Deliverable 3.2 Report on the features of the experimental designs for marine restoration actions on demonstration sites and over reference areas

1 Deliverable 3.2 Report on the features of the experimental designs for marine restoration actions on demonstration sites and over Ref. Ares(2024)8740358 - 06/12/2024 2 Project number: 101093865 Project duration: 1 Dec 2022 – 30 Nov 2025 Project coordinator: Ida Beathe Øverjordet, SINTEF Ocean Web site: www.climarest.eu Deliverable ID: D3.2 Due month: M12 Preparation date: 2023-09-30 Title: D3.2 Report on the features of the experimental designs for marine restoration actions on demonstration sites and over. Lead beneficiary: UNINA Prepared by: Simonetta Fraschetti – Sara De Benedictis – Simone Musumeci UNINA Cristina Gambi, Cinzia Corinaldesi, Antonio Dell’Anno, Alice Premici & Roberto Danovaro - UNIVPM Abstract A literature review has been conducted to assess which experimental designs are adopted for restoration across marine ecosystems along the EU Seas and behind. Documents have been selected for showing the logic and the designs adopted in the most relevant studies showing restoration success. Overall ca 25 references are included in the document to support the description of experimental designs. The output of this literature review is included in the Deliverable 3.2: Report on the features of the experimental designs for marine restoration actions on demonstration sites and over. Dissemination level PU Public X CO Confidential, only for members of the consortium (including the Commission services) CI Classified information as referred to in Commission Decision 2001/844/EC) Deliverable type R Document, report X DEM Demonstrator, pilot, prototype DEC Websites, patent filings, videos, etc. OTHER Software, technical diagram, etc. Authorship information Editor Simonetta Fraschetti, UNINA Contributing partners UNINA & UNIVPM Version history Version number Date Description of changes 1 2023-09-30 Report submitted to the project officer 3 Table of contents 1. Experimental designs in ecological restoration 4 2. Case Studies in Climarest 6 2.1 Svalbard, Norway 6 2.2 Ireland 7 2.3 France 8 2.4 Spain 8 2.5 Madeira, Portugal 9 3. Literature review 9 4. Seagrass meadows 10 4.1 Examples of restoration using different methods 10 4.2 Anchoring of shoots with iron nails 12 4.3 The importance of the context and the effectiveness of seedlings 14 5. Oyster reefs 16 5.1 Effects of cathodically protected steel vs plastic for oyster restoration mats 16 5.2 Substrates and tidal elevation 18 6. Kelp forests 19 6.1 Effect of herbivory and thresholds of change 19 6.2 Use of green gravel as a new restoration method 21 6.3 Artificial structures 22 7. Algal forests 24 7.1 Effect of conspecific adults and grazers 24 7.2 Combining active and passive restoration 25 7.3 In situ vs ex situ restoration 26 7.4 Ex situ restoration technique with facilities far from restoration sites 28 7.5 Testing the effects of culling on reversing barrens 29 8. Saltmarshes 31 8.1 Restoration, sediment types and associated environmental variables 31 9. Final considerations 33 10. References 35 4 Summary Restoration is considered an effective strategy to accelerate the recovery of biological communities at a local scale. Recent initiatives in restoration demonstrated that it can be carried out also at large scale addressing the same scales of human pressures. Despite several success stories, the effects of restoration actions in the marine ecosystems are still unpredictable. Recent reviews highlighted the importance of several factors contributing to restoration effectiveness, able to decrease unpredictability of results (Fraschetti et al., 2021) such as site selection and restoration techniques. In addition, methodologies related to “protecting” the restored site, for example through the removal of invasive species or the use of protective mesh cages, have been found to have a significant positive influence on the outcomes of restoration. However, one issue that should be considered central in restoration, as in all ecological studies, is the importance of rigorous experimental designs to implement restoration interventions, assess restoration success and related drivers. This deliverable has been conceived to show examples of experimental designs adopted to test the role of different interventions and processes (methodologies, substrates, adult presence, herbivory, gradients of pressures) across the habitats included in the project. 1. Experimental designs in ecological restoration Restoration is a costly process (Bayraktarov et al., 2019). This often translates into the choice of limiting the efforts (e.g. replication in space and time) in the field and under laboratory conditions. This should never happen since it corresponds to serious consequences in terms of correct assessments of the effects of restoration interventions, success and failures, with the impossibility to provide the correct guidelines to practitioners. Measuring and explaining the variability inherent in ecological systems is not an easy task and systems needing ecological restoration are featured by the same large variability characterizing most natural systems. International Standards of Ecological Restoration (Gann et al., 2019) recognize the need for carefully planning rigorous experimental designs to support robust restoration interventions. This issue has been recently covered also in specific meta-analyses (Tedford & Castorani, 2022) who showed that understanding the controls on variation in oyster predation strength, able to compromise restoration effectiveness, is impeded by inconsistencies in experimental methodologies. To understand the outcomes of restoration, and to avoid spatial confounding, it is necessary to compare the putative impacted area with more than one reference area several times before and several times after the intervention, with the same number of replicates in each area and at each time. This approach has a long history and is an evolution of the Before-After-Control-Impact or BACI experimental designs. It is called beyond-BACI (Fig. 1) and increases the ability to unambiguously identify cause-effect relationships between anthropogenic disturbance and responses of natural populations (Underwood, 1993). Such formal monitoring has been largely used in environmental 5 impact assessments and in studies of the effectiveness of protection of Marine Protected Areas (MPAs). The use of this logical framework can be also extended to active restoration. When studies of human interventions are carried out in single sites, the analysis becomes asymmetrical since there is only one single affected/protected/restored area to be compared with multiple control areas. The procedure is based on the comparison between the putatively impacted/protected/restored areas and multiple controls, sampled at different times before and after the intervention (Fig. 1). In most cases, however, no data are available before a disturbance occurred, before the establishment of an MPA, or before a restoration intervention is planned. In such a case, it is still possible to analyze the hypothesis that the protected/impacted/restored populations or assemblages differ from those in non-treated reference areas. Also in this case, analogously to what was described above for environmental impact assessment, multiple controls are needed to separate natural patterns of variability from the effects of the factor involved in the analyses (disturbance, protection, restoration). The classification of success, partial success, and failure is impracticable in case studies where the criteria to evaluate restoration success are not based on a formal comparison between restored sites (single or multiple) and replicated controls. This kind of experimental design increases the ability to unambiguously identify cause-effect relationships between specific interventions (e.g., anthropogenic disturbance, protection, other treatments) and responses of natural populations. Spatial confounding or pseudoreplication (Hurlbert, 1984) are a consequence of the absence of multiple ‘control’ areas in studies of human interventions. We also emphasize the need of repeating measurements in time, to estimate temporal variability of response variables. Fig. 1. Beyond-BACI design for the assessment of environmental impact. The procedure is based on the comparison between putatively impacted areas (in this figure only one) and multiple controls sampled at different times before and after the intervention (indicated by the arrow). In case the intervention would be relative to restoration, we expect an increase in the selected response variable, compared to controls (see Underwood, 1993 for further details). It is important to stress that controls must have the same type of habitat as the area selected for restoration and should be featured by the same ecological processes occurring in the restored area. 6 “Ideal” controls do not exist: they should always be randomly selected among those potentially representing the ecological processes of the same area where the intervention is occurring. Here we provide a selection of experiments that have been conceived to test specific hypotheses in restoration intervention. We anticipate that in many cases it was difficult or even impossible to derive with precision how the study was carried out in the field, since most studies lack of relevant information on the logic and the design behind the experiment, and considering the factors included in the analyses. All cited literature comes from peer-reviewed journals. Thus, most of times, rather than specific criticism across the reported examples, final comments will stress common limits encountered in this Deliverable, introducing specific suggestions to increase the consistency of the results. 2. Case Studies in Climarest The project has included five Case Studies conceived to improve the status of different ecosystems. Climarest has the final aim to provide guidelines to solve very common environmental issues adopting different approaches (from restoration to mitigation and monitoring). A description of the activities follows. 2.1 Svalbard, Norway This Case Study develops in two habitats: Shores: preventing further erosion in Arctic beaches with loose sediments. Soft bottom: similar substrate to reference conditions within the end of the project. Species target & habitat: no target species but target ecosystem types: 1) arctic littoral shores and 2) soft sediment bottom. The biodiversity target is to increase diversity of species. Location: Adventfjorden (Svalbard) Criteria for site selection: an urban site that is threatened by coastal erosion (primary site), a cultural heritage site that is threatened by coastal erosion (secondary site). Reference either to a rocky coastline or a stable (no coastal erosion) sedimentary coast. Soft bottom: Sampling stations in a transect away from the sewage outlet and in a reference fjord. The reference needs to have similar features, but it should be without human impact. Adopted experimental design: Experimental design for the assessment of outcomes: BACI (before/after/control/impact) Description of the protocol and activity Shores: flora and fauna biodiversity at the coastal and shore and shoreface. Video survey with blue eye for flora and fauna. Visual survey for fauna (birds) and onshore flora. Yearly frequency of surveys. Soft bottom: Soft bottom fauna diversity characterization (3 grabs from tentatively 5 stations in a transect from the outlet and in the reference fjord), eDNA profiles of sediments and bottom water compared to diversity analysis. Pollution characterization; Sampling of sediments by grab, quantify 7 litter, sample for microplastics and chemical pollution (3 grabs from each station, tentatively 5 stations in the polluted fjord and one station in the reference). Waters samples at 3 depths of all stations for investigations of spreading of the plume. Sampling in the fjord once a year. Sampling of sewage with characteristics of pollution, nutrients and oxygen consumption during degradation before and after the sewage sieve monthly in collaboration with Longyearbyen lokalstyre. Samples of the waste material removed at the sewage sieve. Before and during social campaigns: Sample different branches of the sewage network to see if the campaigns have higher effects in some areas than others. Areas are largely separated into housing, hotels and restaurants, school and kindergarten, industrial area and the airport. Compare before and after levels of pollutants and waste. 2.2 Ireland This Case Study focuses on seagrass restoration. The restoration intervention is presently carried out in three locations with replicated donor meadows and receiving sites. Fig. 2. Distribution of the experimental sites for the restoration of the seagrass meadows in Ireland. At present, activities are focused on the collection of material for study population genetics of Zostera spp. across Ireland (4 out of 20 sites), site selection and characterization of environmental conditions to ensure favorable growth conditions for seagrass restoration. A small-scale seagrass restoration study and monitoring of indicators of transplantation success is on-going in 4 sites and the engagement of relevant stakeholders and formation of a citizen science network for future monitoring and assessment of restoration tools and strategies is on-going. 8 2.3 France The restoration objectives within Climarest are to initiate restoration of native oyster reefs at the scale of 100 to 1000m² within the timeframe of the project Species target & habitat: Ostrea edulis reefs Location: Quiberon bay and bay of Brest Criteria for site selection: Remaining wild populations Adopted experimental design: Deployment of replicated reefs supporting settlement and growth of native oyster and catalyzing the formation of a three-dimensional biogenic reef. Phased deployment of reef structures between year 1 and year 2. Year 1: demonstration and monitoring of upscaled reef implementation. All reefs are deployed in the same site, chosen for its good potential in terms of recruitment. The reefs will face different treatments, two of which at least will be predator removal vs. no predator removal. Year 2: complementary deployment to evaluate conditions and processes for restoration in sites not prone to good recruitment. Material: Artificial reefs used as early colonizing substrates. 2.4 Spain The objective is to develop nature-based solutions to mitigate the impact of mussel farming on benthic ecosystems, secondary production and C, N and P fluxes to higher trophic levels. Expected outcomes are: · Increase the resilience of communities and populations of interest below mussel farms. · Develop restoration techniques for invertebrate and fish populations of commercial interest affected by overfishing and the effect of climate change. · Restocking and stock enhancement of the population of European lobster (Homarus gammarus) providing shelter and increasing food availability. · Analyse the possibility of developing IMTA to promote the use of waste derived from mussel farming in European Waters. · Contribute to climate change mitigation in terms of sequestration of atmospheric carbon, this study will assess the potential of ARs to store and sequester blue carbon below mussel culture sites. Species target & habitat: Species target: macrobenthic invertebrates, benthic macrofauna and the European lobster (Homarus gammarus) as an endanger species; Habitat: Soft bottoms, Sandy bottoms, Maerl bottoms Location: the outer part of the Ría de Vigo, in front of the beach of Limens Criteria for site selection: The locality chosen for restoration actions is representative of the area, with a medium intensity of mussel farming and high fishing pressure Adopted experimental design: The study will be conducted in two different sites: under mussel rafts and on sandy bottom without rafts. Three AR units will be deployed at each study site, resulting 9 in a total of two sites with ARs under mussel rafts (3 replicates for site) and two control sites without ARs, as well as two sites on sandy bottom with and without ARs. Material: A total of twelve artificial reefs units will be installed at soft bottoms in Ría de Vigo (12-15 m depth). Each unit is composed of three cement blocks attached, covered with similar shapes to Maërl beds and holes of different sizes to accelerate the colonisation and provide shelter for fishes and invertebrates. 2.5 Madeira, Portugal The objective of the Case Study is rocky subtidal restoration under protected and not protected conditions to test potential synergistic interaction between passive and active restoration. The idea is also to reduce urchin densities, by culling. The experimental design is featured by two protected and two not protected locations. In each location, three replicated plots will be dedicated to grazing reduction, three to facilitate algal recover with the use of artificial structures, three in which grazing reduction and artificial structures will be combined, three control plots. The experiment will provide relevant information on how active and passive restoration combined with controlled herbivory and adopting NBS can boost the recovery of disturbed systems this supporting theory and practice in ecological restoration. 3. Literature review We used the same literature review carried out to support Deliverable 3.1. The literature review was conducted to identify restoration protocols adopted in different marine ecosystems worldwide with special reference to the habitats identified in the CLIMAREST demonstration sites (seagrasses, oyster reefs, algal and kelp forests). The process of inclusion in the search of both scientific (through WoS and Scopus) and grey literature, is described in the Deliverable 3.1. For the analysis we extracted the following numbers of studies across habitats: ● Seagrass: 5 ● Saltmarshes: 2 ● Oyster reefs: 2 ● Algal and Kelp forests: 8 4. Seagrass meadows 4.1 Examples of restoration using different methods Derrenbacker and Lewis (1982) studied three methods of seagrass planting (use of long steel staples to anchor the seagrass, hand-broadcast Thalassia testudinum seedlings, sections of T. testudinum rhizomes with attached short shoots transplanted). They were evaluated in an area of Lake Surprise, Key Largo, Florida which had been impacted by water pipeline installation. The experimental design includes the selection of three areas with different impacts: moderately impacted (shell hash) area, a severely impacted (fine silt) area, and a severely impacted (rocky) area. This comparison 16 impressed current cathodic protection system to develop mineral accretion. The objective was to determine if mineral accretion mats were as effective at promoting the growth of oysters as plastic oyster mats. It is believed the creation of the mineral accretion structures will enhance oyster growth, which in turn provides increased filtration and improvement to local water quality. The structures will also create habitat for organisms associated with oyster reefs, such as juvenile fish, barnacles, crabs, shrimp, mussels, tunicates, snails, and algae. Response variables were, among the others, oyster recruitment and growth. All steel mats were subjected to pre-chalking to develop a thin mineral accretion layer, prior to deployment. This prevented the steel from rusting and would allow for ease of deployment at more remote locations. After pre-chalking, mats were rinsed and deployed at three distinct locations along the east coast of Florida. This study was carried out along the coast of Florida where three locations were selected at a distance of tens of kilometers. Fig. 8. Steel and plastic mats were deployed at three locations along the east coast of Florida, within the Indian River Lagoon system: Port Canaveral, Melbourne Beach, and Grant (Hunsucker et al., 2021) 17 At each location, three replicates of both steel and plastic mats were placed and the three replicates on steel were connected via marine grade wires to a solar panel to continue mineral accretion. The three locations were not similar each other since, for instance, Port Canaveral is located about 4 km from the Atlantic Ocean, representative of open water conditions. However, based on pilot projects and previous field work, all three sites were considered appropriate for the development of mineral accretion and oyster settlement. All oyster mats were at their final immersion site in early August 2019. However, formal comparisons considered these differences might be difficult. For the description of the methods, please refer to Hunsucker et al. (2021). Fig. 9. Oyster restoration mats (45.7 cm by 45.7 cm) which constructed of a mild steel (left) and aquaculture grade plastic (right) (Hunsucker et al., 2021) Time: Materials were monitored weekly, for three months. In this study, researchers had to deal with extreme events. After five weeks from the experimental set up, all mats were moved to a secure location in preparation for Hurricane Dorian. The hurricane sites provided immersion and protection from the storm, but had a lower salinity (<10 ppt) compared to the regular sites. All mats were returned to the test sites within about a two-week period and continued to be monitored for remainder of the three-month immersion. Oyster recruitment was observed after one month at all three of the test locations. At this time, the number of observed oyster spat was comparable between the plastic and steel mats. At week five, all test sites had oyster settlement on the steel mats, whereas settlement was only observed on the plastic mats at Grant. Main results: at the end of the experiment, there were 70 oysters recorded on the steel mats compared to 12 on the plastic mats. In addition to oyster spat growing on the dead and dried oyster substrate, they were also found growing directly on the steel mesh. The steel mats were effective at promoting the growth of oysters, but the overall recruitment was dependent on the test site and each varied based on environmental and ecological conditions. At the Melbourne Beach location, there was a significantly greater presence of oysters on the steel 18 compared to the plastic. At Port Canaveral and Grant, oyster growth at times was greater on the steel but at the end of the three-month period, numbers were similar between the steel and plastic. Based on the oyster counts alone, it is apparent that the steel mats can promote and sustain oyster growth at a rate greater than or equivalent to the traditionally used plastic mats. 5.2 Substrates and tidal elevation In the study of White et al. (2009) to restore a population of Ostrea lurida, drastically compromised and incapable of natural recovery, various types of substrates and different tidal elevations have been tested, which are identified as factors capable of influencing the restoration of this species. The aim of the project was to assess the limits of the recovery of O. lurida and explore techniques to improve the success of restoration efforts. The study was conducted in the North Bay Oyster Reserve (Puget Sound, Washington, United States) and the initial condition is represented by three different tide levels (+0.3m, 0m, -0.3m) in which 3 rows were distributed, one for each level, of ten experimental units represented by squares of 1 m²; within the squares, six different types of substrate were randomly placed: bare rock, gravel, crushed shell of Crassostrea gigas (Pacific oyster), whole shell of C. gigas, whole shell of O. lurida and live O. lurida, with 5 replicates for each substrate. The experimental design provides, therefore, 30 experimental units, positioned at a distance of about 2 meters from each other. Two monitoring campaigns were carried out after 5 and 11 months from experimental set up. Response variables: O. lurida recruit density was calculated for the 3 tidal heights and the 6 substrates, and post-settlement mortality was evaluated. Main results: Unfortunately, the results obtained do not guarantee the achievement of the restoration success because statistically no significant differences were found in the survival rate or in the number of recruits according to the type of substrate or the level of tide. This work represents, however, a starting point to address new successful restoration experiments that can use as a basis of their experimental designs the lowest tide level (-0.3m) and the substrate formed by shells of Ostrea lurida, which represent, in any case, the factors with the most promising results of this experiment. 6. Kelp forests 6.1 Effect of herbivory and thresholds of change Leinaas et al. (1996) documented the response of the species Laminaria saccharina and Laminaria hyperborea to reduced grazing pressure of Strongylocentrotus droebachiensis through the analysis of the recovery patterns of the kelp forest in an area where historically the two species were recorded. Responses from the sea urchin population structure to increased kelp cover were also studied. During the study, an unexpected mass mortality event, which involved the area, was instrumental in providing valuable information on reducing the density of sea urchins needed to initiate kelp recovery. The response variables were sea urchin density and algal cover. 19 The study was conducted near the island of Vega (Norway) with two sampling sites represented by two skerries separated by a 5 m sand tongue; one skerry is the experimental site, where all sea urchins between 0 and 10 m in depth were manually removed, and the other was the only, unmanipulated control. As recognized also by the Authors, the replication level was actually very limited. Thus, eleven additional sites were monitored to assess whether changes in vegetation at the manipulated site could be linked to natural fluctuations or to the sea urchin removal. A photographic sampling, with five replicated photographs taken within a frame of 0.25 m² at three different depths (2, 5, 10 m) and along transects of 30 m (on the manipulated site), 50 m (in the control site) and 1015 m wide, was conducted to calculate algal coverage at the sampling site, at the control site and at three of the eleven sites already monitored. The density of sea urchins, on the other hand, was estimated randomly with ten squares of 0.25 m² at three depths and along the same transects; at a depth of 5 m, in addition, about 60-120 sea urchins were measured with the callipers to estimate their size, discarding those with a diameter less than 10 mm. For more information on the structure of the vegetation, algae were collected from 0.25 m to 2 m of depth, with 7 replicates per site, in the 5 sites already mentioned; the algae were counted and measured in length. Fig. 10. The study sites in a group of small islets and skerries north of Vega Island: 1 experimental skerry, 2 control, 3, 4 and 5 other sites where sea urchin density and algal cover have been recorded since 1991 (Steinskjær, Nilsarentsskjærand Sandø respectively). Unnumbered arrows point to sites where persistence of the barren state was observed until summer 1991 (Leinaas et al., 1995) Time: Sampling was carried out for a total of five years. After the first three years of the experimental set up a mass mortality event of the sea urchins affected the studied area. In the final analyses, this disturbance event was also included in the ANOVA, to quantify the eventual effect of this disturbance on recovery patterns and population structure. Main results: the recovery of kelps reflects the degree of reduction in the density of sea urchins: moderate reductions in the density of sea urchins allowed only annual algae growth (L. saccharina) 20 with the potential of shifting again towards a barren state, while stronger reductions induced the recovery of the forest of L. hyperborea. The study suggests that about 10 individuals/m² of Strongylocentrotus droebachiensis would prevent the recovery of a forest in the upper subtidal, about one-third of the densities originally observed in the area. A density of 20-25 individuals per m² would maintain the area as barren. At 5-10 m of depth, 5-10 sea urchins per m² are enough to maintain the barren because the algae have a reduced growth linked to the greater depth and are thus more vulnerable to grazing. The article does not mention the success of the restoration since the main purpose was to identify the threshold of the number of sea urchins allowing the recovery of the system, without further human intervention. 6.2 Use of green gravel as a new restoration method Fredriksen et al. (2020) developed and tested “green gravel”, a novel restoration technique for kelp forests, which overcomes many limitations of current approaches. Restoring reefs using green gravel requires little investment and provides potential pathways to propagate resistant genotypes that could ‘future proof ’ vulnerable kelp forests to future stress. Moreover, the technique is applicable to both laminarian and fucoid kelp forests. To test the effectiveness of different transplantation techniques the gravel was out-planted using three different methods: (1) divers placed the gravel in open boxes (30×40 cm) on the seafloor at 7 m depth, (2) divers laid the gravel in open plots on bare rock at 7 m depth, and (3) gravel were dropped to 3 m and 7 m depth from a boat. Gravel was dropped from the surface by overturning a tray of gravel over the side of the boat, while a diver marked the landing site. Fig 11. Work flow for making green gravel. Fertile plants are collected (A) and reproductive tissue isolated for zoospore release (B). Small rocks are seeded in trays by adding spore solution (C). After a few weeks small sporophytes are visible (D). Green gravel are scattered on the reef (E) where they continue to grow (F). Photo credits: H. Steen, all photos. (Fredriksen et al., 2020) The design included one donor site and field sites. The donor site was outside Flødevigen Research Station (southern Norway) where fertile kelp plants (Saccharina latissima) with visible sori were collected in autumn (n = 10–15) and sporogenic tissue was excised for zoospore release. Kelp zoospores were added to small rocks (Fig. 11 B,C) with high, medium and low starting densities, 21 respectively. The field sites were located in a semi-protected area with patchy kelp (2–20% canopy cover) and turf-dominated reefs and mainly coarse sedimentary substrate. The response variables were kelp length, growth and green gravel retention. To test whether gravel could be retained on rocky reefs, divers transplanted green gravel onto bare and turf-covered reefs at 6 m depth at 4 sites in early winter. At each site, a handful of gravel was placed on the seafloor inside 3 cleared 0.25 m2 plots or inside 3 ‘open plots’ with high turf-algae cover (n = 5–7 gravel in each plot). Gravel was covered with small kelp sporelings <2 cm length. Cleared plot treatments were created by removing all turf algae with a scraper from a 0.25 m2 quadrat, leaving only bare rock and encrusting coralline algae. Substrate (turf and bare rock) was treated as a fixed factor and study sites and replicated plots were treated as random factors with plots nested within sites. Fig. 12. Healthy sugar kelp forests (Saccharina latissima, A) disappeared along the southern coast of Norway in 2002. Large tracts of coast once dominated by sugar kelp now only have scattered individuals, and the seafloor is instead covered by a dense carpet of turf algae (B). This loss has been persistent, warranting investigation of restoration techniques - such as green gravel (C) - to aid recovery. Photo credits: K. Filbee-Dexter (A), T. Wernberg (B), S. Fredriksen (C). (Fredriksen et al., 2020) Main Results: The out-planted kelp had high survival and growth over 9 months, even when dropped from the surface. Seeded juvenile kelps can eventually overgrow the green gravel and attach to the surrounding substrate or underlying rock, which suggests the plants could be retained on these reefs. Cost assessment is also available in the study. 6.3 Artificial structures The objectives of study of Oyamada et al. (2008) were to observe the algal succession, the growth of Ecklonia cava and other attached organisms on two types of artificial-reef blocks. The experiment design included in one site, five Marine Blocks (1m3) and five concrete blocks (1m3) installed close to E. cava beds, compared to natural seaweed beds, placed at a depth of 7 m around the coastal frontage of Jogashima, Kanagawa, Japan. They replaced five concrete blocks with five cubicMarine Blocks with a size of 1m3. The Authors stress that a concrete block is an 22 appropriate reference substrate for evaluating artificial substrates for seaweed bed development, because there are already many case studies on concrete blocks in the literature. Fig.13. Marine block (1m × 1m × 1m). (Oyamada et al., 2008) The area consists mainly of rocky ledges, and there is a natural marine forest of Ecklonia cava, a large perennial brown alga, whose lifetime is 3–4 years. The blocks were placed in November 2001 on relatively flat areas of seafloor scattered across this terrain. Scuba-diving observations and measurements, to count the number of E. cava and Sargassum plants and other attached organisms attached to each block, were continued until April 2006. Response variables: the number and length of all E. cava individuals on the upper surface (1×1m) of each of the five blocks and identified the major seaweeds and attached organisms. The observation frequency was set to be once every 2–5 months. Fig.14. a,b Photos of top surfaces in October 2002. a Marine Block. The surface is covered with Ecklonia cava. b Concrete Block. Ecklonia cava is attached to the edge of the block (Oyamada et al., 2008). Main results: the succession in the algal vegetation on blocks was similar to that in the vegetation in the surrounding natural marine forests of E. cava. In the succession, diatoms first covered blocks, and then small annual algae appeared, and ultimately E. cava, a large perennial alga, became dominant and formed stable vegetation on nearby natural rock reefs. Such processes are normally seen in natural seaweed vegetation and so it can be concluded that both Marine Blocks and concrete blocks can form artificial reefs having the same functions as natural rocky reefs. Further study showed that the number and weight of E. cava on natural reefs was smaller than that on blocks, but E. cava on natural reefs was longer. 23 7. Algal forests Different experimental designs used to restore macroalgal forests have been adopted using in situ and ex situ restoration strategies and considering different approaches to affect the . 7.1 Effect of conspecific adults and grazers A large-scale restoration intervention is illustrated in the study by Tamburello et al. (2019), in which an ex situ active restoration approach was used to test the effect of conspecific adult individuals, macrograzers and mesograzers on Cystoseira amentacea germlings. The initial conditions are represented by two different Cystoseira amentacea forests: one donor, characterized by >80% coverage, and the other to be restored, characterized by <10% coverage. For each condition, two locations were randomly identified, at a distance of about ten km from each other. For each location, then, two sites were chosen, distant from each other in the order of hundreds of meters, in which 18 experimental units were fixed (20×20 squares). Five tiles with laboratory-grown Cystoseira amentacea germlings, obtained from fertile fronds of donor populations, were placed in every experimental unit. In order to assess the effect of conspecific adult individuals on germlings, in half of the experimental units (9) of the sites to be restored adult individuals were transplanted, previously taken from half of the experimental units of the donor sites. To assess the influence of the macrograzers, cages were fixed in twelve experimental units for each site, half of which had openings of 4 cm on their side to allow access to mesograzers. Figure 15. Experimental design from Tamburello et al. (2019) 24 Six experimental units, instead, had no cages. For each condition, there were therefore three replicates. The work highlights the need to have nursery facilities in proximity to restoration sites to contribute significantly to the success of the intervention. One potential pitfall of the study was represented by a limited number of sites and the total duration of the study was limited to three months, too short a time to fully assess the success of the restoration actions. Survival of Cystoseira germlins in the field was not affected by the presence of conspecific adults, while the limitation of grazers appears to be crucial to the success of the experiment. 7.2 Combining active and passive restoration In the work of Medrano et al. (2020), the success of active revegetation techniques as a tool to promote forests of Treptacantha elegans reduced to barren of sea urchin was tested, to assess the potential success of active restoration techniques (in situ) combined with passive restoration and marine protection. The design for active revegetation initially includes 10 locations: six degraded barren grounds to revegetate (3 inside the NTZ and 3 outside), two dense T. elegans forests as reference sites of forest state (1 inside the NTZ and 1 outside), and two degraded barren grounds as reference sites of degraded habitat (1 inside the NTZ and 1 outside). Given the small size of the NTZ, they selected a limited number of reference areas. All 10 locations were selected in the 5–10 m depth range (5 inside the NTZ and 5 outside) covering the most diverse range of barren grounds’ sizes and local-scale variability given the limited size of the NTZ area (93 ha). In the six areas of barren to be restored, in addition to having carried out active restoration with the in situ technique, sea urchins were removed completely to eliminate the impact of herbivory. Once the successful technique was estimated in preliminary tests, in spring 2018, the in situ seedling technique was implemented within the 6 degraded barren grounds (3 inside the NTZ and 3 outside) and the 4 reference sites. Taking advantage of this experimental setup, they additionally tested three types of settlement collectors: stone plates (flat manufactured limestone), plots of original substrate (unmanipulated limestone substrate, mainly covered by the representa-tive algal assemblages), and plots of scraped and clearedsubstrate of approximately 25 cm2 area each. Six collectors of each type were randomly placed or delimited around the six seeding bags in each experimental site. Revegetation success was assessed 1 year later in the six barren grounds, but was only achieved after combining active with passive restoration strategies (see also Fig. 16 for details). Results encourage revegetation of barren grounds to shift from less productive habitats to complex T. elegans forests, highlight the potential of the combined passive and active restoration strategies, as well as the important role of marine reserves not only in conservation but also in ecological restoration. The outcomes of the papers are that active revegetation practices alone are insufficient torestore marine degraded grounds and effective whencombined with passive restoration practices. Revegetation of degraded barren grounds in marine protected areas triggers the recovery of marine 25 forests. More and well-enforced No-Take marine reserves areessential for management purposes and ecological restoration. Figure 16. Diagram showing the experimental design and the three-step T. elegans revegetation protocol (Medrano et al., 2020) 7.3 In situ vs ex situ restoration The study carried out by Verdura et al. (2019), focuses on the restoration of algal forests of Cystoseira barbata in sites where this species disappeared in the 80s specifically on the island of Menorca (Balearic Islands, NW Mediterranean) which has been a UNESCO biosphere reserve since 1993. The study includes two different techniques of non-destructive restoration in situ and ex situ. The experimental design of the in situ technique includes: a donor site, located in Fornells Bay where fertile apical branches of wild specimens have been collected, two sites affected by the restoration action, at Cala Teulera, where the species C. barbata was present until the 1970s and then was found extinct. Three control sites also located in the Fornells Bay where 20 squares (20 x 20) randomly placed, were calculated the densities and the distribution of the dimensional structure of 32 to study the sediment dynamics therein. The “restored” area consisted of a mono-specific stand of well-established B. maritimus. The “mudflat” area was an unvegetated area running adjacent to the coast between the two extreme extents of the vegetated areas. Monitoring: Sediment deposition and settlement were sampled on 16 occasions over a year period, on each occasion sampling took place at 12 permanent points. Sampling was carried out four times in each season, this allowing a correct quantification of the temporal variability within each season. The four sampling occasions in each season spanned across a spring to neap tidal change, beginning on a spring tide. On each occasion, sediment deposition and settlement traps were deployed during low tide and remained exposed for two flood tide events, after which they were retrieved. The next sampling occasion took place the following day at low tide, this was repeated four times each season. Sediment elevation measurements were taken once per season, following the sediment deposition and settlement sampling. Sampling aimed to capture information on the sediment regime within the study area. Main results: It is not yet clear how the current sedimentary interactions will affect the long-term survival and evolution of this site, specifically the persistence of the restored area. Although the data suggests the site of restoration is experiencing a different biogeomorphic situation compared to the bare mudflat, which is assumed to reflect its previous state, this situation does not yet mirror its natural counterpart. This study illustrates the subtleties of saltmarsh development and feedbacks, highlighting the importance of site selection in restoration activities. The “established” appearance of the restored planted saltmarsh disguises its possible vulnerability in dynamic sedimentary interactions leading to net erosion across the site. This study highlighted the possible vulnerability of restored areas, which may visually appear established, but whose processes do not indicate a robust ecosystem; specifically, net elevation loss during the study. The delay for restoration to achieve comparable ecological functionality to natural areas is important to consider when carrying out restoration projects and may require additional effort once established to ensure their persistence. Such information could advance approaches and maximize the delivery of conservation objectives which are crucial to prevent the loss of these systems. 9. Final considerations Ecosystem restoration involves innovation and experimentation, and restoration activities often result in surprises and setbacks. Because of this, it is often necessary to conduct initial experimentation to support decision-making (e.g. choice of species and spacing) or to install and test alternative treatments during the project to enable adaptive management (FAO et al., 2023). However, our description has the aim to stress the importance of adopting rigorous experimental design in restoration and the caution needed in analysing and extrapolating results from simplistic approaches. Evaluating the effects of anthropogenic activities on organisms and environments is 33 fundamentally important to ensure adequate management and conservation of natural environments. Logical and methodological criteria underlying evaluation procedures are critical for detecting environmental impacts, or the effectiveness of mitigation efforts (Benedetti-Cecchi, 2004). A key step in this procedure involves the identification of an appropriate sampling designs, capable of measuring a specific effect and to tease it apart from background natural variation. The same logic has to be applied also to restoration which is a human intervention into ecological systems. Restoration should be perceived as an experimental treatment thus following the same criteria of any ecological experiment. Restoration, for its natura, is based on manipulative experiments. This is very relevant also for inferring mechanisms, but the only limit is that often manipulative studies are limited to relatively small spatial scales, few locations, and short durations, limiting the potential for generalities. Coordinated experiments can overcome this limit. In the analyses of the literature, several issues emerge: very often just one control is included in the restoration intervention, and, also, many experiments are pseudoreplicated. This means that an inappropriate source of variability is used as reference to examine the effect of a factor. As an example, analysing effectiveness of restoration along an environmental gradient in which selected sites are also featured by different substrates translates into the consequences that potential differences at the end of the experiment in restoration effectiveness might be linked to the gradient but also to the different substrates, without the possibility to tease apart the effect of one factor from the effect of the other. Our analyses show that despite there is an increasing number of studies improving the spatial and temporal scales of the interventions, still inconsistent methodologies and other idiosyncrasies among studies limit the direct comparisons needed to improve general knowledge. Most studies concur that current and historical presence, site local condition assessment and choice of the actions to be carried out are central in all restoration interventions. In addition: 1Species selection and donor populations are critical elements in restoration whose selection has to be addressed through a solid scientific knowledge. 2Complex experiments dealing with multiple factors assessing the effects, for instance, of substrate, biotic interactions, environmental variables are important to advance restoration. However, it is necessary to specify the relationships among factors (crossed vs nested) and the nature of each factor (fixed or random). This issue is most of times not addressed in the studies (but see for exceptions Guarnieri et al., 2020; Medrano et al., 2020). 3Compare replicated restoration sites to multiple controls, before and after restoration interventions, on both control and restored sites. The need of multiple controls has been largely stressed and should be always clear that reference areas must have the undisturbed assemblages and similar environmental features as disturbed areas that we want to restore. 34 4Repeating measurements in time, to estimate temporal variability of response variables is also very important and long term experiments should be a priority. In conclusion, considering the effort and the costs invested in restoration, rigorous experiments are a priority. In this framework, it is important starting any restoration intervention to develop partnerships among scientists, practitioners, and the local community, to ensure that the project receives an appropriate level of scientific and practical advice and assistance to optimize its success and relevance. 10. References Bayraktarov E, Stewart-Sinclair PJ, Brisbane S, Bostrom-Einarsson L, Saunders MI, Lovelock CE, Possingham HP, Mumby PJ, Wilson KA. (2019). Motivations, success, and cost of coral reef restoration. Restoration Ecology 27(5), 981-991. Benedetti-Cecchi, L. (2004). Experimental design and hypothesis testing in ecology. Biologia Marina Mediterranea. 11. 407-455. Chen W, Ge ZM, Fei BL, Zhang C, Liu QX, Zhang LQ. (2017). Soil carbon and nitrogen storage in recently restored and mature native Scirpus marshes in the Yangtze Estuary, China: Implications for restoration. Ecological Engineering 104, 150-157. Davis and Short (1997). Restoring eelgrass, Zostera marina L., habitat using a new transplanting technique: the horizontal rhizome method. Aquatic Botany 59, 1-15. De La Fuente G, Chiantore M, Asnaghi V, Kaleb S, Falace A. (2019). First ex situ outplanting of the habitat-forming seaweed Cystoseira amentacea var. stricta from a restoration perspective. PeerJ, 7:e7290. doi: 10.7717/peerj.7290. Derrenbacker JA, Lewis RR. (1982). Seagrass habitat restoration in Lake Surprise, Florida Keys. In: Stoval, RH (ed.) Proceedings Ninth Annual Conference on Wetlands Restoration and Creation, May 20-21, Hillsborough Community College, Tampa, FL, 132-154. FAO, SER & IUCN CEM. (2023). Standards of practice to guide ecosystem restoration. A contribution to the United Nations Decade on Ecosystem Restoration. Summary report. Rome, FAO. doi: 10.4060/cc5223en Fonseca MS. (1994). A Guide to Planting Seagrasses in the Gulf of Mexico (Publication No. TAMUSG-94-601). Texas A&M University Sea Grant College Program. Galveston, TX. Fraschetti, S., McOwen, C., Papa, L., Papadopoulou, N., Bilan, M., Boström, C., ... & Guarnieri, G. (2021). Where is more important than how in coastal and marine ecosystems restoration. Frontiers in Marine Science, 8, 626843. Fredriksen S, Filbee-Dexter K, Norderhaug KM, Steen H, Bodvin T, Coleman MA, Moy F, Wernberg T. (2020). Green gravel: a novel restoration tool to combat kelp forest decline. Scientific Reports 10, 1-7. Gann, G. D., McDonald, T., Walder, B., Aronson, J., Nelson, C. R., Jonson, J., ... & Dixon, K. (2019). International principles and standards for the practice of ecological restoration. Restoration Ecology, 27(S1), S1-S46. Guarnieri G, Bevilacqua S, Figueras N, Tamburello L, Fraschetti S. (2020). Large-scale sea urchin culling drives the reduction of subtidal barren grounds in the Mediterranean Sea. Frontiers in Marine Science 7, 519. doi: 10.3389/fmars.2020.00519. Hunsucker K, Melnikov A, Gilligan M, Gardner H, Erdogan C, Weaver R, et al. (2021). Cathodically protected steel as an alternative to plastic for oyster restoration mats. Ecological Engineering 164, 106210. doi: 10.1016/j.ecoleng.2021.106210. Hurlbert, S. H. (1984). Pseudoreplication and the design of ecological field experiments. Ecological monographs, 54(2), 187-211. Lange T, Oncken NS, Svane N, Steinfurth RC, Kristensen E, Flindt MR. (2022). Large-scale eelgrass transplantation: a measure for carbon and nutrient sequestration in estuaries. Marine Ecology Progress Series 685, 97-109. doi: 10.3354/meps13975. 35 Leinaas HP, Christie H. (1996). Effects of removing sea urchins (Strongylocentrotus droebachiensis): stability of the barren state and succession of kelp forest recovery in the east Atlantic. Oecologia 105, 524-536. Medrano A, Hereu B, Cleminson M, Pagès-Escolà M, Rovira G, Solà J, et al. (2020). From marine deserts to algal beds: Treptacantha elegans revegetation to reverse stable degraded ecosystems inside and outside a No-Take marine reserve. Restoration Ecology 28, 632-644. doi: 10.1111/rec.13123. Oyamada K, Tsukidate M, Watanabe K, Takahashi T, Isoo T, Terawaki T. (2008). A field test of porous carbonated blocks used as artificial reef in seaweed beds of Ecklonia cava. In Nineteenth International Seaweed Symposium, 413-418. Pickerell CH, Schott S, Wyllie-Echeverria S. (2005). Buoy-deployed seeding: demonstration of a new eelgrass (Zostera marina L.) planting method. Ecological Engineering 25, 127-136. doi: 10.1016/j.ecoleng.2005.03.005. Tamburello L, Papa L, Guarnieri G, Basconi L, Zampardi S, Scipione MB, et al. (2019). Are we ready for scaling up restoration actions? An insight from Mediterranean macroalgal canopies. PLoS One 14, e0224477. doi: 10.1371/journal.pone.0224477. Taylor BW, Paterson DM, Baxter JM (2019) Sediment dynamics of natural and restored Bolboschoenus maritimus salt marsh. Frontiers in Ecology and Evolution 7, 237. Tedford, K. N., & Castorani, M. C. (2022). Meta-analysis reveals controls on oyster predation. Frontiers in Marine Science, 9, 1055240. Terrados J, Marín A, Celdrán D. (2013). Use of Posidonia oceanica seedlings from beach-cast fruits for seagrass planting. Botanica Marina 56, 185-195. doi: 10.1515/bot-2012-0200. Underwood, A. J. (1993). The mechanics of spatially replicated sampling programmes to detect environmental impacts in a variable world. Australian Journal of ecology, 18(1), 99-116. Verdura J, Sales M, Ballesteros E, Cefalì ME, Cebrian E. (2018). Restoration of a canopy-forming alga based on recruitment enhancement: methods and long-term success assessment. Frontiers in Plant Science 9, 1832. doi: 10.3389/fpls.2018.01832. White JM, Buhle ER, Ruesink JL, Trimble AC. (2009). Evaluation of Olympia oyster (Ostrea lurida carpenter 1864) status and restoration techniques in Puget Sound, Washington, United States.