Skylight’s Ted Schmitt: Technology can turn tide of war against IUU

Ted Schmitt, director of conservation and head of the Skylight program at the Seattle, Washington, U.S.A.-based Allen Institute for AI (AI2).

Perpetrators of illegal, unreported, and unregulated (IUU) fishing are starting to feel the heat from satellite monitoring solutions and artificial intelligence, according to Ted Schmitt, director of conservation and head of the Skylight program at the Seattle, Washington, U.S.A.-based Allen Institute for AI (AI2). The institute, created by the late Microsoft co-founder, Paul Allen, operates Skylight, a free technology platform using maritime monitoring, analysis software, computer vision, and machine learning to “deploy models that can surface suspicious activity in real-time,” according to AI2.

Skylight is also working with satellite imagery from Sentinel 1, a constellation of polar-orbiting satellites operated by the European Space Agency, allowing it “to move from capturing one percent of the ocean once a month to 17 percent of the ocean twice per month.” Using this technology, Skylight can monitor in eight hours what would take a person 800 hours to cover.

Skylight works with developing nations but also with naval enforcement bodies globally, including the U.S. Coast Guard. It recently joined the Joint Analytical Cell, a new collaboration to give lower-income coastal states better access to fisheries intelligence, data analysis, and capacity-building assistance in the battle against illegal, unreported, and unregulated (IUU) fishing.

In an interview with SeafoodSource, Schmitt said the rapid advances in computer and satellite technology are beginning to bear fruit in the fight against illegal fishing.

SeafoodSource: Can you share any practical examples or incidents where your monitoring services have been brought to bear in tracking IUU and/or assisting coastal states to make a successful intervention?

Schmitt: In the western Indian Ocean, fisheries monitoring centers use Skylight to identify, track, and document vessels fishing in restricted areas. In a recent instance, teams monitoring a sensitive coastal area identified several vessels illegally trawling for shrimp. The analysts took screenshots of the vessel’s tracks as evidence, complemented by the vessel monitoring system (VMS) [data] of the illegal activity. The threat of sanctions for a second offense has thus far been enough to observe the vessels respecting the restricted areas.

In West Africa, Skylight is supporting a national parks agency protecting a network of marine protected areas (MPAs). Before implementing Skylight into their operations, the agency was using VMS. This gave them great insights into the movements of their national fleet, but wasn’t designed to track foreign vessels that may be attempting to fish in these protected areas.

Today, any time a vessel enters one of these MPAs, the platform is set up to alert the [relevant] maritime analysts. In one such case, a foreign vessel was identified entering a restricted MPA and the team took immediate action to prevent the vessel from fishing in the protected area. To further support these agencies efforts to tackle the IUU fishing crisis and better understand what’s happening in their waters, Skylight continues to develop methods to detect suspicious behavior, including leveraging satellite imagery to detect vessels who are not transmitting their location. Most recently, this includes vessel detection from Sentinel-1 satellite radar, while additional sources should be available in the Skylight platform within the next couple months.

SeafoodSource: Do you have any indication that perpetrators of IUU are changing their behavior as a result of the enhanced monitoring?

Schmitt: [Recent] behaviors of vessels would indicate yes. We are noticing vessels stop transmitting their locations through vessel tracking systems like automatic identification systems (AIS) to evade detection near protected or restricted areas such as marine protected areas or exclusive economic zones. We are also noticing sophisticated methods such as AIS spoofing or scrambling, resulting in incorrect or missing AIS data. This suspicious behavior is likely tied to illegal activity. This, of course, means we have to up our game … to detect the “dark” vessels, [by using] satellite imagery such as Sentinel-1 [and other] sophisticated computer vision techniques.

SeafoodSource: Do you think Skylight’s monitoring can help improve seafood market traceability efforts at the point of entry to major seafood markets?

Schmitt: Yes, one of the best tools to keep stolen fish out of major seafood markets is the Port State Measures Agreement (PSMA). To give this policy teeth, countries and NGOs are using Skylight to identify suspicious activity, such as surfacing potential transshipment events for port authorities implementing PSMA measures.

An example of this in action is how Stop Illegal Fishing (SIF) uses Skylight to help its partner, South Africa’s State Security Agency, tackle IUU fishing. Skylight’s advanced machine learning algorithm alerted SIF to a dark rendezvous leading to fishing vessel Torng Tay No. 1’s request for entry into the Durban, [South Africa] port. When SIF’s team of analysts took a closer look at the vessel’s history, they found the fishing vessel was loitering for almost four hours, plenty of time for the ship to transport fish to or from another vessel. While most cases of transshipment at sea are legal, this practice can hide IUU fishing practices. When inspected by the South African authorities, it was found that the fishing vessel underreported to the government the amount of fish on board. The fishing vessel was fined by South African authorities. If the country catches the vessel Torng Tay No. 1 illegally fishing again, the vessel will then be fined again at 10 times the original fine.

Photo courtesy of University of Washington

Subscribe

Want seafood news sent to your inbox?

You may unsubscribe from our mailing list at any time. Diversified Communications | 121 Free Street, Portland, ME 04101 | +1 207-842-5500
None