AI-Powered Saildrones Monitor Alaska's Krill

Govind Tekale

Scientists deployed three robot boats to track krill in Alaska's Bering Sea when COVID-19 halted regular ocean research missions in 2020.

Photo Source: Office of Naval Research (CC BY 2.0)

These autonomous saildrones carried special sonar equipment to monitor tiny shrimp-like krill, which feed fish, seabirds, and marine mammals in the ocean ecosystem.

Photo Source: Animalia

Regular research ships used five different sonar frequencies to track krill, but the robot boats operated with just two frequencies - creating a technical challenge for scientists.

Photo Source: Filter Forge (CC BY 2.0)

NOAA scientists solved this limitation by teaching computers to recognize krill patterns using historical data from previous ship surveys since 2004.

Photo Source: Daniel/Dan Eidsmoe (CC BY 2.0)

The machine learning system, called random forest algorithm, successfully matched the accuracy of traditional four-frequency krill detection methods.

Photo Source: Stephanie Stutz (CC BY 4.0)

Recent studies revealed krill populations in the North Atlantic dropped by 50%, making this new tracking technology crucial for monitoring ocean health.

Photo Source: Øystein Paulsen (CC BY-SA 3.0)

This breakthrough keeps tabs on America's largest volume fishery - the Bering Sea's walleye pollock - which depends on krill for survival.

Photo Source: Bob@workboats-usa.com (CC BY-SA 4.0)

Scientists can now use robot boats and smart algorithms to continue vital ocean research even when traditional survey ships cannot sail.

Photo Source: Diego Delso (CC BY-SA 4.0)