Small-scale fishers and coastal communities are increasingly turning to digital tools to help them be more sustainable and address the issues of climate change. Examples of these tools include weather prediction applications and Artificial Intelligence used to monitor the fish they catch. 

Environmental organisations claim that commercial vessels' overfishing, and illicit fishing pose serious harm to fisheries and the ecosystem, depriving millions of people in coastal towns who depend on fishing for food and jobs. 

More than 90% of all capture fishermen and fish workers worldwide are small-scale fishermen. They are also impacted by rising sea levels, an increase in storm frequency, and fish that migrate to new places in search of calmer waters or if their habitat is destroyed. 

What the study elucidates?

A recent study that uses satellite technology and Artificial Intelligence to gather data on fishing, shipping, and energy production on the oceans reveals that, for the most part, everyone has been flying blind. Researchers discovered that over one-third of the transportation and energy sectors and more than three-quarters of industrial fishing activities are not included in public tracking systems, making them virtually invisible on a global scale. 

The research was published in the journal Nature and came from Global Fishing Watch (GFW), a vessel-tracking technology developed by Google in collaboration with the non-profits Oceana and SkyTruth. Using 2 petabytes of satellite imagery and deep learning models, analysts mapped industrial activity over 15% of the ocean between 2017 and 2021. They then contrasted the results with activity found in publicly accessible tracking data. 

The vessels' illegal deactivation of the GPS-based automatic identification systems (AIS), which transmitted their positions, led to the imposition of financial penalties. This is frequently an indication that a boat is fishing illegally. 

"As much as 20% of what's caught is thought to be illegal or unregulated fishing," states David Kroodsma, head of Global Fishing Watch's research team. 

The UN Food and Agriculture Organization estimates that illegal, unreported, and unregulated (IUU) fishing may take up to 26 million tons of fish annually, or one in every five fish. According to reports, the black market has a potential value of up to $23 billion (£18 billion). 

According to the UN, a third of the world's fish stocks are currently being fished at levels that are biologically sustainable, which greatly adds to overfishing. For instance, the population of bluefin tuna is thought to be only 2.6% of what it was when it was historically unfished. 

How does this AI software work? 

Global Fishing Watch is now mapping the movements of over 65,000 commercial fishing vessels, both those with and without AIS, using satellite imagery and increasingly advanced AI technologies in an effort to better monitor and quantify the overfishing problem. 

Millions of gigabytes of satellite footage are analyzed by the AI to find offshore infrastructure and ships. Then, in order to identify vessels that neglect to broadcast their whereabouts, it examines data from ships' AIS signals that is available to the public and combines it with radar and optical pictures. 

Furthermore, even though not all ships are compelled by law to use AIS, the AI's "fishing detection algorithm" attempts to determine which of these "dark" vessels are most likely to be fishing. 

"We use information like the length of the vessel, plus environmental information as to where that vessel is located, the imaging of the area, the density of vessel traffic in the area, the state of the ocean, like the temperature - a bunch of information on where the vessel is operating," says Fernando Paolo, senior machine learning and remote sensing engineer at Global Fishing Watch. 

"And this model infers whether the vessel is likely a fishing vessel or not a fishing vessel, like passenger vessels, oil tankers, shipping vessels and so on." 

Three fourths of the industrial fishing vessels worldwide are not publicly tracked, according to the study, with South Asia and Africa being notable hotspots. 

Offering a higher-resolution imagery 

The goal of Global Fishing Watch's current initiative is to provide higher-resolution imagery to aid in the detection of more tiny fishing boats. 

But eradicating illicit fishing requires gathering information as soon as possible, and the University of Southampton and the nearby company RS Aqua are working together to achieve this goal through a project. 

Underwater robot, sensors and AI 

The team is developing an underwater robot that can recognize the sound of fishing and communicate that information in real time using underwater sensors and Artificial Intelligence. The AI is currently being trained to recognize the noise made by trawlers operating in protected seas. It can already distinguish between the sounds of the ocean naturally. 

Nature

Photo by Cheng Zi on Unsplash