Mediterranean macroalgal forests are typically dominated by canopy-forming Cystoseira sensu lato species, including the genera Cystoseira, Gongolaria and Ericaria (Sauvageau, 1912; Feldmann, 1937; Ercegović, 1952; Molinari Novoa and Guiry, 2020) ( Figures 1A, B ). They are generally considered as the ‘Mediterranean kelps’ (Mangialajo et al., 2008a). Cystoseira s.l. species form dense canopies and create complex three-dimensional structures in rocky coastal ecosystems (Bulleri et al., 2002; Rodríguez-Prieto et al., 2013) providing habitat, food and shelter for many other associated species (Giaccone, 1973; Giaccone and Bruni, 1973; Ballesteros, 1988, Ballesteros, 1990a; Ballesteros, 1990b; Ballesteros et al., 1998; Cheminée et al., 2017; Piazzi et al., 2018; Sant and Ballesteros, 2021a). Cystoseira s.l. forests occur from the upper infralittoral down to the upper circalittoral zone (reported to 50 m depth in Hereu et al., 2008). Their distribution is dependent on a variety of environmental factors such as light intensity, hydrodynamics, temperature and nutrient availability, among others (Feldmann, 1937; Giaccone and Bruni, 1973; Ballesteros, 1989; Ballesteros and Zabala, 1993; Delgado et al., 1995; Arévalo et al., 2007; Hereu et al., 2008; Sales and Ballesteros, 2009; Vergés et al., 2009; Chappuis et al., 2014; Sant and Ballesteros, 2021a; Sant and Ballesteros, 2021b; Ballesteros and Sant, 2022). Cystoseira s.l. species represent one of the most productive and biodiversity-rich habitats of the Mediterranean Sea and underpin important ecosystem services, functions and benefits (e.g., carbon burial and nutrient cycling) (Boudouresque, 1972; Verlaque, 1987; Ballesteros, 1988; Ballesteros, 1989; Ballesteros, 1990a; Ballesteros, 1990b; Sales and Ballesteros, 2012; Piazzi et al., 2018; Pinna et al., 2020). Cystoseira s.l. provides nursery services for fish stocks which in turn support commercial and recreational fisheries, thereby delivering both economic and cultural values (Costa-Domingo et al., 2022; Friedrich et al., 2022). Cystoseira s.l. also provides service as bioindicator for water quality (Ballesteros et al., 2007; Orfanidis et al., 2011).

During the last three decades, most of the Cystoseira s.l. forests have been progressively lost in the Mediterranean Sea (Bellan-Santini, 1965; Munda, 1993; Cormaci and Furnari, 1999; Thibaut et al., 2005; Mangialajo et al., 2008a; Pinedo et al., 2013; Iveša et al., 2016), as with other macroalgal forests around the globe (Steneck et al., 2002; Airoldi and Beck, 2007; Filbee-Dexter and Wernberg, 2018; Bernal-Ibáñez et al., 2021; Martín García et al., 2022). Habitat destruction (loss of suitable substrate from coastal development or other direct seabed contact), changes in water quality following sedimentation, eutrophication and pollution, as well as overgrazing have been the main causes of their decline (Thibaut et al., 2005; Arévalo et al., 2007; Mangialajo et al., 2008b; Sala et al., 2011, Sala et al., 2012; Vergés et al., 2014; Pinedo et al., 2015; Piazzi and Ceccherelli, 2019; Orfanidis et al., 2021) ( Figures 1C, D ). This is expected to be exacerbated by impacts of climate change (such as marine heatwaves; Lejeusne et al., 2010; Celis-Plá et al., 2017; Verdura et al., 2021). Currently, all Cystoseira s.l. species except C. compressa are included in the Annex II of the Barcelona Convention (United Nations Environment Programme/Mediterranean Action Plan) and the establishment of dedicated Marine Protected Areas (MPAs) has been encouraged (Gianni et al., 2013).

Despite a few populations exhibiting natural recovery after a decline (e.g., Iveša et al., 2016), the natural re-establishment of Cystoseira s.l. forests is extremely rare (Chapman, 1995; Soltan et al., 2001; Sales et al., 2011; Capdevila et al., 2018a; Riquet et al., 2021). The lack of a nearby source of propagules and the low dispersal capacity of these species hinder their natural recovery (Perkol-Finkel and Airoldi, 2010). Consequently, active restoration methodologies have become one of the few feasible alternatives to promote the re-establishment of lost Cystoseira s.l. forests, following mitigation of the factors responsible for the decline.

The first records of macroalgal restoration projects go back to 1959, and have substantially been increasing since the 1990s (Eger et al., 2022a). However, these efforts have not been homogeneously distributed across the globe since most projects have been performed in Japan and the USA (Ueda et al., 1963; North, 1976; Wilson and McPeak, 1983; Arai, 2003; Japanese Fisheries Agency, 2009; Japanese Fisheries Agency, 2015; Japanese Fisheries Agency, 2021; see also Fraschetti et al., 2021 Eger et al., 2022a and references therein). Macroalgal restoration efforts targeting Cystoseira species in the Mediterranean Sea only started in 2006 (see Gianni et al., 2013 for a review). Since 2011, collaborative efforts generated knowledge on restoration techniques, protocols and trials ( Figures 1E, F ), as well as complementary actions (Sales et al., 2015; Falace et al., 2018; Verdura et al., 2018; De La Fuente et al., 2019; Tamburello et al., 2019; Medrano et al., 2020; Orlando-Bonaca et al., 2022), roadmaps (Cebrian et al., 2021) and spatial prioritisation (Fabbrizzi et al., 2020; Fabbrizzi et al., 2023). Most of these efforts have been led and developed by academic researchers from public research institutions and universities, at small scales, reflecting the relatively incipient stage of macroalgal restoration (Eger et al., 2022a). In parallel, practitioners have been researching and refining methodologies, exploring the effectiveness of large-scale restoration interventions (e.g., Thibaut et al., 2021). The contribution of other stakeholders, such as governments, private companies, non-governmental organisations (NGOs) or community groups, is now critically needed to go forward with restoration upscaling. Large-scale solutions in restoration actually arise from small-scale successes. These successes inject social values and optimism needed for global investment (McAfee et al., 2021).

The degree of Cystoseira s.l. restoration knowledge is now robust enough to scale up restoration projects (Tamburello et al., 2019). Restoration upscaling requires baseline information (e.g., historic distribution), biological and ecological features (e.g., reproductive phenology, population connectivity), knowledge of mechanisms that promote and dampen the recolonisation process, and indicators for the evaluation of the restoration success (e.g., target species and ecosystem level long-term success). Restoration upscaling also requires a better understanding of the benefits that the restored habitats can deliver to people and the local economy, as well as the costs involved in implementation, monitoring and maintenance. The objective of scaling up restoration actions in the Mediterranean Sea is driven by new ambitious initiatives of the United Nations and the EU: the UN Decade on Ocean Science for Sustainable development (2021-2030), the UN Decade on Ecosystem Restoration (2021-2030) aiming to accelerate restoration of marine ecosystems (UN, 2019) and the proposed EU Nature Restoration Law (EU NRL) (EC, 2022) aiming to repair damage done to European nature by 2050. It is therefore urgent to identify robust guiding principles and practices on macroalgal forest restoration in order to foster stakeholder engagement within science-based restoration interventions.

Restoration in the marine realm is gaining recognition globally, however, it still lags behind terrestrial work due to science gaps, implementation scale, and the appropriate restoration reporting framework to better support decisions on marine restoration (Elliott et al., 2007; Suding, 2011; Blignaut et al., 2013; Bayraktarov et al., 2016; Bayraktarov et al., 2020; Eger et al., 2022b). The importance and increasing practice of marine restoration has driven the need to guide restoration projects towards the best possible outcomes. This has resulted in an increasing number of experience-based publications, particularly in the last few years, with clear guidelines. Whilst large coordinating organisations have taken the role of providing high level generic restoration approaches (IUCN – Keenleyside et al., 2012; SER – Gann et al., 2019; FAO et al., 2021), these are still based primarily on terrestrial restoration. In the Mediterranean, a best practices guide has been developed for site specific case studies, collaboratively with the FAO led Task Force on Ecosystem Restoration (MBPC, 2022). However, on the whole, marine species and habitat specific guidelines have only recently appeared, targeting macroalgae (Gianni et al., 2013; Cebrian et al., 2021; Eger et al., 2022c), seagrasses (van Katwijk et al., 2009; UNEP-Nairobi Convention/USAID/WIOMSA, 2020a; Beheshti and Ward, 2021; Gamble et al., 2021), saltmarshes (Hudson et al., 2021), mangroves (ICRI, 2018; UNEP-Nairobi Convention/USAID/WIOMSA, 2020b), corals (Edwards and Gómez, 2007; Goergen et al., 2020; Hein et al., 2020; Shaver et al., 2020; Quigley et al., 2021; Escovar-Fadul et al., 2022), shellfish (MIT Sea Grant, n.d.; Leonard and Macfarlane, 2011; Fitzsimons et al., 2019; Preston et al., 2020) and multiple habitats (Leocadie et al., 2020). These documents are a mixture of principles, best practices and guidelines for successful restoration and share important key considerations around a restoration action.

Moving further from previous works (Gianni et al., 2013; Cebrian et al., 2021; best principles and guides mentioned above), the aim of the current work was to provide a framework to assist in the restoration decision-making process and to address key considerations of restoration implementation (society, competence, governance and finance). It also completes and improves a previous version of the decision tree proposed by Gianni et al. (2013), appropriately modified to meet new specific considerations for Cystoseira s.l. restoration in the Mediterranean. It simplifies complex decision making and helps to decrease uncertainty in restoration initiatives.

The Cystoseira s.l. decision-support framework ( Figure 2 ) aims at avoiding the initiation of restoration actions where the chances of success are very low and increasing the overall likelihood of restoration success. Those interested in performing a restoration action will have easy-to-follow steps that make the decision-making process smoother whilst science-based.

The framework consists of a sequential decision process with nested elements, giving details on what is needed to be considered when implementing a restoration project, and the steps to follow in the evaluation of success at species and ecosystem levels and long-term monitoring. The decision tree gives insight as to whether an active restoration project should take place or not. The framework also addresses non-academic stakeholders involved in the process of restoration. The following sections give further details on the framework elements.

The key considerations for successful restoration implementation are grouped in four framework pillars concerning; society and support, competence, finance, and governance ( Figure 4 , Box A).

Usually, restoration initiatives first focus on the development of the target species (e.g. vegetation cover, density and biomass) ( Figure 4 , Box B). It is broadly assumed that this is the first step for ecosystem recovery, since other species should benefit from increase in structural complexity (Geist and Hawkins, 2016). Recovering the target species population is essential for the potential re-establishment of ecosystem processes and functions (Geist and Hawkins, 2016). However, the relationship between the recovery of a target (e.g., Cystoseira s.l. species) and the processes and functions of the ecosystem has to be empirically tested (Benayas et al., 2009; Moreno-Mateos et al., 2012; Crouzeilles et al., 2016). Therefore, the assessment of ecosystem processes and functions provided by Cystoseira s.l. forests should include quantifiable properties to describe the community in terms of structure alongside ecosystem functions (Montoya et al., 2012). To date, reported successful macroalgal forest restoration has mainly focused on the recovery of the canopy-forming species (Whitaker et al., 2010; Verdura et al., 2018; Fredriksen et al., 2020; Layton et al., 2020; Gran et al., 2022), and only a few studies have evaluated the re-establishment of associated species (Ling, 2008; Marzinelli et al., 2016; Galobart et al., 2023).

Forest restoration projects need to have clear, time-bound and meaningful objectives, based on which the indicators of restoration success can be specified (Stanturf et al., 2001). Monitoring of the selected indicators is based on rigorous sampling (the duration and periodicity of which will be species- and ecosystem-dependent) and reference conditions. The inclusion of multiple control sites is needed to assess the outcomes of the restoration action. Reference sites should be ecologically similar to the site selected for restoration interventions, except for the absence of anthropogenic pressures. In the case of Cystoseira s.l. forests, finding control sites in the same region may not be an easy task due to their important loss at the local scale. Reference sites may be available in MPAs, effectively protected and largely intact, but this is not generally the case for macroalgal forests in the Mediterranean Sea. In the absence of proper reference sites, reference conditions can be established on the basis of historical data or models (e.g. Thibaut et al., 2005; Fabbrizzi et al., 2023).

In general, success evaluation of marine restoration interventions is based on short-term periods (Bayraktarov et al., 2016; Kollmann et al., 2016; Fraschetti et al., 2021). Bearing in mind that the recovery of many marine ecosystems can take up to 15-25 years (Jones and Schmitz, 2009; Borja et al., 2010; Bekkby et al., 2020) and that canopy-forming macroalgae are mid- to long-lived species (Schiel and Foster, 2006; Smale et al., 2013), longer evaluations should be considered for reliable outputs (e.g., Gran et al. (2022) revisiting a site 10 years after re-introduction, reporting a 3 orders-of-magnitude increase in the extension of the forested area). The life span of most Cystoseira s.l. species is still unknown, and likely highly variable. More long-term ecological studies are needed to establish common protocols and indicators of Mediterranean forest restoration actions.

Restoration occurs as a succession of achievements. In the short-term, this involves the success of the action implemented (e.g., recovery of the target species and population). In the long-term, success is assessed through high level restoration goals, usually through ecosystem level indicators, such as ecosystem functions and services.

When success is achieved at the ecosystem level ( Table 2 ), the next step in the framework is long-term monitoring and adaptive management ( Figure 4 Box C). Here, participatory monitoring should be implemented with the involved stakeholders and, as much as possible, in the framework of an ad hoc long-term program involving citizen science (Gann et al., 2019). Such monitoring, if based on scientific knowledge and robust yet simple methods, is often more beneficial and relevant for stakeholders than conventional scientific approaches (Gann et al., 2019).

An adaptive management approach, suggested in case of uncertainty about which management action is the more appropriate (and this is the case for several Cystoseira s.l. forests) could be based on timely monitoring and an iterative evaluation of results, as well as funding for ongoing restoration (Gann et al., 2019).

The indicators selected for monitoring should not be destructive or invasive, and this is particularly true for restored populations. If possible, they also have to be easy and rapid to assess (including by non-scientists), as well as time- and cost-effective. Target species and ecosystem level indicators ( Table 2 ) and other biotic and abiotic factors potentially threatening the forest ( Table 1 ) should be considered for long-term monitoring in the framework of adaptive management. As an example, the proliferation of herbivores (i.e. sea urchins or herbivorous fish) should be monitored, in order to anticipate the depletion of the restored forest.

Long-term monitoring will contribute to the knowledge of the functioning and evolution of the ecosystem and can point out where further interventions may be required. If during the implementation phase, long-term monitoring shows early warning signals such as a decrease in forest cover or density, an assessment of the causes of degradation ( Table 1 ) should be immediately performed and intervention, mitigation or regulation actions considered.

In order to have a favourable restoration outcome but also to preserve Cystoseira s.l. forests in a good-moderate conservation status, the establishment of an MPA is a proper management tool, prompting increased awareness and better communication of the actions carried out, and the possibility to regulate those activities that could threaten the restored forests (e.g., fishing, trampling, beach management, anchoring). If well managed, MPAs reduce levels of human pressures, allowing long-term stabilisation of essential ecosystem processes, and fostering the resilience of marine communities, such as Cystoseira s.l. forests, to future disturbances (e.g., climate change) (Bevilacqua et al., 2022). MPAs usually offer important support services (e.g., video surveillance monitoring, patrolling, vessels, trained personnel), but their creation should not replace long-term monitoring: disturbance factors may also be present in protected zones, due to fluctuation of other species potentially interacting with the restored Cystoseira s.l. forests. MPA establishment may take considerable time from planning to implementation, and this should be taken into consideration when assessing the priority needs for action towards degraded area recovery.

New methodologies and techniques for the restoration of the Cystoseira s.l. have been developed within the last decade through viable programmes and with the know-how for all major steps now in place (Cebrian et al., 2021; Fabbrizzi et al., 2023). Improvements are still necessary in restoration protocols to ensure optimal success, particularly the refinement of propagule handling, the conditions for ensuring their viability and procedures for transplanting them into the field. Restoration upscaling is possible when the environment and the intensity of human impacts is compatible with restoration goals (e.g., Gran et al., 2022). Current knowledge gaps concern some of the less ‘well-known’ Mediterranean macroalgal species and their requirements for optimal survival and growth, as well as their historical and potential distributions. There are also some gaps on how to deal with stressors that hinder restoration success. All present and potential future impacts have to be assessed (Cebrian et al., 2021), particularly natural ones such as extreme climatic conditions and climate change, which may be solved with identification of future potential restoration areas, and the use of different species strains or more tolerant species (Verdura et al., 2021). More efforts to understand the role of grazers (fish and invertebrates, including mesograzers, Monserrat et al., 2023) in the control of macroalgal forests are also needed to define strategies for reducing grazing pressure to an adequate level, over which the restoration programme would not be viable.

One of the key restoration gaps is in linking existing top-down policy requirements and bottom-up initiatives (Ramírez-Monsalve et al., 2021). Policies give target percentages of degraded habitat to be restored, but do not state what, where and how to restore. In contrast, bottom-up initiatives are often promoted by scientists who have specific research interests and find an opportunity for science-based action (Smith et al., 2021). The proposed EU Nature Restoration Law is expected to partially fix this issue by specifying the need for Member States to restore target percentages of selected habitats that include various Cystoseira s.l. habitats. However, there need to be bodies that can coordinate, prioritise, facilitate and fund actions, whether this is part of a regional organisation (e.g. UNEP PAP-RAC, GFCM, Barcelona Convention), national or local authority. Funding could also be secured through these or other groups (e.g. EU LIFE projects including new projects in support of the EU Nature Restoration Law, national bodies), the prioritisation of the Global Environmental Facility (GEF) or industry, with the premise that restoration is an investment, rather than a cost. Funding also needs to consider the scale of the action and the length of time required to maintain a monitoring cycle covering long-term goals. Standardised restoration elements need to be listed and their cost realistically estimated to depict appropriate budgets enabling complete restoration action and for communicating the level of investment required for recovery of ecosystem services (Verdura et al., 2018; Friedrich et al., 2022). Restoration interacts synergistically with conservation (passive restoration) and can easily be linked to, for example, MPAs or other area-based measures. These help to fulfil the removal of human stressors. The difference between conservation and restoration sits within the intervention framework, with the main distinction between passive and active restoration lying primarily in the timing and extent of human interventions (Chazdon et al., 2021). In the EU, restoration should also be linked to Maritime Spatial Planning, and at the Mediterranean regional scale to the Integrated Coastal Zone Management Protocol and the United Nations Mediterranean Action Plan (UNEP-MAP and UNEP PAP/RAC) where restoration areas are allocated within spatial plans and highlighted for protection. Another mechanism that could be used in the region to prioritise restoration actions on Cystoseira s.l. concerns the designation of “Other Effective area-based Conservation Measures (OECMs)”, agreed under the 14^th Conference of Parties of the Convention on Biological Diversity.

Publicity is important in providing a common understanding of the problems posed by Cystoseira s.l. degradation and loss. Understanding, valuing and communicating the societal benefits arising from healthy Cystoseira s.l. forests are essential to raise public awareness about their importance. Good and widespread publicity will drive public support, which in turn can drive institutional support for further actions. A coordination group could provide the backbone for efficient information. The message of restoration should be very clear, what is meant, what is feasible, what can be done and what is expected. Although restoration is about helping nature recover for the benefit of people and nature, the language of restoration has not always been clear. This has led to ambiguity and misunderstandings among stakeholders, whether restoration is achieved through protection and natural regeneration only, or restoration may also require various direct interventions (e.g., direct removal of grazers, substrate creation, transplantation). In most sites however, a range of protective and restorative actions will be required for the Cystoseira s.l. restoration (Gann et al., 2019; Chazdon et al., 2021).

In a broader context, Fabbrizzi et al. (2023) demonstrated that introducing systematic conservation planning principles and tools in restoration projects is crucial to understanding and defining how much and where an ecosystem or habitat can be recovered. These conservation planning principles and tools allow to effectively manage efforts and assess possibilities for setting region-specific targets. Adopting marine spatial planning leads to accounting for environmental constraints and socio-economic implications affecting restoration activities. The use of prioritisation software (e.g., MARXAN, Zonation 5) informs the allocation of restoration targets identified a priori, by combining spatial information from different sources. Future efforts should be directed to better integrate site prioritisation into marine spatial plans, accounting for ecological, social and economic objectives to enhance system resilience.

CS, JV, NP, LM, SF, EC, EF, MM, MD and SB contributed to the conception and design of the study. CS, JV, NP and LM led the writing of the manuscript with major contributions from SF, EC, EF, MM, MD and RD. LM, JV and MD provided the original decision-support schematic. JV prepared the manuscript figures and tables. All authors contributed to the revising of the manuscript and approved the submitted version.