Mitigation or adaptation? Unveiling the temporal scenario-based priorities of port climate resilience strategies using fuzzy Dombi-Bonferroni best-worst method

With the emergence of the Anthropocene concept, humanity has increasingly recognized that climate change, mainly caused by human emissions, is a major challenge today. Maritime transportation handles more than 80% of international trade (UNCTAD, 2022). Ports play a crucial role by serving as vital links between sea transport and multimodal transport, both of which contribute substantially to carbon emissions (Azarkamand et al., 2020). Around 3.9% of the world's carbon dioxide emissions are contributed by the shipping industry (Budiyanto et al., 2022). Meanwhile, due to their coastal locations, ports are also very vulnerable to climate change (O'Keeffe et al., 2020). The operational disruptions that climate change brings to ports include extreme wind, temperature, and wave heights, as well as physical damage caused by tropical cyclones, coastal flooding, etc. Many ports face operational disruptions due to extreme weather, resulting in costly downtime (Verschuur et al., 2023). Port closures can lead to cargo congestion at ports, vessel queues, backlogs at warehousing and transshipment facilities, and manufacturing shutdowns. A 2017 survey by the United Nations Conference on Trade and Development (UNCTAD) revealed that respondents have been impacted by extreme events, resulting in delays (60%), operational disruptions (76%), or physical damage (45%) (UNCTAD, 2017). Undoubtedly, port disruptions have significant adverse effects on supply and demand, as ports serve as nodes in the global maritime supply chain network.

Disruptive events such as production accidents, natural disasters, and changes in the international trade environment continuously affect the operations and infrastructure of ports worldwide, underscoring the necessity to develop robust resilience plans to ensure continuous operations during crises (González-Solano et al., 2024). Climate change-related risks constitute a key component of port operational risks, prompting port stakeholders and governments to advance effective climate risk responses to mitigate risks, reduce disruptions, and enhance port resilience, thereby minimizing or avoiding vulnerability to climate risks (Wang et al., 2024). Therefore, strengthening the climate resilience of ports and other critical transport infrastructure is essential for implementing the 2030 Agenda for Sustainable Development and advancing the objectives of international agreements such as the Paris Agreement and the Sendai Framework for Disaster Risk Reduction (UNCTAD, 2020).

To improve port resilience to climate change, ports should not only mitigate climate change by reducing greenhouse gas (GHG) emissions in the long term but also increase their capacity to withstand current climate risks through infrastructure and operational adjustments in the short term. These strategies are defined as Climate Change Mitigation (CCM) and Climate Change Adaptation (CCA), respectively (Ng et al., 2019). While existing studies typically focus solely on either mitigation or adaptation (León-Mateos et al., 2021; Puig et al., 2024; Song et al., 2025), practical port decision-making often involves trade-offs between these two strategies due to limited resources. Therefore, prioritizing these two strategies is necessary to facilitate implementation and improve effectiveness (Jiang et al., 2020).

However, a critical yet often overlooked aspect of port climate strategy lies in how the nature of climate actions evolves over time. For instance, short-term decisions are frequently driven by immediate regulatory compliance such as the International Maritime Organization's (IMO) EEXI (Energy Efficiency eXisting ship Index)/CII (Carbon Intensity Indicator) ratings, while long-term strategies require proactive capital investment to achieve deep decarbonization and structural resilience. Consequently, a prioritization framework that evaluates all strategies under a single time horizon may conflate short-term operational priorities with long-term strategic investment needs, increasing the risk of temporal myopia. To address this, this study disaggregates CCM and CCA strategies through short- and long-term time-horizon perspectives. Since short- and long-term strategies differ in their strategic focus, this distinction enables more nuanced insight into how port managers seek to balance current survival and future viability.

Therefore, this study proposes a Multi-criteria Decision Making (MCDM) framework that enables port authorities to make effective short- and long-term strategic choices to enhance port climate resilience (PCR) when evaluating CCM and CCA strategies. Because the prioritization process relies on uncertain expert judgments, a fuzzy MCDM approach is appropriate for this problem. Accordingly, the Fuzzy Dombi-Bonferroni Mean Operators-based Best-Worst Method (FDB-BWM) is employed to quantify the relative importance of port climate resilience strategies (PCRSs) and facilitate the prioritization process. The Dombi and Bonferroni mean operators consider the interrelationships among attributes, thereby mitigating the influence of outliers or extreme data. The combination of the Best-Worst Method (BWM) with Dombi-Bonferroni mean operators in a fuzzy context provides a flexible aggregation framework capable of accommodating interrelationships among criteria through adjustable parameters for decision-making in real-world assessments (Yaran Ögel et al., 2023).

Thus, our contributions are twofold:

1. Developing a novel prioritization framework that incorporates a time-horizon perspective (short-term vs long-term) into the comparative evaluation of CCM and CCA strategies, thereby revealing their relative importance in enhancing port climate resilience.

2. Providing empirical evidence that reveals the temporal prioritization of mitigation and adaptation strategies based on expert evaluations from the Republic of Korea, thereby offering practical guidance for port climate resilience planning.

Holling (1973) originally introduced the concept of resilience in the field of ecology. Generally, resilience refers to the ability of the system to reduce the likelihood of a shock, absorb an unexpected disturbance, and recover after a disruption (Bruneau et al., 2003; Christopher and Peck, 2004; Gaonkar and Viswanadham, 2007).

Hence, port resilience can be understood as the ability of a system to bounce back to the desired operating level within a reasonable period after disruptive events, such as natural disasters, economic downturns, geopolitical shocks, pandemics, technological transformations, and environmental regulations (Hossain et al., 2019; Wendler-Bosco and Nicholson, 2020). Three types of factors influencing port resilience are identified: natural disasters (Christodoulou et al., 2019), production accidents (Gu and Liu, 2025), and fluctuations in the international trade environment (Liu et al., 2023b). Natural disasters such as storms are expected to worsen due to climate change (Becker and Caldwell, 2015). Research on PCR indicates a clear evolution, particularly in terms of methodological approaches.

Early studies mainly relied on qualitative and exploratory methods. Gharehgozli et al. (2017), Becker (2017), and Mutombo et al. (2017) adopted case studies and conceptual frameworks to explore how ports respond to climate-related disruptions, with an emphasis on institutional settings, management practices, and decision-making processes. Semi-structured interviews are also widely used to identify barriers and enabling factors for resilience planning (Mclean and Becker, 2021). More recent studies have shifted toward structured expert-based and multi-criteria decision-making methods. The Delphi method, Analytic Hierarchy Process (AHP), and evidential reasoning are frequently applied to integrate expert knowledge and handle uncertainty in resilience assessment (León-Mateos et al., 2021; Wan et al., 2024). A smaller but growing body of literature adopts quantitative and performance-based methods, such as network analysis and efficiency evaluation, to examine how ports perform under climate stress (Poo et al., 2024).

According to prior literature related to climate change resilience, there are two dimensions: inherent resilience and adaptive resilience (Naderpajouh et al., 2018).

Inherent resilience is sometimes called mitigation resilience. The CCM strategy is a proactive measure intended to reduce GHG emissions and promote environmental stewardship. The increase in GHGs blocks the path of heat from the Earth's surface into outer space, which leads to an increase in the Earth's surface temperature. The greenhouse effect is one of the main causes of global climate change (Xia, 2021). In the long term, it is necessary to take certain measures to reduce pollutant gas emissions to prevent them from having an impact on climate change. Therefore, the CCM strategy has been proposed to slow the pace of global climate change. Puig et al. (2024) identified seven key strategies for GHG reduction in ports. Xia (2021) concluded the CCM strategies of China's ports, for example, shore power, electric RTGs (Rubber-Tired Container Gantry Cranes), alternative fuels, and multimodal transport. EESI (2022) proposes CCM strategies in ports, such as electrification, shore power, truck replacement programs.

Adaptive resilience refers to the climate change adaptation (CCA) strategies, which are reactive strategies implemented after the adverse impacts of climate change have materialized. Specifically, CCA is used to reduce vulnerability and minimize risks associated with the harmful impacts of climate change, such as extreme weather events. León-Mateos et al. (2021) adapted Summers et al. (2017)'s five dimensions of CRSI (Climate change screening index) to assess port adaptation resilience: governance, society, infrastructure and facilities, operational environment, and risk management. Puig et al. (2024) also summarized CCA strategies in ports, including risk assessments and vulnerability studies, port personnel training etc. Within the infrastructure dimension specifically, recent studies have quantified the physical vulnerability of port infrastructure to climate hazards. Mutombo et al. (2017) proposed a framework for assessing port infrastructure resilience and prioritizing adaptive solutions, providing methodological precedent for infrastructure-focused adaptation assessment. Verschuur et al. (2023) quantified multi-hazard risks to global port infrastructure and their cascading trade and logistics losses, underscoring the urgency of long-term structural adaptation. Wang et al. (2024) examined typhoon-related adaptation planning for Chinese container ports, highlighting infrastructure interventions such as breakwater reinforcement and terminal elevation. These studies underscore the continued importance of long-term structural adaptation in safeguarding port infrastructure against intensifying climate hazards.

Although CCM and CCA differ in their climate change objectives, they are proposed to enhance seaport climate resilience in the long term and short term, respectively.

Multi-Criteria Decision Making (MCDM) is a pivotal branch of decision theory designed to facilitate rational decision-making when multiple conflicting evaluation criteria exist (Stewart, 1992; Guo and Zhao, 2017). To date, numerous MCDM methodologies have been established, including AHP, TOPSIS (Technique for Order Performance by Similarity to Ideal Solution), ANP (Analytic Network Process), VIKOR (Vise Kriterijumsa Optimizacija I Kompromisno Resenje), and SWARA (step-wise weight assessment ratio analysis) (Gu et al., 2023; Peykani et al., 2026). Among these, the Best-Worst Method (BWM) which was introduced by Rezaei (2015), has gained prominence. BWM relies on pairwise comparisons between the identified “best” and “worst” criteria and the other criteria. Compared to other MCDM methods, BWM requires fewer comparison data by using only reference comparisons, thereby remedying the inconsistency characteristic of standard pairwise comparisons (Rezaei, 2015). Currently, BWM has been widely applied in supply chain management (Badri Ahmadi et al., 2017; Malek and Desai, 2019; Kumar et al., 2021). However, the crisp 1–9 scale of classical BWM is often insufficient for capturing the ambiguity inherent in qualitative human judgment. Therefore, Guo and Zhao (2017) integrated fuzzy numbers into BWM to create Fuzzy BWM (FBWM). This approach has since been applied to supplier selection (Wu et al., 2019; Fathi and Hassannayebi, 2025), healthcare waste treatment (Goldani et al., 2023), product-service system requirements (Chen et al., 2020), etc.

Aggregation operators are critical tools for information fusion in MCDM contexts (Pamucar et al., 2020). Yet, basic FBWM often relies on assumptions of linearity and independence during aggregation and weight generation. The generalized Dombi operator class addresses the disadvantages of traditional min-max operators in fuzzy environments; its general parameters for T-norms and T-conorms make the aggregation process more flexible (Pamucar et al., 2020). Furthermore, the Bonferroni mean (BM) operator allows for the representation of interconnections between elements, fusing them into a unique score function (Bonferroni, 1950). Pamucar et al. (2020) were the first to combine Dombi and Bonferroni operators (Dombi-Bonferroni), enabling more objective expert evaluations within a subjective environment. Subsequently, this hybrid method (Dombi-Bonferroni based FBWM) has been used to prioritize drivers of retail food waste (Yaran Ögel et al., 2023) and factors affecting airport selection (Tanrıverdi et al., 2022). The Dombi-Bonferroni aggregation handles fuzzy preferences within the Best-Worst framework in an adjustable, non-linear manner that reflects expert interactions, ensuring results align more closely with real-world decision logic (Pamucar et al., 2020; Yaran Ögel et al., 2023). These properties make FDB-BWM a contextually appropriate methodological choice for this study, as it provides a flexible aggregation framework for handling uncertain expert judgments in a heterogeneous decision-making setting.

Based on this review, two significant research gaps emerge. First, previous research has predominantly prioritized CCM. While some scholars acknowledge CCA, few studies in the port sector consider both strategies (Jiang et al., 2020), and virtually none investigate the relative priority between CCM and CCA, let alone consider time horizons. Second, the FBWM, particularly when combined with Dombi-Bonferroni aggregators, has not yet been widely applied to the port field. To bridge these gaps, this study employs a Fuzzy Dombi-Bonferroni-based BWM (FDB-BWM) to explore the prioritization of PCRSs (CCM and CCA).

This study aims to effectively prioritize PCRSs using FDB-BWM. Hence, this section briefly introduces triangular fuzzy numbers (TFNs), the Dombi operator, and the Bonferroni mean (BM) operator.

Uncertainty-driven practical problems can be effectively addressed through fuzzy set theory (Zadeh, 1965). TFN is one of the most preferred fuzzy numbers (Tanrıverdi et al., 2022). Since the study is applied in a fuzzy context, the following definitions of TFNs are relevant.

For the basic operational laws of two TFNs, please refer to Ecer (2015) work.

1. Dombi operator

Where $ρ>0$ and $(p,q)∈[0,1]$⁠.

1. Bonferroni mean operator

1. Dombi-Bonferroni mean operator

Based on the expressions of the Dombi and Bonferroni operators, the TFN Dombi-Bonferroni mean operator (FDB) is defined as follows.

Where $f(φi~)=(f(φil),f(φim),f(φiu))=(φil∑i=1nφil,φim∑i=1nφim,φiu∑i=1nφiu)$ represents a fuzzy function. For the detailed calculation process, please refer to Pamucar et al. (2020).

Following these preliminaries, the steps of the FDB-BWM methodology are shown in Figure 1. The detailed descriptions are as follows.

• Step 1: Conduct a comprehensive literature review and establish criteria pools to identify potential evaluation criteria and sub-criteria.

• Step 2: Finalize the criteria and sub-criteria based on experts' opinions.

• Step 3: Identify the best (most important) criterion $cB$ and worst (least important) criterion $cW$ within each set of main criteria and sub-criteria.

• Step 4: Determine the fuzzy preferences of the best criterion over all criteria based on fuzzy judgment scale presented in Table 1. The resulting fuzzy Best-to-Others vector is expressed as:

Where $a~Bj$ denotes the fuzzy preference of the best criterion $cB$ over criterion $j$⁠, $j=1,2,…,n$⁠, and $a~BB=(1,1,1)$⁠.

Similarly, perform pairwise comparisons to determine the preference of all other criteria over the worst criterion. Then the Others-to-Worst vector is obtained as:

• Step 5: Determine the optimal fuzzy weights $w~=(w~1*,w~2*,…,w~n*)$⁠. The optimal fuzzy weights should perfectly satisfy the relationships $w~Bw~j=a~Bj$ and $w~jw~W=a~jW$ for every criterion. To satisfy these conditions, a solution should be found in which the maximum absolute divergences $|w~Bw~j−a~Bj|$ and $|w~jw~W−a~jW|$ for all $j$ are minimized. Then, simplify the function by transforming it into a nonlinearly constrained optimization problem as follows (Guo and Zhao, 2017).

Where $w~B=(lBw,mBw,uBw)$⁠, $w~j=(ljw,mjw,ujw)$⁠, $w~W=(lWw,mWw,uWw)$⁠, $a~Bj=(lBj,mBj,uBj)$⁠, $a~jW=(ljW,mjW,ujW)$⁠, $ξ~*=(k*,k*,k*)$⁠. By solving Eq. (9), the optimal fuzzy weights can be obtained.

• Step 6: Examine the consistency ratio (CR) for pairwise comparison.

• Step 7: Aggregate the fuzzy weights of main criteria and sub-criteria from different experts with FDB operators (Eq. (6)). Notably, to synthesize the evaluations of multiple experts, this study employs the Aggregation of Individual Priorities (AIP) (Forman and Peniwati, 1998), wherein individual fuzzy weights are first derived and subsequently aggregated. In alignment with established applications of the triangular fuzzy Dombi-Bonferroni BWM methodology (e.g. Tanrıverdi et al., 2022; Yaran Ögel et al., 2023), the CR acts as a strict pre-condition for individual evaluation. By aggregating the derived fuzzy weight vectors $(l,m,n)$ rather than the raw relational matrices, the risk of violating transitivity is reduced. Given that the Dombi-Bonferroni operator adheres to the mathematical principles of boundedness and monotonicity (Yager, 2009), the aggregated outputs remain strictly within the bounds of the highly consistent individual judgments, thereby perfectly preserving the ordinal transitivity of the decision-making group without introducing logical contradictions.

• Step 8: Defuzzify the local weights using the GMIR method (Eq. (2)) to obtain the crisp weights.

• Step 9: Normalize the crisp weights for main criteria level and sub-criteria level.

• Step 10: Rank the criteria.

This study was conducted within the specific context of ports in the Republic of Korea (ROK). The experts' evaluations were obtained through linguistic variables (Bai et al., 2019; Gu et al., 2023). A total of nine experts working in the ROK, with relevant professional backgrounds across diverse organizations, participated in this research. The linguistic evaluation questionnaire distributed to these experts is included as Supplementary_material_1 to ensure transparency and reproducibility. Their experience ranged from 8 to 30 years. Experts were deliberately selected from government (port authority), industry, and academia to ensure applicability and consistency of the findings across a wide range of organizational and industrial contexts (Liu et al., 2023a). The profiles of these maritime experts are detailed in Table 2. Prior studies in MCDM and fuzzy decision frameworks suggest that a panel of 5–10 experts is sufficient to achieve stable and reliable evaluation results when experts are carefully selected and possess substantial domain knowledge (Gu et al., 2023; Tanrıverdi et al., 2022).

The research criteria were selected in two stages. In the first stage, initial sub-criteria pools for CCM and CCA are created through a comprehensive review of relevant literature, as summarized in Tables 3 and 4.

Candidate criteria are selected primarily according to the frequency of their citation in previous studies, as citation frequency reflects the extent to which a strategy has been recognized and discussed in the existing literature. In the second stage, the final selected criteria are validated by maritime experts to confirm their relevance and applicability to the ROK port context. Based on this expert validation, the final research criteria (Tables 5 and 6) are created. To capture the temporal heterogeneity of PCRSs, the main criteria are structured into four distinct dimensions: Short-term CCM (M-ST), Long-term CCM (M-LT), Short-term CCA (A-ST), and Long-term CCA (A-LT). The short-term horizon is defined as the coming decade (<10 years), focusing on immediate compliance and operational adjustments. In contrast, the long-term horizon (>20 years) focuses on avoiding stranded assets and securing future competitiveness under the IMO 2050 Net-Zero framework (IMO, 2023). Specifically, the strategic focus evolves over time: CCM shifts from immediate regulatory compliance to strategic energy transition, while CCA progresses from immediate disaster response to structural survival. Notably, while the nomenclature of these sub-criteria remains the same in the proposed framework, their focus differs depending on the time horizon. For instance, regarding M1 (emission reduction commitment and target setting), the short-term focus lies in establishing baselines to satisfy immediate regulatory mandates (e.g. IMO's EEXI and CII ratings), whereas the long-term focus shifts toward defining strategic pathways to achieve absolute carbon neutrality aligned with the IMO's 2050 Net-Zero ambition. To clearly differentiate these distinct temporal objectives within the evaluation framework, this study applies the prefixes “ST-” and “LT-” to the sub-criteria identifiers to denote short-term and long-term contexts, respectively. For instance, ST-M1 represents the sub-criterion under the short-term CCM framework, while LT-M1 refers to its counterpart in the long-term CCM framework.

Following the finalization of the main criteria and their sub-criteria, the nine invited experts performed linguistic evaluations in accordance with Steps 3 and 4 of the FBWM. The assessment data, including the selection of the best and worst criteria, along with the Best-to-Others (B-t-O) and Others-to-Worst (O-t-W) vectors are presented in Supplementary_material_2 (Table A. 1–Table A. 3). The nine experts differed in their selection of the best and worst criteria, which consequently influenced their preferences over the remaining criteria. For instance, Expert 1 identified Long-term CCM (M-LT) as the best criterion and Long-term CCA (A-LT) as the worst, whereas other experts had different perspectives.

Subsequently, the linguistic variables were converted into TFNs using Table 1. The optimal fuzzy weights for both the main criteria and sub-criteria are computed (see Table A. 4–A. 6 in Supplementary_material_2. For example, the triangular fuzzy weights for Expert 1's main criteria are shown below.

The consistency ratio (CR) for Expert 1's assessment of the main dimensions is calculated as 0.0517 using Eq. (10). Since this value is below the threshold of 0.1, the evaluation is deemed reliable. Similarly, the CR values for all other dimensions are confirmed to be less than 0.1 (Table 7), validating the consistency of the fuzzy pairwise comparisons across the study.

To aggregate the experts' evaluations, the FDB operator (Eq. (6)) is employed to calculate the aggregated fuzzy weights. In this process, the parameters $p$⁠, $q$⁠, and $ρ$ are set to 1 as the baseline. This setting is recommended by Pamucar et al. (2020) for practical decision-making applications, as it is intuitive and simple while still capturing the interrelationships between attributes without artificially emphasizing interaction effects. Finally, the aggregated fuzzy weights are defuzzified into crisp values using the GMIR method (Eq. (2)). These crisp weights are then normalized to determine the local and global rankings, as presented in Table 8.

Based on the results, the four main dimensions are ranked as M-LT (Long-term CCM), M-ST (Short-term CCM), A-ST (Short-term CCA), and A-LT (Long-term CCA), indicating a primary preference for long-term CCM strategies. Furthermore, ST-M1 and LT-M1 are the most significant factors within the M-ST and M-LT dimensions, respectively. Similarly, ST-A4 and LT-A1 emerge as the most important factors in the A-ST and A-LT dimensions. Overall, LT-M1 is identified as the most crucial factor in this study, followed by LT-M5 and LT-M2. Conversely, according to the global ranking, LT-A6, LT-A7, and ST-A7 are the least important factors.

The priority rankings of PCRSs derived from the aggregated expert evaluations are presented in Table 8. The results reveal a distinct hierarchical preference structure that characterizes the current strategic orientation of the ROK maritime sector.

From a strategic perspective, the results indicate that mitigation-oriented strategies (CCM) receive a higher overall priority than adaptation-oriented strategies (CCA). Long-term Mitigation (M-LT) emerges as the most critical dimension with a normalized weight of 0.32954, followed by Short-term Mitigation (M-ST, 0.23932). In contrast, adaptation dimensions received comparatively lower weights, with Short-term Adaptation (A-ST) at 0.22577 and Long-term Adaptation (A-LT) at 0.20537. From a temporal perspective, different priority patterns are observed for mitigation and adaptation strategies. In the mitigation domain, long-term strategies receive higher priority than short-term ones, reflecting a forward-looking strategic investment. Conversely, in the adaptation domain, short-term measures are preferred over long-term ones, which implies a “reactive” approach to physical climate risks that focuses on immediate preparedness rather than long-term infrastructure redesign. The priority given to CCM over CCA is theoretically consistent with the asymmetric regulatory pressures faced by port operators. Unlike CCA investments, which are generally not subject to equivalent binding regulatory mandates, CCM non-compliance carries immediate and quantifiable operational consequences including vessel trading restrictions and reputational penalties under EEXI/CII frameworks. This interpretation is further corroborated by Jiang et al. (2020), who stated that mitigation strategies tend to receive greater strategic emphasis among port stakeholders compared with adaptation.

At the sub-criteria level, a total of 28 indicators are evaluated, which further reinforces the prioritization of CCM as the top eight positions are exclusively occupied by mitigation strategies. LT-M1 (Emission Reduction Commitment and Target Setting: Net-zero pathways) ranks first globally with a weight of 0.06036. Additionally, Puig et al. (2024)'s study indicated that all surveyed ports had established GHG reduction targets, which highlights the importance of LT-M1. This reflects the fact that without a clear long-term carbon-neutrality roadmap, today's infrastructure investments risk becoming technologically obsolete. Since the IMO mandates absolute zero emissions by 2050, port managers must prioritize long-term pathways to avoid the financial trap of stranded assets. LT-M1 is followed by LT-M5 (Alternative Fuel Supply: Future Fuels) and LT-M2 (Regulatory Emission Control Instruments: Zero Emission Zones), underscoring that the tangible transition to hydrogen or ammonia is the core technical challenge (Jesus et al., 2024). The prominence of LT-M5 is independently corroborated by Puig et al. (2024), who identified alternative fuel supply as one of the most strategically critical GHG reduction levers for mega ports globally. Furthermore, regardless of time horizon, M1 (Emission Reduction Target Setting) consistently carries the highest weight, indicating that clear decarbonization targets form the foundation of CCM strategies in ROK ports. Conversely, adaptation strategies rank relatively lower in the hierarchy. The highest-ranked adaptation factor ST-A4 (Contingency Planning: Business Continuity) appears in the global ranking at 9th with a weight of 0.03880. Given the high cost of large-scale physical infrastructure redesign, operationally actionable short-term measures offer more immediate and tangible risk reduction returns under current climate volatility (Mclean and Becker, 2020). This result suggests that, under current resource and institutional constraints, ROK ports place relatively greater emphasis on maintaining operational continuity during extreme events over committing to long-term structural resilience investments.

These empirical findings highlight two critical insights regarding the state of port resilience in the ROK. First, the results reflect a carbon-first orientation in PCRSs. Climate resilience is primarily framed around carbon neutrality objectives, with mitigation measures serving as the central pathway for aligning port development with national and international climate agendas. This orientation follows global decarbonization trends. It also highlights the strategic role of ports in long-term emission reduction through infrastructure investment, energy transition, and regulatory coordination. Second, this study also reveals a clear temporal distinction between CCM and CCA strategies. CCM is mainly linked to long-term strategic planning, which relies on clear targets and established technology pathways. In contrast, CCA focuses on short-term operational preparedness. CCA supports flexible responses and operational continuity during climate-related disruptions. This temporal distinction does not suggest a hierarchy between mitigation and adaptation. Instead, it highlights their functional complementarity. CCM offers long-term strategic direction through a clearly defined carbon-neutrality roadmap. CCA supports short-term operational resilience. It emphasizes organizational capacity building and emergency preparedness. These efforts create a foundation for future decisions on large-scale structural re-engineering, such as terminal height elevation (LT-A1).

First, to examine the methodological consequences of incorporating the FDB operator, a comparative analysis is conducted between the proposed FDB-BWM and the benchmark FBWM approach. As illustrated in Figure 2, the overall rankings obtained from the two methods are generally consistent, particularly at the extremes of the ranking spectrum. Both methods identify LT-M1 and LT-M5 as the top strategic priorities and consistently place long-term adaptation criteria (LT-A6, LT-A7) at the bottom. However, notable differences emerge in the mid-ranking zone due to the FDB operator's sensitivity to inter-expert interactions, which produces a more nuanced differentiation among mid-ranking criteria.

The high level of agreement between the two models is expected under the baseline parameter setting $p=q=ρ=1$⁠, which is recommended as a practical and intuitive setting. Under this setting, the Dombi-Bonferroni operator provides a flexible aggregation framework while producing rankings comparable to those obtained using the FBWM. Thus, the comparison primarily indicates the consistency of the prioritization results under the adopted baseline setting in this study. Meanwhile, FDB-BWM retains the flexibility to represent different preference structures through parameter settings.

Second, to further examine the robustness of the FDB-BWM results, this study tests how changes in the three key parameters (⁠$p,q,ρ$⁠) in the FDB operator (Eq. (6)) affect the ranking results. These parameters represent different weighting emphases in the fuzzy decision framework. To test the robustness of the results, a total of 57 experiments were conducted under four scenarios.

1. Scenario 1: Fixing $p=q=1$ while varying $ρ$ within the range of $[2,15]$⁠.

2. Scenario 2: Fixing $p=ρ=1$ while varying $q$ within the range of $[2,15]$⁠.

3. Scenario 3: Fixing $q=ρ=1$ while varying $p$ within the range of $[2,15]$⁠.

4. Scenario 4: Simultaneously varying all three parameters such that $p=q=ρ∈[2,15]$⁠.

Figure 3 presents the ranking distributions of the 28 sub-criteria across all scenarios using box plots. Each box summarizes the variability of rankings under different parameter settings. The results demonstrate a high degree of robustness, particularly at the extremes. The top-ranked criteria (e.g. LT-M1, LT-M2) and the bottom-ranked criteria (e.g. LT-A6, LT-A7) show minimal deviation. Although some criteria in the middle rankings exhibit moderate dispersion, their relative positions remain largely consistent across scenarios.

To quantitatively validate the consistency of the results, Spearman's Rank Correlation Coefficient is calculated to compare the rankings derived from the sensitivity analysis against the original base case (⁠$p=q=ρ=1$⁠) (Spearman, 1904). The analysis yields an average correlation coefficient of 0.8705, which exceeds the commonly accepted threshold for a strong correlation (>0.8) (Sedgwick, 2014). This result reveals that the proposed model is robust and the final strategic hierarchy remains reliable regardless of parameter changes.

This study proposed a novel decision-making framework for prioritizing port climate resilience strategies (PCRSs) by explicitly distinguishing between climate change mitigation and adaptation across short-term and long-term horizons. This study employs the FDB-BWM as a flexible aggregation framework to prioritize PCRSs based on the judgments of experts from the ROK maritime sector. The empirical results lead to three pivotal conclusions:

First, CCM receives higher priority than CCA overall, although the difference is much more pronounced in the long-term dimension than in the short-term dimension. This indicates that stakeholders view decarbonization as the primary driver of resilience, reflecting strong strategic alignment with international decarbonization agendas and national carbon-neutrality commitments. Second, a general strategic differentiation is observed regarding time horizons. In the mitigation domain, stakeholders prioritize a proactive and long-term strategy as Long-term Mitigation (M-LT) ranks highest. However, a reactive and short-term approach prevails in the adaptation domain, with Short-term Adaptation (A-ST) ranking higher than Long-term Adaptation (A-LT). This reflects a functional complementarity: mitigation provides long-term strategic direction, while adaptation ensures short-term operational continuity under increasing climate volatility. Third, at the operational level, LT-M1 (Emission Reduction Commitment and Target Setting: Net-zero pathways) and LT-M5 (Alternative Fuels: Future Fuels) are identified as the top two global drivers. This suggests that defining a clear policy trajectory and securing future energy supplies are considered key prerequisites for enhancing port climate resilience. Meanwhile, ST-A4 (Contingency Planning: Business Continuity) ranks as the top adaptation strategy, which highlights the critical importance of maintaining operational stability and minimizing downtime during extreme weather events.

The findings offer several practical implications for port authorities and policymakers. First, the dominance of long-term mitigation highlights the need for ports to move beyond incremental emission reduction measures and adopt explicit net-zero pathways. Clear long-term targets help guide infrastructure investment and energy transition decisions, and facilitate coordination with shipping lines and fuel suppliers. Second, the strong emphasis on short-term adaptation highlights the need for operational readiness under growing climate extremes. Ports should continue to invest in contingency planning, emergency response protocols, and early warning systems. These measures are relatively low-cost and can deliver immediate benefits for operational continuity. Third, the relatively lower priority of long-term adaptation suggests a risk of insufficient investment in structural resilience. Policymakers should therefore promote a gradual shift from reactive measures toward long-term adaptive planning to reduce future exposure to climate risks. Finally, port authorities operate under dynamic resource constraints, where immediate operational pressures, such as high Yard Occupancy Ratios (YOR) or surging throughput, may temporarily compete with PCRS investments for budgetary resources. Therefore, the priority rankings derived from this study should be interpreted as a long-term strategic reference rather than as a rigid capital-allocation rule. Port authorities are advised to distinguish between operational expenditure (OPEX) for short-term operational needs and long-term capital expenditure (CAPEX) for climate resilience, treating them as complementary budget streams. During periods of operational stress, OPEX should be prioritized to maintain operational continuity, while planning activities for long-term strategies such as LT-M5, should continue through low-cost actions, including feasibility studies, regulatory pre-approvals, and green financing applications. In this way, port authorities can preserve strategic flexibility while meeting immediate operational demands.

However, this study focuses exclusively on ROK ports. While the expert panel is diverse and experienced, the relatively small sample size (nine experts) and the country-specific context may limit the generalizability of the findings. Future studies could apply the framework to ports in different regions to explore cross-country heterogeneity in PCRSs. Methodologically, developing new quantitative frameworks to evaluate post-aggregation consistency in heterogeneous group decision-making scenarios remains a meaningful direction for future research.

The supplementary material for this article can be found online.