2

Selecting Interventions for Decision Analysis

The committee identified 23 types of coral interventions in its first report (see Table 1.1), and even within these 23 options, there are further options and choices that can be made. As a consequence, coral reef managers are faced with the task of evaluating the benefits and risks of a growing number of interventions, separately and in combination. Several questions will help narrow the array of options. Foremost is whether the state of research has progressed sufficiently that these options are ready to be deployed. However, there is still value in assessing the potential risks, benefits, and utility of interventions that are not immediately “ready” in order to evaluate the value in pursuing a research program to improve their practical readiness. Additionally, a manager will want to consider which interventions are appropriate for a particular reef, community, and environment. Within a region, there may be specific areas better suited for testing interventions, distinct from areas where the interventions are likely to have a greater benefit or chance of success. Lastly, the relationships among interventions—such as shared biological or infrastructure resources—may increase stakeholder interest in investing in a set of interventions. The overarching goal of this chapter is to provide answers to the questions: “What can managers decide to do now and what can they expect to have available in the near future? What interventions are suited for the local ecological and management context?” The answers will help define the set of opportunities for consideration in a decision analysis used by local agencies and communities.

TIMEFRAME TO TECHNICAL READINESS

The coral interventions detailed in the committee’s first report (NASEM, 2019) represent a wide range of strategies that have very different levels of practical development. In this section, the committee first identifies interventions that managers could decide to evaluate for implementation now. Next identified are interventions that seem likely to be available at a local scale in approximately 2-5 years. Last are interventions that may need longer-term (5-20-year) investments, either because the technology to do them is not yet available or because coral growth for multiple generations is required. This assessment is largely based on the state of research described by the committee in their first report. It is important to emphasize that while an intervention may be technically ready to deploy, addressing its acceptability for meeting management goals and other social values is essential as is assessing its anticipated risks and benefits before deployment (discussed further in Chapters 3 and 4). Additionally, its current effectiveness may be more limited in time or space than is required to meet a management objective. Further research may be valuable even for this class of interventions to lessen risks or improve benefits (discussed further in Chapter 5). Research on developing the practicality of interventions that are nearly ready may help provide these tools in a timeframe that is useful for decision makers. Long-term investments in both laboratory-based research and field testing may be needed to make the last set of interventions practical and widely available.

The committee has organized a three-tiered approach to resilience intervention: immediate, short-term development, and long-term development interventions. Interventions in the “immediate” (I) category might help stabilize and improve local reefs in the face of current climate but may not be enough as climate continues to worsen. The interventions in the “short-term” (S) development category, once available, could strengthen resilience even further. Likewise, deployment of the short-term development interventions may help stabilize and improve reef populations while the “long-term” (L) development interventions are developed and come online (see Figure 2.1). Interventions across categories thus build on one another, and effectively may also buy time until greenhouse gas emissions can be reduced or removed.

Interventions Available Immediately (I)

Supportive breeding (I) can be done with many corals by collecting gametes during annual spawning (a broadcast release of gametes), followed by fertilization, larval rearing, and outplanting at the location of collection. For some species, larvae can be collected directly from

colonies on a regular basis. Constraints include limited windows for larval release or spawning, low survival of larval settlers, and the labor involved in large-scale husbandry efforts. Population size enhancement can be achieved with fragmentation and growth of colonies in nursery settings. These restoration efforts are currently the most often used on reefs. When colonies are chosen for breeding or fragmentation based on particular traits, such as heat resistance, then this is also considered managed selection (I). Testing corals for heat resistance and using them in nurseries is now available as a strategy (Morikawa and Palumbi, 2019). Scaling up these approaches requires increases in costs, labor, reef space, and donor colonies.

Assisted gene flow (I) is similarly available and differs from the above in that larvae or fragments available for outplanting are chosen for their high resistance to heat (or other traits) and then moved into new locations within their range where these genes or colonies do not occur or are rarely present. Finding and mapping heat-resistant colonies is currently possible, but requires marine research or aquaculture facilities for testing. Small, inexpensive, and portable tools to assess stress tolerance would expand development of this approach.

Gamete and larval capture and seeding (I) can be accomplished during periodic larval release and annual spawning cycles in most ocean aquaculture or research facilities (Chamberland et al., 2015, 2016, 2017). Its impact is limited by the number of days per year in which larvae or gametes can be collected, and the number of larvae that can be reared. Most corals are spawners, releasing eggs and sperm into the water column

over limited periods each year, ranging from a couple of nights during a single lunar cycle to several nights over multiple months. Brooding corals that release fully developed larvae may do so nearly daily, monthly, throughout the year, or only during several months. As such, the use of gamete and larval collections is both location and species dependent. The products of this approach are used in supportive breeding or assisted gene flow, described above.

Pre-exposure (I) to heated conditions has been demonstrated to increase heat resistance in numerous acclimation experiments in laboratory settings and appears practical for short periods at least. Pre-exposure to particular temperature or irradiance regimes has long been recognized as being an important determinant of stress tolerance in the field (e.g., Brown et al., 2002; McClanahan et al., 2005), and temperature regimes that pre-expose corals to a mild warming episode prior to severe heating have been shown to offer some protective value against bleaching on the Great Barrier Reef (Ainsworth et al., 2016). Depending on the mechanism by which increased resilience is achieved, the effects of pre-exposure may be temporary, declining in days or weeks if due to reversible changes (e.g., temporary changes in gene or protein expression), but may be more long-lived (e.g., Brown et al., 2015) if due to epigenetic, maternal, or microbiome effects (Putnam and Gates, 2015). Wide-scale sublethal reef heating has not yet been developed as a bleaching mitigation intervention strategy, although pre-exposure of corals to bleaching stressors prior to outplanting is being trialed as part of restoration efforts (Cabral et al., 2018; Winter et al., 2018).

Likewise, algal symbiont manipulation (I) of adult corals has been shown to increase heat resistance if Durusdinium symbionts replace Cladocopium and other genera, both in the laboratory (Cunning et al., 2015, 2018; Silverstein et al., 2015) and in the field (Berkelmans and van Oppen, 2006). However, such replacements, when they occur naturally on reefs following bleaching events, may be temporary, lasting only months to years (LaJeunesse et al., 2009), although some shifts may have occurred on reefs that are more long-lived (decades, Edmunds et al., 2019). Algal symbiont communities can also be manipulated at earlier life history stages by raising coral juveniles in different reef environments (e.g., Abrego et al., 2009). Most coral species produce gametes without symbionts, and the larvae or recruits can be introduced to specific symbionts in a laboratory or other controlled environment (e.g., Little et al., 2004; McIlroy and Coffroth, 2017; Williamson et al., 2018). However, the longevity of these directed changes may be relatively short-lived (weeks to months) if the environmental conditions in which the recruit is raised do not favor the new symbionts (Abrego et al., 2009; Coffroth et al., 2001).

Coral cryopreservation (I, S) is poised to help accelerate gamete and larval capture and seeding, selective breeding, and assisted gene flow (Hagedorn et al., 2018). Freezing gametes for future use would greatly expand opportunities for rearing and reduce the risks of coral transportation that limit use of managed relocation (see below). Sperm storage has been shown to be effective, and such gametes can fertilize eggs (Hagedorn et al., 2017). As a result, sperm storage is immediately available. Although coral larvae have been successfully cryopreserved (Daly et al., 2018), egg storage and egg revival for fertilization have yet to be demonstrated.

Interventions Potentially Available in a Short-Term (S) or Long-Term (L) Timeframe

Managed breeding—outcrossing between populations (S) can be done by bringing gametes together from corals from different populations adapted to different habitats. Outcrossing requires genetic differentiation across multiple loci between pairs of parents, and generally this demands medium- or long-distance transport. As a result, this intervention traditionally has been limited by the distances over which corals can be moved during spawning. Programs in laboratory-based spawning have demonstrated success, meaning that corals can first be moved long distances for this work and then brought into spawning condition (Craggs et al., 2017). Alternatively, cryopreservation of gametes can also facilitate this intervention over any distance. Success in these efforts is delayed by the need for multiyear grow-out of larvae to produce sexually reproductive adults and by the need for rapid mapping of coral genetic differentiation across multiple spatial scales for multiple species so that the degree of outcrossing, and the potential genetic effects, can be controlled.

Managed breeding—hybridization between species (S) can be done by cross-fertilizing different species. It does not require long-distance movement of the parent corals if the species are sympatric. Success is limited by cross-fertilization ability of different species (and potentially the fertility of the hybrids), for which there has only been demonstration in relatively few species (Chan et al., 2018; Willis et al., 1997, 2006). Readiness is also limited by the need for multiyear grow-out of hybrid larvae.

Supplements to corals (microbiome manipulation [probiotics], antibiotics, phage therapy, nutritional supplementation, and antioxidants) (S) have demonstrated effectiveness in laboratory or nursery/aquarium settings and in some cases, small-scale field trials (e.g., Atad et al., 2012; Hudson, 2000; Marty-Rivera et al., 2018; Rosado et al., 2018). However, limited knowledge about functional relationships between these interventions and coral health limits the ability to deploy them in a targeted, effective way. Nutritional supplements are in regular use in nurseries,

but delivery methods onto reef scales that would specifically target the coral, but not other taxa, are not developed. Application of antibiotics and phage therapy is also feasible, but limited by knowledge about the specificity of the antibiotics or bacteriophages to pathogenic bacteria. Similarly, limited knowledge of the functional role of particular bacteria and other components of the microbiome make targeted application difficult. Application of antioxidants has been trialed, but knowledge about its feasibility for mitigating reactive oxygen species production during bleaching is rudimentary.

Ocean acidification interventions show feasibility at small scales. Abiotic ocean acidification interventions (S) (the addition of a strong base to elevate pH, or stripping carbon dioxide from the water column) have a demonstrated ability to increase pH (e.g., Albright et al., 2016; Koweek et al., 2016; Rau et al., 2007; Riebesell et al., 2010). However, the approaches are logistically limited by the difficulty of deployment on a reef at scales large enough to impact pH and aragonite saturation. Seagrass meadows or macroalgal beds (S) located near coral reefs have been shown to raise local pH and the aragonite saturation state naturally (Manzello et al., 2012) and in flume experiments (Anthony et al., 2013). However, the benefit achieved from use of seagrasses and macroalgae will be limited by identifying the right water depth, water residence time, seagrass or macroalgae density, coral species, fate of fixed carbon, and other geographical, oceanographic, ecological, geomorphological, and meteorological attributes.

Cool water mixing (S) is conceptually simple but requires infrastructure that is difficult to install at even small reef scales. Reef cooling also faces the difficulty of cooled water being moved away from reefs and diluted by local currents.

Shading interventions, similar to cool water mixing, indirectly address resilience not by increasing heat resistance, but by reducing temperature and light stress that is associated with bleaching (although they could be used as a reef-scale technique during regional thermal stress to maintain sublethal stress levels that contribute to hardening). Marine shading (S) technologies have been proposed to reduce light by creating physical or chemical barriers at the water surface or in the local atmosphere. Water surface barriers are easy to imagine, but the simplest ones such as plastic sheeting could have disastrous consequences due to entanglement and pollution. Thin chemical barriers that are biodegradable and safe, yet also significantly reduce light, require research, development, and testing. Atmospheric shading (L), by using salt aerosols to brighten and increase cloud cover, for example, has been theorized based on natural phenomena and empirical evidence but not tested (Alterskjær et al., 2012; Latham et

al., 2013; Zhang et al., 2018). These interventions have also not yet been thoroughly evaluated for unintended consequences.

Genetic manipulation of corals and algae (L) has attracted a great deal of attention but is likely to be a practical intervention only over longer timeframes. Genetic manipulation of coral larvae has been achieved through use of CRISPR/Cas9 genetic constructs (Cleves et al., 2018), but adding exogenous DNA to symbiont cells has so far been unsuccessful. Nevertheless, rapid technology developments in DNA manipulation occurring across taxa may bring new tools to bear on these problems in the next few years, and dedicated research is likely to progress quickly. Likewise, targets for gene manipulation have not been identified. Assessing success of gene manipulation of corals would be required over multigenerational timescales through a series of growth periods (from manipulated larvae to sexual adults) each likely to be between 2 to 5 years. Genetic manipulation of single-celled symbionts would not require these long multigenerational steps. The combination of technology development, target identification, and grow-out periods adds up to a decade or more of time lag before being ready.

Managed relocation—assisted migration (L) and introduction to new areas (L) would be implemented in a way similar to assisted gene flow. However, increased distances of movement increase costs and decrease effectiveness in ways that are not yet well described (e.g., lower survival of coral individuals transported over great distances). As distances increase, risks from disease or invasive species introductions are probably also likely to increase. In this case, there are also significant barriers to evaluating risk related to invasive species and pathogen introduction because any experimentation in nature would necessitate risky procedures. One remedy is to move coral genes, not colonies, especially if gamete cryopreservation is more successful in the future. Another option might be to focus on translocation across depths rather than latitude to reduce distance moved, but closer distances are more likely to be within natural dispersal ranges such that intervention would be less relevant and impactful.

CONTEXT DEPENDENCY OF INTERVENTIONS

In addition to identifying the range of intervention options based on their technical readiness, the committee also recognized that the suitability of either testing or deploying a particular intervention would vary across geographic regions, given differences in the ecological settings and social systems. Here, the committee provides an expert judgment of the attributes that might dictate which interventions are available in a particular jurisdiction, or alternatively, where within the jurisdiction an

intervention might be tested or deployed. This evaluation is based on the ecological targets of an intervention and the ecological variables that may be at risk from an intervention (the details of which are described more fully in the committee’s first report). This evaluation describes intuitive expectations based on the best available science, and as such, represents hypotheses that could be evaluated in a model of the type developed in Chapter 4.

The details of the local ecological and social setting in which a particular intervention is either tested or deployed will determine its suitability and/or effectiveness. In some cases, these details could have different implications for interventions based on whether an intervention is being tested or deployed. For example, minimizing risks is a higher priority when testing, while maximizing benefits may be a higher priority when deploying at scale (assuming best practices to reduce risk are developed in the testing phase). Described here are some of the principal location-specific criteria that might be used to determine where (and in some cases, when) a particular intervention might be considered.

Biophysical and Ecological Context

Degree of reef degradation Most interventions will likely target degraded reefs, as indicated by reduced amounts of live coral cover or changed composition (e.g., species or intraspecific genetic diversity), because a high degree of degradation may be needed for an intervention to be warranted. Focusing on highly degraded reefs to test interventions also minimizes risks to more intact reef systems, unless both are part of a connected network of reefs. However, there is a tradeoff: interventions might be less successful on degraded reefs due to the presence of the local stressors that led to reef decline. Interventions on more intact reefs may give higher returns because prevention in combination with restoration increases the chance of producing beneficial outcomes (Possingham et al., 2015). Therefore, a focus on highly degraded reefs to test interventions may help minimize the risks of as-yet-untested interventions. But more intact reefs might be preferred for deployment at scale to maximize the likelihood of success and, therefore, benefits. For example, genetic interventions, particularly genetic manipulation of coral or symbionts, rely on a successfully reproducing population of corals in order to spread. Furthermore, reefs that are currently highly degraded due to historical stress, but which have been recently managed to reduce that stress (e.g., improved water quality, fisheries management, substrate availability) might represent both a place with lower risk and higher likelihood of success. Areas that contain both high and low levels of reef degradation allow for the opportunity to conduct a comparative approach for testing

interventions in both degraded and healthy contexts. Such interventions include managed breeding approaches that require comparing the fitness outcomes across populations.

Disease outbreak The majority of interventions are best tested or deployed in areas where disease is not active or has not recently affected the coral population for three reasons. First, the confounding effects of disease are likely to obscure results. Second, some interventions, such as cryopreservation, managed breeding, pre-exposure, or managed relocation, might exacerbate the severity or spread of the disease. Third, interventions that increase stress tolerance (e.g., pre-exposure, managed relocation, or symbiont or microbiome manipulation) might create tradeoffs with disease resistance (Jones and Berkelmans, 2010; Shore-Maggio et al., 2018). Exceptions to this are the disease interventions themselves (e.g., antibiotics and phage therapy), for which disease needs to be present to test effectiveness.

Bleaching history and future projections Most interventions that aim to increase heat tolerance (and thereby reduce the severity of future bleaching events) need to be tested in areas where bleaching has a high probability of occurring in the future to evaluate their efficacy. However, while high past bleaching history can be an indicator of future re-exposure to thermal stress, it might not mean that corals in those areas remain highly susceptible. Prior bleaching events can result in the removal of susceptible coral genotypes, shuffling of symbionts toward more resilient types, and acclimatization that might make interventions less relevant (Brown et al., 2000, 2002; Sully et al., 2019). Depending on the severity of past bleaching and the likely risk of future bleaching, these areas might be considered either low priorities for intervention if recent bleaching was severe and projections of future bleaching are low-to-average or high priorities for intervention if projected future bleaching is high (see Figure 2.2). If past bleaching was high and future bleaching is projected to be high such that reefs are both more degraded and less likely to persist into the future, these reefs might even be considered priority sites for intervention because the state of the resource and its future bleaching exposure might indicate that the risks of intervening are minimized. These conundrums suggest that a reliable and comparative way of testing a set of corals for heat tolerance would be an important part of the monitoring toolbox for most interventions.

A combination of historic observations of bleaching, future expectations of bleaching, and testing for heat tolerance could help define the value of local interventions to increase heat tolerance (see Figure 2.2). Exceptions to this might include interventions that involve sexual reproduction of healthy individuals (e.g., managed breeding, gamete and larval

capture and seeding). Such reproduction could be hindered if bleaching occurred before the intervention and larval survival might be impaired if bleaching occurred after settlement. Areas with high potential for future bleaching are also candidate areas for shading and cooling environmental interventions.

Water quality High water quality is generally preferred for most interventions under the expectation that conventional management approaches to maintaining water quality (e.g., reducing nutrients and sedimentation) will help maximize the health and survivorship of corals on which interventions are being implemented. Some interventions that may relate to overall coral health and function, such as manipulation of algal symbionts and the microbiome, may be appropriate to test across different sites that vary in their water quality to assess how these parameters can influence the success of manipulations. This would then inform the range of acceptable water quality values for deployments designed to maximize intervention success.

Herbivory As with high water quality, high degrees of herbivory (either natural, managed, or enhanced) might maximize the survival of corals on which interventions are implemented by reducing overgrowth and competition by macroalgae (Adam et al., 2015; Ladd et al., 2018). High herbivory may be especially important for interventions that involve the outplanting of recruits and small fragments (e.g., managed breeding, managed relocation, gamete and larval capture and seeding) because of increased coral susceptibility to macroalgal overgrowth at small sizes (Birrell et al., 2008; Box and Mumby, 2007; Martinez et al., 2012), as well

as reduced oxygen levels and pH at night close to the substratum due to algal photorespiration. In contrast, low herbivory might be preferred for testing use of nutritional supplementation to evaluate the impact that these interventions have on surrounding nutrient flow and to evaluate the risk of their stimulation of algal growth.

Recruitment capacity/substrate availability Reproductive interventions such as managed breeding and gamete and larval capture and seeding would typically favor testing and deployment at sites with high substrate availability and quality (such as the presence of crustose coralline algae), which affect recruitment capacity. Water quality is also a key parameter because it affects bacteria known to induce larval recruitment in a number of species. Overall, coral recruit settlement and survival depend on the interactive effect of an array of factors including nutrient levels, algal density and composition, and herbivore density and composition (Birrell et al., 2008; Jouffray et al., 2015; Smith et al., 2010).

Degree of connectivity At the testing stage, isolated sites (low connectivity) are preferred for most interventions to reduce widespread unintended consequences. For example, low-connectivity areas reduce the spread of nonnative corals, pests, and pathogens, which are all risks associated with managed relocation approaches. Low connectivity also reduces the risk of unintended gene flow. Once testing has led to the development of best practices to reduce risks, then reefs that have high connectivity might be targeted for deployment to maximize the likelihood that benefits of the intervention spread, especially for interventions targeted at genetic and community composition (e.g., managed selection, managed breeding, gamete and larval capture and seeding, managed relocation). However, low-connectivity reefs may benefit most from managed relocation (Hewitt et al., 2011).

Spatial extent of reef At the testing stage, replicate sets of small patch reefs (with appropriate replicate control patch reefs) might be a good strategy for leaving a smaller intervention footprint and reducing risks for many interventions. However, once testing has led to the development of best practices for reducing risks, then interventions might be deployed at reefs with a larger spatial extent to maximize benefits. A large reef may be optimal for implementing shading and cooling interventions, where a patchwork design for implementation would be possible. Because environmental interventions have the risk of halting natural adaptation and acclimatization processes, leaving some areas unaffected could allow for reproduction between shaded and unshaded (with a higher likelihood of developing thermal tolerance) corals.

Potential for cold shock At the testing phase, many interventions designed to increase thermal tolerance, such as managed relocation, managed breeding, or pre-exposure, might choose sites based on likelihood of cold shock. If there is the potential for cold shock in individual locations within a region, including such sites in testing will help assess whether an intervention designed to increase high-temperature tolerance comes with a decrease in low-temperature tolerance, making them vulnerable to cold shock. If this is the case, then sites with low potential for cold shock might be prioritized when deploying at scale to increase the likelihood of intervention success.

Spatial variability in temperature Several interventions, such as managed breeding or assisted gene flow, require significant thermal variability among sites at both the testing and deployment phases. This is because the standing genetic variation that is the prerequisite for these interventions likely arises in response to long-term differences in thermal regime among sites. Local-scale variability is also an indicator of the presence of genotypes that may be candidates for managed selection.

Select attributes of target species Different species are likely to have differential success with certain interventions. Specific interventions that may be best used for particular types of species are described here:

• Managed selection: Requires predictable gamete or larval release, with hardy larvae.
• Supportive breeding and outcrossing between populations (managed breeding): Require key reef-building species, especially those with rapid growth traits, to increase feasibility of breeding in captivity.
• Hybridization between species: Requires an ability for two species to form hybrids, potentially fertile hybrids, depending on management goals. In most cases, this ability is rare.
• Gametes and larval capture and seeding: Best with larvae from species that are fast growing or capable of fusing.
• Coral genetic manipulation: Best for species with short generation times to decrease the time needed for testing and proliferation into the population. History of resilience to climate change provides appropriate genetic basis.
• Symbiont genetic manipulation: Requires generalist symbiont known to be capable of forming associations with diverse coral species and a known history of conferring thermal tolerance in local reef environment.

• Algal symbiont manipulation: Requires coral species with the capacity to maintain stable associations with diverse symbionts.
• Microbiome manipulation: Requires coral species with the capacity to maintain stable associations with diverse symbionts.

Management Context

Infrastructure and resources Generally, resource requirements are high for testing and deploying most interventions at large scales. However, in some cases this infrastructure may already exist, such as restoration programs that currently propagate corals in nurseries and outplant corals in large numbers. These existing resources provide an opportunity to rapidly test some interventions, and potentially scale these resources up for deployment. Some of the general requirements for the different intervention types are detailed below.

• Genetic and reproductive interventions typically require an existing sexual restoration program, including the capability to collect gametes or larvae, rear them in ex situ facilities, and settle them on substrates that can be deployed in the field. This typically requires a combination of in-water boat and dive support, field laboratory, and settlement facilities (either in the laboratory or in nearshore in-water systems). Assisted gene flow and managed selection require test facilities that can measure heat tolerance of individual colonies; map them; and return to them for sampling, propagation, or gamete collection. Different clones show different responses to heat, grow-out location, or other conditions such as disease (Morikawa and Palumbi, 2019; Muller et al., 2018), so these interventions require the ability to label and follow clones over a period of years. Some interventions (e.g., genetic manipulation) require extensive technical expertise to develop methods and protocols that are not currently available, in addition to suitable molecular laboratory facilities that are typically available only at large research facilities. For some of these interventions, quarantine facilities for genetically modified organisms will also be needed to prevent their escape to the wild.
• Physiological interventions commonly need existing restoration and monitoring activities (either sexual or asexual) when interventions are performed in a laboratory or nursery setting and then outplanted. They often also require some degree of technical expertise or specialized equipment (e.g., molecular assays or a physiological laboratory) to develop the intervention and/or test its effectiveness. Quarantine facilities may also

• be needed, especially for testing disease interventions such as antibiotics, phage therapy, or probiotics (a form of microbiome manipulation).
• Population and community interventions will usually require an existing asexual propagation program, including coral fragment collection, tolerance testing (thermal and disease susceptibility), mapping, labeling, construction of nurseries, and an outplanting and monitoring program (i.e., boat and dive support). Quarantine facilities for nonnative species or nonlocal genotypes (i.e., species or genotypes not normally found in the site or region where the restoration program is located) will also typically be required to reduce the likelihood that they carry nonlocal pests and pathogens.
• Environmental interventions will typically require significant infrastructure that does not already exist at the site of deployment, such as for atmospheric shading or cool water mixing. However, infrastructural needs for some interventions could be relatively modest, such as mitigation of ocean acidification locally by restoring corals adjacent to seagrass beds (which typically requires infrastructure similar to those for population and community interventions). Reef shading using microfilms (which could deploy a small amount of microfilm over a relatively large area) would require less than most environmental interventions, but nevertheless needs boat deployment, cleanup, and monitoring.

Size of management jurisdiction The size of the management jurisdiction may need to be relatively large for some environmental interventions, such as atmospheric shading, which are likely to impact large areas rather than target particular reefs. In such cases, impact across national borders might be an important management issue. For other interventions, especially physiological interventions, the size of the jurisdiction is less critical because the intervention is intended to operate over much smaller scales. Some interventions, such as managed breeding or relocation, may not necessarily require a large jurisdiction, but may instead depend more on cooperation among jurisdictions, including agreements to provide restoration or breeding stock that are more heat tolerant. For the most part, the larger a management jurisdiction, the more intervention options there will be. Large areas might consider zoning certain areas for the testing and deployment of interventions as a way of managing risk. In contrast, smaller jurisdictions may be more responsive to adaptation needs because they have fewer stakeholders and a simpler decision-making structure.

Regional management consensus As mentioned above, some site-based interventions (e.g., antibiotic treatment of a coral disease outbreak) may require little consensus among adjacent management jurisdictions in order to proceed, whereas others, such as atmospheric shading, may require a broader management consensus before they can be attempted. Achieving consensus is likely to become increasingly challenging as more jurisdictions are involved, making some interventions much more difficult to test and/or deploy. Determining which jurisdictions have standing in a decision-making process may also be challenging. For example, breeding corals sourced from two different locations to produce offspring that are more thermally tolerant may appear to be a question involving only the jurisdictions that provide the parents, and the jurisdiction that outplants the offspring to the wild. However, other management jurisdictions with connectivity to the outplanting site will also necessarily need to be considered in the decision-making process. Consequently, a common set of guidelines will need to be developed to accommodate these different interests. In general, the higher the level of success of an intervention, and the more broadly it is deployed, the more consensus among regions is required. Multijurisdictional consensus is not included explicitly in the decision framework but how these might be reconciled via deliberation when setting multiple objectives and analyzing tradeoffs is discussed.

Societal acceptability An important consideration regarding whether an intervention can be tested or implemented is the degree of perceived social license granted to a management group to pursue an intervention in the first place. Even if best practices have been developed and can be followed, controversial interventions may not be implemented until the intervention in question becomes acceptable to stakeholders. Acceptability will most likely be related to the need for interventions, their likely benefits and risks (real and perceived), and steps being taken to minimize risks and maximize benefits whenever interventions are attempted and the efforts made to communicate this to stakeholders. The role of stakeholder engagement in identifying acceptable interventions is described in Chapter 3.

INTERDEPENDENCE AMONG INTERVENTIONS

It is clear that some interventions are highly linked, and that their advantages are not independent. Managers may be in a position of examining multiple interventions (as well as conventional strategies) in combination, and there are possible efficiencies that may arise due to interdependence across interventions. Interdependence can arise from at least three causes: (1) two interventions may use the same infrastructure or raw

material; (2) using one intervention might make a second one more or less likely to succeed, or make it more or less costly; and (3) two interventions may benefit from the same research and development programs. The goal in the following is to point out some of the strongest interdependent relationships among interventions. The committee did not exhaustively analyze all possible combinations of interventions from its first report. Instead, through expert judgement the committee identifies cases with clear relationships among interventions in infrastructure, raw material, chance of dual success, and joint development needs.

Related infrastructure Many interventions need dedicated workspace in marine laboratories with good running seawater systems so that parent, juvenile, or larval corals can be grown. Managed selection, managed breeding (all three kinds), gamete and larval capture and seeding, algal symbiont manipulation, and microbiome manipulation all might be based in such a laboratory setting. As a result, investing in marine laboratory resources would benefit all of these approaches. However, limited seawater space at marine laboratories might lead to competition among proponents of different interventions for laboratory space. As a result, strong deployment of one of these interventions would provide infrastructure that could facilitate the others, but competition for laboratory space might create delays. In situ larval cultivation in open water can allow the needed scale to be achieved while reducing such competition (Heyward et al., 2002). Likewise, development of genetic tools to alter coral genomes might also generate knowledge, staff, facilities, or bioinformatics tools that would benefit genetic manipulation of symbionts.

Raw material Some interventions require the raw material of certain coral colonies, larvae, or other biological resources. Some of these similarities could enhance the success of similar interventions. For example, managed breeding (supportive breeding and outcrossing between populations), managed selection, and assisted gene flow all depend on the detection and certification of particular coral colonies as either stress resistant or sensitive. As a result, programs and protocols that find and test coral colonies can provide a boost to all of these interventions. Whether these colonies are a limiting resource on native reef habitats or are abundant enough to fuel all of the interventions that need them is a question for future mapping studies. “Corals of opportunity” from permitted activities such as dredging, especially from highly stressed areas such as harbors, can contribute to the supply and provide valuable genotypes.

Likewise, interventions that require larvae will jointly benefit from research and facilities that induce coral spawning beyond their often-limited window. Facilities that can deliver larvae weekly, for example,

could advance research in outcrossing, hybridization, genetic manipulation, pre-exposure, and algal symbiont manipulation. Access to a wide array of known symbiont cultures could also advance pre-exposure and algal symbiont manipulation.

Chance of dual success There are circumstances in which two interventions together could provide strong benefits that are not provided by one alone. In theory, any of the interventions reviewed by the committee could act this way. For example, increasing coral health with nutritional supplementation might make other interventions more successful by decreasing the amount of heat resistance needed. Likewise, deploying environmental interventions that reduce the degree of coral exposure to thermal stress may improve the chances that a thermal tolerance intervention will succeed. However, antagonistic relationships might also occur. For example, the environmental interventions to reduce thermal stress might instead reduce the success of genes or species introduced or promoted in interventions focused on increasing thermal stress tolerance. A key question is whether any of these interventions together have synergistic or antagonistic effects, for which the impact of the two together is greater or less than the sum of the impacts of each alone (an antagonistic interaction might still lead to a net positive outcome but it is less than expected from combining independent expectations). A possible example might be that enhanced herbivory within marine protected areas (MPAs) may promote larval settlement of local progeny of transplanted, heat-resistant colonies. In such a case, the dual interventions of assisted gene flow and MPA designation may promote resilience better than the sum of either intervention alone.

Capacity building—social as well as infrastructure The above examples show how infrastructure development can benefit a number of different interventions. Just as important, though, may be social systems at the state or local levels that promote the use of interventions in real-world settings. Simple systems to test corals for heat resistance, growing them in nonlaboratory settings, creating nurseries that have a variety of species and strains within species, and transplanting them out onto reef surfaces would have a strong impact on the ability of well-researched interventions to be deployed to reefs across the world. Development of the ability of nonscientists to contribute effectively to testing and deploying interventions is a key area that would require a focus on education, citizen science, and the kind of help that might best resemble the knowledge transfer systems on which local application depends. The training of more culturally connected local scientists from coral reef jurisdictions is also an important undertaking for local capacity development.

Conclusion: Multiple, interrelated interventions provide a toolbox for increasing coral reef persistence and resilience. This set of options can be tested and deployed based on community goals, ecological objectives for reef management, and the benefits and risks across multijurisdictional or even multinational boundaries. These efforts are likely to evolve over time as interventions become more feasible and as new interventions are developed.