Satellite imagery combined with machine learning provides an unprecedented, detailed view of the maritime industry, focusing specifically on the number and activities of fishing and transport ships at sea. Surprisingly, there are far more vessels operating than publicly available data indicates, a reality that policymakers must acknowledge.
As a shared global resource, the oceans belong to everyone; however, different countries and regions have diverse customs, laws, and motivations that influence their maritime practices.
A key tool in maritime monitoring is the Automated Identification System (AIS), which utilizes shipborne transponders to accurately track vessel activities. However, AIS adoption is not consistent globally, leading to gaps in crucial data regarding the number of fishing vessels in specific areas, their operators, and fish catch amounts. This fragmented information—derived from local, proprietary, and government-approved sources—poses challenges for informed policy decisions and creates an atmosphere of lawlessness in the industry. Many ships operate in restricted or protected waters or exceed safe harvesting limits, threatening vulnerable fish stocks.
Satellite imagery provides a solution to these challenges, offering a perspective that makes it nearly impossible for vessels to operate unnoticed. The vastness of the maritime industry, coupled with the extensive nature of satellite imagery, demands efficient analysis. Fortunately, advancements in machine learning can facilitate the necessary vessel recognition and tracking operations, enabling us to monitor tens of thousands of ships actively operating at sea.
In a study published in Nature, researchers Fernando Paolo, David Kroodsma, and their team at Global Fishing Watch, along with collaborators from various universities, analyzed two petabytes of satellite imagery collected between 2017 and 2021. They identified millions of vessels at sea, cross-referencing this data with reported coordinates from AIS tracking.
The findings revealed a startling statistic: approximately three-fourths of all industrial fishing vessels remain publicly untracked, as do nearly a third of all transport and energy vessels. The significance of the "dark fishing" industry may even rival that of the documented sector. Furthermore, satellite imagery recorded increases in wind turbine installations and other renewable energy projects, which also pose tracking challenges.
Paolo explained, “There are several reasons why these vessels are not included in public tracking systems.” Smaller boats, those in areas lacking satellite coverage, and vessels deliberately turning off their transponders can all be considered "untracked."
He noted, “Some countries utilize proprietary systems for monitoring vessels within their maritime boundaries. However, these systems are limited to specific vessels, and this information does not extend to other nations.”
As global populations rise and ocean temperatures increase, sharing vital data across borders and addressing issues within national jurisdictions is more critical than ever.
“Fisheries are dynamic resources that move extensively, so transparent tracking of fishing vessels is essential for sustainable stock management. A complete understanding of vessels' ecological impacts is challenging without consistent public position and activity reporting,” Paolo emphasized.
Visualizations from the study show that regions like Iceland and the Nordics have the highest levels of vessel tracking, whereas Southeast Asia fares poorly—often recording virtually no tracking off the coasts of Bangladesh, India, and Myanmar.
This lack of tracking does not imply illegal activity; it merely suggests that vessel activity may not adhere to the publicly shared requirements observed in Nordic countries. The global community has minimal visibility into the fishing activities in these regions. Notably, the study found that the Asian fishing industry is significantly under-represented in data provided by AIS.
Researchers revealed that AIS data might suggest around 36% of fishing activity occurs in European waters and 44% in Asia. However, satellite data paints a strikingly different picture, indicating that only 10% of fishing vessels are in European waters, while an astonishing 71% operate in Asian waters. China alone accounts for nearly 30% of global fishing activity.
This analysis does not aim to place blame on specific countries or regions but instead highlights the severe gaps in our understanding of the global fishing industry’s scale. Without accurate data, both policy formulation and scientific research risk being fundamentally flawed.
Moreover, satellite analysis identified consistent fishing activity in protected areas, such as the Galapagos Islands, where fishing is prohibited under international law, warranting heightened scrutiny of these illicit vessels.
“The next step involves collaborating with authorities in various regions to investigate these new maps. We may have uncovered fishing activity within marine protected or restricted areas that requires further assessment and safeguarding,” Paolo stated.
He remains hopeful that enhanced data will inform policy decisions, but he acknowledges the work is far from finished. “This represents the initial phase of our open data platform. We are continuously processing fresh radar imagery from the Sentinel-1 satellite, identifying activities globally. This data is available on our website, globalfishingwatch.org, and is updated to reflect information collected just three days prior.”
Global Fishing Watch is a nonprofit organization supported by various philanthropies and individual donors, whose contributions are acknowledged on their website.
Ranking
