How Artificial Intelligence Can Help Predict Toxic Algal Blooms


Cawthron Institute scientist Mike Packer in algal growth room.

Cawthron Institute

A team at the Nelson Artificial Intelligence Institute has developed technology they claim can detect tiny algal cells in the ocean before they multiply to create toxic algal blooms.

Toxic algal blooms, also know as ‘red tides’, are are a human health hazard, damage ecosystems, close fisheries, and hurt the economy. But if a bloom is accurately predicted before it begins, fisheries management can plan for closures in advance.

Shellfish fisheries in particular stand to gain a lot from this type of technology. Shellfish like oysters and mussels are filter feeders, meaning they eat by pumping water through their gills and trapping food particles. With the capacity to filter fifty or more gallons of water in a single day, oysters quickly accumulate the toxic chemicals released by harmful algal blooms. Artificial intelligence technology could inform fisheries management of impending blooms, closing fisheries before any toxic shellfish are collected, thereby preventing fisherman from wasting resources while adding extra protection to consumers.

Red tide off of Florida, summer 2018.

The Florida Daily

The technology developed by Nelson can distinguish healthy algal cells from toxic cells, and labels the toxic cells accordingly. Using artificial intelligence is much more efficient than having scientists process samples manually. “In a sample, there’s thousands [of algal cells] and you’ve got to count them and detect the one or two toxic ones,” explains Brian Russell, director of the Nelson Artificial Intelligence Institute. “[It’s a] really challenging job [for] a trained biologist with a science degree and then one to two years training.”

A microscope image of toxic algal cysts from Alexandrium tamarense, which produces the neurotoxin saxitoxin.

David Wall, WHOI

The Nelson Artificial Intelligence Institute is not the first group to develop artificial intelligence to track algae, but their strategy is unique. In addition to automating the identification of algae, Russell’s team, in collaboration with the Cawthron Institute, have developed a system that takes a photo and records a ‘label’ every time a different algae is identified. “So now we’re growing this database of all this labelled data,” says Russel. “Basically, it’s becoming as expert as the scientist and we can scale it and go faster and faster and faster.”

The Cawthron Institute and Nelson Artificial Intelligence Institute recently received $6 million of funding for continued efforts studying and tracking algal blooms, and plan to apply their technology to help understand why whales beach in New Zealand’s Golden Bay.

Articles You May Like

Does Genetic Diversity Aid Starling Invasions Around The World?
Trump to block downloads of TikTok, WeChat on Sunday
A New Survey Identifies The 4 Best (And Worst) Coping Strategies To Deal With The Stress Of The Pandemic
Startup tackles decarbonizing industrial heat processes
What Happens If A Hurricane Named After A Greek Letter Needs To Be Retired?

Leave a Reply

Your email address will not be published. Required fields are marked *