Global Fishing Watch
Illuminating vessel activity on the open ocean
Under Pressure
It has been estimated that only 25% of global fishing activity is being actively tracked or monitored across the world’s oceans. This leaves fleets of so-called “dark vessels” – many of them small in size – who are skirting regulations, sovereign state territorial boundaries, and applying increasing pressure on the ocean’s resources and fragile ecosystems.
Driving the market for these dark fleets is an increasing global market demand for fish and fish byproducts, to the extent that over half of the world’s oceans are now fished on an industrial scale. According to the U.N. Food and Agriculture Organization, over the last 60 years, fish consumption has increased roughly 3% annually – almost double the population growth rate of 1.6%. The largest industrial fishing fleets now circumnavigate the globe for months or years on end, often conducting dozens of transshipments with small nearby fishing vessels to transfer and fill their holds before returning to port.
Large industrial fishing trawlers operating together in the Mergui Archipelago, Myanmar.
The math paints a clear picture: we need to better understand where humans are undertaking fishing activities worldwide through improved transparency – so that we can take action to sustainably protect these fragile marine resources for future generations. The first step to understanding is knowing when and where fishing is happening.
The global fishing footprint had never been quantified… until now.
Mapping Fishing Intensity
Global Fishing Watch is an independent, international nonprofit organization whose purpose is to provide a global real-time view of fishing – making human activity at sea publicly accessible. To further this goal and get this valuable information in the hands of the broader science and conservation community, Global Fishing Watch's Fishing Intensity layers are now available in the ArcGIS Living Atlas of the World .
These new services provide a snapshot of relative fishing intensity for the year 2020, and are available as both a yearly aggregate and a monthly timeseries:
These multidimensional raster layers summarize vessel track density (total vessel hours) and apparent fishing activity (fishing hours) variables, distinguishing between the transit of vessels to or from port, and apparent fishing activity. Each 1 square-kilometer pixel value represents the number of total hours of vessel or fishing activity are occurring annually or monthly.
The swipe map below compares total vessel hours (left) and fishing hours (right) variables. Note the linear tracks of ships transiting from across the planet to fish the waters off the coast of Argentina and Chile.
Global Fishing Watch hours (left) and fish hours (right). Note the linear paths in the hours map, which show vessels transporting to and from ports and fishing grounds.
The basemap used here provides some context for "what's under the boat?" with its deliberate ocean-forward design:
- Land area is muted, accentuating ocean features
- Bathymetric depth classes (shown at right) are guided by meaningful pelagic zone boundaries
- Important bathymetric features such as ridges, canyons, basins, troughs, straights and mounts are prominent
Exploring patterns of global fishing activity can prompt some questions about why certain areas experience high fishing intensity, and why others don't. Take a tour of some of the most heavily-fished parts of the world below, and look to the basemap for clues as to why some oceans see more vessel traffic than others.
Much of the fishing activity in Europe is taking place in the shallow waters around Norway, Denmark, the United Kingdom and Ireland, France, Spain, and Italy.
The Nystrøm, pausing through Brønnøysund harbor, fishes using purse seines and is registered in Norway.
Iceland's remote location makes its economy heavily dependent on the fishing industry. According to Government of Iceland statistics, fishing-related jobs employ 7% of its workforce and account for 12% of the country's GDP.
Top fish-consuming countries by kilograms per person per year (U.N. Food and Agriculture Organization, 2023).
Globally, the most fishing activity occurs in the waters off of South Korea and China, represented by the bright yellow and orange hues of the fishing intensity layer.
China currently holds the #1 spot, accounting for 15% of the total catch annually, while the top 20 producing countries combined account for 74% of global fish production.
Fishing boats on Fushan Bay, Qingdao, China.
The waters off the coast of Argentina also show particularly high intensity of vessel traffic. In the Falkland Islands, squid fishing comprises 75% of the catch there, which finds its way to consumers in Europe and the Far East.
Orange fishing boats in Mar del Plata, Argentina.
A counterpoint to the most heavily-fished waters are areas that have little to no detected fishing activity - like the Bay of Bengal, the Andaman Sea, or the Java Sea.
In these coastal waters, the apparent dearth of fishing activity is more likely explained by the lack of voluntary tracking - or "dark fleet" vessels actively circumventing it.
A large fleet of fishing trawlers working together in a small area of the Andaman Sea.
A look at the more granular monthly 2020 time-series highlights local temporal trends that are lost in the yearly aggregate. These may include abrupt changes in fishing activity according to regulations or moratoriums on fishing activity - like the Gulf of Mexico Shrimp Fishery Management Plan 's temporary hold on brown shrimp fishing in the late spring of 2020:
Monthly fishing intensity can be impacted by the seasonality of a fished species, weather conditions, or regulations.
Methodology
Unfortunately, fishing vessels don’t actively broadcast when they switch between transit or fishing activity, so Global Fishing Watch utilizes some innovative detection tools to classify vessel activity. The routes themselves are tracked through terrestrial- and satellite-based positioning technology that reports participating ship locations and movement attributes several times a minute. These breadcrumb-like trails can then be analyzed and classified by apparent activity.
AIS
Mapping the locations and activities of vessels at sea is vitally important for navigation, collision avoidance, maritime security, search and rescue, economics, and many other use cases. Knowing where vessels at sea are located – fishing or otherwise – is made possible through the use of the Automatic Identification System (AIS), a network of terrestrial receivers which monitors the movement and location of equipped vessels.
A single NEXT SAT-AIS satellite can cover approximately the same ocean area as the contiguous United States.
Near shore, the vessels communicate with terrestrial (or other ship-based) AIS receivers, which have a typical rage of 15-20 nautical miles at sea level up to 40-60 nm or beyond when placed unobstructed at higher elevations. To extend the reach of AIS beyond the ~20 nm buffer on coastlines and around other vessels, space-based AIS (SAT-AIS) fills void of the open ocean and provides position information for every location on earth.
Fishing Gear
Fishing methods and technology has evolved over time to exploit different species of fish and pelagic environments, each with varying levels of bycatch and sustainability. The gallery below illustrates the equipment used to catch different types of fish. Click on each image for more information on each type from the Marine Stewardship Council .
Common fishing methods and equipment used globally. (Modified images courtesy of the Marine Stewardship Council .)
Longline and trawling fishing methods make up a large proportion of the fishing hours - and catch - detected by Global Fishing Watch. When tracked by AIS, the vessels leave a GPS trail that hints at the onboard activity - and the catch they are pursuing - governed by the direction, speed, and pattern of turns.
A timeline of popular fishing methods ranked by total fishing hours in 2020. (Source data from Global Fishing Watch.)
It's possible to make an educated guess about whether a vessel is fishing or not just by looking at the shape of its track. Is it taking a direct route to or from a port in a relatively straight path? Probably in transit. Or, is it creating "loopy" or "figure-8" shapes, following bathymetric features or contours, or skirting nearby regulatory or sovereign boundaries? Probably fishing.
Manually making these activity distinctions for every track would be unfeasible, so Global Fishing Watch leverages automated tools to categorize the activity of every individual segment of a vessel's AIS track.
Machine Learning
Global Fishing Watch is solving the “when is a vessel actively fishing?” question through the use of machine learning. This was accomplished in two parts, using 22 billion AIS vessel positions collected from 2012-2016, while employing Convolutional Neural Networks (CNNs), a type of Deep Learning Network. (More details on the methodology can be found here. )
First, a CNN was used to identify vessel characteristics, placing them into one of six classes of fishing, and an additional six classes of non-fishing vessels. 45,441 marine vessels with known registries were used to train the model, resulting in the correct identification of the vessel's class with 95% accuracy.
A second CNN model was then used to to determine whether, for any individual AIS track segment, the vessel was conducting apparent fishing activity. This model used a vessel's identity, and changes in location, speed, and direction in combination with a training dataset of AIS tracks of 503 known vessels, to correctly identify activity (fishing vs. not fishing) over 90% of the time.
Flow chart showing the aggregation of daily fishing activity into monthly and annual rasters.
The resulting points were then summarized in daily vessel density rasters (one for only fishing activity, one for all vessel activity) and further aggregated into monthly and annual views.
The animation below shows the vessel track of a Florida squid boat at sea for 3 weeks in April, 2020. This is typical of the AIS track data used by Global Fishing Watch for assessing vessel activity, as it transits from between ports in Tampa and Clearwater to nearby schools of squid off the shallow shelf.
AIS track of a squid fishing boat over 3 weeks in April 2020 off the coast of Florida, near the port of Tampa.
As each school of squid is fished, the vessel moves on to new, more active grounds - until a return to port is required to offload or refuel. As the detected fishing hours activity (colored raster cells) appear near the end of the animation above, note the relative accuracy (and omissions) of detected fishing activity by the machine learning model. As mentioned earlier, the model is currently 90% accurate in its detection, leaving some room for future improvement.
Summarizing Activity
Equal-area hexbins can be used to summarize regional trends in fishing intensity at a range of scales.
Creating summaries of fishing activity can highlight broader spatial or temporal patterns in intensity. One way of representing persistent fishing pressure globally is to use the monthly fishing hours layer in combination with equal-area tessellations to count how many months of the year there is any apparent fishing activity in each bin.
Zonal Statistics was used to determine, in each cell, how many months of the year (0 to 12) fishing is happening. While some areas show logical near-shore fishing activity, others prompt questions about why year-round fishing happens in such remote areas - and what they might be fishing for.
This map shows the number of months per year that apparent fishing activity is occurring - summarizing the monthly fishing hours layer using multidimensional Zonal Statistics .
Reunion is a young island with a steep bathymetric profile, only recently emerging from the sea. Just a few kilometers from shore, the ocean floor drops rapidly past 1000 meters depth, making vertical fishing methods, like jigging, popular here.
Fishing on the west of the island happens year-round, but trade winds on the east side of the island in the winter often make these waters inaccessible.
Another way to contextualize fishing activity is to enrich these summary hexbins with other information, like bathymetry - specifically, slope. A majority of the global fishing activity happens in shallower coastal waters, with the highest fishing intensity actually takes place on shelf edges.
Using reference size theme for binning in ArcGIS Online, fishing intensity can be correlated to how much slope is present in the underlying bathymetry within each hexbin. The slope is represented by the color of the symbol, while the size of the interior circle indicates the relative level of fishing intensity.
Bathymetry near the West European Basin from 1x to 10x vertical exaggeration, with fishing hours shown at sea level.
The European Union and Guinea-Bissau have been in a sustainable fisheries partnership since 2007, allowing vessels from Spain, Portugal, Italy, Greece and France to fish in Guinea-Bissau’s coastal waters.
As part of the €17 million per year agreement, about 25% of the funds will be used for sustainable fisheries management, improved control and surveillance, and supporting local fishing communities.
Fish for sale to the European market in Bologna, Italy.
In 2006 an article in the journal Science predicted that, without intervention, a global fishery collapse would occur by 2048. In the 18 years since this shocking projection was made, there are signs that humans are capable of taking this threat seriously, with a growing understanding that strong management, international leadership, and cooperation can turn the tides from an antiquated "conquer the ocean" mindset, to one of protection and local stewardship.
Protection and Monitoring
Ocean resources can be protected from fishing at the local community or state level - through restrictions in the time and type of fishing activities that can occur - or through offshore extensions of a country's sovereign boundary. To prevent over-fishing there are several layers of international and domestic protections in place, all of which are available in the Living Atlas:
Each of these protected areas, explained below, have a meaningful impact on where fishing activity happens globally.
Exclusive Economic Zones (EEZs)
To give sovereign states the right to exploit, manage, and conserve the resources in their own shallow coastal waters, Exclusive Economic Zones prohibit encroachment by international fishing vessels within a defined buffered area from the nearest coastline or projected offshore country boundary.
An Exclusive Economic Zone (EEZ) is a defined maritime area where a coastal state has jurisdiction over both the living and non-living resources within that zone. Prescribed by the United Nations Convention on the Law of the Sea , these boundaries are designed to be used in combination with other marine boundaries in order to help determine areas of trade, commerce and transportation.
This graphic shows the mix of factors and sub-zones that determine EEZ jurisdictions, including territorial boundaries, continental shelf extents, and also the mysterious, unregulated region in the high seas known as The Area.
Complicating the future legal status of these EEZ boundaries is the dynamic nature of the coastlines that define them, particularly due to climate change and sea level rise. The end result is a gradual shrinking of maritime boundaries, leading to potential future international disputes regarding access to ocean resources.
Visit NOAA for more information on these EEZ sub-regions.
Exclusive Economic Zones are helpful in explaining why many fishing areas exhibit such sharp boundaries, with little to no fishing activity just a few kilometers away.
Oman's EEZ is one of many locations where industrial shipping operations in The Area pull right up to the state's sovereign boundary.
World Database of Protected Areas (WDPA)
A joint project between UN Environment Programme and the International Union for Conservation of Nature, the World Database of Protected Areas (Marine and Costal) in Living Atlas is the most comprehensive global database of marine and terrestrial protected areas.
Updated monthly, these boundaries are used to define habitats or sensitive ecosystems that have been deemed important for protection by different governments or managing authorities.
The Great Barrier Reef, off the coast of Queensland in the Coral Sea.
The WDPAs include: marine parks, reefs, nature reserves, dive areas, national parks, and other actively-managed protected areas.
Most importantly for recreational and commercial fishing, they also include information on fishing restrictions. While many areas allow for resource extraction within the protected area, the highest level of protection defines them as "no take" year-round, prohibiting any fishing activity.
In 1982 a group of 8 member states in the Pacific enacted the Nauru Agreement Concerning Cooperation in the Management of Fisheries of Common Interest (commonly called the Parties to the Nauru Agreement or PNA), which sought to limit the overfishing of skipjack tuna by foreign fleets.
The arrangement allows PNA members to sell fishing rights to foreign ships at a set price according to science-based assessments of current fish stocks. This has led to a more sustainable fishing practice that protects the local waters and the livelihoods of these island states who rely on them.
Skipjack tuna being harvested using purse seines.
The interstitial high-seas areas, not protected by the PNA, stand out starkly as bright wedges of intense fishing activity between the PNA member states' EEZs in the Global Fishing Watch layer.
NOAA Marine Protected Areas (MPA)
Protection of U.S. waters, which combines state and federal oversight programs, is defined by NOAA’s Marine Protected Areas (MPA) Inventory , also available in Living Atlas. Updated yearly, these areas meet the specific requirements defined as international protected areas according to the Union for Conservation of Nature (IUCN) .
MPA areas are classified according to 6 levels of protection, shown in the legend below. Visit NOAA for more information on MPAs and their use in marine conservation.
California’s coast and ocean resources are some of the most prized in the United States – and a foundation of the state’s economy and identity. Monterey Bay, and its signature canyon and fan, bring deep-water ocean life close to shore, resulting in a remarkably productive and diverse marine ecosystem.
Established in 1992, Monterey Bay National Marine Sanctuary (MBNMS) is a federally protected marine area offshore of California's central coast, stretching from Marin south to Cambria.
Within this blue multiple-use zone, the Soquel Canyon and Portuguese Ledge State Marine Conservation areas (outlined below) are two of the deepest Marine Protected Areas on the west coast. While fishing is allowed here, commercial and recreational takes are restricted to Pelagic Finfish only.
3D view of Soquel and Portuguese Ledge canyons.
Use the time controls at the bottom of the map to animate monthly fishing intensity in Monterey Bay.
In 2003, the State of California Fish and Game Commission established 13 Marine Protected Areas (MPAs) within the state waters of Channel Islands National Park, off the Southern California coast, providing protection through research, education, conservation, and stewardship.
A view of the Channel Islands, looking west from Inspiration Point on Anacapa Island.
The light green area defines the Channel Islands National Marine Sanctuary, while the orange areas represent a mix of State (within 3 nautical miles of shore) and Federal (offshore) no-take fishing areas.
Unlike Soquel Canyon, absolutely no fishing activity is permitted in the orange no-take zones. Overlaying Global Fishing Watch fishing intensity for February 2020 with AIS tracks from that same period shows that most vessels are careful to cross over into the sanctuary zones.
However, a closer look the edge of the Gull Island (Santa Cruz Island) State Marine Reserve reveals two pixels with apparent fishing activity, according to Global Fishing Watch - supported by the tracks of several vessels in that region at the time.
This is not conclusive proof of any wrong-doing, though if patterns of apparent fishing activity regularly encroach a Marine Protected Area, additional public outreach, monitoring and enforcement may be warranted.
While the Global Fishing Intensity layers represent aggregate, rasterized views of vessel activity, adding individual vessel AIS tracks, as seen in Monterey Bay and the Channel Islands protected areas, can provide an additional level of detail. While we don't know the activities of any vessel from minute to minute, and Global Fishing Watch's detection algorithm is not 100% accurate, it is still a useful resource for exploring fishing patterns and vessel frequency, particularly in the most protected "no-take" areas.
U.S. Vessel Traffic App
Joining the Global Fishing Watch release are updates to the U.S. Vessel Traffic web app, which enables anyone to explore AIS vessel tracks for U.S. waters. Covering the period from January 2015 to June 2024, the app provides monthly views of 8 different categories of vessel traffic, with individual vessel information available through its unique Maritime Mobile Service Identity (MMSI) number.
Use the app below to explore the tracks of fishing (or other) vessels in U.S. coastal waters and rivers.
vessel-traffic
Clicking on an individual line highlights that vessel's track and provides information about the name, date, speed/direction, unique MMSI number, and length, beam, and draft. This information is then conveniently stored, along with the current extent, in the URL for easy sharing.
Pro tip: add &hideBottomPanel=true to the end of the app URL to hide the bottom user interface.
An individual vessel's track off the Oregon Coast in May, 2020. Click on the map to select a different vessel track.
Track of the Seattle Enterprise fishing vessel hugging the California-Oregon regulatory border. ( MMSI 369604000 )
For GIS analysts, one of the most useful functions in the U.S. Vessel Traffic app is the ability to download an entire month of vessel tracks, using the Download option in the lower-right. These are delivered in a zipped geodatabase, and can be used to understand vessel route density, encroachments into prohibited areas, or for other visualization or animation tasks.
As we've seen in this story, Global Fishing Watch has the potential to transform how we protect and manage our oceans, by illuminating activity that was previously hidden from sight. Used in combination with other authoritative Living Atlas layers, these insights can help transform understanding into action, driving local initiatives and partnerships that reduce overfishing and habitat destruction.
The importance of our oceans as a critical global resource cannot be over-stated. As vessel monitoring evolves through improved detection models and technological innovations, so can our social and political behavior, restoring ecosystems and protecting them for future generations.
A school of tuna fish near the Galapagos Islands.
Resources
We're excited to add these new Global Fishing Watch layers to ArcGIS Living Atlas. If you have additional questions or feedback, feel free to contact the Living Atlas Environment Team , we'd love to hear from you!
Finally, click the cards below for additional information and resources mentioned in this story.
