Tracking the global reduction of marine traffic during the COVID-19 pandemic

The study investigates how the COVID-19 pandemic and associated mobility restrictions (the so-called anthropause) affected ship-based human activities at sea. With substantial global changes in transport, energy consumption, and consumer demand, the authors aim to quantify the magnitude, spatial patterns, and temporal dynamics of marine traffic changes across multiple maritime sectors. Leveraging Automatic Identification System (AIS) data, which enable near real-time global monitoring of vessel movements, the research assesses changes during the first half of 2020 relative to 2019 and examines relationships with governmental confinement measures. The work addresses the need to understand implications for the blue economy and potential environmental effects on the marine environment.

Prior work has documented wide-ranging environmental and socioeconomic consequences of COVID-19, including reductions in CO2 emissions and localized declines in underwater noise, water turbidity, and fishing effort. AIS has been widely used to monitor maritime activity and related impacts (e.g., air pollution, invasive species spread, vessel-wildlife collisions) and has been applied to assess COVID-19-related risks and local traffic changes. Historical analyses indicate growing trends in shipping, cruise tourism, and fisheries activity, with cumulative human impacts rising across most EEZs. The literature also highlights that external drivers (e.g., oil prices, regulations, piracy, cultural/political events) can shape maritime activity, underscoring the need to disentangle COVID-19 effects from other factors.

Global analysis: The authors used satellite AIS (S-AIS) density maps from exactEarth Ltd. for January–June in 2019 and 2020, derived from a constellation of 65 microsatellites (average update rate <5 minutes; coverage comparable between years). Data were aggregated to a 0.25° × 0.25° global grid as monthly vessel transits per cell and converted to densities (transits per km² per month) accounting for cell area by latitude. Vessels were categorized as cargo, tanker, passenger, fishing, and other (e.g., service, research, recreational). Changes were computed as absolute and percentage differences between equivalent months in 2020 versus 2019 at cell level. Occupancy (spatial extent of cells used) and proportions of cells with increases/decreases were calculated monthly by sector and globally. Spatial summaries were produced by latitude, distance to coast, and across regions and selected focal areas. Marine chokepoints (n=10) were analyzed within 0.5° radius to assess monthly mean absolute changes and variability by sector. Country-level analyses summarized mean absolute changes within each EEZ. Policy stringency and statistical modeling: The Oxford COVID-19 Government Response Tracker (OxCGRT) Stringency Index (0–100) provided a measure of confinement severity per country. Linear mixed-effects models (LMMs) assessed the effect of stringency and its interaction with World Bank income class on relative changes in vessel traffic density by ship category, with country as a random effect. Models were fit in R (lme4 with REML; significance via lmerTest Satterthwaite method). The number of countries per model varied by category due to differences in occupancy. Regional high-resolution analysis (Western Mediterranean): Terrestrial AIS (T-AIS) data (5-minute intervals) were obtained via SOCIB’s real-time system connected to MarineTraffic, covering 1 January 2016–30 November 2020 (>545 million messages). Data cleaning removed duplicates and invalid MMSI entries; consistent MMSI–vessel attributes were selected annually. Vessel categories matched S-AIS with an additional split of recreational boats (sailing and pleasure craft; codes 36 and 37) from other; excluded irrelevant/invalid codes (20–29, empty/null). To mitigate uneven antenna coverage, analyses were restricted to vessels underway within 24 nautical miles (44.4 km) of EU coasts (Spain, France, Italy) across 164,318.2 km². Daily counts of underway vessels were computed per category. Temporal comparisons used equivalent 2019 periods as baselines to estimate percentage reductions, median and maximum drops, and recovery trajectories across sectors.

• Global declines: During January–February 2020, global traffic generally increased versus 2019; after WHO’s pandemic declaration (11 March), patterns reversed with 52.2% of grid cells showing decreases in March and peaking at 54.8% in April.
• Occupancy: Considering all vessels, global occupancy declined maximally by 1.4% in April 2020 relative to 2019; passenger vessels exhibited the largest sectoral drop in occupancy at 30%.
• Sectoral differences: Cargo decreased from January onward; passenger and other vessels showed marked declines after March (passenger most severe); tanker reductions became evident in May–June; fishing showed indications of recovery from May.
• Spatial patterns: Largest decreases occurred in coastal areas and the Northern Hemisphere. Passenger traffic decreased strongly in tourism hotspots (Caribbean, Mediterranean). Mixed patterns were observed in the East China Sea and along some shipping lanes (e.g., Arabian Sea). Increases occurred in regions like Indonesia.
• Chokepoints: Substantial decreases after March at the Panama Canal, Strait of Gibraltar, Strait of Dover, and Bosphorus (notably in cargo and other vessels), while increases near the Cape of Good Hope were driven by cargo traffic.
• Country-level: 70.2% of analysed EEZs (n=124 countries) showed overall decreases in average traffic density during January–June 2020. China decreased in January–February with early rebound March–April; many countries declined after March; by June, some showed recovery (USA, South Korea, India) while others remained low (UK, Peru).
• Western Mediterranean (daily T-AIS): Following 11 March, counts of underway vessels declined across all sectors versus 2019 baselines, with an overall median drop of 51% during initial national lockdowns until ~22 June. Maximum reductions ranged from 22.2% (tankers) to 93.7% (recreational boats); maximal overall drop across all categories reached 62.2% in mid-April. Passenger vessels had the highest median drop among sectors (47.5%). Recovery was uneven: cargo, tanker, and especially fishing recovered relatively swiftly; recreational rebounded rapidly mid-July to mid-September; passenger remained depressed despite partial summer recovery.
• Stringency effect and income: LMMs found the Stringency Index significantly associated with reductions in traffic density across all categories except fishing. Lower-income countries were less affected by confinement measures.

The findings demonstrate a rapid, unprecedented global and regional contraction of marine traffic following COVID-19 lockdowns, with pronounced declines in passenger and recreational sectors and comparatively smaller impacts on maritime transport (cargo and tankers), reflecting the resilience of global supply chains. European and Mediterranean regions exhibited the most notable decreases, whereas East Asia, notably China, recovered earlier, aligning with the timing of easing restrictions and corroborated by independent CO2 emissions rebounds. Chokepoints analyses highlighted the systemic propagation of disruptions across major routes, while alternate routing (e.g., Cape of Good Hope) reflected changing economic conditions (e.g., oil price, canal toll avoidance). The spatial heterogeneity and sector-specific dynamics underscore how policy stringency and economic context shape maritime mobility. These traffic changes have implications for environmental pressures (e.g., underwater noise, air pollution, risk of wildlife collisions) and for tourism-dependent economies potentially facing sustained impacts. The results validate AIS-based monitoring as a powerful tool for near real-time assessments of human pressures at sea and for identifying natural experiments to study ecological responses during reduced-activity periods.

This study provides the first global, multi-sector, near real-time assessment of marine traffic reductions during COVID-19, integrating satellite and terrestrial AIS with policy stringency data. It quantifies heterogeneity in magnitude, timing, and sectoral impacts, reveals strong global declines peaking in April 2020, and documents pronounced and prolonged disruptions in passenger and recreational traffic. The approach offers a scalable framework for tracking blue economy dynamics and informing environmental and conservation assessments. Future research should: (1) extend longitudinal monitoring to capture long-term shifts and recovery; (2) integrate AIS trajectory-based metrics (e.g., speeds, port calls) to infer pressures like noise, emissions, and anchoring; (3) assess small-scale fisheries not well captured by AIS; (4) disentangle COVID-19 effects from other drivers (e.g., regulations, market dynamics); and (5) conduct targeted regional and sectoral studies to evaluate socioeconomic and ecological outcomes.

AIS data have known constraints: many small vessels lack AIS and transmission gaps occur, limiting detection of small-scale fisheries. Terrestrial AIS coverage is spatially uneven, necessitating coastal filtering and potentially biasing counts. Not all observed changes are attributable solely to COVID-19; concurrent factors include oil price fluctuations, regulatory changes (e.g., Indonesia’s AIS mandate in August 2019), rerouting to avoid canal tolls, seasonal fishing ground shifts, and other socioeconomic or political events. Comparing 2020 to 2019 may underestimate impacts given underlying growth trends in marine traffic occupancy (~3% in 2020). Country-level analyses may be influenced by differences in AIS adoption, enforcement, and economic context.