Oceans and AI: The quest for sustainable fisheries

This is a photo of Fresh caught herring at the local fishing cooperative in Freest, Germany in February 2018. (Image credit: Photo by Bernd Wüstneck/picture alliance via Getty Images)

Fresh caught herring at the local fishing cooperative in Freest, Germany in February 2018. (Image credit: Photo by Bernd Wüstneck/picture alliance via Getty Images)

This article debuts a five-part series: Oceans and AI, which explores the practical application of technology to address threats to our planet’s seas and fish.

Ocean fisheries have been in and out of trouble since the late nineteenth century when large-scale steam-powered fishing fleets first ravaged fish populations. A Royal Commission formed in 1880 placed most of the blame for the decline of fisheries in New South Wales, Australia, on “wanton” netting, in particular, its effects on small fish. Other factors it mentioned included sewage pollution and boat traffic. Sound familiar?

Today those problems seem trivial. Factory-sized, mega-vessels comb the oceans with state-of-the-art technology and plunder a resource that once was considered inexhaustible. Over the past 100 years, the fishing industry has embraced larger, more powerful fishing vessels equipped with sophisticated gadgetry like radar, sonar, LORAN, and GPS. The fish don’t have a chance.

Since 1961, annual fish consumption has been twice the rate of global population growth. A dramatic fall of 74-percent in worldwide stocks of fish including mackerel, tuna, and bonito occurred between 1970 and 2010, according to the Living Blue Planet Report from the World Wildlife Fund and the Zoological Society of London.

Our planet’s growing population and the demand that people are placing on seafood stocks is dramatic. Fisheries supply 3.2 billion people with almost 20-percent of their average intake of animal protein, even more in poor countries, according to United Nations data.

This chart shows Recent data suggests that fish populations may not keep pace with human populations (Source: UN FAO, 2018)

Recent data suggests that fish populations may not keep pace with human populations (Source: UN FAO, 2018)

Yet the same advanced technologies accelerating the decline of fisheries can help monitor fishing practices and capture information to help threatened populations. Better data collection of fish catching practices, as well as ongoing education, promises to help slow the trends threatening global ocean fisheries.

On the hook

Much of the progress in the application of big data and predictive analytics for the monitoring of fisheries comes thanks to entrepreneurs like Amos Barkai, who embarked on an entrepreneurial journey more than 18 years ago in Cape Town, South Africa. He saw that commercial fishing was lagging far behind other industries when it came to adopting information technologies, so he developed electronic monitoring software that would record, report, and manage commercial fishing data.

This photo shows Amos Barkai, CEO of OLSPS, at the 2017 UN Ocean Conference. (Photo credit: IISD Reporting Services)

Amos Barkai, CEO of OLSPS, at the 2017 UN Ocean Conference. (Photo credit: IISD Reporting Services)

But when Barkai approached local fishers with his vision for data capture and analysis, they dismissed his proposal. The thought of having computers on board that would monitor and send data to their regulators seemed ridiculous to them. So Barkai reset his expectations, realizing he’d have to educate fishers on the benefits of using computers, data capture, and cameras to gather information. Nothing less than the health of the local fish populations was at stake. He adopted a more gradual approach.

Amid declining yields due to environmental and human pressures on the oceans, Barkai eventually won over the local fishers. They came to undertand that by reporting data about fish location and yields, they would be able to more precisely target areas for fishing and thereby save valuable fuel in the process. In the end, it was about saving money. Operating large fishing vessels cost anywhere from $10,000 to $50,000 per day to operate.

Today, Barkai’s company OLSPS Analytics provides software for predictive analytics, biological modeling, digital image capture, and electronic monitoring (EM). And as EM has been adopted by regulators and put to work more widely in commercial fishing fleets, Barkai has incorporated machine learning and deep learning into his software.

Barkai’s drive to develop new fishing technologies has helped plot a course for such other fish-minded entrepreneurs as Anchor Lab and Fish-e. But more importantly, products like Barkai’s Olrac Electronic Logbook (eLog) Solution helped win over a generation of fishers, making it easier for younger, tech-savvy people entering the trade today to adopt potential fishery-saving software and monitoring technologies.

Monitor first, save later

In Europe, there was also strong initial resistance by fishers to the use of EM technologies. These consist of onboard video cameras, gear sensors, and image processing software to capture and record information on fishing location, species and quantities of fish caught, and those discarded. But they didn’t gain acceptance until 2008 when the European Union fisheries’ governing body required fishing vessels to use them. Across the E.U. today, fisheries are managed through a combination of quotas, effort restrictions, and the comprehensive documentation of caught fish.

Ten years on results from the current E.U. and U.S. electronic monitoring programs reveal that that EM works. By increasing the geographic areas of monitoring and by improving data collection, fisheries managers to more effectively oversee fish stocks.

Other pioneering work to apply technology to ocean health has been done by the National Ocean and Atmospheric Administration (NOAA) and the Nature Conservancy. On the leading edge of both organizations is the use of AI in their electronic monitoring efforts to identify fish species caught, monitor ocean currents, keep track of dead zones, and help measure pollution levels.

This photo shows Simple video cameras can capture comprehensive data on caught fish (Image credit: NOAA)

Simple video cameras can capture comprehensive data on caught fish (Image credit: NOAA)

This photo shows An onboard video camera records a skate catch (Image credit: NOAA)

An onboard video camera records a skate catch (Image credit: NOAA)

NOAA’s experimental fieldwork involves the electronic monitoring of fish populations while they’re still underwater. One device debuted in 2015 uses a low-powered computer and sonar system attached to the seafloor to automatically survey fish populations. The device wakes up when it detects fish movement and sends out pings to determine the number of fish present.

In the United States, EM programs are becoming more established, and more sophisticated. Onboard cameras already track bluefin tuna catches across five major Atlantic fisheries. In the Northeast, there are pilot projects underway for the groundfish fishery and the midwater-trawl herring mackerel fishery, with full implementation expected in the next few years. EM will be fully implemented for the West Coast whiting and fixed gear fisheries, and for the bottom trawl fishery and non-whiting midwater trawl fisheries by 2019. It will also be implemented this year for Alaska’s small boat fixed gear and pot fisheries.

“Our electronic monitoring program in Alaska has been using these technologies for several years,” says Jennifer Ferdinand, director of NOAA’s Alaska fisheries monitoring and analysis division. “It has not adopted AI yet but it is coming. AI systems are especially good with species identification. Sometimes the systems are more accurate and faster at being able to ID fish than people. Video images are brought back and reviewed on shore.”

“The good news is these fisheries are remarkably resilient systems and if we can take pressure off these systems quickly, they’re capable of a great deal of recovery,” says Mark Zimring, director of the Nature Conservancy’s Tuna Program.

In this effort, monitoring and measurement are critical because, as Zimring notes, “You can’t manage what you can’t measure.”

Of course, technology won’t be a panacea — especially if fishers are unable to afford and equip their boats with it. A recent United Nations report cautioned new technologies will do little for fisheries that lack the capacity or financial resources to adopt them.

Yet, technology may be our best shot at understanding the complex problems that threaten fisheries’ production. From there, these problems will need to be mitigated with smarter, more informed approaches to ensure the planet’s greatest source of food can thrive.