The spatiotemporal heterogeneity of global shipping patterns and chokepoints: evidence from geopolitical events in two regions

Geopolitical disruptions increasingly affect maritime transportation networks, yet their spatially heterogeneous impacts on global shipping activity remain insufficiently understood. This study develops a two-scale analytical framework to investigate the spatiotemporal evolution of global shipping activity and examine regional responses associated with geopolitical disruptions in the Red Sea and the Arctic Ocean. Historical Automatic Identification System (AIS) data were processed to reconstruct vessel trajectories, aggregated into spatial grids, and analyzed using a space-time cube model and emerging hot spot analysis. At the global scale, the framework identifies persistent, intensifying, diminishing, and emerging patterns of shipping activity, revealing substantial spatial heterogeneity across major shipping corridors and maritime chokepoints. At the regional scale, the comparison between the Red Sea and the Northern Sea Route cases demonstrates differentiated response patterns under distinct geopolitical contexts. In the Red Sea region, AIS-derived observations show a rapid redistribution of vessel activity between the Red Sea-Suez Canal corridor and the Cape of Good Hope route. In contrast, Arctic shipping activity exhibited stronger temporal fluctuations and spatial heterogeneity, reflecting the combined influence of geopolitical conditions, environmental variability, infrastructure development, and resource-related activities. The findings indicate that geopolitical disruptions are associated with heterogeneous adjustments in maritime networks, with different regions exhibiting distinct patterns of route redistribution, activity concentration, and temporal persistence. The proposed framework provides a data-driven approach for monitoring changes in maritime activity, identifying vulnerable corridors, and improving understanding of shipping-network resilience under geopolitical uncertainty.

Shipping is a key component of global trade and commerce, carrying approximately 90% of international trade by volume (Wang et al., 2024; Rong, 2020). In the context of continued economic globalization, demand for maritime transport has grown steadily in recent years (Ducruet et al., 2018). A substantial share of maritime trade passes through one or more maritime chokepoints, raising longstanding concerns about the potential consequences of their disruption, blockage, or functional change (O’Neil and Talmadge, 2009; Pratson, 2023). Maritime chokepoints are strategically critical locations along major shipping routes. They are generally characterized by narrow navigable passages, high traffic density, limited throughput capacity, and a lack of readily available alternative routes (Rodrigue, 2005). Consequently, they constitute particularly vulnerable links within global transportation and logistics networks (Bai et al., 2023). Prominent examples include the Bosphorus Strait and the Panama Canal, while emerging Arctic shipping routes are also increasingly regarded as strategically important passages within the global maritime transportation system (Wenjun et al., 2025; Poo et al., 2024). Disruptions at these locations can generate substantial risks to maritime transportation and international trade, potentially reshaping shipping routes, port network configurations, and global trade patterns (UNCTAD, 2024a). In this study, the Northern Sea Route is treated as an emerging strategic maritime corridor rather than as a conventional narrow geographical chokepoint.

The effects of geopolitical crises on maritime transportation have consequently attracted increasing attention from international organizations. In December 2023, the 33rd session of the International Maritime Organization Assembly adopted a resolution addressing the impacts of the Russia-Ukraine conflict on international shipping (IMO, 2023). Similarly, the United Nations Conference on Trade and Development examined how the Red Sea crisis affected shipping traffic through the Suez Canal and disrupted wider maritime trade patterns (UNCTAD, 2024b). These developments highlight the vulnerability of maritime chokepoints to geopolitical uncertainty and the need to better understand the spatial and temporal responses of global shipping activities to regional crises. Recent studies have employed large-scale AIS data to assess the dependence of multiscale maritime networks on individual chokepoints and to quantify their strategic importance using network centrality indicators (Liang et al., 2025).

The term “geopolitics” broadly refers to the interaction between geographical conditions and competition for political power (Antonsich, 2004). Geopolitical risk arises from events such as military interventions, terrorist attacks, economic sanctions, regional conflicts, and political or regime changes in neighboring countries, all of which may pose substantial threats to global economic security and development (Chen et al., 2023). The maritime shipping industry is particularly exposed to such risks because of its dependence on geographically concentrated routes, ports, and strategic passages (Xiao et al., 2024; Caldara and Iacoviello, 2022). Geopolitical disruptions can alter international trade flows, trigger the diversion or reconfiguration of shipping routes, and reshape the spatial distribution of maritime traffic (Modarress et al., 2012). Their impacts may also extend beyond the transportation system by increasing volatility in financial and commodity markets (Pangalos, 2023), and undermining the stability of fuel and energy supply chains (Gupta et al., 2019; Khan, 2025). Geopolitical events can therefore be regarded as observable manifestations of underlying geopolitical risks. Against this background, the present study examines geopolitical events in two strategically important regions: the Red Sea and the Northern Sea Route. By comparing these two cases, the study investigates how different forms of geopolitical uncertainty influence the spatiotemporal patterns of shipping activity and the functional importance of maritime chokepoints.

The impacts of geopolitical uncertainty on global shipping have attracted growing scholarly and policy attention. Verschuur et al. (2025) quantified the systemic consequences of disruptions at major global maritime chokepoints, demonstrating how geopolitical conflicts and natural hazards can propagate through international trade networks. Ducruet and Yoon (2024) examined how the political isolation of North Korea has shaped maritime trade connectivity and regional development, while Jin et al. (2022) assessed the effects of the COVID-19 pandemic on China’s international liner shipping network. More recently, Rodriguez-Diaz et al. (2025) explored the geopolitical and regulatory challenges associated with the operation of shadow fleets in international shipping. The additional costs and risk premiums generated by geopolitical disruptions have also become an important research focus. The European Parliament (2024), for example, assessed the effects of the Red Sea crisis on shipping costs and supply-chain expenses in European economies. Zhao et al. (2025) investigated the implications of geopolitical risk for the decarbonization of maritime transport, while other studies have examined the resulting increase in marine insurance premiums and related financial risks during the Red Sea crisis (Alvarez, 2024; EIU, 2024). These studies indicate that geopolitical events can affect not only route selection and network connectivity, but also the operational costs, environmental transition, and financial resilience of the shipping industry.

Automatic Identification System (AIS) data have been widely employed to investigate global shipping activities because they provide large-scale, high-frequency information on vessel movements, routes, speeds, and operational patterns (Emmens et al., 2021). Massive AIS datasets can therefore be used to identify changes in the activities of different vessel types and to assess the spatial and temporal responses of shipping systems to external disruptions. Existing applications include revealing the adaptive capacity of container shipping networks during disruptive events (Cao et al., 2025; He et al., 2025), estimating black carbon emissions from Arctic shipping (Qi et al., 2025), and modelling multiscale maritime networks from individual ports to global shipping routes (Liu et al., 2024). AIS data have also been applied to reconstruct maritime trade patterns for major bulk commodities, including crude oil (Su et al., 2025; Prochazka et al., 2019) and liquid natural gas (Li and Yang, 2023; Yu et al., 2025).

Despite increasing attention to geopolitical risks in maritime transportation, several research gaps remain. Existing studies have mainly focused on individual cargo categories, specific shipping sectors, or single disruption events, while limited attention has been paid to the long-term spatiotemporal evolution of global shipping activity and the heterogeneous responses of maritime networks under different geopolitical contexts. In particular, how shipping activity patterns vary across different spatial scales following major geopolitical disruptions, and whether different types of disruptions correspond to distinct patterns of route adjustment and activity redistribution, remain insufficiently understood.

To address these gaps, this study develops a multiscale spatiotemporal analytical framework to characterize the evolution of global shipping activity and examine regional variations associated with major geopolitical disruptions. By integrating AIS-derived vessel trajectory data with a space-time cube model and emerging hot spot analysis, this study identifies persistent, intensifying, diminishing, and emerging patterns of shipping activity across global maritime networks. Furthermore, two strategically important regions, the Red Sea and the Northern Sea Route, are selected as case studies to compare different forms of maritime network adaptation under distinct geopolitical contexts. Specifically, this study aims to characterize the global spatiotemporal patterns and trends of shipping activity, identify regional changes in vessel activity associated with the Red Sea crisis and the Russia-Ukraine conflict; and explore how different geopolitical contexts correspond to heterogeneous patterns of route redistribution and maritime network adaptation.

This study makes three main contributions. First, this study links global shipping-pattern analysis with regional case studies to examine maritime-network changes across multiple spatial scales. Second, it compares the Red Sea and Northern Sea Route to reveal contrasting patterns of rapid route redistribution and gradual multi-factor adjustment. Third, it integrates AIS-derived activity with spatiotemporal analysis to identify how shipping patterns emerge, persist, intensify, diminish, or fluctuate over time.

The following sections provide a detailed description of the data and methods employed in this study. As illustrated in Figure 1, the methodological framework comprises four main stages: (1) data collection and preprocessing, (2) raster data processing, (3) space-time cube construction, and (4) emerging hot spot analysis.

Study area and data processing

This study adopts a two-scale analytical framework to investigate the spatiotemporal evolution of global shipping activity and changes associated with geopolitical disruptions. At the global scale, we examine the long-term spatial distribution and temporal evolution of maritime vessel activity. At the regional scale, the Red Sea-Suez Canal corridor and the Northern Sea Route (NSR) are selected as case-study areas to investigate heterogeneous changes associated with different geopolitical contexts, including the Red Sea security crisis and the Russia-Ukraine conflict (Figure 2). This multiscale framework enables the identification of both broad restructuring patterns of global shipping networks and region-specific adaptations around strategically important maritime corridors.

The primary dataset used in this study consists of AIS records provided by Shipxy (https://www.shipxy.com). The purchased dataset integrates terrestrial-based and satellite-based AIS observations, combining high-frequency coastal observations with offshore tracking capability. This integrated AIS configuration improves the spatial representation of vessel activities across coastal waters, open oceans, and high-latitude regions. The AIS dataset covers global maritime activities and contains vessel identification information (Maritime Mobile Service Identity, MMSI), timestamped geographic positions, navigation status and vessel type. This study focuses on three major commercial vessel categories: container ships, bulk carriers, and oil tankers. These vessel types were selected because they represent major components of global maritime trade and account for substantial shares of international cargo transportation. Vessel classification information was obtained from the vessel-type attributes associated with the Shipxy AIS records and linked by MMSI. Only Class A AIS messages were included in the analysis. Receiver-level metadata were not retained in the dataset available for this study; therefore, the relative contributions of terrestrial and satellite reception could not be reconstructed reliably for individual periods or regions. This limitation is considered explicitly when interpreting offshore and Arctic results.

The raw AIS dataset was processed through a multi-step preprocessing procedure before trajectory reconstruction. First, records lacking valid MMSI identifiers, timestamps, or geographic coordinates were removed. Duplicate observations and invalid positioning records were excluded to improve data consistency. Second, abnormal vessel movements caused by transmission errors or positioning anomalies were filtered based on unrealistic movement characteristics, including abnormal speeds and spatial discontinuities.

AIS observations with SOG less than 0.5 knots were classified as stationary and removed prior to trajectory construction. This filtering substantially reduced observations associated with berthing, anchoring, and other stationary activities, thereby focusing the analysis on vessel movements. A maximum temporal gap of 1 h was adopted for trajectory reconstruction. No independent fixed spatial-distance threshold was imposed; instead, implausible spatial jumps were identified through speed- and position-based quality control procedures. AIS records were grouped by MMSI and ordered chronologically to reconstruct vessel trajectories. The full berth-voyage-berth segmentation framework of Liu et al. (2025) was not implemented; instead, trajectory segments with excessive temporal gaps or implausible spatial jumps were removed to avoid artificial connections.

The reconstructed vessel trajectories were converted into line features and used as the basic spatial units for subsequent analysis. Kernel density estimation was then applied using the Kernel Density tool in ArcGIS Pro 3.4 to transform discrete vessel trajectories into continuous spatial representations of shipping activity.

KDE was used to characterize broad shipping corridors and spatial redistribution patterns rather than absolute vessel numbers. For the regional analyses of the Red Sea and Arctic study areas, KDE was performed using an output cell size of 10 km and a search radius/bandwidth of 100 km. The 10-km resolution was selected to preserve spatial detail in narrow chokepoints, while the 100-km bandwidth balanced corridor continuity and excessive smoothing. To reduce edge effects, each regional study area was buffered by 200 km, corresponding to the maximum tested bandwidth, and KDE was calculated using trajectories within the buffered extent before the resulting density surfaces were masked to the study area (Supplementary Table S5).

The KDE raster outputs were subsequently aggregated into regular spatial units for spatiotemporal analysis. The global analysis used a 1° × 1° geographic grid, whereas the Red Sea and Arctic regional analyses used a finer 10 km × 10 km grid. The global analysis was conducted in Plate Carrée (World; ESRI: 54001), while the Red Sea analysis used World Mollweide (ESRI: 54009) and the Arctic analysis used Asia North Albers Equal Area Conic (ESRI: 102025). To assess the robustness of the spatial patterns, sensitivity analyses were conducted using alternative KDE bandwidths (50, 100, and 200 km) and spatial resolutions (10km and 20km) (Supplementary Table S7). Monthly unique-MMSI counts were additionally summarized to characterize temporal variations in vessel activity (Supplementary Tables S1, S2).

Supporting spatial datasets, including coastlines and exclusive economic zone boundaries, were obtained from Marine Regions (https://www.marineregions.org/) and were used for geographic visualization and spatial referencing.

Methods

Space-time cube construction

A Space-time cube framework can be extracted for any given time point, thereby constructing a spatio-temporal integrated cube incorporating spatial and temporal axes. Originally proposed by Hägerstrand (1986), the space-time cube integrates spatial and temporal dimensions into a three-dimensional analytical structure and has been widely applied for identifying dynamic geographical patterns.

In this study, the KDE-derived raster layers were aggregated into monthly spatiotemporal bins, where the x and y dimensions represent geographic locations and the z dimension represents time. The temporal interval was set to one month because monthly aggregation provides a balance between capturing short-term responses to geopolitical disruptions and identifying long-term evolutionary patterns. The temporal windows used for the global, Red Sea, and Arctic analyses were predefined to ensure temporal comparability and are summarized in Supplementary Tables S3 and S4. The spatiotemporal cube was constructed using the Space Time Cube tools implemented in ArcGIS Pro 3.4.

Mann-Kendall trend analysis

The Mann-Kendall (M-K) test was applied to evaluate monotonic temporal trends in shipping activity within each spatial unit. The M-K test is a non-parametric approach widely used for detecting temporal trends without requiring assumptions regarding data distribution.

In this study, positive M-K statistics indicate increasing trends in shipping activity, whereas negative values indicate decreasing trends. Statistical significance was evaluated at the 90%, 95%, and 99% confidence levels. For a time series, the test statistic S is calculated as follows (Hamed, 2008; Xu et al., 2023):

Where

Where x i and x j are the KDE values at time steps i and j, respectively. Positive and negative values indicate increasing and decreasing trends, respectively. Positive S and Z values indicate increasing trends, whereas negative values indicate decreasing trends. Tied observations contribute zero to S and were accounted for in the variance calculation. Statistical significance was evaluated at the 90%, 95%, and 99% confidence levels (|Z| > 1.645, 1.960, and 2.576, respectively), with (p< 0.05) used as the main significance threshold.

Because statistical significance alone does not indicate the magnitude of change, Sen’s slope was additionally calculated to estimate the rate of temporal change:

Positive and negative Sen’s slope values indicate increasing and decreasing rates of change, respectively. For the NSR analysis, months outside July-October were excluded, and the ordered observations within the analyzed seasonal windows were used as the temporal sequence. Sen’s slope is therefore interpreted as relative change per analyzed observation rather than as an annualized rate. No pre-whitening or explicit correction for temporal autocorrelation was applied to the M-K analysis. Residual serial dependence may therefore affect statistical significance, and the trend results should be interpreted cautiously.

Emerging hot spot analysis

Emerging Hot Spot Analysis (EHSA) was used to identify the spatiotemporal evolution patterns of shipping activity. EHSA integrates spatial hotspot detection and temporal trend analysis based on the space-time cube framework. The parameter settings used for the EHSA are provided in Supplementary Table S4. The time-series analysis was further used to assess the temporal evolution of detected hot spots and cold spots.

The Getis-Ord Gi* statistic was used to identify statistically significant hot and cold spots (Wei et al., 2025; Mupfiga et al., 2024; Giuliani, 2024), while the M-K test was applied to evaluate temporal changes within the spatiotemporal bins. Based on the combination of spatial clustering and temporal evolution characteristics, each spatial unit was classified into different spatiotemporal patterns, including persistent, intensifying, diminishing, sporadic, oscillating, and other hotspot or coldspot categories. The Getis-Ord Gi* statistic is calculated as follows.

Where x j refers to the value of feature j, w ij. means the spatial weight between features i and j, n refers to the number of the elements.

Consequently, the space-time cube framework captures changes in hot- and cold-spot patterns across both spatial and temporal dimensions (Zhang et al., 2025). Parameters related to spatial relationships, temporal intervals, and statistical significance thresholds were consistently applied across the study areas to ensure comparability (Supplementary Table S5). The EHSA classification followed the standard classification scheme implemented in ArcGIS Pro 3.4. The standard ArcGIS Pro classification distinguishes categories including new, consecutive, intensifying, persistent, diminishing, sporadic, oscillating, and historical hot/cold spots, together with locations showing no statistically significant pattern (Supplementary Table S6).

Spatiotemporal patterns of global shipping activity

Global shipping activity exhibited a pronounced spatial concentration along major coastal regions, transoceanic routes, and strategically important maritime passages (Figure 3). Based on AIS-derived vessel trajectory density, shipping activity was primarily concentrated in East and Southeast Asian coastal waters, the northern Indian Ocean, trans-Pacific routes, the North Atlantic, and major maritime chokepoints connecting regional seas.

From July 2023 to December 2025, the monthly number of vessel movements summarized for the global analysis remained relatively stable, with an average of approximately 103,000 movements per month. Despite substantial monthly fluctuations, the overall number of vessel movements increased by approximately 5.4% during the study period (Figure 3). However, the spatial distribution of vessel activity exhibited considerable heterogeneity, with some corridors experiencing increasing trajectory density while others showed persistent decreases.

The M-K trend analysis identified statistically significant upward and downward trends in different regions of the global shipping network. Significant increasing trends were mainly distributed along sections of the West African coast, the Cape of Good Hope route, the southwestern Indian Ocean, and parts of the trans-Pacific corridor. In contrast, significant decreasing trends were concentrated along the Red Sea-Suez Canal corridor, the Mediterranean Sea, northwestern European waters, and parts of the South Atlantic route. The annual results showed more spatially extensive significant trends in 2024, whereas the 2025 pattern was dominated by non-significant grid cells with more localized significant changes (Figure 4).

The temporal distribution of significant trends indicated that spatially extensive significant trends were most evident in analysis windows beginning between July 2023 and July 2024. From the window beginning in August 2024 onward, the majority of ocean grid cells exhibited no statistically significant trend, although localized upward and downward trends persisted or intermittently re-emerged along several major shipping corridors (Supplementary Figure S1).

Of the 17 possible EHSA classifications (eight hot-spot categories, eight cold-spot categories, and no pattern detected), 14 were observed in the global analysis (Figure 4a). Hotspot patterns accounted for 18.5% of analyzed grid cells, whereas coldspot patterns accounted for 74.3%. Persistent hotspots were mainly located along established high-density shipping corridors, including East and Southeast Asia, the northern Indian Ocean, the eastern and western coasts of North America, and the Caribbean-Gulf of Mexico region.

Intensifying hotspots were mainly observed along the Cape of Good Hope route, the southwestern Indian Ocean, western African waters, and parts of the trans-Pacific corridor. In contrast, diminishing hotspots were concentrated around the Red Sea-Suez Canal corridor, the Mediterranean Sea, northwestern European waters, and sections of the South Atlantic route. Coldspot patterns were primarily distributed in relatively low-density open-ocean regions, including the central Pacific, southern Atlantic, southern Indian Ocean, and high-latitude waters.

Compared with the full-period pattern, hotspot distributions became more spatially extensive and continuous along several major shipping corridors in 2024, whereas the 2025 pattern showed greater spatial fragmentation and localized redistribution, indicating pronounced interannual variability in global shipping activity (Figure 4). Across the rolling windows, the specific hot- and cold-spot categories and their spatial extents varied, while the broader contrast between corridor-associated hotspots and open-ocean coldspots remained generally consistent (Supplementary Figure S2). Overall, the global analysis demonstrates that maritime shipping activity maintained a stable spatial backbone during the study period, while localized redistribution occurred along several major corridors.

Spatiotemporal changes in shipping activity in the Red Sea region

Following November 2023, AIS-derived vessel trajectory density showed pronounced spatial changes around the Red Sea-Suez Canal corridor and alternative routes around southern Africa (Figure 5). The Red Sea-Suez Canal route exhibited a substantial decline in vessel activity, whereas the Cape of Good Hope route showed increased trajectory density, forming a newly intensified shipping corridor (Supplementary Table S1).

Based on AIS-derived vessel observations, the number of vessels identified within the Suez Canal transit area decreased substantially during 2024 compared with the previous year. Compared with July-December 2023 with July-December 2024, the decline was observed across major vessel categories, including container ships, oil tankers, and bulk carriers, with reductions of approximately 71%, 45%, and 42%, respectively. These changes were accompanied by increased vessel activity along the Cape of Good Hope route, indicating a spatial redistribution of shipping trajectories between the two major Asia-Europe maritime pathways.

From July 2023 to December 2025, the trend analysis showed a clear spatial contrast between increasing and decreasing shipping activity. Significant increases (p < 0.05, 31.10%) were mainly observed along the South Atlantic-Cape of Good Hope-southwestern Indian Ocean corridor, whereas significant decreases (p < 0.05, 7.6%) were concentrated along the Mediterranean-Red Sea-Gulf of Aden corridor (Figure 6a).

The representative moving-window analyses showed that significant changes were most evident during the early window (July 2023-April 2024), when vessel activity patterns along the Red Sea corridor began to diverge from previous conditions. In the later windows (July 2023-April 2024, May 2024-February 2025, March-December 2025), most grid cells showed no significant trend, although localized changes remained visible along several alternative routes (Supplementary Figure S3).

From July 2023 to December 2025, Emerging Hot Spot Analysis identified 16 spatiotemporal patterns of the 17 possible classifications were observed in the Red Sea study area (Figure 6a). Hotspot patterns accounted for 15.8% of analyzed grid cells, coldspot patterns accounted for 65.0%, and non-significant patterns accounted for 19.2%. Hotspots were dominated by intensifying and persistent hotspots, which accounted for 5.5% and 3.7%, respectively, followed by diminishing hotspots (2.8%). Intensifying hotspots were mainly located along the Cape of Good Hope route and the southwestern Indian Ocean, whereas diminishing hotspots were concentrated along the Red Sea-Suez Canal corridor.

Compared with the full-period pattern, hotspot distributions became more extensive and continuous during 2024, while some hotspot areas became fragmented in 2025. These results indicate that the spatial redistribution of vessel activity around the Red Sea corridor persisted beyond the initial disruption period (Figure 6; Supplementary Figure S3).

Spatiotemporal changes in shipping activity along the Northern Sea Route

To minimize the influence of strong seasonal variability, the Arctic analysis focused on the core navigation period from July to October for each year between 2019 and 2025; months outside this window were excluded rather than treated as zero observations. AIS-derived vessel activity along the NSR showed substantial temporal fluctuations and spatial heterogeneity (Figure 7; Supplementary Table S2). During the 2019–2021 window, increasing trends were relatively widespread across the NSR region. During the 2020–2022 window, the overall trend weakened, with most grid cells showing statistically stable patterns and only limited areas exhibiting decreasing trends. During 2023–2025 window, increasing trends became more widely distributed across the NSR, forming a more continuous spatial pattern, while significant decreases became relatively limited (Figure 8). These changes indicate that vessel activity along the NSR experienced multiple phases of expansion, weakening, and renewed concentration during the study period.

Emerging Hot Spot Analysis revealed clear temporal differences in the dominant spatiotemporal patterns. Oscillating hotspots were the dominant category during 2019-2021, suggesting repeated fluctuations in vessel activity along the route. During 2020-2022, oscillating coldspots became more prevalent, reflecting reduced activity intensity in several NSR segments. In the most recent period (2023-2025), oscillating hotspots again became dominant, particularly along sections associated with major Arctic shipping routes (Figure 8).

Spatially, changes were not uniformly distributed across the NSR. Increased activity was concentrated in several key navigation corridors and around important Arctic maritime nodes, whereas other areas exhibited relatively stable or intermittent activity patterns.

The results indicate that NSR shipping activity has undergone a complex temporal evolution characterized by alternating periods of intensification and weakening. Unlike the relatively concentrated redistribution observed in the Red Sea case, Arctic shipping changes displayed stronger temporal fluctuations and spatial heterogeneity.

Robustness and validation analysis

To evaluate the reliability of AIS-derived shipping activity patterns, validation and sensitivity analyses were incorporated. AIS-derived vessel activity around the Suez Canal was compared with publicly available official statistics; the direction of decline was consistent with the reduction in canal transits reported by the Suez Canal Authority (2025). The official statistics are used only as an external consistency check and are not treated as AIS-derived observations.

Second, sensitivity tests were designed using alternative regional grid resolutions (10 and 20 km) and KDE bandwidths (50, 100, and 200 km) to evaluate whether the principal spatial patterns depend strongly on a single parameter setting.

Sensitivity analyses showed that the main spatial patterns were robust to reasonable variations in KDE bandwidth and grid resolution. Spatial correlations between the baseline KDE surface and alternative bandwidth and grid resolution scenarios ranged from 0.78 to 0.91 for hot spots and from 0.91 to 0.96 for cold spots (Supplementary Table S8). Changes in grid resolution and KDE bandwidth produced some local differences in spatial detail and hot- and cold-spot extent but did not alter the overall pattern of declining activity in the Red Sea and increasing activity around the Cape of Good Hope (Supplementary Table S8). Monthly unique-MMSI counts showed the same directional changes as KDE-derived intensity, further supporting the robustness of the observed redistribution (Supplementary Tables S1, S2).

Advantages and limitations of the spatiotemporal framework

Global maritime transportation networks are characterized by strong spatial concentration and complex interdependencies among routes, ports, and strategic passages. Localized disruptions may therefore extend beyond the affected area through route adjustment, traffic redistribution, and changes in the relative importance of alternative corridors. Capturing such processes requires an approach that considers both spatial variation and temporal evolution.

By integrating AIS-derived vessel trajectory density, space-time cube modeling, and emerging hot spot analysis, this study provides a multiscale framework for identifying not only where shipping activity changes occur, but also whether such changes are persistent, intensifying, diminishing, or oscillating over time. Compared with static mapping or simple before-and-after comparisons, the framework offers a more dynamic representation of shipping-network adjustment. The comparison between the Red Sea and the NSR further illustrates that different geopolitical contexts may be associated with distinct response patterns. The Red Sea case is characterized by relatively rapid redistribution between alternative routes, whereas changes along the NSR are more gradual and heterogeneous, reflecting the combined influence of geopolitical, environmental, infrastructural, and commercial factors.

Several limitations should nevertheless be considered. First, AIS observations remain subject to spatial and temporal variation in reporting frequency and completeness. Although the integration of terrestrial and satellite AIS improves coverage, observation gaps may persist, particularly in remote oceanic and high-latitude areas. Receiver-level metadata were not available to reconstruct historical terrestrial/satellite reception proportions, and the detected patterns should therefore be interpreted as changes in sobserved AIS-derived vessel activity rather than a complete census of maritime operations.

Second, KDE and spatial aggregation necessarily smooth fine-scale vessel movements. While this facilitates comparison of broad shipping corridors and regional redistribution patterns, localized changes around narrow waterways, ports, and small navigation areas may be obscured. The results should therefore be interpreted at the corridor and network scale rather than as precise representations of individual vessel behavior.

Third, the observational design does not support definitive causal attribution. The detected changes represent temporal and spatial associations between vessel activity and geopolitical developments, while the underlying mechanisms may also involve operator decisions, commodity flows, insurance conditions, infrastructure availability, and other factors.

Finally, the fixed July-October window used for the Arctic analysis improves seasonal comparability but does not capture possible changes in shoulder-season navigation or fully account for interannual sea-ice variability. No explicit quantitative sea-ice covariate was incorporated. In addition, although descriptive differences among major vessel categories were considered, separate spatiotemporal models were not developed for individual vessel types. Future research could therefore integrate environmental covariates and vessel-type-specific analyses to further clarify the drivers and heterogeneity of shipping responses.

Shipping network adaptation under acute geopolitical disruption: the Red Sea case

The Red Sea case provides an example of how shipping networks may rapidly adjust when a strategically important maritime corridor experiences elevated operational risks. The observed decline in vessel activity along the Red Sea-Suez Canal corridor and the simultaneous increase along the Cape of Good Hope route indicate a substantial redistribution of shipping trajectories during the study period.

This spatial adjustment is consistent with documented changes in carrier routing decisions following increased security concerns in the Bab el-Mandeb Strait and surrounding waters (Arvis et al., 2024; Bogetić, 2024). Rather than representing a complete interruption of maritime trade, the observed pattern reflects a shift of vessel activity from a shorter but higher-risk corridor toward a longer alternative route. Such route substitution is consistent with the adaptive capacity of global shipping networks while also revealing the additional operational pressures transferred to alternative corridors.

The redistribution of vessel activity has several implications for maritime network resilience. First, alternative routes can maintain transportation connectivity when major chokepoints become less reliable. However, route substitution does not eliminate disruption-related costs; instead, it redistributes operational burdens across carriers, vessels, fuel consumption, scheduling reliability, and alternative ports.

Second, the persistence of intensified activity along the Cape of Good Hope route suggests that shipping networks may not immediately return to previous configurations even when disruption conditions change, as operators may consider security risks, insurance conditions, and operational uncertainty.

Nevertheless, the observed changes should not be interpreted as solely determined by the Red Sea crisis. Other factors, including seasonal fluctuations, market conditions, fleet deployment strategies, and changes in energy transportation patterns, may also influence vessel activity. The AIS-based evidence primarily demonstrates the spatial reorganization of shipping trajectories during the disruption period, while the relative contribution of different driving factors requires further investigation.

Gradual and multi-factor restructuring of Arctic shipping: the Northern Sea Route case

Unlike the rapid route redistribution observed in the Red Sea, changes along the Northern Sea Route represent a more gradual and complex process shaped by interactions among geopolitical conditions, environmental variability, infrastructure development, and Arctic resource activities.

The results show that NSR vessel activity experienced alternating phases of intensification and weakening between 2019 and 2025. These fluctuations indicate that Arctic shipping development does not follow a simple linear expansion pathway. Instead, activity levels vary according to navigation conditions, commercial demand, vessel availability, and broader geopolitical circumstances (Ma et al., 2025; Zhao et al., 2024).

The Russia-Ukraine conflict represents an important contextual factor affecting Arctic shipping (Mladenovska, 2025; Yang et al., 2025). Changes in international participation, sanctions, energy trade patterns, and access to specialized vessels may have influenced the composition and intensity of NSR activity. However, these effects occurred alongside other important drivers. In particular, Russian Arctic resource projects, LNG transportation, icebreaker capacity, port infrastructure, and year-to-year sea-ice variability all contribute to shaping shipping patterns in the region.

The NSR therefore differs fundamentally from conventional maritime chokepoints such as the Suez Canal. Rather than functioning primarily as a high-volume global transit corridor, the NSR remains strongly connected to regional resource development and a limited number of commercial flows. The increasing activity observed in recent years reflects selective growth in specific sectors and routes rather than a comprehensive replacement of established Eurasian shipping corridors.

From a resilience perspective, the NSR provides additional geographical flexibility for international transportation but also introduces new dependencies related to Arctic environmental conditions, specialized vessel requirements, infrastructure availability, and regulatory arrangements. Future development of the route will likely depend on the interaction among climate change, technological progress, commercial feasibility, and geopolitical relations.

Implications for maritime chokepoint resilience

The comparison between the Red Sea and NSR cases highlights that maritime chokepoint vulnerability is not determined solely by physical geography but also by the adaptability of transportation networks and the availability of alternative routes. The NSR is therefore interpreted here as an emerging strategic maritime corridor relevant to route diversification, rather than as a conventional narrow chokepoint equivalent to the Suez Canal or Bab el-Mandeb Strait.

The Red Sea case demonstrates that global shipping networks possess substantial rerouting capacity, allowing vessel activity to shift between major corridors when risks increase. However, such flexibility is accompanied by higher operational costs and increased pressure on alternative routes.

The NSR case illustrates a different pathway in which geopolitical and environmental changes gradually reshape the strategic importance of emerging maritime corridors. Although Arctic routes may provide future diversification opportunities, their development remains constrained by environmental uncertainty, infrastructure limitations, and geopolitical factors.

These findings suggest that maritime resilience strategies should move beyond protecting individual chokepoints and consider broader network-level adaptation. Monitoring changes in vessel activity, identifying emerging alternative corridors, and evaluating the capacity of ports and routes to absorb redirected traffic are increasingly important for managing geopolitical uncertainty in maritime transportation systems.

This study developed a multiscale spatiotemporal framework to investigate the evolution of global shipping activity and examine how maritime networks respond to geopolitical uncertainty. By integrating AIS-derived vessel trajectory density, space-time cube modeling, and emerging hot spot analysis, this study identified persistent, intensifying, diminishing, and fluctuating patterns in global shipping activity and explored heterogeneous responses in two strategically important regions: the Red Sea and the Northern Sea Route.

The results indicate that global shipping activity maintains a relatively stable spatial backbone along major coastal regions, transoceanic routes, and maritime chokepoints, while localized redistribution occurs in response to changing operational conditions. At the global scale, increasing and decreasing activity patterns were concentrated along several major corridors, indicating that maritime networks continuously adjust their spatial organization rather than remaining static.

The regional analyses reveal contrasting response patterns under different geopolitical contexts. In the Red Sea case, AIS-derived observations show a rapid redistribution of vessel activity between the Red Sea-Suez Canal corridor and the Cape of Good Hope route, reflecting the ability of global shipping networks to adapt through route substitution when operational risks increase. In contrast, changes along the NSR exhibited stronger temporal fluctuations and spatial heterogeneity, reflecting the combined influence of geopolitical conditions, Arctic environmental variability, infrastructure development, and resource-related shipping activities.

This study contributes to maritime transportation research by demonstrating that geopolitical uncertainty does not generate uniform impacts across shipping networks. Instead, different types of disruptions may produce distinct spatiotemporal response patterns depending on route availability, regional characteristics, and network adaptability. The proposed framework provides a quantitative approach for identifying emerging changes in maritime activity and assessing the resilience of strategically important shipping corridors. Future studies could integrate additional operational, environmental, and vessel-level information to further investigate the mechanisms underlying maritime network adaptation.

Overall, the findings highlight the importance of understanding maritime resilience from a network perspective. Beyond protecting individual chokepoints, effective maritime risk management requires continuous monitoring of emerging route changes, evaluation of alternative corridor capacity, and consideration of the complex interactions between geopolitical uncertainty, environmental constraints, and shipping system adaptation.