Using satellite imagery and AI, researchers have mapped human activity at sea more accurately than ever before. The effort uncovered a huge amount of previously under-the-radar industrial activity, from suspicious fishing operations to a boom in offshore energy development.
The cards were published today in the magazine Nature. The research, led by Google-backed nonprofit Global Fishing Watch, found that as many as three-quarters of the world’s industrial fishing vessels are not publicly tracked. Up to 30 percent of transport and energy ships also escape public tracking.
Those blind spots could hinder global conservation efforts, the researchers say. To better protect the world’s oceans and fisheries, policymakers need a more accurate picture of where people are exploiting marine resources.
“The question is which 30 percent should we protect?”
Nearly every country on Earth has agreed to a shared goal of protecting 30 percent of the Earth’s lands and waters by 2030 under the Kunming-Montreal Global Biodiversity Framework adopted last year. “The question is which 30 percent should we protect? And you can’t have discussions about where the fishing activities are and where the oil platforms are located unless you have this map,” says David Kroodsma, one of the authors of the Nature paper and director of research and innovation at Global Fishing Watch.
Until now, Global Fishing Watch and other organizations have mainly relied on the maritime Automatic Identification System (AIS) to see what was happening at sea. The system tracks ships carrying a box that emits radio signals, and the data has been used in the past to document overfishing and forced labor on ships. However, there are major limitations to the system. Requirements for the carriage of AIS vary by country and vessel type. And it’s quite easy for someone to turn off the box if they want to avoid detection, or cruise through locations where signal strength is spotty.
To fill in the blanks, Kroodsma and his colleagues analyzed 2,000 terabytes of images from the European Space Agency’s Sentinel-1 satellite constellation. Instead of taking traditional optical images, which is like taking photos with a camera, Sentinel-1 uses advanced radar instruments to observe the Earth’s surface. Radar can penetrate clouds and ‘see’ in the dark – and could spot offshore activity that AIS had missed.
Since 2,000 terabytes is a huge amount of data to process, the researchers developed three deep learning models to classify each detected ship, estimate their size and figure out different types of offshore infrastructure. They monitored about 15 percent of the world’s oceans, where 75 percent of industrial activity takes place, paying attention to both ship movements and the development of stationary offshore structures such as oil rigs and wind turbines between 2017 and 2021.
As fishing activity fell at the onset of the Covid-19 pandemic in 2020, they noted heavy vessel traffic in public monitoring systems in areas that “previously showed little to no vessel activity” – especially around South and South-East Asia, and the North and west of the country. coasts of Africa.
A boom in offshore energy development was also visible in the data. At the end of 2020, there were more wind turbines than oil structures. The following year, turbines made up 48 percent of all ocean infrastructure, while oil structures accounted for 38 percent.
Almost all offshore wind development took place off the coasts of Northern Europe and China. In the northeastern US, clean energy opponents have tried to falsely link whale deaths to the coming development of offshore wind energy, even though evidence points to ship attacks as the problem.
Oil structures have many more ships swarming around them than wind turbines. Tankers are sometimes used to transport oil to shore as an alternative to pipelines. The number of oil structures grew by 16 percent in the five years studied. And offshore oil development was linked to five times as much global shipping traffic in 2021 as wind turbines. “The actual amount of ship traffic worldwide originating from wind turbines is small compared to the rest of the traffic,” says Kroodsma.
When asked whether this kind of research would have been possible without artificial intelligence, “the short answer is no, I don’t think so,” said Fernando Paolo, lead author of the study and a machine learning engineer at Global Fishing Watch. “Deep learning excels at finding patterns in large amounts of data.”
New machine learning tools being developed as open source software to process global satellite images “democratize access to data and tools and enable researchers, analysts and policymakers in low-income countries to leverage tracking technologies at a low cost,” says another article published in Nature today in which comments are made on Paolo and Kroodsma’s research. “Until now, no comprehensive, global map of these different types of maritime infrastructure has been available,” says the paper written by Microsoft postdoctoral researcher Konstantin Klemmer and University of Colorado Boulder assistant professor Esther Rolf.
The technological advances come at a crucial time for documenting rapidly evolving changes in maritime activity, as countries try to halt climate change and protect biodiversity before it is too late. “The reason this matters is because it’s getting busier and busier [at sea] and it becomes more and more used and suddenly you have to decide how we are going to manage this gigantic global commons,” says Kroodsma. THe Edge. ‘This can’t be the Wild West. And that’s how it has been historically.”