Ecosystems

, Volume 17, Issue 1, pp 1–13

An Ecosystem-Level Perspective on the Host and Symbiont Traits Needed to Mitigate Climate Change Impacts on Caribbean Coral Reefs

Authors

    • Marine Spatial Ecology Lab, School of Biological SciencesThe University of Queensland
  • Manuel González-Rivero
    • Marine Spatial Ecology Lab, School of Biological SciencesThe University of Queensland
    • Coral Reefs Ecosystems Lab, School of Biological SciencesThe University of Queensland
  • Peter J. Mumby
    • Marine Spatial Ecology Lab, School of Biological SciencesThe University of Queensland
Article

DOI: 10.1007/s10021-013-9702-z

Cite this article as:
Ortiz, J.C., González-Rivero, M. & Mumby, P.J. Ecosystems (2014) 17: 1. doi:10.1007/s10021-013-9702-z

Abstract

Caribbean reefs have steadily declined during the past 30 years. Thermal disturbances that elicit coral bleaching have been identified as a major driver of such coral degradation. It has been suggested that either the evolution of more tolerant symbionts, or shifts in the distribution of existing, tolerant symbionts could ameliorate the effect of rising sea temperatures on Caribbean reefs. Using a spatial ecosystem model we describe the characteristics that new tolerant symbionts, ‘super-symbionts’, and their coral hosts, require for coral cover to be maintained. We also quantify the time necessary for such symbionts to become dominant before their potential beneficial effect is lost. Running scenarios under two levels of greenhouse gas emissions, we find that aggressive action to reduce emissions could almost triple the time available for new super-symbionts to become dominant and potentially mitigate the effect of thermal disturbances. The benefits of thermally tolerant super-symbionts depend on the life-history traits of the host, the number of coral species infected and the present coral assemblage. Corals that are strong competitors with macroalgae are likely to become dominant on future reefs if a super-symbiont appears in the next 25–60 years. In principle, super-symbionts could have ecosystem-level benefits in the Caribbean providing that they become dominant in multiple coral hosts with specific life-history traits within the next 60 years. This potential benefit would only be realized if the appearance of the super-symbiont is combined with drastic reductions of greenhouse gas emissions and maintenance of ecosystem processes such as herbivory.

Keywords

coral reefsecosystem statelife history traitsclimate changethermal tolerancesymbiosisecosystem modeling

Introduction

The endosymbiotic relationship between scleractinian corals and dinoflagellate algae is essential for maintaining the high productivity observed on coral reefs (Hatcher 1988). The symbionts provide coral hosts with energy that is utilized for rapid deposition of calcium carbonate and other metabolic processes (Muscatine 1990). During periods of anomalously high sea surface temperatures, the coral–algal symbiosis is disrupted and corals appear bleached as the concentration of symbiotic algae in their tissue is reduced (Douglas 2003; Enriquez and others 2005). Persistent thermal disturbances generate mass bleaching events and can result in extensive coral mortality over regional and even global scales (Wilkinson and others 1999; Berkelmans and others 2004). In the Caribbean, mortality associated with mass bleaching events was first described in 1987, and the frequency and intensity of bleaching events has increased thereafter (Williams and others 1987; Glynn 1991; Eakin and others 2010).

The discovery of relatively uncommon symbiont types that increase the thermal tolerance of their coral host has been interpreted as a potential safeguard against climate change impacts (Baker and others 2004). A compelling argument is that an increase in the relative abundance of these tolerant symbionts could represent a natural adaptive process that would ameliorate the impact of increasing sea surface temperatures (Berkelmans and van Oppen 2006; Jones and others 2008). An apparent increase in the dominance of tolerant symbionts worldwide, and the discovery that the symbiont assemblage can change during and after bleaching events has been interpreted as evidence in support of these hypotheses (Baker and others 2004; Oliver and Palumbi 2009). Moreover, the potential for rapid adaptation of the symbionts as a consequence of a rapid generation time (Csaszar and others 2010), coupled with a strong selective pressure, has led to suggestions that symbionts may yet achieve even greater thermal tolerance and therefore help counteract the effect of climate change (Baskett and others 2009; Pandolfi and others 2011).

Recent studies focusing on the physiological properties of tolerant symbionts have shown that the increased tolerance observed in some symbionts has an associated metabolic cost (Little and others 2004; Smith 2010). Those symbionts exhibiting greater thermal tolerance (for example, symbiont clade D1a) appear to be less efficient at transferring energy to the coral host and therefore the host grows at a slower rate. Specifically, the benefit of clade D1a—the first symbiont type described as thermo-tolerant—has been found to reduce the mortality rate of hosts by 30% under thermal stress but at a cost of a reduction in coral growth rate of 50–60% (Little and others 2004; Berkelmans and van Oppen 2006; Jones and others 2008; Smith 2010).

A reduction in the somatic growth rate of corals could have deleterious consequences at population and ecosystem scales. Specifically, rates of coral recovery from disturbance could be impaired for several reasons, all of which may act in concert. If individual corals grow slowly, the fundamental rate of population recovery is reduced unless compensated for by an increase in the rate of recruitment or reduction of mortality. However, recruitment is likely to decrease under slower growth, which may exacerbate the impact of slower growth on population recovery. First, juvenile corals are particularly vulnerable to predation (Box and Mumby 2007), though they appear to reach a threshold size escape (Doropoulos and others 2012). If growth rates are reduced, corals will take longer to reach a size escape and therefore remain vulnerable to predators for longer, which increase their per capita mortality rate. Second, smaller corals fare less successfully than larger corals during competitive interactions with fleshy macroalgae (Ferrari and others 2012a, b). Thus, coral mortality arising from competition is likely to increase if coral growth is reduced. Third, coral maturation appears to be size-dependent (Soong and Lang 1992) and therefore slower growth delays the onset of reproduction, potentially reducing reproductive output of the population and future recruitment.

The ecological mechanisms listed above suggest that the benefits of a coral obtaining thermally tolerant symbionts should be considered in light of the ecological cost of impaired growth. In a recent article, Ortiz and others (2013) described the traits required by a symbiont to ameliorate the ecological costs of retarded growth and withstand the increase in thermal stress under climate change. The study focused specifically on Caribbean reefs where changes in growth rate are likely to be critical because of frequent interactions between corals and macroalgae (Roff and Mumby 2012). The study examined trade-offs between the beneficial increase in thermal tolerance conferred by the symbiont and the cost of reduced coral growth (Ortiz and others 2013). Several combinations of thermal tolerance versus growth impairment were identified that exerted a net benefit to the ecosystem. However, the values for each trait fell well outside of the range reported to date. For example, the coral mortality from bleaching has to decrease fourfold beyond levels currently observed, and the costs to growth must drastically reduce from a current level of 50% growth reduction to only a 20% growth reduction. Here we term those hypothetical symbionts that met the required traits ‘super-symbionts’ to highlight the disparity with observed traits.

Two mechanisms could lead to the emergence of super-symbionts: rapid adaptation of existing symbionts, or rapid selection of extant symbionts that have not been detected because of their low relative abundance or constrained geographic distribution. Both mechanisms will incur a time delay before being realized. Unfortunately, a delay in the onset of super-symbionts could reduce their potential benefit. In an extreme case, a Caribbean reef that exhibits hysteresis could degrade to the point where it becomes entrained in a coral-depauperate basin of attraction that is reinforced by negative feedbacks on recruitment (Mumby and Steneck 2008; Mumby and others 2013). Under this scenario, the emergence of a super-symbiont might not be sufficient to enable significant coral recovery. There is, therefore, a trade-off between the length of time taken for establishment of a super-symbiont—whose occurrence is positively correlated with time—and the potential for a beneficial impact on a degrading ecosystem, where the scope for benefit is negatively correlated with time.

To date, the majority of studies exploring the ability of coral reefs to cope naturally with climate change have been mainly focused on the likelihood of the evolution of new symbionts (Little and others 2004; Berkelmans and van Oppen 2006; Jones and others 2008; Smith 2010; Ortiz and others 2013). The potential impact that different life history traits of the coral host could have on the benefit obtained by the acquisition of such hypothetical super-symbionts has not been considered. It has been generally assumed that corals with different life strategies would benefit similarly from these new partnerships (Baker and others 2004). Given the identified trade-off between coral growth and thermal tolerance when acquiring a more thermo-tolerant symbiont, it could be expected that the intrinsic growth rate and thermal sensitivity of the coral host may play a role in the benefit obtained. Similarly, as the ability of corals to compete against macroalgae affects coral growth rate, competitive ability may also affect the benefits provided by a particular super-symbiont.

Our first analysis of symbiont traits focused on the symbiont itself and assumed that traits appeared immediately. Here we extend our earlier study in two ways. First, we examine what combinations of symbiont traits would best facilitate the persistence of corals in a complex ecosystem and under climate change. Second, we ask how quickly such super-symbionts must become dominant given the increasing rate of disturbance facing reefs and the scope for alternate ecosystem attractors (Scheffer and Carpenter 2003). Importantly, our second analysis focuses on how the traits of the coral host and the composition of the coral assemblage affect the population and ecosystem benefit obtained from the appearance of the super-symbiont. Our previous study did not consider the importance of host-symbiont partnerships; it only considered the traits of the symbiont itself. Here, we find that traits of the coral host are particularly important in determining the potential benefits of new symbionts.

The principal aims of this study are:
  • To determine if a reduction in carbon emissions significantly increases the ability of Caribbean coral reefs to overcome the challenges posed by climate change, by reducing the speed with which new thermally tolerant symbionts would have to become dominant.

  • To determine the effect of ecosystem-level parameters, such as community structure and the intensity of coral-macroalgal competition, on the ability of different coral types to benefit from the appearance of new thermo-tolerant super-symbionts.

Methods

The Basic Model

The simulation model employed in the present study was designed to represent mid-depth (6–15 m) Montastraea-dominated forereefs in the Caribbean, which typically have the highest biomass and diversity of reef organisms (Mumby and others 2008). Since white band disease has depleted populations of large, branching corals (Aronson and Precht 2001) stylized massive growth forms of coral were simulated together with rates of recruitment, growth, reproduction, and mortality. The model is a square lattice of 2,500 cells each of which approximates 0.25 m2 of reef, and can be occupied by a mixture of living and dead substrata. Although the reef has a toroidal lattice of 2,500 cells, the lattice structure merely helps define probabilistic rules of coral recruitment and vegetative algal growth. Individual cells comprise multiple coral colonies and algal patches so interactions occur at colony scales as they do in situ. The reef has continuous boundaries, arranged as a torus. Corals can recruit to individual patches of cropped algae but not macroalgae. Macroalgae grow vegetatively and can overgrow corals. Grazing affects all algal classes and always results in the first grazed algal class (cropped algae). An unexploited community of parrotfishes can maintain up to 30–40% of the reef area in a grazed state (Mumby 2006). Competitive interactions between corals and macroalgae reduce the growth rate of each taxon and are the only processes modeled to occur across cell boundaries (within a 4-cell von Neumann neighborhood). The arrangement of elements within an individual cell has no explicit spatial structure, but coral-coral competition can occur at intra-cellular scales. Corals are subjected to size-dependent fecundity and mortality, resulting in three functional categories: recruits (horizontal cross-sectional area 1–60 cm2), pubescents (61–250 cm2), and adults (>250 cm2). All simulations assume a no stock-recruitment relationship and corals recruit at maximum levels irrespective of stock size (that is, up to 4 recruits per 0.25 m−2). Individual cells in the lattice are updated in random sequence using discrete intervals of 6 months. The parameterization was based on reefs with little sediment deposition; therefore no effect of sediment on recruitment is incorporated. All parameters were fitted from empirical studies (Appendix 1 in Supplementary Material).

Fifty simulations were run for each scenario for a duration of 180 time steps (90 years), as this is the length of the available climate projections (Edwards and others 2011). These projections extend to the year 2100.

Thermal Stress

Thermal stress is implemented in the model as bleaching events following Edwards and others (2011). Bleaching events are triggered when the summer modeled sea surface temperature generates more than 4 degree heating weeks in a summer season. When a bleaching event is triggered there is partial and total colony mortality associated with it. Mortality rates are species specific, size-specific, and consider whether a colony has experienced previous bleaching or not. Total mortality due to bleaching is calculated as a function of the intensity and duration of the thermal stress using empirical relationships (Ortiz and others 2013). To focus purely on the benefits of super-symbionts to corals under global warming, we did not consider local disturbances such as hurricanes and epizootics that could obfuscate the link between super-symbionts and reef trajectory.

Climate Scenarios

Two climate change scenarios were considered for the calculation of partial and total coral mortality owing to bleaching events (Edwards and others 2011). The low carbon emissions scenario (Representative Concentration Pathway 2.6) represents the future trajectory of sea surface temperature anomalies considering an immediate drastic reduction in greenhouse gas emissions. This scenario implies a peaking of about 450 ppm (CO2-eq) by 2040 and 380 ppm (CO2-eq) by 2100. The business-as-usual emissions scenario (Representative Concentration Pathways 8.5) represents a high emissions situation where emissions continue to grow and little action is taken to reduce emissions in the near future. In this scenario, CO2 concentrations increase linearly and reach 1,200 ppm (CO2-eq) by 2100. All climate variables were provided as a spatial mean across the Caribbean Sea (Jones and others 2011).

Coral Species

Four representative coral species were included in the model. Two species, Montastraea annularis and Montastraea cavernosa, grow faster and are relatively tolerant to thermal stress compared to Porites astreoides and Mycetophyllia danaana. Of the four species, M. danaana is the strongest competitor against macroalgae (Table 1).
Table 1

Description of the Four Coral Types Included in the Model

Representative taxa

Species code

Growth rate

Thermal tolerance

Competitive ability against macroalgae

Reproduction strategy

Lobophora

Dictyota

Overall

Porites astreoides

P. ast

Slow

Low

Intermediate

Low

Low

Brooder

Montastraea cavernosa

M. cav

Fast

High

Intermediate

High

Intermediate

Spawner

Mycetophyllia danaana

M. dan

Slow

Low

Very high

Low

High

Brooder

Montastraea annularis

M. anu

Fast

High

Low

High

Intermediate

Spawner

Super-Symbiont Types

Symbionts are modeled to have two traits with opposing impacts on the host performance: (1) improved rates of survival (a benefit), and (2) reduced rates of somatic growth (a cost). A previous paper identified six combinations of these traits that had a net ecological benefit to corals (Ortiz and others 2013) (Figure 1). Here, we focus on the three sets of traits that are closest to those that have been described for clade D1a—an existing thermo-tolerant symbiont—and therefore the most likely to occur in nature (Table 2).
https://static-content.springer.com/image/art%3A10.1007%2Fs10021-013-9702-z/MediaObjects/10021_2013_9702_Fig1_HTML.gif
Figure 1

Effect of different tolerant symbionts on reef state after 65 years. High carbon emissions scenario. Color scale represents final cover with tolerant symbiont minus final cover without tolerant symbiont. Difference in reef state plotted as a function of relative growth (proportion of colony growth relative to the growth of holobionts with “normal” symbiont) and relative mortality (proportion of colony bleaching induced mortality relative to the mortality of holobionts with “normal” symbiont). Dotted line represents the upper limit of the empirical range for growth and mortality in corals with a tolerant symbiont similar to clade D1a. Of the 25 combinations of symbiont parameters the three denoted with asterisks are the ones explored in this article (modified from Ortiz and others 2013).

Table 2

Characteristics of the Super-Symbionts Studied and Clade D1a (Known Temperature Tolerant Symbiont)

Symbiont type

COST: Coral host growth (relative to growth of the same coral host harboring a less tolerant symbiont)

BENEFIT: Temperature-induced mortality (relative to mortality of the same coral host harboring a less tolerant symbiont)

Reference clade D1a

50%

30%

Type 1—growth over thermal tolerance

90% (lowest cost)

20% (lowest benefit)

Type 2—thermal tolerance over growth

60% (highest cost)

5% (greatest benefit)

Type 3—growth and thermal tolerance simultaneously

80% (intermediate cost)

10% (intermediate benefit)

The percentages represent the proportion of growth or mortality when compared to a coral inhabited by sensitive symbionts (for example, corals harboring clade D1a present 30% of the bleaching induced mortality of corals inhabited by sensitive symbionts but also present 50% of their growth rate).

Three of the four coral species in this study have been shown to switch symbionts during a bleaching event (LaJeunesse and others 2009a, b); the fourth M. danaana, has not been investigated. Each species showed a different propensity to switch symbionts, and many colonies appear to revert back to their original symbiont within 2 years after bleaching (LaJeunesse and others 2009a, b). Therefore, the most likely scenario involves some corals being more likely to switch symbionts than others, and some being less likely to exhibit a reversal of symbionts after bleaching. Unfortunately, a lack of empirical evidence precludes a species-specific parameterization for either type of symbiont switching. Thus, all coral species were allowed to switch to the new super-symbiont during a bleaching event and remain with the new symbiont thereafter. Although this scenario is unlikely, empirical evidence suggests that it is possible: LaJeunesse and others (2009a, b) report that a small proportion of corals switched to D1a and retained the novel symbiont even 2 years after bleaching.

Initial Conditions

Initial coral cover was set just above representative levels for the Caribbean at 20%. We assumed that the branching coral, Acropora, remained functionally absent on the reef (Bythell and Sheppard 1993). We assumed that the urchin Diadema remained absent (Hughes and others 2010) and that parrotfish populations were not exploited and able to maintain 33% of the reef in a grazed state (Mumby 2006).

Model Runs

Maximum Time Available Before the Benefit of the Super-Symbiont is Lost

To determine the latest date (number of years) by which a super-symbiont must appear to provide an ecosystem-level benefit, the model was initialized with an even composition of corals (5% cover for each coral species = 20% total initial coral cover) and run for both climate change scenarios. At the beginning of each simulation every coral colony was assigned its “normal sensitive” symbiont type. Super-symbionts became available in the environment at years 5, 15, 25, 35, 45, or 60. Once super-symbionts were available, all corals had a 40% probability of becoming colonized by the super-symbiont during individual bleaching events. This rate was based on evidence for clade D1a (LaJeunesse and others 2009a, b). Other probabilities of colonization were considered but the results were not significantly affected (data not shown). The effect of the super-symbiont was considered to be positive or ‘successful’ if the coral cover at the end of the simulated period was at least equal to initial coral cover (that is, no net coral loss under climate change).

Minimum Number of Coral Species That Have to be Infected by Super-Symbionts Before the Benefit is Lost

Fourteen new runs were performed focusing on the “successful” scenario with the latest appearance of super-symbionts. In each run, one of the four coral species or one of the ten possible combinations of two or three species was allowed to switch symbionts (Table 3). The final coral cover in each of these runs was then explored to determine the minimum number of coral species that have to be infected by super-symbionts before the benefit is lost.
Table 3

Final Coral Cover as a Function of the Identity and Number of Coral Species Allowed to Switch Symbionts

Coral species allowed to switch symbionts

Final coral cover

Low carbon emissions

High carbon emissions

P. ast

9.4

4.5

M. cav

12.8

7.5

M. dan

18.8

9.1

M. anu

13.8

8.3

P. ast, M. cav

20.4

8.8

P. ast, M. dan

21.4

11.1

P. ast, M. anu

20.1

10.3

M. cav, M. dan

24.8

14.1

M. cav, M. anu

21.6

13.4

M. dan, M. anu

25.8

14.9

P. ast, M. cav, M. dan

29.2

16.1

P. ast, M. cav, M. anu

26

15.3

P. ast, M. dan, M. anu

30.2

17

M. cav, M. dan, M. anu

33.6

19.9

P. ast, M. cav, M. dan, M. anu

38.1

21.9

All simulations were run with the introduction of super-symbiont type 3 (growth and thermal tolerance simultaneously) 25 years after the start of the simulation. Numbers in italics highlight final coral cover higher or equal to initial coral cover (no net loss).

Effect of Coral Species Traits on the Ecosystem-Level Benefit Provided by Super-Symbionts

To determine which coral species benefited the most from the appearance of the super-symbiont the trajectory of each coral species was examined for the “successful” scenario with the latest appearance of super-symbionts.

To examine the effect of the initial coral assemblage on the benefits provided by super-symbionts’ colonization we explored both climate change scenarios with five different starting coral assemblages giving a total of ten scenarios. For each climate change scenario, we ran an even starting coral assemblage (where each coral had an initial abundance of 5%), and four uneven scenarios, each being initially dominated by one of the four coral species (12% initial cover vs. 2.5% initial cover for the other three species).

To understand why some coral assemblages fared better than others, the population growth rate (PGR) was calculated for each species. The PGR is the rate of change in cover over a single time step, following equation (1) such that
$$ {\text{PG}}{{\text{R}}_{i,t}} = {C_{i,t}} - {C_{i,t - 1}} $$
(1)
where C is the cover of species i at time t and each time step is 6 months. To compare trajectories between pairs of species, the ratio of their PGR was calculated and then plotted against macroalgal abundance to explore whether species with a higher extension rate but lower competitive ability against macroalgae had a relatively low PGR when macroalgal cover was high.

Results

Reducing greenhouse gases had a strong influence on the predicted response of corals with a super-symbiont (Figure 2A, B). Under low carbon emissions, two of the three symbiont types were capable of maintaining coral cover even when their establishment was delayed by up to 60 years (Figure 2A). In contrast, only one of the symbionts was capable of maintaining coral cover under a business-as-usual carbon emissions scenario, and only if established within 25 years (Figure 2B). Final coral cover was consistently higher under low carbon emissions and the improvement ranged considerably between 30 and 400% depending on the symbiont type and onset of the super-symbiont (Figure 2A, B).
https://static-content.springer.com/image/art%3A10.1007%2Fs10021-013-9702-z/MediaObjects/10021_2013_9702_Fig2_HTML.gif
Figure 2

Coral cover after 90 years of simulations, as a function of the time taken for different super-symbionts to become dominant in the system (X-axis). A Low emissions scenario, B high emissions scenario. Different color lines represent different symbiont types. Dotted lines represent the starting coral cover for all scenarios (20%).

Under the low carbon emissions scenario it was only necessary for the super-symbiont to become dominant in any two of the four coral species to maintain coral cover (Table 3). However, under the business-as-usual scenario, the super-symbiont had to become dominant in all four coral species to be capable of maintaining coral cover after 90 years (Table 3).

Of the three super-symbiont types, that with intermediate benefits and costs had the greatest and most consistent positive effect on the coral population (Figure 2A, B). The next most desirable super-symbiont had a lower cost to growth, albeit with reduced benefits to thermal tolerance. Thus, in terms of trade-offs between alternate symbiont traits, it was more desirable for the host to sacrifice some of the survivorship advantage of a super-symbiont to minimize the reduction in the host’s somatic growth (Figure 2A, B).

Of the four coral species, the strongest competitor against macroalgae (M. danaana) benefited the most from the establishment of the super-symbiont (Figure 3A, B). The strength of this benefit occurred despite M. danaana being a slow growing, thermally sensitive coral. The species that benefited the least from a super-symbiont was also a brooder with slow growth and low thermal tolerance (P. astreoides) but its competitive ability against macroalgae was poor (Figure 3A, B). The faster growing, more temperature-tolerant, spawning corals, with intermediate competitive abilities (M. annularis and M. cavernosa) received an intermediate benefit from the establishment of the super-symbiont (Figure 3A, B). Thus, the benefits to a host of increased thermal tolerance are not related to the species’ sensitivity to thermal stress; rather, they appear to be driven by competitive ability against macroalgae.
https://static-content.springer.com/image/art%3A10.1007%2Fs10021-013-9702-z/MediaObjects/10021_2013_9702_Fig3_HTML.gif
Figure 3

Trajectory of the cover of the four coral taxa over time when the intermediate super-symbiont becomes dominant 25 years after the start of the simulation (maximum time necessary for coral cover to be maintained in high emissions scenario) under contrasting carbon emissions scenarios: A low carbon emissions, B business as usual.

Although M. danaana benefited the most from the acquisition of a super-symbiont, this did not mean that assemblages dominated by this species resulted in the most coral. Indeed, when the initial assemblage structure was modified from an even assemblage to one dominated by any of the four coral species, the reef trajectory changed significantly (Figure 4A, B). The highest final coral cover was obtained when either of the two faster-growing spawners was initially dominant, despite them only having intermediate competitive abilities against macroalgae. Final coral cover was lowest when either of the brooders dominated the assemblage. The assemblage of greatest evenness exhibited an intermediate final cover (Figure 4A, B).
https://static-content.springer.com/image/art%3A10.1007%2Fs10021-013-9702-z/MediaObjects/10021_2013_9702_Fig4_HTML.gif
Figure 4

Trajectory of total coral cover as a function of the initial coral host assemblage and contrasting carbon emissions scenarios: A low carbon emissions, B high carbon emissions. The intermediate super-symbiont becomes dominant 25 years after the start of the simulation. Table inserted in shaded area provides slope of the linear part of the recovery phase (area of the curves within the shaded area).

Corals usually benefited from the colonization of a super-symbiont. However, the rate of coral recovery after the symbiont colonization event was sensitive to the absolute abundance of coral at that time: Reefs with more coral—those initially dominated by fast growing spawners—exhibited faster recovery, demonstrated by an increase in the slope of the recovery trajectory of up to 40% (insets in Figure 4A, B).

To investigate why M. danaana fared better than M. annularis even though its growth rate is slower, we plotted the ratio of each species’ realized instantaneous population growth (RIPGR) rate as a function of macroalgal cover (Figure 5C, D). M. danaana was favored when macroalgal cover was high, whereas M. annularis was favored as macroalgal cover decreased (Figure 5A–D). As a consequence, even though M. danaana was the most abundant species at the end of all scenarios, it only dominated the assemblage when final coral cover was low (40% of total coral cover) (Figure 5A, B).
https://static-content.springer.com/image/art%3A10.1007%2Fs10021-013-9702-z/MediaObjects/10021_2013_9702_Fig5_HTML.gif
Figure 5

Upper panels Trajectory of the cover of the four coral taxa over time with two different initial coral assemblages (A dominated by M. danaana or B dominated by M. anularis) when the intermediate super-symbiont becomes dominant 25 years after the start of the simulation (maximum time necessary for coral cover to be maintained in high emissions scenario), under low carbon emissions scenario. Lower panels Ratio of population growth rate (PGR) between M. danaana and M anularis (values greater than one represent higher grow rate of M. danaana) as a function of macroalgal cover (C initial assemblage dominated by M. danaana or D dominated by M. anularis) after the intermediate super-symbiont becomes dominant 25 years after the start of the simulation (maximum time necessary for coral cover to be maintained in high emissions scenario), under low carbon emissions scenario.

Discussion

The model predicts that a super-symbiont would need to dominate reefs fairly rapidly and the time required depends, in part, on the rate of climate change. Under significant reduction of greenhouse gas emissions, a super-symbiont would need to dominate at least two coral species within 60 years to maintain coral cover over time. This super-symbiont would have to provide either 33% more benefit than the existing clade D1a at only 20% of the cost (symbiont type 1 in Table 1), or provide 66% more benefit at only 40% of the cost (symbiont type 3 in Table 1). Even though potential for thermal adaptation in symbionts has been identified (Baskett and others 2009; Csaszar and others 2010), the rate of such adaptation is uncertain and the concomitant costs in terms of coral holobiont growth are unknown.

The business-as-usual greenhouse gas emissions scenario (RCP 8.5) presents a considerable challenge for selection. Under this scenario, only one of the three super-symbionts can provide a net benefit to the reef and its establishment would be required within 25 years, approximately two and a half times faster than under aggressive mitigation of greenhouse gases. Furthermore, the model predicts that all four coral species would need to be dominated by the super-symbiont. Even when three of the four coral species included in our analysis have been found to be capable of switching to a different symbiont during bleaching, some species appear to be more likely to switch than others, and a high proportion of the colonies that switch tend to revert back to the original symbiont type within 2 years of the bleaching event (Thornhill and others 2006; LaJeunesse and others 2009a, b). Therefore the likelihood of all coral species becoming infected and then retaining the super-symbiont is lower than that of only two species, as required under the low emissions scenario. If some species are unable to associate with more thermally tolerant symbionts then they may be placed at a greater evolutionary disadvantage if greenhouse gases continue to follow RCP 8.5.

There has been considerable interest in identifying the traits that make species winners or losers under climate change (Loya and others 2001; Darling and others 2012). The results reveal that species-level traits may be overtaken by community-level interactions in determining the real winners. Here, the main selection pressure and stress is rising sea temperature. Logically, the host species poised to benefit the most from a new thermally tolerant symbiont is that with greatest sensitivity to temperature (that is, the species that experiences the greatest mortality during bleaching). However, in this study the key driver of a species’ benefit was its competitive ability against macroalgae; the strongest competitor benefited the most from a super-symbiont.

The importance of competition with macroalgae reflects the strength of this driver in Caribbean ecosystems, not only in terms of its high abundance (Roff and Mumby 2012) but also the diversity of its deleterious effects on corals (Burkepile and Hay 2006; Nugues and Bak 2006; Mumby and others 2007a, b; Arnold and others 2010; Ferrari and others 2012a, b). Essentially, corals that are weak competitors with macroalgae have relatively little opportunity to take advantage of the benefits accrued from a super-symbiont even if they are relatively thermo-tolerant and grow rapidly. For these corals, their realized growth rate is a combination of their intrinsic growth rate minus the partial mortality generated by their poor competitive ability against macroalgae and the reduction in somatic growth that occurs during coral-algal interactions (Box and Mumby 2007; Ferrari and others 2012a, b). Therefore, as algal cover increases, a point is reached after which the PGR of faster-growing corals becomes less than that of corals like M. danaana with a lower growth rate but higher competitive ability against macroalgae (values >1 in Figure 5C, D). This competitive mechanism explains the link between initial community composition and the subsequent trajectory. When the reef was initially dominated by the slow-growing M. danaana, coral cover decreased faster than when it was initially dominated by faster growing corals. Thus, coral cover was relatively low and macroalgal cover was high when the super-symbiont entered the population. Because M. danaana fares best under high macroalgal cover, it eventually dominated the coral assemblage even though the total coral cover was lower than that in scenarios dominated by faster-growing corals. A similar argument explains the trade-off between symbiont traits in which hosts perform better if growth costs are minimized rather than if survivorship benefits are maximized. Here, any impact on a host’s competitive ability, such as reducing its rate of somatic growth, will have large deleterious consequences at a community scale.

The frequency and duration of competitive interactions between corals and macroalgae will depend, in part, on the intensity of grazing. Grazing intensity depends on the abundance of herbivores (bites per unit time) and the space in which they have to feed (bites per unit time per unit area). As coral cover increases, herbivores confine their feeding into an ever-smaller area which increases its intensity (Williams and others 2001; Mumby and Steneck 2008). It is for this reason that the rate of coral recovery was greater on reefs with more coral. Reefs with low coral cover—regardless of the species—tend to have more macroalgae and more frequent competitive interactions between corals and macroalgae (Hughes and others 1987). Thus, the overall trajectory of a reef was directly sensitive to the amount of coral, irrespective of its identity. If fast-growing species were present then coral cover tended to be higher at any given point in time. Higher coral and higher grazing intensity then facilitated the recovery of any coral once a super-symbiont became established. Thus, the emerging picture of winners and losers is complex. Strong competitors have the greatest potential to benefit from an increase in survivorship during bleaching. Fast-growing corals might not experience the greatest benefit in relative terms, but they will often serve as facilitators for the recovery of other corals by helping to increase grazing intensity. The net outcome of these influences depends on the initial cover of each species. For example, the higher the initial abundance of M. danaana the more it will dominate in future. However, a reef dominated by M. danaana will tend to have lower cover overall because the initial paucity of faster-growing corals reduces their contribution to cover and their role as facilitators.

The results of the model appear to contrast with those of Baskett and others (2009), who suggest that under certain conditions, coral cover could be maintained for the next 100 years through symbiont adaptation. However, it is difficult to draw firm comparisons between studies. The model used in Baskett and others (2009) is a combination of two continuous population models (one for the symbionts and one for the coral host) and a genetic model for the adaptation of the symbionts to thermal stress. The model does not consider coral–macroalgal interactions which this study shows can alter the trajectories of reefs considerably. In addition, Baskett and others (2009) included a fast-growing Acropora type as one of the two corals modeled. The growth rate of acroporids in the Caribbean is about 10 times that of massive corals. Acropora was not included here because of its continued scarcity following epizootics in the 1980s and 1990s (Baums and others 2005; Vollmer and Palumbi 2007; Williams and others 2008). Yet, had Acropora been included the prognosis for Caribbean reefs would be radically better. A recent model suggests that Caribbean reefs do not exhibit alternate attractors in the presence of Acropora (Roff and Mumby 2012) such that coral recovery occurs even when cover and grazing intensity are low. Another difference between the studies is that Baskett and others (2009) did not explicitly include the magnitude of the growth-mortality trade off in hosts associating with thermally tolerant symbionts. Therefore, Baskett and others (2009) effectively conclude that coral cover could be maintained for the next 100 years through symbiont adaptation in the absence of coral–macroalgal competition, and with a fast growing branching coral like Acropora with 50% relative abundance at the beginning of the simulation. This study considers a less resilient state of the ecosystem that lacks Acropora, and factors in the community-level interactions with macroalgae, which together place a greater importance on the cost of a reduction in coral growth rate associated with novel symbionts.

The approach followed here is conservative in that it identifies the maximum time available for the super-symbiont to become dominant to ameliorate the effect of climate change. It further implies that any deterioration of the initial conditions will require a faster invasion of the super-symbiont on the reef. It is important to note that these results are based on a model parameterization for a wave-exposed Caribbean reef. The macroalgal productivity of these reefs is considerably higher than leeward reefs, largely because of relatively high wind-generated turbulence in the water column that increases the flux of nutrients to algae. When grazing is diminished by fishing and diseases, the combination of relatively strong bottom-up forcing and weak top-down control results in a high cover of macroalgae. Indeed, a recent analysis of herbivore exclusion experiments found that macroalgal blooms established faster and with greater magnitude in the Caribbean than the Indo-Pacific, implying that bottom-up forcing is relatively high in this region, possibly because of a lack of iron limitation (Roff and Mumby 2012). Therefore, it would be inappropriate to extrapolate these results to areas where phase shifts are less likely such as offshore reefs in the Pacific.

Our results suggest that the potential benefits of a super-symbiont described in Ortiz and others (2013) would be considerably greater if climate change was mitigated directly by aggressive action to reduce greenhouse gas emissions. Furthermore, the introduction of super-symbionts may lead to a shift in coral community structure towards corals with higher competitive ability against macroalgae even if their intrinsic growth rate is relatively low.

Acknowledgments

We thank Maria Gomez-Cabrera for helping to edit the figures. This study was funded by an ARC Laureate Fellowship to PJM.

Supplementary material

10021_2013_9702_MOESM1_ESM.docx (636 kb)
Supplementary material 1 (DOCX 636 kb)

Copyright information

© Springer Science+Business Media New York 2013