification (p = 0.022). Genotypic values of
plasticity in light calcification and plasticity in Fv/Fm were positively correlated when calculated
as the change between the control and high pCO2 treatments (R = 0.59, p = 0.044, Fig. 5.4a).
However, plastic responses to high temperature and combined stress exposure were not
162
correlated across traits (high temperature: R = -0.074, p = 0.82, combined: R = 0.013, p = 0.97;
Fig. 5.4). Genotypic levels of plasticity in light calcification and Fv/Fm were not significantly
correlated with changes in symbiont cell density, a metric of bleaching, during exposure to any
of the stress treatments.
Plastic responses in a given trait tended to be weakly correlated across treatments (Fig.
S5.1). Plastic responses in exposure to high temperature for light calcification were weakly,
positively correlated with plastic responses to combined stressors (R = 0.31, p = 0.32). Similarly,
light calcification plasticity when exposed to high pCO2 was weakly, positively correlated with
plasticity when exposed to combined stressors (R = 0.27, p = 0.4). Plasticity in Fv/Fm in high
pCO2 was significantly correlated with plasticity in response to combined exposure (R = 0.69, p
= 0.013) but only weak relationships in Fv/Fm plasticity were seen in high pCO2 vs. high
temperature comparisons (R = 0.19, p = 0.69) and high temperature vs. combined exposures (R =
0.49, p = 0.104).
163
Figure 5.4: Relationship of genotypic values of physiological plasticity to high pCO2, high
temperature, or combined stress across traits. Scatter plots show linear relationship between
plasticity in light calcification, Fv/Fm, or symbiont cell density with shaded 95% confidence
intervals colored by direction of exposure for which plasticity was measured. Solid lines of best
fit are present for significant relationships while dashed lines indicate nonsignificant
relationships.
5.5 Discussion
5.5.1 Population-level plasticity to acidification, temperature and combined stress
Responses of A. cervicornis to changes in seawater pCO2, temperature, or the combined
elevated pCO2 and temperature stress add to the abundant evidence for population-level
physiological plasticity in coral (Rocker et al. 2019; Ziegler et al. 2014; Drury and Lirman 2021;
Barott et al. 2021; Hoadley et al. 2015; Borell and Bischof 2008; Wall et al. 2018). Of the traits
that were used to describe holobiont physiology, the change in buoyant weight and calcification
164
rates in the light and dark showed moderate declines as coral were exposed to acidification or
temperature stress, while the largest declines were evident in the combined stress treatment (Fig.
5.1). Calcification is an energy intensive process that can become more expensive under acidic
conditions (Allemand et al. 2011; Cohen and Holcomb 2009) so the reduction in skeletal growth
seen here may represent the strain on energy budgets caused by acidification and thermal stress.
Considering survival and fecundity are size-dependent (Madin et al. 2014; Babcock 1991;
Álvarez-Noriega et al. 2016), reductions in growth are maladaptive and likely symptomatic of a
degraded physical state. Similarly, symbiont traits showed significant decreases in the combined
acidification and temperature treatment, although declines were less pronounced in single stress
conditions (Fig. 5.3). Reductions in photosynthetic efficiency traits like Fv/Fm and ETRmax
may indicate an increase in non-photochemical quenching, a process of energy dissipation that
prevents free-radical formation, suggesting plasticity in these traits may serve a beneficial role
(Murchie and Lawson 2013; Mydlarz, McGinty, and Harvell 2010). Alternatively, decreases in
Fv/Fm and ETRmax are associated with damage to photosynthetic machinery that has been
implicated in the onset of bleaching (Bhagooli and Hidaka 2006; Warner, Fitt, and Schmidt
1999). However, because plasticity in Fv/Fm was not correlated with the maintenance or loss of
symbiont cell density (Section 5.5.2) these changes in photochemical efficiency do not appear to
serve an immediate adaptive or maladaptive role. Potential for lags between declines in Fv/Fm
and alterations in bleaching response may also contribute to the lack of relationship between
these traits. While the single endpoint phenotyping used here cannot address the presence of time
lags between plasticity and changes in fitness, understanding relevant timescales for the costs
and benefits of plasticity will be important considering the timing of environmental triggers,
165
responses, and selection events are all expected to influence its ecology and evolution (Bitter et
al. 2021).
Total protein content was the only host-specific trait to show significant effects of
environmental change with the largest declines occurring in the combined acidification and
temperature treatment (Fig. 5.2). These results are consistent with previous studies showing the
loss of coral protein, along with lipids and carbohydrates, as consequences of heat and
acidification stress (Cziesielski, Schmidt-Roach, and Aranda 2019; Schoepf et al. 2013). While
losses in total protein would be expected under stress, fluctuations in the production of specific
enzymes, such as prophenoloxidase (PPO) may facilitate response to thermal challenges
(Mydlarz, McGinty, and Harvell 2010; Mydlarz et al. 2009). Similarly, antioxidants, like
peroxidase (POX), can mitigate damage from reactive oxygen species produced during periods
of stress (Mydlarz, McGinty, and Harvell 2010; Yakovleva et al. 2004). Despite these potential
roles in acclimation, there were no overall changes to these immunological proteins during
acidification, temperature, or combined stress. The lack of treatment effect on immunological
proteins may be a result of the experimental exposure length. Innate immune responses in coral
are often rapid (van de Water et al. 2015; Palmer and Traylor-Knowles 2012) and may therefore
be an important mechanism for immediate acclimation but will not be maintained during
prolonged stress. Active enzymes, like POX and phenoloxidase (PO) are expected to have short
half-lives (van de Water et al. 2016; Lesser 2006) and therefore plastic changes in their
production will likely disappear over time leaving only signals of variation in constitutive
production among individuals. Reserves of antioxidants may also be less necessary after two
month of stress, especially if increases in non-photochemical quenching within the symbiont
166
were capable of slowing the production of reactive oxygen species (Lesser 2006; Mydlarz,
McGinty, and Harvell 2010).
Taken together, population-level physiological plasticity seen here across holobiont, host,
and symbiont traits does not appear to represent active regulation of processes that are
facilitating acclimation to temperature and acidification stress. Rather, because the majority of
the observed responses are in the direction seen by corals with degrading health, these overall
changes in physiology appear to be passive, maladaptive plasticity. However, it may be
important to consider that although plasticity appears to direct trait values away from fitness
optima during environmental change, mortality was extremely low throughout the 2-month
exposure period even though A. cervicornis fragments were at or above their critical bleaching
threshold of 31.5℃ in some treatments (Manzello, Berkelmans, and Hendee 2007). With this in
mind, cessation of growth and calcification along with reductions in total protein may be
representative of differential allocation of energy between fitness related traits (i.e. growth vs
survival) in order to persist until abiotic conditions promote recovery. The ability to decrease
growth rate may be beneficial especially considering the growing evidence for trade-offs
between holobiont growth and immune response (Schlecker et al. 2022) or bleaching
susceptibility (Cunning et al. 2015; Cornwell et al. 2021). Although the mechanisms underlying
these changes will be difficult to ascertain from this experiment, variation in plasticity among
individuals may provide insight into the overall role of physiological plasticity for A. cervicornis.
5.5.2 Intraspecific variation in physiology
There was significant variation among genotypes in their mean phenotype for all
physiological traits except change in buoyant weight and calcification in the dark, which is
consistent with previous studies in A. cervicornis (Drury, Manzello, and Lirman 2017; Lohr and
167
Patterson 2017). Genotypic differences in mean trait values across experimental treatments
indicates the existence of genetic variation that may contribute to the evolution and adaptation of
A. cervicornis under climate change. The concepts of intraspecific variation and heritability of
mean trait values are extensively discussed in Muller et al. (2021). Beyond the intergenerational
impacts, strong genotype effects in physiology paired with a lack of genotype-by-treatment
interactions have meaningful implications for corals dealing with environmental change
occurring over a single generation. The absence of GxE in the majority of physiological traits
suggests that plastic responses to acidification and temperature stress are consistent across this
population of A. cervicornis. Predictions of future success for individuals may then be possible
because genotypes with relatively high values for fitness related traits, such as symbiont cell
density, in benign conditions appear to maintain their relative rank within the population even
though population-wide reductions may occur during climate change. Restoration of A.
cervicornis may utilize these results to help prioritize individuals for ongoing outplanting and
assisted evolution strategies that rely on standing genetic variation to strengthen the tolerance of
current and future populations (Young, Schopmeyer, and Lirman 2012; van Oppen et al. 2015).
Quantifying the amount of intraspecific variation in physiological plasticity will also be
important for predicting how acclimation ability in A. cervicornis may evolve over time (Hendry
2016). There was no influence of GxE in ten out of twelve physiological traits under
acidification, thermal, or combined stress suggesting limited adaptive potential of physiological
plasticity. Contrary to the majority of traits, light calcification rate and Fv/Fm, did display
significant genotype-by-environment interactions, contributing to the growing body of evidence
supporting the role of GxE in determining coral phenotypes (Barott et al. 2021; Drury and
Lirman 2021). Although intraspecific variation in plasticity of these two traits suggests a
168
potential to evolve, identifying benefits and costs associated with calcification and Fv/Fm
plasticity are necessary for understanding the trajectory of change over time. One form of a
potential cost is a trade-off between plasticity across traits or across types of environmental
stressors (Dewitt, Sih, and Wilson 1998; Murren et al. 2015; van Heerwaarden and Kellermann
2020). The correlation analysis showed no tradeoffs between plasticity across traits in response
to any experimental treatment and in fact, plasticity in calcification and in Fv/Fm were positively
correlated when corals were exposed to high pCO2 in isolation. Similarly, the ability to respond
to each type of environmental stress did not come at the cost of plasticity in response to any other
stressor. Instead, relative sensitivity to one condition tended to be positively correlated with
sensitivity to all other conditions (Fig. S5.1). The lack of trade-offs in trait plasticity mirrors the
positive relationships between trait means across-treatments seen here (Muller et al. 2021) and in
other Acroporids (Wright et al. 2019).
The relationship between plasticity and fitness is especially important when considering
the benefits or costs of being plastic. Here, genotypic values for plasticity in calcification and
Fv/Fm were compared to values of bleaching (i.e. loss of symbiont cells) which is a common
proxy for coral fitness under stress (Douglas 2003; Glynn 1993). There were no significant
correlations between physiological plasticity and loss of symbiont cell density during exposure to
any experimental treatment (Fig. 5.4) suggesting no net benefit or cost to being plastic in these
traits. Previous evidence implicating declines in Fv/Fm in the onset of bleaching, measured by
changes in symbiont cell density and chlorophyll a concentration (Warner, Fitt, and Schmidt
1999), and calcification rate with impaired growth (D’Olivo and McCulloch 2017; Carricart-
Ganivet et al. 2012) suggest plasticity in the direction seen here should be maladaptive.
However, the lack of strong correlations indicated that plasticity in these traits is neither adaptive
169
nor maladaptive under the high temperature, high pCO2, or combined stress conditions used here.
This may be due, in part, to the population-level bleaching responses that dictated individual
responses. Taken together, these results suggest that while intraspecific variation in physiological
plasticity exists in some traits, the lack of benefits or costs will likely prevent directional
selection on plasticity itself, at least under the static mean conditions tested here. However,
present and future environmental conditions are predicted to increase in both the mean and
variability, so understanding the role of adaptive plasticity in response to the frequency and
predictability of environmental change is an important next step (Bitter et al. 2021). Intraspecific
variation in physiological plasticity present within A. cervicornis populations may become more
important for the evolution and adaptation of the species as the type of environmental variation
experienced by coral shifts under climate change.
5.5.3 Conclusions
The phenotypic plasticity observed here in response to elevated temperature, pCO2 and
combined stressors suggests A. cervicornis physiology is particularly susceptible to changes in
abiotic conditions. While this plasticity appeared to move population trait averages away from
local optima, indicating maladaptive plasticity, individual coral genotypes were able to maintain
high survival suggesting population-level plasticity help to manage tradeoffs between physiology
performance and overall survival. While the majority of traits only showed overall treatment and
genotype effects, GxE in light calcification and Fv/Fm suggest intraspecific variation in
physiological plasticity exists in A. cervicornis. The ability to be plastic had no apparent trade-
offs or any effect on coral bleaching in the context of this study, suggesting physiological
plasticity in this system is neutral. Although selection on plasticity in nature will likely be driven
by increased variability in addition to the static changes tested here, this experiment provides
170
insight into the role of physiological plasticity underlying acclimation to climate change. For
example, GxE in calcification rate may help explain the abundant GxE in skeletal morphology
seen in similar genotypes of A. cervicornis (Chapter 3) and in numerous other species (Bruno
and Edmunds 1997; Todd 2008). Future experiments testing GxE under fluctuating
environmental variation that more closely mimic natural variation will be a vital next step for
understanding the adaptive potential of physiological plasticity under climate change.
5.6 Supplementary Material
Figure S5.1: Relationship between plasticity in all traits showing GxE across treatments. Each
scatter plot shows the linear relationship between plasticity in one trait and environmental
change and the other trait and environmental change. Corresponding R values for each linear
relationship are included in the upper right portion of the figure. Red asterisks indicate
significant relationships.
171
References
Agudo-Adriani, Esteban A., Jose Cappelletto, Francoise Cavada-Blanco, and Aldo Croquer.
2016. “Colony Geometry and Structural Complexity of the Endangered Species Acropora
Cervicornis Partly Explains the Structure of Their Associated Fish Assemblage.” PeerJ 4
(April): e1861.
Aguirre, Emily G., Wyatt C. Million, Erich Bartels, Cory J. Krediet, and Carly D. Kenkel. 2022.
“Host-Specific Epibiomes of Distinct Acropora Cervicornis Genotypes Persist after Field
Transplantation.” Coral Reefs 41 (2): 265–76.
Ahn, Seongho, Kevin A. Haas, and Vincent S. Neary. 2020. “Dominant Wave Energy Systems
and Conditional Wave Resource Characterization for Coastal Waters of the United States.”
Energies 13 (12): 3041.
Ainsworth, T. D., O. Hoegh-Guldberg, S. F. Heron, W. J. Skirving, and W. Leggat. 2008. “Early
Cellular Changes Are Indicators of Pre-Bleaching Thermal Stress in the Coral Host.”
Journal of Experimental Marine Biology and Ecology 364 (2): 63–71.
Albright, Rebecca, and Chris Langdon. 2011. “Ocean Acidification Impacts Multiple Early Life
History Processes of the Caribbean Coral Porites Astreoides.” Global Change Biology 17
(7): 2478–87.
Allemand, Denis, Éric Tambutté, Didier Zoccola, and Sylvie Tambutté. 2011. “Coral
Calcification, Cells to Reefs.” In Coral Reefs: An Ecosystem in Transition, edited by Zvy
Dubinsky and Noga Stambler, 119–50. Dordrecht: Springer Netherlands.
Alvarez-Filip, Lorenzo, Isabelle M. Côté, Jennifer A. Gill, Andrew R. Watkinson, and Nicholas
K. Dulvy. 2011. “Region-Wide Temporal and Spatial Variation in Caribbean Reef
Architecture: Is Coral Cover the Whole Story?” Global Change Biology 17 (7): 2470–77.
Alvarez-Filip, Lorenzo, Nicholas K. Dulvy, Jennifer A. Gill, Isabelle M. Côté, and Andrew R.
Watkinson. 2009. “Flattening of Caribbean Coral Reefs: Region-Wide Declines in
Architectural Complexity.” Proceedings. Biological Sciences / The Royal Society 276
(1669): 3019–25.
Álvarez-Noriega, Mariana, Andrew H. Baird, Maria Dornelas, Joshua S. Madin, Vivian R.
Cumbo, and Sean R. Connolly. 2016. “Fecundity and the Demographic Strategies of Coral
Morphologies.” Ecology 97 (12): 3485–93.
Aronson, R. B., A. W. Bruckner, J. Moore, W. F. Precht, and E. Weil. 2008. “IUCN Red List of
Threatened Species: Acropora Cervicornis.” Cambridge: International Union for
Conservation of Nature and Natural Resources.
Auld, Josh R., Anurag A. Agrawal, and Rick A. Relyea. 2010. “Re-Evaluating the Costs and
Limits of Adaptive Phenotypic Plasticity.” Proceedings. Biological Sciences / The Royal
172
Society 277 (1681): 503–11.
Babcock, Russell C. 1991. “Comparative Demography of Three Species of Scleractinian Corals
Using Age‐ and Size‐Dependent Classifications.” Ecological Monographs.
https://doi.org/10.2307/2937107.
Barley, Jordanna M., Brian S. Cheng, Matthew Sasaki, Sarah Gignoux-Wolfsohn, Cynthia G.
Hays, Alysha B. Putnam, Seema Sheth, Andrew R. Villeneuve, and Morgan Kelly. 2021.
“Limited Plasticity in Thermally Tolerant Ectotherm Populations: Evidence for a Trade-
Off.” Proceedings. Biological Sciences / The Royal Society 288 (1958): 20210765.
Barott, Katie L., Ariana S. Huffmyer, Jennifer M. Davidson, Elizabeth A. Lenz, Shayle B.
Matsuda, Joshua R. Hancock, Teegan Innis, Crawford Drury, Hollie M. Putnam, and Ruth
D. Gates. 2021. “Coral Bleaching Response Is Unaltered Following Acclimatization to
Reefs with Distinct Environmental Conditions.” Proceedings of the National Academy of
Sciences of the United States of America 118 (22).
https://doi.org/10.1073/pnas.2025435118.
Barshis, Daniel J., Jason T. Ladner, Thomas A. Oliver, François O. Seneca, Nikki Traylor-
Knowles, and Stephen R. Palumbi. 2013. “Genomic Basis for Coral Resilience to Climate
Change.” Proceedings of the National Academy of Sciences of the United States of America
110 (4): 1387–92.
Barve, Narayani, Craig Martin, Nathaniel A. Brunsell, and A. Townsend Peterson. 2014. “The
Role of Physiological Optima in Shaping the Geographic Distribution of Spanish Moss.”
Global Ecology and Biogeography: A Journal of Macroecology 23 (6): 633–45.
Bates, Douglas, Martin Mächler, Ben Bolker, and Steve Walker. 2014. “Fitting Linear Mixed-
Effects Models Using lme4.” arXiv [stat.CO]. arXiv. http://arxiv.org/abs/1406.5823.
Bates, Douglas, Martin Maechler, Ben Bolker, Steven Walker, Rune Haubo Bojesen
Christensen, Henrik Signmann, Bin Dai, et al. 2022. Linear Mixed-Effects Models Using
“Eigen” and S4 (version 1.1-28). https://cran.r-project.org/web/packages/lme4/lme4.pdf.
Bay, Line K., Aurélie Guérécheau, Nikos Andreakis, Karin E. Ulstrup, and Mikhail V. Matz.
2013. “Gene Expression Signatures of Energetic Acclimatisation in the Reef Building Coral
Acropora Millepora.” PloS One 8 (5): e61736.
Bay, Line K., Karin E. Ulstrup, H. Bjørn Nielsen, Hanne Jarmer, Nicolas Goffard, Bette L.
Willis, David J. Miller, and Madeleine J. H. Van Oppen. 2009. “Microarray Analysis
Reveals Transcriptional Plasticity in the Reef Building Coral Acropora Millepora.”
Molecular Ecology 18 (14): 3062–75.
Bay, Rachael A., and Stephen R. Palumbi. 2015. “Rapid Acclimation Ability Mediated by
Transcriptome Changes in Reef-Building Corals.” Genome Biology and Evolution 7 (6):
1602–12.
173
———. 2017. “Transcriptome Predictors of Coral Survival and Growth in a Highly Variable
Environment.” Ecology and Evolution 7 (13): 4794–4803.
Bellantuono, Anthony J., Camila Granados-Cifuentes, David J. Miller, Ove Hoegh-Guldberg,
and Mauricio Rodriguez-Lanetty. 2012. “Coral Thermal Tolerance: Tuning Gene
Expression to Resist Thermal Stress.” PloS One 7 (11): e50685.
Bhagooli, Ranjeet, and Michio Hidaka. 2006. “Thermal Inhibition and Recovery of the
Maximum Quantum Yield of Photosystem II and the Maximum Electron Transport Rate in
Zooxanthellae of a Reef-Building Coral.” Journal of the Japanese Coral Reef Society 8 (1):
1–11.
Bitter, M. C., J. M. Wong, H. G. Dam, S. C. Donelan, C. D. Kenkel, L. M. Komoroske, K. J.
Nickols, et al. 2021. “Fluctuating Selection and Global Change: A Synthesis and Review on
Disentangling the Roles of Climate Amplitude, Predictability and Novelty.” Proceedings.
Biological Sciences / The Royal Society 288 (1957): 20210727.
Blackman, F. F. 1905. “Optima and Limiting Factors.” Annals of Botany 19 (74): 281–95.
Boersma, M., P. Spaak, and L. De Meester. 1998. “Predator-Mediated Plasticity in Morphology,
Life History, and Behavior of Daphnia: The Uncoupling of Responses.” The American
Naturalist 152 (2): 237–48.
Borell, Esther M., and Kai Bischof. 2008. “Feeding Sustains Photosynthetic Quantum Yield of a
Scleractinian Coral during Thermal Stress.” Oecologia 157 (4): 593–601.
Bottjer, David J. 1980. “Branching Morphology of the Reef Coral Acropora Cervicornis in
Different Hydraulic Regimes.” Journal of Paleontology 54 (5): 1102–7.
Bowden-Kerby, A. 2008. “Restoration of Threatened Acropora Cervicornis Corals: Intraspecific
Variation as a Factor in Mortality, Growth, and Self-Attachment.” In Proceedings of the
11th International Coral Reef Symposium, 2:1200–1204. Citeseer.
Brading, Patrick, Mark E. Warner, Phillip Davey, David J. Smith, Eric P. Achterberg, and David
J. Suggett. 2011. “Differential Effects of Ocean Acidification on Growth and
Photosynthesis among Phylotypes of Symbiodinium(Dinophyceae).” Limnology and
Oceanography 56 (3): 927–38.
Breker, Michal, Melissa Gymrek, and Maya Schuldiner. 2013. “A Novel Single-Cell Screening
Platform Reveals Proteome Plasticity during Yeast Stress Responses.” The Journal of Cell
Biology 200 (6): 839–50.
Brener-Raffalli, Kelly, Jeremie Vidal-Dupiol, Mehdi Adjeroud, Olivier Rey, Pascal Romans,
François Bonhomme, Marine Pratlong, et al. 2022. “Gene Expression Plasticity and
Frontloading Promote Thermotolerance in Pocillopora Corals.” Peer Community Journal 2
174
(e13). https://doi.org/10.24072/pcjournal.79.
Briceño, Henry O., and Joseph N. Boyer. 2019. “2018 Annual Report of the Water Quality
Monitoring Project for the Water Quality Protection Program of the Florida Keys National
Marine Sanctuary,” SERC Research Reports, . https://digitalcommons.fiu.edu/sercrp/117/.
Brown, B. E. 1997. “Coral Bleaching: Causes and Consequences.” Coral Reefs 16 (1): S129–38.
Bruno, John F., and Peter J. Edmunds. 1997. “Clonal Variation for Phenotypic Plasticity in the
Coralmadracis Mirabilis.” Ecology 78 (7): 2177–90.
Burns, Jhr, D. Delparte, R. D. Gates, and M. Takabayashi. 2015. “Integrating Structure-from-
Motion Photogrammetry with Geospatial Software as a Novel Technique for Quantifying
3D Ecological Characteristics of Coral Reefs.” PeerJ 3 (July): e1077.
Bythell, J., P. Pan, and J. Lee. 2001. “Three-Dimensional Morphometric Measurements of Reef
Corals Using Underwater Photogrammetry Techniques.” Coral Reefs 20 (3): 193–99.
Calle-Triviño, Johanna, Camilo Cortés-Useche, Rita Inés Sellares-Blasco, and Jesús Ernesto
Arias-González. 2018. “Assisted Fertilization of Threatened Staghorn Coral to Complement
the Restoration of Nurseries in Southeastern Dominican Republic.” Regional Studies in
Marine Science 18 (February): 129–34.
Camp, Emma F., Verena Schoepf, and David J. Suggett. 2018. “How Can ‘Super Corals’
Facilitate Global Coral Reef Survival under Rapid Environmental and Climatic Change?”
Global Change Biology 24 (7): 2755–57.
Carricart-Ganivet, Juan P., Nancy Cabanillas-Terán, Israel Cruz-Ortega, and Paul Blanchon.
2012. “Sensitivity of Calcification to Thermal Stress Varies among Genera of Massive
Reef-Building Corals.” PloS One 7 (3): e32859.
Cesar, Herman, Lauretta Burke, and And Lida Pet-Soede. 2003. “The Economics of Worldwide
Coral Reef Degradation.” International Coral Reef Action Network.
https://agris.fao.org/agris-search/search.do?recordID=GB2013202743.
Chan, Neil C. S., Daniel Wangpraseurt, Michael Kühl, and Sean R. Connolly. 2016. “Flow and
Coral Morphology Control Coral Surface pH: Implications for the Effects of Ocean
Acidification.” Frontiers in Marine Science 3. https://doi.org/10.3389/fmars.2016.00010.
Cheviron, Zachary A., Gwendolyn C. Bachman, and Jay F. Storz. 2013. “Contributions of
Phenotypic Plasticity to Differences in Thermogenic Performance between Highland and
Lowland Deer Mice.” The Journal of Experimental Biology 216 (Pt 7): 1160–66.
Christensen, Rune Haubo Bojesen. 2019. Regression Models for Ordinal Data (version 2019.12-
10). https://cran.r-project.org/web/packages/ordinal/ordinal.pdf.
175
Cignoni, P., M. Callieri, M. Corsini, M. Dellepiane, F. F. Ganovelli, and G. Ranzuglia. 2008.
MeshLab: An Open-Source Mesh Processing Tool (version Sixth Eurographics Italian
Chapter Conference).
Cohen, Anne L., and Michael Holcomb. 2009. “Why Corals Care about Ocean Acidification:
Uncovering the Mechanism.” Oceanography 22 (4): 118–27.
Cohen, Anne L., and Ted A. McConnaughey. 2003. “Geochemical Perspectives on Coral
Mineralization.” Reviews in Mineralogy and Geochemistry 54 (1): 151–87.
Coker, Darren J., Shaun K. Wilson, and Morgan S. Pratchett. 2014. “Importance of Live Coral
Habitat for Reef Fishes.” Reviews in Fish Biology and Fisheries 24 (1): 89–126.
Cornwell, Brendan, Katrina Armstrong, Nia S. Walker, Marilla Lippert, Victor Nestor, Yimnang
Golbuu, and Stephen R. Palumbi. 2021. “Widespread Variation in Heat Tolerance and
Symbiont Load Are Associated with Growth Tradeoffs in the Coral Acropora Hyacinthus in
Palau.” eLife 10 (August). https://doi.org/10.7554/eLife.64790.
Cramer, Katie L., Jeremy B. C. Jackson, Mary K. Donovan, Benjamin J. Greenstein, Chelsea A.
Korpanty, Geoffrey M. Cook, and John M. Pandolfi. 2020. “Widespread Loss of Caribbean
Acroporid Corals Was Underway before Coral Bleaching and Disease Outbreaks.” Science
Advances 6 (17): eaax9395.
Cunning, R., P. Gillette, T. Capo, K. Galvez, and A. C. Baker. 2015. “Growth Tradeoffs
Associated with Thermotolerant Symbionts in the Coral Pocillopora Damicornis Are Lost in
Warmer Oceans.” Coral Reefs 34 (1): 155–60.
Cunning, Ross, Katherine E. Parker, Kelsey Johnson-Sapp, Richard F. Karp, Alexandra D. Wen,
Olivia M. Williamson, Erich Bartels, et al. 2021. “Census of Heat Tolerance among
Florida’s Threatened Staghorn Corals Finds Resilient Individuals throughout Existing
Nursery Populations.” Proceedings of the Royal Society B: Biological Sciences 288 (1961):
20211613.
Cziesielski, Maha J., Sebastian Schmidt-Roach, and Manuel Aranda. 2019. “The Past, Present,
and Future of Coral Heat Stress Studies.” Ecology and Evolution 9 (17): 10055–66.
Daniels, Camille, Sebastian Baumgarten, Lauren Yum, Craig MIchell, Till Bayer, Chatchanit
Arif, Cornelia Roder, Ernesto Weil, and Christian Voolstra. 2015. “Metatranscriptome
Analysis of the Reef-Building Coral Orbicella Faveolata Indicates Holobiont Response to
Coral Disease.” Frontiers in Marine Science 2. https://doi.org/10.3389/fmars.2015.00062.
Davies, S. P. 1989. “Short-Term Growth Measurements of Corals Using an Accurate Buoyant
Weighing Technique.” Marine Biology 101 (3): 389–95.
Davy Simon K., Allemand Denis, and Weis Virginia M. 2012. “Cell Biology of Cnidarian-
Dinoflagellate Symbiosis.” Microbiology and Molecular Biology Reviews: MMBR 76 (2):
176
229–61.
De Meester, Luc. 1996. “Evolutionary potential and local genetic differentiation in a
phenotypically plastic trait of a cyclical parthenogen, Daphnia manga” Evolution;
International Journal of Organic Evolution 50 (3): 1293–98.
Denis, Vianney, Lauriane Ribas-Deulofeu, Nicolas Sturaro, Chao-Yang Kuo, and Chaolun Allen
Chen. 2017. “A Functional Approach to the Structural Complexity of Coral Assemblages
Based on Colony Morphological Features.” Scientific Reports 7 (1): 9849.
Dennison, William C., and David J. Barnes. 1988. “Effect of Water Motion on Coral
Photosynthesis and Calcification.” Journal of Experimental Marine Biology and Ecology
115 (1): 67–77.
DeSalvo, M. K., C. R. Voolstra, S. Sunagawa, J. A. Schwarz, J. H. Stillman, M. A. Coffroth, A.
M. Szmant, and M. Medina. 2008. “Differential Gene Expression during Thermal Stress
and Bleaching in the Caribbean Coral Montastraea Faveolata.” Molecular Ecology 17 (17):
3952–71.
Des Roches, Simone, Linwood H. Pendleton, Beth Shapiro, and Eric P. Palkovacs. 2021.
“Conserving Intraspecific Variation for Nature’s Contributions to People.” Nature Ecology
& Evolution 5 (5): 574–82.
Dewitt, T. J., A. Sih, and D. S. Wilson. 1998. “Costs and Limits of Phenotypic Plasticity.”
Trends in Ecology & Evolution 13 (2): 77–81.
Dixon, Groves, Yi Liao, Line K. Bay, and Mikhail V. Matz. 2018. “Role of Gene Body
Methylation in Acclimatization and Adaptation in a Basal Metazoan.” Proceedings of the
National Academy of Sciences of the United States of America 115 (52): 13342–46.
D’Olivo, J. P., and M. T. McCulloch. 2017. “Response of Coral Calcification and Calcifying
Fluid Composition to Thermally Induced Bleaching Stress.” Scientific Reports 7 (1): 2207.
Donner, Simon D., William J. Skirving, Christopher M. Little, Michael Oppenheimer, and Ove
Hoegh-Guldberg. 2005. “Global Assessment of Coral Bleaching and Required Rates of
Adaptation under Climate Change.” Global Change Biology 11 (12): 2251–65.
Dornelas, Maria, Joshua S. Madin, Andrew H. Baird, and Sean R. Connolly. 2017. “Allometric
Growth in Reef-Building Corals.” Proceedings. Biological Sciences / The Royal Society 284
(1851). https://doi.org/10.1098/rspb.2017.0053.
Dorn, L. A., E. H. Pyle, and J. Schmitt. 2000. “Plasticity to Light Cues and Resources in
Arabidopsis Thaliana: Testing for Adaptive Value and Costs.” Evolution; International
Journal of Organic Evolution 54 (6): 1982–94.
Doszpot, Neil E., Michael J. McWilliam, Morgan S. Pratchett, Andrew S. Hoey, and Will F.
177
Figueira. 2019. “Plasticity in Three-Dimensional Geometry of Branching Corals Along a
Cross-Shelf Gradient.” Diversity 11 (3): 44.
Douglas, A. E. 2003. “Coral Bleaching––how and Why?” Marine Pollution Bulletin 46 (4): 385–
92.
Drury, Crawford, Justin B. Greer, Iliana Baums, Brooke Gintert, and Diego Lirman. 2019.
“Clonal Diversity Impacts Coral Cover in Acropora Cervicornisthickets: Potential
Relationships between Density, Growth, and Polymorphisms.” Ecology and Evolution 9 (8):
4518–31.
Drury, Crawford, and Diego Lirman. 2021. “Genotype by Environment Interactions in Coral
Bleaching.” Proceedings. Biological Sciences / The Royal Society 288 (1946): 20210177.
Drury, Crawford, Derek Manzello, and Diego Lirman. 2017. “Genotype and Local Environment
Dynamically Influence Growth, Disturbance Response and Survivorship in the Threatened
Coral, Acropora Cervicornis.” PloS One 12 (3): e0174000.
Drury, Crawford, Claire B. Paris, Vassiliki H. Kourafalou, and Diego Lirman. 2018. “Dispersal
Capacity and Genetic Relatedness in Acropora Cervicornis on the Florida Reef Tract.”
Coral Reefs 37 (2): 585–96.
Dubé, Caroline E., Emilie Boissin, Jeffrey A. Maynard, and Serge Planes. 2017. “Fire Coral
Clones Demonstrate Phenotypic Plasticity among Reef Habitats.” Molecular Ecology 26
(15): 3860–69.
D’Urban Jackson, Tim, Gareth J. Williams, Guy Walker-Springett, and Andrew J. Davies. 2020.
“Three-Dimensional Digital Mapping of Ecosystems: A New Era in Spatial Ecology.”
Proceedings. Biological Sciences / The Royal Society 287 (1920): 20192383.
Edmunds, Peter J. 2017. “Intraspecific Variation in Growth Rate Is a Poor Predictor of Fitness
for Reef Corals.” Ecology 98 (8): 2191–2200.
Edmunds, Peter J., and Hollie M. Putnam. 2020. “Science-Based Approach to Using Growth
Rate to Assess Coral Performance and Restoration Outcomes.” Biology Letters 16 (7):
20200227.
Edmunds, P. J. 1994. “Evidence That Reef-Wide Patterns of Coral Bleaching May Be the Result
of the Distribution of Bleaching-Susceptible Clones.” Marine Biology 121 (1): 137–42.
Eierman, Laura E., and Matthew P. Hare. 2016. “Reef-Specific Patterns of Gene Expression
Plasticity in Eastern Oysters (Crassostrea Virginica).” The Journal of Heredity 107 (1): 90–
100.
Enochs, I. C., D. P. Manzello, R. Carlton, S. Schopmeyer, R. van Hooidonk, and D. Lirman.
2014. “Effects of Light and Elevated pCO2 on the Growth and Photochemical Efficiency of
178
Acropora Cervicornis.” Coral Reefs 33 (2): 477–85.
Félix‐Burruel, Ricardo E., Eugenio Larios, Enriquena Bustamante, and Alberto Búrquez. 2019.
“Nonlinear Modeling of Saguaro Growth Rates Reveals the Importance of Temperature for
Size‐dependent Growth.” American Journal of Botany 106 (10): 1300–1307.
Ferrari, Renata, Will F. Figueira, Morgan S. Pratchett, Tatiana Boube, Arne Adam, Tania
Kobelkowsky-Vidrio, Steve S. Doo, Trisha Brooke Atwood, and Maria Byrne. 2017. “3D
Photogrammetry Quantifies Growth and External Erosion of Individual Coral Colonies and
Skeletons.” Scientific Reports 7 (1): 16737.
Figueira, Will, Renata Ferrari, Elyse Weatherby, Augustine Porter, Steven Hawes, and Maria
Byrne. 2015. “Accuracy and Precision of Habitat Structural Complexity Metrics Derived
from Underwater Photogrammetry.” Remote Sensing 7 (12): 16883–900.
Fox, Rebecca J., Jennifer M. Donelson, Celia Schunter, Timothy Ravasi, and Juan D. Gaitán-
Espitia. 2019. “Beyond Buying Time: The Role of Plasticity in Phenotypic Adaptation to
Rapid Environmental Change.” Philosophical Transactions of the Royal Society of London.
Series B, Biological Sciences 374 (1768): 20180174.
Fragata, Inês, Miguel Lopes-Cunha, Margarida Bárbaro, Bárbara Kellen, Margarida Lima,
Gonçalo S. Faria, Sofia G. Seabra, Mauro Santos, Pedro Simões, and Margarida Matos.
2016. “Keeping Your Options Open: Maintenance of Thermal Plasticity during Adaptation
to a Stable Environment.” Evolution; International Journal of Organic Evolution 70 (1):
195–206.
Fuess, Lauren E., Jorge H. Pinzón C, Ernesto Weil, Robert D. Grinshpon, and Laura D. Mydlarz.
2017. “Life or Death: Disease-Tolerant Coral Species Activate Autophagy Following
Immune Challenge.” Proceedings. Biological Sciences / The Royal Society 284 (1856).
https://doi.org/10.1098/rspb.2017.0771.
Gardner, Toby A., Isabelle M. Côté, Jennifer A. Gill, Alastair Grant, and Andrew R. Watkinson.
2003. “Long-Term Region-Wide Declines in Caribbean Corals.” Science 301 (5635): 958–
60.
Gavrilets, Sergey, and Samuel M. Scheiner. 1993. “The Genetics of Phenotypic Plasticity. V.
Evolution of Reaction Norm Shape.” Journal of Evolutionary Biology 6 (1): 31–48.
Ghalambor, Cameron K., Kim L. Hoke, Emily W. Ruell, Eva K. Fischer, David N. Reznick, and
Kimberly A. Hughes. 2015. “Non-Adaptive Plasticity Potentiates Rapid Adaptive Evolution
of Gene Expression in Nature.” Nature 525 (7569): 372–75.
Ghalambor, C. K., J. K. McKAY, S. P. Carroll, and D. N. Reznick. 2007. “Adaptive versus Non-
Adaptive Phenotypic Plasticity and the Potential for Contemporary Adaptation in New
Environments.” Functional Ecology 21 (3): 394–407.
179
Gladfelter, Elizabeth H. 1984. “Skeletal Development in Acropora Cervicornis.” Coral Reefs 3
(1): 51–57.
Glynn, P. W. 1993. “Coral Reef Bleaching: Ecological Perspectives.” Coral Reefs 12 (1): 1–17.
Glynn, P. W., and L. D’Croz. 1990. “Experimental Evidence for High Temperature Stress as the
Cause of El Nino-Coincident Coral Mortality.” Coral Reefs 8 (4): 181–91.
Graham, N. A. J., and K. L. Nash. 2013. “The Importance of Structural Complexity in Coral
Reef Ecosystems.” Coral Reefs 32 (2): 315–26.
Graham, Nicholas A. J. 2014. “Habitat Complexity: Coral Structural Loss Leads to Fisheries
Declines.” Current Biology: CB.
Grottoli, Andréa G., Lisa J. Rodrigues, and James E. Palardy. 2006. “Heterotrophic Plasticity
and Resilience in Bleached Corals.” Nature 440 (7088): 1186–89.
Gruber, Nicolas, Dominic Clement, Brendan R. Carter, Richard A. Feely, Steven van Heuven,
Mario Hoppema, Masao Ishii, et al. 2019. “The Oceanic Sink for Anthropogenic CO2 from
1994 to 2007.” Science 363 (6432): 1193–99.
Hall, V. R., and T. P. Hughes. 1996. “Reproductive Strategies of Modular Organisms:
Comparative Studies of Reef- Building Corals.” Ecology 77 (3): 950–63.
Heerwaarden, Belinda van, and Vanessa Kellermann. 2020. “Does Plasticity Trade Off With
Basal Heat Tolerance?” Trends in Ecology & Evolution 35 (10): 874–85.
Hemond, Elizabeth M., and Steven V. Vollmer. 2010. “Genetic Diversity and Connectivity in the
Threatened Staghorn Coral (Acropora Cervicornis) in Florida.” PloS One 5 (1): e8652.
Hendry, Andrew P. 2016. “Key Questions on the Role of Phenotypic Plasticity in Eco-
Evolutionary Dynamics.” The Journal of Heredity 107 (1): 25–41.
Herlan, J., and D. Lirman. 2008. “Development of a Coral Nursery Program for the Threatened
Coral Acropora Cervicornis in Florida,” January, 1244–47.
Herler, Jürgen, and Markus Dirnwöber. 2011. “A Simple Technique for Measuring Buoyant
Weight Increment of Entire, Transplanted Coral Colonies in the Field.” Journal of
Experimental Marine Biology and Ecology 407 (2): 250–55.
Hernández-Landa, Roberto C., Erick Barrera-Falcon, and Rodolfo Rioja-Nieto. 2020. “Size-
Frequency Distribution of Coral Assemblages in Insular Shallow Reefs of the Mexican
Caribbean Using Underwater Photogrammetry.” PeerJ 8 (April): e8957.
Hippel, Eric von, and Georg von Krogh. 2003. “Open Source Software and the ‘Private-
Collective’ Innovation Model: Issues for Organization Science.” Organization Science 14
180
(2): 209–23.
Hoadley, Kenneth D., D. Tye Pettay, Andréa G. Grottoli, Wei-Jun Cai, Todd F. Melman, Verena
Schoepf, Xinping Hu, et al. 2015. “Physiological Response to Elevated Temperature and
pCO2 Varies across Four Pacific Coral Species: Understanding the Unique Host+symbiont
Response.” Scientific Reports 5 (December): 18371.
Hoegh-Guldberg, O., L. Hughes, S. McIntyre, D. B. Lindenmayer, C. Parmesan, H. P.
Possingham, and C. D. Thomas. 2008. “Assisted Colonization and Rapid Climate Change.”
Science 321 (5887): 345–46.
Hoegh-Guldberg, O., P. J. Mumby, A. J. Hooten, R. S. Steneck, P. Greenfield, E. Gomez, C. D.
Harvell, et al. 2007. “Coral Reefs under Rapid Climate Change and Ocean Acidification.”
Science 318 (5857): 1737–42.
Hoffmann, A. A., and J. Merilä. 1999. “Heritable Variation and Evolution under Favourable and
Unfavourable Conditions.” Trends in Ecology & Evolution 14 (3): 96–101.
Hoogenboom, Mia O., Sean R. Connolly, and Kenneth R. N. Anthony. 2008. “Interactions
between Morphological and Physiological Plasticity Optimize Energy Acquisition in
Corals.” Ecology 89 (4): 1144–54.
Hoogenboom, Mia O., Grace E. Frank, Tory J. Chase, Saskia Jurriaans, Mariana Álvarez-
Noriega, Katie Peterson, Kay Critchell, et al. 2017. “Environmental Drivers of Variation in
Bleaching Severity of Acropora Species during an Extreme Thermal Anomaly.” Frontiers
in Marine Science 4. https://doi.org/10.3389/fmars.2017.00376.
House, Jenny E., Viviana Brambilla, Luc M. Bidaut, Alec P. Christie, Oscar Pizarro, Joshua S.
Madin, and Maria Dornelas. 2018. “Moving to 3D: Relationships between Coral Planar
Area, Surface Area and Volume.” PeerJ 6 (February): e4280.
Howells, Emily J., Ray Berkelmans, Madeleine J. H. van Oppen, Bette L. Willis, and Line K.
Bay. 2013. “Historical Thermal Regimes Define Limits to Coral Acclimatization.” Ecology
94 (5): 1078–88.
Hughes, T. P., D. Ayre, and J. H. Connell. 1992. “The Evolutionary Ecology of Corals.” Trends
in Ecology & Evolution 7 (9): 292–95.
Huntington, Brittany E., and Margaret W. Miller. 2014. “Location-Specific Metrics for Rapidly
Estimating the Abundance and Condition of the Threatened CoralAcropora Cervicornis.”
Restoration Ecology 22 (3): 299–303.
Jimenez, Isabel M., Michael Kühl, Anthony W. D. Larkum, and Peter J. Ralph. 2011. “Effects of
Flow and Colony Morphology on the Thermal Boundary Layer of Corals.” Journal of the
Royal Society, Interface / the Royal Society 8 (65): 1785–95.
181
Jimenez, Isabel M., Michael Kühl, A. W. D. Larkum, and P. J. Ralph. 2008. “Heat Budget and
Thermal Microenvironment of Shallow-Water Corals: Do Massive Corals Get Warmer than
Branching Corals?” Limnology and Oceanography 53 (4): 1548–61.
Johnson, Meaghan E., Caitlin Lustic, Erich Bartels, Iliana B. Baums, David S. Gilliam, Elizabeth
Anne Larson, Diego Lirman, Margaret W. Miller, Ken Nedimyer, and Stephanie
Schopmeyer. 2011. “Caribbean Acropora Restoration Guide: Best Practices for Propagation
and Population Enhancement.”
http://nsuworks.nova.edu/cgi/viewcontent.cgi?article=1076&context=occ_facreports/.
Jokiel, Paul L., Christopher P. Jury, and Ilsa B. Kuffner. 2016. “Coral Calcification and Ocean
Acidification.” In Coral Reefs at the Crossroads, edited by Dennis K. Hubbard, Caroline S.
Rogers, Jere H. Lipps, Stanley, Jr., and George D., 7–45. Dordrecht: Springer Netherlands.
Jokiel, P. L., J. E. Maragos, and L. Franzisket. 1978. “Coral Growth: Buoyant Weight
Technique.” Coral Reefs: Research Methods. https://www.researchgate.net/profile/Paul-
Jokiel/publication/285472773_Coral_growth_buoyant_weight_technique_In_Stoddart_DR_
Johannes_RE_eds_Coral_reefs_research_methods/links/568b229608aebccc4e1a3a5c/Coral-
growth-buoyant-weight-technique-In-Stoddart-DR-Johannes-RE-eds-Coral-reefs-research-
methods.pdf.
Jones, Alison, and Ray Berkelmans. 2010. “Potential Costs of Acclimatization to a Warmer
Climate: Growth of a Reef Coral with Heat Tolerant vs. Sensitive Symbiont Types.” PloS
One 5 (5): e10437.
Kelly, Morgan. 2019. “Adaptation to Climate Change through Genetic Accommodation and
Assimilation of Plastic Phenotypes.” Philosophical Transactions of the Royal Society of
London. Series B, Biological Sciences 374 (1768): 20180176.
Kenkel, Carly D., Albert T. Almanza, and Mikhail V. Matz. 2015. “Fine-Scale Environmental
Specialization of Reef-Building Corals Might Be Limiting Reef Recovery in the Florida
Keys.” Ecology 96 (12): 3197–3212.
Kenkel, Carly D., and Mikhail V. Matz. 2016. “Gene Expression Plasticity as a Mechanism of
Coral Adaptation to a Variable Environment.” Nature Ecology & Evolution 1 (1): 14.
Kenkel, C. D., E. Meyer, and M. V. Matz. 2013. “Gene Expression under Chronic Heat Stress in
Populations of the Mustard Hill Coral (Porites Astreoides) from Different Thermal
Environments.” Molecular Ecology 22 (16): 4322–34.
Kenkel, C. D., S. P. Setta, and M. V. Matz. 2015. “Heritable Differences in Fitness-Related
Traits among Populations of the Mustard Hill Coral, Porites Astreoides.” Heredity 115 (6):
509–16.
Keshavmurthy, Shashank, Morgan Beals, Hernyi Justin Hsieh, Kwang-Sik Choi, and Chaolun
Allen Chen. 2021. “Physiological Plasticity of Corals to Temperature Stress in Marginal
182
Coral Communities.” The Science of the Total Environment 758 (March): 143628.
Kiel, C., B. E. Huntington, and M. W. Miller. 2012. “Tractable Field Metrics for Restoration and
Recovery Monitoring of Staghorn Coral Acropora Cervicornis.” Endangered Species
Research 19 (2): 171–76.
Knowlton, Nancy, Russell E. Brainard, Rebecca Fisher, Megan Moews, Laetitia Plaisance, and
M. Julian and Caley. 2010. “Coral Reef Biodiversity.” In Life in the World’s Oceans:
Diversity, Distribution, and Abundance, edited by Alasdair McIntyre. John Wiley & Sons.
Kroeker, Kristy J., Lauren E. Bell, Emily M. Donham, Umihiko Hoshijima, Sarah Lummis,
Jason A. Toy, and Ellen Willis-Norton. 2020. “Ecological Change in Dynamic
Environments: Accounting for Temporal Environmental Variability in Studies of Ocean
Change Biology.” Global Change Biology 26 (1): 54–67.
Kuffner, I. B., A. Stathakopoulos, L. T. Toth, and L. A. Bartlett. 2020. “Reestablishing a
Stepping-Stone Population of the Threatened Elkhorn Coral Acropora Palmata to Aid
Regional Recovery.” Endangered Species Research 43 (December): 461–73.
Kuffner, Ilsa B., Erich Bartels, Anastasios Stathakopoulos, Ian C. Enochs, G. Kolodziej, Lauren
T. Toth, and Derek P. Manzello. 2017. “Plasticity in Skeletal Characteristics of Nursery-
Raised Staghorn Coral, Acropora Cervicornis.” Coral Reefs 36 (3): 679–84.
Kuznetsova, A., P. B. Brockhoff, R. H. B. Christensen, and S. P. Jensen. 2020. Tests in Linear
Mixed Effects Models [R Package lmerTest Version 3.1-3] (version 3.1-3). Comprehensive
R Archive Network (CRAN). https://cran.r-project.org/web/packages/lmerTest/index.html.
Kvitt, Hagit, Hanna Rosenfeld, and Dan Tchernov. 2016. “The Regulation of Thermal Stress
Induced Apoptosis in Corals Reveals High Similarities in Gene Expression and Function to
Higher Animals.” Scientific Reports 6 (July): 30359.
Lange, Ines D., and Chris T. Perry. 2020. “A Quick, Easy and Non‐invasive Method to Quantify
Coral Growth Rates Using Photogrammetry and 3D Model Comparisons.” Methods in
Ecology and Evolution / British Ecological Society 11 (6): 714–26.
Langfelder, Peter, and Steve Horvath. 2008. “WGCNA: An R Package for Weighted Correlation
Network Analysis.” BMC Bioinformatics.
https://bmcbioinformatics.biomedcentral.com/articles/10.1186/1471-2105-9-559.
———. 2012. “Fast R Functions for Robust Correlations and Hierarchical Clustering.” Journal
of Statistical Software. https://www.jstatsoft.org/v46/i11/.
Lavy, A., G. Eyal, B. Neal, and R. Keren. 2015. “A Quick, Easy and Non‐intrusive Method for
Underwater Volume and Surface Area Evaluation of Benthic Organisms by 3D Computer
Modelling.” Methods in Ecology and Evolution / British Ecological Society.
https://doi.org/10.1111/2041-210X.12331.
183
Leichter, J. J., G. B. Deane, and M. D. Stokes. 2005. “Spatial and Temporal Variability of
Internal Wave Forcing on a Coral Reef.” Journal of Physical Oceanography 35 (11): 1945–
62.
Leon, Javier X., Chris M. Roelfsema, Megan I. Saunders, and Stuart R. Phinn. 2015. “Measuring
Coral Reef Terrain Roughness Using ‘Structure-from-Motion’close-Range
Photogrammetry.” Geomorphology 242: 21–28.
Lesser, Michael P. 2006. “Oxidative Stress in Marine Environments: Biochemistry and
Physiological Ecology.” Annual Review of Physiology 68: 253–78.
Lesser, Michael P., Virginia M. Weis, Mark R. Patterson, and Paul L. Jokiel. 1994. “Effects of
Morphology and Water Motion on Carbon Delivery and Productivity in the Reef Coral,
Pocillopora Damicornis (Linnaeus): Diffusion Barriers, Inorganic Carbon Limitation, and
Biochemical Plasticity.” Journal of Experimental Marine Biology and Ecology 178 (2):
153–79.
Leung, Christelle, Marie Rescan, Daphné Grulois, and Luis-Miguel Chevin. 2020. “Reduced
Phenotypic Plasticity Evolves in Less Predictable Environments.” Ecology Letters 23 (11):
1664–72.
Leuzinger, Sebastian, Kenneth R. N. Anthony, and Bette L. Willis. 2003. “Reproductive Energy
Investment in Corals: Scaling with Module Size.” Oecologia 136 (4): 524–31.
Levis, Nicholas A., Andrew J. Isdaner, and David W. Pfennig. 2018. “Morphological Novelty
Emerges from Pre-Existing Phenotypic Plasticity.” Nature Ecology & Evolution 2 (8):
1289–97.
Levis, Nicholas A., and David W. Pfennig. 2019. “Plasticity-Led Evolution: Evaluating the Key
Prediction of Frequency-Dependent Adaptation.” Proceedings. Biological Sciences / The
Royal Society 286 (1897): 20182754.
———. 2021. “Innovation and Diversification via Plasticity-Led Evolution.” In Phenotypic
Plasticity & Evolution, 211–40. CRC Press.
Libro, Silvia, Stefan T. Kaluziak, and Steven V. Vollmer. 2013. “RNA-Seq Profiles of Immune
Related Genes in the Staghorn Coral Acropora Cervicornis Infected with White Band
Disease.” PloS One 8 (11): e81821.
Libro, Silvia, and Steven V. Vollmer. 2016. “Genetic Signature of Resistance to White Band
Disease in the Caribbean Staghorn Coral Acropora Cervicornis.” PloS One 11 (1):
e0146636.
Liew, Yi Jin, Didier Zoccola, Yong Li, Eric Tambutté, Alexander A. Venn, Craig T. Michell,
Guoxin Cui, et al. 2018. “Epigenome-Associated Phenotypic Acclimatization to Ocean
184
Acidification in a Reef-Building Coral.” Science Advances 4 (6): eaar8028.
Li, Jianping, and Ruiqiang Ding. 2013. “Temporal-Spatial Distribution of the Predictability
Limit of Monthly Sea Surface Temperature in the Global Oceans.” International Journal of
Climatology 33 (8): 1936–47.
Lind, Martin I., Pär K. Ingvarsson, Helena Johansson, David Hall, and Frank Johansson. 2011.
“Gene Flow and Selection on Phenotypic Plasticity in an Island System of Rana
Temporaria.” Evolution; International Journal of Organic Evolution 65 (3): 684–97.
Lines, Emily R., Miguel A. Zavala, Drew W. Purves, and David A. Coomes. 2012. “Predictable
Changes in Aboveground Allometry of Trees along Gradients of Temperature, Aridity and
Competition.” Global Ecology and Biogeography: A Journal of Macroecology 21 (10):
1017–28.
Lirman, Diego, and Peggy Fong. 2007. “Is Proximity to Land-Based Sources of Coral Stressors
an Appropriate Measure of Risk to Coral Reefs? An Example from the Florida Reef Tract.”
Marine Pollution Bulletin 54 (6): 779–91.
Lirman, Diego, Stephanie Schopmeyer, Victor Galvan, Crawford Drury, Andrew C. Baker, and
Iliana B. Baums. 2014. “Growth Dynamics of the Threatened Caribbean Staghorn Coral
Acropora Cervicornis: Influence of Host Genotype, Symbiont Identity, Colony Size, and
Environmental Setting.” PloS One 9 (9): e107253.
Lohr, Kathryn E., and Joshua T. Patterson. 2017. “Intraspecific Variation in Phenotype among
Nursery-Reared Staghorn Coral Acropora Cervicornis (Lamarck, 1816).” Journal of
Experimental Marine Biology and Ecology 486 (January): 87–92.
Lohr, K. E., S. Bejarano, D. Lirman, S. Schopmeyer, and C. Manfrino. 2015. “Optimizing the
Productivity of a Coral Nursery Focused on Staghorn Coral Acropora Cervicornis.”
Endangered Species Research 27 (3): 243–50.
Losos, J. B., D. A. Creer, D. Glossip, R. Goellner, A. Hampton, G. Roberts, N. Haskell, P.
Taylor, and J. Ettling. 2000. “Evolutionary Implications of Phenotypic Plasticity in the
Hindlimb of the Lizard Anolis Sagrei.” Evolution; International Journal of Organic
Evolution 54 (1): 301–5.
Louis, Yohan D., Ranjeet Bhagooli, Carly D. Kenkel, Andrew C. Baker, and Sabrina D. Dyall.
2017. “Gene Expression Biomarkers of Heat Stress in Scleractinian Corals: Promises and
Limitations.” Comparative Biochemistry and Physiology. Toxicology & Pharmacology:
CBP 191 (January): 63–77.
Love, Michael I., Wolfgang Huber, and Simon Anders. 2014. “Moderated Estimation of Fold
Change and Dispersion for RNA-Seq Data with DESeq2.” Genome Biology.
https://doi.org/10.1186/s13059-014-0550-8.
Lugo-Fernández, A., H. H. Roberts, Wiseman, Jr, and W. J. 1998. “Tide Effects on Wave
185
Attenuation and Wave Set-up on a Caribbean Coral Reef.” Estuarine, Coastal and Shelf
Science 47 (4): 385–93.
Lustic, Caitlin, Kerry Maxwell, Erich Bartels, Brian Reckenbeil, Emily Utset, Stephanie
Schopmeyer, Ian Zink, and Diego Lirman. 2020. “The Impacts of Competitive Interactions
on Coral Colonies after Transplantation: A Multispecies Experiment from the Florida Keys,
US.” Bulletin of Marine Science 96 (4): 805–18.
Ly, Delphine, Martha Hamblin, Ismail Y. Rabbi, M. Gedil, M. A. Bakare, H. Gauch Jr, R. U.
Okechukwu, A. Dixon, P. A. Kulakow, and Jean-Luc Jannink. 2013. “Relatedness and
Genotype X Environment Interaction Affect Prediction Accuracies in Genomic Selection: A
Study in Cassava.” Crop Science. https://cgspace.cgiar.org/handle/10568/76637.
Madin, Joshua S., Andrew H. Baird, Maria Dornelas, and Sean R. Connolly. 2014. “Mechanical
Vulnerability Explains Size‐dependent Mortality of Reef Corals.” Edited by Howard
Cornell. Ecology Letters 17 (8): 1008–15.
Majerova, Eva, Fiona C. Carey, Crawford Drury, and Ruth D. Gates. 2021. “Preconditioning
Improves Bleaching Tolerance in the Reef-Building Coral Pocillopora Acuta through
Modulations in the Programmed Cell Death Pathways.” Molecular Ecology 30 (14): 3560–
74.
Manzello, Derek P. 2015. “Rapid Recent Warming of Coral Reefs in the Florida Keys.”
Scientific Reports 5 (November): 16762.
Manzello, Derek P., Ray Berkelmans, and James C. Hendee. 2007. “Coral Bleaching Indices and
Thresholds for the Florida Reef Tract, Bahamas, and St. Croix, US Virgin Islands.” Marine
Pollution Bulletin 54 (12): 1923–31.
Manzello, Derek P., Mikhail V. Matz, Ian C. Enochs, Lauren Valentino, Renee D. Carlton,
Graham Kolodziej, Xaymara Serrano, Erica K. Towle, and Mike Jankulak. 2019. “Role of
Host Genetics and Heat-Tolerant Algal Symbionts in Sustaining Populations of the
Endangered Coral Orbicella Faveolata in the Florida Keys with Ocean Warming.” Global
Change Biology 25 (3): 1016–31.
Mauritsen, Thorsten, and Robert Pincus. 2017. “Committed Warming Inferred from
Observations.” Nature Climate Change 7 (9): 652–55.
Mayfield, Anderson B., Yi-Yuong Hsiao, Tung-Yung Fan, Chii-Shiarng Chen, and Ruth D.
Gates. 2010. “Evaluating the Temporal Stability of Stress-Activated Protein Kinase and
Cytoskeleton Gene Expression in the Pacific Reef Corals Pocillopora Damicornis and
Seriatopora Hystrix.” Journal of Experimental Marine Biology and Ecology 395 (1): 215–
22.
Mayfield, Anderson B., Li-Hsueh Wang, Pei-Ciao Tang, Tung-Yung Fan, Yi-Yuong Hsiao,
Ching-Lin Tsai, and Chii-Shiarng Chen. 2011. “Assessing the Impacts of Experimentally
Elevated Temperature on the Biological Composition and Molecular Chaperone Gene
186
Expression of a Reef Coral.” PloS One 6 (10): e26529.
McClanahan, Timothy R., Mebrahtu Ateweberhan, Christopher A. Muhando, Joseph Maina, and
Mohammed S. Mohammed. 2007. “Effects of Climate and Seawater Temperature Variation
on Coral Bleaching and Mortality.” Ecological Monographs 77 (4): 503–25.
McCullough, Erin L., Kimberly J. Ledger, Devin M. O’Brien, and Douglas J. Emlen. 2015.
“Variation in the Allometry of Exaggerated Rhinoceros Beetle Horns.” Animal Behaviour
109 (November): 133–40.
McKinnon, David, Hu He, Ben Upcroft, and Ryan N. Smith. 2011. “Towards Automated and in-
Situ, near-Real Time 3-D Reconstruction of Coral Reef Environments.” In OCEANS’11
MTS/IEEE KONA, 1–10. ieeexplore.ieee.org.
Meyer, E., G. V. Aglyamova, and M. V. Matz. 2011. “Profiling Gene Expression Responses of
Coral Larvae (Acropora Millepora) to Elevated Temperature and Settlement Inducers Using
a Novel RNA-Seq Procedure.” Molecular Ecology 20 (17): 3599–3616.
Miller, M., A. Bourque, and J. Bohnsack. 2002. “An Analysis of the Loss of Acroporid Corals at
Looe Key, Florida, USA: 1983–2000.” Coral Reefs 21 (2): 179–82.
Miller, N. A., and J. H. Stillman. 2012. “Physiological Optima and Critical Limits.” Nature
Education Knowledge. 2012.
https://www.nature.com/scitable/knowledge/library/physiological-optima-and-critical-
limits-45749376/.
Million, Wyatt C., Sibelle O’Donnell, Erich Bartels, and Carly D. Kenkel. 2021. “Colony-Level
3D Photogrammetry Reveals That Total Linear Extension and Initial Growth Do Not Scale
With Complex Morphological Growth in the Branching Coral, Acropora Cervicornis.”
Frontiers in Marine Science 8: 384.
Moczek, Armin P., Sonia Sultan, Susan Foster, Cris Ledón-Rettig, Ian Dworkin, H. Fred
Nijhout, Ehab Abouheif, and David W. Pfennig. 2011. “The Role of Developmental
Plasticity in Evolutionary Innovation.” Proceedings. Biological Sciences / The Royal
Society 278 (1719): 2705–13.
Monismith, Stephen G. 2007. “Hydrodynamics of Coral Reefs.” Annual Review of Fluid
Mechanics 39 (1): 37–55.
Morgan, M. B., D. L. Vogelien, and T. W. Snell. 2001. “Assessing Coral Stress Responses Using
Molecular Biomarkers of Gene Transcription.” Environmental Toxicology and Chemistry /
SETAC 20 (3): 537–43.
Morgan, Michael B., and Terry W. Snell. 2002. “Characterizing Stress Gene Expression in Reef-
Building Corals Exposed to the Mosquitoside Dibrom.” Marine Pollution Bulletin 44 (11):
1206–18.
187
Morikawa, Megan K., and Stephen R. Palumbi. 2019. “Using Naturally Occurring Climate
Resilient Corals to Construct Bleaching-Resistant Nurseries.” Proceedings of the National
Academy of Sciences 116 (21). https://doi.org/10.1073/pnas.1721415116.
Moya, A., L. Huisman, S. Forêt, J-P Gattuso, D. C. Hayward, E. E. Ball, and D. J. Miller. 2015.
“Rapid Acclimation of Juvenile Corals to CO2 -Mediated Acidification by Upregulation of
Heat Shock Protein and Bcl-2 Genes.” Molecular Ecology 24 (2): 438–52.
Muko, Soyoka, Kohkichi Kawasaki, Kazuhiko Sakai, Fugo Takasu, and Nanako Shigesada.
2000. “Morphological Plasticity in the Coral Porites Sillimaniani and Its Adaptive
Significance.” Bulletin of Marine Science 66 (1): 225–39.
Muller, Erinn M., Erich Bartels, and Iliana B. Baums. 2018. “Bleaching Causes Loss of Disease
Resistance within the Threatened Coral Species Acropora Cervicornis.” eLife 7
(September). https://doi.org/10.7554/eLife.35066.
Muller, Erinn M., Ashley M. Dungan, Wyatt C. Million, Katherine R. Eaton, Chelsea Petrik,
Erich Bartels, Emily R. Hall, and Carly D. Kenkel. 2021. “Heritable Variation and Lack of
Tradeoffs Suggest Adaptive Capacity in Acropora Cervicornis despite Negative Synergism
under Climate Change Scenarios.” Proceedings. Biological Sciences / The Royal Society
288 (1960): 20210923.
Murchie, E. H., and T. Lawson. 2013. “Chlorophyll Fluorescence Analysis: A Guide to Good
Practice and Understanding Some New Applications.” Journal of Experimental Botany 64
(13): 3983–98.
Murren, C. J., J. R. Auld, H. Callahan, C. K. Ghalambor, C. A. Handelsman, M. A. Heskel, J. G.
Kingsolver, et al. 2015. “Constraints on the Evolution of Phenotypic Plasticity: Limits and
Costs of Phenotype and Plasticity.” Heredity 115 (4): 293–301.
Muscatine, L., L. R. McCloskey, and R. E. Marian. 1981. “Estimating the Daily Contribution of
Carbon from Zooxanthellae to Coral Animal respiration1.” Limnology and Oceanography
26 (4): 601–11.
Mydlarz, Laura D., Courtney S. Couch, Ernesto Weil, Garriet Smith, and C. Drew Harvell. 2009.
“Immune Defenses of Healthy, Bleached and Diseased Montastraea Faveolata during a
Natural Bleaching Event.” Diseases of Aquatic Organisms 87 (1-2): 67–78.
Mydlarz, Laura D., Elizabeth S. McGinty, and C. Drew Harvell. 2010. “What Are the
Physiological and Immunological Responses of Coral to Climate Warming and Disease?”
The Journal of Experimental Biology 213 (6): 934–45.
Mydlarz, Laura D., and Caroline V. Palmer. 2011. “The Presence of Multiple Phenoloxidases in
Caribbean Reef-Building Corals.” Comparative Biochemistry and Physiology. Part A,
Molecular & Integrative Physiology 159 (4): 372–78.
188
Nielsen, Matthew E., and Daniel R. Papaj. 2022. “Why Study Plasticity in Multiple Traits? New
Hypotheses for How Phenotypically Plastic Traits Interact during Development and
Selection.” Evolution; International Journal of Organic Evolution, March.
https://doi.org/10.1111/evo.14464.
Noonan, S. H. C., G. P. Jones, and M. S. Pratchett. 2012. “Coral Size, Health and Structural
Complexity: Effects on the Ecology of a Coral Reef Damselfish.” Marine Ecology Progress
Series 456 (June): 127–37.
O’Donnell, Kelli E., Kathryn E. Lohr, Erich Bartels, Iliana B. Baums, and Joshua T. Patterson.
2018. “Acropora Cervicornis Genet Performance and Symbiont Identity throughout the
Restoration Process.” Coral Reefs. https://doi.org/10.1007/s00338-018-01743-y.
O’Donnell, K. E., K. E. Lohr, and E. Bartels. 2017. “Evaluation of Staghorn Coral (Acropora
Cervicornis, Lamarck 1816) Production Techniques in an Ocean-Based Nursery with
Consideration of Coral Genotype.” Journal of Experimental.
https://www.sciencedirect.com/science/article/pii/S0022098116302829.
Oliver, T. A., and S. R. Palumbi. 2011. “Do Fluctuating Temperature Environments Elevate
Coral Thermal Tolerance?” Coral Reefs 30 (2): 429–40.
Oppen, Madeleine J. H. van, James K. Oliver, Hollie M. Putnam, and Ruth D. Gates. 2015.
“Building Coral Reef Resilience through Assisted Evolution.” Proceedings of the National
Academy of Sciences of the United States of America 112 (8): 2307–13.
Palmer, C. V., and N. Traylor-Knowles. 2012. “Towards an Integrated Network of Coral
Immune Mechanisms.” Proceedings. Biological Sciences / The Royal Society 279 (1745):
4106–14.
Palumbi, Stephen R., Daniel J. Barshis, Nikki Traylor-Knowles, and Rachael A. Bay. 2014.
“Mechanisms of Reef Coral Resistance to Future Climate Change.” Science 344 (6186):
895–98.
Pantano, Lorena. 2022. “DEGreport: Report of DEG Analysis.”
http://lpantano.github.io/DEGreport/.
Paradis, Bennett T., Raymond P. Henry, and Nanette E. Chadwick. 2019. “Compound Effects of
Thermal Stress and Tissue Abrasion on Photosynthesis and Respiration in the Reef-
Building Coral Acropora Cervicornis (Lamarck, 1816).” Journal of Experimental Marine
Biology and Ecology 521 (December): 151222.
Parkinson, John Everett, Andrew C. Baker, Iliana B. Baums, Sarah W. Davies, Andréa G.
Grottoli, Sheila A. Kitchen, Mikhail V. Matz, Margaret W. Miller, Andrew A. Shantz, and
Carly D. Kenkel. 2020. “Molecular Tools for Coral Reef Restoration: Beyond Biomarker
Discovery.” Conservation Letters 13 (1). https://doi.org/10.1111/conl.12687.
189
Parkinson, John Everett, Erich Bartels, Meghann K. Devlin-Durante, Caitlin Lustic, Ken
Nedimyer, Stephanie Schopmeyer, Diego Lirman, Todd C. LaJeunesse, and Iliana B.
Baums. 2018. “Extensive Transcriptional Variation Poses a Challenge to Thermal Stress
Biomarker Development for Endangered Corals.” Molecular Ecology 27 (5): 1103–19.
Perkins, R. D., and Paul Enos. 1968. “Hurricane Betsy in the Florida-Bahama Area: Geologic
Effects and Comparison with Hurricane Donna.” The Journal of Geology 76 (6): 710–17.
Pigliucci, Massimo. 2005. “Evolution of Phenotypic Plasticity: Where Are We Going Now?”
Trends in Ecology & Evolution 20 (9): 481–86.
Pigliucci, Massimo, Courtney J. Murren, and Carl D. Schlichting. 2006. “Phenotypic Plasticity
and Evolution by Genetic Assimilation.” The Journal of Experimental Biology 209 (Pt 12):
2362–67.
Poquita-Du, Rosa Celia, Danwei Huang, Loke Ming Chou, Mrinalini, and Peter A. Todd. 2019.
“Short Term Exposure to Heat and Sediment Triggers Changes in Coral Gene Expression
and Photo-Physiological Performance.” Frontiers in Marine Science 6.
https://doi.org/10.3389/fmars.2019.00121.
Porter, James W., and Ouida W. Meier. 1992. “Quantification of Loss and Change in Floridian
Reef Coral Populations.” Integrative and Comparative Biology 32 (6): 625–40.
Pratchett, Morgan S., Kristen D. Anderson, Mia O. Hoogenboom, Elizabeth Widman, Andrew
H. Baird, John M. Pandolfi, Peter J. Edmunds, and Janice M. Lough. 2015. “Spatial,
Temporal and Taxonomic Variation in Coral Growth—implications for the Structure and
Function of Coral Reef Ecosystems.” Oceanography and Marine Biology: An Annual
Review 53: 215–95.
Pratchett, Morgan S., Andrew S. Hoey, and Shaun K. Wilson. 2014. “Reef Degradation and the
Loss of Critical Ecosystem Goods and Services Provided by Coral Reef Fishes.” Current
Opinion in Environmental Sustainability 7 (April): 37–43.
Price, Trevor D., Anna Qvarnström, and Darren E. Irwin. 2003. “The Role of Phenotypic
Plasticity in Driving Genetic Evolution.” Proceedings. Biological Sciences / The Royal
Society 270 (1523): 1433–40.
Putnam, H. M., A. B. Mayfield, T. Y. Fan, C. S. Chen, and R. D. Gates. 2013. “The
Physiological and Molecular Responses of Larvae from the Reef-Building Coral
Pocillopora Damicornis Exposed to near-Future Increases in Temperature and pCO2.”
Marine Biology 160 (8): 2157–73.
Rasband, W. S. n.d. “WS 1997--2014. ImageJ. US National Institutes of Health, Bethesda, MD.”
Rausher, Mark D. 1992. “THE MEASUREMENT OF SELECTION ON QUANTITATIVE
190
TRAITS: BIASES DUE TO ENVIRONMENTAL COVARIANCES BETWEEN TRAITS
AND FITNESS.” Evolution; International Journal of Organic Evolution 46 (3): 616–26.
R Core Team. 2020. R: A Language and Environment for Statistical Computing (version 3.6.3).
https://www.R-project.org/.
Reaka-Kudla, Marjorie L., Don E. Wilson, and Edward O. Wilson. 1996. Biodiversity II:
Understanding and Protecting Our Biological Resources. Joseph Henry Press.
Reed, Thomas E., Robin S. Waples, Daniel E. Schindler, Jeffrey J. Hard, and Michael T.
Kinnison. 2010. “Phenotypic Plasticity and Population Viability: The Importance of
Environmental Predictability.” Proceedings. Biological Sciences / The Royal Society 277
(1699): 3391–3400.
Reidenbach, Matthew A., Jeffrey R. Koseff, Stephen G. Monismith, Jonah V. Steinbuckc, and
Amatzia Genin. 2006. “The Effects of Waves and Morphology on Mass Transfer within
Branched Reef Corals.” Limnology and Oceanography 51 (2): 1134–41.
Rinkevich, Baruch. 1995. “Restoration Strategies for Coral Reefs Damaged by Recreational
Activities: The Use of Sexual and Asexual Recruits.” Restoration Ecology 3 (4): 241–51.
Risk, Michael J. 1972. “Fish Diversity on a Coral Reef in the Virgin Islands.” Atoll Research
Bulletin 153: 1–4.
Ritchie, R. J. 2008. “Universal Chlorophyll Equations for Estimating Chlorophylls A, B, C, and
D and Total Chlorophylls in Natural Assemblages of Photosynthetic Organisms Using
Acetone, Methanol, or Ethanol Solvents.” Photosynthetica 46 (1): 115–26.
Rocker, Melissa M., Carly D. Kenkel, David S. Francis, Bette L. Willis, and Line K. Bay. 2019.
“Plasticity in Gene Expression and Fatty Acid Profiles of Acropora Tenuis Reciprocally
Transplanted between Two Water Quality Regimes in the Central Great Barrier Reef,
Australia.” Journal of Experimental Marine Biology and Ecology 511 (February): 40–53.
Rodriguez-Lanetty, Mauricio, Saki Harii, and Ove Hoegh-Guldberg. 2009. “Early Molecular
Responses of Coral Larvae to Hyperthermal Stress.” Molecular Ecology 18 (24): 5101–14.
Roff, George, Jennifer Joseph, and Peter J. Mumby. 2020. “Multi-Decadal Changes in Structural
Complexity Following Mass Coral Mortality on a Caribbean Reef.” Biogeosciences 17
(23): 5909–18.
Rosic, Nedeljka, Edmund Yew Siang Ling, Chon-Kit Kenneth Chan, Hong Ching Lee, Paulina
Kaniewska, David Edwards, Sophie Dove, and Ove Hoegh-Guldberg. 2015. “Unfolding the
Secrets of Coral-Algal Symbiosis.” The ISME Journal 9 (4): 844–56.
Rumble, Stephen M., Phil Lacroute, Adrian V. Dalca, Marc Fiume, Arend Sidow, and Michael
Brudno. 2009. “SHRiMP: Accurate Mapping of Short Color-Space Reads.” PLoS
191
Computational Biology 5 (5): e1000386.
Safaie, Aryan, Nyssa J. Silbiger, Timothy R. McClanahan, Geno Pawlak, Daniel J. Barshis,
James L. Hench, Justin S. Rogers, Gareth J. Williams, and Kristen A. Davis. 2018. “High
Frequency Temperature Variability Reduces the Risk of Coral Bleaching.” Nature
Communications 9 (1): 1671.
Savary, Romain, Daniel J. Barshis, Christian R. Voolstra, Anny Cárdenas, Nicolas R. Evensen,
Guilhem Banc-Prandi, Maoz Fine, and Anders Meibom. 2021. “Fast and Pervasive
Transcriptomic Resilience and Acclimation of Extremely Heat-Tolerant Coral Holobionts
from the Northern Red Sea.” Proceedings of the National Academy of Sciences of the
United States of America 118 (19). https://doi.org/10.1073/pnas.2023298118.
Schaum, C-Elisa, Björn Rost, and Sinéad Collins. 2016. “Environmental Stability Affects
Phenotypic Evolution in a Globally Distributed Marine Picoplankton.” The ISME Journal
10 (1): 75–84.
Scheiner, Samuel M. 1993. “Genetics and Evolution of Phenotypic Plasticity,” November.
https://doi.org/10.1146/annurev.es.24.110193.000343.
Scheiner, Samuel M., and Richard F. Lyman. 1989. “The Genetics of Phenotypic Plasticity I.
Heritability.” Journal of Evolutionary Biology 2 (2): 95–107.
Schlecker, Louis, Christopher Page, Mikhail Matz, and Rachel M. Wright. 2022. “Mechanisms
and Potential Immune Tradeoffs of Accelerated Coral Growth Induced by
Microfragmentation.” PeerJ 10 (March): e13158.
Schmid, M., and F. Guillaume. 2017. “The Role of Phenotypic Plasticity on Population
Differentiation.” Heredity 119 (4): 214–25.
Schmidt-Nielsen, Knut. 1997. Animal Physiology: Adaptation and Environment. Cambridge
University Press.
Schoepf, Verena, Andréa G. Grottoli, Mark E. Warner, Wei-Jun Cai, Todd F. Melman, Kenneth
D. Hoadley, D. Tye Pettay, et al. 2013. “Coral Energy Reserves and Calcification in a High-
CO2 World at Two Temperatures.” PloS One 8 (10): e75049.
Seneca, Francois O., and Stephen R. Palumbi. 2015. “The Role of Transcriptome Resilience in
Resistance of Corals to Bleaching.” Molecular Ecology 24 (7): 1467–84.
Shaw, Emily C., Robert C. Carpenter, Coulson A. Lantz, and Peter J. Edmunds. 2016.
“Intraspecific Variability in the Response to Ocean Warming and Acidification in the
Scleractinian Coral Acropora Pulchra.” Marine Biology 163 (10): 210.
Simpson, George Gaylord. 1953. “The Baldwin Effect.” Evolution; International Journal of
Organic Evolution 7 (2): 110–17.
192
Somero, G. N. 2010. “The Physiology of Climate Change: How Potentials for Acclimatization
and Genetic Adaptation Will Determine ‘Winners’ and ‘Losers.’” The Journal of
Experimental Biology 213 (6): 912–20.
Stearns, Stephen C. 1989. “The Evolutionary Significance of Phenotypic Plasticity.” Bioscience
39 (7): 436–45.
Stimson, John, and Robert A. Kinzie. 1991. “The Temporal Pattern and Rate of Release of
Zooxanthellae from the Reef Coral Pocillopora Damicornis (Linnaeus) under Nitrogen-
Enrichment and Control Conditions.” Journal of Experimental Marine Biology and Ecology
153 (1): 63–74.
Stinchcombe, J. R., L. A. Dorn, and J. Schmitt. 2004. “Flowering Time Plasticity in Arabidopsis
Thaliana: A Reanalysis of Westerman & Lawrence (1970).” Journal of Evolutionary
Biology 17 (1): 197–207.
Stocking, Jonathan B., Christian Laforsch, Robert Sigl, and Matthew A. Reidenbach. 2018. “The
Role of Turbulent Hydrodynamics and Surface Morphology on Heat and Mass Transfer in
Corals.” Journal of the Royal Society, Interface / the Royal Society 15 (149): 20180448.
Storz, Jay F., Graham R. Scott, and Zachary A. Cheviron. 2010. “Phenotypic Plasticity and
Genetic Adaptation to High-Altitude Hypoxia in Vertebrates.” The Journal of Experimental
Biology 213 (Pt 24): 4125–36.
Su, Yu-Sung, and Masanao Yajima. 2021a. “R2jags: Using R to Run ‘JAGS.’” https://cran.r-
project.org/web/packages/R2jags/index.html.
———. 2021b. “R2jags: Using R to Run ‘JAGS.’” Comprehensive R Archive Network
(CRAN). August 5, 2021. https://cran.r-project.org/web/packages/R2jags/index.html.
Suzuki, Yuichiro, and H. Frederik Nijhout. 2006. “Evolution of a Polyphenism by Genetic
Accommodation.” Science 311 (5761): 650–52.
Szmant, Alina M. 1997. “Nutrient Effects on Coral Reefs: A Hypothesis on the Importance of
Topographic and Trophic Complexity to Reef Nutrient Dynamics.” In Proc 8th Int Coral
Reef Symp, 2:1527–32.
Tambutté, E., A. A. Venn, M. Holcomb, N. Segonds, N. Techer, D. Zoccola, D. Allemand, and
S. Tambutté. 2015. “Morphological Plasticity of the Coral Skeleton under CO2-Driven
Seawater Acidification.” Nature Communications 6 (June): 7368.
Tanhua, Toste, James C. Orr, Laura Lorenzoni, and Lina Hansson. 2015. “Monitoring Ocean
Carbon and Ocean Acidification.” Bulletin N\textordmasculine 64: 1.
Tautz, D., and M. Renz. 1983. “An Optimized Freeze-Squeeze Method for the Recovery of DNA
193
Fragments from Agarose Gels.” Analytical Biochemistry 132 (1): 14–19.
Taylor, Susan S., and Alexandr P. Kornev. 2011. “Protein Kinases: Evolution of Dynamic
Regulatory Proteins.” Trends in Biochemical Sciences 36 (2): 65–77.
Therneau, Terry M. 2020. Coxme: Mixed Effects Cox Models (version 2.2-16). Comprehensive R
Archive Network (CRAN). https://CRAN.R-project.org/package=coxme.
Therneau, Terry M., Thomas Lumley, Elizabeth Atkinso, and Cynthia Crowson. 2022. Survival:
Survival Analysis (version 3.3-1). Comprehensive R Archive Network (CRAN).
https://CRAN.R-project.org/package=survival.
Tiffin, Peter, and Mark D. Rausher. 1999. “Genetic Constraints and Selection Acting on
Tolerance to Herbivory in the Common Morning Glory Ipomoea Purpurea.” The American
Naturalist 154 (6): 700–716.
Todd, P. A., R. J. Ladle, N. J. I. Lewin-Koh, and L. M. Chou. 2004. “Genotype × Environment
Interactions in Transplanted Clones of the Massive Corals Favia Speciosa and Diploastrea
Heliopora.” Marine Ecology Progress Series 271: 167–82.
Todd, Peter A. 2008. “Morphological Plasticity in Scleractinian Corals.” Biological Reviews of
the Cambridge Philosophical Society 83 (3): 315–37.
Traylor-Knowles, Nikki, and Michael T. Connelly. 2017. “What Is Currently Known About the
Effects of Climate Change on the Coral Immune Response.” Current Climate Change
Reports 3 (4): 252–60.
Tufto, Jarle. 2000. “The Evolution of Plasticity and Nonplastic Spatial and Temporal
Adaptations in the Presence of Imperfect Environmental Cues.” The American Naturalist
156 (2): 121–30.
Tunnicliffe, V. 1981. “Breakage and Propagation of the Stony Coral Acropora Cervicornis.”
Proceedings of the National Academy of Sciences of the United States of America 78 (4):
2427–31.
Urbina-Barreto, Isabel, Frédéric Chiroleu, Romain Pinel, Louis Fréchon, Vincent Mahamadaly,
Simon Elise, Michel Kulbicki, et al. 2021. “Quantifying the Shelter Capacity of Coral Reefs
Using Photogrammetric 3D Modeling: From Colonies to Reefscapes.” Ecological
Indicators 121 (February): 107151.
Valladares, Fernando, José Chico, Ismael Aranda, Luis Balaguer, Pierre Dizengremel, Esteban
Manrique, and Erwin Dreyer. 2002. “The Greater Seedling High-Light Tolerance of
Quercus Robur over Fagus Sylvatica Is Linked to a Greater Physiological Plasticity.” Trees
16 (6): 395–403.
Van Buskirk, J., and U. K. Steiner. 2009. “The Fitness Costs of Developmental Canalization and
194
Plasticity.” Journal of Evolutionary Biology 22 (4): 852–60.
Veal, C. J., M. Carmi, M. Fine, and O. Hoegh-Guldberg. 2010. “Increasing the Accuracy of
Surface Area Estimation Using Single Wax Dipping of Coral Fragments.” Coral Reefs 29
(4): 893–97.
Velotta, Jonathan P., and Zachary A. Cheviron. 2018. “Remodeling Ancestral Phenotypic
Plasticity in Local Adaptation: A New Framework to Explore the Role of Genetic
Compensation in the Evolution of Homeostasis.” Integrative and Comparative Biology 58
(6): 1098–1110.
Velotta, Jonathan P., Catherine M. Ivy, Cole J. Wolf, Graham R. Scott, and Zachary A.
Cheviron. 2018. “Maladaptive Phenotypic Plasticity in Cardiac Muscle Growth Is
Suppressed in High-Altitude Deer Mice.” Evolution; International Journal of Organic
Evolution 72 (12): 2712–27.
Via, Sara, and Russell Lande. 1985. “Genotype-Environment Interaction and the Evolution of
Phenotypic Plasticity.” Evolution; International Journal of Organic Evolution 39 (3): 505–
22.
Via, S., R. Gomulkiewicz, G. De Jong, S. M. Scheiner, C. D. Schlichting, and P. H. Van
Tienderen. 1995. “Adaptive Phenotypic Plasticity: Consensus and Controversy.” Trends in
Ecology & Evolution 10 (5): 212–17.
Vollmer, Steven V., and Stephen R. Palumbi. 2007. “Restricted Gene Flow in the Caribbean
Staghorn Coral Acropora Cervicornis: Implications for the Recovery of Endangered Reefs.”
The Journal of Heredity 98 (1): 40–50.
Wall, Christopher B., Contessa A. Ricci, Grace E. Foulds, Laura D. Mydlarz, Ruth D. Gates, and
Hollie M. Putnam. 2018. “The Effects of Environmental History and Thermal Stress on
Coral Physiology and Immunity.” Marine Biology 165 (3): 56.
Walsh, John E., Inna Shapiro, and Timothy L. Shy. 2005. “On the Variability and Predictability
of Daily Temperatures in the Arctic.” Atmosphere-Ocean 43 (3): 213–30.
Ware, Matthew, Eliza N. Garfield, Ken Nedimyer, Jessica Levy, Les Kaufman, William Precht,
R. Scott Winters, and Steven L. Miller. 2020. “Survivorship and Growth in Staghorn Coral
(Acropora Cervicornis) Outplanting Projects in the Florida Keys National Marine
Sanctuary.” PloS One 15 (5): e0231817.
Warner, M. E., W. K. Fitt, and G. W. Schmidt. 1996. “The Effects of Elevated Temperature on
the Photosynthetic Efficiency of Zooxanthellae in Hospite from Four Different Species of
Reef Coral: A Novel Approach.” Plant, Cell & Environment 19 (3): 291–99.
———. 1999. “Damage to Photosystem II in Symbiotic Dinoflagellates: A Determinant of Coral
Bleaching” 96 (12): 8007–12.
195
Water, Jeroen A. J. M. van de, Tracy D. Ainsworth, William Leggat, David G. Bourne, Bette L.
Willis, and Madeleine J. H. van Oppen. 2015. “The Coral Immune Response Facilitates
Protection against Microbes during Tissue Regeneration.” Molecular Ecology 24 (13):
3390–3404.
Water, Jeroen A. J. M. van de, Joleah B. Lamb, Scott F. Heron, Madeleine J. H. van Oppen, and
Bette L. Willis. 2016. “Temporal Patterns in Innate Immunity Parameters in Reef‐building
Corals and Linkages with Local Climatic Conditions.” Ecosphere 7 (11): e01505.
Weis, Virginia M. 2019. “Cell Biology of Coral Symbiosis: Foundational Study Can Inform
Solutions to the Coral Reef Crisis.” Integrative and Comparative Biology 59 (4): 845–55.
Wesselingh, Renate A., Peter G. L. Klinkhamer, Tom J. De Jong, and Laurence A. Boorman.
1997. “Threshold Size for Flowering in Different Habitats: Effects of Size-Dependent
Growth and Survival.” Ecology 78 (7): 2118–32.
West-Eberhard, Mary Jane. 2003. Developmental Plasticity and Evolution. Oxford University
Press.
West-Eberhard, M. J. 1989. “Phenotypic Plasticity and the Origins of Diversity.” Annual Review
of Ecology and Systematics 20 (1): 249–78.
West, G. B., J. H. Brown, and B. J. Enquist. 1997. “A General Model for the Origin of
Allometric Scaling Laws in Biology.” Science 276 (5309): 122–26.
Williams, Dana E., and Margaret W. Miller. 2005. “Coral Disease Outbreak: Pattern, Prevalence
and Transmission in Acropora Cervicornis.” Vol. 301: 119–28.
Williams, D. E., and M. W. Miller. 2006. “Importance of Disease & Predation to the Growth &
Survivorship of Juvenile Acropora Palmata & Acropora Cervicornis: A Demographic
Approach,” January, 1096–1104.
Wilson, Robbie S., and Craig E. Franklin. 2002. “Testing the Beneficial Acclimation
Hypothesis.” Trends in Ecology & Evolution 17 (2): 66–70.
Wilson, Shaun K., Scott C. Burgess, Alistair J. Cheal, Mike Emslie, Rebecca Fisher, Ian Miller,
Nicholas V. C. Polunin, and Hugh P. A. Sweatman. 2008. “Habitat Utilization by Coral
Reef Fish: Implications for Specialists vs. Generalists in a Changing Environment.” The
Journal of Animal Ecology 77 (2): 220–28.
Woesik, Robert, Raymond B. Banister, Erich Bartels, David S. Gilliam, Elizabeth A. Goergen,
Caitlin Lustic, Kerry Maxwell, et al. 2021. “Differential Survival of Nursery‐reared
Acropora Cervicornis Outplants along the Florida Reef Tract.” Restoration Ecology 29 (1).
https://doi.org/10.1111/rec.13302.
Woodley, J. D., E. A. Chornesky, P. A. Clifford, J. B. Jackson, L. S. Kaufman, N. Knowlton, J.
196
C. Lang, et al. 1981. “Hurricane Allen’s Impact on Jamaican Coral Reefs.” Science 214
(4522): 749–55.
Wright, Rachel M., Galina V. Aglyamova, Eli Meyer, and Mikhail V. Matz. 2015. “Gene
Expression Associated with White Syndromes in a Reef Building Coral, Acropora
Hyacinthus.” BMC Genomics 16 (May): 371.
Wright, Rachel M., Carly D. Kenkel, Carly E. Dunn, Erin N. Shilling, Line K. Bay, and Mikhail
V. Matz. 2017. “Intraspecific Differences in Molecular Stress Responses and Coral
Pathobiome Contribute to Mortality under Bacterial Challenge in Acropora Millepora.”
Scientific Reports 7 (1): 2609.
Wright, Rachel M., Hanaka Mera, Carly D. Kenkel, Maria Nayfa, Line K. Bay, and Mikhail V.
Matz. 2019. “Positive Genetic Associations among Fitness Traits Support Evolvability of a
Reef-Building Coral under Multiple Stressors.” Global Change Biology 25 (10): 3294–
3304.
Yakovleva, I., R. Bhagooli, A. Takemura, and M. Hidaka. 2004. “Differential Susceptibility to
Oxidative Stress of Two Scleractinian Corals: Antioxidant Functioning of Mycosporine-
Glycine.” Comparative Biochemistry and Physiology. Part B, Biochemistry & Molecular
Biology 139 (4): 721–30.
Yates, F., and W. G. Cochran. 1938. “The Analysis of Groups of Experiments.” The Journal of
Agricultural Science 28 (4): 556–80.
Yetsko, K., M. Ross, A. Bellantuono, D. Merselis, M. Rodriquez-Lanetty, and M. R. Gilg. 2020.
“Genetic Differences in Thermal Tolerance among Colonies of Threatened Coral Acropora
Cervicornis: Potential for Adaptation to Increasing Temperature.” Marine Ecology Progress
Series 646 (July): 45–68.
Young, Benjamin D., Xaymara M. Serrano, Stephanie M. Rosales, Margaret W. Miller, Dana
Williams, and Nikki Traylor-Knowles. 2020. “Innate Immune Gene Expression in Acropora
Palmata Is Consistent despite Variance in Yearly Disease Events.” PloS One 15 (10):
e0228514.
Young, C. N., S. A. Schopmeyer, and D. Lirman. 2012. “A Review of Reef Restoration and
Coral Propagation Using the Threatened Genus Acropora in the Caribbean and Western
Atlantic.” Bulletin of Marine Science 88 (4): 1075–98.
Zawada, David G., Gregory A. Piniak, and Clifford J. Hearn. 2010. “Topographic Complexity
and Roughness of a Tropical Benthic Seascape.” Geophysical Research Letters 37 (14).
https://doi.org/10.1029/2010gl043789.
Zawada, Kyle J. A., Maria Dornelas, and Joshua S. Madin. 2019. “Quantifying Coral
Morphology.” Coral Reefs 38 (6): 1281–92.
197
Zenni, Rafael Dudeque, Jean-Baptiste Lamy, Laurent Jean Lamarque, and Annabel Josée Porté.
2014. “Adaptive Evolution and Phenotypic Plasticity during Naturalization and Spread of
Invasive Species: Implications for Tree Invasion Biology.” Biological Invasions 16 (3):
635–44.
Ziegler, Maren, Andrea Anton, Shannon G. Klein, Nils Rädecker, Nathan R. Geraldi, Sebastian
Schmidt-Roach, Vincent Saderne, et al. 2021. “Integrating Environmental Variability to
Broaden the Research on Coral Responses to Future Ocean Conditions.” Global Change
Biology 27 (21): 5532–46.
Ziegler, Maren, Cornelia M. Roder, Claudia Büchel, and Christian R. Voolstra. 2014. “Limits to
Physiological Plasticity of the Coral Pocillopora Verrucosa from the Central Red Sea.”
Coral Reefs 33 (4): 1115–29.
Abstract (if available)
Abstract
The potential of phenotypic plasticity to enhance or limit survival, accelerate or hamper evolution, and to even generate novel phenotypic variation make this trait a crucial part of the ecology and evolution of species. Despite these eco-evolutionary implications, our understanding of plasticity’s role in evolution and its potential to evolve has been limited to theory and sparse empirical evidence from choice model systems. While not traditionally considered a model for studying phenotypic plasticity, the sessile and long-lived nature of reef building coral suggest that plasticity, rather than avoidance or migration, may be the predominant strategy to deal with environmental change. Given the threatened or endangered status of coral around the world, understanding the adaptive role of plasticity will be critical to predicting if and how these species will respond to climate change. This dissertation investigates the abundance of intraspecific variation in morphological, physiological, and molecular plasticity in the critically endangered Caribbean coral, Acropora cervicornis. To quantify plasticity’s role on fitness and provide insight into its adaptive potential, I utilized a combination of in situ and lab-based experimentation to induce plasticity in A. cervicornis genotypes across a range of conditions and compared an individual’s potential to be plastic with its overall fitness. I uncover novel intraspecific variation in morphological plasticity and add to previous evidence of variation in physiological and molecular plasticity. While morphological plasticity appears to be adaptive, physiological plasticity showed no impact on coral fitness. Similarly, both fixed and plastic variation in gene expression seem to contribute to survival. These results suggest phenotypic plasticity is an important process in A. cervicornis that contributes to fitness in a trait-specific fashion. Moreover, the work presented here helps establish A. cervicornis as a system for studying the eco-evolutionary dynamics of phenotypic plasticity that also can inform genetic- and environment-based strategies for coral restoration.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Transgenerational inheritance of thermal tolerance in two coral species in the Florida Keys
PDF
Sex differences in aging and the effects of mitochondria
PDF
Application of evolutionary theory and methods to aquatic ecotoxicology
PDF
Disentangling the ecology of bacterial communities in cnidarian holobionts
PDF
Understanding the genetics, evolutionary history, and biomechanics of the mammalian penis bone
PDF
The molecular adaptation of Trichodesmium to long-term CO₂-selection under multiple nutrient limitation regimes
PDF
Genetic architectures of phenotypic capacitance
PDF
Genetic architecture underlying variation in different traits in the Pacific oyster Crassostrea gigas
PDF
Marine protistan diversity, spatiotemporal dynamics, and physiological responses to environmental cues
PDF
Genetic and environmental effects on symbiotic interactions across thermal gradients
PDF
The evolution of pollution tolerance in the marine copepod Tigriopus
PDF
Investigating the potential roles of three mammalian traits in female reproductive investment
PDF
The evolution of gene regulatory networks
PDF
Simulated and field environmental effects on the transcriptome and metabolome of mussel Mytilus californianus
PDF
Thermal acclimation and adaptation of key phytoplankton groups and interactions with other global change variables
PDF
Phylogeography, reproductive isolation, and the evolution of sex determination mechanisms in the copepod Tigriopus californicus
PDF
Deconvolution of circulating tumor cell heterogeneity and implications for aggressive variant prostate cancer
PDF
Male-female conflict after mating: function and dynamics of the copulatory plug in mice (Mus domesticus)
Asset Metadata
Creator
Million, Wyatt Christopher
(author)
Core Title
Phenotypic plasticity and its ecological and evolutionary significance for reef building coral
School
College of Letters, Arts and Sciences
Degree
Doctor of Philosophy
Degree Program
Biology (Marine Biology and Biological Oceanography)
Degree Conferral Date
2022-08
Publication Date
07/25/2022
Defense Date
06/06/2022
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
Acropora cervicornis,gene expression,GxE,morphology,OAI-PMH Harvest,Physiology
Format
application/pdf
(imt)
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Kenkel, Carly (
committee chair
), Dodd, Lynn (
committee member
), Edmands, Suzanne (
committee member
), Ehrenreich, Ian (
committee member
)
Creator Email
wmillion@usc.edu,wyattm300@gmail.com
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-oUC111375218
Unique identifier
UC111375218
Legacy Identifier
etd-MillionWya-10970
Document Type
Dissertation
Format
application/pdf (imt)
Rights
Million, Wyatt Christopher
Type
texts
Source
20220728-usctheses-batch-962
(batch),
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright. The original signature page accompanying the original submission of the work to the USC Libraries is retained by the USC Libraries and a copy of it may be obtained by authorized requesters contacting the repository e-mail address given.
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA
Repository Email
cisadmin@lib.usc.edu
Tags
Acropora cervicornis
gene expression
GxE
morphology