ORIGINAL RESEARCH article

Front. Mar. Sci., 21 October 2026

Sec. Marine Affairs and Policy

Volume 13 - 2026 | https://doi.org/10.3389/fmars.2026.1928781

Cross-domain risk governance of transboundary marine policies for blue carbon ecosystem protection: a fuzzy AHP evaluation framework applied to the nuclear wastewater discharge case

  • School of Finance and Trade, Wenzhou Business College, Wenzhou, China

Abstract

Introduction:

Blue carbon ecosystems sequester carbon at rates that dwarf most terrestrial forests yet occupy less than 0.5% of the global ocean surface, making their protection under transboundary marine policy shocks both ecologically imperative and analytically underserved. Unilateral marine policy decisions in one coastal state, whether discharging pollutants, restructuring fishery regimes, or altering spatial zoning, generate externalities that propagate simultaneously across political, economic, industrial, and environmental channels into the blue carbon systems of neighboring states. No existing risk assessment tool captures this cross-domain, transjurisdictional character in a single integrated procedure.

Methods:

This study develops and applies a Cross-Domain Risk Management Procedure (CDRMP) calibrated explicitly to blue carbon governance contexts. Grounded in polycentric governance theory, ecosystem services economics, and a transboundary externality model, the CDRMP framework embeds 36 evaluation criteria constructed via Nominal Group Technique and weighted via Fuzzy Analytic Hierarchy Process (FAHP) across 114 respondents from government, industry, and public constituencies. The framework is applied to Japan's Fukushima Daiichi nuclear wastewater discharge into the Pacific Ocean as an illustrative transboundary case.

Results:

Three principal findings emerge. First, geopolitical contestation (weight 0.197) nearly matches direct environmental contamination (0.206) in CDRMP severity, indicating that political governance instability and ecological contamination are mutually reinforcing rather than sequential risk events. Second, five of the seven highest-ranked criteria are blue-carbon-relevant, three of them blue-carbon-specific additions introduced by the ecological extension of the standard environmental domain, suggesting that frameworks without such calibration may systematically underdetect the most severe governance risks. Third, policy evaluation under optimistic, pessimistic, and equal-probabilities decision criteria shows no single policy dominating across all rules and analytical levels: marine environmental policy is preferred under two of three criteria at the dimension level and attains the single lowest aggregate deviation observed (0.856 under equal-probabilities), marine spatial policy is preferred under the pessimistic criterion at both levels, and marine economic policy is preferred under the optimistic criterion at the criterion level.

Discussion:

The contingent pattern of policy preference is consistent with, though not a direct test of, complementary rather than substitutable governance roles across environmental, spatial, and economic instruments. The study contributes the first quantitatively operational CDRMP framework to the literatures on polycentric governance, blue carbon ecosystem services, and multi-criteria marine policy evaluation, providing coastal states with a replicable assessment procedure for transboundary marine policy risks.

Graphical Abstract

Highlights

  • A novel 36-criterion Cross-Domain Risk Management Procedure (CDRMP) integrates political, economic, industrial, and environmental risk channels into the first Fuzzy AHP framework explicitly calibrated to blue carbon ecosystem governance.

  • Five of the seven highest-ranked CDRMP criteria are blue-carbon-relevant; of these, three are blue-carbon-specific additions introduced by the ecological extension. Under an illustrative counterfactual in which a generic framework retained only the two of these five criteria with direct standard-environmental analogues, three of five (60%) of the highest-severity blue-carbon-relevant criteria would be unrepresented; this is a hypothetical sensitivity illustration, not an empirical estimate of improved detection capability.

  • Geopolitical contestation (weight 0.197) rivals direct environmental contamination (0.206) in CDRMP severity, indicating that political governance instability and ecological contamination are mutually reinforcing rather than sequential risk events.

  • Marine environmental policy achieves the single lowest aggregate deviation observed (0.856, equal-probabilities); marine spatial policy is preferred instead under the pessimistic criterion at both analytical levels, and marine economic policy under the optimistic criterion at the criterion level.

  • Expert FAHP weights may implicitly reflect ecological irreversibility: respondents assign highest severity to sediment contamination and carbon sink degradation without explicit instruction, a pattern consistent with, though not conclusive proof of, latent preferences for blue carbon protection that standard market mechanisms obscure.

1 Introduction

The ocean’s climatic significance has been radically reappraised over the past decade, and nowhere more consequentially than in the recognition that coastal vegetated habitats, occupying less than 0.5% of the global ocean surface, sequester carbon at rates that dwarf those of most terrestrial forests (Macreadie et al., 2026). Tidal flats alone fix approximately 8.6 Tg C per year and hold 1.2 Pg C within the uppermost meter of sediment (Wang et al., 2026b); mangrove soils across Africa yield median stocks of 323 Mg C per hectare (Adams et al., 2026). Accounting frameworks that expand to incorporate phytoplankton burial, macroalgal beds, and shelf sediment dynamics continue to push global blue carbon estimates upward (Macreadie et al., 2026; Uddin et al., 2026), cementing these systems as irreplaceable components of any credible net-zero pathway. Both the economic valuation (Luisetti et al., 2015) and the financial mobilization (Pendleton et al., 2014) of coastal blue carbon have been advanced as necessary conditions for such protection.

The trajectory of loss, however, accelerates in the opposite direction. Between 2002 and 2019, salt marsh soil organic carbon declined by a net 0.52 million tons globally, with losses concentrated in mature, carbon-dense stands that restoration cannot compensate on policy-relevant timescales (Zheng et al., 2026). Climate warming alone suppresses microbial carbon fixation in subsoils by more than one third, an endogenous destabilization independent of land-use pressure (Xiao et al., 2026), while coastal erosion drives net declines of 55 to 77 percent in total ecosystem carbon stocks of tropical island mangrove over multi-decadal periods (Rodrigues et al., 2026). The compound verdict is that blue carbon ecosystems are losing ground faster than governance systems can protect them (Zhang et al., 2026; Macreadie et al., 2026).

Why governance has failed is as consequential a question as the failure itself. Blue carbon ecosystems are structurally indifferent to jurisdictional boundaries: mangrove estuaries straddle coastlines, seagrass meadows extend across exclusive economic zones, and tidal sediment carbon is continuously exposed to disturbances originating beyond any single sovereign reach. When a coastal state expands marine industry, alters spatial zoning, discharges pollutants, or redistributes fishery pressure, the resulting externalities propagate to neighboring blue carbon systems through oceanographic connectivity and shared biogeochemical cycles (Chen et al., 2025; Karli et al., 2026). This is the very condition under which decentralized governance fails systematically: nationally rational policy choices produce suboptimal ecological outcomes.

Governance fragmentation amplifies the problem. Despite the overarching framework of UNCLOS, marine policy operates through sectoral treaties, regional bodies, and national agencies with inconsistent mandates (Zhang et al., 2026; Khaskheli et al., 2025). Comparative analysis across Australia, the EU, China, and Kenya shows that international law fragmentation simultaneously creates governance voids and stimulates incompatible bottom-up regulatory innovation (Zhang et al., 2026). Baltic Sea evidence confirms that neighboring states’ climate policy choices measurably affect coastal ecosystem stability through carbon emission and eutrophication pathways (Chen et al., 2025). Blue nature-based solutions remain largely absent from formal EU marine policy despite demonstrated effectiveness (Frehen et al., 2025); tenure ambiguities in blue carbon project design expose coastal communities to dispossession (Lawless et al., 2026). What remains structurally absent, across all these contexts, is any cross-domain risk assessment tool capable of translating one country’s transboundary marine policy decisions into systematically evaluated governance risks for the blue carbon systems of another (Jin et al., 2026; Karli et al., 2026).

Existing scholarship has made progress in individual risk domains: political disputes over maritime jurisdiction, economic externalities in fisheries and blue economy sectors (Ma et al., 2025; Santos, 2025); the governance challenges of translating blue carbon into climate commitments remain acute across jurisdictions (Purushotham and Thompson, 2026), and the carbon-management potential of blue economic resources in regional blocs such as ASEAN is only beginning to be quantified (Afzal et al., 2026), industrial risks to coastal habitats (Shen et al., 2025; Du et al., 2026), and environmental pollution impacts (Liu et al., 2026). Taken together, however, they reveal the critical gap that Thomas (2014) likewise foregrounds: no integrated framework captures how a single marine policy decision simultaneously generates cascading cross-domain risks across political, economic, industrial, and environmental dimensions for neighboring blue carbon ecosystems.

Japan’s decision to discharge treated nuclear wastewater from Fukushima Daiichi into the Pacific Ocean exemplifies the governance void with illustrative clarity. The decision generates geopolitical disputes over international law compliance, disrupts regional fishery markets and supply chains, threatens aquaculture and maritime tourism, and poses contamination risks to the mangrove and seagrass systems of East Asian coastal states. No existing assessment tool can evaluate or rank policy response options across these interconnected dimensions, nor identify which instrument type most effectively addresses the full CDRMP spectrum. This analytical absence motivates the present study.

This paper develops and applies a Cross-Domain Risk Management Procedure (CDRMP) designed specifically for blue carbon governance contexts. Grounded in polycentric governance theory, ecosystem services economics, and the theory of transboundary externalities, the study pursues three objectives: to construct a CDRMP evaluation framework integrating political, economic, industrial, and environmental risk dimensions with explicit blue carbon calibration; to apply it to the nuclear wastewater case, quantifying CDRMP severity and comparing three marine policy responses under optimistic, pessimistic, and equal-probability decision criteria; and to identify which policy strategy most effectively aligns institutional capacity with the ecological characteristics and transboundary exposure of blue carbon habitats.

Three contributions follow. Theoretically, the study integrates polycentric governance theory with blue carbon ecosystem services economics to explain the systemic generation of cross-domain risks, offering a mechanistic account that domain-specific literatures have not provided. Methodologically, this study advances the application of Fuzzy AHP in marine governance by incorporating blue-carbon-specific risk indicators, enabling more ecologically grounded policy evaluation than existing frameworks support. Practically, it provides coastal states with a replicable assessment procedure for transboundary marine policy risks that will intensify in frequency and ecological consequence.

2 Theoretical framework: polycentric governance, blue carbon ecosystem services, and the conceptualization of cross-domain risk

The theoretical architecture rests on three interlocking pillars: polycentric governance theory, blue carbon ecosystem services economics, and a transboundary externality model linking unilateral marine policy decisions to blue carbon outcomes across jurisdictional boundaries. Each pillar resolves a distinct explanatory problem. Polycentric governance theory identifies why institutional fragmentation generates cross-domain risk as a structural rather than contingent outcome. Ecosystem services economics establishes the value metrics against which that risk must be calibrated and explains why those values are systematically underweighted in national policy calculi. The transboundary externality model specifies the causal channels through which policy decisions in one state impose concentrated costs on the blue carbon systems of another. Figure 1 presents the integrated theoretical framework, depicting the layered causal pathways through which unilateral marine policies propagate across state boundaries to degrade blue carbon ecosystem services. Table 1 maps each pillar to the governance failure it diagnoses and the scholarship that grounds it empirically; read together, the table and figure establish that the CDRMP framework responds to a well-documented class of governance failure that existing marine policy scholarship has diagnosed but not equipped with appropriate analytical tools.

Figure 1

Table 1

Theoretical pillarCore propositionGovernance failure addressedKey references
Polycentric governanceShared marine ecological goods require coordinated multi-scale decision centers. When authority is fragmented, individually rational choices aggregate into collectively harmful outcomes for transboundary ecosystems.No single coastal state internalizes the cross-border ecological costs its marine policy decisions impose on neighboring systems, creating a structural divergence between national incentives and collective conservation optima.Ostrom (1990, 2009); Zhang et al. (2026); Karli et al. (2026); Khaskheli et al. (2025)
Ecosystem services economicsBlue carbon systems provide provisioning, regulating, supporting, and cultural services whose economic value is systematically underweighted in national policy owing to temporal irreversibility and public-good characteristics.Policy actors discount the full cost of blue carbon degradation, producing a structural bias toward extractive over protective marine policies and leaving carbon stocks exposed to cross-border policy shocks.McCook et al. (2025); Macreadie et al. (2026); Schmitz et al. (2026); Zheng et al. (2026); Wang et al. (2026b)
Transboundary externality modelUnilateral marine policy decisions propagate negative externalities simultaneously across political, economic, industrial, and environmental channels, concentrating costs in receiving states while the originating state retains all policy benefits.No integrated risk assessment tool captures all four cross-domain channels in a single framework calibrated to the ecological characteristics of blue carbon systems, leaving receiving-state policy-makers without instruments for comparative response evaluation.Chen et al. (2025); Jin et al. (2026); Ma et al. (2025); Lawless et al. (2026); Shen et al. (2025)
Marine industrial risk extensionCross-domain industrial risks emerge when neighboring states’ marine energy and industrial policies restructure supply chains, alter resource competition, or introduce hazardous substances that degrade the physical substrates of blue carbon sequestration.Industrial policy analysis in marine governance neglects downstream effects on ecosystem carbon stocks, generating blind spots in environmental impact assessment and cross-border risk management that conventional frameworks do not address.Du et al. (2026); Shen et al. (2025); Ma et al. (2025); Adams et al. (2026)

Theoretical foundations of the CDRMP framework.

2.1 Polycentric governance theory and the structural preconditions for cross-domain risk

Polycentric governance theory holds that effective stewardship of shared ecological goods requires multiple, partially overlapping decision-making centers operating at different scales (Ostrom, 1990; 2009). In marine governance, the framework resolves a structural paradox that single-level analysis cannot accommodate: the proliferation of governance institutions at international, national, and local levels can, without coordination mechanisms, produce outcomes worse than any single institution acting alone. When treaty bodies, regional commissions, national agencies, and local coastal authorities pursue individually coherent objectives without aligned incentive structures, their decisions aggregate into collectively damaging outcomes for the shared ecological systems they nominally protect (Zhang et al., 2026; Khaskheli et al., 2025).

The empirical foundations span multiple national contexts. Longitudinal analysis of China’s marine economic policy from 2003 to 2023 documents a persistent structural imbalance in which economic instruments crowd out environmental ones, eroding conditions for sustained blue carbon sequestration (Shen et al., 2025). Comparative research on marine governance in Turkey identifies six structural pathologies that transcend any single country: the implementation gap between legal frameworks and enforcement, crisis-driven reactive governance, institutional fragmentation across agencies, external funding dependence, limited marine literacy, and contested blue economy discourse (Karli et al., 2026). Khaskheli et al. (2025) synthesize these patterns as “governance fragmentation,” showing that socioeconomic disparities and inconsistent legal implementation jointly undermine coordinated blue carbon management at scale.

The mechanism through which fragmentation generates cross-domain risk is the externalization of ecological costs across jurisdictional and sectoral boundaries. When an originating state’s marine policy generates pollutants or habitat-altering infrastructure that extends into neighboring exclusive economic zones, it bears none of the ecological costs while the receiving state absorbs all of them. This asymmetric externality structure produces a systematic divergence between national policy incentives and the collective optimum for transboundary blue carbon conservation. Resolving that divergence requires integrated assessment models capable of evaluating ocean use benefits and risks across long temporal horizons and broad geographic regions (Jin et al., 2026), which is the function the CDRMP framework is designed expressly to serve.

2.2 Blue carbon ecosystem services economics: from biophysical stock to policy-relevant value

The second pillar situates blue carbon ecosystems within the ecosystem services framework, translating biophysical properties into economic terms that governance actors can operationalize. Blue carbon ecosystems deliver services across all four Millennium Ecosystem Assessment categories: provisioning (fisheries habitat, coastal food security), regulating (carbon sequestration, storm attenuation, water quality), supporting (biodiversity, nutrient cycling), and cultural (coastal community identity, Indigenous marine heritage). The economic magnitude of these services is substantial and increasingly well-quantified, making them viable anchors for cross-domain risk calibration (McCook et al., 2025; Macreadie et al., 2026).

Carbon sequestration is the most strategically significant, and its sensitivity to external policy shocks exceeds what most frameworks recognize. Seagrass systems in the Yellow Sea and Bohai Sea store 7.18 × 107 Mg C in high-suitability habitats; high-emission pathways exacerbate fragmentation and undermine carbon sink stability across decadal timescales (Wang et al., 2026a). Critically, human activity intensity accounts for 24% of spatial variation in seagrass distribution, exceeding the contribution of temperature-salinity parameters combined (Wang et al., 2026a). The implication is direct: policy choices in neighboring jurisdictions constitute quantifiable threats to blue carbon stocks, not merely diffuse background pressures. Indirect ecological pathways amplify this sensitivity further: conserving megafaunal grazers such as dugongs amplifies sediment carbon stocks by a factor of 2.63 through enhanced seagrass productivity and accelerated nutrient recycling (Schmitz et al., 2026), meaning cross-domain policy risks that disrupt trophic networks carry carbon accounting consequences that standard habitat loss models systematically underestimate. Protected areas have been shown to reduce blue carbon emissions where enforcement is credible (Miteva et al., 2015), and credible MRV systems are increasingly identified as the backbone of such governance (Onrizal et al., 2026). Coastal blue carbon is additionally exposed to acute climate extremes such as typhoons, whose disruption to coastal systems is increasingly quantified with fuzzy resilience-assessment methods (Wan et al., 2024).

The ecosystem services framework also clarifies why transboundary policy spillovers are uniquely destructive for blue carbon systems. Unlike fish populations, which redistribute spatially in response to changed conditions, blue carbon stocks are embedded in sedimentary processes that accumulate across decades and are effectively irreversible once disturbed. The net global loss of 0.52 million tons of salt marsh SOC concentrated in mature, carbon-dense habitats (Zheng et al., 2026) confirms that governance arrangements discounting long-term sedimentary carbon values will systematically bias national policy toward extraction over conservation, even when aggregate welfare calculations favor the reverse (McCook et al., 2025). Integrating ecosystem services values into marine policy risk assessment is therefore not methodological refinement but a prerequisite for governance frameworks that accurately reflect the true cost of cross-domain externalities.

2.3 Conceptualizing cross-domain risk: a transboundary externality model for blue carbon governance

This study conceptualizes Cross-Domain Risk from Marine Policies (CDRMP) as the set of political, economic, industrial, and environmental risks that a coastal state’s blue carbon ecosystem services absorb as a consequence of marine policy decisions originating in neighboring or distant states. Three advances distinguish this conceptualization from prior risk frameworks.

The first is its transjurisdictional character. CDRMP arises not from domestic policy failures in isolation but from the interaction of multiple national governance systems operating on ecologically shared systems. Zhang et al.’s (2026) transnational environmental law analysis establishes the foundational legal logic: blue carbon governance requires “a coordinated system shaped by both state and non-state actors across different jurisdictions.” Fragmentation simultaneously creates governance voids and stimulates incompatible bottom-up innovation, producing a regulatory patchwork that undermines transnational carbon accounting (Zhang et al., 2026). The CDRMP framework operationalizes this legal diagnosis as a risk assessment problem, asking which configurations of cross-border policy interaction generate the greatest risks and across which ecological and institutional dimensions those risks are most concentrated.

The second advance is CDRMP’s explicitly multi-domain character. Policy spillovers propagate simultaneously across political, economic, industrial, and environmental channels. Baltic Sea evidence shows that transnational climate policy affects coastal ecosystem stability through compound carbon emission and eutrophication pathways that no single-domain assessment captures (Chen et al., 2025). Marine ecological civilization zone analysis confirms that even domestically oriented environmental policies generate non-linear cross-sectoral trade-offs demanding multi-criteria rather than domain-specific instruments (Ma et al., 2025).

The third advance is CDRMP’s grounding in the distinctive ecological characteristics of blue carbon systems. These ecosystems occupy a structurally anomalous governance position: their primary climate service constitutes a global public good, while the policy decisions capable of destroying it are taken by individual states acting on national interests that do not internalize its global value. Lawless et al.’s (2026) analysis of tenure rights across 122 blue carbon guidance documents illustrates this divergence: even where international normative frameworks exist, national interpretation leaves rightsholders and ecosystems systematically exposed to expropriation risks that cross-border risk assessment frameworks have not previously addressed.

Figure 1 and Table 1 depict and summarize how the three conceptual advances converge into an analytical architecture in which polycentric governance fragmentation, ecosystem service value losses, and transboundary externality transmission are represented as causally connected layers.

2.4 Theoretical synthesis: from conceptual architecture to evaluation design

The three pillars converge on a single practical implication: risk assessment tools confined to a single domain, jurisdiction, or timescale are structurally incapable of capturing the governance problem blue carbon ecosystems face. Polycentric governance theory establishes that institutional fragmentation is the generative root condition of CDRMP. Ecosystem services economics establishes that the values at stake are large, temporally irreversible, and systematically discounted in national policy calculi. The transboundary externality model specifies the causal channels and supplies the logic for why risk severity must be assessed simultaneously across all four domains.

This synthesis generates three design requirements for the CDRMP evaluation framework developed in Section 3. The framework must be multi-dimensional, integrating risks across all four domains simultaneously. It must be ecologically calibrated, incorporating indicators that reflect the irreversibility, trophic sensitivity, and transboundary exposure of blue carbon stocks. And it must be policy-actionable, enabling receiving states to compare the relative performance of spatial, economic, and environmental governance responses against the full CDRMP risk spectrum. The Nominal Group Technique and Fuzzy Analytic Hierarchy Process are adopted as the methodological instruments for meeting these requirements.

3 Methodology: NGT, FAHP, and the blue carbon-calibrated CDRMP evaluation framework

The analytical design rests on the purposeful integration of two instruments whose strengths are complementary and whose limitations are mutually compensating. NGT provides the qualitative architecture for constructing a theoretically defensible, ecologically calibrated indicator system. FAHP furnishes the quantitative mechanism for weighting those indicators in a manner that honestly represents the irreducible imprecision of expert judgment on transboundary ecological risk. NGT without FAHP yields a well-grounded indicator list with no principled basis for cross-domain severity comparison. FAHP without NGT produces numerically precise weights for indicators whose theoretical foundations and ecological calibration have not been independently validated. Their integration operationalizes all three design requirements simultaneously: multi-dimensionality, ecological calibration, and policy-actionability. The four-step procedure is summarized in Figure 2; the complete indicator hierarchy appears in Appendix Table 1.

Figure 2

3.1 Step one: literature-grounded initial indicator pool

A Web of Science Core Collection search for Cross-Domain Risk from Marine Policies returned no prior publications, confirming the CDRMP framework’s conceptual originality. A parallel China National Knowledge Infrastructure search spanning marine policy research from 1980 to February 2023 returned 698 articles distributed across four thematic domains. This search concluded prior to the March 2023 survey distribution described in Section 3.3, ensuring that the indicator pool submitted to the NGT procedure reflects a fixed literature base: maritime politics, maritime economics, marine industry, and marine culture and environment. This four-domain structure maps directly onto the four principal channels through which transboundary marine policy externalities propagate, as established in Section 2, and corresponds to the categories most empirically relevant to blue carbon ecosystem vulnerability (Jin et al., 2026; Chen et al., 2025).

The blue carbon dimension was incorporated at this initial stage through deliberate extension of the standard marine culture and environment domain. Rather than relying on generic coastal environmental proxies, the extended indicator pool included measures specific to the ecosystem services identified in Section 2.2: carbon sequestration capacity, blue carbon habitat integrity, coastal erosion and disaster risk, and marine ecosystem pollution affecting sedimentary carbon dynamics. This extension responds directly to evidence that blue carbon stocks are governed by ecological processes, including temporal irreversibility and trophic amplification, that generic environmental indicators systematically fail to capture (Zheng et al., 2026; Schmitz et al., 2026). The extension is consistent with fuzzy comprehensive evaluation practice applied to ecological carrying capacity assessment (Wu and Hu, 2020).

3.2 Step two: NGT-facilitated indicator refinement

The Nominal Group Technique is a structured consultation procedure (Delbecq and Van de Ven, 1971) designed to elicit and consolidate expert judgment while controlling for conformity pressures and authority differentials that distort unstructured deliberation. Its deployment tests the literature-derived indicator pool against domain knowledge of blue carbon governance and transboundary marine risk and establishes the final evaluation dimensions and criteria through a structured, four-stage consensus process; the following paragraph reports which elements of that process were and were not retained in a separately tabulable form.

The NGT proceeded through four stages calibrated to the CDRMP indicator construction problem. In the individual generation stage, a multi-disciplinary expert panel spanning marine policy, environmental economics, blue carbon ecology, and risk governance was asked in writing to identify which dimensions and criteria most comprehensively capture the cross-domain risks that foreign marine policies impose on a coastal state’s blue carbon ecosystem services. The written format prevents premature convergence around authoritative voices, ensuring the full disciplinary range enters the record before social filtering occurs. Responses were collated without discussion; structured deliberation then clarified the meaning, scope, and ecological boundary conditions of each candidate indicator, with sustained attention to whether environmental and cultural criteria reflected the distinctive dynamics of blue carbon systems rather than generic coastal hazards. Anonymous voting and a consensus threshold rule identified the final evaluation model. The NGT yielded four CDRMP risk dimensions and 36 evaluation criteria: Maritime Politics (MP0), Maritime Economics (ME0), Marine Industry (MI0), and Marine Culture and Environment (MC0), each encompassing three sub-dimensions, each containing three criteria, as presented in full in Appendix Table 1.

The NGT protocol summarized above followed the four-stage structure standard in the method (individual generation, structured deliberation, anonymous voting, and consensus-based closure). Further procedural granularity, the exact consensus threshold applied at the voting stage and the disposition of each individual candidate indicator considered and not retained, was not logged in a form suitable for separate tabulation and is not reconstructed retrospectively here; respondent-level identifying detail such as country of residence and specific institutional affiliation is likewise not reported, protecting the anonymity of individual panel members and survey respondents in a politically sensitive transboundary governance context. The inclusion of general-public respondents alongside government and industry experts in the subsequent FAHP stage follows directly from the polycentric governance rationale developed in Section 2.1: national policy incentives are jointly shaped by administrative, commercial, and civic actors, and a stakeholder pool restricted to credentialed experts would reproduce the elite-centered blind spot that itself contributes to the systematic underweighting of blue carbon ecosystem services in national policy (McCook et al., 2025). Public respondents were not an unfiltered general-population sample but were restricted to individuals with documented coastal engagement, such as residence in a coastal community, occupational dependence on coastal resources, or active participation in coastal environmental monitoring, so that their judgments reflect informed lay experience. Their responses were aggregated separately by stakeholder group before the cross-group geometric mean was taken, preserving rather than diluting the distinctiveness of each perspective.

3.3 Step three: FAHP weighting procedure

Classical AHP’s assumption that experts can assign crisp numerical values to pairwise comparisons across complex, politically contested, and ecologically uncertain domains is rarely warranted in practice. When violated, it generates inconsistent matrices whose weights misrepresent underlying preference structures (Buckley, 1985; Saaty, 1980, 1988). FAHP resolves this by replacing crisp comparisons with triangular fuzzy numbers (TFNs), now a standard instrument in environmental and infrastructure risk assessment (Kubler et al., 2016), whose interval width explicitly encodes judgment imprecision: a narrow interval signals confidence; a wide interval signals genuine uncertainty. This encoding is a theoretically principled response to the epistemic realities of multi-domain risk assessment, validated across marine governance contexts from port service evaluation to offshore infrastructure selection.

A structured questionnaire employing a nine-category linguistic scale encoded as TFNs within [1, 9] was administered to three stakeholder groups: government officials (n = 39), marine industry operators (n = 38), and members of the general public with documented coastal engagement (n = 37), totaling 120 questionnaires distributed in March 2023. The three-group design is theoretically required by the polycentric governance framework: CDRMP is a governance failure arising from incentive misalignment across institutional actors with different objective functions. Aggregating within rather than across groups would embed one stakeholder category’s preference ordering into the assessment, reproducing at the methodological level the governance bias the CDRMP framework is designed to diagnose. The effective response rate of 95% produced 114 usable respondents across 702 pairwise comparison matrices. The linguistic scale is reproduced in Table 2.

Table 2

Linguistic judgmentTriangular fuzzy number (l, m, u)Reciprocal TFN
Equally important(1, 1, 1)(1, 1, 1)
Weakly more important(1, 2, 3)(1/3, 1/2, 1)
Moderately more important(2, 3, 4)(1/4, 1/3, 1/2)
Moderately to strongly more important(3, 4, 5)(1/5, 1/4, 1/3)
Strongly more important(4, 5, 6)(1/6, 1/5, 1/4)
Strongly to very strongly more important(5, 6, 7)(1/7, 1/6, 1/5)
Very strongly more important(6, 7, 8)(1/8, 1/7, 1/6)
Very strongly to absolutely more important(7, 8, 9)(1/9, 1/8, 1/7)
Absolutely more important(7, 9, 9)(1/9, 1/9, 1/7)

Linguistic fuzzy measurement scale for FAHP pairwise comparisons.

Source: Adapted from Buckley (1985). TFN = Triangular Fuzzy Number.

For the CDRMP severity assessment, each respondent was in principle asked to complete pairwise comparisons at the dimension level (one 4×4 matrix), the sub-dimension level (four 3×3 matrices, one nested within each dimension), and the criterion level (twelve 3×3 matrices, one nested within each sub-dimension), and an equivalent matrix set for each of the three policy capacity questionnaires described in Section 3.4. Not every respondent completed every matrix set in full, and the per-respondent completion record was not retained in a form that supports an exact reconciliation between the reported respondent total and the reported matrix total; this data-reporting limitation is acknowledged explicitly in the Limitations subsection, and the underlying respondent-by-matrix completion log is available from the corresponding author on request.

Let denote the TFN assigned by respondent k to the comparison of criterion l against criterion j. The aggregate fuzzy comparison value is the geometric mean across respondents:

The geometric mean of criterion i’s fuzzy comparison values across all criteria is:

The fuzzy weight vector is obtained by normalization:

yielding for each of the 36 CDRMP criteria.

Local and Global Weights. The procedure above is applied independently within each comparison set in the hierarchy, producing a local defuzzified weight for each of the four dimensions relative to one another, a local weight for each sub-dimension relative to the other sub-dimensions within its dimension, and a local weight for each criterion relative to the other criteria within its sub-dimension. The global weight of a given criterion is the product of its own local weight and the local weights of its parent sub-dimension and parent dimension along the same branch of the hierarchy; global weights across all 36 criteria are then renormalized to sum to one. The dimension-level weights reported in Section 4.2 are local dimension weights; the criterion-level weights reported in Section 4.3 and Table 3 are renormalized global weights. The complete local and global weight vectors for every level of the hierarchy are available from the corresponding author on request; the main text reports the subset directly relevant to the dimension- and criterion-level findings discussed in Sections 4 and 5.

Table 3

RankCriterionDomainWeightBlue carbon implication
1Marine environmental pollutionMarine Culture and Environment0.206Radionuclide contamination of blue carbon sediments; disruption of microbial carbon fixation already vulnerable to thermal stress
2Global maritime strategic disputesMaritime Politics0.197Sustained contestation over coastal state rights to precautionary blue carbon protection in Indo-Pacific governance
3Ecosystem pollutionMarine Culture and Environment0.128Trophic cascade disruption suppressing the megafaunal amplification of sediment carbon stocks
4Inter-country governance organization disputesMaritime Politics0.123Institutional fragmentation preventing multilateral coordination of transboundary blue carbon precaution
5Coastal erosion and disasterMarine Culture and Environment0.081Permanent physical removal of blue carbon stock through contamination-accelerated coastal erosion
6Biodiversity loss0.078Loss of species sustaining sedimentary carbon accumulation in mangrove and seagrass systems
7Carbon sink degradation0.068Direct reduction of carbon sequestration capacity in blue carbon habitats proximate to the discharge zone

Seven highest-ranked CDRMP criteria under the nuclear wastewater discharge case.

Normalization is applied at each hierarchical level after defuzzification: within every comparison set the defuzzified local weights are divided by their sum so that the four dimensions, the three sub-dimensions within each dimension, and the three criteria within each sub-dimension each sum to one; the global criterion weights are then the product of the branch-local weights and are renormalized across all 36 criteria to sum to one. Combination and entropy-based schemes for stabilizing composite weights are well established (Chen, 2020; Wu and Hu, 2020) and inform the renormalization applied here. The dimension-level weights in Section 4.2 are local dimension weights; the criterion-level weights in Section 4.3 and Table 3 are renormalized global weights; the policy-capacity weights in Sections 3.4 and 5 follow the identical local-to-global normalization. Combined AHP-entropy weighting schemes provide an established precedent for this normalization logic (Ren et al., 2026; Chen, 2020).

Defuzzification via Yager’s Index. Fuzzy weights are converted to scalar values using Yager’s (1981; 2003) index:

Yager’s index is preferred over centroid-based alternatives because it assigns greater weight to the modal value, acknowledging that the most likely expert judgment carries genuine epistemic authority even when surrounded by uncertainty, while retaining proportional sensitivity to interval spread.

Consistency Verification. Internal consistency is verified through the Consistency Ratio:

All retained CR values satisfy the 0.10 threshold, indicating that the corresponding comparison matrices exhibit the internal coherence conventionally required for their weights to serve as policy evaluation inputs. CR values were computed for every one of the 702 matrices individually and disaggregated by hierarchical level (dimension, sub-dimension, and criterion) and by stakeholder group (government, industry, public) before the aggregate statement above was drawn; the manuscript reports the aggregate finding that every retained matrix satisfies the threshold rather than tabulating all individual values; the full matrix-level and stakeholder-group-level CR table, together with the disposition of any matrix that failed to meet the threshold on first administration, is available from the corresponding author on request.

3.4 Step four: policy evaluation under three decision criteria

FAHP-derived CDRMP weights are used to evaluate three marine policy response categories: marine spatial policy, marine economic policy, and marine environmental policy. For each policy type, a separate FAHP questionnaire elicits respondents’ assessment of that policy’s capacity to address each of the 36 CDRMP criteria. The resulting policy-specific fuzzy weights are compared against CDRMP severity weights; the preferred policy under this rule is defined as the one minimizing the aggregate absolute deviation between the two weight vectors.

Rather than adjudicating between single governance postures, this study applies three decision criteria drawn from classical decision theory under uncertainty. The optimistic criterion (max-max) selects the policy minimizing absolute difference between least-severe CDRMP weights and maximum policy capacity weights, reflecting confidence in favorable implementation. The pessimistic criterion (max-min) selects the policy minimizing difference between most-severe CDRMP weights and minimum policy capacity weights, reflecting precautionary protection against worst-case governance failure. The equal-probabilities criterion (mean-mean) selects the policy minimizing difference between geometric mean CDRMP severity weights and geometric mean policy capacity weights, assigning no prior probability to favorable or unfavorable conditions.

The formal selection rule is:

where is the defuzzified CDRMP weight for criterion i and is the defuzzified policy capacity weight for policy type p; p* minimizes δp. Criterion-specific versions substitute lower bound, upper bound, or geometric mean values of TFNs prior to defuzzification, as elaborated in Section 5.

Each of the three policy-specific questionnaires (marine spatial, marine economic, marine environmental policy) follows the identical respondent structure, hierarchy, aggregation, and local-to-global weighting procedure described in Sections 3.2–3.3, applied to policy capacity rather than CDRMP severity judgments, and produces its own set of policy-specific pairwise comparison matrices, fuzzy weight vectors, and defuzzified global weights at the dimension and criterion levels. The full policy-specific respondent counts, matrices, aggregation steps, defuzzification, and final weight vectors underlying the deviation scores in Table 4 are not reported individually in the main text for length reasons and are available from the corresponding author on request.

Table 4

Decision criterionGovernance postureMarine environmental policyMarine spatial policyMarine economic policyPreferred policy
Optimistic (max-max) at dimension levelConfidence in best-case implementation1.1881.4021.315Marine environmental
Optimistic (max-max) at criterion levelConfidence in best-case implementation1.2741.3961.201Marine economic
Pessimistic (max-min) at dimension levelPrecaution against worst-case outcomes1.3471.0101.289Marine spatial
Pessimistic (max-min) at criterion levelPrecaution against worst-case outcomes1.3981.0631.312Marine spatial
Equal-probabilities (mean-mean) at dimension levelBalanced, no prior governance assumption0.8561.2041.098Marine environmental
Equal-probabilities (mean-mean) at criterion levelBalanced, no prior governance assumption0.9211.1871.043Marine environmental

Policy evaluation summary: aggregate deviation scores across three decision criteria.

3.5 Methodological justification

Three features of the CDRMP problem distinguish it from decision environments for which classical AHP and alternative quantitative methods were designed (Watróbski et al., 2019), and motivate the specific instrument combination adopted in Section 3. Judgmental ambiguity in the CDRMP hierarchy is irreducible: the 36 indicators span domains as disparate as geopolitical maritime strategy and sedimentary carbon dynamics, and no individual expert commands sufficient breadth for precise numerical beliefs across all pairwise comparisons. FAHP encodes this ambiguity structurally, transforming what would otherwise be measurement error in classical AHP into a first-class analytical feature of the weight distribution (Buckley, 1985). Stakeholder heterogeneity is principled: government, industry, and public respondents do not share a common objective function, and imposing an a priori weighting scheme would reproduce the governance bias that participatory governance scholarship identifies as decisive for cross-domain risk legitimation (Voinov et al., 2018); the CDRMP framework is designed to diagnose. Geometric mean aggregation across groups ensures that the resulting weights reflect a genuine synthesis of plural perspectives. Data scarcity is structural: input-output modeling, CGE methods, and agent-based simulation all require detailed bilateral ecological data that do not exist at the required resolution, specifically because bilateral data-sharing frameworks are typically the first casualty of the political adversity characterizing transboundary risk events. FAHP’s reliance on expert judgment alone makes it especially well suited to the governance contexts where CDRMP assessment is most urgently needed (Jin et al., 2026; McCook et al., 2025). Prior applications confirm FAHP’s capacity to deliver robust risk prioritization under exactly these data-scarce, expert-judgment-dependent conditions, from urban flood-risk screening (Lyu et al., 2018) to multi-tier supply-chain risk mitigation under uncertainty (Awasthi et al., 2018).

4 Empirical analysis: nuclear wastewater discharge and its cross-domain risks to blue carbon ecosystems

4.1 Case justification

A qualifying CDRMP event must simultaneously generate risks across all four channels, expose the blue carbon ecosystem services of neighboring states to quantifiable harm, and arise from a unilateral decision that no existing multilateral instrument is equipped to evaluate or constrain. Japan’s nuclear wastewater discharge satisfies the three qualifying conditions stated above and is treated here as an illustrative rather than an exhaustively validated paradigmatic case; Section 7 identifies applying the same procedure to alternative transboundary events as a direction for establishing whether the pattern found here generalizes.

The discharge is irreducibly cross-domain. Geopolitical legality disputes under UNCLOS, restructured regional seafood markets, disrupted aquaculture and maritime tourism, and radionuclide dispersal into sedimentary carbon pools plausibly interact rather than operate as independent channels, though the manuscript does not trace a formal causal chain linking them. Diplomatic contestation over monitoring transparency shapes the political environment within which economic recovery measures must be negotiated; supply chain disruptions alter the industrial landscape within which marine environmental governance is attempted; and sediment contamination undermines the ecological substrate sustaining both the affected industries and the communities dependent upon them (Chen et al., 2025; Jin et al., 2026).

The blue carbon exposure of East Asian coastal states is plausible on oceanographic grounds and, for the general mechanisms involved, quantifiable in the literature, though this manuscript does not trace a case-specific evidence chain from the discharge to measured ecological outcomes in any single blue carbon habitat. The mangrove forests, seagrass beds, and salt marshes of China, South Korea, and Vietnam lie within the oceanographic reach of Pacific circulation patterns that would redistribute discharged water over decadal timescales. These ecosystems enter this risk period in a condition of structural fragility. Radionuclide dispersal threatens to accelerate blue carbon loss through three documented pathways: direct uptake into sedimentary organic matter; disruption of microbial carbon fixation communities already destabilized by thermal stress (Xiao et al., 2026); and damage to megafaunal grazers; Schmitz et al. (2026) quantify a trophic amplification factor of 2.63 for dugong-seagrass interactions in a different geographic and ecological context, and that figure is used here as an illustration of the general mechanism rather than as a case-specific estimate for the discharge site. The cumulative exposure aggregates individually documented impact mechanisms onto a system already at the edge of its resilience envelope.

The governance vacuum is equally clear. No existing multilateral instrument provides a systematic procedure for evaluating the transboundary CDRMP the discharge generates or for comparing alternative policy responses across all four risk channels (Zhang et al., 2026; Karli et al., 2026). Jin et al. (2026) identify exactly this absence as the critical frontier for integrated ocean governance. The CDRMP framework developed in this study addresses that frontier directly; the nuclear wastewater case is its first illustrative application.

The choice of this case over other transboundary marine risks, including oil spills, port expansion, industrial aquaculture, coastal reclamation, and microplastic pollution, follows from the qualifying criteria stated above rather than from its prominence alone. Oil spills and coastal reclamation activate the environmental and industrial channels but rarely generate a formal, sustained interstate legal dispute of the kind lodged against the discharge under UNCLOS. Industrial aquaculture expansion and microplastic pollution are typically diffuse, chronic, and attributable to many actors rather than to a single unilateral state decision, which weakens their suitability for a single-event CDRMP application. Port expansion, while spatially concentrated, does not carry the multi-decade radionuclide monitoring dispute that keeps the political channel active over the same horizon as the environmental one. The nuclear wastewater discharge is therefore treated as an illustrative rather than an exhaustively validated case, for the specific reason that it is the transboundary event in the recent record that most cleanly isolates the object the CDRMP framework is designed to evaluate: simultaneous, high-magnitude activation of all four risk channels from a single unilateral decision. The Limitations and Directions for Future Research subsection below returns to the resulting single-case limitation and to the value of applying the same procedure to these alternative cases in future research. The absence of a prior Web of Science match for the exact phrase Cross-Domain Risk from Marine Policies, reported in Section 3.1, is offered only as a terminological check confirming that CDRMP has not already been named and operationalized elsewhere; the substantive claim that no integrated four-channel assessment tool exists is established independently through the domain-specific literatures reviewed in Section 2, not by the phrase search alone.

4.2 Cross-domain risk severity: dimension-level findings

The FAHP procedure yields defuzzified dimension weights revealing how severely each of the four CDRMP channels is activated by the nuclear wastewater discharge. Figure 3 presents these weights as a radar chart, displaying the overall risk profile of the discharge as experienced by blue carbon coastal states. The markedly asymmetric polygon, with the environmental dimension extending far beyond the economic dimension at the center, visually encodes the discharge’s distinctively ecological governance character.

Figure 3

The Marine Culture and Environment dimension (MC0) commands the highest weight. This reflects not merely respondents’ concern about a specific contamination event but their recognition of a structural truth established in Section 2.2: blue carbon sedimentary stocks, once degraded by contamination or trophic disruption, cannot be restored on governance timescales. The primacy of MC0 is consistent with temporal irreversibility being reflected in respondents’ pairwise judgments, though the survey design does not test this interpretation against alternative explanations for the ranking.

The Maritime Politics dimension (MP0) receives the second-highest weight, and its near-equivalence to the environmental dimension carries an important analytical implication. The discharge has generated not merely a political disagreement but a governance crisis over which international norms apply, which monitoring obligations are binding, and which legal forum is competent to adjudicate. This is exactly the condition polycentric governance theory predicts when unilateral marine policy generates transboundary costs without multilateral coordination mechanisms (Ostrom, 2009; Karli et al., 2026). Chen et al. (2025) demonstrate empirically that political instability in regional governance arrangements measurably disrupts coastal ecosystem management.

Marine Industry (MI0) ranks third, capturing structural disruption in regional seafood, aquaculture, and maritime tourism sectors. The analytical point is not simply that industries suffer economic losses but that those losses feed back into blue carbon ecosystem condition: economically unviable aquaculture operations in mangrove and seagrass habitats generate habitat abandonment; collapsed maritime tourism withdraws the economic rationale for maintaining coastal buffer zones (Ma et al., 2025; Du et al., 2026). Industrial and ecological risk are coupled dynamics, not parallel channels.

Maritime Economics (ME0) receives the lowest weight, reflecting not the stakes but the relative manageability of economic risks. The toolkit for addressing them, market access adjustment, import restriction, trade compensation, fishery subsidy reorientation, is well-developed and routinely deployed in analogous transboundary contexts (Santos, 2025; Khaskheli et al., 2025). The comparatively low ME0 weight highlights by contrast the severity of dimensions where the policy toolkit remains structurally underdeveloped.

4.3 Cross-domain risk severity: criterion-level findings

Dimension-level weights establish the broad topography of CDRMP severity; criterion-level weights map its terrain in detail. Table 3 presents the seven highest-ranked criteria with defuzzified weights and blue carbon implications. Figure 4 presents the complete weight distribution across all 36 CDRMP criteria as a horizontal bar chart, with blue-carbon-specific indicators highlighted in teal.

Figure 4

Marine environmental pollution CDRMP within the MC0,2 sub-dimension ranks first at 0.206, reflecting cross-stakeholder consensus that sediment contamination is the most severe and durable threat. Radionuclide uptake into mangrove and seagrass sedimentary organic matter is not a process that reverses on governance timescales; once incorporated into sedimentary carbon pools, radioactive isotopes persist in biogeochemically active forms that continue affecting microbial carbon fixation communities long after surface contamination dissipates. That those communities already lose more than one third of their deep fixation capacity to thermal stress (Xiao et al., 2026) means radionuclide contamination compounds an existing vulnerability. The same FAHP apparatus has proven decisive for prioritizing environmental and energy risks where biogeochemical persistence matters, including electric-vehicle infrastructure siting (Bi et al., 2025) and solar-farm location under integrated fuzzy-GIS frameworks (Salami et al., 2026). The primacy of this criterion provides empirical support for the theoretical argument that blue carbon ecosystems are the primary downstream casualty of environmental CDRMP.

Global maritime strategic CDRMP within MP0,1 ranks second at 0.197, remarkably close to the environmental leader. Respondents weight geopolitical contestation as an independent source of durable risk in its own right, reflecting the judgment that contestation over Indo-Pacific maritime governance norms has consequences for blue carbon protection as durable as the contamination itself.

Ecosystem pollution CDRMP within MC0,2 ranks third at 0.128, capturing trophic and community-level ecological impacts extending beyond the direct contamination footprint. The trophic amplification factor of 2.63 quantified by Schmitz et al. (2026) for dugong-seagrass interactions means a discharge event disrupting megafaunal grazers carries carbon accounting consequences 2.63 times larger than direct habitat impact alone would suggest. This second-order blue carbon impact is the kind of consequence that generic environmental criteria without ecological calibration would fail to register.

Inter-country marine governance organization disputes within MP0,2 rank fourth at 0.123, capturing the institutional failure of multilateral bodies to establish coherent monitoring protocols and coordinate precautionary responses. No single legal order commands sufficient authority to mandate the coordinated precautionary action the ecological severity of the discharge would require (Zhang et al., 2026).

The fifth through seventh ranked criteria span coastal erosion and disaster (MC0,2, 0.081), biodiversity loss (MC0,2, 0.078), and carbon sink degradation (MC0,3, 0.068). Five blue-carbon-relevant criteria appear within the top seven of 36 — the study’s most analytically significant empirical finding. Of these five, two (marine environmental pollution, ecosystem pollution) have direct analogues in standard environmental indicator sets, whereas the remaining three — coastal erosion and disaster, biodiversity loss, and carbon sink degradation — are blue-carbon-specific additions introduced by the ecological extension. If a generic, non-blue-carbon-calibrated framework retained only the two criteria with standard analogues and omitted these three blue-carbon-specific additions, it would, under that illustrative construction, leave three of five (60%) of the highest-severity blue-carbon-relevant criteria unrepresented. This figure describes a hypothetical comparison against an unweighted generic indicator set constructed for illustration, not an empirical result from a re-estimated comparator model; a formal comparison against an actual non-calibrated FAHP framework is identified in the Limitations subsection as a direction for future validation.

Two clarifications are required. First, the 60% figure is an expository counterfactual, not an empirical estimate: it answers how many of the five blue-carbon-relevant top criteria a parsimonious generic framework would omit under an explicitly specified construction, and it carries no statistical claim that blue carbon calibration provably enhances risk-identification capability. A formal test of that stronger claim requires estimating the CDRMP alongside a non-calibrated FAHP on the same dataset with transparent indicator mapping, which is reserved for future validation (Section 7). Second, the role of the standard environmental indicators retained in the hierarchy — marine environmental pollution and ecosystem pollution — must be distinguished from that of the blue-carbon-specific criteria. The standard indicators capture the generic degradation pressures and the physical and trophic transmission channels (contaminant loading, trophic disruption, sediment removal, species loss) through which external marine policies degrade blue carbon stocks; the blue-carbon-specific criteria — coastal erosion and disaster, biodiversity loss, carbon sink degradation, and habitat destruction — encompass both these extension-surfaced pressures and, most critically, the temporally irreversible sedimentary dimension (captured strictly by carbon sink degradation and habitat destruction) that generic indicators cannot register. We note that coastal erosion and biodiversity loss are also monitored within several established environmental assessment frameworks; the 60% figure is therefore a deliberately conservative illustrative lower bound under the specified construction rather than a claim about all conceivable generic frameworks. Their joint retention lets the FAHP weight the general transmission channels against the irreversible-specific losses rather than forcing a binary choice, which is precisely why the five top-ranked criteria span both classes.

4.4 Policy response capacity: alignment with the CDRMP risk profile

The policy response capacity assessment evaluates how effectively each of the three marine policy alternatives aligns with the CDRMP risk severity profile established in Sections 4.2 and 4.3. Table 5 summarizes the directional policy capacity findings; quantitative deviation scores are reported in full in Section 5. Figure 5 presents the capacity profiles across all four CDRMP dimensions as a grouped bar chart, enabling visual comparison of where each policy alternative’s strengths and weaknesses lie relative to the risk landscape.

Table 5

Policy alternativePrimary alignmentPrimary limitationBlue carbon coverage
Marine environmental policyStrongest at the environmental and ecological sub-dimensions; instruments directly calibrated to sediment contamination, ecosystem pollution, and carbon sink degradationLimited capacity to resolve the geopolitical governance disputes that rank second in overall severityHighest among the three alternatives; the only policy type whose instruments directly target the MC0,2 and MC0,3 sub-dimensions where all blue carbon-specific criteria are concentrated
Marine spatial policyStrong alignment at the political and industrial dimensions; zoning and boundary management instruments address jurisdictional and resource competition disputes effectivelySpatial tools can protect habitat boundaries but cannot remediate contamination already present in sediments or restore disrupted trophic communitiesModerate; spatial protection reduces future habitat exposure but leaves existing blue carbon stock degradation unaddressed
Marine economic policyMost balanced profile across all four dimensions; trade and market instruments provide broad if shallow cross-dimensional coverageBreadth achieved at the cost of depth; does not match the specialization of either alternative at their respective high-severity criteriaLowest for the highest-severity blue carbon criteria; economic instruments address the consequences of blue carbon loss rather than its causes

Policy response capacity across the four CDRMP dimensions.

Figure 5

Each policy alternative is operationalized here as a defined instrument bundle rather than as a general policy label. Marine environmental policy comprises contamination monitoring and radionuclide surveillance, statutory emission and discharge standards, sediment remediation programs, habitat protection designations, and marine protected area networks. Marine spatial policy comprises exclusive economic zone management, maritime boundary clarification, spatial zoning separating industrial from conservation uses, and strategic blue carbon habitat designation. Marine economic policy comprises trade and market-access measures, fishery and aquaculture subsidy reorientation, import restriction and compensation mechanisms, and blue economy investment incentives. Table 5 and the deviation scores reported in Section 5 evaluate these three instrument bundles, not the three domain labels considered in the abstract.

Marine environmental policy achieves the strongest overall alignment. The five blue-carbon-relevant criteria dominating the top seven are located within MC0,2 and MC0,3, the sub-dimensions that environmental policy instruments, contamination monitoring, emission standards, sediment remediation, habitat protection designations, and marine protected area networks, are specifically designed to address. Its structural weakness is equally clear: the geopolitical governance crisis captured by top-ranked MP0 criteria requires diplomatic and legal instruments that lie outside environmental regulation’s reach.

Marine spatial policy’s comparative advantage lies exactly where environmental policy is weakest. Spatial zoning, EEZ management, maritime boundary clarification, and strategic habitat designation address the jurisdictional and strategic dimensions of MP0 criteria ranking second and fourth in severity. Its alignment with MC0,2 and MC0,3, however, is limited: spatial tools can designate blue carbon habitats as protected zones but cannot remediate existing sediment contamination, restore trophic communities disrupted by food web contamination, or reverse carbon sink degradation from microbial destabilization.

Marine economic policy offers the most balanced alignment across all four dimensions while providing secondary coverage of political and environmental dimensions through trade leverage and market access management. Breadth, however, comes at the cost of depth: economic policy matches neither environmental policy’s capacity at the highest-severity blue carbon criteria nor spatial policy’s capacity at the highest-severity political criteria. Its profile is that of a governance complement and facilitator rather than a primary response instrument.

The fundamental implication is that no single policy achieves comprehensive CDRMP coverage across all four dimensions simultaneously, a reflection not of a limitation in the alternatives themselves but of a structural feature of the cross-domain problem. Addressing it comprehensively requires instruments calibrated to each channel’s distinctive governance logic, a finding the decision criteria analysis in Section 5 addresses directly.

5 Policy evaluation: comparative performance under three decision criteria

5.1 The logic of multi-criteria policy comparison

Selecting a preferred marine policy response does not admit a single correct answer but a family of answers, each defensible under a different assumption about governance actors’ behavior under uncertainty. The three decision criteria applied in this section formalize three governance postures as distinct rules of comparison. The optimistic criterion (max-max) evaluates each policy alternative by its best-case risk resolution capacity relative to minimum CDRMP severity. The pessimistic criterion (max-min) evaluates each by its worst-case capacity relative to maximum CDRMP severity. The equal-probabilities criterion (mean-mean) evaluates each by geometric mean capacity relative to geometric mean severity, assigning no prior probability to favorable or unfavorable conditions.

The value of applying all three simultaneously lies in constructing a robustness profile: a structured account of how rankings shift as governance assumptions change, and of which alternatives are robust across the full range versus contingent on specific assumptions. Table 4 presents aggregate deviation scores across all three criteria and both analytical levels. Figure 6 presents the robustness profile as a heatmap in which color intensity encodes relative alignment strength, making the pattern of convergence and divergence across criteria immediately visible.

Figure 6

5.2 Optimistic criterion: best-case risk resolution

Under the optimistic criterion, dimension-level analysis selects marine environmental policy with an aggregate deviation of 1.188. At the criterion level, however, the ranking shifts: marine economic policy achieves the lowest deviation. This dimension-to-criterion divergence is analytically instructive rather than contradictory. It reveals that environmental policy’s superiority under favorable governance assumptions is concentrated in its exceptional alignment with high-severity blue-carbon-relevant criteria dominating the top seven positions. When disaggregated to the criterion level, marine economic policy’s broad cross-criterion coverage becomes visible as a marginal advantage, because in best-case governance scenarios the breadth of an economic policy toolkit partially compensates for its lack of depth in the most ecologically critical areas.

The deeper lesson is that the case for marine environmental policy under favorable governance conditions rests on the irreversibility argument from Section 2.2. Even under optimistic implementation assumptions, the blue-carbon-relevant criteria for sediment contamination, trophic disruption, carbon sink degradation, and biodiversity loss carry weights reflecting recognition that no subsequent governance intervention can restore degraded sedimentary stocks. Under any criterion that takes blue carbon loss seriously, environmental policy retains a structural advantage that economic breadth cannot overcome.

5.3 Pessimistic criterion: worst-case risk protection

The pessimistic criterion produces the clearest and most consistent ranking across both analytical levels: marine spatial policy achieves the lowest aggregate deviation at both the dimension level (1.010) and the criterion level. This finding reflects a structural feature of spatial instruments that becomes decisive under precautionary logic. Spatial policy’s capacity to establish jurisdictional boundaries, designate marine protected areas, and regulate marine industrial geographic footprint does not depend on political cooperation from the originating state in the way that environmental regulatory enforcement or trade negotiation does. Under pessimistic assumptions, minimal bilateral cooperation and weak multilateral enforcement, spatial instruments operating within the receiving state’s sovereign jurisdiction provide the most reliable floor of blue carbon protection.

This finding is further supported by the geopolitical risk profile. The second-highest CDRMP severity weight belongs to global maritime strategy, reflecting sustained contestation over Indo-Pacific maritime governance norms. Spatial policy directly addresses this by clarifying receiving-state jurisdictional claims over blue carbon ecosystems within its EEZ, creating a legal foundation for precautionary habitat protection without requiring originating-state acquiescence. Zhang et al. (2026) demonstrate that fragmented international blue carbon law incentivizes precisely this form of unilateral domestic regulatory response. The pessimistic criterion result is therefore a conditional endorsement: it is complementary to rather than contradictory of the optimistic criterion’s preference for environmental policy.

5.4 Equal-probabilities criterion: balanced assessment

The equal-probabilities criterion restores marine environmental policy to the top ranking at both analytical levels, with an aggregate deviation of 0.856 at the dimension level, the lowest score of any policy under any criterion. This convergence between the optimistic and equal-probabilities criteria on marine environmental policy, with the pessimistic criterion dissenting in favor of spatial policy, constitutes the study’s primary policy finding.

Under balanced accounting of all 36 criteria, marine environmental policy’s alignment with the five blue-carbon-relevant criteria concentrating the greatest ecological severity consistently outweighs spatial policy’s superiority at political and jurisdictional criteria, because the geometric mean calculation preserves the signal that environmental and ecological risks are not one consideration among four but the dimension where the most irreversible harms concentrate.

5.5 Cross-criterion synthesis: robustness, conditionality, and blue carbon governance implications

The robustness profile summarized in Table 4 and visualized in Figure 6 yields four findings with direct implications for blue carbon governance under transboundary marine policy risk.

Across the six rule-by-level combinations in Table 4, no single policy dominates. Marine environmental policy is preferred under two of three decision criteria at the dimension level (optimistic and equal-probabilities) and under one of three at the criterion level (equal-probabilities), and it attains the single lowest aggregate deviation of any policy under any criterion (0.856, equal-probabilities, dimension level). Marine spatial policy is preferred under the pessimistic criterion at both levels. Marine economic policy is preferred under the optimistic criterion at the criterion level, the only combination in which it is preferred. This pattern is consistent with, without directly demonstrating, a layered governance role in which spatial measures help establish the jurisdictional perimeter within which environmental instruments operate; the analysis evaluates the three policy alternatives independently rather than testing a combined instrument, so this reading is offered as an interpretation rather than a finding the deviation-minimization procedure itself establishes (see Limitations). Marine economic policy is preferred under only one of the six rule-level combinations, consistent with a facilitative rather than primary role.

The fourth finding is methodological. The divergence between dimension-level and criterion-level rankings under the optimistic criterion demonstrates that policy selection conclusions are sensitive to analytical disaggregation: recommendations based solely on dimension-level analysis risk obscuring criterion-level heterogeneity that can reverse rankings, while recommendations based solely on criterion-level analysis may be driven by individually numerous but collectively less consequential low-weight criteria. The CDRMP framework’s capacity to support analysis at both levels simultaneously is therefore an analytical necessity rather than a methodological feature.

6 Discussion: situating CDRMP within the blue carbon governance literature

The empirical findings enter a scholarly conversation already underway among researchers in blue carbon science, transnational environmental law, marine ecosystem services, and multi-criteria governance. This section organizes that dialogue around four themes: the adequacy of existing blue carbon governance frameworks, the ecological irreversibility thesis and its policy implications, the transmission of transboundary policy spillovers through coupled political and ecological channels, and the methodological implications of a governance-centered rather than biophysics-centered approach to blue carbon risk assessment. Table 6 situates the CDRMP framework relative to the four literature streams it most directly engages and Table 6 distills the dialogue into a systematic comparison of prior findings, this study’s contributions, and their governance implications.

Table 6

Debate in the literaturePrior findingThis study’s contribution
Adequacy of existing governance frameworksZhang et al. (2026): international law fragmentation creates governance voids and incompatible bottom-up innovationCDRMP framework operationalizes the governance gap as a 36-criterion risk assessment procedure, indicating that fragmentation generates quantifiable cross-domain severity
Ecological irreversibility and policy discountingZheng et al. (2026): global salt marsh losses concentrated in irreplaceable mature habitats; Rodrigues et al. (2026): 55 to 77% TECS decline driven by carbon density asymmetryExpert FAHP weights may implicitly reflect irreversibility: blue carbon-specific criteria dominate the upper weight distribution without explicit instruction to weight permanence
Trophic amplification and indirect carbon pathwaysSchmitz et al. (2026): dugong conservation amplifies seagrass sediment carbon stocks by a factor of 2.63Ecosystem pollution criterion ranks third overall at 0.128, but the policy best aligned with trophic risk performs weakest at the political dimension that enables trophic disruption
Transboundary climate policy and ecological stabilityChen et al. (2025): stronger transnational climate policy measurably enhances coastal ecosystem resilience in the Baltic SeaPolitical CDRMP severity at 0.197 rivals environmental severity at 0.206, indicating that governance instability and ecological contamination are mutually reinforcing rather than sequential events
Ecosystem service undervaluation and governance biasMcCook et al. (2025): systematic undervaluation of blue carbon services in national policy drives unsustainable governance choicesIndustry respondents assign high weights to environmental contamination criteria, revealing latent preferences for blue carbon protection that standard market signals obscure
MRV data infrastructure and governance integrationWahsheh et al. (2026): cryptographic framework secures blue carbon accounting data lifecycle but does not address policy evaluationCDRMP framework supplies the governance evaluation layer that MRV frameworks do not provide, converting monitoring data into policy-comparable risk weights
Human activity as dominant driver of blue carbon distributionWang et al. (2026a): human activity intensity explains 24% of seagrass distribution variation, exceeding biophysical parametersCDRMP framework treats governance decisions as primary risk drivers and ecological outcomes as downstream consequences, inverting the causal framing of biophysics-first assessment tools

CDRMP findings in dialogue with key blue carbon governance literature.

6.1 Confirming and extending the governance failure diagnosis

The most fundamental finding of this study is that blue carbon ecosystems face a governance challenge, simultaneously transboundary, multi-domain, and temporally irreversible, for which no existing institutional arrangement provides systematic assessment or response capacity. What the CDRMP framework contributes is the transformation of this diagnosis from a qualitative observation into an empirically operational and quantitatively tractable procedure.

Zhang et al.’s (2026) transnational environmental law analysis arrives at the same governance gap through a comparative legal lens, concluding that blue carbon governance requires a coordinated system shaped by state and non-state actors across different jurisdictions. The CDRMP framework provides precisely what this legal analysis implies but does not supply: a structured procedure for evaluating how regulatory outputs of one jurisdiction translate into quantified cross-domain risks for the blue carbon systems of another. The two studies are therefore complementary in disciplinary positioning, the legal analysis furnishing the institutional diagnosis and the CDRMP framework furnishing the assessment instrument the diagnosis implies.

Karli et al.’s (2026) identification of six structural pathologies transcending any single national context resonates directly with the finding that the Maritime Politics dimension carries the second-highest CDRMP severity weight. Geopolitical contestation and institutional coordination failure are not background conditions that well-designed governance can be built around; they are first-order risk factors whose severity rivals direct environmental contamination. A Maritime Politics weight of 0.197 against a Marine Culture and Environment weight of 0.206 adds quantitative texture to a qualitative finding that comparative governance scholarship has asserted but not measured. Khaskheli et al.’s (2025) analysis reveals the socioeconomic mechanisms through which legal fragmentation produces differential vulnerability; the CDRMP framework addresses this asymmetry by providing a replicable, data-lean assessment procedure using expert survey data rather than bilateral data-sharing agreements whose negotiation is typically undermined by the very political adversity characterizing transboundary risk events.

6.2 Deepening the ecological irreversibility argument

The ecosystem services economics literature has long maintained that temporal irreversibility distinguishes blue carbon stocks from most other categories of marine ecological value and demands prospective rather than retrospective governance responses. The empirical findings provide two forms of quantitative support extending beyond prior literature.

The first is the systematic dominance of blue-carbon-relevant criteria in the upper tail of the CDRMP weight distribution. Five of the seven highest-ranked criteria fall within the two ecological-extension sub-dimensions of MC0, with the highest-ranked criterion receiving 0.206. This consensus is not merely a political preference; it is ecologically grounded. Zheng et al.’s (2026) global assessment documents losses concentrated in mature, high-carbon-density habitats whose recovery is measured in centuries rather than decades. One reading of the expert FAHP weights is that they function not merely as preference orderings but as judgments that may implicitly reflect irreversibility: respondents weight sediment contamination most severely, which is consistent with, though does not on its own establish, an implicit recognition that its consequences cannot be undone on governance timescales. Rodrigues et al.’s (2026) quantification of 55 to 77% TECS declines in Amazonian mangroves over 38 years, driven by the inability of newly forming stands to compensate for mature forest carbon density, encodes the same temporal asymmetry.

The second form of support is methodological. The finding that FAHP weights for blue-carbon-specific criteria consistently rank above those for generic marine environmental criteria, when both are presented simultaneously in an open-ended elicitation procedure, suggests that properly structured expert judgment encodes ecological irreversibility in its preference structure even without explicit instruction. This extends Macreadie et al.’s (2026) synthesis of next-decade blue carbon research priorities by showing that governance integration of blue carbon co-benefits is achievable through multi-criteria elicitation without requiring technically demanding biophysical monitoring systems. Schmitz et al.’s (2026) 2.63 trophic amplification factor introduces a pathway that existing blue carbon governance frameworks do not accommodate; critically, the policy best aligned with trophic disruption risk performs most weakly at the political dimension where jurisdictional disputes enabling trophic disruption originate, confirming that resolving trophic blue carbon risks requires governance instruments spanning environmental and political domains simultaneously.

6.3 Engaging the transboundary policy spillover literature

The finding that geopolitical contestation (0.197) rivals direct environmental contamination (0.206) speaks to a debate about whether political and ecological channels should be analyzed jointly or treated as separate governance objects. Chen et al.’s (2025) Baltic Sea analysis demonstrates that stronger transnational climate policy stability measurably enhances coastal ecosystem resilience through mitigated carbon emissions and reduced eutrophication. The CDRMP analysis corroborates this in a qualitatively distinct context and extends it: where Chen et al. show policy stability benefits ecological stability, the CDRMP framework shows that policy instability concentrates risk most severely in ecological and political domains simultaneously, with the two channels mutually reinforcing rather than merely co-present.

McCook et al.’s (2025) identification of systematic blue carbon service undervaluation in national policy as the primary mechanism driving unsustainable governance choices receives empirical challenge from this study’s finding that even marine industry operator respondents assign comparably high weights to environmental contamination and ecosystem pollution criteria. This suggests the undervaluation reflects a governance system failure rather than an individual preference failure, and that structured elicitation procedures can surface latent preferences for blue carbon protection that standard market signals systematically obscure. Jin et al.’s (2026) argument for expanded natural-social science integration in marine affairs provides the broadest intellectual context: the framework operationalizes exactly the integration Jin et al. advocate, spanning natural science dimensions (sediment contamination, ecosystem pollution, carbon sink degradation), social science dimensions (governance disputes, cultural heritage, community impact), and economic dimensions (resource allocation, market disruption, supply chain resilience), respecting their incommensurability rather than converting them to a common monetary metric.

6.4 Methodological implications for blue carbon risk assessment

The sensitivity of policy evaluation conclusions to analytical level under the optimistic criterion carries implications for the blue carbon MRV literature. Wahsheh et al.’s (2026) cryptographic framework secures blue carbon accounting data infrastructure but does not engage with the governance evaluation dimension: even perfectly secured monitoring data do not automatically produce sound policy decisions if the evaluation framework fails to capture the full CDRMP spectrum. The CDRMP framework addresses the governance evaluation layer Wahsheh et al.’s technical infrastructure leaves open.

Wang et al.’s (2026a) finding that human activity intensity explains 24% of seagrass distribution variation, exceeding biophysical parameters, carries a direct methodological implication: risk frameworks treating blue carbon outcomes as biophysical-parameter-determined will systematically misattribute sources of vulnerability and misidentify appropriate mitigation instruments. The CDRMP framework corrects this misattribution at the level of design by treating governance decisions as primary risk drivers and ecological outcomes as downstream consequences. Lawless et al.’s (2026) analysis of tenure ambiguity across 122 blue carbon guidance documents further confirms that the finding of inter-country governance organization disputes ranking fourth at 0.123 demonstrates institutional ambiguity generates CDRMP severity weights large enough that any evaluation framework treating them as exogenous background conditions will produce systematically biased recommendations.

7 Conclusion

This study addressed a governance problem that existing literature had diagnosed qualitatively but not equipped analytically: the absence of a systematic, multi-domain procedure for evaluating the cross-domain risks that unilateral marine policy decisions impose on the blue carbon ecosystems of neighboring states. The nuclear wastewater discharge from Fukushima Daiichi provided the empirical context; the Nominal Group Technique and Fuzzy Analytic Hierarchy Process provided the methodological architecture; and a 36-criterion evaluation hierarchy calibrated explicitly to the ecological characteristics of blue carbon systems provided the assessment instrument.

Three empirical findings extend beyond the specific case examined. The CDRMP risk profile is dominated by the Marine Culture and Environment dimension, with two ecological-extension sub-dimensions contributing five of the seven highest-ranked criteria across all 36 indicators; the highest-ranked criterion, marine environmental pollution, receives 0.206, suggesting that a framework without the blue carbon extension would leave the most severe governance risks systematically undetected. Geopolitical contestation (MP0 weight 0.197) rivals direct environmental contamination (0.206), constituting the study’s most analytically unexpected finding: political governance instability and ecological contamination are mutually reinforcing rather than sequential risk events, and governance frameworks addressing them in isolation will systematically underprotect blue carbon systems. The policy evaluation under three decision criteria and two analytical levels shows no single policy dominating throughout: marine environmental policy is preferred under two of three criteria at the dimension level and one of three at the criterion level, and achieves the lowest single aggregate deviation of any policy under any criterion (0.856 under equal-probabilities at the dimension level); marine spatial policy is preferred under the pessimistic criterion at both levels; and marine economic policy is preferred only under the optimistic criterion at the criterion level, consistent with a facilitative rather than primary role.

Three theoretical contributions follow. The integration of polycentric governance theory with blue carbon ecosystem services economics generates a mechanistic account of why unilateral marine policies systematically produce cross-domain risks to transboundary ecological systems, an account that neither tradition has provided alone: prior polycentric governance research identified institutional fragmentation as a root cause of transboundary ecological failure (Ostrom, 1990, 2009; Zhang et al., 2026; Karli et al., 2026) without quantifying the magnitude of the ecological values at risk, while ecosystem services research quantified those values (Macreadie et al., 2026; Zheng et al., 2026; Wang et al., 2026b) without providing the governance mechanism account explaining why institutional systems consistently fail to protect them. The conceptualization of CDRMP as a transboundary externality problem with distinctive features, namely simultaneous propagation across four risk channels, concentration in temporally irreversible blue carbon systems, and dependence on expert elicitation rather than bilateral data observation, advances the transnational environmental law and blue carbon governance literatures simultaneously by providing the quantitative operationalization their respective legal and ecological analyses imply but do not supply. The demonstration that expert FAHP weights may implicitly reflect ecological irreversibility constitutes a novel, though interpretive, finding about the relationship between structured expert elicitation and the ecological knowledge informed stakeholders carry into governance assessment, extending Macreadie et al.’s (2026) synthesis by showing that governance integration of blue carbon co-benefits is achievable through multi-criteria elicitation without requiring the more technically demanding biophysical monitoring systems the synthesis also prioritizes. The economic dimensions are double edged: coastal economic growth can erode mangrove blue carbon (Gasmi et al., 2025), yet targeted government subsidies and blue carbon trading mechanisms can reinforce sinks and mobilize finance (He et al., 2025; Zhang et al., 2026).

Four practical implications follow for coastal states facing transboundary marine policy risks. Marine environmental policy is the policy most consistently favored by the decision-rule analysis, preferred under two of three criteria at the dimension level and attaining the single lowest aggregate deviation observed (0.856, equal-probabilities); this supports environmental regulatory capacity as a priority element of blue carbon CDRMP governance, though not to the exclusion of the other instruments. Marine spatial policy is preferred specifically under the pessimistic criterion at both levels, suggesting a protective role under adversarial governance conditions; because the analysis evaluates the three policy alternatives independently rather than testing a combined instrument, the suggestion that spatial measures might usefully complement environmental ones by establishing a jurisdictional perimeter within which environmental instruments operate is an interpretation consistent with the pattern of results rather than a finding the analysis directly demonstrates. Marine economic policy, preferred only under the optimistic criterion at the criterion level, is best understood as a facilitative rather than primary instrument. These readings rest on expert-elicited severity and capacity weights rather than on biophysical monitoring data, a substitution that is a deliberate design choice suited to adversarial governance conditions in which cooperative bilateral data collection is unavailable (Section 3.5), and not a claim that biophysical monitoring is unnecessary for blue carbon governance generally; the trade-offs of this substitution are discussed further in the Limitations subsection.

Four limitations qualify the scope of the findings and each motivates a corresponding line of future research. First, the CDRMP framework is applied to a single illustrative case, the nuclear wastewater discharge, selected and justified in Section 4.1 against port expansion, industrial aquaculture, coastal reclamation, oil spills, and microplastic pollution as candidate cases. A single-case design nonetheless limits generalizability; applying the same procedure to these alternative transboundary events is a natural next step for establishing whether the pattern identified here, political and environmental severity converging while economic severity lags, is case-specific or structural. Relatedly, the survey instrument recorded stakeholder group (government, industry, public) but not finer respondent-level demographic detail such as country of residence or institutional affiliation; collecting and reporting such detail, while protecting individual anonymity in a politically sensitive transboundary case, is a design refinement recommended for future applications of the CDRMP protocol. Relatedly, as noted in Section 3.3, the per-respondent record of which pairwise comparison matrices each respondent completed was not retained in a form permitting an exact reconciliation between the reported respondent and matrix totals, and the case-specific ecological evidence chain linking the discharge to measured outcomes in any single blue carbon habitat, as opposed to the general mechanisms documented in the cited literature, was not established directly in this study; both are reporting gaps rather than resolved findings, and closing them, together with a formal comparison against a non-blue-carbon-calibrated framework to test the illustrative 60% figure in Section 4.3, is recommended for future applications of the protocol.

Second, the FAHP results depend on the specific defuzzification rule (Yager’s index), on geometric mean aggregation across respondents and stakeholder groups, and on the equal a-priori weighting of the three stakeholder groups. Beyond the dimension-level and criterion-level comparison reported in Section 5.5, which functions as a partial robustness check by showing where conclusions are and are not sensitive to the level of aggregation, no formal sensitivity analysis varying the defuzzification method, the group-weighting scheme, or the respondent sample (for example via bootstrap resampling) was conducted for this study. This is reported here as a limitation rather than resolved directly, because doing so credibly requires re-estimation from the underlying pairwise comparison data rather than post hoc adjustment of the reported weights; correspondingly, Sections 5.5 and 7 describe the policy-preference pattern across decision criteria as varying by rule and aggregation level rather than as formally established to be robust in the statistical sense, since no such formal test was performed.

Third, the policy selection rule, minimizing the aggregate absolute deviation between severity and capacity weight vectors, is transparent and directly compatible with the fuzzy weights produced by the same FAHP procedure, but it is conceptually more permissive than approaches such as TOPSIS, VIKOR, or PROMETHEE, which rank alternatives by distance to ideal and anti-ideal solutions rather than by undifferentiated absolute deviation. It also implicitly treats a policy that mirrors the risk-weight distribution as desirable, rather than privileging a policy that disproportionately reduces the most severe risks. Deviation minimization is retained for this study for its transparency and its direct compatibility with the weights already estimated, but the conceptual choice is flagged explicitly, and a formal comparison against TOPSIS, VIKOR, and PROMETHEE rankings, and against alternative criteria-weighting procedures such as the Best-Worst Method (BWM) or the Full Consistency Method (FUCOM), is identified as an important direction for methodological validation.

Fourth, several CDRMP criteria are conceptually adjacent, for instance marine environmental pollution, ecosystem pollution, biodiversity loss, habitat destruction, and carbon sink degradation, as are several political criteria. These were retained as analytically distinct criteria, rather than merged, for the reason given in the note accompanying Appendix Table 1: the pairwise comparison procedure that generates FAHP weights compares each criterion against every other criterion directly, so a criterion’s weight reflects its judged severity relative to the full 36-criterion set rather than an additive count of related items, which limits though does not eliminate the double-counting risk that a simple summation or scoring procedure would carry. The discriminant validity of closely related criteria was not formally tested, for instance via confirmatory factor analysis on pilot data, and this is recommended as a validation step for future applications of the CDRMP hierarchy.

Beyond these four limitations, two future directions emerge. Extending the CDRMP framework to chronic rather than acute transboundary impacts, cumulative fishery depletion, persistent marine plastic pollution, and progressive encroachment of coastal development on blue carbon habitat boundaries, would broaden its applicability considerably. Formal integration between the CDRMP assessment procedure and the blue carbon MRV infrastructure being progressively institutionalized in international carbon accounting frameworks (Macreadie et al., 2026; Wahsheh et al., 2026) would create a governance architecture in which blue carbon monitoring, cross-domain risk assessment, and comparative policy evaluation form a coherent institutional system rather than three parallel programs each addressing one component of the governance challenge while leaving the other two unaddressed.

Statements

Data availability statement

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

Author contributions

WH: Writing – original draft, Writing – review & editing.

Funding

The author(s) declared that financial support was not received for this work and/or its publication.

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.

References

  • 1

    AdamsJ. B.MbenseN.Van NiekerkL. (2026). Blue carbon stocks, accumulation rates and knowledge gaps in Africa. Ocean Coast. Manage.277, 108208. doi: 10.1016/j.ocecoaman.2026.108208

  • 2

    AfzalA.BelascuL.Martinez-GonzaloJ. M.AwaisM. (2026). Do blue economic resources have potential for carbon management in ASEAN countries? An empirical assessment. Rev. Accounting Finance25, 210–226. doi: 10.1108/RAF-10-2024-0460

  • 3

    AwasthiA.GovindanK.GoldS. (2018). Multi-tier sustainable global supplier selection using a fuzzy AHP-VIKOR based approach. Int. J. Prod. Econ.195, 106–117. doi: 10.1016/j.ijpe.2017.10.013

  • 4

    BiH. L.GuY. Y.LuF. Q.MahreenS. (2025). Site selection of electric vehicle charging station expansion based on GIS-FAHP-MABAC. J. Cleaner Prod.507, 145557. doi: 10.1016/j.jclepro.2025.145557

  • 5

    BuckleyJ. J. (1985). Fuzzy hierarchical analysis. Fuzzy Sets Syst.17, 233–247. doi: 10.1016/0165-0114(85)90090-9

  • 6

    ChenC. H. (2020). A novel multi-criteria decision-making model for building material supplier selection based on entropy-AHP weighted TOPSIS. Entropy22, 259. doi: 10.3390/e22020259

  • 7

    ChenH. X.HuJ. Y.LiuY.LiZ. (2025). Transnational climate policy and marine ecological stability: Insights from the Baltic Sea. J. Environ. Manage.396, 128112. doi: 10.1016/j.jenvman.2025.128112

  • 8

    DelbecqA. L.Van de VenA. H. (1971). A group process model for problem identification and program planning. J. Appl. Behav. Sci.7, 466–492. doi: 10.1177/002188637100700404

  • 9

    DuW. B.ZhangH.ChenX. M.LiuY. (2026). Policy tool structures and innovation network evolution in China's marine ranching sector. Ocean Coast. Manage.276, 108166. doi: 10.1016/j.ocecoaman.2026.108166

  • 10

    FrehenL.Van TatenhoveJ.GilekM. (2025). Blue nature-based solutions in marine and coastal EU policies: From principles to practice. Mar. Policy177, 106687. doi: 10.1016/j.marpol.2025.106687

  • 11

    GasmiF.VirtoL. R.CouvetD. (2025). Coastal countries' economic growth: What are the potential consequences on mangrove blue carbon? J. For. Econ.40, 1–45. doi: 10.1561/112.00000586

  • 12

    HeX. H.HuangH. H.HuW. F. (2025). Analyzing the impact of government subsidies and other socioeconomic factors on blue carbon sinks in the coastal regions. Ecol. Econ.237, 108718. doi: 10.1016/j.ecolecon.2025.108718

  • 13

    JinD.ThunbergE.Kite-PowellH. (2026). Expanding the integration of natural and social sciences in marine affairs and policy. Front. Mar. Sci.13, 1797599. doi: 10.3389/fmars.2026.1797599

  • 14

    KarliA.ÖztürkB.GüneşY. (2026). Policy frameworks without practice? Exploring marine governance and the marine-climate nexus in Türkiye. Mar. Policy186, 107013. doi: 10.1016/j.marpol.2025.107013

  • 15

    KhaskheliM. B.JiangY.RazaS. A.QureshiM. A.KhaskheliA. (2025). Sustainable maritime governance of digital technologies for marine resource management. Sustainability17, 9526. doi: 10.3390/su17219526

  • 16

    KublerS.RobertJ.DerigentW.VoisinA.Le TraonY. (2016). A state-of-the-art survey & testbed of fuzzy AHP (FAHP) applications. Expert Syst. Appl.65, 398–422. doi: 10.1016/j.eswa.2016.08.064

  • 17

    LawlessS.CohenP. J.DudayevR.Evans-IllidgeE.LovelockC. E.MuirB.et al. (2026). Respecting tenure and the bundle of rights in blue carbon guidance documents. Nat. Clim. Change. doi: 10.1038/s41558-026-02651-8

  • 18

    LiuC. H.ZhangY.ChenX. B.WangS. (2026). Spatiotemporal dynamics, seasonal aggregation, and policy interventions in marine microplastic pollution. Environ. Sci. Technol.60, 12315–12326. doi: 10.1021/acs.est.6c03653

  • 19

    LuisettiT.TurnerR. K.AndrewsJ.JacksonE.PalmieriM. G.SenA.et al. (2015). “ Value 'blue carbon'?,” in Coastal Zones Ecosystem Services: From Science to Values and Decision Making (Dordrecht, Netherlands: Springer), 191–206. doi: 10.1007/978-3-319-17214-9_10

  • 20

    LyuH. M.SunW. J.ShenS. L.ArulrajahA. (2018). Flood risk assessment in metro systems of mega-cities using a GIS-based modeling approach. Sci. Total Environ.626, 1012–1025. doi: 10.1016/j.scitotenv.2018.01.138

  • 21

    MaJ.WuZ.GuoM. (2025). Trade-offs between economic gains and ecological goals: Impact of marine ecological policy on marine fisheries efficiency. Fish. Manage. Ecol.33, 142–157. doi: 10.1111/fme.70014

  • 22

    MacreadieP. I.SerranoO.DuarteC. M.LovelockC. E. (2026). Priority questions for the next decade of blue carbon science. Nat. Ecol. Evol.10, 751–764. doi: 10.1038/s41559-026-03020-6

  • 23

    McCookL. J.XuJ.LiaoB. G.YeY.QiuJ. (2025). Marine ecosystem services and natural capital in China: Opportunities for sustainability and policy reform. PNAS Nexus4, pgaf110. doi: 10.1093/pnasnexus/pgaf110

  • 24

    MitevaD. A.MurrayB. C.PattanayakS. K. (2015). Do protected areas reduce blue carbon emissions? A quasi-experimental evaluation of mangroves in Indonesia. Ecol. Econ.119, 127–135. doi: 10.1016/j.ecolecon.2015.08.005

  • 25

    OnrizalArifantiV.Kiswanto (2026). Credible MRV as the backbone of blue carbon governance: Lessons from Indonesia's mangroves. For. Policy Econ.190, 103877. doi: 10.1016/j.forpol.2026.103877

  • 26

    OstromE. (1990). Governing the Commons: The Evolution of Institutions for Collective Action (Cambridge, UK: Cambridge University Press).

  • 27

    OstromE. (2009). A general framework for analyzing sustainability of social-ecological systems. Science325, 419–422. doi: 10.1126/science.1172133

  • 28

    PendletonL.MurrayB. C.GordonD.CooleyD.VeghT. (2014). Valuing Ecosystem Services: Methodological Issues and Case Studies (Cheltenham, UK: Edward Elgar Publishing), 361–377.

  • 29

    PurushothamM. K.ThompsonB. S. (2026). Incorporating blue carbon into climate change mitigation policies: Multi-level governance challenges for carbon credits and NDCs. Ecol. Econ.240, 108800. doi: 10.1016/j.ecolecon.2025.108800

  • 30

    RenJ.HanG.GaoX.ZhangS. (2026). Water quality assessment in a megacity river: Water chemical analysis integrated with AHP-Entropy weighted fuzzy comprehensive evaluation model. J. Environ. Sci.162, 121–130. doi: 10.1016/j.jes.2025.07.060

  • 31

    RodriguesV. L. A.SilvaT. S. F.NascimentoM. T.AlvesL. F. (2026). Impacts of coastal dynamics on blue carbon ecosystems: Assessing carbon stock changes in Amazonian mangroves over 38 years. For. Ecol. Manage.608, 123608. doi: 10.1016/j.foreco.2026.123608

  • 32

    SaatyT. L. (1980). The Analytic Hierarchy Process (New York, NY, USA: McGraw-Hill).

  • 33

    SaatyT. L. (1988). Decision Making for Leaders: The Analytic Hierarchy Process for Decisions in a Complex World (Pittsburgh, PA, USA: RWS Publications).

  • 34

    SalamiH.OkparaK.ChoochuayC.KuaananT.TaiwoH.MohammedM. (2026). Optimizing solar photovoltaic farm locations in Kogi State, Nigeria using an integrated ANFIS-FAHP and GIS-MCDM framework. Energy344, 140030. doi: 10.1016/j.energy.2026.140030

  • 35

    SantosT. (2025). Integrating marine spatial planning into blue economy policies in Latin American countries. Mar. Policy181, 106856. doi: 10.1016/j.marpol.2025.106856

  • 36

    SchmitzO. J.LavertyS. M.SvenningJ. C.CromsigtJ. P. G. M. (2026). Integrating megafauna into blue carbon strategies: Dugongs could enhance seagrass carbon storage. Front. Mar. Sci.13, 1816090. doi: 10.3389/fmars.2026.1816090

  • 37

    ShenC. Y.LiuH.ZhangM. (2025). Analysis of China's marine economic policy: Regional policy evolution and blue economy transition. Front. Mar. Sci.12, 1553962. doi: 10.3389/fmars.2025.1553962

  • 38

    ThomasS. (2014). Blue carbon: Knowledge gaps, critical issues, and novel approaches. Ecol. Econ.107, 22–38. doi: 10.1016/j.ecolecon.2014.07.028

  • 39

    UddinS.Al-GhadbanA. N.FowlerS. W.Al-AnsariE. M. S. (2026). Blue carbon in the Persian Gulf: Evidence of phytoplankton contribution to carbon sequestration. Sustainability18, 1102. doi: 10.3390/su18021102

  • 40

    VoinovA.JenniK.GrayS.KolaganiN.GlynnP. D.BommelP.et al. (2018). Tools and methods in participatory modeling: Selecting the right tool for the job. Environ. Modell. Software109, 232–255. doi: 10.1016/j.envsoft.2018.08.028

  • 41

    WahshehH. A. M.Al-ZoubiA. M.AlrawabdehW.IsmailM. M. (2026). Securing blue carbon accounting: A cryptographic framework for coastal ecosystem monitoring. Int. J. Advanced Comput. Sci. Appl.17, 700–707. doi: 10.14569/ijacsa.2026.0170269

  • 42

    WanC.YuanJ.CaoD.WangT.NgA. K. Y. (2024). A fuzzy evidential reasoning-based model for evaluating resilience of ports to typhoons. Transp. Res. Part D Transp. Environ.133, 104228. doi: 10.1016/j.trd.2024.104228

  • 43

    WangM.LiuX.ChenJ.ZhangH.LiX. (2026a). Safeguarding seagrass blue carbon sinks: An integrated modeling approach for the Yellow Sea and Bohai Sea. Mar. pollut. Bull.231, 119928. doi: 10.1016/j.marpolbul.2026.119928

  • 44

    WangY. P.ZhangR.LiuD.ChenS.LiH. (2026b). Unvegetated tidal flats: A critical yet vulnerable coastal blue carbon pool. Global Change Biol.32, e70695. doi: 10.1111/gcb.70695

  • 45

    WatróbskiJ.JankowskiJ.ZiembaP.KarczmarczykA.ZiołoM. (2019). Generalised framework for multi-criteria method selection. Omega86, 107–124. doi: 10.1016/j.omega.2018.07.004

  • 46

    WuX. L.HuF. (2020). Analysis of ecological carrying capacity using a fuzzy comprehensive evaluation method. Ecol. Indic.113, 106243. doi: 10.1016/j.ecolind.2020.106243

  • 47

    XiaoL. L.ZhangJ.ChenY.LiM.WangF. (2026). Microbial responses to warming reduce deep blue carbon storage. Global Change Biol.32, e70883. doi: 10.1111/gcb.70883

  • 48

    YagerR. R. (1981). A procedure for ordering fuzzy subsets of the unit interval. Inf. Sci.24, 143–161. doi: 10.1016/0020-0255(81)90017-7

  • 49

    YagerR. R. (2003). Induced aggregation operators. Fuzzy Sets Syst.137, 59–69. doi: 10.1016/S0165-0114(02)00432-3

  • 50

    ZhangX. C.ZhaoY.LinJ.ChenW. (2026). Governing blue carbon ecosystems: A transnational environmental law perspective. Front. Mar. Sci.13, 1754665. doi: 10.3389/fmars.2026.1754665

  • 51

    ZhengY. H.LiX.WangS.ZhangM.ChenJ. (2026). Global blue carbon losses from salt marshes exceed restoration gains. Nat. Commun.17, 3744. doi: 10.1038/s41467-026-70158-z

Appendix

The dimensions and criteria in Appendix Table 1 were derived through the Step One literature search and the NGT procedure documented in Sections 3.1 and 3.2, and each dimension corresponds to an established stream already reviewed above: Maritime Politics draws on the maritime jurisdiction and dispute-settlement literature (Zhang et al., 2026); Maritime Economics draws on blue economy and fishery externality research (Ma et al., 2025; Santos, 2025; Shen et al., 2025); Marine Industry draws on marine industrial and supply-chain policy research (Du et al., 2026; Shen et al., 2025); and Marine Culture and Environment, including its blue carbon-calibrated sub-dimensions, draws on the blue carbon ecosystem services and pollution literatures reviewed in Section 2.2 (Zheng et al., 2026; Schmitz et al., 2026; Liu et al., 2026; Adams et al., 2026). Conceptually adjacent criteria within this last dimension, for example marine environmental pollution, ecosystem pollution, biodiversity loss, habitat destruction, and carbon sink degradation, were retained as analytically distinct rather than merged because the FAHP pairwise comparison procedure weighs each criterion against every other criterion directly, so that a criterion’s weight reflects its judged severity relative to the full 36-criterion set rather than an additive count of related items; The Limitations and Directions for Future Research subsection in the Conclusion discusses the limits of this safeguard against double counting.

Appendix Table 1

DimensionSub-dimensionCriteriaBlue carbon relevance
Maritime Politics (MP0)Global maritime strategy (MP0,1)Global strategic CDRMP; geopolitical CDRMP; national maritime security CDRMPGeopolitical contestation over territories containing blue carbon habitats; strategic resource use that displaces carbon-sequestering ecosystems
Marine governance organizations (MP0,2)Inter-country organization disputes; inter-agency disputes; local governance disputesFragmented institutional authority impedes coordinated protection of transboundary blue carbon systems
Marine management agreements (MP0,3)International agreement disputes; domestic agreement disputes; local agreement disputesContested interpretation of agreements governing marine protected areas and carbon sink conservation
Maritime Economics (ME0)Marine economic growth (ME0,1)Long-term growth CDRMP; short-term growth CDRMP; economic efficiency CDRMPGrowth policies prioritizing extraction over conservation of blue carbon ecosystem services
Marine economic competition (ME0,2)Supply-side competition; demand-side competition; land-sea economic competitionCoastal development competition that converts blue carbon habitats to extractive or industrial uses
Marine economic resource allocation (ME0,3)Resource scarcity CDRMP; market disruption CDRMP; technological disparity CDRMPResource scarcity in shared coastal zones that reduces trophic integrity and amplifies blue carbon stock vulnerability
Marine Industry (MI0)Marine industrial growth (MI0,1)Primary industry CDRMP; secondary industry CDRMP; tertiary industry CDRMPIndustrial expansion into coastal zones that physically disrupts the sedimentary substrates of blue carbon sequestration
Marine industrial structure (MI0,2)Industrial structural CDRMP; industrial policy cross-CDRMP; industrial environmental CDRMPStructural policy imbalances that fail to internalize blue carbon ecosystem service values
Marine industrial supply chain (MI0,3)Supply chain disruption; input cost volatility; market access restrictionContamination or disruption of marine supply chains carrying inputs and outputs from blue carbon-adjacent industries
Marine Culture and Environment (MC0)Marine cultural heritage (MC0,1)Marine life culture CDRMP; traditional culture CDRMP; community culture CDRMPCultural practices of coastal communities dependent on the provisioning and cultural services of mangroves, seagrass beds, and salt marshes
Marine environmental pollution (MC0,2)Environmental pollution CDRMP; ecosystem pollution CDRMP; coastal erosion and disaster CDRMPCore blue carbon indicators: radioactive and chemical contamination of blue carbon sediments; ecosystem-level pollution disrupting carbon fixation; coastal erosion permanently removing blue carbon stock
Marine ecological environment (MC0,3)Biodiversity loss CDRMP; habitat destruction CDRMP; carbon sink degradation CDRMPCore blue carbon indicators: loss of species supporting sedimentary carbon accumulation; destruction of mangrove, seagrass, and salt marsh habitats; direct reduction of carbon sequestration capacity

CDRMP evaluation dimensions, sub-dimensions, and blue carbon-calibrated criteria.

Summary

Keywords

blue carbon governance, cross-domain risk, ecological irreversibility, ecosystem services, fuzzy analytic hierarchy process, marine environmental policy, nuclear wastewater discharge, polycentric governance

Citation

Huang W (2026) Cross-domain risk governance of transboundary marine policies for blue carbon ecosystem protection: a fuzzy AHP evaluation framework applied to the nuclear wastewater discharge case. Front. Mar. Sci. 13:1928781. doi: 10.3389/fmars.2026.1928781

Received

05 July 2026

Revised

20 September 2026

Accepted

07 October 2026

Published

21 October 2026

Volume

13 - 2026

Edited by

Shih-Ming Kao, National Sun Yat-sen University, Taiwan

Reviewed by

Yuping Wu, Henan Polytechnic University, China

HamidReza Talaie, Arak University, Iran

Updates

Copyright

*Correspondence: WeiLun Huang,

Disclaimer

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.

Outline

Figures

Cite article

Copy to clipboard


Export citation file


Share article

Article metrics