Precision fisheries of the future will rely on data and AI to improve profits
The fishing industry is catching up to the data-driven digital revolution already remaking land-based agriculture and other industries.
And the fishing industry doesn’t need to wait for the future to arrive: the technology and advanced analytics that could improve environmental sustainability and increase profitability are available today, a recent report from global consulting and management firm McKinsey & Company concludes.
Fishing companies could start by mining efficiencies out of existing data, such as vessel catch totals, GPS location, and fuel consumption patterns, the report advises. Advanced analytics could uncover performance drivers, lowering costs. Fuel consumption data could help companies discover the most energy efficient routes and maneuvers, while data from onboard sensors could be processed to predict breakdowns in advance.
But the fishing industry has some catching up to do, especially compared to other primary production industries such as farming and forestry, according to McKinsey senior partner and report co-author Philip Christiani.
"Precision agriculture and forestry are quite well developed globally, while precision fisheries remains limited," Christiani told SeafoodSource. "There is already a sea of data out there [including] vessel position, catch data, sea conditions. The big challenge is how to better organize them and bind them together to create insights."
Though the use of advanced analytics in fisheries is currently limited to small-scale projects, McKinsey estimates that large-scale fishing companies could lower their annual operating costs by USD 11 billion (EUR 9.9 billion) by using precision fishing methods and technologies. Precision fishing – and the resulting improvements in ocean management – has the potential to raise profits for the industry by as much as USD 53 billion (EUR 47.5 billion) by 2050, while also doubling fish biomass from its current level.
The precision fishing revolution is possible because of the growing torrent of data from compact low-cost sensors, satellite images, cameras, drones, and other technologies. Smart phones and Internet-of-Things devices communicate this data across wireless cellular or satellite networks to robust data processing centers that rely on machine learning and artificial intelligence.
Data come from optical sensors on satellites that can detect sea temperature, turbidity and other variables, while radar sensors can gain information about topography, winds, sea ice, and vessel movement. Radar can collect information even when the sky is dark or cloudy. Airborne and floating drones can provide additional information about oceanographic conditions.
These tools have the potential to drive better decision-making tools that help balance complex goals such as profitability and sustainability. They could address uncertainty around catches and help manage risks, while improving reporting methods – tamping down illegal fishing and labor abuses.
Image recognition using onboard cameras paired with software could help fishermen document catch in real time. Predictive models of fish distribution that rely on data about ocean conditions could help fishermen expend less effort to catch fish, while better avoiding bycatch.
One company that is leading the digital transformation of fishing is Aker BioMarine, which catches and processes Antarctic krill for aquaculture ingredients and products for human consumption.
The company is using machine learning to predict where krill biomass is at any given time, eliminating the 15 days at sea that vessels currently spend burning fuel while searching for krill rather than harvesting it. The software promises to improve on captains’ decades of experience by meshing past and present data from satellite images, algae blooms, weather, water conditions, and actual fish catches.
"We are at the level where the model predicts as good as our best captains … As we get more data into the model, we feel pretty confident that we will be able to predict better than the humans,” CEO Matts Johansen told SeafoodSource. "With this model we don't have those search days any more, so we can harvest more and have a lower carbon dioxide footprint doing it.”
Aker BioMarine is also using artificial intelligence to optimize vessel engines and factory operations, collecting data from onboard sensors and feeding recommendations to vessel captains and factory managers.
"If you can tweak the temperatures just a little bit, you can optimize the speed to the most fuel-efficient speed and you can use self-learning models to do this," Johansen said. "The end goal is to fully automate it, so you can set the specification and it runs by itself.”
Autonomous, solar-powered sail buoy drones 1.5 meters long collect data about ocean conditions and biomass and transmit it via satellite to Aker BioMarine, which then runs it through its algorithms to predict biomass. The company is partnering with fishery managers to share this data and improve management.
By equipping regulators with better data, Aker BioMarine hopes be able to maintain current catch levels; Johansen said the krill fishery is managed very conservatively because regulators don't have enough data to allow higher catch levels.
"Our aim here now is not to increase the quota, but at least make sure we can continue fishing at the current level and make sure everyone feels comfortable with the current level," Johansen said.
Photo courtesy of Aker BioMarine