From pilots to practice: a governance-technology integration framework for circular water in the baltic sea region

The Baltic Sea Region (BSR) is a major arena for circular water innovation, supported by EU investment exceeding €20.5 million in recent years. Yet moving from technically successful pilots to wider regional implementation remains difficult. This study inquires into how governance–technology integration is framed and enacted across policy, practitioner discourse, and project implementation in circular water management, and what coordination processes appear to support scaling. We combine a three-tier, mixed-method analysis of a 372,856-token corpus. Tier 1 covers 12 EU/HELCOM/EUSBSR policy documents (supranational and macro-regional framing). Tier 2 analyses seven Europe Forum Turku transcripts (135,151 tokens; 2024–2025), comprising 526 coded practitioner and policymaker statements, using keyword/collocation analysis and semantic network metrics. Tier 3 compares five Interreg BSR projects representing diverse technology–governance configurations, drawing on project documentation and implementation records. Results show a temporal shift in practitioner discussions away from stand-alone technical optimisation language toward framing that more explicitly links technology to regulation, institutional responsibility, and strategic priorities such as resilience and security. Across policy, forum, and project levels, three coordination mechanisms recur as central to scaling: (i) structural coordination through cross-border policy platforms and regulatory alignment, (ii) knowledge coordination via practitioner forums that translate policy priorities into shared technical and operational repertoires, and (iii) temporal coordination that aligns policy cycles and funding horizons with technology development and deployment timelines. This data-driven study illustrates that scaling in circular water management depends on adaptive coordination across institutional scales, not only on technological performance. Making coordination requirements explicit can improve the effectiveness of public investment and strengthen pathways from pilots to regional practice. In doing so, it proposes a cross-scale coordination lens to locate where scaling stalls between policy intent, practitioner translation, and post-pilot implementation.

1 Introduction

The Baltic Sea Region (BSR) faces persistent water-quality pressures from eutrophication and hazardous substances, increasingly compounded by climate-driven hydrological extremes. These shared, long-standing challenges have also fostered unusually strong cross-border cooperation to address transboundary impacts at regional scale, making the BSR a distinctive governance space where problems are jointly diagnosed and progressively addressed through coordinated policy, science, and implementation. In this context, circular nutrients and water management has emerged as a guiding strategy for sustainable development.

This governance landscape rests on two complementary pillars: HELCOM (the Baltic Marine Environment Protection Commission), which provides science-based coordination for marine protection and thematic strategies, and the EU Strategy for the Baltic Sea Region (EUSBSR), which connects macro-regional policy priorities with EU funding instruments, including the Interreg Baltic Sea Region programme (EUSBSR, 2021; HELCOM, 2016; HELCOM, 2021). Within HELCOM, the Baltic Sea Action Plan (BSAP) has since 2007 served as the central science-based framework for regional marine protection, predating many EU initiatives that later adopted circular economy perspectives. Building on this foundation, the Baltic Sea Regional Nutrient Recycling Strategy (HELCOM, 2021) articulates a shared vision to improve nutrient use efficiency, reduce nutrient losses, and enable safe nutrient recycling across agriculture, aquaculture, forestry, and municipal systems. Its development was enabled by sustained project-based cooperation, notably the Interreg BSR platform project SuMaNu (Sustainable Manure and Nutrient Management), which consolidated knowledge from earlier initiatives on nutrient use efficiency in manure, digestate, and other organic nutrient streams. Through this platform approach, Interreg funding supported institutional capacity, shared analytical frameworks, and policy dialogue among authorities, researchers, and practitioners, providing technical and governance groundwork that informed the drafting of the Nutrient Recycling Strategy and its adoption by the HELCOM Contracting Parties. These developments illustrate how long-standing regional cooperation has progressively intersected with evolving EU priorities through iterative learning and capacity building rather than top-down policy alignment. In parallel, HELCOM has advanced a regional strategic approach for hazardous substances, updating the BSAP toward a more holistic and prioritised process that interfaces with EU and national policy processes to ensure coherence and mutual reinforcement, while remaining grounded in HELCOM’s regional mandate and agreed measures (HELCOM, 2025b; HELCOM, 2025a). Together, these developments reaffirm the BSAP as HELCOM’s enduring scientific and strategic backbone for addressing nutrient and hazardous-substance pressures in the Baltic Sea.

At macro-regional level, the EUSBSR Action Plan (2021) provides an overarching framework organised around “Save the Sea, Connect the Region, Increase Prosperity”. It links regional priorities with EU policies and the 2021–2027 Multiannual Financial Framework (European Union, 2025) and seeks to reduce fragmentation among existing networks without creating new legal structures. For the environmental dimension, the EUSBSR relies strongly on HELCOM’s BSAP for scientific targets and priority measures; in practice, it functions as a vehicle for embedding BSAP objectives into EU macro-regional policy and funding programmes. Water-related challenges are central: Policy Area (PA) Nutri targets nutrient emissions from urban and point sources and advances safe, sustainable nutrient recycling consistent with HELCOM’s strategy and the Green Deal’s circular economy agenda (European Commission, 2019a; European Commission 2019b; EUSBSR, 2021). In parallel, PA Hazards addresses hazardous substances by aligning BSAP commitments with EU regulatory frameworks, including the Water Framework Directive (WFD), Marine Strategy Framework Directive (MSFD), REAC, and the Chemicals Strategy for Sustainability, emphasizing prevention, mitigation of legacy pollution, and systematic feedback into EU and national policy processes (EUSBSR, 2021). This architecture is reinforced by cooperation and funding instruments, most visibly the Interreg Baltic Sea Region 2021–2027 programme (European Union, 2021), which translates strategic intentions into transnational projects and pilots. The EUSBSR Action Plan embeds thematic Policy Areas to support EU programmes, enabling projects to act as implementation vehicles while leaving space for regional adaptation. HELCOM operates in parallel as an autonomous, science-based regional governance body that develops BSAP targets, assessments, and recommendations, but it is not an EU implementation authority; EU policy operationalisation is primarily pursued through EUSBSR Policy Areas and EU funding programmes. While institutionally distinct, EUSBSR Policy Areas frequently build on HELCOM strategies, assessments, and targets, resulting in practical alignment across project platforms, monitoring activities, and policy feedback processes.

Together, these elements position the BSR water agenda within a multi-level governance system that provides coherence and strong links to EU regulatory frameworks. Yet the critical challenge remains converting this dense architecture into sustained uptake of circular water technologies at regional scale. Beyond engineering performance, scaling depends on alignment with regulatory compliance and oversight arrangements, an issue also documented in digital innovation contexts where data governance and legal risk management become integral to implementation (Li et al., 2024). While the BSR demonstrates remarkable success in developing innovative circular water technologies at pilot scale, translating these achievements to regional implementation reveals challenges that extend beyond technical feasibility. Across environmental management literature, technology readiness does not automatically translate to implementation readiness (Salvador-Carulla et al., 2024; Van Cauwenbergh et al., 2022), and substantial EU investments can yield proven technical solutions that still struggle to achieve the transformative impacts envisioned in policy frameworks (EPRS, n.d.).

Recent environmental governance scholarship increasingly recognizes that implementation requires systematic integration of technological capabilities with institutional frameworks, stakeholder engagement mechanisms, and economically sustainable models (HELCOM, 2021; Newig et al., 2023; Valentinov et al., 2025). This governance–technology integration approach moves beyond linear models where policy “creates a framework” and technology “fills it,” toward dynamic systems in which technical possibilities and governance structures co-evolve through iterative learning (Odebode, 2023). The BSR presents particularly complex integration challenges because of institutional diversity, varying national implementation capacities, and the transboundary nature of water management. HELCOM’s nutrient recycling strategy development exemplifies this complexity, requiring coordination between technical innovation, regulatory harmonization, and economic incentive alignment across multiple national jurisdictions (HELCOM, 2021). Understanding these integration dynamics is therefore essential for optimizing public investments already committed to circular water management and for designing future programmes that more effectively bridge the gap between technical innovation and regional-scale environmental impact (European Commission, 2025). The discourse evolution observed in practitioner forums offers additional insight into how the field itself recognizes and adapts to these challenges (Backer et al., 2010; Boström et al., 2016; European Commission, 2025).

Against this background, the study builds on critiques of linear implementation accounts (Ames, 1961; Godin, 2006), in which governance is treated as a static backdrop and outcomes hinge largely on technical or economic merits (Geels, 2002; Smith et al., 2010). While useful, such framings have been acknowledged in the literary canon to underestimate institutional complexity, feedback dynamics, and social contestation, especially in transboundary environmental governance (Ejdemo and Örtqvist, 2020; Hikkaduwa Liyanage et al., 2022). Sustainability transitions research further explains why technically mature solutions can fail to generate systemic impact (Köhler et al., 2019; Meissner et al., 2024; Turnheim and Sovacool, 2020), emphasizing iterative alignment between technological possibilities, governance architectures, and user practices (Kok et al., 2022; Meissner et al., 2024), and co-evolutionary dynamics producing contingent outcomes (Cataldo et al., 2025; Keller et al., 2022). These shifts motivate a multi-level governance perspective that can account for cross-scale coordination and feedbacks.

Multi-level governance (MLG) offers a useful lens for analysing coordination challenges across territorial and functional boundaries, where authority and capacity are distributed across supranational, national, regional, and local arenas, and across sectoral networks (Conteh and Harding, 2023; Sajida, 2025). Analyses distinguish between arrangements with clearer territorial responsibilities and those with more overlapping, task-specific roles (Benz, 2000; Hooghe and Marks, 2003). For environmental issues, effective coordination typically requires both vertical linkages (connecting higher-level frameworks to local implementation) and horizontal collaboration (aligning actors at the same level) (Yi et al., 2019). More recent studies, however, emphasize MLG as a dynamic process shaped by cross-scale interactions, shifting coalitions, and sectoral interdependencies (Maggetti and Trein, 2019). MLG can foster innovation when it enables experimentation and polycentric linkages, but it can also exacerbate fragmentation and incoherence when roles remain unclear (Di Gregorio et al., 2019; Kern and Rogge, 2018). In the BSR, coordination effectiveness depends not only on institutional architecture but also on the ability to link governance mechanisms to technological opportunities in nutrient recovery and wastewater treatment (EUSBSR, 2021; HELCOM, 2021). Classic implementation research identifies persistent gaps between policy intentions and outcomes (Pressman and Wildavsky, 1984; Winter, 2013), and contemporary perspectives argue these gaps often reflect misalignments between governance cycles and technology development pathways (Beck et al., 2021). Funding periods, regulatory reform, and cross-border coordination may operate on timescales that diverge from technology demonstration or market uptake, creating friction points where innovations stall. The challenge is therefore less about choosing governance “forms” and more about fostering adaptive coordination sensitive to institutional and technological dynamics (Carlisle and Gruby, 2019; Di Gregorio et al., 2019). Because coordination alone does not explain how technologies advance, we complement MLG with a transitions lens that focuses on how innovations move from protected niches into incumbent systems.

The multi-level perspective (MLP) on sustainability transitions explains how innovations evolve from niches toward adoption through interactions with regimes and landscape pressures (Geels, 2011; Köhler et al., 2019; Markard et al., 2012). Scaling depends on alignment with institutions, market structures, and societal expectations, and may involve hybridization and integration rather than substitution (Balanzó-Guzmán and Ramos-Mejía, 2023; Köhler et al., 2019). In water and nutrient governance, where infrastructures are long-lived and distributed across jurisdictions, innovations may complement and reshape legacy systems (Damman et al., 2023; Sánchez-Silva, 2018). At the same time, MLP categories can appear too sharply delineated in contexts of overlapping jurisdictions and shared responsibilities (El Bilali, 2019; Kuhmonen et al., 2024), reinforcing the need for an applied lens that foregrounds how governance arrangements and technologies co-evolve in fragmented, cross-jurisdictional settings.

Rather than proposing a novel theoretical framework, this study develops an integrative analytical lens that operationalizes existing coordination theory for circular water technology contexts. We synthesize three established coordination mechanisms, structural (Kellner et al., 2024; Ostrom, 2010), knowledge (Ferrari, 2020; Roux et al., 2023), and temporal (Haddad et al., 2022; Wiegant et al., 2024), to analyse how they function simultaneously across policy, practice, and project levels in transboundary environmental governance. The contribution is threefold: (i) demonstrating multi-scale simultaneity by tracing how mechanisms interact across tiers; (ii) extending adaptive governance theory by specifying how coordination demands vary with technological materiality (e.g., standardization, regulatory sensitivity, local embeddedness) under identical policy frameworks (Chaffin et al., 2014; Folke et al., 2004); and (iii) providing computational diagnostics to identify coordination mechanisms through discourse analysis, enabling systematic cross-context comparison beyond narrative case description.

1.1 Research questions (RQs)

This study asks: RQ1 How do EU and macro-regional policy frameworks in the BSR linguistically and structurally operationalise integration across circular economy, water quality, nutrient management, and hazardous-substance governance (Tier 1)? RQ2 How does practitioner discourse evolve over 2024–2025 in the Europe Forum Turku transcripts with respect to linking circular water technologies (e.g., nutrient recycling, wastewater/WWTP) to strategic and governance frames such as resilience and security (Tier 2)? RQ3 Across five Interreg BSR projects, what recurring governance–technology integration modes and coordination bottlenecks are visible in project texts, and how do these map onto circular economy scaling pathways from pilots toward broader implementation (Tier 3)? RQ4 How can structural, knowledge, and temporal coordination mechanisms be operationalised as a cross-scale analytical lens to interpret where scaling advances or stalls in transboundary circular water governance?

1.2 Hypotheses (H, testable expectations)

Because this study is observational and corpus-based, hypotheses are formulated as testable expectations about detectable cross-tier and temporal patterns (not causal claims). H1 Tier 1 policy texts will show higher prevalence and stronger clustering of structural integration language (cross-framework targets, coordination, monitoring) than Tier 2 and Tier 3. H2 Tier 2 discourse will shift from predominantly technical optimisation framing in 2024 toward stronger governance/strategic linking (including resilience/security framing) in 2025. H3 Tier 3 project texts will over-represent implementation constraints and integration bottlenecks (e.g., regulatory compliance, standards, procurement, financing/business models) relative to Tier 1. H4 Cross-tier comparison will reveal that structural, knowledge, and temporal coordination mechanisms co-occur but manifest differently across policy, practice, and projects, helping explain where scaling advances or stalls.

The structure of this paper is as follows. Section 2 describes the three-tier corpus design (Tier 1 policy documents; Tier 2 Europe Forum Turku 2024–2025 transcripts; Tier 3 five Interreg BSR projects), data cleaning and preprocessing, and the combined AntConc and InfraNodus workflow used to generate frequency, collocate, KWIC, co-mention, and semantic-network indicators. Section 3 reports findings across tiers, including discourse evolution and security–technology convergence (Tier 2), policy-level integration mechanisms and clustered policy architectures (Tier 1), project-level implementation testing and integration modes (Tier 3), and cross-scale coordination mechanisms mapped across the portfolio. Section 4 discusses implications for governance–technology integration, including discursive integration patterns, scaling pathways from pilot success to broader implementation, transferability conditions, and study limitations and future research directions.

2 Materials and methods

This study applies a three-tier corpus design, Tier 1 policy documents, Tier 2 Europe Forum Turku 2024–2025 transcripts, and Tier 3 texts from five Interreg BSR projects, using a fully computational workflow that combines AntConc lexico-collocational mining with InfraNodus semantic-network diagnostics. Following cleaning and preprocessing (stoplists, lowercasing, pmw normalisation, and consistent collocation windows), Tier 1 analysis identifies governance signatures and policy integration terms (RQ1); Tier 2 analysis tracks temporal discourse change and bridging vocabularies linking circular water technologies to strategic frames such as resilience and security (RQ2); Tier 3 analysis maps recurring governance–technology integration modes and bottlenecks around focal technologies (RQ3). Cross-tier concordance and cross-network comparisons then operationalise structural, knowledge, and temporal coordination mechanisms as a cross-scale analytical lens for interpreting where scaling advances or stalls between policy intent, practitioner translation, and project implementation (RQ4). The methodological workflow and RQ mapping are shown in Figure 1.

Figure 1

Figure 1. Methodological workflow and RQ mapping.

Table 1 summarises how each research question maps to the tier(s), tools, and outputs in the workflow.

Table 1

Table 1. RQ-to-method matrix table.

2.1 Tri-level integration analysis framework

This study employs a tri-level integration framework to examine governance-technology coordination processes across interconnected scales of decision-making and implementation, adapting MLP-oriented thinking to focus specifically on coordination mechanisms that enable or constrain circular water technology scaling. Tier 1 (supranational/policy) examines policy frameworks and enabling conditions for innovation, including EU circular economy and water governance instruments, HELCOM BSAP provisions relevant to nutrient recycling, and national implementation strategies developed between 2020 and 2025. Tier 2 (transnational/forum discourse) treats practitioner forums as knowledge-integration hubs that translate policy into operational strategies; Europe Forum Turku is used to analyse how practitioners adapt discourse based on implementation experience. Tier 3 (implementation/projects) analyses projects as controlled experimentation spaces where governance-technology integration mechanisms are tested across diverse technology-governance combinations.

2.2 Data sources and selection criteria

This study employs a tri-level integration framework to examine governance-technology coordination processes across interconnected scales of decision-making and implementation, adapting MLP-oriented thinking to focus specifically on coordination mechanisms that enable or constrain circular water technology scaling. Tier 1 (supranational/policy) examines policy frameworks and enabling conditions for innovation, including EU circular economy and water governance instruments, HELCOM BSAP provisions relevant to nutrient recycling, and national implementation strategies developed between 2020 and 2025. Tier 2 (transnational/forum discourse) treats practitioner forums as knowledge-integration hubs that translate policy into operational strategies; Europe Forum Turku is used to analyse how practitioners adapt discourse based on implementation experience. Tier 3 (implementation/projects) analyses projects as controlled experimentation spaces where governance-technology integration mechanisms are tested across diverse technology-governance combinations.

2.3 Analytical methods

We use a fully computational workflow with AntConc (Anthony, 2004; 2019) for lexico-collocational mining and InfraNodus (Paranyushkin, 2019; Tursunkulova et al., 2024) for semantic network analysis. All cross-subset comparisons report normalized frequencies per 1,000,000 tokens (pmw). A core stoplist is applied to all tiers; a transcript add-on stoplist is applied only to Tier 2 (forums) to remove speech disfluencies. Modals and negations (e.g., must/should/can/will; not/no/never) are retained. Unless we analyze dates, numbers are ignored. The wildcard * was used for pattern matching in node words and regex queries.

2.3.1 RQ1 (enabling governance): Tier 1 policy discovery

RQ1 uses a four-step discovery pass (Wordlist → Collocates → N-grams → KWIC) with a governance/technology node set capturing institutional levers and circular-water targets: policy, strate, actio, plan*, framewor*, directive*, regulation, HELCOM, fund*, monitor*, standard*, procure*, water*, wastewater, nature, scal***. Collocates are extracted with a symmetric L5-R5 window, ranked by log-likelihood (LL) for robust high-frequency ties, with mutual information (MI; AntConc “Effect”) used to surface informative lower-frequency ties. Thresholds are min frequency ≥3 and range ≥2 files; pmw is reported to three decimals. We then mine 2–3-g (min frequency ≥2) and run regex-guided clusters (nature, phosphorus|nitrogen, recovery, recyc, green, monitor, requirement**) to capture canonical policy formulations and mechanism phrases. Finally, KWIC contexts (25 tokens; sorted by Node +1 and Node −1) document 2-3 evidentiary lines per mechanism family, including funding, standards/certification, monitoring/assessment, procurement, permitting, and explicit policy-to-technology linkages. Together, collocates, n-grams, and KWIC operationalise enabling governance as a reproducible distributional signature.

2.3.2 RQ1 specificity - Tier 1 vs. (Tier 2 + Tier 3)

We run an AntConc keyword comparison with Tier 1 as target and Tiers 2 + 3 as reference using LL, p ≤ 0.001, min frequency ≥5, and a reference range filter ≤2 files to emphasise Tier-1 salience. Tier-1-specific keywords (e.g., embedding, managing authority, steering group, priority substances, assessment cycle) are then examined in Tier 1 using Collocates (L5–R5, ranked by LL with MI as complement) and documented with 2–3 KWIC lines per keyword (context 50–80, sorted by Node ±1). InfraNodus triangulates the same partitions: policy mechanisms are expected to appear within policy communities and/or as high-betweenness bridges linking governance terms (e.g., directive/plan/funding) to technology targets (e.g., wastewater/nutrient/nature-based). Where InfraNodus suggests structural gaps, KWIC is used to distinguish upstream specification without downstream uptake (fragmentation) from practice-led emergence without explicit policy anchoring.

2.3.3 RQ2 (integration and learning) - Tier 2 forums

Tier 2 analysis applies the transcript add-on stoplist and focuses on practice verbs/instruments using the node set pilot, project, procure, standard, monitor*, permit*, financ*, fund*, risk, challenge, cost*, border, knowledg*, capacity, resilien*, security***. Collocates use L5–R5, with MI as the primary ranking for informative lower-frequency ties and LL to confirm robust associations; thresholds are min frequency ≥5 and range ≥2 sessions. 2–3-g (min frequency ≥3) capture recurring practice idioms (procurement, water/nutrient discussions, activated carbon, feasibility, nature-based). Additional stance/learning constructions are retrieved via regex patterns for the same node set, including co-occurrence with barrier/challenge near permit*/standar*/fund*. For temporal learning, the corpus is split into 2024 vs 2025, and keyword tests run in both directions using LL, p ≤ 0.001, min frequency ≥5. Shift terms are then checked via Collocates and Concordance Plot dispersion to ensure changes reflect broad uptake rather than single-session effects.

2.3.4 RQ3 (project-level coordination) - Tier 3 projects

Tier 3 maps coordination mechanisms co-occurring with focal technologies using the node set water, pilot, nutrient, phosphorus, sludge, business, scal***. Collocates use L5-R5, ranked by LL (robust ties) with MI as complement (informative ties), with thresholds min frequency ≥3 and range ≥3 files for cross-project generalisation. The goal is to detect technology ↔ coordination linkages (e.g., nutrient with procurement/permitting/standards/monitoring; business* with offtake/pricing/contract; pilot* with scale/funding/extension). Barrier/enabler patterns are retrieved via regex (barrier|challenge|risk|resilien|securit|national***). KWIC is sorted by Node +1 to capture governing verbs (e.g., “procurement enables … ”, “permitting delays … ”, “standards require … ”, “funding supports … ”), distinguishing obstructive vs facilitative contexts. Clusters are captured via targeted 2–3-g around public, responsibility, standard, program, feasibilit, investment, stakeholder*** (e.g., feasibility study, investment plan, stakeholder engagement, program criteria).

2.3.5 Cross-tier mapping (mechanism concordance)

Cross-tier concordance tests whether policy-named mechanisms are expressed consistently in Tier 2 and Tier 3. A mechanism seed list derived from Tier 1 outputs includes embedding, procurement, certification, monitor, permit, nutrient recovery, nutrient***. Concordance is run separately per tier (core stoplist; plus Tier 2 add-on) and exported with File columns to compute file-level coverage (files hit/total files per tier), indicating dispersion rather than within-text intensity. KWIC (25 tokens; Node ±1 sorting) is sampled to exclude list/bibliographic noise, and LL/MI collocates are checked to confirm operational contexts (e.g., procurement with criteria/tender/life-cycle; monitor* with indicator/compliance/assessment). Interpretation follows: (i) high Tier 1 + medium/high Tier 2/3 = coherent enabling; (ii) high Tier 1 only = policy–practice gap; (iii) high Tier 2/3 with low Tier 1 = practice-led emergence; (iv) mixed patterns = partial translation or sector constraints.

2.3.6 InfraNodus semantic network analysis (features only)

Semantic co-occurrence networks are built separately for each tier (Tier 2 uses the transcript add-on). Terms are nodes; co-occurrences within the InfraNodus default window form edges (weighted by co-occurrence counts). Community detection identifies thematic communities; nodes are ranked by betweenness centrality (bridging potential) and degree (local salience). For each tier and year-slice we report nodes/edges, modularity, top communities (labelled by representative terms), top-10 betweenness nodes, and top-10° nodes. Cross-network comparisons examine (i) whether policy communities (directive/plan/funding) align with practice/project communities (wastewater/nutrient/pilot/procurement), and (ii) whether betweenness shifts in Tier 2 (2024→2025) indicate emerging integration hubs (e.g., procurement, permitting, business model, monitoring). Where gaps appear, shortest-path inspection identifies missing or underexpressed bridging terms and KWIC is used to contextualise the absence.

2.3.7 Reporting conventions

All frequency comparisons use pmw over preprocessed token counts; files are lowercased; core stoplist applied (plus Tier 2 add-on); modals/negations retained; numbers ignored unless dates are analysed. Collocations report LL (robust, high-frequency ties) and MI (AntConc “Effect”; informative lower-frequency ties). We use a symmetric L5-R5 window as a mid-range span suited to mixed genres, capturing clause-level mechanism phrases without collapsing into topical co-presence; a robustness check with L3-R3 yielded consistent top-ranked patterns, so results are reported using L5-R5 with L3-R3 noted as sensitivity. Keyword tests use LL, p ≤ 0.001 with stated minimum frequency/range filters. Cross-tier mechanism coverage is reported as file-level proportion rather than pmw. N-grams (2–3) include raw counts and pmw with noted regex patterns. Precision: pmw to three decimals; association measures to two.

2.4 Methodological innovation

This study delivers a data driven, computational account of governance-technology integration using only AntConc and InfraNodus. AntConc provides parameterized, reproducible evidence of enabling mechanisms (via collocates, n-grams, and pmw-normalized keywords) at policy, forum, and project levels. InfraNodus contributes semantic network clusters and betweenness-based bridging diagnostics that reveal how integration hubs emerge and strengthen across tiers and over time. Together, these tools quantify coordination patterns, identify gaps, and surface concrete mechanism-technology linkages that inform scalable implementation.

3 Findings

3.1 Discourse evolution: toward integration strategies

Across 2024–2025, the discourse evolves from a predominantly technical framing (e.g., wastewater treatment plants, recovery technologies, recycled fertilizers) toward an explicitly strategic/policy framing that embeds nutrient recycling and wastewater management within regional resilience and security agendas.

Figure 2 shows the discourse security-technology convergence from 2024–2025. Quantitative evidence reveals the trajectory of this evolution: In 2024, the session transcripts lend little clear indication of usage of core technical terms such as nutrient recycling or wastewater/WWTP, while security and geopolitics dominate its discussion. Across the full Tier 2 2024 set (including the EU keynote), security appears at ≈26 pmw, and the technical bundle at ≈5.9 pmw. In the span of a year, by 2025, technical terms clearly emerge (e.g., nutrient recycling, wastewater) and we observe the first documented co-presence within the same discussions of technical and strategic language (e.g., wastewater/WWTP items in sessions that foreground security and resilience), a coupling not observed in the 2024 sample. See Figure 2 for per-10,000-word rates (Security ≈14.5 pmw; Technical bundle ≈9.5 pmw in 2025).

Figure 2

Figure 2. 2024–2025 discourse security-technology convergence in Tier 2 data.

The collocate analysis of nutrient underscores a technical semantic core, with strongest associates as recycling (MI 6.994), fertilizing (MI 6.526), nitrogen (MI 4.941) and notably lack immediate collocates such as security, resilience, or strategy. This indicates that strategic links form via broader discursive bridges rather than immediate word-window proximity. KWIC (±12) shows these bridges in practice: in the forum transcripts, nutrient recycling appears alongside regional coordination and policy-area roles, and wastewater within planning/stakeholder debates; in the EU water-security speech, nutrient recycling is tied to strategic autonomy, a European water resilience strategy, and Interreg BSR as instruments for security and resilience.

KWIC evidence (Table 2), proximity windows were tested at ±12 words, suggests how these bridges manifest in practice:

• In forum transcripts, nutrient recycling is discussed alongside regional coordination and policy area roles, while wastewater appears within debates about regional planning and stakeholders (see representative KWIC snippets, Forum transcripts).

• In the EU water security speech, nutrient recycling is linked directly to Europe’s strategic autonomy, a European water resilience strategy, and programs like EU Interreg BSR (Baltic Sea Region) as instruments for security and resilience (see representative KWIC snippets, EU Water Security speech).

Table 2

Table 2. Representative KWIC snippets window at ±12 words, Europe Forum 2024 and 2025 transcripts.

These indicators suggest co-presence in discourse and discursive re-framing: technical terms retain process-oriented collocates while appearing in KWIC contexts that include policy and strategic framing. Wastewater and nutrient recycling are no longer narrated solely as engineering tasks; they are increasingly positioned as levers of regional strategy, autonomy, and resilience through a discourse that layers technical competence with strategic necessity.

3.2 Multi-scale integration mechanisms

3.2.1 Policy level: enabling frameworks (Tier 1 corpus data)

The analysis reveals a multi-level governance architecture that enables circular economy integration across multiple sectoral frameworks while maintaining regional differentiation. This integration operates through three primary mechanisms: cross-sectoral policy harmonization, adaptive implementation pathways, and coordinated monitoring systems.

3.2.1.1 EU circular economy integration across sectors

The policy corpus (Tier 1, comprising 12 documents) shows lexical patterns consistent with circular economy integration across EU water, agricultural, and marine policy domains. The analysis combines frequency data (AntConc) with semantic network patterns (InfraNodus).

3.2.1.1.1 Lexical evidence of policy integration

The corpus contains 84 references to “circular economy” alongside prominent mentions of the three major policy frameworks: Water Framework Directive (WFD: 16 occurrences), Common Agricultural Policy (CAP: 32 occurrences), and Baltic Sea Action Plan (BSAP: 23 occurrences). The terms “directive” (269) and “framework” (213) appear frequently, indicating policy-oriented discourse. Water-Related Discourse Cluster: KWIC concordance search for water-related terms (water) yields 541 hits, with quality management terminology prominently featured (water: 405, body: 145, quality: 244). Of these water-focused contexts, 84 co-occur with circular economy language, suggesting thematic linkage between water governance and circular economy principles. Action-Oriented Implementation Language: High-frequency implementation terms show substantial co-occurrence: “action” (500), “policy” (346), and “plan” (182). The term “project” appears 101 times. Semantic network analysis (InfraNodus) identifies four primary topical clusters: (1) Sustainable Policy (eu-policy-food), (2) Water Quality (water-body), (3) Circular Action (action-plan-project), and (4) Substance Management (substance-framework-management).

These lexical patterns suggest three integration mechanisms: cross-directive coordination (evidenced by co-occurrence of policy framework terms with water and circular economy language), sectoral bridging through action plans (high frequency of action-plan-project collocations), and implementation coherence (systematic pairing of action-policy-plan terminology). However, these remain analytical interpretations of frequency and co-occurrence patterns rather than explicitly named frameworks in the source documents.

3.2.1.2 Regional differentiation within coordinated frameworks

KWIC analysis of Baltic Sea Region governance reveals a multi-stage coordination process. Commission references predominantly feature legislative-procedural language (will, shall, propose, submit; likelihoods 52–305), consistent with its role in initiating policy frameworks. HELCOM collocates center on institutional coordination (regional [113.6], contracting parties [30–31], BSAP [30.9], actions [56.9]), indicating its function in organizing regional implementation through structured policy instruments rather than technical specification. The strong association between “plan” and action (likelihood: 636.6) suggests that action plans constitute the interface where Commission-level policy frameworks are coordinated across the region through multiple channels (notably EUSBSR Policy Areas and programmes), while drawing on HELCOM assessments and the BSAP as a shared regional scientific and strategic baseline. The core lexical signals supporting these mechanisms are summarised in Table 3, with the full indicator matrix reported in Supplementary Table S1.

Table 3

Table 3. Summary of policy integration mechanisms evidenced in Tier 1 (full table in Supplementary Table S1).

3.2.1.3 Textual evidence of policy integration mechanisms

Systematic coverage across the policy architecture is evidenced by the distribution of policy-related terminology throughout the corpus. The wildcard search for “polic*” yields 448 occurrences distributed across 11 of 12 documents, representing 91.7% corpus coverage. This distribution pattern suggests comprehensive integration of policy discourse across different framework documents rather than concentration in isolated policy statements. The EUSBSR Action Plan 2021 contains the highest concentration with 108 occurrences (24.1% of total policy terminology), indicating its role as a coordinating framework that synthesizes multiple policy streams. Secondary sources include the Farm to Fork Action Plan (14 occurrences), HELCOM hazardous substances framework (5 occurrences), and EU Circular Economy documents (3 occurrences). Notably, one document contains no policy terminology, suggesting either a technical focus or a potential gap in policy integration coverage.

Institutional coherence emerges through systematic collocational patterns that link action-oriented implementation language with policy frameworks. The term “action” frequently collocates with “plan” with mutual information score of 2.66, appearing 126 times within a 5-word window of action terminology. This “action plan” construction shows asymmetric positioning, with “plan” appearing 111 times to the right of “action” versus only 15 times to the left, indicating a conventionalized phrase structure that serves as the primary vehicle for translating policy objectives into implementation mechanisms. The term “policy” itself co-occurs 64 times in close proximity to action terminology, suggesting systematic integration rather than parallel development of policy frameworks and implementation strategies.

Regional coordination mechanisms are visible through the collocational patterns of institutional and geographic terminology. The term “helcom” appears 51 times in contexts associated with action terminology, indicating its prominence as a regional coordination platform where action-oriented objectives and governance often intersect with broader EU policy agendas. Similarly, “sea” appears 78 times near action terms, reflecting the geographic specificity of implementation while maintaining connection to broader policy objectives. This pattern suggests that regional bodies operate as translation mechanisms that maintain policy coherence while adapting frameworks to local contexts. The procedural language “under” appears 151 times in action-related contexts across multiple documents, indicating systematic use of directive-based compliance framing that links regional actions to authorizing policy frameworks.

Sectoral integration patterns reveal cross-domain coordination through the co-occurrence of sectoral terminology with implementation language. The term “agricultural” appears 24 times near action terminology across 7 documents, suggesting integration of agricultural policy considerations within broader environmental governance frameworks. The collocate “best” appears 22 times with a mutual information score of 2.19, likely reflecting “best practices” discourse that facilitates knowledge transfer across sectors and scales. The term “key” appears 27 times in action contexts, functioning as an emphasis marker that highlights priority integration points, while “include” appears 21 times, suggesting frameworks that deliberately incorporate multiple sectoral perspectives.

Table 3 highlights the most central indicators; the complete evidence base (all indicators, frequencies, MI values, and document coverage) is provided in Supplementary Table S1. The semantic network analysis conducted through InfraNodus reveals a modularity score of 0.43, indicating moderate clustering of discourse into distinct but interconnected thematic areas. Four primary topical clusters emerge: Sustainable Policy (eu-policy-food), Water Quality (water-body), Circular Action (action-plan-project), and Substance Management (substance-framework-management). The moderate modularity suggests neither complete integration nor fragmentation, but rather a structured architecture where distinct policy domains maintain some autonomy while remaining connected through shared terminology. A critical gap appears between the Sustainable Policy and Substance Management clusters, suggesting potential weakness in linking broad sustainability frameworks with specific hazardous substance governance mechanisms. However, 13 bridge concepts (including body, substance, European, management, framework, data, water, deadline, priority, target, baltic, baseline, and quality) connect these clusters, providing potential pathways for enhanced integration.

Implementation effectiveness is suggested by the prevalence of operational language throughout the corpus. The terms “this” (188 occurrences) and “other” (75 occurrences) appear frequently near action terminology, indicating extensive cross-referencing between policy instruments that may facilitate coordinated implementation. The appearance of temporal markers such as “deadline” and outcome-oriented terms such as “baseline” and “target” in bridge positions within the semantic network suggests the presence of coordinated monitoring and accountability mechanisms that enable tracking of implementation progress across multiple governance levels. The term “data” similarly occupies a bridge position, indicating potential integration of monitoring systems that could support adaptive management across scales.

The corpus analysis identifies textual patterns that characterize the architecture of multi-scale integration mechanism. These patterns collectively (Table 3; Supplementary Table S1) suggest three primary integration outcomes visible at the lexical level. First, framework coordination is evidenced by the systematic co-occurrence of policy terminology with implementation language across multiple documents, indicating deliberate efforts to maintain coherence while translating objectives into action. Second, institutional intermediation appears through the collocational patterns associated with regional bodies like HELCOM, which demonstrate linguistic patterns consistent with translation between EU directives and regional implementation contexts. Third, sectoral bridging is visible in the co-occurrence of domain-specific terminology (agricultural, water, substance management) with shared implementation frameworks, suggesting mechanisms that enable cross-sectoral coordination while maintaining sectoral specificity.

3.2.2 Forum level: knowledge integration (Tier 2 corpus data)

The Europe Forum in Turku serves as an intermediary space where macro-regional governance frameworks can, via discourse and narrative exchanges, translate into actionable technology deployment strategies. Analysis of forum discourse across 2024–2025 reveals a systematic progression from security-framed policy articulation to technical implementation pathways, showing how participants discuss translation from policy framing toward implementation language. The dynamic process by which institutional frameworks shape and are shaped by technological innovation trajectories.

3.2.2.1 Practitioner-driven solution adaptation

Forum discussions exhibit a clear temporal shift in knowledge integration patterns. The 2024 sessions established water explicitly within security and resilience paradigms. Official framing positioned “water security as an integral part of citizens’ general safety” and linked nutrient recycling to Europe’s “strategic autonomy” through reduced fertilizer import dependence exemplifies this security-technology nexus. This framing creates institutional legitimacy for technological interventions, the alignment of policy narratives with technological deployment possibilities. The Baltic Sea Region Forum’s focus on “security of maritime transport” alongside “nutrient recycling” indicates how practitioners adapt broad policy frames to concrete implementation contexts. Corpus analysis reveals this transition through vocabulary shifts: while security terms maintain prominence (security: 236 occurrences across 7 files), implementation language increases substantially (implementation: 31 occurrences across 6 files), indicating practitioner focus on translating policy intent into technical action.

3.2.2.2 Cross-sector learning and technology transfer

The forum architecture facilitates multi-stakeholder dialogue across institutional boundaries. The “Comprehensive Security and EU’s Research and Innovation” session exemplifies cross-sector knowledge exchange by integrating dual-use technologies with security frameworks. Forum discourse bridges traditionally separate domains, environmental management, security policy, and innovation systems, as evidenced by the co-occurrence of terms from these distinct fields.

Corpus analysis reveals high-frequency integration vocabulary: “research” (193 occurrences, range 7), “innovation” (128, range 6), “development” (100, range 7), and “cooperation” (39, range 6). These terms appear across nearly all forum sessions, indicating systematic cross-sector engagement. The co-occurrence of “EU” policy language (469 instances) with “security” discourse (236 instances, range 7) alongside technical terminology (“technology” 63, “technologies” 59) suggests knowledge boundary spanning between governance, security, and technological domains. The temporal distinction is also indicated by sentence co-mention evolution in the data:

• • 2024: 52 sentences, 0 water-security co-mentions

• • 2025: 429 sentences, emergence of cross-domain co-mentions

Network analysis (modularity 0.25) identifies four conceptual clusters in forum discourse: EU Governance, Security Priorities, Global Economy, and Temporal Dynamics. Security-related nodes occupy central network positions with connections spanning multiple clusters, functioning as discursive bridges between governance frameworks and economic/technological discussions. However, the moderate modularity score indicates clusters maintain some structural separation, the network shows integration mechanisms alongside persistent domain boundaries, reflecting ongoing processes of cross-sector learning rather than complete institutional convergence. These cross-sector integration themes and their 2024–2025 shift are summarised in Table 4.

Table 4

Table 4. Cross-sector integration themes from practitioner discussions.

3.2.2.3 Public engagement strategies and acceptance building

As shown in Table 4, forum discourse increasingly links technical water measures to security and strategic autonomy, reframing nutrient recycling and wastewater upgrades as responses to geopolitical and economic risks. In 2025, new co-mentions emerge (wastewater↔security; resilience↔security), while the vocabulary shifts from broad “water” talk (180→32, −82%) to targeted “nutrient” discourse (7→62, +786%), indicating a move toward actionable interventions. Speakers emphasize EU-level instruments and funding levers (e.g., Interreg BSR; EU regulatory files) as pathways from policy to implementation, consistent with the Tier-2 network where EU sits as the key bridge (highest betweenness; modularity ≈0.25). This progression may be interpreted as legitimacy transfer: established priorities (security, strategic autonomy) lending credibility to specific technological interventions (nutrient recycling, wastewater upgrades). Claims about “resilience” should be tempered, as its frequency declined overall (21→15, −29%), even though it appears in the new 2025 co-mentions.

3.2.3 Project level: implementation testing (Tier 3 corpus data)

Based across five Interreg BSR projects, three distinct technology-governance integration modes recur, each probing different scaling paths beyond pilots. AntConc collocates and clusters emphasize practice-proximate themes such as recycling, treatment, management, flood, sewage, nitrogen, fertilizers, and governance-adjacent bigrams, such as stakeholder engagement (freq 18, range 4), public perception, public health, standards for/and, investments in, indicating that technical deployment is consistently entangled with stakeholder, standards, and investment work (Tier-3 top collocates/top bigrams). These project-level integration approaches are compared across the five cases in Table 5.

Table 5

Table 5. Integration approach comparison across 5 interreg projects.

Where governance enables vs. constrains

• National-level constraints. Regulatory fragmentation across BSR countries surfaces as a recurring barrier (e.g., divergent water-quality standards affecting membrane configurations; uneven acceptance of nutrient recovery). This limits programmes’ support function even when EU frameworks are shared.

• Municipal-level adaptation. City-scale projects (e.g., City Blues, WaterMan) seem to indicate that coordination capacity is pivotal: where dedicated resources and cross-department routines exist, uptake of NBS/real-time control is feasible; elsewhere, expertise and bandwidth gaps slow transferability.

• Regional-level gaps. The operational separation between Priority 2 (Water-smart societies) and Priority 3 (Climate-neutral societies) creates coordination seams: nutrient-circularity efforts (e.g., CiNURGi) are governed separately from water-focused projects, these seams could potentially cur integrated regional planning.

To consolidate these recurring modes and their scaling implications, Table 5 summarises programme context, governance integration, technology adaptation, and business-model framing across the five projects.

3.2.3.1 Technology adaptation to governance contexts

Drawing on the cross-project comparison in Table 5, technology adaptation to governance contexts follows three recognizable patterns. A regulatory-compliance strategy, evident in ReNutriWater and NURSECOAST-II, prioritizes fit-to-standard designs, such as membrane technology configurations and seasonal nanobubble modules, that deliver new value while remaining within existing legal and water-quality frameworks. A policy-integration strategy, seen in CiNURGi and WaterMan, depends on creating new institutional arrangements, including cross-border agreements, integrated emergency planning, and multi-municipal coordination, to make deployment viable at scale. A stakeholder-engagement strategy, exemplified by City Blues, relies on co-creation and the routine participation of citizens and municipal departments; here, governance adaptation is inseparable from technical performance because operation and maintenance practices must be embedded locally.

3.2.3.2 Business model development for post-pilot sustainability

Business-model and financing language in the project texts is present but often generic rather than specifying revenue mechanisms. CiNURGi frames post-pilot sustainability in terms of creating business opportunities around nutrient recycling. ReNutriWater materials explicitly reference ongoing work to refine business model concepts, but do not detail the revenue logic in the public texts captured here. NURSECOAST-II discussion highlights financing challenges and points to incentives and income-generation narratives as potential enablers. City Blues outputs emphasize operational models and lifecycle implementation guidance for NBS rather than a defined business model in the available project texts.

3.3 Circular economy scaling pathways

Figure 3 presents the integration pathway typology (P1-P3) used to classify circular-economy scaling dynamics across the portfolio.

Figure 3

Figure 3. Integration pathway typology.

Example anchors:

• ReNutriWater, NURSECOAST-II → Pathway 1 (P1)

• City Blues → Pathway 2 (P2)

• CiNURGi, WaterMan → Pathway 3 (P3)

3.3.1 Successful scaling examples: CiNURGi nutrient recycling networks

CiNURGi illustrates a mature scaling pathway in which technology development is coupled to governance innovation and then propagated toward system-level change. The project’s use of modular, standardized processing units enables deployment across heterogeneous local conditions while cross-border cooperation agreements resolve coordination problems that single municipalities cannot fix alone. Other projects offer a fee-for-service arrangements, so that nutrient recovery becomes financially legible within municipal budgets, allowing existing wastewater streams to be valorized without new core infrastructure. The combination of modularity, intermunicipal governance, and a clear payment mechanism, creates network effects: participants share operating know-how and procurement costs, which lowers barriers for late joiners and accelerates diffusion across countries.

3.3.2 Adaptive management in ReNutriWater and WaterMan projects

Figure 3 also highlights how ReNutriWater aligns with P1, whereas WaterMan aligns with P3, illustrating adaptive management under different scaling logics. ReNutriWater and WaterMan collectively illustrate adaptive management across two strands of the typology: ReNutriWater aligns with the technical-optimization pathway (P1), while WaterMan aligns with the technology-development → governance-innovation pathway (P3). ReNutriWater iteratively tunes membrane configurations to local water-quality baselines and prevailing regulatory standards, with successful set-ups circulated among partner utilities. WaterMan’s infrastructure is deliberately modular so assets can be reconfigured for drought or flood contexts and then embedded in multi-municipal emergency planning. In both cases, adaptation is real but bounded by programme accountability: fixed deliverables and reporting cycles can narrow the space for ongoing experimentation, creating a practical tension between flexibility and compliance that slows scaling even when the engineering is sound.

Across the portfolio, recurrent barriers map cleanly onto the figure’s transition points. Regulatory fragmentation hampers movement from technical optimization to stable compliance, while limited long-term financing interrupts progression from compliance to market integration. In addition, weak coupling between technical work and stakeholder processes constrains the shift from policy innovation to durable network effects. The prominence of standards-, stakeholder-, and investment-related terms in the Tier 3 results is consistent with these choke points, underscoring that successful scaling hinges as much on governance design and revenue logic as on the performance of the technologies themselves.

3.4 Coordination mechanisms across scales

Across the five projects, coordination failures seem to stem less from intent than from the absence of durable, cross-scale mechanisms. Regional platforms such as Interreg BSR and HELCOM (Baltic Marine Protection Commission) provide agenda setting, alignment with environmental objectives, and periodic coordination, but they are not designed to internalize business-model risk or to coordinate operations once projects end. National interfaces such as utilities and environmental agencies, translate EU and regional aims into country rules; they work well for fit-to-standard deployments but become bottlenecks where cross-border market rules or new policy arrangements are needed. At municipal level, departments function as effective boundary spaces during funded pilots, especially where cross-department routines and co-creation processes exist, yet this capacity is fragile when dedicated resources lapse.

Three mechanism types emerge from practitioner experience. Boundary organizations facilitate multi-scale integration but with uneven mandate fit: regional bodies align goals, national bodies enforce compliance, and municipal bodies enable day-to-day implementation, however, none consistently bridge economic integration for circular water technologies. Knowledge translation operates through technical documentation, practitioner networks, and policy integration. Technical reporting captures implementation, but often lacks the specificity and legal framing needed for rapid policy uptake; peer networks spread practices horizontally yet depend on continuing institutional participation; and policy recommendations translate lessons upward but require contextual adaptation frameworks to land in different municipal settings.

The InfraNodus semantic network (Figure 4) visualizes coordination and translation dynamics across the Tier 3 project corpus. Two dominant hubs, project and water, structure the network, indicating that implementation discourse is organized around project-based coordination linked to a shared water challenge. A macro-regional programme spine (interreg, baltic, sea, region) connects tightly to project, reflecting how Interreg framing and regional identity anchor implementation narratives. A technical operations community clusters around water (wastewater, treatment, nutrient, plant, reuse, technology), representing the operational and technology-facing content of deployment. Cross-cluster ties concentrate on practice-proximate bridge terms, especially area, pilot, policy, solution, management, and development, which connect programme framing to technical action, indicating where governance arrangements and implementation routines must meet for scaling to occur. Named project nodes (e.g., ReNutriWater, NURSECOAST-II, CiNURGi) attach to the central hub, while sector/place terms (e.g., tourism, city) sit on connected branches, signalling context-specific constraints and uptake conditions. Overall, the network makes visible a translation pathway from regional programme framing → project coordination → pilot/management routines → technical implementation in wastewater and nutrient recovery, clarifying where scaling advances or stalls when coordination, compliance, and post-project market formation are weakly linked.

Figure 4

Figure 4. InfraNodus semantic network for Tier 3 corpus data, visualizing coordination and translation dynamics across the projects.

Three coordination patterns recur with distinct scaling implications. Hub-and-spoke arrangements arise when a single project coordinates a transnational network, effective for momentum, but dependent on external programme scaffolding. Peer-to-peer consortia enable rapid knowledge diffusion among municipalities, yet coordination fades when funding cycles end. Hierarchical separation between thematic priorities (for example, water versus circular economy) creates institutional seams that hinder integration despite overlapping objectives. These patterns map neatly to the typology’s transitions and explain where progress slows.

Limitations are structural. Resource dependency means the most effective coordination appears during projects and wanes after closure. Temporal misalignment places multi-year regulatory cycles against shorter project and municipal budgeting horizons, constraining iterative adaptation. Scale mismatches persist because technical pilots can mature faster than the regional regulatory and market harmonization they require. Taken together, the evidence indicates that scaling circular water technologies will require purpose-built coordination mechanisms that persist beyond individual projects, explicitly target post-project market formation, and formalize the translation pathways identified in the network.

4 Discussion

4.1 Governance–technology integration as an analytical lens

Building on the empirical findings in Section 3, governance–technology integration can be interpreted as an emerging analytical lens for understanding environmental technology implementation in multi-level governance settings. Across the Tier 2 corpus, the discourse shifts between 2024 and 2025 from predominantly technical descriptions (peak 178 mentions) toward more explicitly integrative discussions that situate circular water technologies within broader strategic and governance contexts (79 mentions in 2025). In this evolving discourse, nutrient recycling and wastewater technologies increasingly appear in proximity to strategic language linked to resilience and regional coordination, indicating a broadening of how technical interventions are framed within policy-relevant narratives (Fig. 3.1; Table 2). At the same time, the Tier 2 collocate evidence shows that technical concepts retain a process-oriented semantic core (e.g., nutrient collocates with recycling, fertilizing, nitrogen), while strategic links more often emerge through broader discursive bridges visible in KWIC rather than as immediate collocates (Section 3.1). The empirical pattern seems to support an interpretation that implementation challenges are increasingly being discussed not only as problems of technical optimisation, but as coordination challenges spanning institutional levels, policy domains, and temporal cycles (Sections 3.1–3.4). These observations are specific to the analysed corpus and should be interpreted as indicative of changing practitioner perspectives rather than as evidence of a general theoretical transition.

The contribution of this study is in illustrating the operationalisation of existing governance and transition theories into a coordination-focused analytical lens that clarifies how technologies may be positioned for scaling within complex, transboundary governance systems (RQ4). While multi-level perspective (MLP) theory provides valuable insights into niche–regime dynamics, the Baltic Sea evidence suggests that environmental governance in practice often operates through overlapping jurisdictions and shared responsibilities that blur categorical boundaries. In the Tier 2 forum network, the modularity score (0.25) indicates structured clustering alongside persistent separation between governance and technical communities, and this configuration is consistent with integration being discussed through dynamic bridging processes rather than fixed institutional arrangements (Table 4). In parallel, Tier 1 policy patterns show a structured but interconnected policy architecture (modularity 0.43) in which action-plan language and cross-referencing (“action–plan–project,” “under,” “data,” “deadline,” “target”) function as translation vehicles linking frameworks, domains, and monitoring logics (Section 3.2.1; Table 3; Supplementary Table S1). Taken together, the evidence supports treating governance–technology integration as a coordination lens that can capture fluid, network-based forms of alignment that are not readily explained by purely hierarchical or level-based models alone (Sections 3.2.1–3.4).

However, several tensions warrant explicit acknowledgement. First, the discourse analysis methodology, while well suited for tracking practitioner learning and framing dynamics, may conflate linguistic evolution with deeper institutional transformation. The observed shift toward strategic framing could reflect political positioning or agenda alignment rather than substantive improvements in coordination practices. Second, the empirical scope is constrained to a single regional context characterised by high institutional density, namely, the EU’s multi-level governance (MLG) architecture and the Baltic Sea Region’s long-standing traditions of transnational cooperation. The applicability of the lens to contexts lacking similar institutional capacity or cooperative traditions therefore requires further investigation (see also Section 4.4).

A further implication emerging across tiers is a temporal paradox in governance–technology integration. While policy frameworks typically operate on multi-year cycles and technological development follows iterative pathways, the project-level evidence suggests that effective integration often depends on short-lived and difficult-to-anticipate alignment windows (Sections 3.2.3–3.4). Across the five EU Interreg BSR projects, capacities to navigate this temporal coordination challenge appear uneven: CiNURGI’s modular design and cross-border governance arrangements align with a comparatively mature scaling pathway, while other cases remain more constrained by resource dependency and post-project institutional fragmentation (Table 3.2.3; Section 3.3). This reinforces the value of a coordination-focused approach that attends not only to structural and knowledge-based mechanisms, but also to temporal alignment processes that enable conjunctural opportunities for scaling and system-level change (Section 3.4).

4.1.1 Discursive integration patterns in Tier 2 (interpretive propositions)

To clarify how governance–technology integration is constructed in practitioner discourse, without implying verified institutional change, the Tier 2 corpus suggests three recurrent rhetorical patterns (Sections 3.1–3.2.2; Table 4). First, policy-technology coupling is visible in the simultaneous intensification of regulatory language and technical terminology. Regulatory discourse rises markedly between years (EU +394%, regulation +490%, commission +255%) while co-occurring with more implementation-specific technical vocabulary (nutrient +786%, wastewater +64%). This pattern is consistent with practitioners positioning technological interventions within institutional frameworks through linguistic association, rather than treating technology as separate from governance (Section 3.2.2; Table 5).

Second, strategic bridging discourse indicates how established priority frames are used to connect domains that are otherwise discussed in parallel. Security terminology remains relatively stable in frequency (115→121), yet gains new discursive connections: in 2025, co-mentions between security and wastewater, and between security and resilience, appear where none were observed in the 2024 sample. With EU exhibiting the highest betweenness centrality (≈0.338), security language functions as a rhetorical connector linking environmental, economic, and strategic domains, enabling technical discussions to be articulated within recognised governance priorities rather than as purely sectoral engineering issues (Section 3.2.2; Fig. 3.2.2; Table 5).

Third, targeted reframing is reflected in a narrowing from broad environmental references to more specific, actionable intervention language. The discourse shifts from generic “water” talk (−82%) toward targeted nutrient-oriented framing (nutrient +786%). This narrowing-and-intensification pattern, together with the emergence of new strategic co-mentions, is consistent with practitioners increasingly framing technical solutions as responses to higher-level policy priorities, such as resilience, security, and strategic autonomy, rather than as standalone environmental measures (Section 3.2.2; Table 5).

These patterns describe discourse structures that may facilitate or reflect integration processes; however, corpus evidence alone cannot verify whether such linguistic strategies translate into coordination mechanisms, technology deployment, or institutional change. They therefore function as interpretive propositions that generate hypotheses requiring validation through implementation studies and institutional analysis beyond textual data (see Section 4.4).

4.2 From pilot success to regional implementation

The transition from pilot-scale technical success to broader implementation highlights how EU-funded projects can function as practical laboratories for testing governance–technology integration under real-world constraints. The five EU Interreg BSR projects examined represent over €20.5 million in strategic public investment and collectively show a range of technology–governance configurations (Section 3.2.3; Table 3.2.3). Interpreted as a portfolio, the cases suggest how different integration strategies may be shaped for broader regional adoption while building on demonstrated technical achievements. At the same time, because most projects are ongoing or only recently completed, these configurations should be treated as emerging designs rather than demonstrated long-term outcomes.

Within that boundary, the project evidence suggests three candidate scaling pathways that are inferred from project design emphases and public materials (Section 3.2.3; Section 3.3; Table 3.2.3). ReNutriWater and NURSECOAST-II align with a regulatory-compliance pathway that aims for technical validation within existing frameworks while establishing conditions for later market development. CiNURGI aligns with a policy-integration pathway, where cross-border cooperation agreements and modular processing units form an institutional infrastructure that may reduce barriers for subsequent adopters and support network-based expansion. City Blues aligns with a stakeholder-engagement pathway, where co-creation processes can build social license and municipal commitment that support continued implementation. These are analytic categories that organise observed project strategies; they do not, on their own, demonstrate durable post-project scaling.

Business model innovation appears as an increasingly prominent design concern across the portfolio, with projects experimenting with fee-for-service, cost-offset, and hybrid financing approaches tailored to different technology–governance combinations (Section 3.2.3; Table 3.2.3). For example, CiNURGI’s intermunicipal governance arrangements combined with standardised processing units illustrate how institutional innovation may be structured to support future economic viability through shared services and service-based revenue logics. At the same time, the Tier 3 public texts often remain generic on revenue mechanisms, and the business-model language should therefore be interpreted cautiously as an intention or design orientation rather than as verified financial performance (Section 3.2.3).

Knowledge transfer mechanisms developed through projects nevertheless suggest opportunities for regional scaling, particularly where documentation practices, practitioner networks, and macro-regional platforms create channels for diffusion beyond a single project lifecycle (Sections 3.2.2–3.4). In the Tier 2 corpus, horizontal exchanges and cross-sector dialogue provide a plausible pathway through which successful solutions may inspire adaptation across partner organisations and regional contexts (Section 3.2.2). At the macro-regional level, institutional channels associated with HELCOM, EUSBSR and Interreg BSR provide coordination and documentation infrastructures that can extend learning beyond individual projects, although post-project continuity remains a recurring constraint in the evidence (Section 3.4). Overall, the empirical pattern suggests that EU project funding can create conditions for experimentation, learning, and investment in circular water innovation, while also revealing the governance and temporal bottlenecks that shape whether pilots translate into durable regional implementation (Sections 3.3–3.4).

4.3 Framework applications and transferability

The governance–technology integration lens, structured around three coordination mechanisms (structural, knowledge, and temporal coordination), addresses widely observed challenges in environmental governance while remaining adaptable to different regional characteristics. Its potential transferability stems from its focus on coordination processes rather than fixed institutional arrangements, enabling application across varied governance architectures. In particular, the temporal coordination component foregrounds the frequent misalignment between policy cycles, funding horizons, and technology development–adoption timelines, an issue that emerges in the empirical findings as a key constraint on scaling beyond pilots (Sections 3.3–3.4). The knowledge coordination component similarly aligns with the observed role of intermediary arenas and documentation practices in translating policy intent into implementation-relevant framing and practice-oriented exchange (Section 3.2.2). Structural coordination is likewise implicated in Tier 1 and Tier 3 patterns where integration depends on cross-framework linkage and on boundary organisations that can broker translation across levels and sectors (Section 3.2.1; Section 3.4).

Several factors nevertheless limit transferability. The BSR benefits from high institutional density through HELCOM, mature EU integration mechanisms, and established traditions of regional cooperation that may not exist elsewhere. The effectiveness of the lens may therefore depend on minimum levels of institutional capacity and cooperative culture, which may need to be developed before similar coordination mechanisms can function effectively. Additionally, the temporal scope captured here (2024–2025) provides sensitivity to short-term framing dynamics but may not capture longer institutional cycles that shape technology scaling trajectories in other settings (see Section 4.4).

The lens may extend beyond water management to other environmental technology domains that require multi-scale coordination, but such extensions should be treated as hypotheses requiring sectoral and regional validation. Energy transition initiatives, biodiversity conservation programmes, and climate adaptation technologies face analogous coordination problems in aligning technical interventions with governance arrangements across jurisdictions. In these contexts, structural coordination may be particularly salient for cross-border infrastructure deployment, knowledge coordination for learning and translation across practitioner communities, and temporal coordination for aligning policy, funding, and adoption timelines. Taken together, the evidence suggests that the lens captures coordination challenges that recur across environmental governance domains, while remaining sensitive to sectoral and contextual variation.

4.4 Study limitations and future research

This study’s analytical scope, while providing insights into governance–technology integration, presents several limitations that indicate directions for future research. The geographic focus on the BSR, though justified by its advanced policy and institutional engagement in water management, limits claims about framework universality. The region’s institutional characteristics, including mature EU integration mechanisms, long-established HELCOM cooperation traditions, and relatively homogeneous democratic governance systems, may not be representative of other contexts. Comparative analysis across regions with different institutional densities, governance cultures, and economic development levels would strengthen generalisability.

The temporal scope presents additional constraints. While the 2024–2025 discourse analysis captures short-term evolution in practitioner framing, this timeframe may be insufficient to assess longer-term institutional change or technology adoption cycles. Environmental governance transformations often require decades to mature, and scaling pathways may exhibit different patterns over extended periods. Longitudinal studies extending beyond project funding cycles would provide more robust evidence of sustainability and institutional embedding.

Methodologically, discourse analysis relies on linguistic indicators that may not fully capture behavioural or institutional change. The relationship between evolving practitioner language and actual coordination improvements requires verification through complementary approaches. Future research could integrate economic impact assessments, stakeholder network analysis, and quantitative measures of technology adoption rates to triangulate discourse findings with material outcomes.

The focus on successful EU-funded projects may introduce selection bias that limits insight into coordination failures or alternative scaling pathways. Comparative work including projects that struggled to achieve integration or pursued different strategies would provide a fuller picture of enabling and constraining factors. Investigation of non-EU funded initiatives or bottom-up scaling approaches could also illuminate coordination mechanisms operating outside formal institutional frameworks.

Future research may also expand the lens to incorporate social and economic dimensions of governance–technology integration. While this study emphasises institutional coordination, scaling success depends on public acceptance, distributional effects, and economic sustainability factors that warrant systematic investigation. Integrating social network analysis could further illuminate how practitioner communities influence diffusion beyond formal policy channels.

Finally, applying the lens to other environmental sectors, climate adaptation, renewable energy systems, biodiversity conservation, offers a promising research direction. Sectoral comparative analysis would refine understanding of which coordination mechanisms are most critical under varying technological and institutional conditions, contributing to more targeted policy design and implementation strategies.

5 Conclusion

This study develops and applies a computational, cross-scale coordination lens to explain why circular water technologies in the Baltic Sea Region often struggle to scale from pilots to sustained regional uptake. Using a three-tier corpus design (Tier 1 policy frameworks, Tier 2 Europe Forum Turku 2024–2025 practitioner discourse, and Tier 3 documentation from five Interreg BSR projects) and a combined AntConc–InfraNodus workflow, the analysis provides quantitative evidence of field maturation from predominantly technical optimisation toward more explicit governance–technology integration. In Tier 2, practitioner framing shifts from purely technical emphasis (peak 178 mentions in 2024) toward strategic integration discourse (79 mentions in 2025), alongside the first observed co-occurrence of technical and security language, indicating an emergent tendency to articulate implementation challenges and solutions within broader strategic frames rather than as standalone engineering issues.

Across tiers, three coordination mechanisms recur and operate dynamically rather than hierarchically. Structural coordination is expressed through cross-border platforms and regulatory alignment that create enabling conditions while maintaining regional differentiation. Knowledge coordination is visible in translation and information-sharing processes that connect technical expertise to institutional practice, including bridging vocabularies around procurement, permitting, business models, and monitoring that link strategic “why” framing to operational “how” steps. Temporal coordination concerns the alignment (or misalignment) between policy cycles, funding periods, and technology development timelines, shaping whether learning can be carried forward beyond pilot phases. When operationalised jointly as a cross-scale lens, these mechanisms explain where scaling repeatedly stalls. Three recurrent pinch points are visible across the corpus: (i) policy–practice gaps, where high-level targets and platforms are insufficiently translated into locally actionable rules, permitting routines, and procurement templates; (ii) post-project market formation and operational continuity, where demonstration success is not matched by durable ownership, financing, and maintenance arrangements once programme cycles end; and (iii) temporal misalignment, where regulatory updates, funding horizons, and technology iteration timelines do not synchronise, narrowing the space for adaptation and slowing replication beyond pilots.

These findings matter because the BSR combines dense multi-level governance architecture with substantial EU investment and strong technical innovation capacity yet still exhibits persistent “pilot-to-practice” frictions. By tracing how coordination mechanisms appear (or fail to appear) across policy intent, practitioner translation, and project implementation, the study shifts the scaling problem from a narrow question of technical readiness to a diagnosis of coordination conditions that determine whether investments translate into durable regional impact. For EU programming, the implication is that improving investment effectiveness depends not only on funding more pilots, but on strengthening the cross-scale mechanisms that carry successful configurations into stable compliance pathways, enduring operations, and replicable uptake across jurisdictions.

5.1 Implications and future research

Two practical implications follow. First, project selection and monitoring should explicitly assess post-pilot readiness, including permitting/procurement pathways, standards and monitoring alignment, and credible ownership, financing, and maintenance arrangements, rather than treating these as downstream concerns. Second, to reduce replication friction across jurisdictions, programmes and regional actors should prioritise reusable translation assets (e.g., templates, shared baselines and indicators, and structured feedback loops) that stabilise knowledge coordination and make implementation routines portable without requiring new institutional structures. A key limitation is that the evidence is text-based: discourse and documentation can indicate integration patterns and likely bottlenecks, but cannot by themselves verify institutional change or long-term deployment outcomes. Future research should therefore test transferability through comparative multi-regional analysis across EU cooperation areas, validate durability through longitudinal tracking beyond current project cycles, and quantify value through economic comparisons of governance–technology integration approaches versus technology-only investments to support evidence-based programme design.

Data availability statement

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/Supplementary Material.

Author contributions

CC: Conceptualization, Data curation, Formal Analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Software, Supervision, Validation, Visualization, Writing – original draft, Writing – review and editing. ES: Conceptualization, Data curation, Formal Analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Software, Supervision, Validation, Visualization, Writing – original draft, Writing – review and editing.

Funding

The author(s) declared that financial support was received for this work and/or its publication. This research was conducted within the CiNURGi project, funded from the EU Interreg Baltic Sea Region Programme 2021–2027, grant #C049. CiNURGI is co-funded by the European Regional Development Fund (ERDF) with a total budget of €6.54 million, of which €5.23 million is provided through ERDF support.

Acknowledgements

The authors gratefully acknowledge this financial support.

Conflict of interest

The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Author disclaimer

The views expressed in this publication are solely those of the authors and do not necessarily reflect the official position of the EU Interreg Baltic Sea Region Programme or the European Union.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/feart.2026.1750984/full#supplementary-material

References

Ames, E. (1961). Research, invention, development and innovation. Am. Econ. Rev. 51 (3), 370–381.

Google Scholar

Anthony, L. (2004). AntConc: a learner and classroom friendly, multi-platform corpus analysis toolkit. Proc. IWLeL 2004 An Interact. Workshop Lang. E-Learning, 7–13.

Google Scholar

Anthony, L. (2019). AntConc (version 3.5.8). Tokyo: Waseda University. Available online at: https://www.laurenceanthony.net/software/antconc/.

Google Scholar

Backer, H., Leppänen, J. M., Brusendorff, A. C., Forsius, K., Stankiewicz, M., Mehtonen, J., et al. (2010). HELCOM Baltic sea action plan – a regional programme of measures for the marine environment based on the ecosystem approach. Mar. Pollut. Bull. 60 (5), 642–649. doi:10.1016/J.MARPOLBUL.2009.11.016

PubMed Abstract | CrossRef Full Text | Google Scholar

Balanzó-Guzmán, A., and Ramos-Mejía, M. (2023). Towards epistemic diversity in sustainability transitions: an exploration of hybrid socio-technical systems. Sustain. Sci. 18 (6), 2511–2531. doi:10.1007/S11625-023-01370-9/FIGURES/3

CrossRef Full Text | Google Scholar

Beck, S., Jasanoff, S., Stirling, A., and Polzin, C. (2021). The governance of sociotechnical transformations to sustainability. Curr. Opin. Environ. Sustain. 49, 143–152. doi:10.1016/J.COSUST.2021.04.010

CrossRef Full Text | Google Scholar

Benz, A. (2000). Two types of multi-level governance: intergovernmental relations in German and EU regional policy. Regional Fed. Stud. 10 (3), 21–44. doi:10.1080/13597560008421130/ASSET//CMS/ASSET/4122A704-A7D5-4DC3-BD0C-4CDB3EE02661/13597560008421130.FP.PNG

CrossRef Full Text | Google Scholar

Boström, M., Grönholm, S., Hassler, B., Boström, M., Grönholm, S., and Hassler, B. (2016). The ecosystem approach to management in Baltic sea governance: towards increased reflexivity? MARE Publ. Ser. 10, 149–172. doi:10.1007/978-3-319-27006-7_7

CrossRef Full Text | Google Scholar

Carlisle, K., and Gruby, R. L. (2019). Polycentric systems of governance: a theoretical model for the commons. Policy Stud. J. 47 (4), 927–952. doi:10.1111/PSJ.12212

CrossRef Full Text | Google Scholar

Cataldo, F., Redecker, C., Montanari, E., Galariotis, I., Greidanus, H., and Quetel, C. (2025). Rethinking societal resilience in a time of polycrisis. Available online at: https://publications.jrc.ec.europa.eu/repository/handle/JRC142772.

Google Scholar

Chaffin, B. C., Gosnell, H., and Cosens, B. A. (2014). A decade of adaptive governance scholarship: synthesis and future directions. Ecol. Soc. 19 (3), art56. doi:10.5751/ES-06824-190356

CrossRef Full Text | Google Scholar

Conteh, C., and Harding, B. (2023). Boundary-spanning in public value co-creation through the lens of multilevel governance. Public Manag. Rev. 25 (1), 104–128. doi:10.1080/14719037.2021.1942529

CrossRef Full Text | Google Scholar

Damman, S., Schmuck, A., Oliveira, R., Koop, S., Stef, H. A., Almeida, M. do C., et al. (2023). Towards a water-smart society: progress in linking theory and practice. Util. Policy 85, 101674. doi:10.1016/J.JUP.2023.101674

CrossRef Full Text | Google Scholar

Di Gregorio, M., Fatorelli, L., Paavola, J., Locatelli, B., Pramova, E., Nurrochmat, D. R., et al. (2019). Multi-level governance and power in climate change policy networks. Glob. Environ. Change 54, 64–77. doi:10.1016/J.GLOENVCHA.2018.10.003

CrossRef Full Text | Google Scholar

Ejdemo, T., and Örtqvist, D. (2020). Related variety as a driver of regional innovation and entrepreneurship: a moderated and mediated model with non-linear effects. Res. Policy 49 (7), 104073. doi:10.1016/J.RESPOL.2020.104073

CrossRef Full Text | Google Scholar

El Bilali, H. (2019). The multi-level perspective in research on sustainability transitions in agriculture and food systems: a systematic review. Agric. 2019 9, 74. doi:10.3390/AGRICULTURE9040074

CrossRef Full Text | Google Scholar

EPRS. (n.d.). An EU strategy for the Baltic sea region (EUSBSR). Available online at: https://www.europarl.europa.eu/thinktank/en/document/EPRS_BRI(2022)733703

Google Scholar

European Commission. (2019a). Economic and social committee and the committee of the regions the European green deal.

Google Scholar

European Commission (2019b). EU green deal COM(2019)640 final. Available online at: https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=celex:52019DC0640.

Google Scholar

European Commission (2025). Riding the wave – sustainable water management in the Baltic sea region - environment. Available online at: https://environment.ec.europa.eu/news/riding-wave-sustainable-water-management-baltic-sea-region-2024-09-09_en.

Google Scholar

European Union (2021). Interreg Baltic sea region programme 2021-2027 - mobility and transport. Available online at: https://transport.ec.europa.eu/transport-modes/maritime/ship-financing-portal/interreg-baltic-sea-region-programme-2021-2027_en.

Google Scholar

European Union (2025). Multiannual financial framework | fact sheets on the European Union. Brussels: European Parliament. Available online at: https://www.europarl.europa.eu/factsheets/en/sheet/29/multiannual-financial-framework.

Google Scholar

EUSBSR (2021). Action plan - EUSBSR. Available online at: https://eusbsr.eu/about/action-plan/.

Google Scholar

Ferrari, M. (2020). Reflexive governance for infrastructure resilience and sustainability. Sustainability 12 (23), 10224. doi:10.3390/SU122310224

CrossRef Full Text | Google Scholar

Folke, C., Carpenter, S., Walker, B., Scheffer, M., Elmqvist, T., Gunderson, L., et al. (2004). Regime shifts, resilience, and biodiversity in ecosystem management. Annu. Rev. Ecol. Evol. Syst. 35 (1), 557–581. doi:10.1146/annurev.ecolsys.35.021103.105711

CrossRef Full Text | Google Scholar

Geels, F. W. (2002). Technological transitions as evolutionary reconfiguration processes: a multi-level perspective and a case-study. Res. Policy 31, 1257–1274. doi:10.1016/s0048-7333(02)00062-8

CrossRef Full Text | Google Scholar

Geels, F. W. (2011). The multi-level perspective on sustainability transitions: responses to seven criticisms. Environ. Innovation Soc. Transitions 1 (1), 24–40. doi:10.1016/J.EIST.2011.02.002

CrossRef Full Text | Google Scholar

Godin, B. (2006). The linear model of innovation. Sci. Technol. Hum. Values 31 (6), 639–667. doi:10.1177/0162243906291865

CrossRef Full Text | Google Scholar

Haddad, C. R., Nakić, V., Bergek, A., and Hellsmark, H. (2022). Transformative innovation policy: a systematic review. Environ. Innovation Soc. Transitions 43, 14–40. doi:10.1016/J.EIST.2022.03.002

CrossRef Full Text | Google Scholar

HELCOM. (2016). State of the Baltic sea 2023. Third HELCOM holistic assessment 2016-2021.

Google Scholar

HELCOM (2021). Baltic marine environment protection commission HELCOM Baltic sea regional nutrient recycling strategy 2 Baltic sea regional nutrient recycling strategy. Available online at: www.helcom.fi.

Google Scholar

HELCOM (2025a). Thematic assessment on hazardous submerged objects in the Baltic sea potentially polluting shipwrecks in the Baltic sea 2 thematic assessment on hazardous submerged objects in the Baltic sea (volume 2) potentially polluting shipwrecks in the Baltic sea. Helsinki.

Google Scholar

HELCOM (2025b). “Thematic assessment on hazardous submerged objects in the Baltic sea warfare materials in the Baltic sea 2 thematic assessment on hazardous submerged objects in the Baltic sea (volume 1) warfare materials,” in The Baltic sea.

Google Scholar

Hikkaduwa Liyanage, S. I., Netswera, F., Meyer, J., and Botha, C. (2022). Four pillars of the green university soft infrastructure. Int. J. Knowl. Manag. 18 (1), 1–16. doi:10.4018/IJKM.305225

PubMed Abstract | CrossRef Full Text | Google Scholar

Hooghe, L., and Marks, G. (2003). Unraveling the central state, but how? Types of multi-level governance. Am. Political Sci. Rev. 97 (2), 233–243. doi:10.1017/S0003055403000649

CrossRef Full Text | Google Scholar

Keller, M., Sahakian, M., and Hirt, L. F. (2022). Connecting the multi-level-perspective and social practice approach for sustainable transitions. Environ. Innovation Soc. Transitions 44, 14–28. doi:10.1016/j.eist.2022.05.004

CrossRef Full Text | Google Scholar

Kellner, E., Petrovics, D., and Huitema, D. (2024). Polycentric climate governance: the state, local action, democratic preferences, and power - emerging insights and a research agenda. Glob. Environ. Polit. 24 (3), 24–47. doi:10.1162/GLEP_A_00753

CrossRef Full Text | Google Scholar

Kern, F., and Rogge, K. S. (2018). Harnessing theories of the policy process for analysing the politics of sustainability transitions: a critical survey. Environ. Innovation Soc. Transitions 27, 102–117. doi:10.1016/J.EIST.2017.11.001

CrossRef Full Text | Google Scholar

Köhler, J., Geels, F. W., Kern, F., Markard, J., Onsongo, E., Wieczorek, A., et al. (2019). An agenda for sustainability transitions research: state of the art and future directions. Environ. Innovation Soc. Transitions 31, 1–32. doi:10.1016/J.EIST.2019.01.004

CrossRef Full Text | Google Scholar

Kok, K. P. W., de Hoop, E., Sengers, F., Broerse, J. E. W., Regeer, B. J., and Loeber, A. M. C. (2022). Governing translocal experimentation in multi-sited transition programs: dynamics and challenges. Environ. Innovation Soc. Transitions 43, 393–407. doi:10.1016/J.EIST.2022.05.001

CrossRef Full Text | Google Scholar

Kuhmonen, T., Kuhmonen, I., and Huuskonen, A. (2024). Sustainability-driven regime shifts in complex adaptive systems: the case of animal production and food system. Sustain. Prod. Consum. 52, 469–486. doi:10.1016/J.SPC.2024.11.022

CrossRef Full Text | Google Scholar

Li, Y., Wang, S., and Khaskheli, M. B. (2024). Integrating artificial intelligence into service innovation, business development, and legal compliance: insights from the Hainan free trade port era. Systems 12 (11), 463. doi:10.3390/SYSTEMS12110463

CrossRef Full Text | Google Scholar

Maggetti, M., and Trein, P. (2019). Multilevel governance and problem-solving: towards a dynamic theory of multilevel policy-making? Public Adm. 97 (2), 355–369. doi:10.1111/PADM.12573

CrossRef Full Text | Google Scholar

Markard, J., Raven, R., and Truffer, B. (2012). Sustainability transitions: an emerging field of research and its prospects. Res. Policy 41 (6), 955–967. doi:10.1016/J.RESPOL.2012.02.013

CrossRef Full Text | Google Scholar

Meissner, D., Harms, R., Kratzer, J., and Zhou, J. (2024). The dead end of classical innovation management and unsustainable innovation. Technovation 129, 102916. doi:10.1016/J.TECHNOVATION.2023.102916

CrossRef Full Text | Google Scholar

Newig, J., Jager, N. W., Challies, E., and Kochskämper, E. (2023). Does stakeholder participation improve environmental governance? Evidence from a meta-analysis of 305 case studies. Glob. Environ. Change 82, 102705. doi:10.1016/J.GLOENVCHA.2023.102705

PubMed Abstract | CrossRef Full Text | Google Scholar

Odebode, J. (2023). Embedding adaptive project governance to integrate sustainability, circular economy practices, and innovative material technologies in architectural built environment projects. 17 (3), 1107–1123. doi:10.30574/WJARR.2023.17.3.0511

CrossRef Full Text | Google Scholar

Ostrom, E. (2010). Polycentric systems for coping with collective action and global environmental change. Glob. Environ. Change 20 (4), 550–557. doi:10.1016/J.GLOENVCHA.2010.07.004

CrossRef Full Text | Google Scholar

Paranyushkin, D. (2019). “InfraNodus: generating insight using text network analysis,” in The web conference 2019 - proceedings of the world wide web conference, WWW 2019, 3584–3589. doi:10.1145/3308558.3314123

CrossRef Full Text | Google Scholar

Pressman, J. L., and Wildavsky, A. (1984). Implementation. 3rd Edition. Available online at: https://www.ucpress.edu/books/implementation/paper.

Google Scholar

Roux, D. J., Taplin, M., Smit, I. P. J., Novellie, P., Russell, I., Nel, J. L., et al. (2023). Co-Producing narratives and indicators as catalysts for adaptive governance of a common-pool resource within a protected area. Environ. Manag. 72 (6), 1111–1127. doi:10.1007/S00267-023-01884-Z/FIGURES/7

PubMed Abstract | CrossRef Full Text | Google Scholar

Sajida (2025). Three decades of multilevel governance research: a scientometric and conceptual mapping in the social sciences. Soc. Sci. Humanit. Open 12, 101745. doi:10.1016/J.SSAHO.2025.101745

CrossRef Full Text | Google Scholar

Salvador-Carulla, L., Woods, C., de Miquel, C., and Lukersmith, S. (2024). Adaptation of the technology readiness levels for impact assessment in implementation sciences: the TRL-IS checklist. Heliyon 10 (9), e29930. doi:10.1016/J.HELIYON.2024.E29930

PubMed Abstract | CrossRef Full Text | Google Scholar

Sánchez-Silva, M. (2018). Managing infrastructure systems through changeability. J. Infrastructure Syst. 25 (1), 04018040. doi:10.1061/(ASCE)IS.1943-555X.0000467

CrossRef Full Text | Google Scholar

Smith, A., Voß, J. P., and Grin, J. (2010). Innovation studies and sustainability transitions: the allure of the multi-level perspective and its challenges. Res. Policy 39 (4), 435–448. doi:10.1016/J.RESPOL.2010.01.023

CrossRef Full Text | Google Scholar

Turnheim, B., and Sovacool, B. K. (2020). Exploring the role of failure in socio-technical transitions research. Environ. Innovation Soc. Transitions 37, 267–289. doi:10.1016/J.EIST.2020.09.005

CrossRef Full Text | Google Scholar

Tursunkulova, I., de Castell, S., and Jenson, J. (2024). Sorcerer’s apprentice? Exploring an AI-Driven tool to analyze academic texts. 41–63. doi:10.1007/978-3-031-66462-5_3

CrossRef Full Text | Google Scholar

Valentinov, V., de Oliveira Santos Jhunior, R., and de Araujo Góes, H. A. (2025). Corporate environmental sustainability via stakeholder collaboration: insights from classical institutional economics. J. Bus. Ethics, 1–18. doi:10.1007/S10551-025-06023-8/TABLES/1

CrossRef Full Text | Google Scholar

Van Cauwenbergh, N., Dourojeanni, P. A., van der Zaag, P., Brugnach, M., Dartee, K., Giordano, R., et al. (2022). Beyond TRL – understanding institutional readiness for implementation of nature-based solutions. Environ. Sci. Policy 127, 293–302. doi:10.1016/J.ENVSCI.2021.09.021

CrossRef Full Text | Google Scholar

Wiegant, D., Dewulf, A., and Van Zeben, J. (2024). Alignment mechanisms to effectively govern the sustainable development goals. World Dev. 182, 106721. doi:10.1016/J.WORLDDEV.2024.106721

CrossRef Full Text | Google Scholar

Winter, S. (2013). Implementation perspectives: status and reconsideration. The SAGE Handbook of Public Administration, 265–278. doi:10.4135/9781446200506.N17

CrossRef Full Text | Google Scholar

Yi, H., Huang, C., Chen, T., Xu, X., and Liu, W. (2019). Multilevel environmental governance: vertical and horizontal influences in local policy networks. Sustainability 11 (8), 2390. doi:10.3390/SU11082390

CrossRef Full Text | Google Scholar