Abstract
In the marine environment, greening of grey infrastructure (GGI) is a rapidly growing field that attempts to encourage native marine life to colonise marine artificial structures to enhance biodiversity, thereby promoting ecosystem functioning and hence service provision. By designing multifunctional sea defences, breakwaters, port complexes and offshore renewable energy installations, these structures can yield myriad environmental benefits, in particular, addressing UN SDG 14: Life below water. Although GGI has shown great promise and there is a growing evidence base, there remain many criticisms and knowledge gaps, and some feel that there is scope for GGI to be abused by developers to facilitate harmful development. Given the surge of research in this field in recent years, it is timely to review the literature to provide an update on the state of the art of the field in relation to the many criticisms and identify remaining knowledge gaps. Despite the rapid and significant advances made in this field, there is currently a lack of science and practice outside of academic sectors in the developed world, and there is a collective need for schemes that encourage intersectoral and trans-sectoral research, knowledge exchange and capacity building to optimise GGI in the pursuit of contributing to sustainable development.
The rapid growth of the human population (8 billion in 2022) has driven the intertwined global crises of climate change and biodiversity loss. Much of this growth and associated development has been in the coastal zone, leading to proliferating land-claim and construction of so-called grey infrastructure along the coastline and increasingly offshore into the oceans (Bugnot et al., 2021; Firth et al., 2013, Box 1, Figure 1), with many major iconic megadevelopments likely inspired by the construction of the Palms, Dubai (in the sense of the ‘palm effect’, Box 2). In many parts of the world, densely populated land-scarce regions (e.g. Penang, Figure 2) may be approaching a form of Malthusian trap (Malthus, 1798) – where population growth is outpacing land availability and thus expanding built infrastructure into the sea. Such situations may be viewed as valuable for concentrating people in urban areas, but they also present major challenges in terms of resource management and environmental impacts.
Figure 2. Penang Island: population growth outpacing land availability? (a) Map showing natural terrestrial habitats (light green = forest; dark green = mangrove), urbanised (grey) and reclaimed land (brown = already reclaimed, red = proposed reclamation). (b) Bars illustrating population density (number of people/km2, 2020 World Bank data) and lines illustrating population density growth rate (1961–2020) of Penang Island compared with the seven most densely populated countries. (c) Image of tall buildings condensed into a narrow coastal strip between mountains and sea (photo credit Hong Chern Wern) (source: (a) and (b) redrawn from Chee et al. (2017))
The resultant urbanised and degraded environments have been likened to novel ecosystems – hybrids of nature and technology that have been irrevocably deflected from their natural trajectories (Bulleri et al., 2020). The United Nations Decade of Ocean Science and Sustainable Development provides impetus for humans to reverse declines in ocean health and to use science to facilitate sustainable development. Sustainable development is typically managed through the environmental impact assessment and/or the mitigation hierarchy processes (Green, 1979). The mitigation hierarchy is widely accepted as the current best practice for achieving sustainable development (CSBI, 2015), wherein practitioners sequentially seek to limit negative environmental impacts and achieve net biodiversity gain (Figure 3). ‘Offsetting’ is intended only as a last resort for developers seeking to compensate for unavoidable damage, after having applied all other steps. With current projections for human population increase, further coastal development is inevitable. Consequently, far greater attention should be given to developing new tools and improving implementation of the mitigation hierarchy (Bigard et al., 2020; Pioch et al., 2017).
Figure 3. Four sequential steps of the mitigation hierarchy: avoid, minimise, restore/rehabilitate and offset/other conservation actions. The goal is to achieve no net loss or net gain of biodiversity (not just species richness) (above the green line) compared with the pre-disturbance landscape. Management actions to the right of the red line should be perceived as a last resort and only be implemented when all other steps have been applied to maximum effect (source: figure and text adapted from CSBI (2015))
In coastal environments, ecological engineering (eco-engineering) can be used for greening of grey infrastructure (GGI) by incorporating ecological knowledge in designing and engineering multifunctional marine infrastructure (Bergen et al., 2001). The goal is to encourage native marine life to colonise artificial structures to enhance biodiversity, thereby promoting ecosystem functioning and hence service provision (Figure 4). GGI has emerged as a promising tool for achieving biodiversity and environmental benefits (addressing UN SDG 14: Life below water) with multifunctional infrastructure, such as sea defences, breakwaters, port complexes and offshore renewable energy installations (see Airoldi et al., 2021; Evans et al., 2021; O'Shaughnessy et al., 2020a; Strain et al., 2018a for global reviews).
Figure 4. Continuum of marine eco-engineering solutions between conventional grey engineering and natural systems. The blue area shows where GGI sits along this continuum (source: Naylor et al. (2023); adapted from Naylor et al. (2020) and Suedel et al. (2022))
Box 1. The palm effect
For millennia, people have been fascinated by the idea of artificial islands. In Neolithic Ireland and Scotland, people lived on defensive wooden islands in lakes called crannógs (Wood-Martin, 1886). Construction of the ancient island site of Nan Madol, Micronesia began as early as the eighth century (Athens, 1980). Indeed, Thomas More's (1516) Utopia was an imagined artificial island in which everything was perfect; a concept that has transcended time with many recent artificial islands being designed and promoted as utopian paradises where the global elite can live, work and play in eco-friendly, green and smart cities (Grydehøj and Kelman, 2016). More recently, this has even gone as far as the concept of ‘seasteading’; with claims of how floating nations ‘will restore the environment, enrich the poor, cure the sick and liberate humanity from politicians’ (Quirk and Friedmann, 2017). Nowhere exemplifies this concept better than Dubai, with its myriad of artificial islands and where one can own their very own island in the shape of their country of choice.
More than 479 artificial islands are known to exist (Bugnot et al., 2021), ranging in scope and design from simple rounded and fractal designs (Chee et al., 2017) to national symbols (e.g. The Pearl and Qatar), and even maps of the whole world (i.e. The World and Dubai). Incidentally, the construction of Dubai's The Universe was halted due to the 2008 global economic crash (El-Sheshtawy, 2010). Jackson and della Dora (2011) describe contemporary artificial islands as ‘metageographical terra-forms that normalise and naturalise people's expectations, knowledge and interactions with the world’. Indeed, there is no doubt that the Palm Islands have inspired many of these designs through what is referred to here as ‘the palm effect’. The trend first caught on in the Arabian Gulf States but has since spread to other regions (Figure 1). For instance, the developers behind the plans for a megadevelopment on the Malaysian island of Langkawi were quoted to claim that the development ‘would be akin to that in Dubai’, encouraging tourism and investment (The Star, 2018). Similarly, the developers behind Lagos’ Eko Atlantic refer to it as ‘Africa's Dubai’ (Eko Atlantic, 2020). Jackson and della Dora (2009) further point out that artificial islands are the ‘new cultural icons’, are a ‘must’ for aspiring ‘global cities’, or for countries at the fringes of capitalist economy striving to get international attention. All artificial islands will inevitably be low-lying and potentially vulnerable to future water intrusion by storms or sea level rise. In the face of pervasive global climate change, rising and stormier seas and population increases, these hybrids of nature and technology are likely to become even more pervasive, perpetuating the palm effect.
Figure 1. The palm effect – coastal artificialisation on the Arabian Peninsula and Asia: (a) The Palms and The World, Dubai, UAE; (b) The Pearl, Qatar; (c) Durrat al Bahrain, Bahrain; (d) Center Point of Indonesia (representing a Garuda (eagle), Sulawesi, Indonesia; (e) Qinhuangdao, Hebei, China; (f) Ocean Flower Island, Hainan, China). The numbers in the bottom right indicate the year that construction started (all images from google earth)
Box 2. Artificialisation of the global coastline
Coastal development is the anthropogenic change in a seascape through the construction of artificial structures within sight of the coastline. Firth et al. (2020) stated that Asian and Middle Eastern countries have constructed some of the most ambitious and iconic land reclamation projects. Since then, several empirical large-scale and even global studies have emerged that can attest to this. For instance, Bugnot et al. (2021) found that China was responsible for 40% of global construction in the marine environment, followed by South Korea (10% of global) and the Philippines (8% of global). Sengupta et al. (2020) found that, since 1988, >700 km2 of land reclamation has occurred in just eight Asian cities, with >35% of this happening in Shanghai alone. Similarly, Chee et al. (2023) found that, since 1991, >82 km2 has been or will be reclaimed in Malaysia. Penang Island, Malaysia, is one of the most densely populated and fastest-growing urban settlements globally (Chee et al., 2017) (see Figure 2). Bounded by the Penang Strait to the east and hills to the west, space for further development is severely limited. The solution to tackle this social problem has been to build upwards (i.e. construction of high-density, high-rise buildings) and outwards (i.e. land reclamation and the construction of five large artificial islands).
Firth et al. (2020) also suggested that coastal African countries that are experiencing rapid population growth are the most vulnerable to future habitat loss. While Claassens et al. (2022) only found that 2.9% of the 3000 km South African coastline is armoured, this number is likely to be much higher given the large number of informal settlements. Further armouring is likely in the future due to growing urban populations and additional pressure from Operation Phakisa to transform the South African coastal environment to unlock the ‘blue economy’ (Loureiro et al., 2022). With 10 million m2 of reclaimed land protected by an 8.5 km long seawall, Nigeria is currently undertaking one of the world's largest land reclamation and construction projects (Ajibade, 2017). Once complete, Eko Atlantic City will be the size of Manhattan (Eko Atlantic, 2022). Incidentally, whilst its name uses an old term for the Nigerian city of Lagos (‘Eko’) it may incite perceptions of eco-friendly development. There is potential for Eko Atlantic to act as a regional catalyst (see Box 1 on the palm effect) for similar coastal megadevelopment in other rapidly growing African countries. Collectively, all this evidence reinforces the assertions by Firth et al. (2020) that continued artificialisation of the coastline is inevitable, particularly in Africa and the Middle East.
In the first global estimation of the footprint of marine construction, Bugnot et al. (2021) reported that 32 000 km2 of seafloor had been reclaimed. In their audit of artificial shoreline extent, Sempere-Valverde et al. (2023) estimated that ∼16% of the global coastline was armoured, with most development in lagoons, estuaries and bays, and also in regions characterised by middle-to-high income countries. Many estimates of coastal armouring exist for developed nations (Table 3), but the figures are less robust for developing nations as many coastal settlements are ‘informal’ and coastal protection and land reclamation practices are haphazard, opportunistic and unregulated (Palmer et al., 2010). Most of the predicted population growth across the globe by 2050 is expected to occur in low- and middle-income countries (Cohen, 2003; Nieves et al., 2017) and, by 2030, an estimated 38% of the global population will live in the near-coast zone (the area located <100 km from the coast and below 100 m elevation (Kummu et al., 2016)). While this zone only occupies 9% of global land area, it supports 62% of cities with over 5 million inhabitants (Firth et al., 2016a) and 42% of global gross domestic product (Kummu et al., 2016). Many regions already claim that >50% of their coastlines are artificial (Table 3), and it seems inevitable that coastal armouring will continue apace.
Firth et al. (2020) highlighted a range of limitations and unknowns, noting that GGI interventions could be assigned to one of three categories: Trojan horses, projects that cause environmental damage either through deliberate or misguided intent; fig leaves, projects that merely cover up or deflect attention from environmental damage caused by the development; and laurel wreaths, win–win projects with measurable benefits for humans and nature (see e.g. Box 3). While laurel wreaths are more challenging to achieve, this is not to say that all GGI interventions that cannot be classified as ‘laurel wreaths’ are a deliberate ‘greenwashing’ attempt, as the field is still relatively new and requires more guidance and experimentation.
Box 3. Examples of Trojan horses, fig leaves and laurel wreaths from artificial reefs (ARs), rigs-to-reefs and GGI of artificial structures.
Firth et al. (2020) highlighted a range of limitations and unknowns, noting that GGI experiments could be assigned to one of three categories: Trojan horses, projects that cause environmental damage either through deliberate or misguided intent; fig leaves, projects that merely cover up environmental damage caused by the development; and laurel wreaths, win–win projects with measurable benefits for humans and nature. The authors suggested that much could be learnt from the AR literature, a more established discipline subject to intense critical evaluation (e.g. Castello and Bartholomew et al., 2022; Komyakova et al., 2021; Pioch et al., 2011; Reis et al., 2021; Salaün et al., 2022; Tickell et al., 2019; Vivier et al., 2021b). Using ARs, rigs-to-reefs (decommissioned oil and gas structures re-purposed as ARs) and GGI on coastal infrastructure, illustrations of Trojan horses, fig leaves and laurel wreaths are provided.
ARs can simultaneously benefit nature and humans. ARs can increase fish production, including commercially important species, by providing refuge, breeding sites and increased feeding opportunities (Cresson et al., 2014; Folpp et al., 2020; Pondella et al., 2002; Roa-Ureta et al., 2019). ARs can provide additional hard-bottom habitat as attachment sites for commercially important bivalves (Goelz et al., 2020) and important habitat-forming species like macroalgae and corals (Campos et al., 2020; Higgins et al., 2022). ARs can be used to protect important habitats from trawling (Relini et al., 2007). ARs built for tourism and diving can also support diverse biota (Jackson et al., 2005; Smith et al., 2022) and relieve pressure on natural reefs (Belhassen et al., 2017; Firth et al., 2023b).
An AR was created to compensate mudflat loss following the creation of a dredged material disposal site. Burton et al. (2002) found that the annual secondary production/unit area from sessile invertebrates at the offset AR was 11–67 times higher than natural soft-sediment habitat, but total annual secondary production was 1.3–7.6 times lower. The authors concluded that the AR improved secondary production, but not enough reef was created to fully offset the lost habitat. The Palm Jebel Ali was built on natural coral habitat (Burt et al., 2008), and research found that coral diversity was lower and fish communities were different on breakwaters, although coral per cent coverage was higher (Burt et al., 2009). ARs were created in Qatar as transplantation sites for coral that would be destroyed by a new oil pipeline, and initial coral survivorship results were promising (Deb et al., 2014). However, survivorship of transplanted coral is generally low, with only 20% of studies reporting survivorship over 90% (Boström-Einarsson et al., 2020).
Some ARs are simply ocean dumping (Chou, 1997), either inadvertently if the AR fails to support much life or through the deliberate disposal of waste material at sea under the guise of ‘reef creation’ (e.g. The Osborne Reef, Florida, Figure 6). Scrap materials used as ‘reefs’ may have chemical pollutants associated with them (Dodrill et al., 2011; Gaylarde et al., 2021), and they may move during storms and impact natural habitats (Morley et al., 2008; Sherman and Spieler, 2006; Turpin and Bortone, 2002). ARs attract both fish and fishers, and can increase catch rates in the short term but contribute to regional overfishing in the long term (Simon et al., 2011; Watanuki and Gonzales, 2006). ARs can also facilitate the spread of invasive hard-bottom species (Adams et al., 2014; Airoldi et al., 2015).
The authors of this paper do not believe that there are examples of laurel wreaths when it comes to the decommissioning of oil and gas platforms as ARs (so-called ‘rigs-to-reefs’). The negative impacts of the placement of such large structures on the environment are too great for any compensatory efforts to be viewed as positive.
All retrospective rigs-to-reefs are fig leaves. It is undeniable that they can provide substantial habitat for a diversity of marine life (e.g. Friedlander et al., 2014; Gass and Roberts, 2006; Sammarco et al., 2012). However, this is often based on descriptive studies that have no formal comparison with natural habitats, and the perceived benefits are often weighted more towards humans than nature. For example, the developers save money from not having to remove the structure (Bressler and Bernstein, 2015) and fishing communities and marine recreationists may benefit from enhanced activities following decommissioning (e.g. Ajemian et al., 2015).
Rigs-to-reefs can become Trojan horses through a range of different pathways. For example: (a) a ‘successfully’ decommissioned rig that has been converted to a ‘reef’ and continues to function as a habitat and provide human recreational activities can retrospectively cause further harm if it suffers a leak, leading to the release of oil into the environment; (b) assertions that offshore platforms support better fisheries than natural habitats (Claisse et al., 2014) may prospectively cause environmental damage through influencing governments to relax regulations or develop policies (e.g. the US National Fishing Enhancement Act (Public Law 98-623, Title II) (BSSE, 2018)), which may play a role in facilitating increased oil production and the proliferation of artificial structures on the seabed.
Disused docks that are redeveloped for commercial/residential purposes. As part of the process, water quality needs to be improved. The encouragement of native mussels increased natural biofiltration which improved water quality (Figure 6(c); Wilkinson et al., 1996) and led to colonisation of diverse assemblage of marine life (Allen et al., 1995; Firth et al., 2023b; Hawkins et al., 1999). Artificial water circulation further enhanced water movement and quality.
A boulder on a breakwater that is located in a sedimentary environment is replaced with a precast concrete habitat-enhancement unit (Firth et al., 2014). The unit encourages colonisation of greater diversity of reef taxa than the adjacent boulders. The enhancement of rocky-reef taxa is only compensation for the loss of sedimentary habitat and biodiversity in the footprint of the structure.
A shore-parallel breakwater is enhanced through the addition of novel habitats such as pits or rockpools (e.g. Evans et al., 2015). Hypothetically, the habitats become dominated by invasive species that prevent establishment of native taxa. This could be exacerbated if the breakwater is outside of the known range for the invasive species, thus facilitating spread through the steppingstone concept (Airoldi et al., 2015; Mineur et al., 2012). This could equally be true if the eco-engineering interventions facilitated establishment and spread of parasites, pathogens, pests or other harmful organisms that have been demonstrated to be spread by artificial structures (e.g. Duarte et al., 2012; Ishii and Katsukoshi, 2010; Lo et al., 2008; Vila et al., 2001; Villareal et al., 2007). To date, no such reports have been made from eco-engineering trials, but researchers are urged to monitor eco-engineered habitats over longer timescales and report on any negative impacts. An even more serious example might be where a local government views a planning application for the construction of an artificial island more favourably based on the promise of the inclusion of eco-engineering. Even if the eco-engineering goals are perceived to be successful later down the line, it was partly responsible for the development getting approved which led to the destruction of natural coastal habitats.
Figure 6. (a) Trojan horse: the Osborne reef constructed out of tyres, Florida, USA (photo credit by Navy Combat Camera Dive Ex-East); (b) fig leaf: the placement of disused plane in the sea as an AR, Malaysia (photo credit Quek Yew Aun), (c) laurel wreath: mussels supporting high biodiversity and filtering water on the walls of the Albert Dock, Liverpool, UK (photo credit Louise Firth)
|
| (b) Regional and local | % Artificial | Source | |
|---|---|---|---|
| Region | Location | ||
| Middle East | Dubai, UAE | 130 | Burt et al. (2013) |
| Australasia | Great Barrier Reef, Australia | 9.4 | Waltham and Sheaves (2015) |
| Adelaide, Australia | 53 | Floerl et al. (2021) | |
| Gladstone, Australia | 23 | Floerl et al. (2021) | |
| Darwin, Australia | 10 | Floerl et al. (2021) | |
| Hobart, Australia | 22 | Floerl et al. (2021) | |
| Geelong, Australia | 56 | Floerl et al. (2021) | |
| Sydney Harbour, Australia | 50 | Chapman (2003) | |
| Sydney Harbour, Australia | 85 | Strain et al. (2019) | |
| Port Phillip Bay, Melbourne, Australia | 25 | Strain et al. (2019) | |
| Various estuaries in New South Wales, Australia | 3.8–46 | Creese et al. (2009) | |
| Derwent Estuary, Hobart, Tasmania, Australia | 19.2 | Strain et al. (2019) | |
| Waitemata Harbour, Auckland, New Zealand | 20.6 | Strain et al. (2019) | |
| Auckland, New Zealand | 47 | Floerl et al. (2021) | |
| Dunedin, New Zealand | 77 | Floerl et al. (2021) | |
| Tauranga, New Zealand | 20 | Floerl et al. (2021) | |
| Wellington, New Zealand | 77 | Floerl et al. (2021) | |
| Whangarei, New Zealand | 25 | Floerl et al. (2021) | |
| Nelson, New Zealand | 50 | Floerl et al. (2021) | |
| Bluff, New Zealand | 21 | Floerl et al. (2021) | |
| Lyttelton, New Zealand | 25 | Floerl et al. (2021) | |
| Picton, New Zealand | 71 | Floerl et al. (2021) | |
| Napier, New Zealand | 93 | Floerl et al. (2021) | |
| Westport, New Zealand | 51 | Floerl et al. (2021) | |
| New Plymouth, New Zealand | 65 | Floerl et al. (2021) | |
| Opua, New Zealand | 5 | Floerl et al. (2021) | |
| Greymouth, New Zealand | 65 | Floerl et al. (2021) | |
| Europe | Emilia Romagna, Italy | 60 | Bacchiocchi and Airoldi (2003) |
| Belfast, UK | 88 | Floerl et al. (2021) | |
| Ravenna Port, Italy | 70 | Strain et al. (2019) | |
| Varna Bay, Bulgaria | 10 | Stancheva et al. (2011) | |
| Plymouth Sound, UK | 33 | Knights et al. (2016) | |
| French Mediterranean | 11 | http://www.medam.org | |
| Weser Estuary, Germany | 60 | Wetzel et al. (2014) | |
| Santander Harbour, Spain | 42 | Strain et al. (2019) | |
| Galway Bay, Ireland | 7 | Firth et al. (2016b) | |
| Galway City | 45 | Firth et al. (2016b) | |
| Cardiff, UK | 87 | Floerl et al. (2021) | |
| Edinburgh, UK | 89 | Floerl et al. (2021) | |
| Africa | eThekwini, South Africa | 11 | Stretch (2012) |
| Asia | Zhoushan Islands, China | 65 | Zhang et al. (2014) |
| Xiamen Harbour, China | 90 | Strain et al. (2019) | |
| Victoria Harbour, Hong Kong | 95 | Lam et al. (2009) | |
| Penang Island (east coast), Malaysia | 88 | Chee et al. (2017) | |
| Balneário Camboriú, Brazil | 36 | Piatto and Polette (2012) | |
| Okinawa Island, Japan | 63 | Masucci and Reimer (2019) | |
| Keelung Harbour, Taiwan | 80 | Strain et al. (2019) | |
| Americas | Bermuda | 15 | Meyer et al. (2015) |
| Arraial do Cabo Harbour, Brazil | 5 | Strain et al. (2019) | |
| Cartagena, Columbia | 12 | Stancheva et al. (2011) | |
| Guam | 15 | NOAA (2012) | |
| American Samoa | 7 | NOAA (2012) | |
| Alabama, USA | 11 | NOAA (2012) | |
| Alaska, USA | 2 | NOAA (2012) | |
| California, USA | 12 | Griggs (1998) | |
| California, USA | 14 | NOAA (2012) | |
| County comparisons, California, USA | 0.5–61 | Hanak and Moreno (2012) | |
| Californian cities of Long Beach, Seal Beach, San Clemente, Oceanside | >70 | Dugan et al. (2011) | |
| Connecticut, USA | 18 | NOAA (2012) | |
| New Haven, Connecticut, USA | 74 | Floerl et al. (2021) | |
| Delaware, USA | 12 | NOAA (2012) | |
| Florida, USA | 21 | Florida DEP (1990) | |
| Florida, USA | 20 | NOAA (2012) | |
| Georgia, USA | 1 | NOAA (2012) | |
| Hawaii, USA | 11 | NOAA (2012) | |
| Oahu, Hawaii, USA | 26 | Campbell et al. (1988) | |
| Louisiana, USA | 3 | NOAA (2012) | |
| Maryland, USA | 15 | NOAA (2012) | |
| Baltimore, Maryland, USA | 71 | Floerl et al. (2021) | |
| Massachusetts, USA | 11 | NOAA (2012) | |
| Boston Harbor, Massachusetts, USA | 58 | Strain et al. (2019) | |
| Mississippi, USA | 12 | NOAA (2012) | |
| New Hampshire, USA | 4 | NOAA (2012) | |
| New Jersey, USA | 17 | NOAA (2012) | |
| New Jersey, USA | 17 | Lathrop and Love (2007) | |
| New York, USA | 24 | NOAA (2012) | |
| North Carolina, USA | 8 | NOAA (2012) | |
| Oregon, USA | 6 | Dugan et al. (2011) | |
| Oregon, USA | 4 | NOAA (2012) | |
| Portland, Oregon, USA | 77 | Floerl et al. (2021) | |
| Pennsylvania, USA | 36 | NOAA (2012) | |
| Puerto Rico, USA | 10 | NOAA (2012) | |
| Rhode Island, USA | 14 | NOAA (2012) | |
| Providence, Rhode Island, USA | 62 | Floerl et al. (2021) | |
| South Carolina, USA | 1 | NOAA (2012) | |
| Texas, USA | 15 | NOAA (2012) | |
| Virginia, USA | 8 | NOAA (2012) | |
| Norfolk, Virginia, USA | 56 | Floerl et al. (2021) | |
| Virgin Islands, USA | 8 | NOAA (2012) | |
| Washington, USA | 6 | NOAA (2012) | |
| Duwamish Estuary, Washington, USA | 66 | Morley et al. (2012) | |
| Puget sound, Washington, USA | 30 | Dugan et al. (2011) | |
| Vancouver, Canada | 75 | Floerl et al. (2021) | |
| Nanaimo, Canada | 34 | Floerl et al. (2021) | |
| Halifax, Nova Scotia, Canada | 59 | Floerl et al. (2021) | |
To prevent GGI from being used for greenwashing to facilitate coastal development, Firth et al. (2020) made several recommendations. Given the upsurge in publications since 2020, Section 2 reviews how scientists and practitioners have addressed these limitations and recommendations. It provides a comprehensive update on the state of the art and identifies remaining knowledge gaps. Section 3 addresses how GGI can be used for greenwashing purposes. Section 4 provides a critical summary of the importance of building the evidence base. Finally, Section 5 provides some conclusions and makes the case for the need for a paradigm shift of current funding strategies and research programmes to encourage the development, implementation and translation of GGI science at global scales.
Firth et al. (2020) made several criticisms and recommendations on how to improve the science and evidence to optimise GGI approaches to and prevent it from deliberately or unknowingly facilitating harmful development. This paper expands on and provides a summary of the criticisms and recommendations (Figure 5). To address the major concern that past experiments have been relatively limited in scope, designing bigger and better experiments is recommended. The limited number and scope of response variables typically measured can be addressed by going beyond simple biodiversity metrics. Furthermore, measuring ecosystem functioning along with socio-economic metrics of service delivery should be incorporated. As most experiments are limited in environmental scope, being conducted in a single location or over short timeframes, revisiting experiments beyond the lifetime of the original project, repeating experiments in the same location where possible, and replicating them under different environmental conditions will yield invaluable additional information about how GGI interventions do under different environmental scenarios. While it may be impossible to fully address the unpredictability of the natural environment, reporting on failures and unintended outcomes of experiments (both positive and negative) will inform better design and save money. Much research to date has been led by the scientific community, and not the practitioner. Science and practice must progress in tandem through collaboration and co-design to ensure that experiments are appropriately scaled-up to ‘real-world’ scenarios while testing their efficacy. Finally, the criticism that there is limited guidance available to practitioners can be addressed through formulating appropriate policies and guidance. Often, practice proceeds ahead of science, especially in large-scale projects. Hence developers, planners, architects and engineers need to be encouraged to engage with scientists to set targets and objectively measure outcomes of designs and interventions intended to lead to environmental mitigation. Such an approach will inform any adjustments or fine-tuning required post-construction or commissioning as well as future developments.
Figure 5. Building the evidence base to optimise GGI. The limitations and unknowns are located at the base of the columns. Suggestions and recommendations for a way forward are in the body of the columns (concepts adapted and expanded on from Firth et al. (2020))
The following sub-sections provide a comprehensive update on the state of the art to address these criticisms and suggestions. See Table 1 for a quick-use reference guide to the most appropriate recent literature that has addressed the various recommendations. A number of comprehensive and large-scale mapping studies have further evidenced statements made by Firth et al. (2020) on the prevalence of ambitious large-scale reclamation projects in the Middle-East and Asia, and the vulnerability of African countries to developments stemming from their rapidly rising populations. The major outputs from these are summarised in Boxes 1 and 2.
|
| Limitations/unknowns | Suggested actions | Key papers and recent advances/progress | ||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Experiments typically consider limited numbers of treatments/material types, ignore differences in surface area and are small scale (i.e. few metres) and short duration (e.g. less than 12 months). | Researchers should design better experiments (see Section 2.1). | Testing the independent and interactive effects of habitat area and spatial configuration (Loke et al., 2019a, 2019b).
| ||||||||||||||||||||||||||||||
| Limited metrics measured. | Go beyond simple biodiversity metrics and capture other biological, environmental and societal information (see Section 2.2). | |||||||||||||||||||||||||||||||
| Uncertainty on how GGI will perform under different environmental scenarios. | Repeat, revisit and replicate experiments (see Section 2.3). |
| ||||||||||||||||||||||||||||||
| Unintended outcomes such as whether GGI facilitates spread of invasive species or acts as an ecological trap or environmental/biological filter. | Report on unintended outcomes (both positive and negative) (see Section 2.4). |
| ||||||||||||||||||||||||||||||
| Dearth of large demonstration tests that show how interventions will perform when scaled-up operationally in ‘real’ developments. | Research and practice should move together in tandem (see Section 2.5). | Limited progress in this area but see Box 4 on Living Seawalls.
| ||||||||||||||||||||||||||||||
| Limited guidance for practitioners | Develop appropriate policies and guidance (see Section 2.6). |
| ||||||||||||||||||||||||||||||
Up until recently, many GGI projects that have primarily been led by the scientific community have been relatively smallscale, simple or potentially confounded by poor experimental design. For instance, only deploying a small number of units or considering comparisons of a small number of treatments (e.g. microhabitat types or materials) or combining factors that do not enable the researcher to disentangle single effects. Similarly, ignoring differences in surface area between treatments can yield misleading results. While it is well recognised that the incorporation of topographic complexity and microhabitats can provide refugia for organisms from both physical stress and biological pressure, this has rarely been quantified. The sections below focus on the state of understanding of the efficacy of each of these approaches for GGI.
Based on well-known positive effects of habitat complexity and diversity of refugia on biodiversity across scales, many GGI experiments have focused on incorporating a variety of small-scale microhabitats to increase the settlement, establishment and survival of different organisms onto artificial structures. Common strategies include retrofitting human-made infrastructure with concrete tiles moulded with topographically complex designs (Bishop et al., 2022; Loke et al., 2019a; Perkol-Finkel et al., 2018) or incorporating small-scale complexity into newly built structures (e.g. addition of holes, pits and water-retaining features such as rockpools (Bender et al., 2020; Evans et al., 2015)). Early studies enhancing habitat complexity failed to assess the contribution of the increase in surface area (rather than change in complexity) to biodiversity. Loke et al. (2019b) developed a novel system for testing the independent and interactive effects of habitat area and spatial configuration (i.e. fragmentation pattern) on intertidal species richness and revealed an optimal tile density and spatial configuration to maximise biodiversity on tropical seawalls. While a number of recent studies either use small-scale sampling units (Bishop et al., 2022) or compartmentalise the data by microhabitat (e.g. ridges against crevices on experimental tiles (Strain et al., 2021)) in an effort to standardise for surface area, greater attempts to disentangle the effects of increased surface area and addition of topographic complexity are recommended.
Arguably, the protection arising from mitigation of environmental stressors and/or predator exclusion by micro habitat complexity may be more important than the increase in surface area. Outcomes will, however, depend on the type of complexity provided as well as environmental context (Bishop et al., 2022; Strain et al., 2018a, 2021). For example, crevices incorporated into artificial structures can promote oyster recruitment (Strain et al., 2018b), with pits and narrow spaces under overhangs, aiding coral recruitment (Burt and Bartholomew, 2019; Strain et al., 2018a). For fish, increased structural complexity and provision of diversity of refugia can alter biological interactions (predation and competition), leading to greater fish recruitment and survival rates on more complex artificial structures (Bartholomew et al., 2022; Burt and Bartholomew, 2019; Hayes et al., 2022; Komyakova and Swearer, 2019; Komyakova et al., 2021; Morris et al., 2018a). Understanding the extent to which each of these factors drives biodiversity–habitat complexity relationships is critical to understanding the range of environmental conditions across which habitat complexity enhancement will provide biological benefit (e.g. MacArthur et al., 2020; see Box 4 on Living seawalls).
Box 4. Living Seawalls
Ecologists at the Sydney Institute of Marine Science have partnered with industrial designers from Reef Design Lab to develop Living Seawalls – a modular system by which critical habitats for marine life can be added to marine constructions at scale (Figure 7(a)). Using three-dimensional printing technology, the pits, crevices and pools often present on natural shorelines are recreated on modular panels. The panels, which can be manufactured from upcycled materials, are then fitted to new or existing marine developments in customisable Living Seawalls mosaics designed to last at least 20 years in locations of low wave exposure.
To date, ten different panel designs have been developed, each of which supports a distinct ecological community (Bishop et al., 2022). After only 2 years, Living Seawalls in Sydney Harbour have been colonised by over 100 species (see supplementary information in Bishop et al. (2022) for full details of the species found) and up to 35% more fish, seaweed and invertebrates than unmodified seawalls (Vozzo et al., 2021b).
Since the first Living Seawall was installed under the Sydney Harbour Bridge in late 2018, panels have been installed at 20 relatively sheltered sites in Australia and in Plymouth, UK (Figure 7(b)), Gibraltar, Wales and Singapore, with installations also planned for Boston, USA. These include urban renewal projects spanning hundreds of metres and private or public water frontages of tens of metres or less. Uptake of Living Seawalls has been assisted by the team's development of frameworks for ecologically enhancing marine infrastructure (Dafforn et al., 2015, 2016; Mayer-Pinto et al., 2017). These were used by the multinational company Lendlease in the Barangaroo urban renewal project and by the State Government of New South Wales in the Sydney Fish Markets redevelopment.
The project has also generated awareness of eco-friendly construction through public seminars and outreach events, stakeholder workshops and social media. Living Seawalls has been featured in over 20 international print, television and radio pieces and displayed in seven leading design museums, globally. Living Seawalls was a finalist for the 2021 Earthshot Prize, the winner of the 2022 Banksia Foundation Biodiversity Award and a Top Innovator in the World Economic Forum's 2022 UpLink BiodiverCities challenge.
Figure 7. (a) The Sydney Living Seawall. Image on right shows close up of the water retaining panel (photo credits living seawalls.com.au). (b) Green algae dominating early successional stages of the Living Seawalls in Plymouth only three months after installation in August 2023. Image on the right shows the first limpet to colonise the panels (November 2023). It is expected that once limpets arrive, they will graze down the algae making space for other species to colonise (photo credits Aeden Cooper)
No single GGI intervention is a ‘one-size-fits-all-solution’. What might be a good habitat for one species may be harmful to another (see Section 2.4.1). Similarly, what might be a good habitat for one species under one set of environmental conditions may be harmful to it under different conditions. Each species is susceptible to a particular set of predators, competitors and physical stressors that may vary with environmental context. To date, the majority of GGI installations have measured outcomes for biodiversity generally. Few studies have focused on particular species, traits or role in ecosystem functioning. For species that live in the intertidal zone, GGI that provides protection from physical stressors such as temperature fluctuations and solar radiation may be more important on the upper shore, while GGI interventions that provide a refuge from predation or grazing pressure (e.g. Martins et al., 2010) may be more important on the low shore and in the subtidal zone (Bishop et al., 2022). Much can be learnt by considering how individual species or functional groups respond to different habitat types under different environmental conditions (Aguilera et al., 2023a, 2023b; Strain et al., 2021; see also Section 2.3). While some progress has been made in the consideration of environmental stressors and biotic interactions, this remains a major knowledge gap. Successfully identifying species and functional groups that have important structuring roles (e.g. predators, competitive dominants and habitat-formers) in a local context is of critical importance. For this to be successful, designers of GGI interventions need to work closely with ecologists who have an in-depth understanding of the local ecosystem.
The growing body of interdisciplinary research involving ecologists and materials scientists is revealing limited differences in the biological colonisation of materials, when cement content, cement replacements (pulverised fly ash, ground granulated blast-furnace slag and fibre concrete), or aggregates and admixtures on biodiversity were examined (Becker et al., 2020; Bone et al., 2022a; Hayek et al., 2022; Hsiung et al., 2020; Kress et al., 2002; Lapinski et al., 2022; McManus et al., 2018; Vivier et al., 2021a). While an early study by Perkol-Finkel and Sella (2014) reported greater colonisation on lower pH concretes (pH 9–10.5) than ‘standard’ values (pH 12.5–13.5), this study was confounded as it was impossible to disentangle the separate effects of complexity and material. Recent experiments in temperate and tropical regions (Hsiung et al., 2020; Lapinski et al., 2022) have found no effect of concrete pH reduction on colonisation.
Natural rock, metal, oysters and wood are also extensively used in marine infrastructure. Experiments suggest limited differences in the communities settling on and establishing these materials (reviewed by Dodds et al. (2022) and on coastal concrete by Bone et al. (2022a)). Hartanto et al. (2022) – for example, reported no tropical intertidal faunal differences and minor algal differences in colonisation of granite, limestone, sandstones and concrete. Little is known about how materials affect microbial community composition. Natanzi et al. (2021) reported differences in the relative abundances of cyanobacteria, diatoms associated with different concrete mixtures after one month. Conversely, after a similar deployment time (31 days), Summers et al. (2022) found no differences in microbial diversity among the same stone types tested in Hartanto et al. (2022). While it appears that surface properties may influence microbial communities early on, it appears that communities converge later in the ecological succession process. Other factors, such as site or surface aspect (north–south directionality) and orientation of material (e.g. horizontal/vertical), may have greater influence on community composition (Amstutz et al., 2021, 2024; Firth et al., 2015), with effects potentially even larger than those from surface complexity (Grasselli and Airoldi, 2021). This evolving picture can help industry focus resources on other aspects of material choice, including carbon dioxide footprint (Dauvin et al., 2022; Dennis et al., 2018) and chemical pollution of concrete (Kress et al., 2002; McManus et al., 2018).
The transplanting of mobile invertebrates has received little attention, but Morris et al. (2018a) found that most individuals migrated out of experimental units. Species-specific behaviour and the use of enclosures should be considered for transplantation studies. Conversely, Firth et al. (2023a) suggest that the small-scale removal of limpets from carefully managed patches may yield biodiversity benefits with implications for ecosystem functioning and service provision. Importantly, this suggestion is relevant in a northwest European context, where limpets have a key structuring role (Coleman et al., 2006; Firth et al., 2021a). Removal of limpets in other systems may have little to no benefit and may be detrimental. Any manipulation of species requires consultation with local ecological experts.
The benefits of GGI interventions for biodiversity are typically assessed using metrics comparing differences in mean numbers of species (i.e. alpha diversity (Evans et al., 2021)). Much can be learnt from consideration of a wider variety of diversity metrics such as beta diversity, which gives a measure of variation in community structure among a set of sample units and hence the importance of variety of microhabitats on a more landscape scale (Firth et al., 2016b; O'Shaughnessy et al., 2023). More worrisome is the lack of attention to species identity, since it implies that attracting a pest, non-native or ephemeral species is potentially given the same positive weight of attracting species that are rare, on the brink of extinction and/or recognised as in need of protection. Also, an unbiased evaluation of the success of a given eco-engineering strategy requires the set of species one wishes to attract to be clearly identified beforehand.
Limited research efforts have focused on the influence of GGI interventions on ecological functionality, including fitness and reproduction, trophic structure (but see Espinosa et al., 2021; Raoux et al., 2022; Sedano et al., 2020a, 2020b) or ecological connectivity. GGI interventions may, for example, serve as hubs for sources of propagules (Reddy, 2022) or facilitate novel interactions among species (Klein et al., 2011) through a failure to support viable populations of key intertidal grazers (Moreira et al., 2006). A key opportunity of GGI is the capacity to deliver multi-faceted ecosystem functioning that can, in turn, support ecosystem services and mitigate stressors common in urban coastal environments (Pioch and Souche, 2021). For example, GGI can assist in reducing wave overtopping on seawalls, increasing their capacity for coastal protection (O'Sullivan et al., 2020). GGI can also enhance the abundance of filter feeders, such as oysters, mussels and sponges, or microbes involved in nutrient cycling and potentially improving water quality (Bulleri et al., 2022; Dodds, 2022; Rouse et al., 2020; Vozzo et al., 2021a). Conversely, added topographic complexity may have positive benefits on species, but have little or no effect on productivity (Mayer-Pinto et al., 2023); and results can be location-specific (Mayer-Pinto et al., 2024). Artificial structures impact surrounding habitats by way of changes in the characteristics and biodiversity of proximal sediments (Hanley et al., 2014; Heery et al., 2018; Martinez et al., 2022), transport of wrack detritus (Critchley et al., 2021), litter accumulation (Aguilera, 2018; Aguilera et al., 2016, 2023a, 2023b) and novel use of structures by terrestrial (pest) predators (Aguilera et al., 2023a). Initial work suggests that GGI can ameliorate some of these impacts, by retaining wrack (Strain et al., 2018b), but the influence of GGI on sediment properties (see Section 2.4.3), biotic assemblages and predator–prey interactions remain largely unknown.
GGI interventions may furthermore provide social and economic benefits such as aesthetic landscape appreciation (Pioch et al., 2011) and fisheries enhancement (Burt and Bartholomew, 2019; Morris et al., 2018a), especially when incorporating indigenous knowledge (Porri et al., 2023). A large-scale example is increased habitat use and feeding by juvenile salmon on large-scale GGI in Seattle (Accola et al., 2022a, 2022b; Sawyer et al., 2020). Generally, however, the economic valuation of ecosystem services associated with GGI is in its infancy (e.g. Mehvar et al., 2018), but is nonetheless essential to support or challenge economic justifications by enabling cost–benefit analysis (Fairchild et al., 2022). Economic values include societal benefits by way of greater knowledge gained of GGI (Strain et al., 2019), a perceived increase in naturalness and biodiversity, as well as an association with healthy environments (Fairchild et al., 2022; Salaün et al., 2022). Perspectives may, however, vary among stakeholders and regions; hence further studies are required for a more comprehensive understanding of perceived benefits and potential social and ecological conflicts in implementing GGI (Aguilera et al., 2023b; Kienker et al., 2018; Morris et al., 2016; Pearson et al., 2016; Salaün et al., 2022). It is important to acknowledge that trade-offs between positive ecological outcomes should be considered. In an experiment where both physical and biological complexity were added to seawalls, trade-offs between species richness and functional outcomes were observed (Mayer-Pinto et al., 2023). Moreover, GGI may negatively affect surrounding natural habitats. If, for example, GGI promotes novel communities with high filtration capacity as observed on artificial structures (Layman et al., 2014), it can potentially impact marine food webs beyond the footprint of the structure (Malerba et al., 2019; Raoux et al., 2022); conversely, societal benefits may be delivered by increasing water quality, especially in enclosed urban water bodies (Wilkinson et al., 1996). Further research is therefore needed in this area (Riascos et al., 2020), including differing viewpoints of different constituencies of stakeholders.
A growing number of studies are testing the efficacy of GGI interventions across local gradients and biogeographic regions (e.g. Clifton et al., 2022; Mayer-Pinto et al., 2024; Strain et al., 2021). Spatially replicated experiments have revealed that the impacts of many (but not all) GGI interventions are highly context-dependent, with results ranging from positive to neutral to negative, across spatial scales of centimetres to hundreds of kilometres (Chee et al., 2021a, 2021b; Clifton et al., 2022; O'Shaughnessy et al., 2021; Strain et al., 2021). While the work of Strain et al. (2021) is exemplary in its spatial extent (28 sites in 14 cities, spanning five continents globally), this was a short-term experiment that relied on each partner independently ‘buying in’ and having the financial and human resources to contribute within a particular timeframe. Consequently, the temporal extent was limited to just 12 months. Nonetheless, this model of global replication is the ‘gold standard’ and should be aspired to when and if global funding opportunities enable such projects. Replication of previous experiments at new locations also shows that results cannot be generalised; in contrast to experiments conducted by Perkol-Finkel and Sella (2014) in Israel, experiments by Hsiung et al. (2020) in the UK and Singapore showed concrete pH to be unimportant in determining biodiversity. Similarly, whereas artificial rockpools have strong positive effects on biodiversity in many temperate settings (Browne and Chapman, 2014; Evans et al., 2015; Hall et al., 2019; Ostalé-Valriberas et al., 2018), in some tropical settings, extreme high temperatures, precipitation and sedimentation may limit their benefits (Firth and Williams, 2009; Waltham and Sheaves, 2018, 2020; but see Chee et al., 2020). In a study spanning five locations across three continents, Mayer-Pinto et al. (2024) found conflicting results for GGI interventions on productivity and respiration metrics. Many GGI interventions (especially those that need to be affixed to infrastructure, such as tiles, panels and precast units such as vertipools and pots) are susceptible to damage from wave action (Browne and Chapman, 2011), interference by sediment inundation (see Section 2.4.3) and even total burial by sand, particularly if the intervention is associated with shore perpendicular groynes (e.g. the BioBlock (Firth et al., 2020)). Conducting experiments along wave exposure gradients will yield invaluable information about limiting conditions and engineering constraints.
For eco-engineering interventions based on habitat complexity enhancement, key correlates of spatial variation include latitude, tidal height, size of the local species pool and locally dominant stressors (Clifton et al., 2022; O'Shaughnessy et al., 2021; Strain et al., 2021). These factors are likely to reflect the varying importance of habitat complexity in mitigating key environmental stressors (e.g. extreme temperatures and desiccation) and intensity of biological interactions (predation, competition and facilitation) in different settings (Strain et al., 2018a). They will also reflect environmental variability in the identity, diversity and supply of colonists on which habitat complexity can act (Clifton et al., 2022). The type of habitat complexity provided (e.g. holes, crevices and water-retaining features) also influences eco-engineering outcomes (Bishop et al., 2022; Strain et al., 2018a). Thus, identification of natural complexity patterns on different rock types and how they vary across latitudes, wave exposure gradients and tidal ranges are relevant in this context to inform eco-engineering approaches (e.g. latitudinal gradients in microhabitat availability (Aguilera et al., 2022; Bracewell et al., 2018)). Biomimicry-inspired designs, by learning from nature, can inform both the ecological functioning and aesthetics of GGI (Pioch and Souche, 2021); habitat-building species may provide the most profitable avenues to explore (Byers, 2022).
Despite increasing spatial replication (see Box 4), most GGI studies remain limited to <1 year (reviewed by Dodds et al. (2022) and Strain et al. (2018a), but see Bender et al. (2020), Bishop et al. (2022), Chee et al. (2020) and Wilkinson et al. (1996) for exceptions). The importance of GGI interventions for biodiversity can change over time. Both Martins et al. (2016) and Bender et al. (2020) revealed similar positive results on revisiting GGI installations after 6 and 12 years, respectively (see Martins et al. (2010) and Langhamer and Wilhelmsson (2009) for original studies). Importantly, both studies specifically targeted enhancing species of commercial interest. Less is known about the long-term influence of GGI interventions on biodiversity more broadly. Where rates of succession are slower (e.g. in temperate regions), early colonisation processes remain over-represented in the GGI literature. Meta-analyses report weak (non-significant) patterns of diminishing substrate property and habitat complexity effects through time, and more multi-year studies are needed to adequately explore this (Dodds et al., 2022; Strain et al., 2018a). Indeed, in addition to the confounding impacts of competition noted above, Bishop et al. (2022) reported diminishing effects of physical habitat complexity on species richness after one year, because habitat-forming taxa themselves became the key determinant of habitat complexity. Early and mid-successional opportunistic species may also inhibit later colonisers, unless their dominance is broken by physical or biological disturbance (Sousa, 1979). Besides evaluating the benefits of eco-engineering at ecologically meaningful scales, long time series are critical in assessing GGI performance under rapidly changing environmental conditions (e.g. warming associated with climate change (Sun et al., 2022; Waltham et al., n.d.)).
All experimental trials and installations have the potential to become damaged and/or lost due to a range of natural and anthropogenic factors. For instance, tiles or units that are affixed to seawalls may suffer damage from wave action and storms (Browne and Chapman, 2011) or vandalism from members of the public (de Moraes et al., 2022). It is important that researchers and practitioners alike report on such incidents, particularly in relation to factors that can be controlled, such as choosing a location that is less likely to be damaged by wave action. Arguably, it is more important to know when schemes fail than when they succeed, as greater knowledge is gained from failure than from success. The sections below review some recent advances in some of the unintended outcomes that were identified by Firth et al. (2020).
‘Ecological traps’ occur when the links between habitat quality and habitat selection are decoupled, resulting in negative fitness outcomes (Battin, 2000; Komyakova and Swearer, 2019; Swearer et al., 2021). Depending on the scale of impact and the species in question, ecological traps may cause local and regional extinctions (Hale et al., 2015). The question of ecological trap formation due to GGI installations has received little attention. Recent research on ARs (Komyakova et al., 2021) and fish farms (Barrett et al., 2018, 2019) provides some evidence for concerns. Habitat selection is generally based on suitable habitat detection using a range of cues (Kingsford et al., 2002). Increased temperatures and pollution (e.g. chemical, light and noise) can impede the ability of larvae to differentiate positive and negative selection cues (Doney et al., 2009; Nilsson et al., 2012), leading to poor habitat selection decisions with negative fitness consequences (Fobert et al., 2019; Komyakova et al., 2022; Marangoni et al., 2022). Consequently, GGI installations deployed in polluted environments (e.g. ports, marina and harbours) may enhance the likelihood of ecological trap formation (Komyakova et al., 2022). Additionally, if GGI is applied to marine infrastructure where recreational fishing is popular, higher mortality rates for certain species may result (Swearer et al., 2021), perpetuating the attraction over production argument (Pickering and Whitmarsh, 1997). Further research is urgently needed to understand the implications of increased attractiveness of GGI installations and potential fitness consequences. Importantly, artificial habitats can act as ecological traps for some species and population sources for others, and hence a multi-species approach is needed (Komyakova et al., 2021).
Some coastal infrastructure can have unintended positive outcomes or ‘happy accidents’ (Rosenzweig, 2003) for species of conservation concern. In South African estuaries, marinas are constructed with gabions (wire cages filled with rocks) and lined with reno mattresses (flattened gabion boxes used to line canals for erosion control). Originally chosen for aesthetic purposes, the gabions and reno baskets have serendipitously provided habitat to the endangered Knysna seahorse (Hippocampus capensis) (Claassens, 2016; Claassens and Hodgson, 2018) and are even used by H. capensis (and other species) in preference to natural eelgrass habitat (Claassens et al., 2018). Similarly, seawalls in the port of Ceuta, North Africa, appear to be important refuges for the endangered limpet Patella ferruginea from human harvesting (Rivera-Ingraham et al., 2011). Where coastal infrastructure supports species of conservation concern (see Firth et al. (2016a) and Ostalé-Valriberas et al. (2022) for reviews), attempts should be made to deploy GGI interventions to boost populations (e.g. Langhamer and Wilhelmsson, 2009; Martins et al., 2010) and to manage such sites as part of wider networks (e.g. artificial marine micro reserve networks (García-Gómez et al., 2011, 2015; Ostalé-Valriberas et al., 2022)).
Convention suggests that GGI installations can promote the diversity and abundance of native species, with the view to offsetting invasion success of NNS (Arenas et al., 2006; Stachowicz et al., 1999). There is no reason, however, why an increase in habitat complexity (either physical or biological by way of seeding/transplanting of habitat-formers) should not also favour NNS (Gauff et al., 2023). Evidence of NNS responses to GGI approaches is growing (e.g. O'Shaughnessy et al., 2020a; Perkol-Finkel et al., 2018; Peters et al., 2017) with variable outcomes across geographic locations, tidal height, types of interventions and functional groups or species. NNS responses to GGI interventions are inevitably species-specific depending on the environmental tolerances and traits expressed by individual species. Rather than looking at the diversity of NNS across experimental units, researchers could consider diversity/abundance of native species as covariates in analyses, and how patterns of colonisation by natives interacts with experimental manipulation in NNS response.
Early evidence that GGI interventions could enhance native biodiversity over NNS came from ‘bio-enhanced panels’ (Perkol-Finkel et al., 2018), although, as both surface complexity and concrete composition were manipulated in the bio-enhanced panels, individual effects cannot be distinguished. More recent GGI studies replicated in different geographical locations (World Harbour Project) isolated complexity from other factors with contrasting results. For example, in Plymouth, UK, O'Shaughnessy et al. (2021) reported reduced NNS abundance on complex compared with flat tiles deployed subtidally; a similar result emerged for intertidal treatments in Sydney, Australia (Vozzo et al., 2021a). By contrast, no treatment-specific differences emerged for NNS in either sub- or intertidal experiments in East London, South Africa (Mafanya, 2020) despite the presence of NNS in the local area (Peters et al., 2017). Using the same panels in Sydney, Australia, Schaefer et al. (2023) found that more complex tiles supported greater abundance of invasive ascidians, particularly when manufactured with oyster shells compared with controls or those manufactured with vermiculite. Similarly, Gauff et al. (2023), working in Toulon, France, found that complex habitats engineered to protect juvenile fish from predation increased NNS numbers. Less information on GGI and NNS is available from the tropics, but reports from several major shipping ports have found few to no NNS (Tan et al., 2018; Waltham and Sheaves, 2020; Wells, 2018; Wells and Bieler, 2020; Wells et al., 2019).
The ‘priming’ of grey infrastructure by the addition or ‘seeding’ of native habitat-forming species to reduce NNS recruitment has also been investigated but has again generated mixed results. O'Shaughnessy et al. (2021) found no difference in numbers of NNS across seeding treatments, while Vozzo et al. (2021a) reported how GGI seeding with a native oyster increased the abundance of a non-native isopod. Morris et al. (2018b), however, found that the seeding of water-retaining rockpools with mobile invertebrates limited NNS establishment.
Despite limited evidence that GGI promotes NNS diversity or abundance, practitioners and government agencies remain concerned that GGI may contribute to the introduction and spread of NNS. While it might be impossible to predict or prevent colonisation of NNS, to minimise risk of colonisation, it is necessary for GGI to be designed to inhibit colonisation of pest species of local concern (see Dafforn (2017) for a review). This requires knowledge of the local species pool and potential impacts to native biodiversity. Furthermore, knowledge of NNS hotspots linked to shipping patterns may be informative (O'Shaughnessy et al., 2020b; Tidbury et al., 2016).
GGI interventions can be prone to sedimentation, limiting their capacity to host typical hard-substrate biota (Bone et al., 2022b; Firth et al., 2016b; Hall et al., 2019; Waltham and Sheaves, 2018). Although considered a potential management issue, due to the perceived costs and need for responsibility associated with sediment removal (Waltham and Sheaves, 2018), sediment collected in artificial rockpools in the UK contained infauna typical of surrounding estuary mudflats (Bone et al., 2022b). Consequently, it is important to consider retained sediment as potential habitat and to sample accordingly, even if the retained sediment is perceived to be less ‘interesting’ than rockpool habitats by the public or practitioners (J. R. Bone, personal communication). Changes in attitude towards retained sediment begin with scientists themselves (often with backgrounds in rocky shore ecology) viewing sedimentation as ‘potential’ instead of ‘problem’. Nonetheless, undesirable sedimentation could be avoided by greater understanding of the local sediment supply and the depositional environment prior to GGI installation. The inherent dynamism of coastal environments means that features previously free of sedimentation may suddenly become inundated, and vice versa. Given that infrastructure often causes the loss of sedimentary habitats (Heery et al., 2017), much greater attention should be given to GGI interventions that encourage the accumulation of sediments, but on sufficiently large scales that they support fully functional habitats.
The GGI concept has been driven since the early 2000s by an ecological research perspective showing that shoreline armouring has significant impacts on biodiversity and functioning (Airoldi et al., 2005; Chapman, 2003; Mayer-Pinto et al., 2018). Integrating GGI into everyday practice requires interdisciplinary collaboration and co-design – combining the necessary engineering and ecological expertise to ensure GGI success and incorporation into technical standards (Pioch et al., 2018). Unfortunately, logistical and financial constraints have limited GGI experiments to small-scale (i.e. few metres of the complete infrastructure) projects retrofitted onto existing structures, with few examples globally of research uptake into industry-led projects (but see Box 4 on Living Seawalls).
The Seattle seawall project showcases the integration of research into practice. Small-scale trials of complex concrete panels and a bench installation at the seawall base to test GGI approaches provided shallow-water habitat and enhanced prey for migrating juvenile salmon (Cordell et al., 2017; Sawyer et al., 2020; Toft et al., 2013). These techniques were later upscaled to nearly 1 km of upgraded seawall, with the addition of skylights into the boardwalk to increase light penetration to the migration corridor. It was estimated that the incorporation of GGI added 2% cost to the US$410 million build (Accola et al., 2022a; Sawyer et al., 2020), but the costs of long-term maintenance are unknown. This project was co-designed by ecologists and engineers from the onset, with a clear objective to improve fish habitat while providing the coastal protection required and human access to and appreciation of the marine environment. As such, it provides an aspirational benchmark and standard, not only for the application of GGI initiatives but also multifunctional infrastructure globally.
Key to avoiding greenwashing is the co-creation of guidance, case studies and policy instruments with practitioners (Naylor et al., 2012) in several domains (e.g. flood risk, green infrastructure) where eco-engineering interventions and implementation measures are placed on a spectrum (Figure 3; Naylor et al., 2023; Perkol-Finkel and Sella, 2014; Pioch et al., 2011; Taljaard et al., 2019). Levers that could support increased uptake of eco-engineering are documented (e.g. Claassens et al., 2022; Evans et al., 2019; Mayer-Pinto et al., 2017; Naylor et al., 2012; Pioch et al., 2018) and policy gaps identified (e.g. Victoria's (Australia) Coastal Strategy, 2020 and South Africa's Coastal Management Act, 2014; RSA, 2014; Global Inventory of Biodiversity Offsetting Policies, GIBOP, 2019). To avoid greenwashing, specific guidance and co-created case studies (e.g. International Guidelines, Suedel et al., 2022; NOAA's Habitat Blueprint), strategic plans (e.g. National Marine Science, 2014), state of science (Australian Government, 2021), regulatory requirements (e.g. Hydraulic Project Approval) and policy instruments (e.g. Welsh Government, 2021) are crucial. Flood management is the policy domain where the greatest efforts have been undertaken to limit greenwashing and encourage the uptake of green–grey eco-engineering (e.g. The Washington State Hydraulic Project Approval, USA). The UK is a leader in this space, with eco-engineering (and GGI) guidance (Naylor et al., 2012, 2017) and co-produced case studies (Estuary Edges, 2008; Naylor et al., 2017) leading to statutory flood policies (Welsh Government, 2019, 2021), explicitly stating that GGI should only be implemented where other nature-based options are not suitable. Similar inventories of case studies and guidance exist for Singapore (Lai et al., 2022) and Malaysia (Chee et al., 2021b). In 2023, the French ministry published a guideline to favour GGI and eco-engineering practices in future French ports (DGITM, 2022). Other policy domains (e.g. conservation, infrastructure and marine) would benefit from explicitly mentioning green–grey eco-engineering.
Central to Firth et al. (2020) was a concern that GGI could be misused for greenwashing, as consultants, developers and local authorities implement GGI to expedite, facilitate and reduce costs of regulatory processes. Not only does this remain the case but reports from municipal engineers (L. B. Firth, personal communication) also indicate how, by strategically incorporating GGI onto marine infrastructure on existing reclaimed land, public opinion can be swayed in support of proposals for future large-scale land reclamation. This it is argued, is a clear example of GGI being used as a Trojan horse for environmentally harmful development. In response to the greater take up of GGI applications on coastal infrastructure, private companies that design environmental products have emerged globally. Although vital to successful GGI implementation and the development of local economies, self-evidently, companies that design and produce environmental products such as habitat enhancement units, novel concretes and AR products stand to gain financially from greater uptake of GGI schemes. Objective and independent evaluations of proposed GGI installations are recommended and, for a critical assessment of development, monitoring and assessments funded or initiated by those with a conflict of interest. Furthermore, any data generated should be publicly available to ensure transparency.
It is imperative for environmental management decisions and actions, such as GGI, to be evidence-based (Downey et al., 2022; Lemasson et al., 2023; Sheaves et al., 2021) in order to maximise beneficial outcomes and minimise waste of time and resources, mainly from public funds (Sutherland and Wordley, 2017). In the nascent field of GGI, caution must be exercised when making decisions to avoid potential evidentiary dissonance (where a paradigm is supported by an apparent abundance of evidence, but with little basis in actual reported scientific findings and a ‘too big to fail’ complex; see discussion by Sheaves et al. (2020)). This phenomenon has been reported in other fields of environmental management (Lemasson et al., 2023; Sheaves et al., 2020). Here, this might occur if the paradigm that GGI is beneficial was promoted without establishing evidentiary scientific support first (Gauff et al., 2023).
GGI may still suffer from a lack of underpinning science that can be used to support evidence-based management. Firth et al. (2020) highlighted some of the limitations to the evidence base, including difficulty generalising the findings (different metrics, timescales, environmental conditions and sampling protocols). Significant progress has been made in response to the call for action expressed in their 2020 paper (see Table 1 for summary; see also the work discussed in this paper). Importantly, many new schemes have emerged since 2020 that have been driven by practitioners and not the scientific community. Ecological responses from such practitioner-led projects will not be subjected to the same rigour and criticism as peer-reviewed scientific literature and may be difficult to find online. Crucial evidence may also be ‘hidden’ from the general scientific literature, their access restricted due to institutional or corporate confidentiality, or due to a lack of transparency or willingness to share from their authors (Sheaves et al., 2016). It is essential to learn from other fields of environmental management where collaborative sharing across sectors is becoming more standard practice (such as in the field of oil and gas decommissioning; e.g. the InSite programme, Table 2).
|
The past few years have seen a clear effort to improve the GGI evidence base, but also to consolidate it, with evidence syntheses starting to emerge (Evans et al., 2021; O'Shaughnessy et al., 2020a; Strain et al., 2018a). These efforts are promising but become rapidly out of date to the rapidly evolving field and proliferation of the literature. Some fall somewhat short of the standards required of evidence syntheses due to not being based on robust, transparent and repeatable methods. Others may not reflect accurately on the state of the art, being selective of only a few specific GGI interventions, rather than encompassing all possible known GGI interventions, and only incorporate published literature, excluding the grey literature (thus likely biasing the results towards positive results and ‘successes’ rather than ‘failures’). Future syntheses should thus aim to follow the gold standards (see Lemasson et al. (2021, 2022) for a systematic protocol and map of evidence available to inform oil and gas decommissioning), incorporate all possible GGI interventions and include both published and grey literature where possible, to provide the best possible scientific support for management and decision making when it comes to applying GGI. If ‘reefing’ of oil and gas and offshore windfarms as a decommissioning option becomes more and more popular in the future (Schläppy et al., 2021), how these structures could be modified or treated prior to reefing to maximise benefits (such as biodiversity enhancement or targeted species overexploited for instance) should be investigated (see Knights et al., 2024a, 2024b e.g. using expert scientific consensus). GGI could be applied to decommissioned offshore structures (either prospectively or retrospectively) to achieve greater ecological benefits. While some studies do exist that touch on this topic for existing operational structures (Bos et al., 2021; Langhamer, 2012; Langhamer and Wilhelmsson, 2009; Roach et al., 2022), this is a major knowledge gap that should be investigated for both existing and decommissioned structures in the future.
This paper provides an update on state of the art of GGI research since 2020 (summarised in Table 1). Researchers are designing better experiments by testing the independent and interactive effects of habitat area and spatial configuration, standardising comparisons between units by using small sampling units or compartmentalising data by microhabitats, considering greater numbers of materials, experimental units, surface orientations and context-specific refuge provision (see Section 2.1). Researchers are also going beyond simple biodiversity metrics and capturing other biological, environmental and societal information (see Section 2.2.). In many places, existing installations and experiments are being revisited to test responses over longer timescales, and repeated and replicated under different environmental contexts (see Section 2.3). Far greater attention is being given to reporting on unintended outcomes such as GGI performing as ecological traps, supporting invasive species and being inundated by sediments or indeed occasional happy accidents with unexpected benefits (see Section 2.4). While limited evidence is currently available on ‘real-world’ scaled-up installations (but see Living Seawalls, Box 4), some large-scale projects are beginning to emerge (see Section 2.5). Finally, appropriate policies and guidance are beginning to emerge (see Section 2.6). Table 2 provides links to some of the major tools and resources that are available to practitioners and educators.
The global evidence base for GGI is rapidly expanding. Researchers and practitioners are urged not to oversell their results and report all findings in a nuanced manner, reflective of the short-term duration and scale of interventions in the wider context of experimental damage caused by the construction, particularly if it is a new development. Synthesising this evidence base using systematic mapping approaches is recommended, and learning from other fields of environmental management where collaborative sharing across sectors is becoming more standard practice. Importantly, some major knowledge gaps remain. Critically, properly scaled-up examples of GGI are still lacking. While Living Seawalls (Box 4) is a great example of how science is being put into practice at real-world scales, the driver behind this example is still very much the research community and not the practitioner. Similarly, how such installations alter connectivity patterns across seascape scales remains unknown. Looking to the future, a major challenge will be to predict how GGI will interact with climate change for both NNS and native species that are on the move (Cannizzo and Griffen, 2019; Cannizzo et al., 2020). This is particularly true if future GGI efforts are applied to structures spanning natural barriers to dispersal and biogeographic provinces (e.g. Forbes’ Line in the UK and Ireland (Firth et al., 2021b; see also Lacroix and Pioch, 2011)). While great advances have been made through large-scale interdisciplinary projects, these are often spatially restricted to national or regional scales based on funding mechanisms. By addressing the above-mentioned challenges, it is possible to dramatically improve our ability to implement GGI in appropriate ways that can be classed as laurel wreaths over fig leaves or Trojan horses and achieve true win–wins for humanity and nature.
The 51 authors on this paper represent institutes in 14 countries spanning five continents. This large collective voice argues the need for a paradigm shift of current funding strategies and research programmes that will encourage the development, implementation and translation of GGI science at global scales. There is currently a lack of science and practice outside of academic sectors in the developed world, and insufficient global funding mechanisms that can support such collaborations. This rationale further evokes the need for equitable north–south partnerships in science informing GGI that is well embedded in the UN Local2030 agenda, with a key focus on sharing of tools and demand-driven research and action. There is thus a collective need for schemes that encourage intersectoral and trans-sectoral research, knowledge exchange and capacity building to optimise GGI in the pursuit of contributing to sustainable development.
Acknowledgements
This study was supported by the Natural Environment Research Council (NERC) and the ARIES Doctoral Training Partnership (NE/S007334/1) and the Natural Environment Research Council Growing Roots Funding (GR303).
