Use of the Delphi method to assess the potential role of enhanced information systems in Mediterranean groundwater management and governance

Sustainable groundwater management is one of the most significant water challenges facing humanity in the 21st century (IAH, 2017). Climate and demographic projections generate double pressure on groundwater (Guppy et al., 2018), and there are many well-documented experiences of unsuccessful groundwater management across the world. Many states have failed to develop groundwater command-and-control policies for a variety of reasons, frequently combined, including user resistance; the logistical difficulty for administrations to maintain up to date and accurate systems of monitoring and sanctioning, or to fund financial compensations; or cases of corruption and lack of legitimacy (Molle & Closas, 2020). A key element in many of these negative experiences is the lack of transparency and mismanagement of groundwater-related information.

This is largely because groundwater has some unique characteristics compared to surface water, such as complexity, multiple independent users, distributed occurrence and use with open access opportunities, and, mainly, invisibility. It is no coincidence that the UN used ‘visibility’ as the focus and motto of the recent international groundwater year (UNESCO, 2023).

Data (raw records captured/stored) are the basis for information (processed relevant data with a message). Knowledge is produced by combining information with previous experience and personal perceptions, which leads to wisdom from which projections and management solutions are derived (the so-called ‘DIKW’ hierarchy, see Liew, 2007; Van der Gun, 2017; Clark, 2019).

Information and transparency on aquifer management could be considered a determinant factor in bringing light to groundwater exploitation (Reddy & Syme, 2014; Villholth & Conti, 2019). This is because it strongly influences the perceptions, decisions, and, thus, behaviours of the different groundwater users. Sharing information on the nature and status of an aquifer and ensuring transparency of data flows is critical to building trust in management to facilitate collective action in groundwater management (Shalsi et al., 2019; Stone, 2019).

Besides, long-term monitoring of aquifer water level is essential for sustainable groundwater management, yet often, this information is still insufficient and not easily accessible. Despite the progress made since the second half of the 20th century, the lack of robust information on groundwater and its use is probably one of the most important data gaps in a global context. Groundwater is a fundamental – yet often unknown – element on the hydrological cycle. This is particularly serious in some arid and semi-arid regions of the world (WMO, 2023), often with a greater dependence on groundwater resources. Moreover, traditional means of obtaining and sharing groundwater data, such as water meters and piezometers controlled by state agencies (with or without the cooperation of stakeholders), have been widely proven to be largely ineffective for good groundwater governance and the prevention of overexploitation (Molle & Closas, 2020, 2021).

Beyond these traditional methods, new ways of collecting data have been developed in recent years, opening the door to improving the quality and quantity of hydrological information in general and specifically for groundwater. We are living through a revolution in the availability and use of information. Computers, mobile phones, the internet, remote sensing, dashboards, platforms, and other technologies and their associated institutional arrangements have already brought profound changes that are still working their way through society. In agricultural management, groundwater uses, and other domains, this information revolution promises to continue, and it could even be compounded by further changes such as the availability of cheap sensors, networked on the Internet of Things (IoT), precision agriculture, computer-based modelling, and, as a recent trend advances in machine learning and artificial intelligence.

New technologies, particularly in telecommunications and IoT sensors, greatly increase the amount of groundwater data that can be collected, especially in remote and hard-to-access areas (Sharples et al., 2020). This set of techniques and methods is often integrated and referred to, with greater or lesser precision, by the term ‘enhanced information systems’ (EISs). In relation to groundwater management, we use the term when (i) traditional means of data gathering (basically water meters and piezometers) are complemented by innovative techniques such as earth observation systems (drones, remote sensing), automatic sensors (soil moisture capacitance FDR, surface renewal), use of tools based on Information and Communication Technologies (ICTs) (e.g., mobile apps) and (ii) particularly when there is a strong engagement of citizens and/or stakeholders in data collection and use (citizen science). Therefore, EIS for groundwater is characterised by a fluent exchange and use of information and knowledge between stakeholders, modellers, and decision-makers on the status of the groundwater systems and possible management strategies. Engagement in citizen science is pivotal in EIS for groundwater because it can provide in-depth learning opportunities and promote the users' ability to understand and deal with variability and uncertainties to help manage complex systems such as groundwater.

The emergence of these techniques may turn groundwater management and governance at large on its head in the immediate or medium-term future. But there are still important unknowns, many of them inherent to any process of social adoption of new technologies, to the necessary bricolage processes, and their insertion in different social, agro-climatic, and cultural contexts. There is still a significant degree of uncertainty on the success or failure of these solutions, which move from the risk of falling into an excessive technoptimism that leads to an overestimation of their capacities (Huesemann & Huesemann, 2011; Danaher, 2022) to the resistances of some stakeholders and institutions to underestimate user capabilities and thus fail to see EIS's potential to help increase institutional capacity.

For these reasons, in this paper, we analyse the potentiality of these tools and how they can improve groundwater management and governance, considering their potential benefits and the obstacles to their implementation and development. To carry out this prospective analysis, it is necessary to consider aspects related to the interaction between technology and information (socio-technical systems) and those associated with the institutional context (socio-ecological systems) (Table 1). The first interaction leads us to consider the characteristics of the data and their availability on the one hand and the available technologies and their potential for adoption by users on the other. The second interaction forces us to explore the relationship of information with institutional tools (normative, organisational, and procedural), such as command-and-control policies, capacities for groundwater monitoring, rule enforcement and institutional design, co-management formulas and the potential for user empowerment. For this prospective analysis, we have used the Delphi method, which uses expert predictions to identify trends and provides a different and complementary perspective to other methodological approaches (Marchais-Roubelat & Roubelat, 2011; Rowe & Wright, 2011; Beiderbeck et al., 2021). In this way, we aim to deepen the understanding of the role that information – arising ‘from social realities created by humans’– can play in the use and management of natural resources (in our case, groundwater) (Von Wehrden et al., 2017).

We carried out this Delphi survey on a global scale, but also with a particular focus on the Mediterranean region. The Mediterranean region is very vulnerable from the point of view of climate change, as reflected in the recent IPCC AR6 report, with an extreme dependence on groundwater (EASAC, 2010; UNESCO-IHP, 2015; Fader et al., 2020). Through the Delphi survey, we tapped into expert knowledge, gathering information by asking experts to reflect on different elements of the complex interface of groundwater and information systems. The following section describes the methodology used, the third section presents the results obtained, and finally, a discussion of the potential and limits of EISs is elaborated, and key conclusions are drawn.

Our methodological approach consists of a Delphi study conducted in collaboration with PROSPEKTIKER, part of the Millennium Project, a global reference project in foresight studies by the UN and the EU. A Delphi survey is a participatory foresight methodology for collecting and synthesising expert opinions originally developed by the RAND Corporation in the late 1950s (Renzi & Freitas, 2015). The objective of a Delphi survey is to identify convergences of opinion and/or generate a reliable consensus opinion of a group of experts via a process of questionnaires interspersed with controlled feedback. The method is intended to provide clarity to experts in areas of uncertainty and assist in decision-making.

The Delphi survey method has been the basis for soliciting inputs and feedback from experts on groundwater using both long-term and foresight perspectives (Taylor & Ryder, 2003). There are very limited Delphi studies or Delphi related methods in relation to groundwater (Arasteh & Farjami, 2021; Dang Tuyet, 2021; Majidipour et al., 2021). However, the focus of our Delphi study is different since it is centred on the role that EIS can play in relation to groundwater management. Therefore, we aim to generate potential recommendations on how EIS can support sustainable participatory groundwater management (the role that information plays in groundwater – i.e., an interface of two commons). We focus directly on information and data through Delphi survey questions on the type of groundwater data used, the uses of technologies for monitoring groundwater, the adoption by users of different types of emergent technology, data availability, and citizen science helping to compensate for the lack of groundwater data.

In terms of the strengths and weaknesses of the Delphi method, Delphi studies are challenging to perform well. The questionnaires must be well prepared and tested to avoid ambiguity. The primary strength of Delphi is its ability to explore issues that require expert judgment. A weakness of the Delphi method is the ease with which questions can be asked for which better techniques exist. In short, Delphi is a powerful technique when used to seek answers to appropriate questions. Another issue to consider when analysing Delphi results is that it can engage with ‘epistemic communities’, as Haas (1992) put it, where experts despite a diverse background share a knowledge base.

We opted for the Delphi method in comparison to other foresight study methods, such as scenario analysis (for understanding the robustness of strategies) and integrated models (for understanding complex systems), because we wanted to focus more on how to gather the most possible scenarios of a specific topic based on specialists' knowledge. Other examples of Delphi method application to a diverse range of water governance objectives can be found in the literature, such as Ward et al. (2018), which examined expert opinion on the current capacities and future priorities on the connections between hydrometric data/databases, water resources models and decision-making, and Antonelli et al. (2022) which analyse the trends and policy interventions in the agri-food system from North and South Mediterranean countries or how to integrate landscapes into land use planning (Martín & Yepes, 2022). While these studies bring valuable insights on related topics, our Delphi was designed to frame the main trends and relationship between groundwater policy at the macrolevel (Mediterranean region), groundwater data, information and knowledge for decision-making and collective action.

Reflecting on the method, the key to a successful Delphi lies in the selection of participants. Since Delphi's results depend on the knowledge and expertise of panellists, it was essential to include persons who were likely to contribute valuable ideas and expertise. In implementing the Delphi survey in this project, a database including experts on groundwater from different backgrounds (academic, policy makers, private sector) was created.

We then formulated the statements and questions in the survey (see supplementary material). The survey statements were selected based on the most critical topics identified through a research consortium workshop in June 2021. At the workshop, partners had the opportunity to propose and discuss key topics on groundwater and EIS and to formulate questions for inclusion in the survey. As an outcome of the workshop, we identified the most important topics in the interface between groundwater and EISs. We also looked at potential disruptive events (i.e., events with low probability and high impact), an important part of the Delphi methodology as a foresight technique.

Survey questions were designed to be comprehensive, seeking to solicit wholly or mainly quantitative answers, in conjunction with open text for additional qualitative comments. Questions in a Delphi method are based on specific hypotheses, wherein experts reflect on those hypotheses and several other elements. These elements included the possibility for the respondents to reflect on their level of expertise on the topic, the relevance of the topic/issues, the impact of EIS on groundwater management, and the probability of occurrence. Follow-up questions sought more specific details where possible. The questionnaire was tested to receive feedback and improvement proposals before launching the final survey to participants.

Herein, we report the results of the questions that addressed how information and data affect groundwater commons management and the probability that these informational instruments will occur. These questions looked at the type of groundwater data, the adoption by users of emergent technologies for monitoring, the different technologies themselves, data availability, and how citizen science could help when faced with a lack of data.

Here again, we looked a bit deeper on the type of data to have a more detailed understanding of the type of data relevant to an enhanced information system in groundwater (see Table 2) based on a typology developed by SADC-GMI.

Furthermore, experts commented how the combination of all three types of data is essential. Only complete data sets will allow us to fully understand the underlaying causes, and therefore, would permit taking corrective actions. In practice, however, information is usually limited, and has many uncertainties, including both errors in measurement and differences in epistemology, i.e., different models used by different experts are common (Bruns, pers. comm). In addition, information on aquifer characteristics (specific yield, transmissivity, storage coefficient, etc.) is also required. Data on groundwater use can be easily obtained (through meters) and shared, but this is not the case for data on water levels and aquifer properties. Some experts also argue against weather data and stream-flow data (supporting information) collected by citizens/users, which, in their opinion, should be collected by technical agencies. There are developments regarding the different types of data in Mediterranean countries. However, in the case of well-abstraction monitoring, the situation is different. As a rule of thumb, experts commented that information should be collected and used if it is easy to collect and does not interfere with usual activities.

Regarding qualitative results, experts refer to the importance of adequate staff and funding. Experts provide an important insight by differentiating between data collection vs. data availability vs. EISs as defined in this paper. It refers to the interaction between different ‘information’ elements in the system and how it is the combination of these different data types (piezometry, pumping and weather data) that are needed to make better decisions, i.e., the ‘enhanced’ part of information systems. Furthermore, users should ideally have sufficient knowledge to select the type of information system that is best for them to use. It is not just a matter of having access to or collecting the data themselves but of choosing the best way to share the information with users or between users. This is one of the keys to EIS's successful use and development.

The main reason is that citizen science can provide complementary data but cannot be the source of the primary data. In other words, data scarcity cannot be resolved with voluntarism. Citizen science has an important role to play in education. In some cases, for example, like the promotion of interests and the visualisation and socialisation of problems. However, experts stress the need for data to be collected in a replicable manner. Success stories must be created.

Experts, therefore, expressed how a clear differentiation must be made between the generation of data and the use/processing of data for evaluating scientific problems and solutions to the problem. Thus, data collection must be done through a superstructure above all single collectors, i.e., a task or function of, e.g., national agencies.

87% of experts consider that the impact of EISs on the adoption of emerging technologies (e.g., on groundwater quality and quantity) is high or very high. Only 3% consider it as low. This is because the data to be received will be useful for groundwater management.

In relation to the probability of adoption of emerging technologies for monitoring groundwater quantity and quality and dynamics, answers show that for most experts surveyed, the adoption of emerging technologies is probable in the future, especially when talking about the Mediterranean region.

In fact, in the Mediterranean region, 86% of experts consider the probability of the adoption of emerging technologies to be high or very high in 2100 (only 22% for the 2030 horizon). When assessing this probability at global scale, the percentage decreases to 69% in 2100 (23% in 2030).

There are several reasons given for these forecasts on technology adoption, like for example, new technologies replacing old with cheaper ones, more precise tools, etc. Experts concur that emerging technologies will be adopted as an element of understandable progress.

Where experts disagree and there are divergent opinions is whether this adoption will take time to upscale, or whether it will be fast.

This is partly explained by barriers that were identified. For example, even if the technology is available and the ratio $/unit of captured data decreases, the cost and capacity for management and quality control will increase. These technologies may require important amounts of investment, and most of the countries have limited economic means to invest and/or have significantly restricted access to emerging technologies. An interesting comment from one of the experts is the challenge with this topic to deal with the ‘Internet of Threats’ not just the IoT from the various technologies. A further insightful comment was provided by another expert whose data itself are not sufficient, i.e., a lot of scientific data is generated through technologies; however, it is hardly used for decision-making.

Here, we report the results on the Delphi survey questions that were asked that dealt with groundwater and other institutional tools for groundwater governance. Specifically, two survey questions looked at regulation (norms and rules) by looking at the role of the State in groundwater management (command and control), and the monitoring, rule enforcement, and institutional design by groundwater users; two survey questions looked at organisational tools like co-management and bottom-up inclusive groundwater management, and finally two survey questions that looked at procedural/social learning aspects like investing in capacity building and stakeholder empowerment and citizen science.

One expert pointed out that a great diversity of situations must be considered, depending on each country's socio-political-legal conditions. This analysis refers to the connection between this enhanced information and the increased ability of users to enforce rules, users' involvement in monitoring, and the involvement of users in institutional design.

Overall, as a bundle on capacity development, experts see a possibility (at around 63–65% probability of occurrence) that users will engage in rule enforcement, monitoring, and institutional design for the longer time horizon to 2100, while evaluating the possibility that they will do so to 2030 as still relatively low. In fact, 18% of respondents think that enforcement of rules and monitoring by users will never actually happen. Where there is a high level of consensus – presenting a strong foundation for government support – is around investment in users' capacities. Thus, in terms of policy design for groundwater management, investing in EISs for users and for the administration itself seems a safe bet, even though it is unlikely to be sufficient in and by itself, unless it is reinforced with other key elements, such as strong or clear regulation and norms on groundwater use.

It is the contrast between the desirability of this happening versus the probability that it would happen, which is striking. Comments such as ‘romantic view’, and ‘optimistic answer’ revealed expert scepticism that inclusive policies would really mean effective power sharing and are not just formal or even tokenistic improvements in management.

It is unclear whether the sectors that currently have power over water resources will allow the sharing of power. It may become ‘mainstream’ in policy documents and official discourses, but it is less clear how this would be implemented in practice. According to an expert, increasing scarcity will lead to a struggle for resources and not to participatory procedures. In fact, whether scarcity leads to more struggle/conflict, more cooperation, or both, and how this may vary is a crucial question that probably needs further research (Bruns, pers. comm.). In short, bottom-up approaches and power-sharing may grow but it will not easily become ‘mainstream’. One expert shared an experience that cast doubt on the endurance of change: ‘We have experience from India with farmer groundwater schools, etc. It has not produced any positive outcomes. Once the project is over, everything falls apart’.

Therefore, and to summarise, in relation to how groundwater users and governments can improve institutions for governing groundwater, our results indicate that EISs can indeed help but are not sufficient by themselves.

Thus, there is a relatively high level of agreement that suggests that EIS will facilitate co-management, i.e., experts agree that information will probably contribute to a better understanding and promotion of co-management.

In terms of qualitative comments, an expert states that all mechanisms of public participation will require an EIS since co-management implies a wide political understanding that also depends on the managers' decisions.

In relation to stakeholder engagement in the collection and sharing of information, as part of an enhanced information system, several factors were raised by experts that would need to be considered. These are: (1) the need to engage authorities, (2) the lack of mutual trust between stakeholders and institutions, (3) the need for adequate resources in terms of time and education (e.g., piezometers have to be protected and read, and this need and skill must be integrated into school curricula, etc.), and (4) how the use of technology could empower citizens and thus, increase social justice.

Experts, in fact, predict that capacity building is becoming a key important topic in groundwater management. For example, in some countries like India, Pakistan, Spain, Mexico, or the Western US where groundwater use is intensive, the governments are already investing a lot in citizen education. New technologies will probably make it easier to collect and coordinate efforts for citizen science, and therefore, capacity building for citizens will increase in the future. Each year, education is a higher priority where effots are put to be constantly improved.

Other experts, though, seem to think that this is wishful thinking, and that public institutions (in general) have other priorities for the coming years. Where there is agreement is that EISs – if developed – can help increase capacity.

Our results provide a series of expert foresight insights, some of which confirm empirical observations from the scientific literature – from the few groundwater EIS experiences studied and documented – and others show interesting discrepancies or nuances. Overall, the experts agree on a positive assessment and highlight the key role that EISs will play in groundwater management. Emerging technologies are seen as a likely and important future trend for processes and functions such as monitoring groundwater quantity, quality and dynamics. Beyond this Delphi analysis, there is widespread consensus on the role of ‘Fourth Industrial Revolution’ technologies in articulating socio-technical systems that can improve natural resource management (World Economic Forum, 2018).

However, what emerges from this foresight study is that the processes of adoption of EIS are slower than might have appeared, and carry the not-negligible risk (nearly 1 in 5) that they may never materialise. Experiences documented in the scientific literature confirm this. The adoption and development of EIS and citizen science will take time. Projects may require a slow process of trust-building and making visible outputs and results, which can take more than a decade. This makes EIS particularly challenging for governments and potentially unattractive if quick results are expected. Experts see a possibility (at around 63–65% probability of occurrence) that users will engage in rule enforcement, monitoring, and institutional design for the longer time horizon to 2100, while evaluating the possibility that they will do so to 2030 as still relatively low.

In fact, 17–18% of respondents think that enforcement of rules and monitoring by users will never happen. This assessment is not surprising, as in some recent experiences, stakeholder engagement processes have only elicited positive responses from a small proportion of target users, due to high coordination costs or different physical barriers (Cooperman et al., 2021). Literature highlights the importance of providing feedback to users to encourage them to participate in EIS projects, because in most cases users will not voluntarily engage in groundwater monitoring unless users derive direct benefit from this (Chang et al., 2022; Chu et al., 2023; Ouedraogo & Rinaudo, 2023). Other contextual factors can facilitate the engagement of stakeholders. Critical situations such as a severe drought or an increase in water salinity levels have been shown to be cathartic events that generate collective action by users. This predisposes users to incorporate EISs that can help streamline water resources management and provide greater transparency (Barnett & Rinaudo, 2023; Ouedraogo & Rinaudo, 2023; Rinaudo et al., 2023).

This brings us to our concept of EIS as a system, since the adoption of enhanced information cannot be limited to the reception of parachuted technology. Without an appropriate social and institutional context, which needs to be designed, incentivised, and proactively nurtured, these EIS initiatives by themselves will remain as technoptimistic wishful thinking that distract from actual governance issues. Therefore, the adoption of EIS needs a deeper analysis to better understand a heterogeneous and multi-factorial process, which is deeply dependent on several – often context specific – variables. Here the general principles and procedures linked to the characteristics of the data and the properties of the devices used to obtain this information provide a clear window for sharing and disseminating data and information and thus, sharing power or re-arranging structures in a different constellation; yet these windows can only truly open if due attention is given to the social and institutional frameworks in which the adoption takes place; and the characteristics of the groundwater resources system.

While the Delphi analysis does not provide a complete characterisation of these issues, it does point to some key factors.

Regarding data characteristics, the experts interviewed in the Delphi survey indicated that not all data are equal. It is much easier to collect and share some types of data, such as supporting information (i.e., river flow gauging, meteorological observations, and satellite land use surveys), than other types. Half the experts agree that it is probable that collecting and sharing data on groundwater occurrence and aquifer properties (i.e., groundwater level monitoring and groundwater quality monitoring) occur. Meanwhile sharing data on groundwater use (i.e., water well-abstraction monitoring and groundwater level variations of wells), will take much longer to materialise. This could be explained because data on wells, or data on groundwater levels can be very sensitive. Thus, it is cautionary to note that one-fifth of experts consider that it is probable that collecting and sharing these types of data will never happen, not even when considering a longer time frame.

The key differentiating factor may lie in their influence on power balances and on the ability of users to maintain or lose control over the resource. As another expert comments, ‘it is more uncertain for those devices that disrupt the social organisation/distribution of economic interests’. Put another way, although technology allows the sharing of data and information; information is power and this greater access to information can also provide greater access to power (or empowerment of more/other actors). According to experts therefore, users ought to become aware of the importance of information for effective and sustainable groundwater management.

Besides, EIS adoption is much more likely for low-cost sensors. Our experts provide valuable insights when they comment that ‘technologies that allow performing the same task by the users with lower cost and greater accuracy will be adopted’. This is a reality which has already been observed in other water management technologies, and which has a greater weight in the regions of the global south. Of greater interest is the fact that the Delphi highlights, after the cost factor, the value of the open-source data and then the smart water management. Some participants attribute this to the fact that emergent technologies are ‘used in the collection/ generation/ sharing of data where there is larger scope for co-data generation and management’. This comment hints at the potential advantages from shared management and to its essential role in processes of trust-building, determinant to stimulate user participation and to facilitate the acceptance of control and self-control policies (Barnett & Rinaudo, 2023; Sanchis-Ibor et al., 2023).

Consequently, where there is a high level of consensus – presenting a strong foundation for government support – is around investment in users' capacities. This no-lose strategy also emerged from our results. Thus, in terms of policy design for groundwater management, investing in EIS for users and for the administration itself seems a safe bet, even though it is unlikely to be sufficient by itself for better groundwater governance. It requires reinforcement with other key elements, such as strong or clear regulations and norms on groundwater use.

There is a broad consensus in the literature that the institutional processes of adopting new technology shape complex socio-technical systems (Callon, 1991; Latour, 1991; Law, 1999), with mutual interactions between hardware, software and orgware (Poblador et al., 2021). These interactions take place in different institutional settings, which, in terms or groundwater management, sometimes are dominated by State control and sometimes by co-management frameworks. In this Delphi analysis, experts concur on a sequential narrative on the importance of strong institutional frameworks in terms of rules and regulations, supported and strengthened by enhanced information, which can then help raise awareness and potentially lead to higher levels of self-management.

Our results indicate that state command-and-control is perceived as the most prominent approach, and that this will remain important for the foreseeable future, especially in the Mediterranean region. Furthermore, command-and-control are also key to the EIS system because it provides a necessary ‘rules’ framework for other approaches, such as self-governance, to emerge and endure. Therefore, experts see the command-and-control approach – or the existence of regulation and/or rules, generally – as an important aspect for the self-organisation of users. State command-and-control is not perceived or presented as an antagonistic force for co-management or stakeholders' empowerment but as a potentially useful basis or foundation to facilitate users' participation. Indeed, this idea was well summarised by one expert: ‘the macrolevel institutional regimes enable community institutions to regulate the use of groundwater’ and has been clearly identified in groundwater management literature (Closas et al., 2017).

Following this, experts have identified some important potential interactions connecting command-and-control and EIS. First, better information can help state agencies to make better decisions in relation not only to command but also to control. This can relate, for example, to the number of illegal and/or informal wells (Loch et al., 2020). Second, investment in staffing is important to enable the oversight of groundwater management and to take better decisions based on good information, such as, for example, on aquifer levels or pollution incidents.

That all data are not equal, as we stated above, is an important finding that implies that various types and qualities of data should be approached or treated according to context and need. Some of these data can be very politically sensitive and often also difficult because it can open the door to making evident illegal use, informal uses, etc (Loch et al., 2020). The state agencies would be bound to follow the law regarding reporting, permit issuance, whatever its jurisdiction encompasses. Molle & Closas (2020) show a pattern of laws not being enforced, or loopholes found, amnesties enacted, monitoring and enforcement that is incomplete or selective, etc, which has also occurred in countries with sophisticated capabilities –like in the case of environmental water allocation in the Australian Murray-Darling Basin (Beasley, 2021; Bruns, pers. comm), or in the case of illegal groundwater use around the Doñana National Park or the Mar Menor in Spain (Bea Martínez et al., 2021). However, if this more sensitive type of data and information is also collected and shared by the users, this could help to build trust, share responsibility, and potentially ensure better compliance, as users demand in some cases (Sanchis-Ibor et al., 2023).

In relation to organisational institutional tools, there is very strong agreement from our experts on the high impact that EIS would have for co-management and bottom-up inclusive governance. Experts emphasise not just information but the effectiveness of these systems. Regarding how governments can support collective action and coordination across wider waterscapes, our results suggest that EIS are a good starting point but insufficient in and by themselves. Here, we mainly analyse what insights we gain in relation to the role of EIS for collective action and, specifically, for government support to users to increase their capacity for co-management and self-management in the case of groundwater. Our results indicate a very high level of agreement among experts, with almost 2/3 of survey respondents agreeing that EIS is very important for co-management. Experts agree that better information will reduce rivalry between users by bringing transparency, potential for discussion and, ultimately, better-informed decisions. Importantly, however, other experts argue that information alone is not sufficient to reduce rivalry. EIS must be reinforced with other elements like well-designed incentives and rules. That is, there are potential difficulties and possible downsides to greater transparency and surveillance, including who is monitored and who has access to information, potential resistance and opposition, manipulation, and disinformation (Bruns, person. Comm). For example, in a separate survey question on illegal groundwater use, experts commented how monitoring is a key aspect by agencies at the local level, and how EIS increases transparency provided these are coupled with adequate policy interventions. Thus, in our complex socio-ecological systems, this confirms the abovementioned importance of rules and incentives.

In relation to procedural and social learning institutional tools, it is striking that experts are sceptical that stakeholder empowerment (e.g., through citizen science and other EIS) will happen. A main finding is the key gap between the desirability of EIS use versus the probability of it occurring. Again, information, although needed, is not sufficient in and of itself for good groundwater governance. Generating information together with users and citizens needs other key elements to succeed. These elements may be the permanent engagement of authorities or investment in adequate resources for staffing, capacity building, and other operational needs so that technology can empower users and citizens and thus help to increase social justice, as competition for resources increases and rules are increasingly needed (e.g., procedures and rights of allocation). This support must be constant and permanent, or at least until the users' institutions are consolidated and achieve financial autonomy. A good illustration of this is the breakdown in collective action and EIS observed when government, research projects or NGO funding support ceases, which has been demonstrated on several occasions (Baalbaki et al., 2019; Rinaudo et al., 2023).

Based on the collective expert knowledge and experience of experts who participated in our Delphi survey, and their understanding of groundwater governance, we draw the following conclusions and takeaways.

First, the processes of adoption of EIS are slow and carry the risk that they may never materialise. EIS must be understood and conceptualised as systems, and without an appropriate social and institutional context, they cannot progress by themselves. Yet with the right institutional and organisational accompaniment, EISs offer many opportunities for enhancing groundwater governance and management.

Second, state command and control will remain a central paradigm for groundwater governance. Here, EIS can indeed increase the effectiveness of and trust in regulation and law, but these command-and-control systems are needed by themselves as a key element in the whole system.

Third, EISs can help stakeholder empowerment and bottom-up, inclusive governance, but experts think it is unlikely this will happen due to the strong competition and rivalry over resources. Here it is particularly interesting that the interface between citizens/users and science can help tip the balance so that – if there is more trust in science (and it is used via EIS) – stakeholder empowerment could materialise.

Fourth, EISs can help in collective action, although some areas, such as monitoring, or rule enforcement will likely remain very challenging. This is where investment in high quality, accurate, shared information is key, and part and parcel of capacity building that will benefit both users and the administration.

Finally, EISs can benefit from the interface between socioecological and socio-technical systems through emergent technologies. Data can be categorised in a hierarchy. Data and information that are easier to generate constitute information that is better suited to be generated in a formalised manner. Data that are harder to generate is often the product of emergent technologies that are context and culturally sensitive. These technologies, however, offer much more scope and potential to shift the system through to data cogeneration and co-monitoring, i.e., a full enhanced information system that covers all data types.

In short, our foresight study indicates that an enhanced information system needs to be framed by a strong command-and-control system. These systems need to have clear rules, with a vision for stakeholder empowerment and inclusive governance that relies on the citizen–science interface as a key tipping point for social justice. As groundwater becomes scarcer, public bodies must engage in capacity building for both their own internal strengthening and that of their users, capitalising on clear data and a data hierarchy with some tasks for specific agencies, and some co-created and managed with users and citizens via emergent technologies.

This study has been funded by the eGROUNDWATER project (GA no. 1921), part of the PRIMA programme supported by the European Union's Horizon 2020 research and innovation programme.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.