Introduction

Rivers are the lifelines of the planet, supporting biodiversity and providing essential contributions to people and their well-being1,2,3. Yet, river ecosystems are suffering from rapid biodiversity loss caused by fragmentation, habitat degradation, invasive alien species, and pollution, all exacerbated by the current context of climate change4,5,6. For centuries, infrastructure has been built to harness goods and services provided by rivers, at a high ecological cost. Structures, such as dams, weirs, dykes, and channelized river sections, have not only disrupted the natural flow regimes along river courses, but have also altered sediment transport, and negatively impact biodiversity and ecosystem functioning along all connectivity dimensions7,8,9,10,11,12,13. Connectivity, defined as the movement and transformation of materials, energy, and organisms, is vital to river health and operates along four key dimensions: longitudinal (from headwaters to river mouth), lateral (between the channel and floodplains), vertical (from surface water to groundwater), and temporal (seasonal or annual)14,15,16. Thus, disruptions to this connectivity impede, for instance, access to critical habitats for resident and migratory species17,18, with cascading effects that ripple far beyond the rivers themselves19,20. Today, over 63% of the world’s rivers longer than 1,000 km do not flow freely15, and global migratory freshwater fish populations have declined by 81% over the last 50 years21. In Europe, over one million physical barriers continue to impede river connectivity22, with 90% of the original floodplains degraded23, and only 44% of European rivers meeting the supposedly binding targets of the EU Water Framework Directive (WFD; 2000/60/EC)24.

Global initiatives such as the UN Decade on Ecosystem Restoration (2021–2030), the Ramsar Convention on Wetlands, and the Convention on Biological Diversity seek to guide and coordinate freshwater ecosystem recovery5,25. At the European level, the Nature Restoration Regulation26 was enforced with the aim of restoring at least 25,000 km of rivers to a free-flowing state by 2030, among other goals26,27. Although the NRR builds on established policies like the WFD and the Habitats Directive (1992/43/EEC), its success depends on the ability of responsible parties to translate its ambitious objectives into practical, effective action on an unprecedented scale. Penca and Tănăsescu28 warn that restoration initiatives will be unsuccessful unless the root causes of biodiversity loss are addressed. Other authors also consider it important to look beyond ecological aspects: Hering et al.29 emphasize that, while the NRR takes a traditional conservationist approach, meeting its objectives requires adequate funding, institutional support, and stakeholder engagement. In fact, responsible parties must go beyond resource allocation to address fundamental conflicts regarding values and priorities30. Therefore, to successfully implement the NRR, restoration efforts must integrate ecological, social, and political dimensions through adaptive, cross-sectoral planning, aligning with recent calls for “fundamental, system-wide shifts in views, structures, and practices”31,32 and supported by growing empirical evidence that socio-ecohydrological and multi-actor restoration initiatives yield more resilient and lasting outcomes27,33,34,35,36,37.

In light of these persistent policy obstacles to implementation and complex socio-ecohydrological interactions, recent advances in structured methodologies have emerged as promising tools to bridge the gap between (and within) research and the practical implementation of restoration initiatives. Techniques such as horizon scanning38,39,40, collaborative research prioritization41,42, and structured expert elicitation43,44 have been demonstrated to be effective in identifying key research needs. For instance, Harper et al.45 distilled 25 essential research questions aimed at reversing freshwater biodiversity decline. This underscores the imperative for evidence-based management and strong policy frameworks. Similarly, Van Rees et al.46 emphasized that freshwater biodiversity is disproportionately threatened and under-prioritized relative to marine and terrestrial systems, advocating for policies that address environmental flows, water quality, and integrated water management. While their work emphasizes that research and conservation actions should mutually reinforce one another through sound evidence and multidisciplinary collaboration, our study extends this approach by explicitly aligning research priorities with tangible conservation and restoration objectives for free-flowing rivers. These are defined as four-dimensional fluvial systems, extending from source to sea, in which ecosystem functions and services are affected by minimal human-induced disruptions in connectivity15,47.

Yet, despite growing recognition of the need for integration, genuine collaboration between natural and social sciences remains rare. Persistent barriers such as disciplinary paradigms, insufficient institutional support, and limited collaborative skills continue to constrain integration48. Empirical analyses show that multidisciplinary research still represents only a small fraction of publications, even in fields where coupled human-natural processes are central49. Engaging social sciences is key to producing outcomes that are more legitimate, salient, and effective50. Building on these insights, our study advances an implementation-oriented, cross-disciplinary research agenda that explicitly links scientific priorities to (i) the typology and context of river barriers, (ii) mechanisms for collaboration across natural and social domains, and (iii) actionable policy frameworks such as the NRR and BDS 2030. In doing so, we move beyond conceptual appeals for interdisciplinarity toward a context-sensitive roadmap for restoring free-flowing rivers and translating research into practice, building on recent European initiatives that bridge scientific, policy, and stakeholder domains in river restoration27,33,34,35. This roadmap explicitly acknowledges that restoration decisions are shaped by stakeholder perspectives, governance constraints, and political trade-offs across spatial scales.

Building on this need for more effective integration, we identify research priorities that emerge from the practical requirements of implementing river restoration in Europe as mandated by the legally binding NRR, which is a direct outcome of the EU Biodiversity Strategy 203047. Our aim is to identify (inter)disciplinary research priorities for successfully and efficiently restoring free-flowing rivers in Europe, thereby bridging the gap between scientific knowledge and practical implementation and accelerating evidence-based restoration efforts. To achieve this, we employed a four-step approach: (1) an online scoping survey to identify key research topics and priorities, (2) an expert workshop to refine these topics, (3) systematic ranking of research areas, and (4) advanced analyses of the ranking results through correspondence analysis to visualize topic clustering and assess how participant characteristics influence topic selection. This integrative methodology, incorporating horizon-scanning, collaborative prioritization, and best practices in expert-elicitation, ensures that the research agenda is scientifically robust and aligned with the needs of policymakers and practitioners working to restore Europe’s river ecosystems. We hypothesize that achieving transformative change in the restoration of free-flowing rivers, as well as in the restoration of other ecosystem types, requires an interdisciplinary approach that integrates insights from the natural and social sciences. Our approach therefore serves as a best-practice example by linking research priorities to diverse restoration contexts and barrier-related challenges, aligning them with EU policy frameworks. This provides a clear, operational roadmap for translating research priorities into context-dependent restoration action.

Methods

To establish a research agenda that bridges the gap between scientific insights and practical policy implementation, we conducted a multi-step expert consultation and ranking process using a modified Delphi protocol40,44. The Delphi method is a structured, iterative process used to obtain consensus among experts through multiple rounds of questionnaires. We applied a modified Delphi approach that combined open-ended scoping, in-person refinement, and structured ranking in a single cycle to balance inclusivity and feasibility. The following sections outline our approach (Fig. 1).

Fig. 1: Outline of the multi-step approach used to develop the research agenda for restoring free-flowing rivers.
Fig. 1: Outline of the multi-step approach used to develop the research agenda for restoring free-flowing rivers.
Full size image

The process began with an online scoping survey followed by an expert workshop to refine the identified topics and categorize them into natural and social science disciplines. A subsequent ranking survey prioritized the topics using weighted measures.

Collecting research topics

An online scoping survey was developed to identify key themes or questions (i.e. research topics) to be addressed to support the NRR target of restoring at least 25,000 km of European rivers to be in free-flowing condition by 2030. Free-flowing rivers are defined here as four-dimensional fluvial systems from source to sea with minimal human-induced disruptions15,47. In this study, we use the term “barrier-related challenges” broadly to include structural and operational disruptions to river connectivity (e.g., dams, weirs, sluices, culverts, embankments, and hydropeaking). The survey was distributed within established networks of experts in river ecology, river restoration and environmental management. An overview of the networks through which the scoping survey was initially distributed is provided in Supplementary Material 1. We ensured broad disciplinary and geographical representation by targeting researchers, practitioners, water managers, policymakers, and NGO representatives. Respondents were asked to list up to three research topics or questions, specify the spatial scale at which each topic should be addressed, and provide information regarding their scientific background, professional position, and country of residence (see Suppl. Mat. 2 for the full survey instrument).

A pilot test of the scoping survey was conducted at the Norwegian Institute for Nature Research via internal email lists targeted at both social and natural scientists working on aquatic biodiversity. This pilot yielded 30 responses and was deemed successful, requiring no substantial refinements. The survey was subsequently launched on SurveyMonkey (SurveyMonkey Inc., San Mateo, CA) from 4 to 23 March 2024, using targeted emails, newsletters, list-servers, and social media in a snowball sampling approach initiated by the group of core authors and their freshwater networks (Suppl. Mat. 1). In addition, the authors distributed the initial scoping survey within their own organizations to further increase disciplinary and geographical reach. Although snowball sampling may introduce biases, this method is widely recognized as an effective way to reach the diverse, transdisciplinary community of scientists and practitioners in this field51.

A total of 714 responses from 237 respondents were manually reviewed for typographical errors, linguistic clarity, and standardized terminology. Responses containing multiple research topics were split into distinct entries, whereas similar responses were combined into single entries. This consolidation process resulted in 425 non-redundant submissions. Using a hierarchical analysis, these were then categorized into eight broad categories, which were further subdivided into 27 preliminary research topics.

Expert workshop and refining research topics

On 18 April 2024, an expert workshop was convened at the Free Flow 2024 Conference (https://freeflowconference.eu/) in Groningen, the Netherlands. The workshop brought together 18 participants from 13 countries, including 15 European and three non-European participants. Participants represented diverse scientific disciplines (14 from the natural sciences and four from the social sciences), and professional roles, including researchers (10), NGO representatives (5), water management professionals (2), and one ecological consultant. The workshop featured an interactive brainstorming session to refine the preliminary research topics. Participants were also guided by three questions: (1) What are the most pressing research needs for restoring free-flowing rivers? (2) Which of these research needs could be addressed through interdisciplinary collaboration? and (3) What are the potential political and practical barriers to implementing our research topics? Their responses were recorded and incorporated into the discussion section of this paper to help prioritize research needs. Throughout the workshop, participants also contributed to the enhancement of our initial list by formulating concise descriptions for the key research topics and categorizing these topics into scientific disciplines (either “natural science” or “social science”) (see Suppl. Mat. 3). Topics classified within natural science predominantly examined ecological processes, while those within social science focused on governance, stakeholder engagement, and economic trade-offs. We acknowledge that the distinctions between natural and social sciences can often be ambiguous, as numerous subjects inherently encompass multiple dimensions, including economic factors.

In addition, our author group assigned each topic to one or more spatial scales (local, regional, national, and global) based on its scope, potential impact, and capacity to influence policies or ecosystems (see Suppl. Mat. 4 for more details). The local scale captures direct, community-level interactions and impacts; the regional scale covers implications across multiple municipalities or within a specific watershed; the national scale reflects country-wide policies and initiatives; and the global scale addresses issues relevant to international governance and cross-border concerns.

Ranking of research priorities

After a preliminary test run at the Leibniz Institute of Freshwater Ecology and Inland Fisheries (IGB) (see Suppl. Mat. 2 for survey details), the ranking survey was distributed via SurveyMonkey to the 175 respondents who had expressed interest in further involvement in the study and kept open for participation from 22 October to 15 November 2024. Respondents were asked to rank their top 10 research topics from the list of 27 (randomized for each respondent), each accompanied by a detailed description (Suppl. Mat. 3). Only the top 10 rankings from each respondent were retained for analysis.

Analysis and synthesis of ranking data

The ranking data were analyzed, and, for each topic, we calculated the following measures: average ranking score, the frequency of appearance in the top 10, and the number of times the topic was ranked first, second, and third. Once calculated, each measure was ordered by priority, with 1 representing the highest priority and 27 the lowest. For the average ranking score, lower values indicated higher priority and were assigned a better rank (i.e., rank 1). In contrast, higher values received better ranks for frequency-based measures (top 10 appearance, and #1, #2, #3 ranking).

To generate the final overall ranking, we applied weighted percentages to reflect the relative importance of each measure: average score (30%), top 10 frequency (25%), #1 frequency (20%), #2 frequency (15%), and #3 frequency (10%). This weighting approach assumes that the average score best represents the general perceived importance of a topic, thereby receiving the highest weight. In contrast, the top 10 frequency captures the consensus across respondents and was assigned a moderately high weight. The top three positions were weighted progressively lower (20%, 15%, and 10%, respectively) to reflect that #1 ranking is more influential than lower rankings. This approach is consistent with methods used in similar research prioritization exercises41,42. The weighted values were then summed and normalized to produce a final priority score for each topic, with lower scores indicating higher overall priority. Finally, topics were sorted from 1 (lowest score; highest priority) to 27 (highest score; lowest priority). Ranking data, including calculations, are provided in Supplementary Material 5.

To assess the robustness of our weighting choices, we recalculated the composite ranking using two alternative schemes that modified the weight assigned to the average-score criterion while redistributing the remaining weight proportionally across the other four criteria. In Scenario B, the average-score weight was reduced to 20%, with the remaining weight reallocated proportionally (28.57%, 22.86%, 17.14%, 11.43%). In Scenario C, the average-score weight was increased to 40%, again with proportional redistribution (21.43%, 17.14%, 12.86%, 8.57%). These proportional percentages reflect the same relative differences between the top-10 frequency and the #1, #2, and #3 frequencies as in the original weighting scheme, scaled to the reduced or expanded remaining weight. We compared the resulting rankings using Spearman’s ρ and Kendall’s τ to evaluate consistency (see Suppl. Mat. 6).

To further explore the relationships among the research topics and assess the influence of participant characteristics, we conducted a Correspondence Analysis (CA). A binary matrix was constructed from the topic selections, where each respondent’s inclusion of a topic in their top 10 priorities was coded as 1 and non-selection as 0. The analysis was performed using the ca package52 in R53, which simultaneously generated scores for both respondents and topics. Eigenvalues were calculated to determine the percentage of total inertia explained by the first two dimensions, revealing the primary patterns in research priorities. For example, differences along these axes may indicate contrasts between topics that are prioritized by different experts, regardless of traditional disciplinary boundaries. In fact, proximity of topics from distant disciplines would suggest a perceived need for interdisciplinary work to drive transformative change or resolve trade-offs. Respondent scores were merged with participant characteristics (discipline, profession, and country), and topic scores were linked with this background information. Finally, convex hulls were overlaid on the biplots to visualize group patterns.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Results

Participant information

The scoping survey generated 425 unique responses from 237 respondents from 45 countries (Suppl. Mat. 7), which were consolidated into 27 research topics across the natural and social sciences (Fig. 2; Suppl. Mat. 3). Most respondents were based in Europe, with 185 individuals (78.5%) coming from EU Member States. Among these, Germany had the highest representation with 32 participants (13.5%), followed by Austria (21; 8.9%), Sweden (21; 8.9%), the Netherlands (20; 8.4%), and Spain (17; 7.2%). In terms of professional roles, 152 respondents (64.1%) identified as researchers, 13.1% were NGO staff, 7.2% were water management professionals, and the remainder included ecological consultants, policy specialists, engineers, and science communicators. With regard to their scientific discipline, 209 respondents (88.2%) had a natural-science background, 17 (7.2%) a social-science background, and 11 did not specify. Finally, 104 respondents (43.9%) reported active involvement in river restoration efforts.

Fig. 2: Overview of the 27 research priority topics identified for restoring free-flowing rivers.
Fig. 2: Overview of the 27 research priority topics identified for restoring free-flowing rivers.
Full size image

The topics were derived from 425 unique submissions provided by 237 respondents in the initial step of our scoping study. Circle size indicates the relative frequency with which elements contributing to each topic were mentioned, with larger circles reflecting above-average representation. Topics are organized along a conceptual gradient from social science-oriented themes (lavender) to natural science-oriented themes (turquoise), with integrative or cross-cutting topics shown in grey. Topics that ranked into the top 10 (see Fig. 3) are highlighted with yellow borders. More details on the topics can be found in Supplementary Materials 3 and 8.

Our subsequent ranking survey involved 75 participants from 25 countries and exhibited a similar distribution: 77.3% were from EU Member States, with Germany again leading in representation, followed by Austria, the Netherlands, and Sweden. Professional roles were dominated by researchers (81.3%), and 94.7% of respondents came from natural sciences, with only 5.3% representing the social sciences (Suppl. Mat. 7).

Topics of relevance

Among the 425 unique submissions, social science topics were most frequently mentioned, with ‘developing prioritization strategies for targeted restoration’ (82 mentions) and ‘enhancing public awareness and communication strategies’ (80 mentions). The leading topic in natural sciences, ‘enhancing riverine biodiversity and ecosystem functioning’, received 79 mentions. Following these top three, the next four most frequently mentioned topics were ‘identifying best practices and addressing hydropower impacts’ and ‘engaging communities and stakeholders (56 mentions each), followed by ‘improving species migration through connectivity restoration’ (41 mentions) and ‘developing effective monitoring frameworks for restored rivers’ (39 mentions). Natural science topics were predominantly associated with regional or local scales (15 out of 18 topics), while social science topics were distributed across all spatial scales, with no clear predominance at any particular level (Suppl. Mat. 3).

Ranked list of research priorities

The ranking analysis of all 27 research topics highlighted the top three as clear priorities: the highest-ranked topic ‘enhancing riverine biodiversity and ecosystem functioning’ was included in the priority top 10 by 44 of the 75 respondents, the second-ranked ‘developing prioritization strategies for targeted restoration’ by 46 respondents, and the third-ranked ‘establishing restoration standards for free-flowing rivers’ by 33 respondents (Fig. 3; Suppl. Mat. 3). Overall, the top 14 topics represent an even balance between natural and social science topics, despite the scoping survey yielding more natural (18) than social science topics (9).

Fig. 3: Final ranking of the 27 research priorities for restoring free-flowing rivers, arranged from highest rank (top) to lowest (bottom).
Fig. 3: Final ranking of the 27 research priorities for restoring free-flowing rivers, arranged from highest rank (top) to lowest (bottom).
Full size image

The bar charts on the right show how many times each topic was originally mentioned in the scoping survey (see also Fig. 1). Grey bars denote social science topics, while black bars represent natural science topics. The dots on the left indicate the spatial scales (local, regional, national, and global) assigned to each topic based on its scope, potential impact, and capacity to influence policies or ecosystems. Light grey dots indicate the probable spatial scale of action for each topic, while dark grey dots denote the scale most frequently assigned by the author group (see Suppl. Mat. 4 for further details).

Spatial scale also played an important role in the prioritization of topics. Most of the top 10 topics were considered relevant at larger scales (national to global), while many of the lower-ranked topics were viewed as more pertinent at local and regional scales (Fig. 3). Importantly, the frequency with which a topic was mentioned in the scoping survey did not necessarily correspond with its final ranking. For example, topics such as ‘securing long-term financial sustainability’, ‘improving governance and stakeholder collaboration’, and ‘developing cross-boundary governance models’ received relatively few mentions initially, but were later ranked highly (ranks 7, 9, and 14, respectively). Similarly, ‘establishing restoration standards for free-flowing rivers’ was mentioned only 17 times in the first round but ranked third highest in the ranking survey (Fig. 3).

Our sensitivity analysis showed that rankings proved highly stable to alternative weightings (scenario B: ρ = 0.996, τ = 0.972; scenario C: ρ = 0.985, τ = 0.937). The composition of the top-10 priorities was unchanged in both scenarios, with only minor within-set reordering (maximum shift of two ranks under scenario B and one rank under scenario C: Suppl. Mat. 6). This indicates that our main conclusions are not driven by the specific weighting of the five criteria.

Clustering of topics

The correspondence analysis (CA) provided deeper insight into the clustering of topics, revealing which topics were often selected together in the top-10 priority lists. CA space clearly distinguishes between natural and social science topics, with minimal overlap. Topics within the same disciplinary category tended to be co-selected in respondents’ top 10 lists, except for ‘prioritization strategies’ and ‘innovative technical solutions’, which spanned disciplinary boundaries (Fig. 4A). Analysis based on respondent background showed that while respondents with a natural science background formed a large cluster (n = 71), those with a social science background (n = 4) were concentrated within the social science portion of the CA space (Fig. 4B). This suggests that respondents with social science expertise preferentially selected topics related to ‘stakeholder cooperation’, ‘community and stakeholder involvement’, and the ‘financing and viability’ of river restoration in their top 10 lists, though the low number of responses from social science experts limits broader interpretation of these findings.

Fig. 4: Results of the correspondence analysis of 75 research priority lists, indicating how topics cluster in CA space and how different respondent groups are distributed across these dimensions.
Fig. 4: Results of the correspondence analysis of 75 research priority lists, indicating how topics cluster in CA space and how different respondent groups are distributed across these dimensions.
Full size image

The first two CA axes explain a total of 22.1% of the inertia (Axis 1: 12.4%, Axis 2: 9.7%). A shows the 27 research topics (labeled by short descriptions found in Suppl. Mat. 8), with social science topics in grey triangles and natural science topics in black circles. B–D plot respondent scores (indicated by topic numbers) over the same CA space (social science topics: triangles, natural science: circles), grouped by discipline, profession, and country, respectively, with convex hulls outlining each group. For the country plot (panel D), only countries with at least five respondents are highlighted with distinct colors and convex hulls. Countries with fewer than five respondents are combined into an “other countries” group, with their CA positions displayed as grey dots.

Professional affiliation also influenced topic selection. Researchers (n = 61) generally endorsed a broad range of topics. In contrast, NGO-affiliated respondents (n = 7) tended to prioritize issues related to ‘prioritization strategies’, ‘financing and viability of restoration’, ‘climate change adaptation’, ‘cross-boundary governance’, and ‘science-policy and legal frameworks.’ In contrast, water managers (n = 3) focused more on natural science topics, including ‘invasive alien species management’, ‘riparian ecology’, and ‘habitat availability’ (Fig. 4C). No clear relationship was observed between participants’ country of residence and the topics they included in their top 10 list. However, German respondents more frequently prioritized specialized natural science topics related to ‘water quality’, ‘pollutant and sediment transport’, and ‘riparian ecology’ (Fig. 4D).

Top 10 research priorities to restore free-flowing rivers

The ranking study revealed key insights regarding research priorities for restoring free-flowing rivers. In Table 1, we present the top 10 highest-ranked research priorities in more detail, while the full list of 27 topics and their descriptions is provided in Supplementary Material 8.

Table 1 Final ranking of the top 10 research priorities for restoring European free-flowing rivers
Full size table

Discussion

Our study helps address global challenges faced by river ecosystems through bridging the gap between scientific research and the practical implementation of river restoration policies54. Building on previous interdisciplinary agenda-setting efforts45,46, it provides the first policy-aligned, implementation-oriented research roadmap specifically tailored to restoring free-flowing rivers. Unlike earlier syntheses, our approach links research priorities to diverse restoration contexts and barrier-related challenges and aligns them directly with the EU NRR and BDS 2030. By integrating diverse expert opinions, our analysis not only maps current research fields in river restoration but also highlights where future scientific work can better support stakeholders in politics, governance, administration, and practice. This integration of ecological knowledge with governance, stakeholder participation, and multi-scale coordination closely reflects Elinor Ostrom’s foundational work on the governance of common-pool resources, which emphasises polycentric institutions, participatory decision-making, and adaptive learning as prerequisites for sustainable environmental management55,56.

Notably, we identify three top research priorities: (1) enhancing riverine biodiversity and ecosystem functioning, (2) developing prioritization strategies for targeted restoration, and (3) establishing restoration standards. Our survey reconfirms the critical role of these often-mentioned research priorities in river restoration, but it also reveals additional, emerging topics that promise to reshape current paradigms, particularly for environmental scientists. Enhancing riverine biodiversity and ecosystem functioning, which strengthens the natural resilience of river ecosystems, represents the overarching target of free-flowing river restoration under the NRR, and also aligns with global restoration aims in the UN Decade on Ecosystem Restoration (2021–2030), the Ramsar Convention on Wetlands, and the Convention on Biological Diversity. Meanwhile, prioritization strategies and restoration standards serve as essential adaptive management tools guiding effective interventions. Importantly, our work underscores that translating these priorities into actionable policies may benefit from improved understanding of robust stakeholder involvement and communication. In the following sections, we detail how these priorities integrate interdisciplinary approaches, address technical and social challenges, and inform policy.

The presence and future of interdisciplinarity and stakeholder integration

Effective restoration of free-flowing rivers has repeatedly been argued to require an integrated approach that combines ecological, hydrological, sociocultural, and economic dimensions57,58. This integration enables restored rivers to deliver a broad array of ecosystem services, ranging from provisioning (e.g., water and food supply) and regulating (e.g., self-purification and flow regulation) to cultural services (e.g., recreation and heritage preservation)3,59, which together have the potential to strengthen community well-being and to support local economies. The portfolio of top-ranked research topics allows us to recognize the importance of this link: the top-ranked priority of enhancing riverine biodiversity is indirectly but closely associated with many ecosystem services, while additional priorities addressing flood, drought, and climate change mitigation (ranks 8 and 10) more directly illustrate the interdependence between ecological integrity and socio-economic resilience (Fig. 3; Table 1). By incorporating these broader societal benefits into restoration strategies, policymakers can design interventions that restore ecological integrity and enhance the adaptive capacity of both aquatic ecosystems and local communities. However, translating this conceptual integration into practice requires dedicated mechanisms that connect disciplines with management and policy processes48,55,56.

To address this challenge, several workable models have emerged that illustrate how interdisciplinary collaboration can be effectively operationalized. First, structural frameworks such as the “Radically Inter- and Trans-disciplinary Environments (RITE)” model proposed by Holm et al.60 emphasize the need for long-term institutional support and funding strategies that enable scientists from social, human, natural, and technical disciplines to collaborate from the outset. Such frameworks also underscore that interdisciplinary collaboration requires stable and sustained funding arrangements, as short-term or sequential project cycles rarely provide the continuity needed for shared concepts, methods, and goals to develop. Rather than treating interdisciplinarity as ad hoc or actor-led, such models call for creating dedicated environments where shared knowledge and methods can evolve. Second, boundary-spanning consortia (e.g., the European Centre for River Restoration) help sustain collaboration beyond project cycles by maintaining shared databases, learning platforms, and regular exchanges49. Third, embedding researchers directly in policy programs through evidence pipelines and co-production initiatives can enhance the uptake of scientific findings in decision-making50.

In line with these frameworks, our study confirms that technical and conceptual advances alone are insufficient unless they are supported by social and economic considerations. Research focused on developing prioritization strategies (ranked 2nd) and defining restoration standards (ranked 3rd) remains essential for guiding restoration efforts. However, these measures must be complemented by addressing societal well-being and economic viability to ensure that restoration is cost effective, socially desirable, and fundamentally successful in achieving long-term ecological outcomes22,46,61,62. Frameworks for barrier removal that integrate societal aspects63,64 demonstrate how such approaches can reconcile conflicting interests and coordinate actions across different spatial scales, ensuring that broad-scale policies translate into context-specific, locally driven initiatives. In practice, this translation often depends on bargaining and compromise among actors with different values, risk perceptions, and objectives. These negotiations are not purely technical or administrative, but inherently political, as they involve competing interests, power relations, and decisions about whose values and priorities guide river management. Compromise can involve quid pro quo agreements between those who support and those who resist removal, for example through staged implementation, partial removal, additional safety or heritage measures, alternative infrastructure solutions, or compensation. It may also be required within the scientific community when evidence supports multiple plausible outcomes and trade offs. Making these political and scientific trade offs explicit, by documenting acceptable compromises, non negotiable ecological thresholds, and clear rules for revisiting decisions as new evidence emerges, can increase legitimacy, reduce conflict, and help move barrier removal decisions from debate to implementation.

Restoration strategies and stakeholder constellations vary substantially across barrier types and regional settings. Recent data from the Dam Removal Europe coalition show that 542 barriers were removed across 23 European countries in 2024, the majority being small weirs and culverts rather than large dams65. This highlights how barrier typology and scale influence restoration feasibility, governance complexity, and stakeholder dynamics. The AMBER project advanced adaptive management approaches by mapping more than one million artificial barriers across Europe and developing decision-support tools to prioritize connectivity restoration based on ecological, social, and economic criteria22. The MERLIN project66 complements this perspective by testing large-scale, nature-based solutions through 18 flagship case studies, emphasizing context-tailored restoration portfolios and blended public-private funding mechanisms. Together, these initiatives demonstrate that restoring free-flowing rivers requires flexible, case-specific approaches that recognize differences in barrier typologies, governance systems, and social license to operate, rather than a one-size-fits-all model.

Our findings indicate that these collaborative mechanisms are most successful when clear coordination roles and communication protocols are established early on49, and when knowledge exchange is embedded in adaptive management loops that link research directly to policy cycles50. Importantly, several of the top-ranked social science priorities can be operationalized as structured research programs rather than remaining conceptual or awareness-based. For example, research on long-term financial sustainability (rank 7) can involve evaluating blended finance models, payment-for-ecosystem-services schemes, and outcome-based conservation contracts, as demonstrated in MERLIN’s nature-based investment pilots. Likewise, improving science-policy interaction and legal support (rank 11) can be examined through evidence pipeline analysis and policy feedback evaluations50, while engaging communities and stakeholders (rank 4) can be operationalized through co-production workflows and participatory governance trials, as applied in the AMBER project and in the various activities organised by Dam Removal Europe. These cases illustrate that social and governance-related priorities can be approached through structured, empirical research with clear evaluation criteria, showing that they are substantive research areas in their own right rather than add-on considerations.

Taken together, these insights highlight the need to view interdisciplinarity not as an end in itself but as a dynamic, multi-scalar process of transformation. These scales have also been theorized as distinct levels of change: the practical, structural and paradigmatic67,68. The practical sphere refers to tangible, implementation-oriented actions such as collaboration, co-production, and adaptive management. The structural sphere involves shifts in institutional arrangements, governance, and power relations, while the paradigmatic sphere encompasses deeper changes in values, worldviews, and knowledge systems. The sharp rise in the ranking of “establishing restoration standards” (priority 3, increase of 11 positions) further illustrates that many respondents may not initially perceive epistemic clarity as a core priority, until confronted with the need for shared definitions, criteria, and conceptual alignment across disciplines. Our results suggest that current restoration research and policy agendas predominantly operate within the practical sphere, where coordination needs and concrete management challenges are most visible, while systemic and paradigmatic transformations remain less explicit. This helps explain why respondents emphasized operational solutions and coordination tools over deeper forms of knowledge integration.

Building on this foundation, our results highlight the central role of stakeholder involvement in bridging research, policy-making, and practical implementation. Effective restoration requires sustained collaboration among researchers of multiple disciplines, policy makers, water managers, NGOs, economic sectors, and the public36,37,69,70. While stakeholder involvement is formally mandated under EU funding schemes and the WFD, implementation remains uneven at local scales. Strengthening participatory communication with educational efforts to promote broader “river ecosystem literacy”71 can help reconcile conflicting views and improve support for free-flowing rivers. Recent recommendations from the EU Commission further emphasize the need for broader public engagement and multi-level communication to better integrate local knowledge into restoration planning72.

Fundamental challenges and multi-scale coordination

In addition to integrating socio-economic dimensions and stakeholder engagement, our work identifies several fundamental challenges within the natural sciences that must be addressed to ensure effective river restoration under the NRR. A primary challenge is to develop clear, parameterized definitions for free-flowing rivers as well as for the reference conditions used to evaluate restoration. Recently, we have seen increasing recognition of the importance of the dendritic, hierarchical spatial structure of river networks, a view reflected in priorities related to both technical (e.g., prioritization strategies (rank 2), connectivity and species migration (rank 6)) and adaptive responses (e.g., drought and flood resilience (rank 8), climate change adaptation (rank 10)). Moreover, experts call for incorporating meta-ecosystem and meta-community perspectives10,73 to better understand species movement and the transport of nutrients and organic matter across river networks. Restoring entire free-flowing river networks, that is, achieving full connectivity across all fluvial dimensions through strategies such as barrier removal, floodplain reconnection, environmental flow restoration, and water quality improvement10,74,75, is ecologically ideal. However, extensive modifications in many European rivers render complete restoration unrealistic76. Compromise-seeking across spatial scales in restoration planning may also be motivated by different scale-dependency of social compared to ecological effects, which is an unexplored research territory.

Building on these challenges, our results show that research priorities span multiple spatial scales, from basin-wide prioritization and connectivity assessments to locally grounded restoration actions (Fig. 3). Although high-level policies are essential for setting standards and coordinating responses to large-scale drivers such as biodiversity loss and climate change, restoration ultimately depends on locally feasible measures and stakeholder support. However, practitioners frequently encounter contradictions between regulatory frameworks that simultaneously promote river regulation and call for restoration, underscoring the need for better alignment between NRR objectives and sectoral policies. This reinforces the importance of multi-scale coordination, where local implementation is embedded with coherent national and EU-level strategies77,78,79.

These challenges also call for adaptive strategies that reflect the dynamic nature of expert perspectives77,78. While our scoping survey primarily highlighted ecological and technical challenges, successive ranking revealed an increasing recognition of social research and governance dimensions (Fig. 3). This evolving perspective may explain why some recommendations remain unimplemented, often due to limited integration of socio-political factors80. For example, while researchers focused on technical aspects, NGO representatives and water managers stressed the need for robust governance frameworks, community engagement, and practical measures such as invasive species management and habitat availability (Fig. 4). In practice, aligning these perspectives requires co-designing restoration objectives and performance metrics with stakeholders before technical decisions are made. Such coordination can be operationalized through participatory planning processes with local authorities and communities, the joint development of decision-support tools with practitioners, and restoration portfolios that combine ecological monitoring with socio-economic assessments. Meaningful integration is feasible when collaborative structures are embedded early in project design rather than appended after ecological planning.

Study robustness and limitations

Our study employed a multi-step expert consultation and ranking process anchored in a modified Delphi protocol40,44. This process engaged a diverse international community from 45 countries, with strong representation from EU Member States, which not only enhanced the credibility of our findings but also enabled us to map both current and future research fields, thereby supporting an integrated ‘local-to-global’ perspective. Although our scoping survey predominantly reflected natural science topics (Fig. 2), the final top 10 list achieved a balance between ecological knowledge and social science insights. A primary gap lies in efficient implementation strategies rather than in further understanding ecological processes.

We further reveal a nuanced interplay among natural, social, and socio-economic considerations in shaping research priorities. Stakeholder backgrounds and spatial context influenced the balance between ecological imperatives and governance, community engagement, and economic needs81,82. Notably, even though our respondent pool was overwhelmingly composed of natural scientists (only 5.3% were social science experts), those with primarily ecological and practical expertise recognized the need for broader social, humanities, and governance dimensions when provided with an integrative framework (Table 1; ranks 7 and 9 collectively increased 24 positions). This observation underscores the challenge of connecting natural and social science topics in river restoration and conservation, where many issues inherently carry normative and political dimensions80. However, this recognition does not necessarily imply agreement on how such integration should occur. Ecologists often frame the challenge in terms of improving policy uptake of ecological evidence, whereas social scientists emphasize the design of the governance, incentive, and decision-making processes that enable feedback between knowledge and action. This difference in orientation represents a core barrier to operational interdisciplinarity, highlighting the need to study and co-develop the mechanisms that structure evidence uses, negotiation, and accountability in river restoration.

Moreover, the predominance of natural scientists may reflect the current structure of scientific networks in river restoration, a factor likely reinforced by our snowball sampling approach51. Leveraging embedded researchers and practitioners as transdisciplinary boundary spanners may further facilitate knowledge transfer and the implementation of integrated restoration strategies83. In light of these findings, future research on the topic should actively recruit more social scientists, practitioners, and scholars from emerging fields such as environmental humanities to enhance interdisciplinary robustness and foster the creation of interdisciplinary research networks. Scalable analytical approaches that support such collaboration already exist, such as multi-criteria prioritization frameworks applied in connectivity restoration22, participatory governance approaches that facilitate co-design and negotiation among stakeholders69,70, and embedded science–policy evidence pipelines that strengthen feedback between monitoring and decision-making50. Together, these examples show that interdisciplinary integration can be operationalized as structured, empirical research programs, rather than remaining conceptual aims.

Despite these strengths, several limitations should be acknowledged to clarify the scope of inference. First, the participant pool was strongly European, reflecting both the regional focus of the NRR and the concentration of active river restoration networks; as such, perspectives from regions where restoration unfolds under different governance systems, cultural relationships to rivers, and socio-ecological pressures may be underrepresented. Second, although the study applied a modified Delphi approach, its outcomes depend on the granularity and disciplinary composition of expert input, with fewer contributions from social sciences and policy practitioners. Third, the interdisciplinary integration models and policy linkages discussed in this study are literature-derived rather than empirically tested within the prioritization process itself. They therefore function as conceptual pathways rather than validated implementation mechanisms, aligning with earlier calls for long-term embedded collaboration frameworks that can be trialed and evaluated in practice48,49,50,60. Consequently, our roadmap should be interpreted as an evolving framework to guide research and policy alignment, rather than a prescriptive blueprint.

Policy implications and future directions

Our study offers a roadmap supporting the EU and its Member States in implementing the NRR and achieving the ambitious target of restoring at least 25,000 km of free-flowing rivers by 2030, building on a structured prioritization process that links diverse restoration contexts and barrier-related challenges to actionable policy guidance. The prioritized research topics, such as establishing restoration standards and developing targeted prioritization strategies, directly align with Article 9 of the NRR. Our results underscore that integrating ecological, hydrological, and socio-economic perspectives is essential for identifying actionable research priorities that can guide policy and restoration planning. To operationalize our research agenda within current EU policy frameworks, we mapped the ten highest-ranked research priorities against relevant legislative articles and potential indicators (Table 2). This alignment is bidirectional: research priorities can be tailored to meet concrete policy needs, while policy frameworks can, in turn, guide the formulation of targeted research questions and associated indicators. This mapping demonstrates how scientific priorities directly support the implementation of the NRR and BDS2030 and provides a concrete entry point for monitoring and adaptive management under National Restoration Plans (NRPs) and river basin management plans. Crucially, under the NRR, Member States are required to inventory and remove artificial barriers (Art. 9), restore floodplain functions, and incorporate these measures into National Restoration Plans (Art. 15), supported by monitoring and regular plan revision (Art. 19–21). Because the NRR specifically targets river connectivity, its implementation must operate in coherence with the wider set of EU environmental laws and strategies, including the WFD, the Habitats and Birds Directives, and the BDS2030, to ensure that connectivity gains translate into basin-scale ecological integrity and species recovery.

Table 2 Alignment of the top 10 research priorities with European policy frameworks and monitoring indicators
Full size table

We conceptualize a practical roadmap for translating these research priorities into tangible restoration outcomes as six iterative stages: (1) diagnose, to assess river system conditions, barriers, and socio-political contexts using harmonized ecological and governance indicators; (2) co-design, where scientists, policymakers, and stakeholders jointly define restoration targets and success criteria through co-production and boundary work50; (3) finance, which mobilizes resources via public-private partnerships and green investment models aligned with NRR Article 19; (4) permit and engage, involving coordinated governance, legal authorization, and local participation to ensure legitimacy and social license; (5) implement and adapt, where adaptive management and context-specific technical solutions guide restoration on the ground; and (6) monitor and report, integrating ecological and socio-economic indicators to inform iterative learning and policy reporting under the NRR and WFD. This staged, participatory approach closely aligns with Elinor Ostrom’s principles for governing common-pool resources, which emphasise polycentric governance, stakeholder participation, monitoring, and adaptive learning as essential components of successful environmental management55,56. This roadmap represents an operational framework for moving from scientific priorities to coordinated policy action, supported by programmatic models such as RITE that foster long-term, cross-disciplinary collaboration60.

Despite these promising frameworks, conflicting policies pose major challenges. Economic interests driven by the Common Agricultural Policy and renewable energy targets often override long-term conservation goals84,85,86. These sectoral misalignments hinder the protection and restoration of freshwater ecosystems, which require integrated policy frameworks that recognize the unique, transboundary nature of river systems87,88. Such coordination challenges reflect the need for polycentric governance structures, where multiple interacting authorities and stakeholder groups operate across scales to manage shared environmental resources effectively89. Reconciling these conflicting policy objectives is crucial for ensuring that our research priorities translate into coherent restoration strategies that are both effective and scalable.

We further highlight the need for enhanced platforms and research tenders that foster long-term collaboration among natural scientists, social scientists, practitioners, and emerging fields such as environmental humanities. This collaborative approach must emphasize participative, transparent communication and long-term stakeholder involvement, as mandated by frameworks like the WFD, to capture the normative, political, and socio-cultural dimensions of restoration. As emphasized in Table 2, several of the top-ranked priorities, particularly stakeholder involvement, awareness, and governance, correspond closely with participatory obligations under the NRR, WFD, and BDS2030. The high rankings for community and stakeholder involvement (rank 4), public awareness and communication (rank 5), and stakeholder cooperation (rank 9), underscore this need. In addition, recent EU initiatives, as reflected in the 10th Framework Programme (FP10), stress the importance of transparent science service portals to facilitate effective communication among scientists, policymakers, and stakeholders. As political discourse evolves in response to geopolitical shifts, climate change pressures, and changes in governance, our research prioritization approach should be revisited in four to five years to track the impact of inter- and transdisciplinary efforts and adapt our agenda to these dynamic challenges.

Our findings reveal key aspects of a comprehensive restoration process, such as establishing clear, multi-scale targets and prioritizing actions based not only on ecological potential and cost-effectiveness but also on enhancing ecosystem services and social outcomes. Monitoring both ecosystem functions and socio-economic impacts remains insufficiently addressed in current research priorities and is not consistently implemented in practice90,91. Although many European restoration projects are gradually incorporating these features, our survey highlights a persistent knowledge gap regarding how to integrate these elements into a fully scalable restoration framework. Even respondents with predominantly natural science backgrounds recognize the need to further incorporate governance and socio-economic dimensions. Moreover, growing societal recognition of free-flowing rivers as essential ecosystems underscores the importance of increasing public awareness, especially regarding the plight of iconic and migratory species that have become locally extirpated, to help transform isolated pilot projects into mainstream river management practices. Embedding joint ecological and socio-economic monitoring indicators within the policy levers identified in Table 2 would strengthen accountability and provide a basis for adaptive learning at basin and national scales. In this way, the research priorities identified here, provide a shared agenda that can guide coordinated implementation under the NRR, ensuring that river restoration in Europe advances not only barrier removal but also long-term ecological resilience and socially meaningful outcomes.