Driven by significant concerns about the risk of ship collisions within the maritime industry, this paper aims to propose an enhanced and intelligent method for assessing collision risks utilizing deep learning techniques. A novel research framework is introduced for assessing potential collision risk of ships entering monitored waterways by leveraging the Automatic Identification System (AIS) data, deep learning methods, and expert knowledge. Specifically, this framework includes 1) the selection of multifaceted ship collision assessment indicators. 2) The development of an intelligent model for risk assessment under limited data. A deep learning model, based on variational autoencoder (VAE) and incorporating multilayer perceptron (MLP) and convolutional neural network (CNN), is proposed. The MLP is employed to upsample features and address issues related to small sample sizes and variable correlations, while the CNN serves as the feature extractor. To improve iterative effectiveness, a novel loss function is introduced. Comparative experiments have shown that the proposed model outperforms existing baselines in predictive performance. By applying the proposed method in the Yangtze Estuary Deepwater Channel, the study reveals several findings and provides recommendations. Results indicate that the most critical risk indicators are the degree of ship course deviation, ship size, and ship speed.
There is a limited understanding of how port-hinterland connections via inland shipping respond to major disruptions. This study investigates how containerized inland waterway transport in the Rhine-Alpine Corridor adapted after global disruptions between 2007 and 2022. Employing panel data modelling and network analysis, the paper examines regional transport hubs and reveals how their connections responded to events such as the financial crisis, critically low water levels, and geopolitical instability. The results provide insights into the vulnerability of cargo flows during disruptive years, along with distinct shifts in network connectivity following each event. The assessment reports the degree of robustness and redundancy across the corridor, offering practical guidance to support the development of regional responses to future disruptions.
China's coordinated development reform in 2006 is the starting point of contemporary port reforms and the best platform for exploring the persistent barriers to China's port coordination and the causes behind repeated reform restarts. This study assesses the geographical impact of 2006 reform in coastal port clusters and explains why the reform fell short of expectation. Using 1998-2019 panel data covering 59 coastal ports and 13 types of cargoes, reform impacts are evaluated through spatial distribution, specialization, and growth fluctuations. The results reveal that the reform modestly slowed throughput decentralization across port clusters, it failed to reverse the overall trend. Moreover, intra-cluster specialization also declined slightly since 2006. To further analyze the general and specific factors that influenced the performance, a quasi-DID approach was combined with policy text analysis. Compared with smaller ports, the main ports designated under the National Coastal Port Layout Plan (hereafter 'the Plan') experienced slower throughput growth,a sustained loss of market share, and no notable increase in dominant cargoes. These outcomes are attributable to institutional constraints and lagging concepts. The findings illuminate the constraints and dynamics shaping China's coastal port geography since 2006 and offer actionable policy insights for future regional port integration.
Regional cruise networks possess a multi-scale structure, with several localized hub-and-spoke systems and a limited number of connector ports linking these local systems. Multi-scalar connectivity substantially expands a port's potential to assemble diverse itineraries, a capability that plays a decisive role in determining cruise port performance. Existing studies typically rely on single-scale indicators that overlook the multi-scalar nature and differentiated market access it creates. This study addresses this gap by examining how port roles vary across spatial scales and how these roles influence passenger flows. Using the Mediterranean as a case study, a three-scale structure is identified, which is sub-basin, subregion and entire-basin. Results show that multi-scalar centrality provides markedly stronger explanatory power for passenger distribution. Sub-basin centrality directly increases a port's own passenger volume, whereas subregional and basin-wide centrality generate positive spillover effects for neighboring ports. The spatial division of these roles differs across sub-basins, producing three patterns: a Unipolar Homeport System, a Gateway-Homeport System, and a Multi-Connector System. The study accordingly advances a multi-scalar perspective on cruise networks and underscores the need for planning frameworks based on multi-scale communities and differentiated port development strategies.
This study focuses on two potential consequences of burnout in seafaring: turnover intention, an important issue given the concerns about the shortage of seafarers, and quiet quitting, where employees reduce engagement without formally resigning. Quiet quitting has not yet been studied among seafarers but given the increasing attention to employee disengagement in contemporary workplaces, it represents a significant potential risk, particularly for safety-critical operations. The aim of the study was to test whether turnover intention mediates the relationship between burnout and its four dimensions (exhaustion, mental distance, cognitive impairment, and emotional impairment) and quiet quitting among seafarers. Data were collected from an international sample of seafarers (N = 508) via an online survey using the Burnout Assessment Tool, Quiet Quitting Scale, and Turnover Intention Scale. Results indicate that turnover intention significantly mediates the relationship between all four burnout dimensions and quiet quitting, with the strongest mediation observed for exhaustion. All direct and indirect effects were statistically significant, supporting the hypothesized partial mediation model. These findings underscore the significance of turnover intention as a mechanism linking burnout to quiet quitting, highlighting the need for preventive measures to mitigate negative outcomes.
Against the backdrop of advancing the United Nations sustainable development goals SDG 7 (affordable and clean energy) and 13 (climate action), measuring and reducing carbon emissions has been critical for ports' decarbonization and sustainable development. This paper integrates data from multiple sources to establish a comprehensive carbon emission evaluation framework for ports, employing system dynamic simulation models to analyze the impact of various emission reduction strategies. And a dual modeling approach integrates bottom-up and top-down methodologies to respectively estimate emissions from in-port ships and port facilities, predicting forward emissions for multiple scenarios using artificial neural networks. A case study of the world's busiest container port-Shanghai Port is presented to demonstrate the proposed evaluation framework. The result shows that strategies, such as adopting alternative fuels, implementing government subsidies, and promoting carbon trading schemes, are effective methods for reducing carbon emissions. This study develops a data-driven framework to identify the most effective emission-reduction strategies providing a foundation for port managers and operators to assess and improve the sustainability performance of their operations. The results also offer insights for policymakers to reduce port carbon emissions and enhance sustainable port development.
Coastal shipping systems play a vital role in sustaining global trade but face growing maritime safety challenges under increasing traffic density and complex operating environments. This study develops an integrated Geographic Information System (GIS) and machine learning (ML) framework to map coastal accident risks and support differentiated governance. Using China's coastal waters as a representative case, 648 reported accidents (2013-2024) were analyzed. Spatial clustering identified 13 high-risk zones, while interpretable ML models highlighted crew allocation, voyage planning, and improper operation as dominant factors associated with collision severity. Interactive risk maps visualize accident types and spatial heterogeneity, enabling intuitive exploration of risk hotspots. Beyond technical contributions, the framework offers practical insights for maritime policy and management by linking spatial analytics with differentiated governance. It enables regulators and port authorities to prioritize inspections, allocate resources efficiently, and tailor safety interventions to local risk contexts. Overall, the study demonstrates how data-driven and interpretable methods can strengthen adaptive maritime governance and contribute to safer, more resilient coastal shipping systems.
Since the inception of the Green Shipping Corridor (GSC) concept in 2021, 126 GSCs have been initiated worldwide as of January 2026. However, significant misunderstandings have occurred regarding the naming of GSCs, their governance frameworks, the type of zero-emission vessels (ZEVs) with alternative fuels and ZEV fleet deployment on GSCs, and the greenhouse gas emissions computation from the fleet. To address these, this paper aims to conduct a systematic review of the development of GSCs by incorporating the most recent literature and policy developments. By doing so, it clarifies common misunderstandings of the GSC concept. This paper contributes to further refinement of GSC concept to enhance the implementation of GSCs in advancing the decarbonisation of the maritime industry. The paper also proposes research and policy agendas in implementing GSCs and optimizing ZEV fleet deployment.
With the increasing complexity of the global container shipping network, it exhibits multi-level collaborative relationships and high-order interactions among multiple ports. This study proposes a directed weighted hypergraph model analysis framework based on global AIS data, considering port berthing order and throughput capacity. By applying weighted K-core and motif methods, it analyzes the structural hierarchy and connectivity patterns of global ports. Results reveal a 7-level global container port hierarchy. Between 2015 and 2023, the network evolved into a more decentralized and structurally stable system. Strong dependencies surged from 31.5% to 50.5%, concentrating in the top three tiers, while direct subordinates of mega-hubs decreased, fostering balanced, collaborative clusters. Furthermore, transportation between higher-level ports (Levels 1-3) primarily forms hub-and-spoke structures, whereas lower-level ports rely on bridging structures. Consequently, we suggest lower-level ports strengthen hinterland routes to enhance hub collaboration. This research provides theoretical support for authorities to clarify port positioning and plan development. It also identifies smaller ports with locational advantages, encouraging them to develop local import and export industries.
Against the backdrop of intensifying global seafarer shortages, enhancing seafarers' retention intention to extend their service periods at sea has become an imperative task for shipping enterprises. Based on the job demands-resources theory (JD-R), this study explored the mechanism by which career transformation support from organisations affects officer-seafarers' retention intention by analysing a cross-level mediation model involving 317 shipping enterprises and 954 Chinese officer-seafarers. The results indicated that career transformation support enhanced retention intention through a fully mediated pathway by increasing the officer-seafarers' career transformation certainty (beta = 0.316, p < 0.001) and reducing their perceived stress related to career transformation (beta = -0.249, p < 0.001). Notably, the mediating effect of perceived stress due to career transformation existed solely at the between-group level, while within-group individual stress variations showed no significant impact. This conclusion breaks through the traditional binary framework of 'onboard or onshore,' and provides a practical paradigm of psychological resource management for shipping enterprises to achieve the goal of retaining officer-seafarers by building a systematic career pathway and a standardised stress buffer system at the organisational level.
Among various shipping risks, piracy incidents are characterized by selectivity, directness, randomness, and violence, posing a significant threat to the safety of vessels, crews, and cargo. Understanding the spatiotemporal evolution and influencing factors of piracy incidents is essential for maritime security policymaking and resource allocation. This study adopts a multi-scale approach, utilizing spatiotemporal analysis and machine learning methods to identify high-incidence piracy regions and their spatiotemporal evolution patterns. Additionally, the study quantitatively assesses the marginal effects of external environmental factors on the formation and development of piracy incidents. At the global scale, high-incidence piracy regions include Maritime Southeast Asia, Somali waters, and the Gulf of Guinea. The intensity and spatial extent of piracy clustering vary across these regions. At the regional scale, the spatiotemporal evolution of piracy incidents exhibits regional differentiation. At the local scale, the formation and development of piracy incidents are influenced by national political and socio-economic contexts, local land-sea interface environments, and specific triggering factors. Indicators such as the socio-economic impact levels from climate disasters, energy investment scale, and coastline complexity are the most significant drivers. This study provides data-driven support and a theoretical framework for maritime transport policymaking.
Ports are major energy hubs and significant sources of GHG emissions, facing growing pressure to improve economic and environmental performance. This paper proposes a distributed demand-response framework using a Multi-Agent System (MAS) in the JADE platform to coordinate flexible port loads-refrigerated containers (reefers), plug-in electric vehicles (PEVs), and shore power supply (SPS)-with ships at berth modeled as prosumers under explicit emission constraints. Price-aware fuzzy controllers allocate active power in real time, while an integrated, real-time MAS for decentralized voltage control shares reactive power regulation. In the examined case studies, terminal-scale simulations show a 8.949% reduction in total energy cost per kWh for flexible port loads (reefers, PEVs, SPSs). Adding ships' generators yields a 6.666% reduction in overall port operating cost per kWh, with reductions of 6.947% (per-ship emission limits) and 6.343% (port-wide) under emissions control. The MAS with voltage support maintains bus voltages within +/- 5% while preserving optimal costs and real-time runtimes. The results indicate a practical pathway to cost-efficient, emissions-conscious, and voltage-compliant smart-port operations.
The resilience of Regional Comprehensive Economic Partnership (RCEP) shipping network is crucial for supply chain security amidst growing geopolitical and operational risks. This study proposes a quantitative framework, integrating complex network theory and resilience triangle model, to assess and compare resilience of six key RCEP sub-networks. We simulate network dynamics across four phases-initial, disruption, recovery, and stabilisation-under four attack modes (targeting degree, betweenness, strength, and random) and recovery strategies. Resilience is quantified through the metrics of network efficiency, connectivity, and independent paths. Findings reveal significant regional heterogeneity. Bohai Rim, China-Australia-New Zealand, and China-Japan-Korea sub-networks exhibit superior resilience in efficiency and independent paths, while Bohai Rim, Guangdong-Hong Kong-Macao, and Yangtze River Delta excel in connectivity. Conversely, China-ASEAN network demonstrates relative vulnerability. Strength-based and Random Recovery strategies are identified as the most effective for most regions. Accordingly, we recommend enhancing hub functionality and path diversity for high-degree/strength ports (e.g. Shanghai , Hong Kong) in GHM and YRD regions, and strengthening transhipment capacity of high-betweenness ports (e.g. Busan, Melbourne) in CJK and CANZ networks. This study provides a robust analytical tool and strategic insights for building resilient maritime infrastructures within the RCEP framework.
Using monthly panel data on crude and product tankers from 2018 to 2025, we examine how the eco-vessel time-charter rate premium (ETCP) relates to the eco-vessel price premium (EPP) and how this relationship changes under decarbonization policy. Higher ETCP is associated with lower EPP, indicating that short-term freight market advantages do not translate mechanically into higher asset values. When decarbonization regulations, such as the IMO 2020 sulfur cap, are in force, they reduce the negative effect of ETCP on EPP by partially aligning the incentives of owners and charterers. Tightening environmental regulation increases eco-vessel values without a commensurate improvement in earnings, shifting valuation risk and financial burdens onto shipowners.
