Policy supports for the deployment of solar fuels: islands as test-beds for a rapid green transition

Coastal areas, particularly islands, are especially vulnerable to climate change due to their geographic and climate conditions. Reaching decarbonisation targets is a long process, which will require radical changes and ‘out of the box’ thinking. In this context, islands have become laboratories for the green transition by providing spaces for exploring possibilities and alternatives. Here we explore how hydrogen (H2) energy technologies can be a critical ally for island production of renewable electricity in part by providing a storage solution. However, given the abundance of sunlight on many islands, we also note the huge potential for a more profound engagement between renewables and hydrogen technologies via the co-generation of ‘green hydrogen’ using solar fuels technology. Solar hydrogen is a clean energy carrier produced by the direct or indirect use of solar irradiation for water-splitting processes such as photovoltaic systems coupled with electrolysers and photoelectrochemical cells. While this technology is fast emerging, we question to what extent sufficient policy support exists for such initiatives and how they could be scaled up. We report on a case study of a pilot H2 plant in the Canary Islands, and we offer recommendations on early-stage policy implications for hydrogen and other solar fuels in an island setting. The paper draws on the literature on islands as policy laboratories and the multi-level perspective on energy transitions. We argue that particular attention needs to be given to discrete issues such as research and planning, and better synchronising between emerging local technology niches, the various regulatory regimes for energy, together with global trends.


Introduction
Given the scale of the climate change challenge, developing new solutions for cleaner energy will require 'out of the box' thinking. Islands can be ideal laboratories for new energy solutions as they provide an opportunity to explore such transitions. At the island level we can study if new energy systems are fair, inclusive and sustainable, and we have opportunities to resolve the basic engineering problems (intermittency of supply, constraints over land use, and a dependency on rare minerals) at a manageable scale. Although most islands around the world still heavily rely on fossil fuels, many also naturally hold an enormous range of renewable potential including solar, wind and tidal energy. These island communities are increasingly focused on the development of offshore energy projects as an answer to their lack of available land space.
Critical to such developments is the question of energy storage, which traditionally has been delivered by expensive pumped water storage or conventional batteries. Some islands, notably the Canaries and Madeira have invested in such infrastructure. The results have sometimes been mixed [1]. However, other innovative technologies are emerging for electricity storage, involving batteries and hydrogen [2,3].
Developing alternative, efficient, and cost-effective ways for storing solar energy is the main way to fight island-based renewables intermittency and curtailment. In this context, solar water splitting, through its  [4]. © IOP Publishing Ltd. All rights reserved. multiple processes (figure 1), provides a scalable route to store and convert solar energy in the form of hydrogen or solar chemicals.
From a technical standpoint, hydrogen can be produced and labelled in several different ways, in what is commonly described as the hydrogen rainbow 4 . Since the focus of this paper examines only green transition pathways for the production of hydrogen, other solar fuels based on carbon are not considered in this analysis. Moreover, islands will tend to have a small amount of CO 2 emitters which could directly affect the adoption of a carbon-based solar fuels industry.
Amongst thermal, electrolytic, and photolytic processes, solar hydrogen production has been progressing through advancements in photovoltaic (PV) water electrolysis (PV + E) technologies. PV + E has been demonstrated with high solar-to-hydrogen conversion efficiencies but it is still too expensive to compete with traditional methods. Meanwhile, the concentrated solar power (CSP) sector has been emerging, although photocatalytic (PC) and photoelectrochemical (PEC) systems are still mostly at a low technological readiness level (TRL). Nonetheless, scientists continue to pursue effective ways to chemically convert water and carbon dioxide into fuels and bring Artificial Photosynthesis up to a higher TRL.
As part of a worldwide effort to reduce greenhouse gas (GHG) emissions, governments are developing a mix of policies and strategies to support the emergence of clean hydrogen projects. Central to such policies are guidelines and standards for a successful demonstration and commercialisation of hydrogen technologies. Planning laws also need to be considered to match the potential of H 2 technologies on local, regional, national, and international scales. A distinctive feature of hydrogen policy development is that deployments tend to spatially co-locate value chains within traditional industrial sites, usually called 'hydrogen clusters/hubs/valleys' . This scenario is evolving as the energy transition becomes more widespread and affects a greater number of sectors (e.g. mobility, built environment). Increasingly, there is interest in coastal industrial clusters, due to their proximity to planned large offshore wind deployments, and the possible conversion of industrial ports into hydrogen hubs to realise emissions savings in ferries, marine shipping, and portside activities [5]. Islands can offer test beds for these initiatives.
Nonetheless, the pattern of hydrogen development in specific clusters raises important challenges at the local site level from an environmental planning perspective. Such projects typically require very proactive local government or agency leadership. Moreover, without local public support, sites of innovation concerning hydrogen and renewable energy developments will likely face political opposition. As Shaw and Mazzucchelli (2010) argue: 'Independent of the community type, political will and acceptance in the community are crucial for successful hydrogen integration and thus for a community to become a hydrogen community' [6]. Islands have the potential of becoming just such 'hydrogen communities' as is detailed in the next section which provides some specific examples and situates these within the wider literature on islands as policy laboratories.

Islands and littorals as policy laboratories for energy transitions
There is a long history of islands being employed as policy laboratories, especially for green technologies [7,8], and a renewed interest in recent years that islands could become key sites for energy transition solutions [9]. Small islands offer relatively limited 'safe' spaces to try out new technologies. Electricity grids on islands are often fragile and energy supply can be problematic hence stimulating new energy projects [1].
A more profound rationale is that islands are often intrinsically liminal places, typically on the margins of established infrastructure and networks. They often enjoy substantial political and regulatory autonomy, permitting exceptions, derogations (notably from some EU norms) and solutions that are not attempted in the principal jurisdiction.
That intrinsic liminality, between sea and land, between the natural world and society, invites numerous opportunities to innovate. Islands are in many ways structurally forced to be innovative. For instance, Europe's premier test centre for wave and tidal research, the European Marine Energy Centre is located in the Orkney Islands [10], justified by the locally optimum physical parameters. This siting also reflects these islands' long history of experiments in renewable energy (wind turbines). For example, the Orkneys are home to several hydrogen pilot projects as part of a wider ReFLEX low-carbon energy transition plan [11], which deliberately includes hydrogen as a key medium for energy storage and mobility 5 . Today, numerous island communities are involved in the production of renewable fuels. A table containing more examples of islands involved in renewable hydrogen production can be found in the supporting information section [12][13][14][15][16][17][18][19][20][21][22][23][24][25].
The North Sea has been a key site of innovation for marine renewables, and islands have often featured as strategic sites for the offshore wind industry. The German island of Heligoland has played a noteworthy role in this regard. Islands in the Aegean, the Mediterranean and the wider Atlantic 'arc' , including the Azores, have also heavily influenced the energy transition through involvement in key projects [1,9,26]. Some projects, notably the EU-funded SMILE [27], have united several far-apart islands (e.g. Samsøe, Madeira, and Orkneys) together into an innovation network. The next logical step in the 'islands as laboratory' paradigm is the possibility of artificial islands, which have been proposed in the North Sea, for both renewable energy and hydrogen production. Danish, Belgian, Dutch and British electricity providers have plans for several of these 'islands' which would in some cases utilise existing islands (e.g. Bornholm) and in others develop entirely new artificial structures for energy storage as well as the associated electricity converters and substations [28,29]. Many of these plans feature hydrogen generation and storage aspects.
For now, the dominant policy narrative regarding islands and energy transitions is to secure a nexus with a booming offshore wind sector, together with novel storage and grid technologies. Therefore, the fundamental challenge for islands will be the broadening of their focus and capabilities to support a more diverse and innovative range of energy sources, including notably solar fuels.
A key observation at this juncture is that innovation projects which involve hydrogen and renewables may differ from island to island as regards their ambition levels, geographic resources, and strategic focus. What we are interested in are the more profound synergies; not just using hydrogen as a way to store island-generated renewables, but making green hydrogen through electricity supplied by local renewable sources. Two recent projects, notably SEAFUEL [18] and GREEN HYSLAND [30], display a higher level of ambition in this regard. The SEAFUEL project aims at first producing green hydrogen locally, and then using it to fuel the local public transport fleets and tourist rental vehicles, which is a unique mobility sector application of the fuel. This approach increases the visibility of hydrogen technologies to the general public, who in the case of Tenerife has large numbers of tourists. People get to see and experience first-hand their mobility needs being powered by the mix of renewables and hydrogen, thereby improving public acceptance of hydrogen and its applications, and increasing the likelihood of future projects. The project is also intrinsically ambitious by not being limited to a single island but rather brokers an 'innovation network' across the Atlantic, with case studies involving Irish and Portuguese islands.
By way of contrast, GREEN HYSLAND is focused initially on just the Spanish island of Mallorca 6 , but aims for the creation of a complete 'H 2 ecosystem' there. This project will explicitly involve solar renewables both to generate electricity and some of that will be used for green hydrogen production. Downstream, this will be used as an energy carrier throughout the island, including injection into Mallorca's natural gas grid. changes in any technology come about when mutually reinforcing trends at multiple social levels come together to achieve a critical mass for change. Change takes time because it requires reinforcement and a certain degree of synchronicity, which is seldom planned or foreseen. Actors at various levels will try to change perceptions, engage in power struggles, lobby for favourable regulations, and compete with one another [32]. Hence, systemic transitions should involve all system levels. The approach uses three specific levels from which one can locate drivers for change: niche, regime and landscape levels.
At the lowest level there are niches, which are individual and specific pioneers of new technologies, often at the start-up firm or laboratory level. They are sometimes individual inventors or pilot projects. In fact, a key insight of the MLP approach is that to have any realistic chance of success most niche innovators require a type of sheltering and protection from competition and the inevitable early setbacks of technology failures and higher costs. This is often achieved institutionally, by being folded up within a larger organisation or project, and usually sheltering takes the form of extensive subsidies and sometimes special derogations and rules as well. Such sheltering often involves regime-level actors (governments, regulators, grid operators, etc) providing support, subsidies, and technical assistance. Here, we are interested in specific examples of islands which themselves are niches of innovation, a special case where the geographic features of the technology niche are very distinctive and this shapes how the niche can innovate.
It is important to point out that at the niche level we do not just find the technology innovators by themselves. There may be competing or rival niches offering different technology solutions, and energy users and consumers are crucial actors who may react by endorsing and accepting the new technology or just the opposite, rejecting it as costly or impractical for other reasons.
Here we note that social science perspectives on islands as key sites for an energy transition stress that an important consideration is the extent to which local island communities are active participants in novel energy infrastructure and whether such investments are matched with local needs [9]. To be successful, there should be local 'buy-in' and support as regards the planning dimension. Shaw and Mazzucchelli (2010) argue that well-targeted multi-stakeholder demonstration and deployment of H 2 or related projects can create a sense of ownership amongst public authorities and citizens [6].
Equally, island projects must be connected to a broader network of opportunities and innovation. Otherwise, they risk being isolated as 'exotic projects' . Therefore, a key evaluative criterion in assessing islands as test beds for a rapid green transition is to what extent islands exhibit strong niche-level activity. These islands should not just have specific pilot projects exploring the technology but these niches should enjoy local community support and the engagement of local governing elites. Also, one should assess to what extent they are sheltered by subsidies and special rules.
At the regime level we find the rule-makers and the overall system in which technologies are governed in what become complex and overlapping socio-technical regimes. These include for example the national electricity grid system, the national system for regulating energy firms and suppliers, and the various fiscal and tax policies in place on energy. Networks of regimes are no longer found to be exclusive at the national level, as globalisation has meant complex international regimes exist which govern technologies in many domains. Within the European context, the EU is a vital actor at the regime level, establishing the broad parameters for the energy marketplace, whilst also being active in setting policy ambitions more broadly, especially regarding climate change and the energy transition.
Established regulatory frameworks and regimes may not be suitable for specific island settings, and thus they may require either adaption or derogations [33]. Supporting policies for island-based energy projects sometimes need to be tailored in a bespoke way, as in a few cases energy projects have resulted in technological failures, more expensive energy prices and no improvement in grid stability. In the short-term island energy projects will likely require significant subsidies and other interventionist measures such as guaranteed preferential feed-in tariffs. According to Tsagkari and Jusmet [1], a common policy weakness is a reluctance to properly fund or legislate for energy storage elements within any planned transition projects on islands. Also, interconnectors to mainland grids are critical where they exist. Equally, because islands have complex governance arrangements, there is a requirement to ensure that the various levels of regulatory oversight (EU, national, regional, archipelagic, and individual island levels) work to support rather than undermine each other.
A key insight made here is then that individual islands adopt innovative energy projects as part of a wider regional, national, and international policy 'regime': the regulatory and policy system of rules, subsidies and decisions on energy types and fuel mix. Crucial here is the extent to which any regulatory regime sends a clear signal that it supports islands as sites of energy innovation.
The good news is that the EU as a regulatory actor is very distinctive in having specific policies for the promotion of energy transitions at the level of islands which are supportive in comparison to traditional policies. The Commission has promoted a Clean Energy for EU Islands initiative [34] which is part of the Clean Energy for All Europeans package. This policy notes that there are more than 2000 inhabited islands within the EU states' jurisdiction and establishes a secretariat and a forum 7 to facilitate collaborations for the sharing of best practices to encourage energy transitions at the level of islands. More importantly, this policy sets up a dedicated financial instrument of over €100 m to fund 60 projects out to 2023 under the New Energy Solutions Optimised for Islands (NESOI) project team [35]. In this regard, the EU is dedicated to furthering islands in their ambitions to become sites embedded in the ongoing energy transition. This positive 'top down' signal and support is arguably vital as it empowers influential local and regional stakeholders (financial, political etc) with the confidence to be ambitious.
What is less evident, however, is EU-level specific support for solar fuels on island sites, rather than a more general backing of innovation that moves towards some type of energy transition. Some NESOI-funded projects [36] do have a solar dimension: the Croatian Solar Islands plan; the CEL-EBRe project at La Palma, Spain; the DGReS-Aegean project, Greece; FESOL project in Lipari, Italy; and the Canary islands; SoFIA for the city of Adeje. While these projects explore PV technologies, they do not appear to go further and explore the potential of solar fuels.
Aside from the role of the EU, path dependencies in national energy policy are also a significant factor and they can determine the scope for ambition at the island level. Some islands will likely choose blue hydrogen pathways, producing H 2 using carbon capture and storage (CCS) over green hydrogen, produced by electrolysis from renewable energy because this fits better with national policy. In some countries, notably the UK, underground storage is abundant (caverns or salt mines) and British national policy favours a 'twin-track' approach of blue and green H 2 production as does the partially autonomous Scottish government [37]. However, for the majority of islands, technological conditions for CCS implementation are scarce whereas renewable sources are plenty. Therefore, the scope for green hydrogen is usually greater. Similarly, some islands may struggle with grid connection infrastructure. Here decentralized and innovative forms of hydrogen production such as PEC systems would seem ideal.
We can easily see here that the national stage remains critical at the regime level, and in the context of the case study here, Spanish energy policy has many unique characteristics. Spain already has a high level of renewable generation of electricity but additional investments must be made to fulfil Spain's target forecast of increasing its renewables capacity to reach 74% of electricity generation by 2030 [38]. In 2016, Spain imported nearly 1.3 million barrels of oil per day, leading to an energy dependency of 73.9% in 2017-much higher than the EU average.
In the case of Spain's islands, the Canaries are unique as they are not connected by cable to the Spanish mainland grid, due to their distance from the mainland. Instead, the archipelago exhibits a system of six distinct island grids [39]. Moreover, most of the Canaries' energy needs are generated by conventional oil and gas-fired plants. The national Spanish grid company, Red Eléctrica de España, remain the dominant actor responsible for the electricity system on the islands and has accordingly focused on achieving grid stability by mainly investing in very expensive pumped water storage systems [40]. More generally, Spain at the regime level exhibits rather typical solidarity rules such as imposing common electricity tariffs across its jurisdictions regardless of the higher economic costs associated with lower economies of scale in island settings, which means that mainland Spanish consumers subsidise island electricity consumers who otherwise would have to pay more. This is a good example of how regime-level rules support island-level energy niches. We can also note that as regards regime-level policy responsibility, Spanish islands usually enjoy some level of political and legal autonomy. In the case of the Canary Islands, an archipelago-wide energy plan has been in place since 2015 [41], which is being augmented in 2022 by an Islands Sustainable Energy Strategy valued at €467 m, drawing on national Spanish and EU post-COVID-19 recovery funding [42].
At the landscape level within the MLP analysis, we find major structural and social changes which have macro-level effects as drivers; these could be major geopolitical events (globalisation) or technological changes (internet, mobile digital devices, etc) but also wider social shocks such as climate change and more recently, the COVID-19 pandemic. The argument here is that system-level drivers for change exert pressures and send signals to regime and niche-level actors which can either support or slow any technology transition. The most obvious landscape driver relevant to solar fuels is the ongoing climate change crisis and the need for industrial societies to find new sources of energy that do not generate GHGs.
These landscape trends rarely bring about change simply by themselves. It usually requires actors at the regime and local niches to engage in innovation that is 'reinforced' by landscape events. Much of this comes down to timing: 'If landscape pressure occurs at a time when niche innovations are not yet fully developed, the transition path will be different than when they are fully developed' [31]. Nevertheless, there is some scope for actors at the niche and regime level to consciously form alliances which anticipate landscape pressures or at least 'shelter' niches until the time is ripe for further exploitation. A key part of such 7 The forum has been meeting since 2017.
nurturing is the running of pilot projects that trial technologies at an early stage of pre-commercial development. It is in that context that we turn to the case study of this paper, to see to what extent it exhibits some of the positive trends identified in the MLP approach: • Support and participation from local island communities; • Sheltering behaviour at niche level through subsidies and legal permissions; • Engagement with a favourable national and EU-level regime through supportive rules, laws and access to funding; • Engagement with wider landscape effects, notably the urgent need for climate change-neutral energy; • A focus on developing new technologies (solar fuels) in anticipation of future opportunities.

The Canary Islands as a hydrogen laboratory
The case study discussed in this section stems from the funding support from EU-level agencies such as the INTERREG programme, which is focused on interregional collaboration amongst geographical areas with common challenges and opportunities. SEAFUEL is an INTERREG Atlantic Area project under the 'Resource Efficiency' priority topic. The project began in December 2017 with the primary objective of demonstrating that natural resources in islands could be used to power local fleets of vehicles using green hydrogen as a zero-emission fuel. The project was particularly interested in isolated islands such as the Canary Islands since there is a significant cost to the regional and national governments for fuel imports. By utilising indigenous resources, such islands have the potential to become sustainable and self-resilient without the need for fuel and/or energy imports.
However, initially, the project was complicated due to limited interest in hydrogen technologies back in 2017 and 2018. Even with the climate change pressure, Paris Agreement and IPCC reports, there was still not a significant drive for renewable energy, renewable fuels and chemicals. The challenge of being a small island, with relatively low amounts of industry activity, prevented stakeholder engagement with the project. Larger industries saw a niche market, with no short-term return on investment, leading to limited interest in becoming a partner.
Nonetheless, a major driver for the change in perceptions, interest, and potential investments came from the EU level, with a definitive push towards renewables and hydrogen as a way to achieve the 2030 and 2050 climate goals. For hydrogen, a crucial moment was the publication in 2020 of the EU Hydrogen Strategy (2 × 40 GW), followed by the Fit for 55 package, all aligned towards achieving carbon neutrality in Europe by 2050. Many Member States followed this with the publication of their hydrogen national strategies, including Spain in 2020. However, these national strategies typically focused their efforts on continental territories, with a strong focus on decarbonising industry and industrial hubs. Spain was no different, with the national strategy's premier target being the replacement of 25% of industrially consumed hydrogen with green alternatives by 2030. This has again become a new challenge for the implementation of hydrogen technologies in isolated islands such as the Canaries. A more bespoke strategy needs to be put in place to tackle the unique challenges apparent in such island territories.
Between 2018 and 2021, intensified efforts were carried out by SEAFUEL project partners to disseminate and communicate the objectives and results of the project across the different regions covering the Atlantic Area. This was implemented via in-person and online workshops, reaching a wide variety of stakeholders across the whole supply chain. Figure 2 shows a scheme of the project's demonstration plant in the Canary Islands, which is located on the island of Tenerife, close to Tenerife South airport, within the facilities of the SEAFUEL partner, Instituto Tecnologico de Energias Renovables (ITER).
The choice of location was extremely important to showcase that green hydrogen can be produced only using solar energy and seawater. SEAFUEL's plant is the first example worldwide of such as installation, where the hydrogen production facility is contained with the refuelling infrastructure. ITER's facilities have over 30 MW of PVs installed as well as two desalination plants which offer potable water to their facilities. In addition, ITER has a wind park and is in charge of the maintenance of several solar and wind parks across the island of Tenerife. For the maintenance work, diesel-powered vans are used to reach those locations, some of them with difficult access and steep roads, making the replacement of those vans via battery-electric analogues very challenging. On the other hand, fuel cell electric vehicles (FCEVs) or range-extended electric vehicles offer a suitable alternative-using hydrogen to extend the range of the vehicle as well as utilising the power of a fuel cell for the van.
Some of the most important barriers for the project were directly related to the niche market in Tenerife, which is an example of where the geography of islands works against them. For example, car manufacturers did not show any interest to supply FCEVs for the project since the cost of transporting the vehicles, and setting up a skilled workshop on the island for the maintenance and repair of the FCEV, etc was deemed too Figure 2. Schematic representation of the SEAFUEL project, using solar energy to power a hydrogen fuelling station using seawater, and using the fuel for a fleet of vehicles [18]. Reproduced with permission from SEAFUEL. Image Credit: Lucia Villalba. Figure 3. Hydrogen production at the SEAFUEL refuelling station in Tenerife [18]. Reproduced with permission from SEAFUEL. Image Credit: Lucia Villalba. high to be commercially viable. As the project evolved and with the increased interest in hydrogen technologies worldwide, Hyundai Canarias and Toyota Spain were attracted to join the project and offer available Hyundai Nexo and Toyota Mirai models for the project implementation. These vehicles are particularly attractive for the second goal of the SEAFUEL project, which emerged in the early stages of the project, and targeted the tourism sector. In the Canary Islands, over 75% of the local economy comes from tourism in the islands. Transportation is the major carbon dioxide emitter, with a significant portion of vehicles directly or indirectly connected to this sector. Therefore, SEAFUEL aimed to use the cars to study the potential interest of tourists visiting the island, with a series of surveys. They also offered, for the first time, the option to test drive these models so that tourists could acquire first-hand the experience of driving a FCEV. The main driver for such is that the car rental sector in tourist islands is enormous and offers an excellent real-life laboratory to test these new technologies to a range of communities. A SEAFUEL study shows that even though some early adopters would happily pay a premium to drive a FCEV while on holiday, both the novelty of the approach, the relatively low cost of renting a car for a few days, and potential incentives from the regional government, opens up the impact of FCEV to a much wider population.
Another major challenge for the deployment of the technology was the installation itself ( figure 3). Firstly, being the only example of a hydrogen production and delivery facility in Tenerife meant that planning permission was delayed several months since the technical staff in charge of the certification had no training in hydrogen technologies. This barrier was circumvented with collaboration and communication, a major aim of the INTERREG programme. Partners from other regions in Europe, with experience in hydrogen project delivery and implementation, prepared an extensive set of documents detailing all the parts and operational knowhow of the installation, along with the relevant EU certificates, usually seen in countries such as Germany and UK, where a series of hydrogen refuelling stations are already in operation. All technical questions for the administration were answered to get the final approval for the civil works required for the installation. Secondly, the fact that the Canary Islands are isolated from continental Europe also means that they have their own taxation and import system, which means that any new import needs to be validated before being transported to the Canary Islands. For SEAFUEL, additional challenges in this area were met. The partner responsible for the construction of this infrastructure was based in the UK, and due to Brexit all imports to EU member states suddenly became much more complicated. Finally, once the installation was in situ and the commissioning started, there emerged an unexpected technical problem in one component of the system. Here, the third major challenge emerged because there were no locally available trained technicians to fix hydrogen technologies, and even with small faults, experienced technical staff needed to be flown in from continental Europe, with a concomitant large cost for any maintenance or repair works.
Apart from the technical demonstration of the project, SEAFUEL wanted to draw attention to the specific requirements of islands to achieve climate change targets as well as to decarbonise the transport sector. For small vehicles the electric option is preferred, as long as the grid infrastructure can withstand the increasing electrical demand and the needed increase in renewables. However, the car rental market for FCEVs offers an alternative to local populations by potentially driving demand for hydrogen production and refilling infrastructure led by the domestic touristic sector. This can bring FCEVs for the tourists while creating a new second-hand market considering the average lifetime of car rentals is a maximum of 5 years. In addition, hydrogen is a clear option for public transportation in Tenerife, where the roads and connections make it unsuitable for electric buses to operate. This and some other challenges and opportunities specific to Tenerife are explained in detail in the SEAFUEL hydrogen roadmap specifically developed for the island of Tenerife. In that sense, it is extremely important to have close contact with the regional authorities so that a set of recommendations can potentially be considered and brought within the Sustainable Energy and Climate Action Plans.

Conclusion
To what extent does the experience of the Canaries confirm the MLP perspective employed here that the successful adoption of new technologies, in the case of solar fuels, requires drivers at multiple levels? Moreover, does the case study also fit with the broader idea of islands as being laboratories for new renewables technologies? We would affirm that in both cases the answer should be broadly affirmative.
The most significant feature of the SEAFUEL project in the Canaries is that it has delivered the first European-integrated green hydrogen production plant based on electricity produced locally on the island from solar or wind. That there have been rather predictable problems in executing this complex feat of engineering or other delays is in many ways to be expected. Moreover, these challenges reveal important lessons that can be learned about both the roll-out of further solar fuel plants and the production of green hydrogen in Island settings. The importance of building a local technical support and maintenance infrastructure is one clear lesson as is the integration of the energy output (in this case vehicle mobility) within a wider island economic context of greening tourism mobility. This case study also confirms the importance of planning as an important constraint, although here it served more as a delaying factor rather than a variable which blocked progress. However, in other islands, especially where ecologically sensitive zones are numerous or adjacent to proposed infrastructure, then planning may present more formidable obstacles if we consider that what is required is not just the hydrogen production and storage facilities but also solar and perhaps wind generation sites. The latter might be distributed on land or even in coastal waters which permits some flexibility, albeit at likely higher project costs.
As regards the MLP perspective, we see here that major landscape events, such as increasingly climate change-driven energy policies, by themselves lack the force to cause innovation. What was critical in this case study was the opportunity presented at the regime level, to be precise through EU-funded programmes and norms which both empower and stimulate local, regional, and national energy policy entrepreneurs to propose demonstration projects. EU-level support remains very tangible and salient such as the Clean Energy for EU islands secretariat launched in 2017. Also vital at the regime level was the support of national electricity grid operators, the Spanish REdE, who were prepared to support the project and its integration into a very complex and quite precarious local islands' grid which is not connected to the mainland. The increased ambition of these regime-level actors, based on successful demonstration projects, can now be seen in the development of larger integrated island-based hydrogen projects such as Green Hysland [30]. The latter will implement solar-based green hydrogen production in Mallorca to reduce emissions in mobility, the built environment, and industry. The project has directly benefited from the groundwork laid by SEAFUEL [18]-in terms of the policy, planning & regulation, and the operational know-how required to realise further island-based deployments of green hydrogen solutions. Therefore SEAFUEL can be said to have directly contributed to increasing the interest and confidence in the hydrogen niches to unlock higher levels of funding and further regime-level expansion from both the public and private sectors.
We also see 'sheltering' behaviour from another important regime-level actor, global car manufacturers, who have pioneered hydrogen fuel cell vehicles and were prepared to step in and offer models when this threatened to be a major problem for the project. Finally, there was a supportive policy niche on the Canaries which was led by the local ITER institute in conjunction with the local government for the islands. This is an example of a non-commercial niche actor strong in technical know-how, although they were also able to work with local businesses, notably in the tourism sector.
In summary, there is considerable scope for synergies and co-location of solar renewables on islands together with the production of solar hydrogen. For now, this appears to be a long-term prospect, yet islands provide a developmental space to explore this opportunity. We note however, that islands are not without their own challenges because of scale and remoteness: in the Canaries, there were problems in getting partners to be interested in deploying FCEVs to the islands for these reasons, although these were resolved.
As previously discussed, islands are more likely to be fossil-fuel dependent. Solar infrastructure on islands must inevitably overcome constrained spatial availability. Moreover, islanders tend to have greater place attachment, and opinions can vary wildly from considering any intervention to be either a detraction of an island's characteristics or as a symbol of progress towards clean energy and a more sustainable economy. Nevertheless, the same literature suggests that eco-design and initiatives involving cleaner industrial processes and energy can compensate for any possible support loss due to visual impact [43].
Ultimately, islanders, and island visitors, have a profound connection with nature and green initiatives that is more immediately experienced than on the mainland. We note that in the Canaries, like in other islands, a decisive feature in support of the project was its promise of at least partly greening local tourism transport demands. Seeing tangible outcomes for the local population is probably decisive in securing local support. Furthermore, the lack of opportunities to host new industries, population size and distance from major markets makes islanders more likely to accept possible disruptions and negative aesthetic impacts provided by marine renewables [44].
In conclusion, green hydrogen and solar fuels provide an excellent opportunity for the decarbonisation of islands, considering their abundant renewable resources. Their typically precarious connections to mainland grids, often the source of energy restrictions, can drive islands to seek more disruptive and decentralized forms of energy technology, for both electricity and fuels production. Islands will need integrated and trustworthy systems featuring comprehensive models including hydrogen production, storage, transportation and utilisation. Beyond early-stage outputs such as fuel cells for tourist vehicles, more ambitious goals will surely evolve. However, balancing supply and demand on an annual basis for islands is not simple, yet renewable and hydrogen storage should be used to balance seasonal inputs, bringing stability to the network. Getting this right will be a major challenge for all countries and the wider energy sector over the coming years, and islands may well provide some of the most useful lessons as part of that journey.

Data availability statement
All data that support the findings of this study are included within the article (and any supplementary files).