All posts by Maddy Bergum

(Image courtesy of the New York Times)

If diverse datasets and geospatial information can help fine-tune global distribution networks, they can also be used to hunt them. In the fight to uphold arms embargoes around the world, the geospatial industry’s proven track record of weaving together disparate data and imagery to form a clear picture provides major assistance.

In 2018, analysis of satellite imagery was able to establish an arms embargo breach in South Sudan. In 2015, before any embargo was in place, Ukraine sold four or more Mi-24 attack helicopters to South Sudan. When an embargo went into place in July 2018, the helicopters were grounded at two different locations – a government security base in Luri and at the Juba International Airport. None were in flyable condition, and any repairs to them would violate the embargo.

Using a total of 308 satellite images, Amnesty International constructed a timeline of the helicopters’ locations since the embargo’s beginning. The analysis showed that, in October and November of 2018, the helicopters had undergone maintenance, and been moved between bases. New rotor blades were one visual clue they used to make this determination.

As discussed in our post about the geospatial industry and illegal fishing, location data from boats at sea is also helpful when searching for unlawful behavior. This applies to ships carrying embargoed weapons, too.

In March 2020, a BBC investigative report used satellite images to show that Turkey sent tanks and weapons to Libya via ship shortly after agreeing to a UN arms embargo. GIS Tracking indicated the ships left Turkey, on their way to Tunisia, but the ships’ transponders were shut off three days after its departure. Satellite images from that day showed the ships off the coast of Libya. When the ship docked at Genoa, Italy, authorities questioned the sailors, and one sailor confessed to being part of the arms trade.

Sometimes confessions of agreement-breaking are not so easy to obtain, though. Fortunately, geospatial data doesn’t lie.

Jeffrey Lewis and Melissa Hanham are experts on North Korea’s nuclear program with the East Asia Nonproliferation Program. The program is located at the Center for Nonproliferation Studies (CNS) at Middlebury’s Institute for International Studies at Monterey.

They used seismic signatures from the ground motion after five of North Korea’s nuclear tests. Combining this data with satellite imagery, they were able to build a three-dimensional map of the test site. Their ability to show the approximate dimensions of the North Korea site indicated that it was being modeled after US test tunnels. Their research utilizing geospatial data uncovered a significant amount of highly confidential information.

Key Takeaways:

• Arms embargoes are often difficult to monitor, which makes them hard to enforce. Thanks to the geospatial industry’s advances in satellite imaging and diverse data analysis, these kinds of situations can be monitored and tracked remotely.
• A creative combination of geospatial data, ship tracking records, and on-the-ground personnel can make a big difference, and stop embargo-breaking before it leads to violence.
• Satellite images can help piece together the movements of otherwise unsupervised equipment, while geospatial information can give analysts ideas about larger infrastructure, like nuclear testing sites.

(Image courtesy of Rumbos)

The effects of illegal fishing are felt worldwide: from environments to economies. Until recently, tracking illegal fishing vessels, or even legal ones, was difficult for a number of reasons. The biggest: oceans are enormous – one ship can lose itself easily in such a vast arena. Only a fraction of their waters have ever been explored, and there are few ‘landmarks’ visible to the human eye.

That’s no longer the case, thanks to a convergence of radar, geospatial data, AIS, and data industry players.

Automatic Identification System, or AIS, was initially developed as a safety tool to track ships on their voyages. Since its initial use in the 1990s, AIS has become increasingly common, and as ships’ systems automatically check-in, that data is collected by satellite. The data grows by the day, and, over time, it can reveal patterns.

(Image Courtesy of Cult of Sea)

Ships only report an identification number and location through AIS, though. To make tracking more accurate, machine learning is being taught algorithms that can tell a commercial vessel from a yacht. To build these algorithms, human observers had to parse immense amounts of data and compile models for patterns of different kinds of vessels.

In September of 2016, SkyWatch, Oceana, and Google teamed up to analyze this immense dataset (the project began with over 3.7B data points). Their goal was to examine illegal fishing over time and provide the public access to this data, too. The initiative is called Global Fishing Watch (GFW).

The initiative is working to push back against the overfishing of 33% of the world’s major commercial fish species and to put a halt to the yearly loss of over 23 billion dollars in stolen, unreported, or unregistered seafood. Their biggest weapon: diverse datasets and geospatial information.

Today, satellites can see through clouds, which helps researchers track the movements of ships not using AIS. Trouble is, while satellite radar can see through clouds, it can’t automatically differentiate between fixed pieces of infrastructure (oil rigs, wind turbines, etc) and moving ships. Using a series of algorithms and maps, Global Fishing Watch is working to establish a database that includes fixed infrastructure, which will make ships easier to pick out of the data.

GFW can also help determine the efficacy of policies & regulations, such as bans. A recent example: their data indicated that Indonesia’s recent ban on foreign fishing vessels has resulted in an enormous drop in fishing activity there. This outcome might not be immediately recognizable based on data from fish imports or population counts. In short, geospatial data and innovative technologies like machine learning are allowing for a crackdown on illegal fishing like never before. With the help of Global Fishing Watch, the illegal fishing industry is dwindling rapidly with the hopes of completely eradicating the practice on a long term scale.

Key Takeaways:

• Illegal fishing is a problem with global implications touching everything from the environment to the economy.
• AIS was developed to keep track of ships for safety, but can now be used to detect illegal fishing patterns.
• When combined with satellite radar and data analysis, AIS data can be put into machine learning algorithms that will allow researchers, activists, and governments to act on solid information about what kinds of vessels are fishing in their waters.