News

Researchers to use ‘big data’ to predict sea crimes


Researchers using artificial intelligence and “big data” plan to develop new algorithms that they say will enable them to identify, locate – and eventually predict – crimes committed in the world’s oceans, from illegal fishing off the Patagonia shelf to drug smuggling in Central America to slave labor and human trafficking in the Indian Ocean.

Mark Floyd
mark.floyd@oregonstate.edu

Dark Fleet
The perpetrators of these illegal, unreported and unregulated (IUU) activities collectively use vessels called “the dark fleet,” not just because of their criminal activity, but because they try to hide their location by turning off their GPS tracking systems and navigating between legally operating and visible boats.
“IUUs include all kinds of terrible things,” said James Watson, a marine scientist expert at Oregon State University, and a principal investigator on the project. “We came into this thinking primarily about illegal fishing, but that turns out to be just the tip of the iceberg. It is much, much bigger.”
The problem often begins, however, with illegal fishing, which is widely recognized by governments around the world, the seafood industry and many citizens as a major problem, according to Jane Lubchenco, an Oregon State researcher and co-principal investigator on the project who formerly served as the National Oceanic and Atmospheric Administration (NOAA) Administrator.
“Illegal fishing threatens food security, compromises ocean health, penalizes honest fishermen and fishing companies and is all-too-often linked to slave labor, human trafficking and drug trafficking,” Lubchenco said. “The issue has been what to do about it. Fortunately, the world has recently made huge strides in creating policy agreements, mobilizing political will and harnessing technology and artificial intelligence to ‘see,’ crack down on, and deal with some of the illegal fishing.
“In truth, these changes have happened at a breakneck pace for governments and industry – a pace that has surprised me and that gives me hope. It is a tough problem, but it is solvable.”

Grant funding
The researchers, all from Oregon State, are using a three-year, $860,000 grant from the National Aeronautics and Space Administration (NASA) to develop their computational tools. Their first step is to synthesize the data on the location of a majority of the world’s vessels – data that only has become available within the last year.
They believe they can identify illegal activities at sea in part by watching the behavior of nearby vessels operating legally – much in the same way that astronomers identify far-away planets without really “seeing” them, but instead based on the behavior of other celestial bodies in response.
Once they have their algorithms, they will test them by using training datasets on known instances of maritime IUU activities, and then develop a model for predicting these activities in real time, and developing appropriate intervention tools.
Watson said one trick is to look at when a suspect vessel goes into a suspicious area, turning off its GPS system as it enters.
“One can infer that they are participating in illegal fishing from their activity,” he said. “There are tons of data out there – it’s a matter of synthesizing it and recognizing patterns. There is a behavioral psychology aspect to it. When I study fish, I look not only at the behavior of an individual fish, but also how the school moves.
“When a vessel begins an illegal activity, the legal boats often begin acting differently. They get out of the area. Those anomalous responses can be telling.”

Suspicious areas
Many “suspicious” areas of illegal activity are well-known. In a matter of weeks, giant trawlers from Europe and China illegally harvested an amount of fish off the Somalian coast equal to five years of harvests by legal, local fishermen. Much of the narcotics trafficking from Central America is handled by boat. And slavery at sea has been an issue in Indonesia, the South China Sea and coastal communities in southeast Asia, the researchers say.
“We have solved part of the problem through politics and political will, and now we need to tackle the rest,” Lubchenco said. “We can use our current tools to see and clamp down on only part of the illegal fishing fleet. That’s where this project is so exciting – we can devise novel ways to vastly expand the numbers and kinds of fishing vessels that might be engaged in illegal activities.”
Watson said the data will be made available to nation states and their coast guards and navies, fishermen, conservation groups, humanitarian NGOs and shipping companies – “anyone interested in, or has a vested interest in, the oceans and what happens at sea.”
Maria Kavanaugh and Jamon Van Den Hoek of Oregon State University also are principal investigators on the project.

About the OSU College of Earth, Ocean, and Atmospheric Sciences
CEOAS is internationally recognized for its faculty, research and facilities, including state-of-the-art computing infrastructure to support earth/ocean/atmosphere observation and prediction. The college is a leader in the study of the Earth as an integrated system, providing scientific understanding to complex environmental challenges.