A roadmap to integrating resilience into the practice of coral reef restoration

Abstract Recent warm temperatures driven by climate change have caused mass coral bleaching and mortality across the world, prompting managers, policymakers, and conservation practitioners to embrace restoration as a strategy to sustain coral reefs. Despite a proliferation of new coral reef restoration efforts globally and increasing scientific recognition and research on interventions aimed at supporting reef resilience to climate impacts, few restoration programs are currently incorporating climate change and resilience in project design. As climate change will continue to degrade coral reefs for decades to come, guidance is needed to support managers and restoration practitioners to conduct restoration that promotes resilience through enhanced coral reef recovery, resistance, and adaptation. Here, we address this critical implementation gap by providing recommendations that integrate resilience principles into restoration design and practice, including for project planning and design, coral selection, site selection, and broader ecosystem context. We also discuss future opportunities to improve restoration methods to support enhanced outcomes for coral reefs in response to climate change. As coral reefs are one of the most vulnerable ecosystems to climate change, interventions that enhance reef resilience will help to ensure restoration efforts have a greater chance of success in a warming world. They are also more likely to provide essential contributions to global targets to protect natural biodiversity and the human communities that rely on reefs.


| INTRODUC TI ON
The future of coral reefs is dependent on the rapid reduction of global greenhouse gas emissions and actions that enhance reef resilience to climate change . Across the globe, coral reefs are degrading due to human-derived local threats (e.g., changes in land and sea use, pollution, overfishing) and anthropogenic climate change such as ocean warming and acidification (Cheal et al., 2017;Hughes et al., 2017;Shantz et al., 2020;Wear & Thurber, 2015). Recently, severe thermal stress events have caused over 70% of the world's reefs to suffer consecutive or prolonged bleaching events resulting in widespread losses of living corals (Eakin et al., 2019). For example, 14% of the world's coral reefs were lost in the decade from 2009 to 2018, due in part to successive and severe bleaching events from 2014 to 2017 that caused up to 95% coral mortality in some areas in the eastern Pacific (Brainard et al., 2018;Souter et al., 2021;Vargas-Angel et al., 2019).
Unless urgent action is taken to keep global mean temperatures from increasing beyond 1-1.5°C, most of the world's reefs are predicted to experience frequent bleaching, threatening the future of coral reefs and the human communities that depend on them (IPCC, 2022). In response, coral reef managers globally are increasingly turning to restoration to slow coral loss, rescue endangered species, and accelerate reef recovery processes .
Ecological restoration is generally defined as the process of "assisting the recovery of an ecosystem that has been degraded, damaged, or destroyed" (Society for Ecological Restoration, 2004).
However, this definition, along with current principles of ecological restoration, implies that the causes of ecosystem degradation and loss can be removed (Gann et al., 2019). While in local-scale contexts this may be true for coral reefs (e.g., removal of blast fishing, herbivore overfishing, or wastewater pollution), global-scale climate processes will likely continue to pose a significant threat to reefs for decades even if current targets for greenhouse gas emissions are met (IPCC, 2021). Thus, many scientists and governments now see restoration as a necessary management intervention to maintain coral reef ecosystem processes, functions, and services through the next few decades of climate change Hein et al., 2021;Kleypas et al., 2021;Knowlton et al., 2021;Vardi et al., 2020).
Restoration also has been identified as a key component in resilience-based management for coral reefs (Anthony et al., 2015;Knowlton et al., 2021;Mcleod et al., 2019). Resilience-based management (RBM) focuses on prioritizing and implementing management actions to enhance reef resilience using knowledge of current and future threats (Mcleod et al., 2019). Underlying RBM is the theory of resilience, defined as the ability of a system to maintain key functions and processes in the face of stress by resisting, recovering, and/or adapting to change (Folke et al., 2010). More recently, resilience has been expanded to describe coupled social-ecological systems that can persist and transform to change (Keck & Sakdapolrak, 2013), where social resilience includes the ability of individuals, organizations, or communities to tolerate, absorb, and adapt to disturbances linked to changing environmental conditions and losses in resources (Keck & Sakdapolrak, 2013). Within the context of coral reefs, resilience refers to reef ecosystems that are less likely to be driven into persistent depauperate (e.g., algal-dominated) states through: (1) resistance, where negative responses of corals to disturbances are reduced, limiting ecosystem change (e.g., less bleaching or less coral cover loss during warm temperature events); (2) recovery, where reef ecosystems more readily return to a predisturbance state (e.g., through rapid coral growth and coral recruitment); and (3) adaptation, where reef ecosystems are altered in response to changing conditions but continue to function and provide ecosystem services (e.g., due to changes in the dominance of coral species or taxa over time).
Research into emerging restoration techniques also increasingly focuses on improving coral or reef ecosystem resilience Van Oppen et al., 2015. Both the US National Academy of Sciences and Medicine (NASEM) and Australia's Reef Restoration and Adaptation Program (RRAP) conducted recent largescale reviews to identify current and future interventions with the potential to promote resilience and assess their potential feasibility, scale, and risks NASEM, 2019). Most recently, Suggett and van Oppen (2022) illustrate how these novel approaches (e.g., probiotics, selective breeding, assisted evolution, bio-banking) can be used in the asexual-sexual coral life cycle to improve restoration success. Meanwhile, other recent publications have provided broadscale guidance for coral reef restoration (e.g., Hein et al., 2021;Quigley et al., 2022;Shaver et al., 2020), which include recommendations aligning with resilience theory (e.g., maximizing biodiversity and promoting connectivity: Nyström et al., 2008), but do not directly relate restoration practice to resilience.
Despite clear recognition in the scientific literature for resilience and climate-focused restoration techniques, there remains a critical gap in implementation. For example, in a global review of over 350 reef restoration projects up to 2018, only five projects included the word "climate" in the project description or goals . While resilience calls for increasing diversity (i.e., species, habitat) to spread the risk of loss from a disturbance event (McLeod et al., 2012;Mcleod et al., 2019;Nyström et al., 2008), nearly a third (28%) of projects in this review focused on just one coral species, the majority of which (59%) were branching corals that are generally less resilient to climate change-related bleaching Loya et al., 2001;van Woesik et al., 2011). Coral reef restoration efforts are also often led by local community-based organizations or management agencies that may not have scientists on staff or have access to scientific publications. Thus, for most practitioners, it is likely not clear how restoration should be conducted to promote reef resilience, and indeed no current resources exist that synthesize the science to describe approaches that are currently available for resilience-based coral reef restoration design.
Here, we address this implementation gap by providing guidance for how coral restoration practitioners, managers, and communities can incorporate resilience principles and climate considerations into coral reef restoration practice. We organize our guidance into four categories: (1) project planning and design, (2) coral selection, (3) site selection, and (4) broader ecosystem context (Table 1, Figure 1). As scientists warn that coral reefs TA B L E 1 Recommendations for incorporating resilience principles and considerations into the design and implementation of coral reef restoration. "Operational status" refers to the ability of practitioners to implement the recommendation in restoration programs at this current time (scale: 1 = operational with many challenges; 2 = operational with some challenges; 3 = operational with few challenges), determined by averaging the ratings of coral reef experts (n = 9). "Implementation needs or dependencies" includes any data, information, or processes that are to be likely required by restoration practitioners to implement the recommendation may be the first ecosystem to be lost to climate change , we present these recommendations with the goal of supporting and catalyzing the coral reef restoration community to shift toward more climate-smart and resilience-focused coral reef restoration.

| Project planning and design
An important principle in ecological restoration includes the consideration of natural variation and anticipated future environmental change when identifying restoration targets (Gann et al., 2019).
Despite the systematic incorporation of climate change impacts into marine spatial planning (Beyer et al., 2018;McLeod et al., 2012), marine reserve design (Mumby et al., 2011), and watershed management (Gibbs et al., 2021), only recently has guidance for coral reef restoration included climate change data in project planning and design . As global warming will continue for decades regardless of near-term reductions in greenhouse gas emissions, current coral reef restoration projects must be designed for predicted future climate change impacts, including how climate change could affect restored coral species, methods used (e.g., storm impacts on artificial reefs), and the location of efforts. The use of climate change adaptation tools, developed for designing other reef management strategies (West et al., 2018), can also be used for coral reef restoration planning (e.g., Shaver et al., 2020). Other tools that examine the role of climate change on local social and ecological conditions or resilience at local sites (i.e., climate vulnerability assessments, reef resilience assessments) also can help to ensure that climate change considerations are embedded in early project planning and design ( Table 1).
Considerations of social resilience for reef-dependent communities should also be included in restoration planning and design, such as how restoration programs can provide increased food security (e.g., improved fisheries), infrastructure security to support reef resilience (e.g., new interventions, predictive coral traits for resilience) will be critical to informing these efforts.

| Coral selection
One of the most common approaches in coral reef restoration utilizes underwater nurseries to grow branching corals, such as the Caribbean staghorn coral Acropora cervicornis (Young et al., 2012).
This single-species approach to growing, propagating, and outplanting corals stemmed from work in Florida focused on repopulating A. cervicornis, a once dominant reef builder that is now critically endangered throughout the Caribbean (Aronson et al., 2008).
Branching coral species like acroporids are commonly used in restoration because they can be easily fragmented and grow rapidly, allowing practitioners to experiment with coral propagation and nursery methods that are now foundational approaches to coral reef restoration. While there may be instances where a single-species approach is appropriate based on a specific restoration goal (e.g., planting branching corals such as Acropora palmata on the reef crest to improve coastal protection services), in general scientists are raising the alarm that coral reef restoration practices must move from a focus on single species and coral outplanting to ecosystem-wide approaches to ensure reef survival to climate change Vardi et al., 2021). Indeed, restoration programs are increasingly incorporating multiple coral species and growth forms, though most efforts still center around coral outplanting .
This shift to incorporate multiple coral species in reef restoration is essential as increased diversity and functional redundancy are core components of ecosystem resilience (Biggs et al., 2012;Elmqvist et al., 2003;Mcleod et al., 2019;Nyström et al., 2008). Specifically, diversity promotes a varied response to disturbance, potentially conferring increased resistance (e.g., less bleaching for some species or genotypes) and recovery (e.g., faster growth rates of some species) of the reef system to climate change impacts. Functional redundancy, where different species provide similar ecological functions (i.e., multiple branching coral species that provide habitat complexity to fish and invertebrates), allows the ecosystem to recover, adapt, and continue functioning after a disturbance even if one species is lost.
Practitioners should seek to incorporate diversity into their restoration programs through using different coral species and genotypes representing a variety of growth forms (and thus, ecological functions that can be used in the field will be critical to support practitioners in these efforts. Importantly, there are possible trade-offs between resilience traits (e.g., heat tolerance) and growth in corals (Cornwell et al., 2021), although some traits appear to be independent (e.g., heat stress and disease resistance: Muller et al., 2018). Resistance to heat stress and ocean acidification, for instance, have been positively associated for endangered A. cervicornis (Muller et al., 2021).
Therefore, ensuring that a wide diversity of coral genotypes, species, and growth forms are used in restoration efforts is likely the best course of action until potential trade-offs can be identified through additional research. Methods to enhance genetic variation will be needed in combination with outplanting diverse corals, including the integration of larvae from as many parental donors as possible or the use of as many heat-resistant corals as possible in nurseries (Cornwell et al., 2021). Ultimately, practitioners should monitor different coral species, genotypes, growth forms, and sizes before and after disturbance events (both in the short and longterm) to determine the best coral assemblage to use for their specific restoration goals and context.

| Site selection
Site selection for restoration is another key area where resilience components should be factored into restoration design. For instance, models show that prioritizing habitat diversity can protect heat-resistant coral populations and promote coral adaptation (Walsworth et al., 2019). Practitioners should seek to conduct restoration in sites that span a variety of reef types (i.e., fringing, barrier, and patch reefs) and conditions, including differences in depths, oceanographic features, and thermal regimes, with replication across site types whenever possible (Nyström et al., 2008;van Nes & Scheffer, 2005;Walsworth et al., 2019) (Table 1). Sites with high diversity and functional redundancy of herbivores (which reduce macroalgae and/or promote substrate conditioning for coral larval settlement) could also be used as site selection criteria (Burkepile & Hay, 2008;Elmqvist et al., 2003) to support increased coral recovery by keeping macroalgal cover in check.
To identify sites that have the highest potential for resilience, practitioners should work with marine managers or scientists to conduct resilience assessments to identify and prioritize locally resilient reefs for restoration outplanting . Resilience assessments, for example, have been used since 2007 by reef managers and scientists in every coral reef region to identify reefs with a higher potential to survive future climate change and prioritize them for management actions . Yet, in a review of how resilience assessments have been used to inform reef management actions to date, only one project used resilience assessment results to identify and select sites for restoration .
Resilience assessments provide critical information on underlying factors leading to higher or lower resilience in different sites (e.g., oceanographic features, water quality conditions, herbivory, and recruitment rates), and therefore can be used to enhance restoration outcomes by identifying resilient sites and informing management activities that should be conducted prior to restoration (Table 1).
To mitigate future risks to restoration brought about by changing environmental conditions, local and global knowledge of predicted climate impacts at potential restoration sites should also be incorporated. Ideally, climate change refuges that are the least at risk from future climate change would be identified and prioritized for restoration (see Chollett et al., 2022 and 50 Reefs, 50reefs.org). This could include areas that (1) are reliably cooled, (2) regularly experience high thermal variability or extreme conditions, (3) do not experience regular intense storm activity, or (4) are projected to be less impacted by future warming or acidification (Fine et al., 2013;McLeod et al., 2009McLeod et al., , 2012Oliver & Palumbi, 2011;Randall et al., 2020) ( Table 1). Recent research on consecutive bleaching events on the Great Barrier Reef shows there is consistency in thermal regimes of reefs, suggesting the locations of refugia and hotspots can be robust and predictable (Cheung et al., 2021). Using data on thermal stress patterns (i.e., historical and projected sea surface temperatures) can help practitioners select restoration sites with a greater likelihood of success in a changing climate, as has been used in marine reserve design (Mumby et al., 2011) and most recently for coral reef restoration (Chollett et al., 2022). For practitioners in the Caribbean and Florida, this information is available for coral reefs down to the 1-km scale through The Nature Conservancy's Caribbean Coral Climate Refugia Data Explorer (Coral Refug ia.tnc.org). One potential low-cost approach to identifying resilient reef sites includes rapid and standardized testing of coral thermal tolerance using portable devices (Oliver & Palumbi, 2011;Thomas et al., 2018). Voolstra et al. (2020) reengineered these as the Coral Bleaching Automated Stress System (CBASS), which tests the responses of small coral samples to acute thermal stress in the field. This system could also identify naturally heat-resistant corals for use as donor colonies for restoration or direct transplantation. However, comparisons between ecologically relevant scenarios and portable stress-test systems such as CBASS will also require further exploration.
Identifying larval connectivity patterns at potential restoration sites is also an important consideration for designing restoration to promote resilience ( Table 1). The value of locally protected, thermally resilient reefs is enhanced when these corals act as sources of larvae to nearby areas (Hock et al., 2017;Mumby, Mason, & Hock, 2021;Mumby, Steneck, et al., 2021), spreading heat resilience traits. For instance, sites identified as thermal refugia may be capable of providing coral larvae to 58% of the Great Barrier Reef, highlighting the importance of restoring such sites to provide system-wide reef resilience (Cheung et al., 2021). Restoration projects should be lo- cations (see Chollett et al., 2022), as have been used to design marine protected area networks (Magris et al., 2016;Schill et al., 2015). For instance, sink reefs (e.g., that receive a large portion of larvae from other areas) in theory may be good candidates for donor coral collection because coral diversity may be higher in these sites. In contrast, source reefs (e.g., that export a large portion of larvae to other areas) may be good sites for outplanting because restored colonies in these areas could support higher recovery to nearby connected reefs. While methods do currently exist for use in restoration, more research and investment are needed to develop and make available predictive larval connectivity studies at local scales for use by restoration practitioner groups (e.g., Frys et al., 2020).

| Broader ecosystem context
Coral reef restoration projects aimed at supporting resilience and climate adaptation of corals cannot be fully realized without considering the broader context within which reef ecosystems function, including connections between adjacent marine habitats and human populations. Ecological connections between ecosystems across the land and seascape are well known to affect reef resilience (e.g., Guannel et al., 2016;Mumby & Hastings, 2008). Recent research highlights how restoration outcomes are improved when multiple degraded and ecologically connected ecosystems are restored together (Milbrandt et al., 2015;van de Koppel et al., 2015) ( Table 1).
For example, intact mangroves and seagrasses may benefit nearby coral reef restoration efforts by improving water quality and alkalinity (e.g., Guannel et al., 2016;Manzello et al., 2012). Restoring mangroves may help to mitigate the effects of lost coral reef structural complexity on reef fish biomass and fisheries productivity, offsetting some of the impacts of climate change on neighboring reefs in terms of fish biodiversity (e.g., Rogers & Mumby, 2019). Likewise, restoring terrestrial forests and riparian vegetation could reduce sediment flow into adjacent coral reefs, supporting improved survivorship and fitness of coral outplants (e.g., Carlson et al., 2019).
An intact ecosystem that has redundancy and feedback systems in place is more likely to show increased resilience compared with single-species monocultures (Downing & Leibold, 2010;Nyström et al., 2008;Vogel et al., 2012). Thus, as coral reef restoration projects mature and increase in scale, methods used should transition from a focus on single species and coral outplanting to approaches that improve ecological processes and functioning Vardi et al., 2021). One way may be to incorporate non-coral species, particularly those known to facilitate coral recovery, recruitment, and health (Ladd et al., 2018, Shaver & Silliman, 2017 (Table 1). For example, herbivores that graze algae and provide suitable substrate for coral settlement could potentially enhance the success of restoration projects (Ceccarelli et al., 2018;Spadaro & Butler IV, 2021;Williams, 2022). In Hawaii, the cultivation and transplantation of the urchin Tripneustes gratilla, in combination with manual removal methods, has been used by the Hawaii Division of Aquatic Resources (USA) to control invasive macroalgae and rehabilitate reefs (Conklin & Smith, 2005;Neilson et al., 2018). Herbivorous snails, used in coculture with ex situ sexually propagated coral recruits, were found to increase coral survival 23-fold (Neil et al., 2021). In another example, encrusting sponges and coralline algae were investigated as natural mechanisms to secure coral rubble and promote recruitment on damaged reefs (Biggs, 2013). Conversely, practitioners may need to incorporate restoration interventions or designs that mitigate the impacts of non-coral species that reduce coral recovery potential, such as corallivores like Crown-of-Thorns Starfish (COTS) or Drupella snails. However, more research on processes and species that promote coral health and resilience, as well as interventions and techniques for restoring non-coral species, is required for practitioners to utilize interspecific relationships to promote coral reefs through restoration.
Landscape-level connections to local human populations are also critical considerations for coral reef restoration designs to support resilience. For instance, anthropogenic stressors, particularly nutrients and other pollution from terrestrial sources, are known to reduce reef resilience (Carilli et al., 2009;Donovan et al., 2020;Vega Thurber et al., 2014), and management actions to mitigate such local stressors can improve resilience Shaver et al., 2018). To ensure local threats are mitigated in existing or potential restoration areas, restoration should be embedded within a broader management framework and deployed in areas where local threats can be controlled (Mcleod et al., 2019) (Table 1).
This could include marine protected areas, other effective areabased conservation measures, or coastal zone management areas where the impacts of overfishing, tourism, coastal development, or marine vessels are reduced. Efforts to manage or restore watersheds to reduce nutrient pollution, sedimentation, and sewage should be undertaken alongside, but ideally before, restoration begins, to improve coral outplant success Shaver et al., 2020).
These efforts should also ensure consideration of the socialecological context in reef management to strengthen social adaptive capacity, resilience, and thus overall compliance with management and restoration actions (McLeod et al., 2012).

| FUTURE DIREC TIONS
Building resilience into coral reef restoration will require new partnerships and the testing and integration of novel biological, ecological, social, and oceanographic methods that specifically target and enhance the mechanisms of coral reef recovery, resistance, and adaptation to local and global disturbances. For example, mechanisms that improve coral recruitment (i.e., survivorship rates of recruitment) could be an important research frontier to enhance coral population recovery after disturbance. Examples include innovations in new materials (e.g., hydrogels) to protect corals in vulnerable early life stages and increase survivorship of coral recruits , the incorporation of crustose coralline algae and biofilms in restoration projects (Heyward & Negri, 1999), or the use of acoustic playback of a healthy reef to enhance coral settlement in degraded sites (Gordon et al., 2019;Lillis et al., 2016).

| CONCLUSION
As the UN Decade on Ecosystem Restoration (2021-2030) begins and nations seek to meet ambitious conservation and biodiversity targets, it is necessary to conduct restoration as part of broader resiliencebased management of coral reefs and incorporate resilience principles and climate change adaptation into restoration practice. The recommendations presented here provide guidance to help the coral reef restoration community enhance reef resilience to climate change and other reef threats (e.g., disease) ( Figure 1). Recommendations are in line with key principles for the practice of ecological restoration that guide all practitioners involved in restoring degraded habitats (i.e., Gann et al., 2019), suggesting this guidance could be applied to efforts in other terrestrial, freshwater, coastal, or marine ecosystems.
Ideally, restoration projects would implement most or all of these recommendations ( Figure 1); however, it is likely that projects will need to prioritize recommendations depending on their local context and needs, including logistical constraints or different stakeholder objectives. Potential strategies to prioritize and select recommendations include multicriteria analysis, deliberative democracy, or codesign approaches, which would allow organizations or institutions to integrate as many recommendations as possible over time to enhance local reef resilience to climate change.
These recommendations also support international initiatives focused on biodiversity and conservation targets (e.g., CBD Post-2020 Global Biodiversity Framework; 30 × 30; UNFCCC COP 27), which are increasingly recognizing the use of restoration for achieving social and ecological outcomes. Key to supporting these global efforts is the demonstration of how countries can meet their biodiversity and climate adaptation goals through targeted coral reef restoration. Global conservation and climate change commitments are transformational opportunities to use restoration to stimulate social-ecological recovery, and the strategic integration of resilience and climate change adaptation into restoration practices in the coming decade is likely to be crucial to this effort. This work provides a first opportunity to address the gap in implementation of restoration to promote reef resilience and climate adaptation and seeks to assist coral reef managers and restoration practitioners to deliver on local and global commitments to sustain coral reefs in the coming decades. While the future of coral reefs is critically dependent on the strongest possible global reductions in greenhouse gas emissions and climate change mitigation, resilience-based coral reef restoration plays an essential role in maintaining these valuable ecosystems while global climate action is achieved.

ACK N OWLED G M ENTS
The authors thank the International Coral Reef Initiative's ad hoc Restoration and Resilience-based Management Working Groups for early discussions and feedback on this topic. In particular, they also thank the following people for providing valuable feedback on initial versions of the recommendations: Mohd Kushairi Mohd Rajuddin, Jennifer Koss, Tom Moore, Scott Winters, Michelle Loewe, and Anderson Mayfield.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data sharing is not applicable to this article as no new data were created or analyzed in this study.